diff --git a/contrib/icu-cmake/CMakeLists.txt b/contrib/icu-cmake/CMakeLists.txt
index adeaa7dcf33..5714fef8347 100644
--- a/contrib/icu-cmake/CMakeLists.txt
+++ b/contrib/icu-cmake/CMakeLists.txt
@@ -15,162 +15,64 @@ set(ICUDATA_SOURCE_DIR "${ClickHouse_SOURCE_DIR}/contrib/icudata/")
 # These lists of sources were generated from build log of the original ICU build system (configure + make).
 
 set(ICUUC_SOURCES
-"${ICU_SOURCE_DIR}/common/errorcode.cpp"
-"${ICU_SOURCE_DIR}/common/putil.cpp"
-"${ICU_SOURCE_DIR}/common/umath.cpp"
-"${ICU_SOURCE_DIR}/common/utypes.cpp"
-"${ICU_SOURCE_DIR}/common/uinvchar.cpp"
-"${ICU_SOURCE_DIR}/common/umutex.cpp"
-"${ICU_SOURCE_DIR}/common/ucln_cmn.cpp"
-"${ICU_SOURCE_DIR}/common/uinit.cpp"
-"${ICU_SOURCE_DIR}/common/uobject.cpp"
-"${ICU_SOURCE_DIR}/common/cmemory.cpp"
-"${ICU_SOURCE_DIR}/common/charstr.cpp"
-"${ICU_SOURCE_DIR}/common/cstr.cpp"
-"${ICU_SOURCE_DIR}/common/udata.cpp"
-"${ICU_SOURCE_DIR}/common/ucmndata.cpp"
-"${ICU_SOURCE_DIR}/common/udatamem.cpp"
-"${ICU_SOURCE_DIR}/common/umapfile.cpp"
-"${ICU_SOURCE_DIR}/common/udataswp.cpp"
-"${ICU_SOURCE_DIR}/common/utrie_swap.cpp"
-"${ICU_SOURCE_DIR}/common/ucol_swp.cpp"
-"${ICU_SOURCE_DIR}/common/utrace.cpp"
-"${ICU_SOURCE_DIR}/common/uhash.cpp"
-"${ICU_SOURCE_DIR}/common/uhash_us.cpp"
-"${ICU_SOURCE_DIR}/common/uenum.cpp"
-"${ICU_SOURCE_DIR}/common/ustrenum.cpp"
-"${ICU_SOURCE_DIR}/common/uvector.cpp"
-"${ICU_SOURCE_DIR}/common/ustack.cpp"
-"${ICU_SOURCE_DIR}/common/uvectr32.cpp"
-"${ICU_SOURCE_DIR}/common/uvectr64.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_bld.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_cnv.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_io.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_cb.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_err.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvlat1.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_u7.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_u8.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_u16.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_u32.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvscsu.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvbocu.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_ext.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvmbcs.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv2022.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvhz.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_lmb.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvisci.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvdisp.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_set.cpp"
-"${ICU_SOURCE_DIR}/common/ucnv_ct.cpp"
-"${ICU_SOURCE_DIR}/common/resource.cpp"
-"${ICU_SOURCE_DIR}/common/uresbund.cpp"
-"${ICU_SOURCE_DIR}/common/ures_cnv.cpp"
-"${ICU_SOURCE_DIR}/common/uresdata.cpp"
-"${ICU_SOURCE_DIR}/common/resbund.cpp"
-"${ICU_SOURCE_DIR}/common/resbund_cnv.cpp"
-"${ICU_SOURCE_DIR}/common/ucurr.cpp"
-"${ICU_SOURCE_DIR}/common/localebuilder.cpp"
-"${ICU_SOURCE_DIR}/common/localeprioritylist.cpp"
-"${ICU_SOURCE_DIR}/common/messagepattern.cpp"
-"${ICU_SOURCE_DIR}/common/ucat.cpp"
-"${ICU_SOURCE_DIR}/common/locmap.cpp"
-"${ICU_SOURCE_DIR}/common/uloc.cpp"
-"${ICU_SOURCE_DIR}/common/locid.cpp"
-"${ICU_SOURCE_DIR}/common/locutil.cpp"
-"${ICU_SOURCE_DIR}/common/locavailable.cpp"
-"${ICU_SOURCE_DIR}/common/locdispnames.cpp"
-"${ICU_SOURCE_DIR}/common/locdspnm.cpp"
-"${ICU_SOURCE_DIR}/common/loclikely.cpp"
-"${ICU_SOURCE_DIR}/common/locresdata.cpp"
-"${ICU_SOURCE_DIR}/common/lsr.cpp"
-"${ICU_SOURCE_DIR}/common/loclikelysubtags.cpp"
-"${ICU_SOURCE_DIR}/common/locdistance.cpp"
-"${ICU_SOURCE_DIR}/common/localematcher.cpp"
-"${ICU_SOURCE_DIR}/common/bytestream.cpp"
-"${ICU_SOURCE_DIR}/common/stringpiece.cpp"
-"${ICU_SOURCE_DIR}/common/bytesinkutil.cpp"
-"${ICU_SOURCE_DIR}/common/stringtriebuilder.cpp"
-"${ICU_SOURCE_DIR}/common/bytestriebuilder.cpp"
-"${ICU_SOURCE_DIR}/common/bytestrie.cpp"
-"${ICU_SOURCE_DIR}/common/bytestrieiterator.cpp"
-"${ICU_SOURCE_DIR}/common/ucharstrie.cpp"
-"${ICU_SOURCE_DIR}/common/ucharstriebuilder.cpp"
-"${ICU_SOURCE_DIR}/common/ucharstrieiterator.cpp"
-"${ICU_SOURCE_DIR}/common/dictionarydata.cpp"
-"${ICU_SOURCE_DIR}/common/edits.cpp"
 "${ICU_SOURCE_DIR}/common/appendable.cpp"
-"${ICU_SOURCE_DIR}/common/ustr_cnv.cpp"
-"${ICU_SOURCE_DIR}/common/unistr_cnv.cpp"
-"${ICU_SOURCE_DIR}/common/unistr.cpp"
-"${ICU_SOURCE_DIR}/common/unistr_case.cpp"
-"${ICU_SOURCE_DIR}/common/unistr_props.cpp"
-"${ICU_SOURCE_DIR}/common/utf_impl.cpp"
-"${ICU_SOURCE_DIR}/common/ustring.cpp"
-"${ICU_SOURCE_DIR}/common/ustrcase.cpp"
-"${ICU_SOURCE_DIR}/common/ucasemap.cpp"
-"${ICU_SOURCE_DIR}/common/ucasemap_titlecase_brkiter.cpp"
-"${ICU_SOURCE_DIR}/common/cstring.cpp"
-"${ICU_SOURCE_DIR}/common/ustrfmt.cpp"
-"${ICU_SOURCE_DIR}/common/ustrtrns.cpp"
-"${ICU_SOURCE_DIR}/common/ustr_wcs.cpp"
-"${ICU_SOURCE_DIR}/common/utext.cpp"
-"${ICU_SOURCE_DIR}/common/unistr_case_locale.cpp"
-"${ICU_SOURCE_DIR}/common/ustrcase_locale.cpp"
-"${ICU_SOURCE_DIR}/common/unistr_titlecase_brkiter.cpp"
-"${ICU_SOURCE_DIR}/common/ustr_titlecase_brkiter.cpp"
-"${ICU_SOURCE_DIR}/common/normalizer2impl.cpp"
-"${ICU_SOURCE_DIR}/common/normalizer2.cpp"
-"${ICU_SOURCE_DIR}/common/filterednormalizer2.cpp"
-"${ICU_SOURCE_DIR}/common/normlzr.cpp"
-"${ICU_SOURCE_DIR}/common/unorm.cpp"
-"${ICU_SOURCE_DIR}/common/unormcmp.cpp"
-"${ICU_SOURCE_DIR}/common/loadednormalizer2impl.cpp"
-"${ICU_SOURCE_DIR}/common/chariter.cpp"
-"${ICU_SOURCE_DIR}/common/schriter.cpp"
-"${ICU_SOURCE_DIR}/common/uchriter.cpp"
-"${ICU_SOURCE_DIR}/common/uiter.cpp"
-"${ICU_SOURCE_DIR}/common/patternprops.cpp"
-"${ICU_SOURCE_DIR}/common/uchar.cpp"
-"${ICU_SOURCE_DIR}/common/uprops.cpp"
-"${ICU_SOURCE_DIR}/common/ucase.cpp"
-"${ICU_SOURCE_DIR}/common/propname.cpp"
-"${ICU_SOURCE_DIR}/common/ubidi_props.cpp"
-"${ICU_SOURCE_DIR}/common/characterproperties.cpp"
-"${ICU_SOURCE_DIR}/common/ubidi.cpp"
-"${ICU_SOURCE_DIR}/common/ubidiwrt.cpp"
-"${ICU_SOURCE_DIR}/common/ubidiln.cpp"
-"${ICU_SOURCE_DIR}/common/ushape.cpp"
-"${ICU_SOURCE_DIR}/common/uscript.cpp"
-"${ICU_SOURCE_DIR}/common/uscript_props.cpp"
-"${ICU_SOURCE_DIR}/common/usc_impl.cpp"
-"${ICU_SOURCE_DIR}/common/unames.cpp"
-"${ICU_SOURCE_DIR}/common/utrie.cpp"
-"${ICU_SOURCE_DIR}/common/utrie2.cpp"
-"${ICU_SOURCE_DIR}/common/utrie2_builder.cpp"
-"${ICU_SOURCE_DIR}/common/ucptrie.cpp"
-"${ICU_SOURCE_DIR}/common/umutablecptrie.cpp"
 "${ICU_SOURCE_DIR}/common/bmpset.cpp"
-"${ICU_SOURCE_DIR}/common/unisetspan.cpp"
-"${ICU_SOURCE_DIR}/common/uset_props.cpp"
-"${ICU_SOURCE_DIR}/common/uniset_props.cpp"
-"${ICU_SOURCE_DIR}/common/uniset_closure.cpp"
-"${ICU_SOURCE_DIR}/common/uset.cpp"
-"${ICU_SOURCE_DIR}/common/uniset.cpp"
-"${ICU_SOURCE_DIR}/common/usetiter.cpp"
-"${ICU_SOURCE_DIR}/common/ruleiter.cpp"
-"${ICU_SOURCE_DIR}/common/caniter.cpp"
-"${ICU_SOURCE_DIR}/common/unifilt.cpp"
-"${ICU_SOURCE_DIR}/common/unifunct.cpp"
-"${ICU_SOURCE_DIR}/common/uarrsort.cpp"
-"${ICU_SOURCE_DIR}/common/brkiter.cpp"
-"${ICU_SOURCE_DIR}/common/ubrk.cpp"
 "${ICU_SOURCE_DIR}/common/brkeng.cpp"
+"${ICU_SOURCE_DIR}/common/brkiter.cpp"
+"${ICU_SOURCE_DIR}/common/bytesinkutil.cpp"
+"${ICU_SOURCE_DIR}/common/bytestream.cpp"
+"${ICU_SOURCE_DIR}/common/bytestrie.cpp"
+"${ICU_SOURCE_DIR}/common/bytestriebuilder.cpp"
+"${ICU_SOURCE_DIR}/common/bytestrieiterator.cpp"
+"${ICU_SOURCE_DIR}/common/caniter.cpp"
+"${ICU_SOURCE_DIR}/common/characterproperties.cpp"
+"${ICU_SOURCE_DIR}/common/chariter.cpp"
+"${ICU_SOURCE_DIR}/common/charstr.cpp"
+"${ICU_SOURCE_DIR}/common/cmemory.cpp"
+"${ICU_SOURCE_DIR}/common/cstr.cpp"
+"${ICU_SOURCE_DIR}/common/cstring.cpp"
+"${ICU_SOURCE_DIR}/common/cwchar.cpp"
 "${ICU_SOURCE_DIR}/common/dictbe.cpp"
+"${ICU_SOURCE_DIR}/common/dictionarydata.cpp"
+"${ICU_SOURCE_DIR}/common/dtintrv.cpp"
+"${ICU_SOURCE_DIR}/common/edits.cpp"
+"${ICU_SOURCE_DIR}/common/emojiprops.cpp"
+"${ICU_SOURCE_DIR}/common/errorcode.cpp"
 "${ICU_SOURCE_DIR}/common/filteredbrk.cpp"
+"${ICU_SOURCE_DIR}/common/filterednormalizer2.cpp"
+"${ICU_SOURCE_DIR}/common/icudataver.cpp"
+"${ICU_SOURCE_DIR}/common/icuplug.cpp"
+"${ICU_SOURCE_DIR}/common/loadednormalizer2impl.cpp"
+"${ICU_SOURCE_DIR}/common/localebuilder.cpp"
+"${ICU_SOURCE_DIR}/common/localematcher.cpp"
+"${ICU_SOURCE_DIR}/common/localeprioritylist.cpp"
+"${ICU_SOURCE_DIR}/common/locavailable.cpp"
+"${ICU_SOURCE_DIR}/common/locbased.cpp"
+"${ICU_SOURCE_DIR}/common/locdispnames.cpp"
+"${ICU_SOURCE_DIR}/common/locdistance.cpp"
+"${ICU_SOURCE_DIR}/common/locdspnm.cpp"
+"${ICU_SOURCE_DIR}/common/locid.cpp"
+"${ICU_SOURCE_DIR}/common/loclikely.cpp"
+"${ICU_SOURCE_DIR}/common/loclikelysubtags.cpp"
+"${ICU_SOURCE_DIR}/common/locmap.cpp"
+"${ICU_SOURCE_DIR}/common/locresdata.cpp"
+"${ICU_SOURCE_DIR}/common/locutil.cpp"
+"${ICU_SOURCE_DIR}/common/lsr.cpp"
+"${ICU_SOURCE_DIR}/common/lstmbe.cpp"
+"${ICU_SOURCE_DIR}/common/messagepattern.cpp"
+"${ICU_SOURCE_DIR}/common/mlbe.cpp"
+"${ICU_SOURCE_DIR}/common/normalizer2.cpp"
+"${ICU_SOURCE_DIR}/common/normalizer2impl.cpp"
+"${ICU_SOURCE_DIR}/common/normlzr.cpp"
+"${ICU_SOURCE_DIR}/common/parsepos.cpp"
+"${ICU_SOURCE_DIR}/common/patternprops.cpp"
+"${ICU_SOURCE_DIR}/common/pluralmap.cpp"
+"${ICU_SOURCE_DIR}/common/propname.cpp"
+"${ICU_SOURCE_DIR}/common/propsvec.cpp"
+"${ICU_SOURCE_DIR}/common/punycode.cpp"
+"${ICU_SOURCE_DIR}/common/putil.cpp"
 "${ICU_SOURCE_DIR}/common/rbbi.cpp"
+"${ICU_SOURCE_DIR}/common/rbbi_cache.cpp"
 "${ICU_SOURCE_DIR}/common/rbbidata.cpp"
 "${ICU_SOURCE_DIR}/common/rbbinode.cpp"
 "${ICU_SOURCE_DIR}/common/rbbirb.cpp"
@@ -178,166 +80,180 @@ set(ICUUC_SOURCES
 "${ICU_SOURCE_DIR}/common/rbbisetb.cpp"
 "${ICU_SOURCE_DIR}/common/rbbistbl.cpp"
 "${ICU_SOURCE_DIR}/common/rbbitblb.cpp"
-"${ICU_SOURCE_DIR}/common/rbbi_cache.cpp"
+"${ICU_SOURCE_DIR}/common/resbund.cpp"
+"${ICU_SOURCE_DIR}/common/resbund_cnv.cpp"
+"${ICU_SOURCE_DIR}/common/resource.cpp"
+"${ICU_SOURCE_DIR}/common/restrace.cpp"
+"${ICU_SOURCE_DIR}/common/ruleiter.cpp"
+"${ICU_SOURCE_DIR}/common/schriter.cpp"
 "${ICU_SOURCE_DIR}/common/serv.cpp"
-"${ICU_SOURCE_DIR}/common/servnotf.cpp"
-"${ICU_SOURCE_DIR}/common/servls.cpp"
 "${ICU_SOURCE_DIR}/common/servlk.cpp"
 "${ICU_SOURCE_DIR}/common/servlkf.cpp"
+"${ICU_SOURCE_DIR}/common/servls.cpp"
+"${ICU_SOURCE_DIR}/common/servnotf.cpp"
 "${ICU_SOURCE_DIR}/common/servrbf.cpp"
 "${ICU_SOURCE_DIR}/common/servslkf.cpp"
-"${ICU_SOURCE_DIR}/common/uidna.cpp"
-"${ICU_SOURCE_DIR}/common/usprep.cpp"
-"${ICU_SOURCE_DIR}/common/uts46.cpp"
-"${ICU_SOURCE_DIR}/common/punycode.cpp"
-"${ICU_SOURCE_DIR}/common/util.cpp"
-"${ICU_SOURCE_DIR}/common/util_props.cpp"
-"${ICU_SOURCE_DIR}/common/parsepos.cpp"
-"${ICU_SOURCE_DIR}/common/locbased.cpp"
-"${ICU_SOURCE_DIR}/common/cwchar.cpp"
-"${ICU_SOURCE_DIR}/common/wintz.cpp"
-"${ICU_SOURCE_DIR}/common/dtintrv.cpp"
-"${ICU_SOURCE_DIR}/common/ucnvsel.cpp"
-"${ICU_SOURCE_DIR}/common/propsvec.cpp"
-"${ICU_SOURCE_DIR}/common/ulist.cpp"
-"${ICU_SOURCE_DIR}/common/uloc_tag.cpp"
-"${ICU_SOURCE_DIR}/common/icudataver.cpp"
-"${ICU_SOURCE_DIR}/common/icuplug.cpp"
 "${ICU_SOURCE_DIR}/common/sharedobject.cpp"
 "${ICU_SOURCE_DIR}/common/simpleformatter.cpp"
-"${ICU_SOURCE_DIR}/common/unifiedcache.cpp"
-"${ICU_SOURCE_DIR}/common/uloc_keytype.cpp"
-"${ICU_SOURCE_DIR}/common/ubiditransform.cpp"
-"${ICU_SOURCE_DIR}/common/pluralmap.cpp"
 "${ICU_SOURCE_DIR}/common/static_unicode_sets.cpp"
-"${ICU_SOURCE_DIR}/common/restrace.cpp"
-"${ICU_SOURCE_DIR}/common/emojiprops.cpp"
-"${ICU_SOURCE_DIR}/common/lstmbe.cpp")
+"${ICU_SOURCE_DIR}/common/stringpiece.cpp"
+"${ICU_SOURCE_DIR}/common/stringtriebuilder.cpp"
+"${ICU_SOURCE_DIR}/common/uarrsort.cpp"
+"${ICU_SOURCE_DIR}/common/ubidi.cpp"
+"${ICU_SOURCE_DIR}/common/ubidi_props.cpp"
+"${ICU_SOURCE_DIR}/common/ubidiln.cpp"
+"${ICU_SOURCE_DIR}/common/ubiditransform.cpp"
+"${ICU_SOURCE_DIR}/common/ubidiwrt.cpp"
+"${ICU_SOURCE_DIR}/common/ubrk.cpp"
+"${ICU_SOURCE_DIR}/common/ucase.cpp"
+"${ICU_SOURCE_DIR}/common/ucasemap.cpp"
+"${ICU_SOURCE_DIR}/common/ucasemap_titlecase_brkiter.cpp"
+"${ICU_SOURCE_DIR}/common/ucat.cpp"
+"${ICU_SOURCE_DIR}/common/uchar.cpp"
+"${ICU_SOURCE_DIR}/common/ucharstrie.cpp"
+"${ICU_SOURCE_DIR}/common/ucharstriebuilder.cpp"
+"${ICU_SOURCE_DIR}/common/ucharstrieiterator.cpp"
+"${ICU_SOURCE_DIR}/common/uchriter.cpp"
+"${ICU_SOURCE_DIR}/common/ucln_cmn.cpp"
+"${ICU_SOURCE_DIR}/common/ucmndata.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv2022.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_bld.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_cb.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_cnv.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_ct.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_err.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_ext.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_io.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_lmb.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_set.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_u16.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_u32.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_u7.cpp"
+"${ICU_SOURCE_DIR}/common/ucnv_u8.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvbocu.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvdisp.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvhz.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvisci.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvlat1.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvmbcs.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvscsu.cpp"
+"${ICU_SOURCE_DIR}/common/ucnvsel.cpp"
+"${ICU_SOURCE_DIR}/common/ucol_swp.cpp"
+"${ICU_SOURCE_DIR}/common/ucptrie.cpp"
+"${ICU_SOURCE_DIR}/common/ucurr.cpp"
+"${ICU_SOURCE_DIR}/common/udata.cpp"
+"${ICU_SOURCE_DIR}/common/udatamem.cpp"
+"${ICU_SOURCE_DIR}/common/udataswp.cpp"
+"${ICU_SOURCE_DIR}/common/uenum.cpp"
+"${ICU_SOURCE_DIR}/common/uhash.cpp"
+"${ICU_SOURCE_DIR}/common/uhash_us.cpp"
+"${ICU_SOURCE_DIR}/common/uidna.cpp"
+"${ICU_SOURCE_DIR}/common/uinit.cpp"
+"${ICU_SOURCE_DIR}/common/uinvchar.cpp"
+"${ICU_SOURCE_DIR}/common/uiter.cpp"
+"${ICU_SOURCE_DIR}/common/ulist.cpp"
+"${ICU_SOURCE_DIR}/common/uloc.cpp"
+"${ICU_SOURCE_DIR}/common/uloc_keytype.cpp"
+"${ICU_SOURCE_DIR}/common/uloc_tag.cpp"
+"${ICU_SOURCE_DIR}/common/ulocale.cpp"
+"${ICU_SOURCE_DIR}/common/ulocbuilder.cpp"
+"${ICU_SOURCE_DIR}/common/umapfile.cpp"
+"${ICU_SOURCE_DIR}/common/umath.cpp"
+"${ICU_SOURCE_DIR}/common/umutablecptrie.cpp"
+"${ICU_SOURCE_DIR}/common/umutex.cpp"
+"${ICU_SOURCE_DIR}/common/unames.cpp"
+"${ICU_SOURCE_DIR}/common/unifiedcache.cpp"
+"${ICU_SOURCE_DIR}/common/unifilt.cpp"
+"${ICU_SOURCE_DIR}/common/unifunct.cpp"
+"${ICU_SOURCE_DIR}/common/uniset.cpp"
+"${ICU_SOURCE_DIR}/common/uniset_closure.cpp"
+"${ICU_SOURCE_DIR}/common/uniset_props.cpp"
+"${ICU_SOURCE_DIR}/common/unisetspan.cpp"
+"${ICU_SOURCE_DIR}/common/unistr.cpp"
+"${ICU_SOURCE_DIR}/common/unistr_case.cpp"
+"${ICU_SOURCE_DIR}/common/unistr_case_locale.cpp"
+"${ICU_SOURCE_DIR}/common/unistr_cnv.cpp"
+"${ICU_SOURCE_DIR}/common/unistr_props.cpp"
+"${ICU_SOURCE_DIR}/common/unistr_titlecase_brkiter.cpp"
+"${ICU_SOURCE_DIR}/common/unorm.cpp"
+"${ICU_SOURCE_DIR}/common/unormcmp.cpp"
+"${ICU_SOURCE_DIR}/common/uobject.cpp"
+"${ICU_SOURCE_DIR}/common/uprops.cpp"
+"${ICU_SOURCE_DIR}/common/ures_cnv.cpp"
+"${ICU_SOURCE_DIR}/common/uresbund.cpp"
+"${ICU_SOURCE_DIR}/common/uresdata.cpp"
+"${ICU_SOURCE_DIR}/common/usc_impl.cpp"
+"${ICU_SOURCE_DIR}/common/uscript.cpp"
+"${ICU_SOURCE_DIR}/common/uscript_props.cpp"
+"${ICU_SOURCE_DIR}/common/uset.cpp"
+"${ICU_SOURCE_DIR}/common/uset_props.cpp"
+"${ICU_SOURCE_DIR}/common/usetiter.cpp"
+"${ICU_SOURCE_DIR}/common/ushape.cpp"
+"${ICU_SOURCE_DIR}/common/usprep.cpp"
+"${ICU_SOURCE_DIR}/common/ustack.cpp"
+"${ICU_SOURCE_DIR}/common/ustr_cnv.cpp"
+"${ICU_SOURCE_DIR}/common/ustr_titlecase_brkiter.cpp"
+"${ICU_SOURCE_DIR}/common/ustr_wcs.cpp"
+"${ICU_SOURCE_DIR}/common/ustrcase.cpp"
+"${ICU_SOURCE_DIR}/common/ustrcase_locale.cpp"
+"${ICU_SOURCE_DIR}/common/ustrenum.cpp"
+"${ICU_SOURCE_DIR}/common/ustrfmt.cpp"
+"${ICU_SOURCE_DIR}/common/ustring.cpp"
+"${ICU_SOURCE_DIR}/common/ustrtrns.cpp"
+"${ICU_SOURCE_DIR}/common/utext.cpp"
+"${ICU_SOURCE_DIR}/common/utf_impl.cpp"
+"${ICU_SOURCE_DIR}/common/util.cpp"
+"${ICU_SOURCE_DIR}/common/util_props.cpp"
+"${ICU_SOURCE_DIR}/common/utrace.cpp"
+"${ICU_SOURCE_DIR}/common/utrie.cpp"
+"${ICU_SOURCE_DIR}/common/utrie2.cpp"
+"${ICU_SOURCE_DIR}/common/utrie2_builder.cpp"
+"${ICU_SOURCE_DIR}/common/utrie_swap.cpp"
+"${ICU_SOURCE_DIR}/common/uts46.cpp"
+"${ICU_SOURCE_DIR}/common/utypes.cpp"
+"${ICU_SOURCE_DIR}/common/uvector.cpp"
+"${ICU_SOURCE_DIR}/common/uvectr32.cpp"
+"${ICU_SOURCE_DIR}/common/uvectr64.cpp"
+"${ICU_SOURCE_DIR}/common/wintz.cpp")
 
 set(ICUI18N_SOURCES
-"${ICU_SOURCE_DIR}/i18n/ucln_in.cpp"
-"${ICU_SOURCE_DIR}/i18n/fmtable.cpp"
-"${ICU_SOURCE_DIR}/i18n/format.cpp"
-"${ICU_SOURCE_DIR}/i18n/msgfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/umsg.cpp"
-"${ICU_SOURCE_DIR}/i18n/numfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/unum.cpp"
-"${ICU_SOURCE_DIR}/i18n/decimfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/dcfmtsym.cpp"
-"${ICU_SOURCE_DIR}/i18n/fmtable_cnv.cpp"
-"${ICU_SOURCE_DIR}/i18n/choicfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/datefmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/smpdtfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/reldtfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/dtfmtsym.cpp"
-"${ICU_SOURCE_DIR}/i18n/udat.cpp"
-"${ICU_SOURCE_DIR}/i18n/dtptngen.cpp"
-"${ICU_SOURCE_DIR}/i18n/udatpg.cpp"
-"${ICU_SOURCE_DIR}/i18n/nfrs.cpp"
-"${ICU_SOURCE_DIR}/i18n/nfrule.cpp"
-"${ICU_SOURCE_DIR}/i18n/nfsubs.cpp"
-"${ICU_SOURCE_DIR}/i18n/rbnf.cpp"
-"${ICU_SOURCE_DIR}/i18n/numsys.cpp"
-"${ICU_SOURCE_DIR}/i18n/unumsys.cpp"
-"${ICU_SOURCE_DIR}/i18n/ucsdet.cpp"
-"${ICU_SOURCE_DIR}/i18n/ucal.cpp"
-"${ICU_SOURCE_DIR}/i18n/calendar.cpp"
-"${ICU_SOURCE_DIR}/i18n/gregocal.cpp"
-"${ICU_SOURCE_DIR}/i18n/timezone.cpp"
-"${ICU_SOURCE_DIR}/i18n/simpletz.cpp"
-"${ICU_SOURCE_DIR}/i18n/olsontz.cpp"
+"${ICU_SOURCE_DIR}/i18n/alphaindex.cpp"
+"${ICU_SOURCE_DIR}/i18n/anytrans.cpp"
 "${ICU_SOURCE_DIR}/i18n/astro.cpp"
-"${ICU_SOURCE_DIR}/i18n/taiwncal.cpp"
+"${ICU_SOURCE_DIR}/i18n/basictz.cpp"
+"${ICU_SOURCE_DIR}/i18n/bocsu.cpp"
+"${ICU_SOURCE_DIR}/i18n/brktrans.cpp"
 "${ICU_SOURCE_DIR}/i18n/buddhcal.cpp"
-"${ICU_SOURCE_DIR}/i18n/persncal.cpp"
-"${ICU_SOURCE_DIR}/i18n/islamcal.cpp"
-"${ICU_SOURCE_DIR}/i18n/japancal.cpp"
-"${ICU_SOURCE_DIR}/i18n/gregoimp.cpp"
-"${ICU_SOURCE_DIR}/i18n/hebrwcal.cpp"
-"${ICU_SOURCE_DIR}/i18n/indiancal.cpp"
-"${ICU_SOURCE_DIR}/i18n/chnsecal.cpp"
+"${ICU_SOURCE_DIR}/i18n/calendar.cpp"
+"${ICU_SOURCE_DIR}/i18n/casetrn.cpp"
 "${ICU_SOURCE_DIR}/i18n/cecal.cpp"
-"${ICU_SOURCE_DIR}/i18n/coptccal.cpp"
-"${ICU_SOURCE_DIR}/i18n/dangical.cpp"
-"${ICU_SOURCE_DIR}/i18n/ethpccal.cpp"
+"${ICU_SOURCE_DIR}/i18n/chnsecal.cpp"
+"${ICU_SOURCE_DIR}/i18n/choicfmt.cpp"
 "${ICU_SOURCE_DIR}/i18n/coleitr.cpp"
 "${ICU_SOURCE_DIR}/i18n/coll.cpp"
-"${ICU_SOURCE_DIR}/i18n/sortkey.cpp"
-"${ICU_SOURCE_DIR}/i18n/bocsu.cpp"
-"${ICU_SOURCE_DIR}/i18n/ucoleitr.cpp"
-"${ICU_SOURCE_DIR}/i18n/ucol.cpp"
-"${ICU_SOURCE_DIR}/i18n/ucol_res.cpp"
-"${ICU_SOURCE_DIR}/i18n/ucol_sit.cpp"
 "${ICU_SOURCE_DIR}/i18n/collation.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationsettings.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationbuilder.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationcompare.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationdata.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationtailoring.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationdatabuilder.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationdatareader.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationdatawriter.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationfastlatin.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationfastlatinbuilder.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationfcd.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationiterator.cpp"
-"${ICU_SOURCE_DIR}/i18n/utf16collationiterator.cpp"
-"${ICU_SOURCE_DIR}/i18n/utf8collationiterator.cpp"
-"${ICU_SOURCE_DIR}/i18n/uitercollationiterator.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationsets.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationcompare.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationfastlatin.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationkeys.cpp"
-"${ICU_SOURCE_DIR}/i18n/rulebasedcollator.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationroot.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationrootelements.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationdatabuilder.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationweights.cpp"
 "${ICU_SOURCE_DIR}/i18n/collationruleparser.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationbuilder.cpp"
-"${ICU_SOURCE_DIR}/i18n/collationfastlatinbuilder.cpp"
-"${ICU_SOURCE_DIR}/i18n/listformatter.cpp"
-"${ICU_SOURCE_DIR}/i18n/ulistformatter.cpp"
-"${ICU_SOURCE_DIR}/i18n/strmatch.cpp"
-"${ICU_SOURCE_DIR}/i18n/usearch.cpp"
-"${ICU_SOURCE_DIR}/i18n/search.cpp"
-"${ICU_SOURCE_DIR}/i18n/stsearch.cpp"
-"${ICU_SOURCE_DIR}/i18n/translit.cpp"
-"${ICU_SOURCE_DIR}/i18n/utrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/esctrn.cpp"
-"${ICU_SOURCE_DIR}/i18n/unesctrn.cpp"
-"${ICU_SOURCE_DIR}/i18n/funcrepl.cpp"
-"${ICU_SOURCE_DIR}/i18n/strrepl.cpp"
-"${ICU_SOURCE_DIR}/i18n/tridpars.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationsets.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationsettings.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationtailoring.cpp"
+"${ICU_SOURCE_DIR}/i18n/collationweights.cpp"
+"${ICU_SOURCE_DIR}/i18n/compactdecimalformat.cpp"
+"${ICU_SOURCE_DIR}/i18n/coptccal.cpp"
 "${ICU_SOURCE_DIR}/i18n/cpdtrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/rbt.cpp"
-"${ICU_SOURCE_DIR}/i18n/rbt_data.cpp"
-"${ICU_SOURCE_DIR}/i18n/rbt_pars.cpp"
-"${ICU_SOURCE_DIR}/i18n/rbt_rule.cpp"
-"${ICU_SOURCE_DIR}/i18n/rbt_set.cpp"
-"${ICU_SOURCE_DIR}/i18n/nultrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/remtrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/casetrn.cpp"
-"${ICU_SOURCE_DIR}/i18n/titletrn.cpp"
-"${ICU_SOURCE_DIR}/i18n/tolowtrn.cpp"
-"${ICU_SOURCE_DIR}/i18n/toupptrn.cpp"
-"${ICU_SOURCE_DIR}/i18n/anytrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/name2uni.cpp"
-"${ICU_SOURCE_DIR}/i18n/uni2name.cpp"
-"${ICU_SOURCE_DIR}/i18n/nortrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/quant.cpp"
-"${ICU_SOURCE_DIR}/i18n/transreg.cpp"
-"${ICU_SOURCE_DIR}/i18n/brktrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/regexcmp.cpp"
-"${ICU_SOURCE_DIR}/i18n/rematch.cpp"
-"${ICU_SOURCE_DIR}/i18n/repattrn.cpp"
-"${ICU_SOURCE_DIR}/i18n/regexst.cpp"
-"${ICU_SOURCE_DIR}/i18n/regextxt.cpp"
-"${ICU_SOURCE_DIR}/i18n/regeximp.cpp"
-"${ICU_SOURCE_DIR}/i18n/uregex.cpp"
-"${ICU_SOURCE_DIR}/i18n/uregexc.cpp"
-"${ICU_SOURCE_DIR}/i18n/ulocdata.cpp"
-"${ICU_SOURCE_DIR}/i18n/measfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/currfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/curramt.cpp"
-"${ICU_SOURCE_DIR}/i18n/currunit.cpp"
-"${ICU_SOURCE_DIR}/i18n/measure.cpp"
-"${ICU_SOURCE_DIR}/i18n/utmscale.cpp"
 "${ICU_SOURCE_DIR}/i18n/csdetect.cpp"
 "${ICU_SOURCE_DIR}/i18n/csmatch.cpp"
 "${ICU_SOURCE_DIR}/i18n/csr2022.cpp"
@@ -346,60 +262,80 @@ set(ICUI18N_SOURCES
 "${ICU_SOURCE_DIR}/i18n/csrsbcs.cpp"
 "${ICU_SOURCE_DIR}/i18n/csrucode.cpp"
 "${ICU_SOURCE_DIR}/i18n/csrutf8.cpp"
-"${ICU_SOURCE_DIR}/i18n/inputext.cpp"
-"${ICU_SOURCE_DIR}/i18n/wintzimpl.cpp"
-"${ICU_SOURCE_DIR}/i18n/windtfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/winnmfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/basictz.cpp"
-"${ICU_SOURCE_DIR}/i18n/dtrule.cpp"
-"${ICU_SOURCE_DIR}/i18n/rbtz.cpp"
-"${ICU_SOURCE_DIR}/i18n/tzrule.cpp"
-"${ICU_SOURCE_DIR}/i18n/tztrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/vtzone.cpp"
-"${ICU_SOURCE_DIR}/i18n/zonemeta.cpp"
-"${ICU_SOURCE_DIR}/i18n/standardplural.cpp"
-"${ICU_SOURCE_DIR}/i18n/upluralrules.cpp"
-"${ICU_SOURCE_DIR}/i18n/plurrule.cpp"
-"${ICU_SOURCE_DIR}/i18n/plurfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/selfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/curramt.cpp"
+"${ICU_SOURCE_DIR}/i18n/currfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/currpinf.cpp"
+"${ICU_SOURCE_DIR}/i18n/currunit.cpp"
+"${ICU_SOURCE_DIR}/i18n/dangical.cpp"
+"${ICU_SOURCE_DIR}/i18n/datefmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/dayperiodrules.cpp"
+"${ICU_SOURCE_DIR}/i18n/dcfmtsym.cpp"
+"${ICU_SOURCE_DIR}/i18n/decContext.cpp"
+"${ICU_SOURCE_DIR}/i18n/decNumber.cpp"
+"${ICU_SOURCE_DIR}/i18n/decimfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/displayoptions.cpp"
+"${ICU_SOURCE_DIR}/i18n/double-conversion-bignum-dtoa.cpp"
+"${ICU_SOURCE_DIR}/i18n/double-conversion-bignum.cpp"
+"${ICU_SOURCE_DIR}/i18n/double-conversion-cached-powers.cpp"
+"${ICU_SOURCE_DIR}/i18n/double-conversion-double-to-string.cpp"
+"${ICU_SOURCE_DIR}/i18n/double-conversion-fast-dtoa.cpp"
+"${ICU_SOURCE_DIR}/i18n/double-conversion-string-to-double.cpp"
+"${ICU_SOURCE_DIR}/i18n/double-conversion-strtod.cpp"
+"${ICU_SOURCE_DIR}/i18n/dtfmtsym.cpp"
 "${ICU_SOURCE_DIR}/i18n/dtitvfmt.cpp"
 "${ICU_SOURCE_DIR}/i18n/dtitvinf.cpp"
-"${ICU_SOURCE_DIR}/i18n/udateintervalformat.cpp"
-"${ICU_SOURCE_DIR}/i18n/tmunit.cpp"
-"${ICU_SOURCE_DIR}/i18n/tmutamt.cpp"
-"${ICU_SOURCE_DIR}/i18n/tmutfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/currpinf.cpp"
-"${ICU_SOURCE_DIR}/i18n/uspoof.cpp"
-"${ICU_SOURCE_DIR}/i18n/uspoof_impl.cpp"
-"${ICU_SOURCE_DIR}/i18n/uspoof_build.cpp"
-"${ICU_SOURCE_DIR}/i18n/uspoof_conf.cpp"
-"${ICU_SOURCE_DIR}/i18n/smpdtfst.cpp"
-"${ICU_SOURCE_DIR}/i18n/ztrans.cpp"
-"${ICU_SOURCE_DIR}/i18n/zrule.cpp"
-"${ICU_SOURCE_DIR}/i18n/vzone.cpp"
+"${ICU_SOURCE_DIR}/i18n/dtptngen.cpp"
+"${ICU_SOURCE_DIR}/i18n/dtrule.cpp"
+"${ICU_SOURCE_DIR}/i18n/erarules.cpp"
+"${ICU_SOURCE_DIR}/i18n/esctrn.cpp"
+"${ICU_SOURCE_DIR}/i18n/ethpccal.cpp"
+"${ICU_SOURCE_DIR}/i18n/fmtable.cpp"
+"${ICU_SOURCE_DIR}/i18n/fmtable_cnv.cpp"
+"${ICU_SOURCE_DIR}/i18n/format.cpp"
+"${ICU_SOURCE_DIR}/i18n/formatted_string_builder.cpp"
+"${ICU_SOURCE_DIR}/i18n/formattedval_iterimpl.cpp"
+"${ICU_SOURCE_DIR}/i18n/formattedval_sbimpl.cpp"
+"${ICU_SOURCE_DIR}/i18n/formattedvalue.cpp"
 "${ICU_SOURCE_DIR}/i18n/fphdlimp.cpp"
 "${ICU_SOURCE_DIR}/i18n/fpositer.cpp"
-"${ICU_SOURCE_DIR}/i18n/ufieldpositer.cpp"
-"${ICU_SOURCE_DIR}/i18n/decNumber.cpp"
-"${ICU_SOURCE_DIR}/i18n/decContext.cpp"
-"${ICU_SOURCE_DIR}/i18n/alphaindex.cpp"
-"${ICU_SOURCE_DIR}/i18n/tznames.cpp"
-"${ICU_SOURCE_DIR}/i18n/tznames_impl.cpp"
-"${ICU_SOURCE_DIR}/i18n/tzgnames.cpp"
-"${ICU_SOURCE_DIR}/i18n/tzfmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/compactdecimalformat.cpp"
+"${ICU_SOURCE_DIR}/i18n/funcrepl.cpp"
 "${ICU_SOURCE_DIR}/i18n/gender.cpp"
-"${ICU_SOURCE_DIR}/i18n/region.cpp"
-"${ICU_SOURCE_DIR}/i18n/scriptset.cpp"
-"${ICU_SOURCE_DIR}/i18n/uregion.cpp"
-"${ICU_SOURCE_DIR}/i18n/reldatefmt.cpp"
-"${ICU_SOURCE_DIR}/i18n/quantityformatter.cpp"
+"${ICU_SOURCE_DIR}/i18n/gregocal.cpp"
+"${ICU_SOURCE_DIR}/i18n/gregoimp.cpp"
+"${ICU_SOURCE_DIR}/i18n/hebrwcal.cpp"
+"${ICU_SOURCE_DIR}/i18n/indiancal.cpp"
+"${ICU_SOURCE_DIR}/i18n/inputext.cpp"
+"${ICU_SOURCE_DIR}/i18n/islamcal.cpp"
+"${ICU_SOURCE_DIR}/i18n/iso8601cal.cpp"
+"${ICU_SOURCE_DIR}/i18n/japancal.cpp"
+"${ICU_SOURCE_DIR}/i18n/listformatter.cpp"
+"${ICU_SOURCE_DIR}/i18n/measfmt.cpp"
 "${ICU_SOURCE_DIR}/i18n/measunit.cpp"
-"${ICU_SOURCE_DIR}/i18n/sharedbreakiterator.cpp"
-"${ICU_SOURCE_DIR}/i18n/scientificnumberformatter.cpp"
-"${ICU_SOURCE_DIR}/i18n/dayperiodrules.cpp"
+"${ICU_SOURCE_DIR}/i18n/measunit_extra.cpp"
+"${ICU_SOURCE_DIR}/i18n/measure.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_arguments.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_checker.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_data_model.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_errors.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_evaluation.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_formattable.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_formatter.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_function_registry.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_parser.cpp"
+"${ICU_SOURCE_DIR}/i18n/messageformat2_serializer.cpp"
+"${ICU_SOURCE_DIR}/i18n/msgfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/name2uni.cpp"
+"${ICU_SOURCE_DIR}/i18n/nfrs.cpp"
+"${ICU_SOURCE_DIR}/i18n/nfrule.cpp"
+"${ICU_SOURCE_DIR}/i18n/nfsubs.cpp"
+"${ICU_SOURCE_DIR}/i18n/nortrans.cpp"
+"${ICU_SOURCE_DIR}/i18n/nultrans.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_affixutils.cpp"
+"${ICU_SOURCE_DIR}/i18n/number_asformat.cpp"
+"${ICU_SOURCE_DIR}/i18n/number_capi.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_compact.cpp"
+"${ICU_SOURCE_DIR}/i18n/number_currencysymbols.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_decimalquantity.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_decimfmtprops.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_fluent.cpp"
@@ -407,7 +343,9 @@ set(ICUI18N_SOURCES
 "${ICU_SOURCE_DIR}/i18n/number_grouping.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_integerwidth.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_longnames.cpp"
+"${ICU_SOURCE_DIR}/i18n/number_mapper.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_modifiers.cpp"
+"${ICU_SOURCE_DIR}/i18n/number_multiplier.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_notation.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_output.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_padding.cpp"
@@ -415,46 +353,125 @@ set(ICUI18N_SOURCES
 "${ICU_SOURCE_DIR}/i18n/number_patternstring.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_rounding.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_scientific.cpp"
-"${ICU_SOURCE_DIR}/i18n/number_utils.cpp"
-"${ICU_SOURCE_DIR}/i18n/number_asformat.cpp"
-"${ICU_SOURCE_DIR}/i18n/number_mapper.cpp"
-"${ICU_SOURCE_DIR}/i18n/number_multiplier.cpp"
-"${ICU_SOURCE_DIR}/i18n/number_currencysymbols.cpp"
+"${ICU_SOURCE_DIR}/i18n/number_simple.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_skeletons.cpp"
-"${ICU_SOURCE_DIR}/i18n/number_capi.cpp"
-"${ICU_SOURCE_DIR}/i18n/double-conversion-string-to-double.cpp"
-"${ICU_SOURCE_DIR}/i18n/double-conversion-double-to-string.cpp"
-"${ICU_SOURCE_DIR}/i18n/double-conversion-bignum-dtoa.cpp"
-"${ICU_SOURCE_DIR}/i18n/double-conversion-bignum.cpp"
-"${ICU_SOURCE_DIR}/i18n/double-conversion-cached-powers.cpp"
-"${ICU_SOURCE_DIR}/i18n/double-conversion-fast-dtoa.cpp"
-"${ICU_SOURCE_DIR}/i18n/double-conversion-strtod.cpp"
-"${ICU_SOURCE_DIR}/i18n/string_segment.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_parsednumber.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_impl.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_symbols.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_decimal.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_scientific.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_currency.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_affixes.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_compositions.cpp"
-"${ICU_SOURCE_DIR}/i18n/numparse_validators.cpp"
-"${ICU_SOURCE_DIR}/i18n/numrange_fluent.cpp"
-"${ICU_SOURCE_DIR}/i18n/numrange_impl.cpp"
-"${ICU_SOURCE_DIR}/i18n/erarules.cpp"
-"${ICU_SOURCE_DIR}/i18n/formattedvalue.cpp"
-"${ICU_SOURCE_DIR}/i18n/formattedval_iterimpl.cpp"
-"${ICU_SOURCE_DIR}/i18n/formattedval_sbimpl.cpp"
-"${ICU_SOURCE_DIR}/i18n/formatted_string_builder.cpp"
-"${ICU_SOURCE_DIR}/i18n/measunit_extra.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_symbolswrapper.cpp"
 "${ICU_SOURCE_DIR}/i18n/number_usageprefs.cpp"
+"${ICU_SOURCE_DIR}/i18n/number_utils.cpp"
+"${ICU_SOURCE_DIR}/i18n/numfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_affixes.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_compositions.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_currency.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_decimal.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_impl.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_parsednumber.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_scientific.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_symbols.cpp"
+"${ICU_SOURCE_DIR}/i18n/numparse_validators.cpp"
 "${ICU_SOURCE_DIR}/i18n/numrange_capi.cpp"
+"${ICU_SOURCE_DIR}/i18n/numrange_fluent.cpp"
+"${ICU_SOURCE_DIR}/i18n/numrange_impl.cpp"
+"${ICU_SOURCE_DIR}/i18n/numsys.cpp"
+"${ICU_SOURCE_DIR}/i18n/olsontz.cpp"
+"${ICU_SOURCE_DIR}/i18n/persncal.cpp"
 "${ICU_SOURCE_DIR}/i18n/pluralranges.cpp"
+"${ICU_SOURCE_DIR}/i18n/plurfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/plurrule.cpp"
+"${ICU_SOURCE_DIR}/i18n/quant.cpp"
+"${ICU_SOURCE_DIR}/i18n/quantityformatter.cpp"
+"${ICU_SOURCE_DIR}/i18n/rbnf.cpp"
+"${ICU_SOURCE_DIR}/i18n/rbt.cpp"
+"${ICU_SOURCE_DIR}/i18n/rbt_data.cpp"
+"${ICU_SOURCE_DIR}/i18n/rbt_pars.cpp"
+"${ICU_SOURCE_DIR}/i18n/rbt_rule.cpp"
+"${ICU_SOURCE_DIR}/i18n/rbt_set.cpp"
+"${ICU_SOURCE_DIR}/i18n/rbtz.cpp"
+"${ICU_SOURCE_DIR}/i18n/regexcmp.cpp"
+"${ICU_SOURCE_DIR}/i18n/regeximp.cpp"
+"${ICU_SOURCE_DIR}/i18n/regexst.cpp"
+"${ICU_SOURCE_DIR}/i18n/regextxt.cpp"
+"${ICU_SOURCE_DIR}/i18n/region.cpp"
+"${ICU_SOURCE_DIR}/i18n/reldatefmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/reldtfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/rematch.cpp"
+"${ICU_SOURCE_DIR}/i18n/remtrans.cpp"
+"${ICU_SOURCE_DIR}/i18n/repattrn.cpp"
+"${ICU_SOURCE_DIR}/i18n/rulebasedcollator.cpp"
+"${ICU_SOURCE_DIR}/i18n/scientificnumberformatter.cpp"
+"${ICU_SOURCE_DIR}/i18n/scriptset.cpp"
+"${ICU_SOURCE_DIR}/i18n/search.cpp"
+"${ICU_SOURCE_DIR}/i18n/selfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/sharedbreakiterator.cpp"
+"${ICU_SOURCE_DIR}/i18n/simpletz.cpp"
+"${ICU_SOURCE_DIR}/i18n/smpdtfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/smpdtfst.cpp"
+"${ICU_SOURCE_DIR}/i18n/sortkey.cpp"
+"${ICU_SOURCE_DIR}/i18n/standardplural.cpp"
+"${ICU_SOURCE_DIR}/i18n/string_segment.cpp"
+"${ICU_SOURCE_DIR}/i18n/strmatch.cpp"
+"${ICU_SOURCE_DIR}/i18n/strrepl.cpp"
+"${ICU_SOURCE_DIR}/i18n/stsearch.cpp"
+"${ICU_SOURCE_DIR}/i18n/taiwncal.cpp"
+"${ICU_SOURCE_DIR}/i18n/timezone.cpp"
+"${ICU_SOURCE_DIR}/i18n/titletrn.cpp"
+"${ICU_SOURCE_DIR}/i18n/tmunit.cpp"
+"${ICU_SOURCE_DIR}/i18n/tmutamt.cpp"
+"${ICU_SOURCE_DIR}/i18n/tmutfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/tolowtrn.cpp"
+"${ICU_SOURCE_DIR}/i18n/toupptrn.cpp"
+"${ICU_SOURCE_DIR}/i18n/translit.cpp"
+"${ICU_SOURCE_DIR}/i18n/transreg.cpp"
+"${ICU_SOURCE_DIR}/i18n/tridpars.cpp"
+"${ICU_SOURCE_DIR}/i18n/tzfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/tzgnames.cpp"
+"${ICU_SOURCE_DIR}/i18n/tznames.cpp"
+"${ICU_SOURCE_DIR}/i18n/tznames_impl.cpp"
+"${ICU_SOURCE_DIR}/i18n/tzrule.cpp"
+"${ICU_SOURCE_DIR}/i18n/tztrans.cpp"
+"${ICU_SOURCE_DIR}/i18n/ucal.cpp"
+"${ICU_SOURCE_DIR}/i18n/ucln_in.cpp"
+"${ICU_SOURCE_DIR}/i18n/ucol.cpp"
+"${ICU_SOURCE_DIR}/i18n/ucol_res.cpp"
+"${ICU_SOURCE_DIR}/i18n/ucol_sit.cpp"
+"${ICU_SOURCE_DIR}/i18n/ucoleitr.cpp"
+"${ICU_SOURCE_DIR}/i18n/ucsdet.cpp"
+"${ICU_SOURCE_DIR}/i18n/udat.cpp"
+"${ICU_SOURCE_DIR}/i18n/udateintervalformat.cpp"
+"${ICU_SOURCE_DIR}/i18n/udatpg.cpp"
+"${ICU_SOURCE_DIR}/i18n/ufieldpositer.cpp"
+"${ICU_SOURCE_DIR}/i18n/uitercollationiterator.cpp"
+"${ICU_SOURCE_DIR}/i18n/ulistformatter.cpp"
+"${ICU_SOURCE_DIR}/i18n/ulocdata.cpp"
+"${ICU_SOURCE_DIR}/i18n/umsg.cpp"
+"${ICU_SOURCE_DIR}/i18n/unesctrn.cpp"
+"${ICU_SOURCE_DIR}/i18n/uni2name.cpp"
 "${ICU_SOURCE_DIR}/i18n/units_complexconverter.cpp"
 "${ICU_SOURCE_DIR}/i18n/units_converter.cpp"
 "${ICU_SOURCE_DIR}/i18n/units_data.cpp"
-"${ICU_SOURCE_DIR}/i18n/units_router.cpp")
+"${ICU_SOURCE_DIR}/i18n/units_router.cpp"
+"${ICU_SOURCE_DIR}/i18n/unum.cpp"
+"${ICU_SOURCE_DIR}/i18n/unumsys.cpp"
+"${ICU_SOURCE_DIR}/i18n/upluralrules.cpp"
+"${ICU_SOURCE_DIR}/i18n/uregex.cpp"
+"${ICU_SOURCE_DIR}/i18n/uregexc.cpp"
+"${ICU_SOURCE_DIR}/i18n/uregion.cpp"
+"${ICU_SOURCE_DIR}/i18n/usearch.cpp"
+"${ICU_SOURCE_DIR}/i18n/uspoof.cpp"
+"${ICU_SOURCE_DIR}/i18n/uspoof_build.cpp"
+"${ICU_SOURCE_DIR}/i18n/uspoof_conf.cpp"
+"${ICU_SOURCE_DIR}/i18n/uspoof_impl.cpp"
+"${ICU_SOURCE_DIR}/i18n/utf16collationiterator.cpp"
+"${ICU_SOURCE_DIR}/i18n/utf8collationiterator.cpp"
+"${ICU_SOURCE_DIR}/i18n/utmscale.cpp"
+"${ICU_SOURCE_DIR}/i18n/utrans.cpp"
+"${ICU_SOURCE_DIR}/i18n/vtzone.cpp"
+"${ICU_SOURCE_DIR}/i18n/vzone.cpp"
+"${ICU_SOURCE_DIR}/i18n/windtfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/winnmfmt.cpp"
+"${ICU_SOURCE_DIR}/i18n/wintzimpl.cpp"
+"${ICU_SOURCE_DIR}/i18n/zonemeta.cpp"
+"${ICU_SOURCE_DIR}/i18n/zrule.cpp"
+"${ICU_SOURCE_DIR}/i18n/ztrans.cpp")
 
 file(GENERATE OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/empty.cpp" CONTENT " ")
 enable_language(ASM)
@@ -464,6 +481,11 @@ if (ARCH_S390X)
 else()
     set(ICUDATA_SOURCE_FILE "${ICUDATA_SOURCE_DIR}/icudt75l_dat.S" )
 endif()
+# ^^ you might be confused how for different little endian platforms (x86, ARM) the same assembly files can be used.
+# These files are indeed assembly but they only contain data ('.long' directive), which makes them portable accross CPUs.
+# Only the endianness and the character set (ASCII, EBCDIC) makes a difference, also see
+# https://unicode-org.github.io/icu/userguide/icu_data/#sharing-icu-data-between-platforms, 'Sharing ICU Data Between Platforms')
+# (and as an experiment, try re-generating the data files on x86 vs. ARM, ... you'll get exactly the same files)
 
 set(ICUDATA_SOURCES
     "${ICUDATA_SOURCE_FILE}"
diff --git a/contrib/libarchive b/contrib/libarchive
index 0c21691b177..313aa1fa10b 160000
--- a/contrib/libarchive
+++ b/contrib/libarchive
@@ -1 +1 @@
-Subproject commit 0c21691b177fac5f4cceca2a1ff2ddfa5d60f51c
+Subproject commit 313aa1fa10b657de791e3202c168a6c833bc3543
diff --git a/contrib/libarchive-cmake/CMakeLists.txt b/contrib/libarchive-cmake/CMakeLists.txt
index e89770da5f6..aa6dd9638b6 100644
--- a/contrib/libarchive-cmake/CMakeLists.txt
+++ b/contrib/libarchive-cmake/CMakeLists.txt
@@ -1,6 +1,6 @@
 set (LIBRARY_DIR "${ClickHouse_SOURCE_DIR}/contrib/libarchive")
 
-set(SRCS 
+set(SRCS
     "${LIBRARY_DIR}/libarchive/archive_acl.c"
     "${LIBRARY_DIR}/libarchive/archive_blake2sp_ref.c"
     "${LIBRARY_DIR}/libarchive/archive_blake2s_ref.c"
@@ -135,7 +135,7 @@ set(SRCS
 )
 
 add_library(_libarchive ${SRCS})
-target_include_directories(_libarchive PUBLIC 
+target_include_directories(_libarchive PUBLIC
     ${CMAKE_CURRENT_SOURCE_DIR}
     "${LIBRARY_DIR}/libarchive"
 )
@@ -157,7 +157,7 @@ if (TARGET ch_contrib::zlib)
 endif()
 
 if (TARGET ch_contrib::zstd)
-    target_compile_definitions(_libarchive PUBLIC HAVE_ZSTD_H=1 HAVE_LIBZSTD=1 HAVE_LIBZSTD_COMPRESSOR=1)
+    target_compile_definitions(_libarchive PUBLIC HAVE_ZSTD_H=1 HAVE_LIBZSTD=1 HAVE_ZSTD_compressStream=1)
     target_link_libraries(_libarchive PRIVATE ch_contrib::zstd)
 endif()
 
@@ -179,4 +179,4 @@ if (OS_LINUX)
     )
 endif()
 
-add_library(ch_contrib::libarchive ALIAS _libarchive)
\ No newline at end of file
+add_library(ch_contrib::libarchive ALIAS _libarchive)
diff --git a/contrib/libarchive-cmake/config.h b/contrib/libarchive-cmake/config.h
index 0b0cab47a52..9696cfb112d 100644
--- a/contrib/libarchive-cmake/config.h
+++ b/contrib/libarchive-cmake/config.h
@@ -334,13 +334,16 @@ typedef uint64_t uintmax_t;
 /* #undef ARCHIVE_XATTR_LINUX */
 
 /* Version number of bsdcpio */
-#define BSDCPIO_VERSION_STRING "3.7.0"
+#define BSDCPIO_VERSION_STRING "3.7.4"
 
 /* Version number of bsdtar */
-#define BSDTAR_VERSION_STRING "3.7.0"
+#define BSDTAR_VERSION_STRING "3.7.4"
 
 /* Version number of bsdcat */
-#define BSDCAT_VERSION_STRING "3.7.0"
+#define BSDCAT_VERSION_STRING "3.7.4"
+
+/* Version number of bsdunzip */
+#define BSDUNZIP_VERSION_STRING "3.7.4"
 
 /* Define to 1 if you have the `acl_create_entry' function. */
 /* #undef HAVE_ACL_CREATE_ENTRY */
@@ -642,8 +645,8 @@ typedef uint64_t uintmax_t;
 /* Define to 1 if you have the `getgrnam_r' function. */
 #define HAVE_GETGRNAM_R 1
 
-/* Define to 1 if platform uses `optreset` to reset `getopt` */
-#define HAVE_GETOPT_OPTRESET 1
+/* Define to 1 if you have the `getline' function. */
+#define HAVE_GETLINE 1
 
 /* Define to 1 if you have the `getpid' function. */
 #define HAVE_GETPID 1
@@ -750,6 +753,12 @@ typedef uint64_t uintmax_t;
 /* Define to 1 if you have the `pcreposix' library (-lpcreposix). */
 /* #undef HAVE_LIBPCREPOSIX */
 
+/* Define to 1 if you have the `pcre2-8' library (-lpcre2-8). */
+/* #undef HAVE_LIBPCRE2 */
+
+/* Define to 1 if you have the `pcreposix' library (-lpcre2posix). */
+/* #undef HAVE_LIBPCRE2POSIX */
+
 /* Define to 1 if you have the `xml2' library (-lxml2). */
 #define HAVE_LIBXML2 1
 
@@ -765,9 +774,8 @@ typedef uint64_t uintmax_t;
 /* Define to 1 if you have the `zstd' library (-lzstd). */
 /* #undef HAVE_LIBZSTD */
 
-/* Define to 1 if you have the `zstd' library (-lzstd) with compression
-   support. */
-/* #undef HAVE_LIBZSTD_COMPRESSOR */
+/* Define to 1 if you have the ZSTD_compressStream function. */
+/* #undef HAVE_ZSTD_compressStream */
 
 /* Define to 1 if you have the <limits.h> header file. */
 #define HAVE_LIMITS_H 1
@@ -923,6 +931,9 @@ typedef uint64_t uintmax_t;
 /* Define to 1 if you have the <pcreposix.h> header file. */
 /* #undef HAVE_PCREPOSIX_H */
 
+/* Define to 1 if you have the <pcre2posix.h> header file. */
+/* #undef HAVE_PCRE2POSIX_H */
+
 /* Define to 1 if you have the `pipe' function. */
 #define HAVE_PIPE 1
 
@@ -1029,6 +1040,12 @@ typedef uint64_t uintmax_t;
 /* Define to 1 if you have the `strrchr' function. */
 #define HAVE_STRRCHR 1
 
+/* Define to 1 if the system has the type `struct statfs'. */
+/* #undef HAVE_STRUCT_STATFS */
+
+/* Define to 1 if `f_iosize' is a member of `struct statfs'. */
+/* #undef HAVE_STRUCT_STATFS_F_IOSIZE */
+
 /* Define to 1 if `f_namemax' is a member of `struct statfs'. */
 /* #undef HAVE_STRUCT_STATFS_F_NAMEMAX */
 
@@ -1077,6 +1094,9 @@ typedef uint64_t uintmax_t;
 /* Define to 1 if you have the `symlink' function. */
 #define HAVE_SYMLINK 1
 
+/* Define to 1 if you have the `sysconf' function. */
+#define HAVE_SYSCONF 1
+
 /* Define to 1 if you have the <sys/acl.h> header file. */
 /* #undef HAVE_SYS_ACL_H */
 
@@ -1273,13 +1293,13 @@ typedef uint64_t uintmax_t;
 /* #undef HAVE__MKGMTIME */
 
 /* Define as const if the declaration of iconv() needs const. */
-#define ICONV_CONST 
+#define ICONV_CONST
 
 /* Version number of libarchive as a single integer */
-#define LIBARCHIVE_VERSION_NUMBER "3007000"
+#define LIBARCHIVE_VERSION_NUMBER "3007004"
 
 /* Version number of libarchive */
-#define LIBARCHIVE_VERSION_STRING "3.7.0"
+#define LIBARCHIVE_VERSION_STRING "3.7.4"
 
 /* Define to 1 if `lstat' dereferences a symlink specified with a trailing
    slash. */
@@ -1333,7 +1353,7 @@ typedef uint64_t uintmax_t;
 #endif /* SAFE_TO_DEFINE_EXTENSIONS */
 
 /* Version number of package */
-#define VERSION "3.7.0"
+#define VERSION "3.7.4"
 
 /* Number of bits in a file offset, on hosts where this is settable. */
 /* #undef _FILE_OFFSET_BITS */
diff --git a/docker/test/base/setup_export_logs.sh b/docker/test/base/setup_export_logs.sh
index e544397dd0c..a39f96867be 100755
--- a/docker/test/base/setup_export_logs.sh
+++ b/docker/test/base/setup_export_logs.sh
@@ -124,6 +124,8 @@ function setup_logs_replication
     check_logs_credentials || return 0
     __set_connection_args
 
+    echo "My hostname is ${HOSTNAME}"
+
     echo 'Create all configured system logs'
     clickhouse-client --query "SYSTEM FLUSH LOGS"
 
@@ -184,7 +186,17 @@ function setup_logs_replication
             /^TTL /d
             ')
 
-        echo -e "Creating remote destination table ${table}_${hash} with statement:\n${statement}" >&2
+        echo -e "Creating remote destination table ${table}_${hash} with statement:" >&2
+
+        echo "::group::${table}"
+        # there's the only way big "$statement" can be printed without causing EAGAIN error
+        # cat: write error: Resource temporarily unavailable
+        statement_print="${statement}"
+        if [ "${#statement_print}" -gt 4000 ]; then
+          statement_print="${statement::1999}\n…\n${statement:${#statement}-1999}"
+        fi
+        echo -e "$statement_print"
+        echo "::endgroup::"
 
         echo "$statement" | clickhouse-client --database_replicated_initial_query_timeout_sec=10 \
             --distributed_ddl_task_timeout=30 --distributed_ddl_output_mode=throw_only_active \
diff --git a/docs/en/engines/table-engines/integrations/azure-queue.md b/docs/en/engines/table-engines/integrations/azure-queue.md
new file mode 100644
index 00000000000..b5259336a8b
--- /dev/null
+++ b/docs/en/engines/table-engines/integrations/azure-queue.md
@@ -0,0 +1,72 @@
+---
+slug: /en/engines/table-engines/integrations/azure-queue
+sidebar_position: 181
+sidebar_label: AzureQueue
+---
+
+# AzureQueue Table Engine
+
+This engine provides an integration with [Azure Blob Storage](https://azure.microsoft.com/en-us/products/storage/blobs) ecosystem, allowing streaming data import.
+
+## Create Table {#creating-a-table}
+
+``` sql
+CREATE TABLE test (name String, value UInt32)
+    ENGINE = AzureQueue(...)
+    [SETTINGS]
+    [mode = '',]
+    [after_processing = 'keep',]
+    [keeper_path = '',]
+    ...
+```
+
+**Engine parameters**
+
+`AzureQueue` parameters are the same as `AzureBlobStorage` table engine supports. See parameters section [here](../../../engines/table-engines/integrations/azureBlobStorage.md).
+
+**Example**
+
+```sql
+CREATE TABLE azure_queue_engine_table (name String, value UInt32)
+ENGINE=AzureQueue('DefaultEndpointsProtocol=http;AccountName=devstoreaccount1;AccountKey=Eby8vdM02xNOcqFlqUwJPLlmEtlCDXJ1OUzFT50uSRZ6IFsuFq2UVErCz4I6tq/K1SZFPTOtr/KBHBeksoGMGw==;BlobEndpoint=http://azurite1:10000/devstoreaccount1/data/')
+SETTINGS
+    mode = 'unordered'
+```
+
+## Settings {#settings}
+
+The set of supported settings is the same as for `S3Queue` table engine, but without `s3queue_` prefix. See [full list of settings settings](../../../engines/table-engines/integrations/s3queue.md#settings).
+
+## Description {#description}
+
+`SELECT` is not particularly useful for streaming import (except for debugging), because each file can be imported only once. It is more practical to create real-time threads using [materialized views](../../../sql-reference/statements/create/view.md). To do this:
+
+1.  Use the engine to create a table for consuming from specified path in S3 and consider it a data stream.
+2.  Create a table with the desired structure.
+3.  Create a materialized view that converts data from the engine and puts it into a previously created table.
+
+When the `MATERIALIZED VIEW` joins the engine, it starts collecting data in the background.
+
+Example:
+
+``` sql
+  CREATE TABLE azure_queue_engine_table (name String, value UInt32)
+    ENGINE=AzureQueue('<endpoint>', 'CSV', 'gzip')
+    SETTINGS
+        mode = 'unordered';
+
+  CREATE TABLE stats (name String, value UInt32)
+    ENGINE = MergeTree() ORDER BY name;
+
+  CREATE MATERIALIZED VIEW consumer TO stats
+    AS SELECT name, value FROM azure_queue_engine_table;
+
+  SELECT * FROM stats ORDER BY name;
+```
+
+## Virtual columns {#virtual-columns}
+
+- `_path` — Path to the file.
+- `_file` — Name of the file.
+
+For more information about virtual columns see [here](../../../engines/table-engines/index.md#table_engines-virtual_columns).
diff --git a/docs/en/engines/table-engines/integrations/s3.md b/docs/en/engines/table-engines/integrations/s3.md
index 48a08dfa499..f02d0563491 100644
--- a/docs/en/engines/table-engines/integrations/s3.md
+++ b/docs/en/engines/table-engines/integrations/s3.md
@@ -35,7 +35,7 @@ CREATE TABLE s3_engine_table (name String, value UInt32)
     [SETTINGS ...]
 ```
 
-### Engine parameters
+### Engine parameters {#parameters}
 
 - `path` — Bucket url with path to file. Supports following wildcards in readonly mode: `*`, `**`, `?`, `{abc,def}` and `{N..M}` where `N`, `M` — numbers, `'abc'`, `'def'` — strings. For more information see [below](#wildcards-in-path).
 - `NOSIGN` - If this keyword is provided in place of credentials, all the requests will not be signed.
diff --git a/docs/en/engines/table-engines/integrations/s3queue.md b/docs/en/engines/table-engines/integrations/s3queue.md
index 06325fa15fb..f1957cfe1ce 100644
--- a/docs/en/engines/table-engines/integrations/s3queue.md
+++ b/docs/en/engines/table-engines/integrations/s3queue.md
@@ -5,6 +5,7 @@ sidebar_label: S3Queue
 ---
 
 # S3Queue Table Engine
+
 This engine provides integration with [Amazon S3](https://aws.amazon.com/s3/) ecosystem and allows streaming import. This engine is similar to the [Kafka](../../../engines/table-engines/integrations/kafka.md), [RabbitMQ](../../../engines/table-engines/integrations/rabbitmq.md) engines, but provides S3-specific features.
 
 ## Create Table {#creating-a-table}
@@ -16,27 +17,25 @@ CREATE TABLE s3_queue_engine_table (name String, value UInt32)
     [mode = '',]
     [after_processing = 'keep',]
     [keeper_path = '',]
-    [s3queue_loading_retries = 0,]
-    [s3queue_processing_threads_num = 1,]
-    [s3queue_enable_logging_to_s3queue_log = 0,]
-    [s3queue_polling_min_timeout_ms = 1000,]
-    [s3queue_polling_max_timeout_ms = 10000,]
-    [s3queue_polling_backoff_ms = 0,]
-    [s3queue_tracked_file_ttl_sec = 0,]
-    [s3queue_tracked_files_limit = 1000,]
-    [s3queue_cleanup_interval_min_ms = 10000,]
-    [s3queue_cleanup_interval_max_ms = 30000,]
+    [loading_retries = 0,]
+    [processing_threads_num = 1,]
+    [enable_logging_to_s3queue_log = 0,]
+    [polling_min_timeout_ms = 1000,]
+    [polling_max_timeout_ms = 10000,]
+    [polling_backoff_ms = 0,]
+    [tracked_file_ttl_sec = 0,]
+    [tracked_files_limit = 1000,]
+    [cleanup_interval_min_ms = 10000,]
+    [cleanup_interval_max_ms = 30000,]
 ```
 
-Starting with `24.7` settings without `s3queue_` prefix are also supported.
+:::warning
+Before `24.7`, it is required to use `s3queue_` prefix for all settings apart from `mode`, `after_processing` and `keeper_path`.
+:::
 
 **Engine parameters**
 
-- `path` — Bucket url with path to file. Supports following wildcards in readonly mode: `*`, `**`, `?`, `{abc,def}` and `{N..M}` where `N`, `M` — numbers, `'abc'`, `'def'` — strings. For more information see [below](#wildcards-in-path).
-- `NOSIGN` - If this keyword is provided in place of credentials, all the requests will not be signed.
-- `format` — The [format](../../../interfaces/formats.md#formats) of the file.
-- `aws_access_key_id`, `aws_secret_access_key` - Long-term credentials for the [AWS](https://aws.amazon.com/) account user.  You can use these to authenticate your requests. Parameter is optional. If credentials are not specified, they are used from the configuration file. For more information see [Using S3 for Data Storage](../mergetree-family/mergetree.md#table_engine-mergetree-s3).
-- `compression` — Compression type. Supported values: `none`, `gzip/gz`, `brotli/br`, `xz/LZMA`, `zstd/zst`. Parameter is optional. By default, it will autodetect compression by file extension.
+`S3Queue` parameters are the same as `S3` table engine supports. See parameters section [here](../../../engines/table-engines/integrations/s3.md#parameters).
 
 **Example**
 
diff --git a/docs/en/engines/table-engines/mergetree-family/mergetree.md b/docs/en/engines/table-engines/mergetree-family/mergetree.md
index 74f1295ef96..0bbee5f86f3 100644
--- a/docs/en/engines/table-engines/mergetree-family/mergetree.md
+++ b/docs/en/engines/table-engines/mergetree-family/mergetree.md
@@ -995,34 +995,42 @@ They can be used for prewhere optimization only if we enable `set allow_statisti
 
     The minimum and maximum column value which allows to estimate the selectivity of range filters on numeric columns.
 
+    Syntax: `minmax`
+
 - `TDigest`
 
     [TDigest](https://github.com/tdunning/t-digest) sketches which allow to compute approximate percentiles (e.g. the 90th percentile) for numeric columns.
 
+    Syntax: `tdigest`
+
 - `Uniq`
 
     [HyperLogLog](https://en.wikipedia.org/wiki/HyperLogLog) sketches which provide an estimation how many distinct values a column contains.
 
-- `count_min`
+    Syntax: `uniq`
 
-    [Count-min](https://en.wikipedia.org/wiki/Count%E2%80%93min_sketch) sketches which provide an approximate count of the frequency of each value in a column.
+- `CountMin`
+
+    [CountMin](https://en.wikipedia.org/wiki/Count%E2%80%93min_sketch) sketches which provide an approximate count of the frequency of each value in a column.
+
+    Syntax `countmin`
 
 
 ### Supported Data Types {#supported-data-types}
 
-|           | (U)Int* | Float* | Decimal(*) | Date* | Boolean | Enum* | (Fixed)String    |
-|-----------|---------|--------|------------|-------|---------|-------|------------------|
-| count_min | ✔       | ✔      | ✔          | ✔     | ✔       | ✔     | ✔                |
-| MinMax    | ✔       | ✔      | ✔          | ✔     | ✔       | ✔     | ✗                |
-| TDigest   | ✔       | ✔      | ✔          | ✔     | ✔       | ✔     | ✗                |
-| Uniq      | ✔       | ✔      | ✔          | ✔     | ✔       | ✔     | ✔                |
+|           | (U)Int*, Float*, Decimal(*), Date*, Boolean, Enum* | String or FixedString |
+|-----------|----------------------------------------------------|-----------------------|
+| CountMin  | ✔                                                  | ✔                     |
+| MinMax    | ✔                                                  | ✗                     |
+| TDigest   | ✔                                                  | ✗                     |
+| Uniq      | ✔                                                  | ✔                     |
 
 
 ### Supported Operations {#supported-operations}
 
 |           | Equality filters (==) | Range filters (>, >=, <, <=) |
 |-----------|-----------------------|------------------------------|
-| count_min | ✔                     | ✗                            |
+| CountMin  | ✔                     | ✗                            |
 | MinMax    | ✗                     | ✔                            |
 | TDigest   | ✗                     | ✔                            |
 | Uniq      | ✔                     | ✗                            |
diff --git a/docs/en/interfaces/formats.md b/docs/en/interfaces/formats.md
index 8892c6d8d3f..df96b8129f1 100644
--- a/docs/en/interfaces/formats.md
+++ b/docs/en/interfaces/formats.md
@@ -39,6 +39,7 @@ The supported formats are:
 | [JSONCompact](#jsoncompact)                                                               | ✔    | ✔     |
 | [JSONCompactStrings](#jsoncompactstrings)                                                 | ✗    | ✔     |
 | [JSONCompactColumns](#jsoncompactcolumns)                                                 | ✔    | ✔     |
+| [JSONCompactWithProgress](#jsoncompactwithprogress)                                       | ✗    | ✔     |
 | [JSONEachRow](#jsoneachrow)                                                               | ✔    | ✔     |
 | [PrettyJSONEachRow](#prettyjsoneachrow)                                                   | ✗    | ✔     |
 | [JSONEachRowWithProgress](#jsoneachrowwithprogress)                                       | ✗    | ✔     |
@@ -988,6 +989,59 @@ Example:
 
 Columns that are not present in the block will be filled with default values (you can use  [input_format_defaults_for_omitted_fields](/docs/en/operations/settings/settings-formats.md/#input_format_defaults_for_omitted_fields) setting here)
 
+## JSONCompactWithProgress (#jsoncompactwithprogress)
+
+In this format, ClickHouse outputs each row as a separated, newline-delimited JSON Object.
+
+Each row is either a metadata object, data object, progress information or statistics object:
+
+1. **Metadata Object (`meta`)**
+    - Describes the structure of the data rows.
+    - Fields: `name` (column name), `type` (data type, e.g., `UInt32`, `String`, etc.).
+    - Example: `{"meta": [{"name":"id", "type":"UInt32"}, {"name":"name", "type":"String"}]}`
+    - Appears before any data objects.
+
+2. **Data Object (`data`)**
+    - Represents a row of query results.
+    - Fields: An array with values corresponding to the columns defined in the metadata.
+    - Example: `{"data":["1", "John Doe"]}`
+    - Appears after the metadata object, one per row.
+
+3. **Progress Information Object (`progress`)**
+    - Provides real-time progress feedback during query execution.
+    - Fields: `read_rows`, `read_bytes`, `written_rows`, `written_bytes`, `total_rows_to_read`, `result_rows`, `result_bytes`, `elapsed_ns`.
+    - Example: `{"progress":{"read_rows":"8","read_bytes":"168"}}`
+    - May appear intermittently.
+
+4. **Statistics Object (`statistics`)**
+    - Summarizes query execution statistics.
+    - Fields: `rows`, `rows_before_limit_at_least`, `elapsed`, `rows_read`, `bytes_read`.
+    - Example: `{"statistics": {"rows":2, "elapsed":0.001995, "rows_read":8}}`
+    - Appears at the end.
+
+5. **Exception Object (`exception`)**
+    - Represents an error that occurred during query execution.
+    - Fields: A single text field containing the error message.
+    - Example: `{"exception": "Code: 395. DB::Exception: Value passed to 'throwIf' function is non-zero..."}`
+    - Appears when an error is encountered.
+
+6. **Totals Object (`totals`)**
+    - Provides the totals for each numeric column in the result set.
+    - Fields: An array with total values corresponding to the columns defined in the metadata.
+    - Example: `{"totals": ["", "3"]}`
+    - Appears at the end of the data rows, if applicable.
+
+Example:
+
+```json
+{"meta": [{"name":"id", "type":"UInt32"}, {"name":"name", "type":"String"}]}
+{"progress":{"read_rows":"8","read_bytes":"168","written_rows":"0","written_bytes":"0","total_rows_to_read":"2","result_rows":"0","result_bytes":"0","elapsed_ns":"0"}}
+{"data":["1", "John Doe"]}
+{"data":["2", "Joe Doe"]}
+{"statistics": {"rows":2, "rows_before_limit_at_least":8, "elapsed":0.001995, "rows_read":8, "bytes_read":168}}
+```
+
+
 ## JSONEachRow {#jsoneachrow}
 
 In this format, ClickHouse outputs each row as a separated, newline-delimited JSON Object.
diff --git a/docs/en/interfaces/third-party/gui.md b/docs/en/interfaces/third-party/gui.md
index 8d9dce983bc..16d9b66e17f 100644
--- a/docs/en/interfaces/third-party/gui.md
+++ b/docs/en/interfaces/third-party/gui.md
@@ -233,6 +233,16 @@ Features:
 - Useful tools: Zookeeper data exploration, query EXPLAIN, kill queries, etc.
 - Visualization metric charts: queries and resource usage, number of merges/mutation, merge performance, query performance, etc.
 
+### CKibana {#ckibana}
+
+[CKibana](https://github.com/TongchengOpenSource/ckibana) is a lightweight service that allows you to effortlessly search, explore, and visualize ClickHouse data using the native Kibana UI.
+
+Features:
+
+- Translates chart requests from the native Kibana UI into ClickHouse query syntax.
+- Supports advanced features such as sampling and caching to enhance query performance.
+- Minimizes the learning cost for users after migrating from ElasticSearch to ClickHouse.
+
 ## Commercial {#commercial}
 
 ### DataGrip {#datagrip}
diff --git a/docs/en/operations/server-configuration-parameters/settings.md b/docs/en/operations/server-configuration-parameters/settings.md
index 9fce83a0dc4..ccc8cf017ca 100644
--- a/docs/en/operations/server-configuration-parameters/settings.md
+++ b/docs/en/operations/server-configuration-parameters/settings.md
@@ -1463,26 +1463,29 @@ Examples:
 
 ## logger {#logger}
 
-Logging settings.
+The location and format of log messages.
 
 Keys:
 
-- `level` – Logging level. Acceptable values: `trace`, `debug`, `information`, `warning`, `error`.
-- `log` – The log file. Contains all the entries according to `level`.
-- `errorlog` – Error log file.
-- `size` – Size of the file. Applies to `log` and `errorlog`. Once the file reaches `size`, ClickHouse archives and renames it, and creates a new log file in its place.
-- `count` – The number of archived log files that ClickHouse stores.
-- `console` – Send `log` and `errorlog` to the console instead of file. To enable, set to `1` or `true`.
-- `console_log_level` – Logging level for console. Default to `level`.
-- `use_syslog` - Log to syslog as well.
-- `syslog_level` - Logging level for logging to syslog.
-- `stream_compress` – Compress `log` and `errorlog` with `lz4` stream compression. To enable, set to `1` or `true`.
-- `formatting` – Specify log format to be printed in console log (currently only `json` supported).
+- `level` – Log level. Acceptable values: `none` (turn logging off), `fatal`, `critical`, `error`, `warning`, `notice`, `information`,
+  `debug`, `trace`, `test`
+- `log` – The path to the log file.
+- `errorlog` – The path to the error log file.
+- `size` – Rotation policy: Maximum size of the log files in bytes. Once the log file size exceeds this threshold, it is renamed and archived, and a new log file is created.
+- `count` – Rotation policy: How many historical log files Clickhouse are kept at most.
+- `stream_compress` – Compress log messages using LZ4. Set to `1` or `true` to enable.
+- `console` – Do not write log messages to log files, instead print them in the console. Set to `1` or `true` to enable. Default is
+  `1` if Clickhouse does not run in daemon mode, `0` otherwise.
+- `console_log_level` – Log level for console output. Defaults to `level`.
+- `formatting` – Log format for console output. Currently, only `json` is supported).
+- `use_syslog` - Also forward log output to syslog.
+- `syslog_level` - Log level for logging to syslog.
 
-Both log and error log file names (only file names, not directories) support date and time format specifiers.
+**Log format specifiers**
 
-**Format specifiers**
-Using the following format specifiers, you can define a pattern for the resulting file name. “Example” column shows possible results for `2023-07-06 18:32:07`.
+File names in `log` and `errorLog` paths support below format specifiers for the resulting file name (the directory part does not support them).
+
+Column “Example” shows the output at `2023-07-06 18:32:07`.
 
 | Specifier   | Description                                                                                                         | Example                  |
 |-------------|---------------------------------------------------------------------------------------------------------------------|--------------------------|
@@ -1537,18 +1540,37 @@ Using the following format specifiers, you can define a pattern for the resultin
 </logger>
 ```
 
-Writing to the console can be configured. Config example:
+To print log messages only in the console:
 
 ``` xml
 <logger>
     <level>information</level>
-    <console>1</console>
+    <console>true</console>
+</logger>
+```
+
+**Per-level Overrides**
+
+The log level of individual log names can be overridden. For example, to mute all messages of loggers "Backup" and "RBAC".
+
+```xml
+<logger>
+    <levels>
+        <logger>
+            <name>Backup</name>
+            <level>none</level>
+        </logger>
+        <logger>
+            <name>RBAC</name>
+            <level>none</level>
+        </logger>
+    </levels>
 </logger>
 ```
 
 ### syslog
 
-Writing to the syslog is also supported. Config example:
+To write log messages additionally to syslog:
 
 ``` xml
 <logger>
@@ -1562,14 +1584,12 @@ Writing to the syslog is also supported. Config example:
 </logger>
 ```
 
-Keys for syslog:
+Keys for `<syslog>`:
 
-- use_syslog — Required setting if you want to write to the syslog.
-- address — The host\[:port\] of syslogd. If omitted, the local daemon is used.
-- hostname — Optional. The name of the host that logs are sent from.
-- facility — [The syslog facility keyword](https://en.wikipedia.org/wiki/Syslog#Facility) in uppercase letters with the “LOG_” prefix: (`LOG_USER`, `LOG_DAEMON`, `LOG_LOCAL3`, and so on).
-    Default value: `LOG_USER` if `address` is specified, `LOG_DAEMON` otherwise.
-- format – Message format. Possible values: `bsd` and `syslog.`
+- `address` — The address of syslog in format `host\[:port\]`. If omitted, the local daemon is used.
+- `hostname` — The name of the host from which logs are send. Optional.
+- `facility` — The syslog [facility keyword](https://en.wikipedia.org/wiki/Syslog#Facility). Must be specified uppercase with a “LOG_” prefix, e.g. `LOG_USER`, `LOG_DAEMON`, `LOG_LOCAL3`, etc. Default value: `LOG_USER` if `address` is specified, `LOG_DAEMON` otherwise.
+- `format` – Log message format. Possible values: `bsd` and `syslog.`
 
 ### Log formats
 
@@ -1588,6 +1608,7 @@ You can specify the log format that will be outputted in the console log. Curren
   "source_line": "192"
 }
 ```
+
 To enable JSON logging support, use the following snippet:
 
 ```xml
diff --git a/docs/en/operations/utilities/clickhouse-keeper-client.md b/docs/en/operations/utilities/clickhouse-keeper-client.md
index 2cd7be99dfb..6f026766750 100644
--- a/docs/en/operations/utilities/clickhouse-keeper-client.md
+++ b/docs/en/operations/utilities/clickhouse-keeper-client.md
@@ -47,6 +47,8 @@ keeper foo bar
 
 -   `ls '[path]'` -- Lists the nodes for the given path (default: cwd)
 -   `cd '[path]'` -- Changes the working path (default `.`)
+-   `cp '<src>' '<dest>'`  -- Copies 'src' node to 'dest' path
+-   `mv '<src>' '<dest>'`  -- Moves 'src' node to the 'dest' path
 -   `exists '<path>'` -- Returns `1` if node exists, `0` otherwise
 -   `set '<path>' <value> [version]` -- Updates the node's value. Only updates if version matches (default: -1)
 -   `create '<path>' <value> [mode]` -- Creates new node with the set value
diff --git a/docs/en/sql-reference/functions/date-time-functions.md b/docs/en/sql-reference/functions/date-time-functions.md
index b65fb3d7e95..d3db0e43041 100644
--- a/docs/en/sql-reference/functions/date-time-functions.md
+++ b/docs/en/sql-reference/functions/date-time-functions.md
@@ -1617,45 +1617,348 @@ The calculation is performed relative to specific points in time:
 
 If unit `WEEK` was specified, `toStartOfInterval` assumes that weeks start on Monday. Note that this behavior is different from that of function `toStartOfWeek` in which weeks start by default on Sunday.
 
-**See Also**
+**Syntax**
 
+```sql
+toStartOfInterval(value, INTERVAL x unit[, time_zone])
+toStartOfInterval(value, INTERVAL x unit[, origin[, time_zone]])
+```
+
+The second overload emulates TimescaleDB's `time_bucket()` function, respectively PostgreSQL's `date_bin()` function, e.g.
+
+``` SQL
+SELECT toStartOfInterval(toDateTime('2023-01-01 14:45:00'), INTERVAL 1 MINUTE, toDateTime('2023-01-01 14:35:30'));
+```
+**See Also**
 - [date_trunc](#date_trunc)
 
 ## toTime
 
 Converts a date with time to a certain fixed date, while preserving the time.
 
+**Syntax**
+
+```sql
+toTime(date[,timezone])
+```
+
+**Arguments**
+
+- `date` — Date to convert to a time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+- `timezone` (optional) — Timezone for the returned value. [String](../data-types/string.md).
+
+**Returned value**
+
+- DateTime with date equated to `1970-01-02` while preserving the time. [DateTime](../data-types/datetime.md).
+
+:::note
+If the `date` input argument contained sub-second components,
+they will be dropped in the returned `DateTime` value with second-accuracy.
+:::
+
+**Example**
+
+Query:
+
+```sql
+SELECT toTime(toDateTime64('1970-12-10 01:20:30.3000',3)) AS result, toTypeName(result);
+```
+
+Result:
+
+```response
+┌──────────────result─┬─toTypeName(result)─┐
+│ 1970-01-02 01:20:30 │ DateTime           │
+└─────────────────────┴────────────────────┘
+```
+
 ## toRelativeYearNum
 
-Converts a date, or date with time, to the number of the year, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of years elapsed since a certain fixed point in the past. 
+
+**Syntax**
+
+```sql
+toRelativeYearNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of years from a fixed reference point in the past. [UInt16](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+    toRelativeYearNum(toDate('2002-12-08')) AS y1,
+    toRelativeYearNum(toDate('2010-10-26')) AS y2
+```
+
+Result:
+
+```response
+┌───y1─┬───y2─┐
+│ 2002 │ 2010 │
+└──────┴──────┘
+```
 
 ## toRelativeQuarterNum
 
-Converts a date, or date with time, to the number of the quarter, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of quarters elapsed since a certain fixed point in the past.
+
+**Syntax**
+
+```sql
+toRelativeQuarterNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of quarters from a fixed reference point in the past. [UInt32](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+  toRelativeQuarterNum(toDate('1993-11-25')) AS q1,
+  toRelativeQuarterNum(toDate('2005-01-05')) AS q2
+```
+
+Result:
+
+```response
+┌───q1─┬───q2─┐
+│ 7975 │ 8020 │
+└──────┴──────┘
+```
 
 ## toRelativeMonthNum
 
-Converts a date, or date with time, to the number of the month, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of months elapsed since a certain fixed point in the past.
+
+**Syntax**
+
+```sql
+toRelativeMonthNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of months from a fixed reference point in the past. [UInt32](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+  toRelativeMonthNum(toDate('2001-04-25')) AS m1,
+  toRelativeMonthNum(toDate('2009-07-08')) AS m2
+```
+
+Result:
+
+```response
+┌────m1─┬────m2─┐
+│ 24016 │ 24115 │
+└───────┴───────┘
+```
 
 ## toRelativeWeekNum
 
-Converts a date, or date with time, to the number of the week, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of weeks elapsed since a certain fixed point in the past.
+
+**Syntax**
+
+```sql
+toRelativeWeekNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of weeks from a fixed reference point in the past. [UInt32](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+  toRelativeWeekNum(toDate('2000-02-29')) AS w1,
+  toRelativeWeekNum(toDate('2001-01-12')) AS w2
+```
+
+Result:
+
+```response
+┌───w1─┬───w2─┐
+│ 1574 │ 1619 │
+└──────┴──────┘
+```
 
 ## toRelativeDayNum
 
-Converts a date, or date with time, to the number of the day, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of days elapsed since a certain fixed point in the past.
+
+**Syntax**
+
+```sql
+toRelativeDayNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of days from a fixed reference point in the past. [UInt32](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+  toRelativeDayNum(toDate('1993-10-05')) AS d1,
+  toRelativeDayNum(toDate('2000-09-20')) AS d2
+```
+
+Result:
+
+```response
+┌───d1─┬────d2─┐
+│ 8678 │ 11220 │
+└──────┴───────┘
+```
 
 ## toRelativeHourNum
 
-Converts a date, or date with time, to the number of the hour, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of hours elapsed since a certain fixed point in the past.
+
+**Syntax**
+
+```sql
+toRelativeHourNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of hours from a fixed reference point in the past. [UInt32](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+  toRelativeHourNum(toDateTime('1993-10-05 05:20:36')) AS h1,
+  toRelativeHourNum(toDateTime('2000-09-20 14:11:29')) AS h2
+```
+
+Result:
+
+```response
+┌─────h1─┬─────h2─┐
+│ 208276 │ 269292 │
+└────────┴────────┘
+```
 
 ## toRelativeMinuteNum
 
-Converts a date, or date with time, to the number of the minute, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of minutes elapsed since a certain fixed point in the past.
+
+**Syntax**
+
+```sql
+toRelativeMinuteNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of minutes from a fixed reference point in the past. [UInt32](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+  toRelativeMinuteNum(toDateTime('1993-10-05 05:20:36')) AS m1,
+  toRelativeMinuteNum(toDateTime('2000-09-20 14:11:29')) AS m2
+```
+
+Result:
+
+```response
+┌───────m1─┬───────m2─┐
+│ 12496580 │ 16157531 │
+└──────────┴──────────┘
+```
 
 ## toRelativeSecondNum
 
-Converts a date, or date with time, to the number of the second, starting from a certain fixed point in the past.
+Converts a date, or date with time, to the number of the seconds elapsed since a certain fixed point in the past.
+
+**Syntax**
+
+```sql
+toRelativeSecondNum(date)
+```
+
+**Arguments**
+
+- `date` — Date or date with time. [Date](../data-types/date.md)/[DateTime](../data-types/datetime.md)/[DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- The number of seconds from a fixed reference point in the past. [UInt32](../data-types/int-uint.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+  toRelativeSecondNum(toDateTime('1993-10-05 05:20:36')) AS s1,
+  toRelativeSecondNum(toDateTime('2000-09-20 14:11:29')) AS s2
+```
+
+Result:
+
+```response
+┌────────s1─┬────────s2─┐
+│ 749794836 │ 969451889 │
+└───────────┴───────────┘
+```
 
 ## toISOYear
 
@@ -3884,19 +4187,29 @@ Result:
 └───────────────────────────────────────────────────────────────────────┘
 ```
 
-## timeSlots(StartTime, Duration,\[, Size\])
+## timeSlots
 
 For a time interval starting at ‘StartTime’ and continuing for ‘Duration’ seconds, it returns an array of moments in time, consisting of points from this interval rounded down to the ‘Size’ in seconds. ‘Size’ is an optional parameter set to 1800 (30 minutes) by default.
 This is necessary, for example, when searching for pageviews in the corresponding session.
 Accepts DateTime and DateTime64 as ’StartTime’ argument. For DateTime, ’Duration’ and ’Size’ arguments must be `UInt32`. For ’DateTime64’ they must be `Decimal64`.
 Returns an array of DateTime/DateTime64 (return type matches the type of ’StartTime’). For DateTime64, the return value's scale can differ from the scale of ’StartTime’ --- the highest scale among all given arguments is taken.
 
-Example:
+**Syntax**
+
+```sql
+timeSlots(StartTime, Duration,\[, Size\])
+```
+
+**Example**
+
 ```sql
 SELECT timeSlots(toDateTime('2012-01-01 12:20:00'), toUInt32(600));
 SELECT timeSlots(toDateTime('1980-12-12 21:01:02', 'UTC'), toUInt32(600), 299);
 SELECT timeSlots(toDateTime64('1980-12-12 21:01:02.1234', 4, 'UTC'), toDecimal64(600.1, 1), toDecimal64(299, 0));
 ```
+
+Result:
+
 ``` text
 ┌─timeSlots(toDateTime('2012-01-01 12:20:00'), toUInt32(600))─┐
 │ ['2012-01-01 12:00:00','2012-01-01 12:30:00']               │
diff --git a/docs/en/sql-reference/functions/type-conversion-functions.md b/docs/en/sql-reference/functions/type-conversion-functions.md
index f7fd2d68cf7..0a40ddcc2c2 100644
--- a/docs/en/sql-reference/functions/type-conversion-functions.md
+++ b/docs/en/sql-reference/functions/type-conversion-functions.md
@@ -3906,7 +3906,7 @@ Result:
 
 ## toDateTime64
 
-Converts the argument to the [DateTime64](../data-types/datetime64.md) data type.
+Converts an input value to a value of type [DateTime64](../data-types/datetime64.md).
 
 **Syntax**
 
@@ -3918,7 +3918,7 @@ toDateTime64(expr, scale, [timezone])
 
 - `expr` — The value. [String](../data-types/string.md), [UInt32](../data-types/int-uint.md), [Float](../data-types/float.md) or [DateTime](../data-types/datetime.md).
 - `scale` - Tick size (precision): 10<sup>-precision</sup> seconds. Valid range: [ 0 : 9 ].
-- `timezone` - Time zone of the specified datetime64 object.
+- `timezone` (optional) - Time zone of the specified datetime64 object.
 
 **Returned value**
 
@@ -3977,10 +3977,137 @@ SELECT toDateTime64('2019-01-01 00:00:00', 3, 'Asia/Istanbul') AS value, toTypeN
 
 ## toDateTime64OrZero
 
+Like [toDateTime64](#todatetime64), this function converts an input value to a value of type [DateTime64](../data-types/datetime64.md) but returns the min value of [DateTime64](../data-types/datetime64.md) if an invalid argument is received.
+
+**Syntax**
+
+``` sql
+toDateTime64OrZero(expr, scale, [timezone])
+```
+
+**Arguments**
+
+- `expr` — The value. [String](../data-types/string.md), [UInt32](../data-types/int-uint.md), [Float](../data-types/float.md) or [DateTime](../data-types/datetime.md).
+- `scale` - Tick size (precision): 10<sup>-precision</sup> seconds. Valid range: [ 0 : 9 ].
+- `timezone` (optional) - Time zone of the specified DateTime64 object.
+
+**Returned value**
+
+- A calendar date and time of day, with sub-second precision, otherwise the minimum value of `DateTime64`: `1970-01-01 01:00:00.000`. [DateTime64](../data-types/datetime64.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT toDateTime64OrZero('2008-10-12 00:00:00 00:30:30', 3) AS invalid_arg
+```
+
+Result:
+
+```response
+┌─────────────invalid_arg─┐
+│ 1970-01-01 01:00:00.000 │
+└─────────────────────────┘
+```
+
+**See also**
+
+- [toDateTime64](#todatetime64).
+- [toDateTime64OrNull](#todatetime64ornull).
+- [toDateTime64OrDefault](#todatetime64ordefault).
+
 ## toDateTime64OrNull
 
+Like [toDateTime64](#todatetime64), this function converts an input value to a value of type [DateTime64](../data-types/datetime64.md) but returns `NULL` if an invalid argument is received.
+
+**Syntax**
+
+``` sql
+toDateTime64OrNull(expr, scale, [timezone])
+```
+
+**Arguments**
+
+- `expr` — The value. [String](../data-types/string.md), [UInt32](../data-types/int-uint.md), [Float](../data-types/float.md) or [DateTime](../data-types/datetime.md).
+- `scale` - Tick size (precision): 10<sup>-precision</sup> seconds. Valid range: [ 0 : 9 ].
+- `timezone` (optional) - Time zone of the specified DateTime64 object.
+
+**Returned value**
+
+- A calendar date and time of day, with sub-second precision, otherwise `NULL`. [DateTime64](../data-types/datetime64.md)/[NULL](../data-types/nullable.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+    toDateTime64OrNull('1976-10-18 00:00:00.30', 3) AS valid_arg,
+    toDateTime64OrNull('1976-10-18 00:00:00 30', 3) AS invalid_arg
+```
+
+Result:
+
+```response
+┌───────────────valid_arg─┬─invalid_arg─┐
+│ 1976-10-18 00:00:00.300 │        ᴺᵁᴸᴸ │
+└─────────────────────────┴─────────────┘
+```
+
+**See also**
+
+- [toDateTime64](#todatetime64).
+- [toDateTime64OrZero](#todatetime64orzero).
+- [toDateTime64OrDefault](#todatetime64ordefault).
+
 ## toDateTime64OrDefault
 
+Like [toDateTime64](#todatetime64), this function converts an input value to a value of type [DateTime64](../data-types/datetime64.md), 
+but returns either the default value of [DateTime64](../data-types/datetime64.md)
+or the provided default if an invalid argument is received.
+
+**Syntax**
+
+``` sql
+toDateTime64OrNull(expr, scale, [timezone, default])
+```
+
+**Arguments**
+
+- `expr` — The value. [String](../data-types/string.md), [UInt32](../data-types/int-uint.md), [Float](../data-types/float.md) or [DateTime](../data-types/datetime.md).
+- `scale` - Tick size (precision): 10<sup>-precision</sup> seconds. Valid range: [ 0 : 9 ].
+- `timezone` (optional) - Time zone of the specified DateTime64 object.
+- `default` (optional) - Default value to return if an invalid argument is received. [DateTime64](../data-types/datetime64.md).
+
+**Returned value**
+
+- A calendar date and time of day, with sub-second precision, otherwise the minimum value of `DateTime64` or the `default` value if provided. [DateTime64](../data-types/datetime64.md).
+
+**Example**
+
+Query:
+
+```sql
+SELECT
+    toDateTime64OrDefault('1976-10-18 00:00:00 30', 3) AS invalid_arg,
+    toDateTime64OrDefault('1976-10-18 00:00:00 30', 3, 'UTC', toDateTime64('2001-01-01 00:00:00.00',3)) AS invalid_arg_with_default
+```
+
+Result:
+
+```response
+┌─────────────invalid_arg─┬─invalid_arg_with_default─┐
+│ 1970-01-01 01:00:00.000 │  2000-12-31 23:00:00.000 │
+└─────────────────────────┴──────────────────────────┘
+```
+
+**See also**
+
+- [toDateTime64](#todatetime64).
+- [toDateTime64OrZero](#todatetime64orzero).
+- [toDateTime64OrNull](#todatetime64ornull).
+
 ## toDecimal32
 
 Converts an input value to a value of type [`Decimal(9, S)`](../data-types/decimal.md) with scale of `S`. Throws an exception in case of an error.
diff --git a/docs/en/sql-reference/statements/select/from.md b/docs/en/sql-reference/statements/select/from.md
index 7a6e2ab054c..f319e7b1357 100644
--- a/docs/en/sql-reference/statements/select/from.md
+++ b/docs/en/sql-reference/statements/select/from.md
@@ -15,7 +15,14 @@ The `FROM` clause specifies the source to read data from:
 
 Subquery is another `SELECT` query that may be specified in parenthesis inside `FROM` clause.
 
-`FROM` clause can contain multiple data sources, separated by commas, which is equivalent of performing [CROSS JOIN](../../../sql-reference/statements/select/join.md) on them.
+The `FROM` can contain multiple data sources, separated by commas, which is equivalent of performing [CROSS JOIN](../../../sql-reference/statements/select/join.md) on them.
+
+`FROM` can optionally appear before a `SELECT` clause. This is a ClickHouse-specific extension of standard SQL which makes `SELECT` statements easier to read. Example:
+
+```sql
+FROM table
+SELECT *
+```
 
 ## FINAL Modifier
 
@@ -45,19 +52,19 @@ As an alternative to using `FINAL`, it is sometimes possible to use different qu
 
 ### Example Usage
 
-**Using the `FINAL` keyword**
+Using the `FINAL` keyword
 
 ```sql
 SELECT x, y FROM mytable FINAL WHERE x > 1;
 ```
 
-**Using `FINAL` as a query-level setting**
+Using `FINAL` as a query-level setting
 
 ```sql
 SELECT x, y FROM mytable WHERE x > 1 SETTINGS final = 1;
 ```
 
-**Using `FINAL` as a session-level setting**
+Using `FINAL` as a session-level setting
 
 ```sql
 SET final = 1;
diff --git a/programs/keeper/Keeper.cpp b/programs/keeper/Keeper.cpp
index ced661d9772..3007df60765 100644
--- a/programs/keeper/Keeper.cpp
+++ b/programs/keeper/Keeper.cpp
@@ -11,6 +11,7 @@
 #include <Core/ServerUUID.h>
 #include <Common/logger_useful.h>
 #include <Common/CgroupsMemoryUsageObserver.h>
+#include <Common/MemoryWorker.h>
 #include <Common/ErrorHandlers.h>
 #include <Common/assertProcessUserMatchesDataOwner.h>
 #include <Common/makeSocketAddress.h>
@@ -384,6 +385,9 @@ try
         LOG_INFO(log, "Background threads finished in {} ms", watch.elapsedMilliseconds());
     });
 
+    MemoryWorker memory_worker(config().getUInt64("memory_worker_period_ms", 0));
+    memory_worker.start();
+
     static ServerErrorHandler error_handler;
     Poco::ErrorHandler::set(&error_handler);
 
@@ -425,8 +429,9 @@ try
             for (const auto & server : *servers)
                 metrics.emplace_back(ProtocolServerMetrics{server.getPortName(), server.currentThreads(), server.refusedConnections()});
             return metrics;
-        }
-    );
+        },
+        /*update_jemalloc_epoch_=*/memory_worker.getSource() != MemoryWorker::MemoryUsageSource::Jemalloc,
+        /*update_rss_=*/memory_worker.getSource() == MemoryWorker::MemoryUsageSource::None);
 
     std::vector<std::string> listen_hosts = DB::getMultipleValuesFromConfig(config(), "", "listen_host");
 
@@ -655,7 +660,6 @@ try
     GWPAsan::initFinished();
 #endif
 
-
     LOG_INFO(log, "Ready for connections.");
 
     waitForTerminationRequest();
diff --git a/programs/server/Server.cpp b/programs/server/Server.cpp
index fb5717ba33f..c69d822e383 100644
--- a/programs/server/Server.cpp
+++ b/programs/server/Server.cpp
@@ -11,7 +11,6 @@
 #include <Poco/Util/HelpFormatter.h>
 #include <Poco/Environment.h>
 #include <Poco/Config.h>
-#include <Common/Jemalloc.h>
 #include <Common/scope_guard_safe.h>
 #include <Common/logger_useful.h>
 #include <base/phdr_cache.h>
@@ -25,6 +24,7 @@
 #include <base/Numa.h>
 #include <Common/PoolId.h>
 #include <Common/MemoryTracker.h>
+#include <Common/MemoryWorker.h>
 #include <Common/ClickHouseRevision.h>
 #include <Common/DNSResolver.h>
 #include <Common/CgroupsMemoryUsageObserver.h>
@@ -111,6 +111,8 @@
 #include <filesystem>
 #include <unordered_set>
 
+#include <Common/Jemalloc.h>
+
 #include "config.h"
 #include <Common/config_version.h>
 
@@ -449,9 +451,12 @@ void checkForUsersNotInMainConfig(
     }
 }
 
+namespace
+{
+
 /// Unused in other builds
 #if defined(OS_LINUX)
-static String readLine(const String & path)
+String readLine(const String & path)
 {
     ReadBufferFromFile in(path);
     String contents;
@@ -459,7 +464,7 @@ static String readLine(const String & path)
     return contents;
 }
 
-static int readNumber(const String & path)
+int readNumber(const String & path)
 {
     ReadBufferFromFile in(path);
     int result;
@@ -469,7 +474,7 @@ static int readNumber(const String & path)
 
 #endif
 
-static void sanityChecks(Server & server)
+void sanityChecks(Server & server)
 {
     std::string data_path = getCanonicalPath(server.config().getString("path", DBMS_DEFAULT_PATH));
     std::string logs_path = server.config().getString("logger.log", "");
@@ -590,6 +595,8 @@ static void sanityChecks(Server & server)
     }
 }
 
+}
+
 void loadStartupScripts(const Poco::Util::AbstractConfiguration & config, ContextMutablePtr context, Poco::Logger * log)
 {
     try
@@ -906,6 +913,8 @@ try
         LOG_INFO(log, "Background threads finished in {} ms", watch.elapsedMilliseconds());
     });
 
+    MemoryWorker memory_worker(global_context->getServerSettings().memory_worker_period_ms);
+
     /// This object will periodically calculate some metrics.
     ServerAsynchronousMetrics async_metrics(
         global_context,
@@ -924,8 +933,9 @@ try
             for (const auto & server : servers)
                 metrics.emplace_back(ProtocolServerMetrics{server.getPortName(), server.currentThreads(), server.refusedConnections()});
             return metrics;
-        }
-    );
+        },
+        /*update_jemalloc_epoch_=*/memory_worker.getSource() != MemoryWorker::MemoryUsageSource::Jemalloc,
+        /*update_rss_=*/memory_worker.getSource() == MemoryWorker::MemoryUsageSource::None);
 
     /// NOTE: global context should be destroyed *before* GlobalThreadPool::shutdown()
     /// Otherwise GlobalThreadPool::shutdown() will hang, since Context holds some threads.
@@ -1204,6 +1214,8 @@ try
 
     FailPointInjection::enableFromGlobalConfig(config());
 
+    memory_worker.start();
+
     int default_oom_score = 0;
 
 #if !defined(NDEBUG)
@@ -1547,15 +1559,6 @@ try
             total_memory_tracker.setDescription("(total)");
             total_memory_tracker.setMetric(CurrentMetrics::MemoryTracking);
 
-            if (cgroups_memory_usage_observer)
-            {
-                double hard_limit_ratio = new_server_settings.cgroup_memory_watcher_hard_limit_ratio;
-                double soft_limit_ratio = new_server_settings.cgroup_memory_watcher_soft_limit_ratio;
-                cgroups_memory_usage_observer->setMemoryUsageLimits(
-                    static_cast<uint64_t>(max_server_memory_usage * hard_limit_ratio),
-                    static_cast<uint64_t>(max_server_memory_usage * soft_limit_ratio));
-            }
-
             size_t merges_mutations_memory_usage_soft_limit = new_server_settings.merges_mutations_memory_usage_soft_limit;
 
             size_t default_merges_mutations_server_memory_usage = static_cast<size_t>(current_physical_server_memory * new_server_settings.merges_mutations_memory_usage_to_ram_ratio);
@@ -1584,8 +1587,6 @@ try
             background_memory_tracker.setDescription("(background)");
             background_memory_tracker.setMetric(CurrentMetrics::MergesMutationsMemoryTracking);
 
-            total_memory_tracker.setAllowUseJemallocMemory(new_server_settings.allow_use_jemalloc_memory);
-
             auto * global_overcommit_tracker = global_context->getGlobalOvercommitTracker();
             total_memory_tracker.setOvercommitTracker(global_overcommit_tracker);
 
diff --git a/src/AggregateFunctions/AggregateFunctionGroupConcat.cpp b/src/AggregateFunctions/AggregateFunctionGroupConcat.cpp
index 636ac80e350..8fb0b645096 100644
--- a/src/AggregateFunctions/AggregateFunctionGroupConcat.cpp
+++ b/src/AggregateFunctions/AggregateFunctionGroupConcat.cpp
@@ -116,15 +116,17 @@ class GroupConcatImpl final
     SerializationPtr serialization;
     UInt64 limit;
     const String delimiter;
+    const DataTypePtr type;
 
 public:
     GroupConcatImpl(const DataTypePtr & data_type_, const Array & parameters_, UInt64 limit_, const String & delimiter_)
         : IAggregateFunctionDataHelper<GroupConcatData<has_limit>, GroupConcatImpl<has_limit>>(
             {data_type_}, parameters_, std::make_shared<DataTypeString>())
-        , serialization(this->argument_types[0]->getDefaultSerialization())
         , limit(limit_)
         , delimiter(delimiter_)
+        , type(data_type_)
     {
+        serialization = isFixedString(type) ? std::make_shared<DataTypeString>()->getDefaultSerialization() : this->argument_types[0]->getDefaultSerialization();
     }
 
     String getName() const override { return name; }
@@ -140,7 +142,14 @@ public:
         if (cur_data.data_size != 0)
             cur_data.insertChar(delimiter.c_str(), delimiter.size(), arena);
 
-        cur_data.insert(columns[0], serialization, row_num, arena);
+        if (isFixedString(type))
+        {
+            ColumnWithTypeAndName col = {columns[0]->getPtr(), type, "column"};
+            const auto & col_str = castColumn(col, std::make_shared<DataTypeString>());
+            cur_data.insert(col_str.get(), serialization, row_num, arena);
+        }
+        else
+            cur_data.insert(columns[0], serialization, row_num, arena);
     }
 
     void merge(AggregateDataPtr __restrict place, ConstAggregateDataPtr rhs, Arena * arena) const override
diff --git a/src/AggregateFunctions/AggregateFunctionUniq.h b/src/AggregateFunctions/AggregateFunctionUniq.h
index cef23f766c7..35d6e599e38 100644
--- a/src/AggregateFunctions/AggregateFunctionUniq.h
+++ b/src/AggregateFunctions/AggregateFunctionUniq.h
@@ -459,6 +459,8 @@ public:
 
     bool isParallelizeMergePrepareNeeded() const override { return is_parallelize_merge_prepare_needed; }
 
+    constexpr static bool parallelizeMergeWithKey() { return true; }
+
     void parallelizeMergePrepare(AggregateDataPtrs & places, ThreadPool & thread_pool, std::atomic<bool> & is_cancelled) const override
     {
         if constexpr (is_parallelize_merge_prepare_needed)
diff --git a/src/AggregateFunctions/IAggregateFunction.h b/src/AggregateFunctions/IAggregateFunction.h
index ee227db6d9d..f8e7051d635 100644
--- a/src/AggregateFunctions/IAggregateFunction.h
+++ b/src/AggregateFunctions/IAggregateFunction.h
@@ -145,6 +145,8 @@ public:
 
     virtual bool isParallelizeMergePrepareNeeded() const { return false; }
 
+    constexpr static bool parallelizeMergeWithKey() { return false; }
+
     virtual void parallelizeMergePrepare(AggregateDataPtrs & /*places*/, ThreadPool & /*thread_pool*/, std::atomic<bool> & /*is_cancelled*/) const
     {
         throw Exception(ErrorCodes::NOT_IMPLEMENTED, "parallelizeMergePrepare() with thread pool parameter isn't implemented for {} ", getName());
@@ -169,7 +171,7 @@ public:
 
     /// Merges states (on which src places points to) with other states (on which dst places points to) of current aggregation function
     /// then destroy states (on which src places points to).
-    virtual void mergeAndDestroyBatch(AggregateDataPtr * dst_places, AggregateDataPtr * src_places, size_t size, size_t offset, Arena * arena) const = 0;
+    virtual void mergeAndDestroyBatch(AggregateDataPtr * dst_places, AggregateDataPtr * src_places, size_t size, size_t offset, ThreadPool & thread_pool, std::atomic<bool> & is_cancelled, Arena * arena) const = 0;
 
     /// Serializes state (to transmit it over the network, for example).
     virtual void serialize(ConstAggregateDataPtr __restrict place, WriteBuffer & buf, std::optional<size_t> version = std::nullopt) const = 0; /// NOLINT
@@ -499,11 +501,15 @@ public:
                 static_cast<const Derived *>(this)->merge(places[i] + place_offset, rhs[i], arena);
     }
 
-    void mergeAndDestroyBatch(AggregateDataPtr * dst_places, AggregateDataPtr * rhs_places, size_t size, size_t offset, Arena * arena) const override
+    void mergeAndDestroyBatch(AggregateDataPtr * dst_places, AggregateDataPtr * rhs_places, size_t size, size_t offset, ThreadPool & thread_pool, std::atomic<bool> & is_cancelled, Arena * arena) const override
     {
         for (size_t i = 0; i < size; ++i)
         {
-            static_cast<const Derived *>(this)->merge(dst_places[i] + offset, rhs_places[i] + offset, arena);
+            if constexpr (Derived::parallelizeMergeWithKey())
+                static_cast<const Derived *>(this)->merge(dst_places[i] + offset, rhs_places[i] + offset, thread_pool, is_cancelled, arena);
+            else
+                static_cast<const Derived *>(this)->merge(dst_places[i] + offset, rhs_places[i] + offset, arena);
+
             static_cast<const Derived *>(this)->destroy(rhs_places[i] + offset);
         }
     }
diff --git a/src/AggregateFunctions/UniqExactSet.h b/src/AggregateFunctions/UniqExactSet.h
index 2ae8c3a8386..25c6f7ac55f 100644
--- a/src/AggregateFunctions/UniqExactSet.h
+++ b/src/AggregateFunctions/UniqExactSet.h
@@ -101,6 +101,13 @@ public:
 
     auto merge(const UniqExactSet & other, ThreadPool * thread_pool = nullptr, std::atomic<bool> * is_cancelled = nullptr)
     {
+        /// If the size is large, we may convert the singleLevelHash to twoLevelHash and merge in parallel.
+        if (other.size() > 40000)
+        {
+            if (isSingleLevel())
+                convertToTwoLevel();
+        }
+
         if (isSingleLevel() && other.isTwoLevel())
             convertToTwoLevel();
 
diff --git a/src/Backups/RestorerFromBackup.cpp b/src/Backups/RestorerFromBackup.cpp
index 278af9d4eb3..5e6beec791d 100644
--- a/src/Backups/RestorerFromBackup.cpp
+++ b/src/Backups/RestorerFromBackup.cpp
@@ -913,11 +913,15 @@ void RestorerFromBackup::createTable(const QualifiedTableName & table_name)
             table_info.database = DatabaseCatalog::instance().getDatabase(table_name.database);
         DatabasePtr database = table_info.database;
 
+        auto query_context = Context::createCopy(context);
+        query_context->setSetting("database_replicated_allow_explicit_uuid", 3);
+        query_context->setSetting("database_replicated_allow_replicated_engine_arguments", 3);
+
         /// Execute CREATE TABLE query (we call IDatabase::createTableRestoredFromBackup() to allow the database to do some
         /// database-specific things).
         database->createTableRestoredFromBackup(
             create_table_query,
-            context,
+            query_context,
             restore_coordination,
             std::chrono::duration_cast<std::chrono::milliseconds>(create_table_timeout).count());
     }
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 1889bba3b39..f0410eee9fe 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -176,7 +176,7 @@ add_library (clickhouse_new_delete STATIC Common/new_delete.cpp)
 target_link_libraries (clickhouse_new_delete PRIVATE clickhouse_common_io)
 if (TARGET ch_contrib::jemalloc)
     target_link_libraries (clickhouse_new_delete PRIVATE ch_contrib::jemalloc)
-    target_link_libraries (clickhouse_common_io PRIVATE ch_contrib::jemalloc)
+    target_link_libraries (clickhouse_common_io PUBLIC ch_contrib::jemalloc)
     target_link_libraries (clickhouse_storages_system PRIVATE ch_contrib::jemalloc)
 endif()
 
diff --git a/src/Client/ConnectionPoolWithFailover.cpp b/src/Client/ConnectionPoolWithFailover.cpp
index fb895d17763..a5c14dc9957 100644
--- a/src/Client/ConnectionPoolWithFailover.cpp
+++ b/src/Client/ConnectionPoolWithFailover.cpp
@@ -168,7 +168,7 @@ std::vector<ConnectionPoolWithFailover::TryResult> ConnectionPoolWithFailover::g
     { return tryGetEntry(pool, timeouts, fail_message, settings, &table_to_check, /*async_callback=*/ {}); };
 
     return getManyImpl(settings, pool_mode, try_get_entry,
-        /*skip_unavailable_endpoints=*/ std::nullopt,
+        /*skip_unavailable_endpoints=*/ false, /// skip_unavailable_endpoints is used to get the min number of entries, and we need at least one
         /*priority_func=*/ {},
         settings.distributed_insert_skip_read_only_replicas);
 }
diff --git a/src/Client/ConnectionPoolWithFailover.h b/src/Client/ConnectionPoolWithFailover.h
index 308644ebbdb..6db52140854 100644
--- a/src/Client/ConnectionPoolWithFailover.h
+++ b/src/Client/ConnectionPoolWithFailover.h
@@ -42,7 +42,7 @@ public:
             size_t max_error_cap = DBMS_CONNECTION_POOL_WITH_FAILOVER_MAX_ERROR_COUNT);
 
     using Entry = IConnectionPool::Entry;
-    using PoolWithFailoverBase<IConnectionPool>::checkTryResultIsValid;
+    using PoolWithFailoverBase<IConnectionPool>::getValidTryResult;
 
     /** Allocates connection to work. */
     Entry get(const ConnectionTimeouts & timeouts) override;
@@ -98,7 +98,7 @@ public:
 
     std::vector<Base::ShuffledPool> getShuffledPools(const Settings & settings, GetPriorityFunc priority_func = {}, bool use_slowdown_count = false);
 
-    size_t getMaxErrorCup() const { return Base::max_error_cap; }
+    size_t getMaxErrorCap() const { return Base::max_error_cap; }
 
     void updateSharedError(std::vector<ShuffledPool> & shuffled_pools)
     {
diff --git a/src/Client/HedgedConnectionsFactory.cpp b/src/Client/HedgedConnectionsFactory.cpp
index be7397b0fad..df63a124539 100644
--- a/src/Client/HedgedConnectionsFactory.cpp
+++ b/src/Client/HedgedConnectionsFactory.cpp
@@ -327,7 +327,7 @@ HedgedConnectionsFactory::State HedgedConnectionsFactory::processFinishedConnect
         ShuffledPool & shuffled_pool = shuffled_pools[index];
         LOG_INFO(log, "Connection failed at try №{}, reason: {}", (shuffled_pool.error_count + 1), fail_message);
 
-        shuffled_pool.error_count = std::min(pool->getMaxErrorCup(), shuffled_pool.error_count + 1);
+        shuffled_pool.error_count = std::min(pool->getMaxErrorCap(), shuffled_pool.error_count + 1);
         shuffled_pool.slowdown_count = 0;
 
         if (shuffled_pool.error_count >= max_tries)
diff --git a/src/Common/AsynchronousMetrics.cpp b/src/Common/AsynchronousMetrics.cpp
index 9b6a7428411..a92d321f8aa 100644
--- a/src/Common/AsynchronousMetrics.cpp
+++ b/src/Common/AsynchronousMetrics.cpp
@@ -1,5 +1,3 @@
-#include <Common/AsynchronousMetrics.h>
-
 #include <IO/MMappedFileCache.h>
 #include <IO/ReadHelpers.h>
 #include <IO/UncompressedCache.h>
@@ -8,8 +6,10 @@
 #include <base/find_symbols.h>
 #include <base/getPageSize.h>
 #include <sys/resource.h>
+#include <Common/AsynchronousMetrics.h>
 #include <Common/CurrentMetrics.h>
 #include <Common/Exception.h>
+#include <Common/Jemalloc.h>
 #include <Common/filesystemHelpers.h>
 #include <Common/formatReadable.h>
 #include <Common/logger_useful.h>
@@ -69,10 +69,14 @@ static void openCgroupv2MetricFile(const std::string & filename, std::optional<R
 
 AsynchronousMetrics::AsynchronousMetrics(
     unsigned update_period_seconds,
-    const ProtocolServerMetricsFunc & protocol_server_metrics_func_)
+    const ProtocolServerMetricsFunc & protocol_server_metrics_func_,
+    bool update_jemalloc_epoch_,
+    bool update_rss_)
     : update_period(update_period_seconds)
     , log(getLogger("AsynchronousMetrics"))
     , protocol_server_metrics_func(protocol_server_metrics_func_)
+    , update_jemalloc_epoch(update_jemalloc_epoch_)
+    , update_rss(update_rss_)
 {
 #if defined(OS_LINUX)
     openFileIfExists("/proc/cpuinfo", cpuinfo);
@@ -411,9 +415,7 @@ Value saveJemallocMetricImpl(
     const std::string & jemalloc_full_name,
     const std::string & clickhouse_full_name)
 {
-    Value value{};
-    size_t size = sizeof(value);
-    mallctl(jemalloc_full_name.c_str(), &value, &size, nullptr, 0);
+    auto value = getJemallocValue<Value>(jemalloc_full_name.c_str());
     values[clickhouse_full_name] = AsynchronousMetricValue(value, "An internal metric of the low-level memory allocator (jemalloc). See https://jemalloc.net/jemalloc.3.html");
     return value;
 }
@@ -768,8 +770,11 @@ void AsynchronousMetrics::update(TimePoint update_time, bool force_update)
     // 'epoch' is a special mallctl -- it updates the statistics. Without it, all
     // the following calls will return stale values. It increments and returns
     // the current epoch number, which might be useful to log as a sanity check.
-    auto epoch = updateJemallocEpoch();
-    new_values["jemalloc.epoch"] = { epoch, "An internal incremental update number of the statistics of jemalloc (Jason Evans' memory allocator), used in all other `jemalloc` metrics." };
+    auto epoch = update_jemalloc_epoch ? updateJemallocEpoch() : getJemallocValue<uint64_t>("epoch");
+    new_values["jemalloc.epoch"]
+        = {epoch,
+           "An internal incremental update number of the statistics of jemalloc (Jason Evans' memory allocator), used in all other "
+           "`jemalloc` metrics."};
 
     // Collect the statistics themselves.
     saveJemallocMetric<size_t>(new_values, "allocated");
@@ -782,10 +787,10 @@ void AsynchronousMetrics::update(TimePoint update_time, bool force_update)
     saveJemallocMetric<size_t>(new_values, "background_thread.num_threads");
     saveJemallocMetric<uint64_t>(new_values, "background_thread.num_runs");
     saveJemallocMetric<uint64_t>(new_values, "background_thread.run_intervals");
-    saveJemallocProf<size_t>(new_values, "active");
+    saveJemallocProf<bool>(new_values, "active");
     saveAllArenasMetric<size_t>(new_values, "pactive");
-    [[maybe_unused]] size_t je_malloc_pdirty = saveAllArenasMetric<size_t>(new_values, "pdirty");
-    [[maybe_unused]] size_t je_malloc_pmuzzy = saveAllArenasMetric<size_t>(new_values, "pmuzzy");
+    saveAllArenasMetric<size_t>(new_values, "pdirty");
+    saveAllArenasMetric<size_t>(new_values, "pmuzzy");
     saveAllArenasMetric<size_t>(new_values, "dirty_purged");
     saveAllArenasMetric<size_t>(new_values, "muzzy_purged");
 #endif
@@ -814,41 +819,8 @@ void AsynchronousMetrics::update(TimePoint update_time, bool force_update)
             " It is unspecified whether it includes the per-thread stacks and most of the allocated memory, that is allocated with the 'mmap' system call."
             " This metric exists only for completeness reasons. I recommend to use the `MemoryResident` metric for monitoring."};
 
-        /// We must update the value of total_memory_tracker periodically.
-        /// Otherwise it might be calculated incorrectly - it can include a "drift" of memory amount.
-        /// See https://github.com/ClickHouse/ClickHouse/issues/10293
-        {
-            Int64 amount = total_memory_tracker.get();
-            Int64 peak = total_memory_tracker.getPeak();
-            Int64 rss = data.resident;
-            Int64 free_memory_in_allocator_arenas = 0;
-
-#if USE_JEMALLOC
-            /// According to jemalloc man, pdirty is:
-            ///
-            ///     Number of pages within unused extents that are potentially
-            ///     dirty, and for which madvise() or similar has not been called.
-            ///
-            /// So they will be subtracted from RSS to make accounting more
-            /// accurate, since those pages are not really RSS but a memory
-            /// that can be used at anytime via jemalloc.
-            free_memory_in_allocator_arenas = je_malloc_pdirty * getPageSize();
-#endif
-
-            Int64 difference = rss - amount;
-
-            /// Log only if difference is high. This is for convenience. The threshold is arbitrary.
-            if (difference >= 1048576 || difference <= -1048576)
-                LOG_TRACE(log,
-                    "MemoryTracking: was {}, peak {}, free memory in arenas {}, will set to {} (RSS), difference: {}",
-                    ReadableSize(amount),
-                    ReadableSize(peak),
-                    ReadableSize(free_memory_in_allocator_arenas),
-                    ReadableSize(rss),
-                    ReadableSize(difference));
-
-            MemoryTracker::setRSS(rss, free_memory_in_allocator_arenas);
-        }
+        if (update_rss)
+            MemoryTracker::updateRSS(data.resident);
     }
 
     {
diff --git a/src/Common/AsynchronousMetrics.h b/src/Common/AsynchronousMetrics.h
index 78d07ef4b6c..215dc6e1337 100644
--- a/src/Common/AsynchronousMetrics.h
+++ b/src/Common/AsynchronousMetrics.h
@@ -1,15 +1,14 @@
 #pragma once
 
+#include <Common/CgroupsMemoryUsageObserver.h>
 #include <Common/MemoryStatisticsOS.h>
 #include <Common/ThreadPool.h>
 #include <Common/Stopwatch.h>
 #include <IO/ReadBufferFromFile.h>
 
 #include <condition_variable>
-#include <map>
 #include <mutex>
 #include <string>
-#include <thread>
 #include <vector>
 #include <optional>
 #include <unordered_map>
@@ -69,7 +68,9 @@ public:
 
     AsynchronousMetrics(
         unsigned update_period_seconds,
-        const ProtocolServerMetricsFunc & protocol_server_metrics_func_);
+        const ProtocolServerMetricsFunc & protocol_server_metrics_func_,
+        bool update_jemalloc_epoch_,
+        bool update_rss_);
 
     virtual ~AsynchronousMetrics();
 
@@ -112,6 +113,9 @@ private:
     MemoryStatisticsOS memory_stat TSA_GUARDED_BY(data_mutex);
 #endif
 
+    [[maybe_unused]] const bool update_jemalloc_epoch;
+    [[maybe_unused]] const bool update_rss;
+
 #if defined(OS_LINUX)
     std::optional<ReadBufferFromFilePRead> meminfo TSA_GUARDED_BY(data_mutex);
     std::optional<ReadBufferFromFilePRead> loadavg TSA_GUARDED_BY(data_mutex);
diff --git a/src/Common/CgroupsMemoryUsageObserver.cpp b/src/Common/CgroupsMemoryUsageObserver.cpp
index 83b04360164..28bb861865a 100644
--- a/src/Common/CgroupsMemoryUsageObserver.cpp
+++ b/src/Common/CgroupsMemoryUsageObserver.cpp
@@ -14,239 +14,21 @@
 #include <fmt/ranges.h>
 
 #include <cstdint>
-#include <filesystem>
-#include <memory>
-#include <optional>
-
-#include "config.h"
-#if USE_JEMALLOC
-#    include <jemalloc/jemalloc.h>
-#define STRINGIFY_HELPER(x) #x
-#define STRINGIFY(x) STRINGIFY_HELPER(x)
-#endif
 
 using namespace DB;
-namespace fs = std::filesystem;
-
-namespace DB
-{
-
-namespace ErrorCodes
-{
-extern const int FILE_DOESNT_EXIST;
-extern const int INCORRECT_DATA;
-}
-
-}
-
-namespace
-{
-
-/// Format is
-///   kernel 5
-///   rss 15
-///   [...]
-using Metrics = std::map<std::string, uint64_t>;
-
-Metrics readAllMetricsFromStatFile(ReadBufferFromFile & buf)
-{
-    Metrics metrics;
-    while (!buf.eof())
-    {
-        std::string current_key;
-        readStringUntilWhitespace(current_key, buf);
-
-        assertChar(' ', buf);
-
-        uint64_t value = 0;
-        readIntText(value, buf);
-        assertChar('\n', buf);
-
-        auto [_, inserted] = metrics.emplace(std::move(current_key), value);
-        chassert(inserted, "Duplicate keys in stat file");
-    }
-    return metrics;
-}
-
-uint64_t readMetricFromStatFile(ReadBufferFromFile & buf, const std::string & key)
-{
-    const auto all_metrics = readAllMetricsFromStatFile(buf);
-    if (const auto it = all_metrics.find(key); it != all_metrics.end())
-        return it->second;
-    throw Exception(ErrorCodes::INCORRECT_DATA, "Cannot find '{}' in '{}'", key, buf.getFileName());
-}
-
-struct CgroupsV1Reader : ICgroupsReader
-{
-    explicit CgroupsV1Reader(const fs::path & stat_file_dir) : buf(stat_file_dir / "memory.stat") { }
-
-    uint64_t readMemoryUsage() override
-    {
-        std::lock_guard lock(mutex);
-        buf.rewind();
-        return readMetricFromStatFile(buf, "rss");
-    }
-
-    std::string dumpAllStats() override
-    {
-        std::lock_guard lock(mutex);
-        buf.rewind();
-        return fmt::format("{}", readAllMetricsFromStatFile(buf));
-    }
-
-private:
-    std::mutex mutex;
-    ReadBufferFromFile buf TSA_GUARDED_BY(mutex);
-};
-
-struct CgroupsV2Reader : ICgroupsReader
-{
-    explicit CgroupsV2Reader(const fs::path & stat_file_dir)
-        : current_buf(stat_file_dir / "memory.current"), stat_buf(stat_file_dir / "memory.stat")
-    {
-    }
-
-    uint64_t readMemoryUsage() override
-    {
-        std::lock_guard lock(mutex);
-        current_buf.rewind();
-        stat_buf.rewind();
-
-        int64_t mem_usage = 0;
-        /// memory.current contains a single number
-        /// the reason why we subtract it described here: https://github.com/ClickHouse/ClickHouse/issues/64652#issuecomment-2149630667
-        readIntText(mem_usage, current_buf);
-        mem_usage -= readMetricFromStatFile(stat_buf, "inactive_file");
-        chassert(mem_usage >= 0, "Negative memory usage");
-        return mem_usage;
-    }
-
-    std::string dumpAllStats() override
-    {
-        std::lock_guard lock(mutex);
-        stat_buf.rewind();
-        return fmt::format("{}", readAllMetricsFromStatFile(stat_buf));
-    }
-
-private:
-    std::mutex mutex;
-    ReadBufferFromFile current_buf TSA_GUARDED_BY(mutex);
-    ReadBufferFromFile stat_buf TSA_GUARDED_BY(mutex);
-};
-
-/// Caveats:
-/// - All of the logic in this file assumes that the current process is the only process in the
-///   containing cgroup (or more precisely: the only process with significant memory consumption).
-///   If this is not the case, then other processe's memory consumption may affect the internal
-///   memory tracker ...
-/// - Cgroups v1 and v2 allow nested cgroup hierarchies. As v1 is deprecated for over half a
-///   decade and will go away at some point, hierarchical detection is only implemented for v2.
-/// - I did not test what happens if a host has v1 and v2 simultaneously enabled. I believe such
-///   systems existed only for a short transition period.
-
-std::optional<std::string> getCgroupsV1Path()
-{
-    auto path = default_cgroups_mount / "memory/memory.stat";
-    if (!fs::exists(path))
-        return {};
-    return {default_cgroups_mount / "memory"};
-}
-
-std::pair<std::string, CgroupsMemoryUsageObserver::CgroupsVersion> getCgroupsPath()
-{
-    auto v2_path = getCgroupsV2PathContainingFile("memory.current");
-    if (v2_path.has_value())
-        return {*v2_path, CgroupsMemoryUsageObserver::CgroupsVersion::V2};
-
-    auto v1_path = getCgroupsV1Path();
-    if (v1_path.has_value())
-        return {*v1_path, CgroupsMemoryUsageObserver::CgroupsVersion::V1};
-
-    throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "Cannot find cgroups v1 or v2 current memory file");
-}
-
-}
 
 namespace DB
 {
 
 CgroupsMemoryUsageObserver::CgroupsMemoryUsageObserver(std::chrono::seconds wait_time_)
     : log(getLogger("CgroupsMemoryUsageObserver")), wait_time(wait_time_)
-{
-    const auto [cgroup_path, version] = getCgroupsPath();
-
-    cgroup_reader = createCgroupsReader(version, cgroup_path);
-
-    LOG_INFO(
-        log,
-        "Will read the current memory usage from '{}' (cgroups version: {}), wait time is {} sec",
-        cgroup_path,
-        (version == CgroupsVersion::V1) ? "v1" : "v2",
-        wait_time.count());
-}
+{}
 
 CgroupsMemoryUsageObserver::~CgroupsMemoryUsageObserver()
 {
     stopThread();
 }
 
-void CgroupsMemoryUsageObserver::setMemoryUsageLimits(uint64_t hard_limit_, uint64_t soft_limit_)
-{
-    std::lock_guard<std::mutex> limit_lock(limit_mutex);
-
-    if (hard_limit_ == hard_limit && soft_limit_ == soft_limit)
-        return;
-
-    hard_limit = hard_limit_;
-    soft_limit = soft_limit_;
-
-    on_hard_limit = [this, hard_limit_](bool up)
-    {
-        if (up)
-        {
-            LOG_WARNING(log, "Exceeded hard memory limit ({})", ReadableSize(hard_limit_));
-
-            /// Update current usage in memory tracker. Also reset free_memory_in_allocator_arenas to zero though we don't know if they are
-            /// really zero. Trying to avoid OOM ...
-            MemoryTracker::setRSS(hard_limit_, 0);
-        }
-        else
-        {
-            LOG_INFO(log, "Dropped below hard memory limit ({})", ReadableSize(hard_limit_));
-        }
-    };
-
-    on_soft_limit = [this, soft_limit_](bool up)
-    {
-        if (up)
-        {
-            LOG_WARNING(log, "Exceeded soft memory limit ({})", ReadableSize(soft_limit_));
-
-#    if USE_JEMALLOC
-            LOG_INFO(log, "Purging jemalloc arenas");
-            mallctl("arena." STRINGIFY(MALLCTL_ARENAS_ALL) ".purge", nullptr, nullptr, nullptr, 0);
-#    endif
-            /// Reset current usage in memory tracker. Expect zero for free_memory_in_allocator_arenas as we just purged them.
-            uint64_t memory_usage = cgroup_reader->readMemoryUsage();
-            LOG_TRACE(
-                log,
-                "Read current memory usage {} bytes ({}) from cgroups, full available stats: {}",
-                memory_usage,
-                ReadableSize(memory_usage),
-                cgroup_reader->dumpAllStats());
-            MemoryTracker::setRSS(memory_usage, 0);
-
-            LOG_INFO(log, "Purged jemalloc arenas. Current memory usage is {}", ReadableSize(memory_usage));
-        }
-        else
-        {
-            LOG_INFO(log, "Dropped below soft memory limit ({})", ReadableSize(soft_limit_));
-        }
-    };
-
-    LOG_INFO(log, "Set new limits, soft limit: {}, hard limit: {}", ReadableSize(soft_limit_), ReadableSize(hard_limit_));
-}
-
 void CgroupsMemoryUsageObserver::setOnMemoryAmountAvailableChangedFn(OnMemoryAmountAvailableChangedFn on_memory_amount_available_changed_)
 {
     std::lock_guard<std::mutex> memory_amount_available_changed_lock(memory_amount_available_changed_mutex);
@@ -300,35 +82,6 @@ void CgroupsMemoryUsageObserver::runThread()
                 std::lock_guard<std::mutex> memory_amount_available_changed_lock(memory_amount_available_changed_mutex);
                 on_memory_amount_available_changed();
             }
-
-            std::lock_guard<std::mutex> limit_lock(limit_mutex);
-            if (soft_limit > 0 && hard_limit > 0)
-            {
-                uint64_t memory_usage = cgroup_reader->readMemoryUsage();
-                LOG_TRACE(log, "Read current memory usage {} bytes ({}) from cgroups", memory_usage, ReadableSize(memory_usage));
-                if (memory_usage > hard_limit)
-                {
-                    if (last_memory_usage <= hard_limit)
-                        on_hard_limit(true);
-                }
-                else
-                {
-                    if (last_memory_usage > hard_limit)
-                        on_hard_limit(false);
-                }
-
-                if (memory_usage > soft_limit)
-                {
-                    if (last_memory_usage <= soft_limit)
-                        on_soft_limit(true);
-                }
-                else
-                {
-                    if (last_memory_usage > soft_limit)
-                        on_soft_limit(false);
-                }
-                last_memory_usage = memory_usage;
-            }
         }
         catch (...)
         {
@@ -337,13 +90,6 @@ void CgroupsMemoryUsageObserver::runThread()
     }
 }
 
-std::unique_ptr<ICgroupsReader> createCgroupsReader(CgroupsMemoryUsageObserver::CgroupsVersion version, const fs::path & cgroup_path)
-{
-    if (version == CgroupsMemoryUsageObserver::CgroupsVersion::V2)
-        return std::make_unique<CgroupsV2Reader>(cgroup_path);
-    else
-        return std::make_unique<CgroupsV1Reader>(cgroup_path);
-}
 }
 
 #endif
diff --git a/src/Common/CgroupsMemoryUsageObserver.h b/src/Common/CgroupsMemoryUsageObserver.h
index 7f888fe631b..3de83d6b437 100644
--- a/src/Common/CgroupsMemoryUsageObserver.h
+++ b/src/Common/CgroupsMemoryUsageObserver.h
@@ -3,53 +3,27 @@
 #include <Common/ThreadPool.h>
 
 #include <chrono>
-#include <memory>
 #include <mutex>
 
 namespace DB
 {
 
-struct ICgroupsReader
-{
-    virtual ~ICgroupsReader() = default;
-
-    virtual uint64_t readMemoryUsage() = 0;
-
-    virtual std::string dumpAllStats() = 0;
-};
-
-/// Does two things:
-/// 1. Periodically reads the memory usage of the process from Linux cgroups.
-///    You can specify soft or hard memory limits:
-///    - When the soft memory limit is hit, drop jemalloc cache.
-///    - When the hard memory limit is hit, update MemoryTracking metric to throw memory exceptions faster.
-///    The goal of this is to avoid that the process hits the maximum allowed memory limit at which there is a good
-///    chance that the Limux OOM killer terminates it. All of this is done is because internal memory tracking in
-///    ClickHouse can unfortunately under-estimate the actually used memory.
-/// 2. Periodically reads the the maximum memory available to the process (which can change due to cgroups settings).
-///    You can specify a callback to react on changes. The callback typically reloads the configuration, i.e. Server
-///    or Keeper configuration file. This reloads settings 'max_server_memory_usage' (Server) and 'max_memory_usage_soft_limit'
-///    (Keeper) from which various other internal limits are calculated, including the soft and hard limits for (1.).
-///    The goal of this is to provide elasticity when the container is scaled-up/scaled-down. The mechanism (polling
-///    cgroups) is quite implicit, unfortunately there is currently no better way to communicate memory threshold changes
-///    to the database.
+///  Periodically reads the the maximum memory available to the process (which can change due to cgroups settings).
+///  You can specify a callback to react on changes. The callback typically reloads the configuration, i.e. Server
+///  or Keeper configuration file. This reloads settings 'max_server_memory_usage' (Server) and 'max_memory_usage_soft_limit'
+///  (Keeper) from which various other internal limits are calculated, including the soft and hard limits for (1.).
+///  The goal of this is to provide elasticity when the container is scaled-up/scaled-down. The mechanism (polling
+///  cgroups) is quite implicit, unfortunately there is currently no better way to communicate memory threshold changes
+///  to the database.
 #if defined(OS_LINUX)
 class CgroupsMemoryUsageObserver
 {
 public:
-    using OnMemoryLimitFn = std::function<void(bool)>;
     using OnMemoryAmountAvailableChangedFn = std::function<void()>;
 
-    enum class CgroupsVersion : uint8_t
-    {
-        V1,
-        V2
-    };
-
     explicit CgroupsMemoryUsageObserver(std::chrono::seconds wait_time_);
     ~CgroupsMemoryUsageObserver();
 
-    void setMemoryUsageLimits(uint64_t hard_limit_, uint64_t soft_limit_);
     void setOnMemoryAmountAvailableChangedFn(OnMemoryAmountAvailableChangedFn on_memory_amount_available_changed_);
 
     void startThread();
@@ -60,32 +34,22 @@ private:
     const std::chrono::seconds wait_time;
 
     std::mutex limit_mutex;
-    size_t hard_limit TSA_GUARDED_BY(limit_mutex) = 0;
-    size_t soft_limit TSA_GUARDED_BY(limit_mutex) = 0;
-    OnMemoryLimitFn on_hard_limit TSA_GUARDED_BY(limit_mutex);
-    OnMemoryLimitFn on_soft_limit TSA_GUARDED_BY(limit_mutex);
 
     std::mutex memory_amount_available_changed_mutex;
     OnMemoryAmountAvailableChangedFn on_memory_amount_available_changed TSA_GUARDED_BY(memory_amount_available_changed_mutex);
 
-    uint64_t last_memory_usage = 0;        /// how much memory does the process use
     uint64_t last_available_memory_amount; /// how much memory can the process use
 
     void stopThread();
 
     void runThread();
 
-    std::unique_ptr<ICgroupsReader> cgroup_reader;
-
     std::mutex thread_mutex;
     std::condition_variable cond;
     ThreadFromGlobalPool thread;
     bool quit = false;
 };
 
-std::unique_ptr<ICgroupsReader>
-createCgroupsReader(CgroupsMemoryUsageObserver::CgroupsVersion version, const std::filesystem::path & cgroup_path);
-
 #else
 class CgroupsMemoryUsageObserver
 {
diff --git a/src/Common/Jemalloc.cpp b/src/Common/Jemalloc.cpp
index d7cc246db6a..d8ff9268cca 100644
--- a/src/Common/Jemalloc.cpp
+++ b/src/Common/Jemalloc.cpp
@@ -5,7 +5,6 @@
 #include <Common/Exception.h>
 #include <Common/Stopwatch.h>
 #include <Common/logger_useful.h>
-#include <jemalloc/jemalloc.h>
 
 #define STRINGIFY_HELPER(x) #x
 #define STRINGIFY(x) STRINGIFY_HELPER(x)
@@ -26,7 +25,6 @@ namespace ErrorCodes
 
 void purgeJemallocArenas()
 {
-    LOG_TRACE(getLogger("SystemJemalloc"), "Purging unused memory");
     Stopwatch watch;
     mallctl("arena." STRINGIFY(MALLCTL_ARENAS_ALL) ".purge", nullptr, nullptr, nullptr, 0);
     ProfileEvents::increment(ProfileEvents::MemoryAllocatorPurge);
@@ -46,20 +44,6 @@ void checkJemallocProfilingEnabled()
             "set: MALLOC_CONF=background_thread:true,prof:true");
 }
 
-template <typename T>
-void setJemallocValue(const char * name, T value)
-{
-    T old_value;
-    size_t old_value_size = sizeof(T);
-    if (mallctl(name, &old_value, &old_value_size, reinterpret_cast<void*>(&value), sizeof(T)))
-    {
-        LOG_WARNING(getLogger("Jemalloc"), "mallctl for {} failed", name);
-        return;
-    }
-
-    LOG_INFO(getLogger("Jemalloc"), "Value for {} set to {} (from {})", name, value, old_value);
-}
-
 void setJemallocProfileActive(bool value)
 {
     checkJemallocProfilingEnabled();
diff --git a/src/Common/Jemalloc.h b/src/Common/Jemalloc.h
index 499a906fd3d..22a94a44eba 100644
--- a/src/Common/Jemalloc.h
+++ b/src/Common/Jemalloc.h
@@ -5,6 +5,8 @@
 #if USE_JEMALLOC
 
 #include <string>
+#include <Common/logger_useful.h>
+#include <jemalloc/jemalloc.h>
 
 namespace DB
 {
@@ -21,6 +23,59 @@ void setJemallocBackgroundThreads(bool enabled);
 
 void setJemallocMaxBackgroundThreads(size_t max_threads);
 
+template <typename T>
+void setJemallocValue(const char * name, T value)
+{
+    T old_value;
+    size_t old_value_size = sizeof(T);
+    mallctl(name, &old_value, &old_value_size, reinterpret_cast<void*>(&value), sizeof(T));
+    LOG_INFO(getLogger("Jemalloc"), "Value for {} set to {} (from {})", name, value, old_value);
+}
+
+template <typename T>
+T getJemallocValue(const char * name)
+{
+    T value;
+    size_t value_size = sizeof(T);
+    mallctl(name, &value, &value_size, nullptr, 0);
+    return value;
+}
+
+/// Each mallctl call consists of string name lookup which can be expensive.
+/// This can be avoided by translating name to "Management Information Base" (MIB)
+/// and using it in mallctlbymib calls
+template <typename T>
+struct JemallocMibCache
+{
+    explicit JemallocMibCache(const char * name)
+    {
+        mallctlnametomib(name, mib, &mib_length);
+    }
+
+    void setValue(T value)
+    {
+        mallctlbymib(mib, mib_length, nullptr, nullptr, reinterpret_cast<void*>(&value), sizeof(T));
+    }
+
+    T getValue()
+    {
+        T value;
+        size_t value_size = sizeof(T);
+        mallctlbymib(mib, mib_length, &value, &value_size, nullptr, 0);
+        return value;
+    }
+
+    void run()
+    {
+        mallctlbymib(mib, mib_length, nullptr, nullptr, nullptr, 0);
+    }
+
+private:
+    static constexpr size_t max_mib_length = 4;
+    size_t mib[max_mib_length];
+    size_t mib_length = max_mib_length;
+};
+
 }
 
 #endif
diff --git a/src/Common/MemoryTracker.cpp b/src/Common/MemoryTracker.cpp
index 7c0115467c6..7bf665ea7a0 100644
--- a/src/Common/MemoryTracker.cpp
+++ b/src/Common/MemoryTracker.cpp
@@ -20,13 +20,9 @@
 #if USE_JEMALLOC
 #    include <jemalloc/jemalloc.h>
 
-#define STRINGIFY_HELPER(x) #x
-#define STRINGIFY(x) STRINGIFY_HELPER(x)
-
 #endif
 
 #include <atomic>
-#include <cmath>
 #include <random>
 #include <cstdlib>
 #include <string>
@@ -115,8 +111,6 @@ void AllocationTrace::onFreeImpl(void * ptr, size_t size) const
 namespace ProfileEvents
 {
     extern const Event QueryMemoryLimitExceeded;
-    extern const Event MemoryAllocatorPurge;
-    extern const Event MemoryAllocatorPurgeTimeMicroseconds;
 }
 
 using namespace std::chrono_literals;
@@ -126,15 +120,13 @@ static constexpr size_t log_peak_memory_usage_every = 1ULL << 30;
 MemoryTracker total_memory_tracker(nullptr, VariableContext::Global);
 MemoryTracker background_memory_tracker(&total_memory_tracker, VariableContext::User, false);
 
-std::atomic<Int64> MemoryTracker::free_memory_in_allocator_arenas;
-
 MemoryTracker::MemoryTracker(VariableContext level_) : parent(&total_memory_tracker), level(level_) {}
 MemoryTracker::MemoryTracker(MemoryTracker * parent_, VariableContext level_) : parent(parent_), level(level_) {}
+
 MemoryTracker::MemoryTracker(MemoryTracker * parent_, VariableContext level_, bool log_peak_memory_usage_in_destructor_)
-    : parent(parent_)
-    , log_peak_memory_usage_in_destructor(log_peak_memory_usage_in_destructor_)
-    , level(level_)
-{}
+    : parent(parent_), log_peak_memory_usage_in_destructor(log_peak_memory_usage_in_destructor_), level(level_)
+{
+}
 
 MemoryTracker::~MemoryTracker()
 {
@@ -204,10 +196,14 @@ void MemoryTracker::debugLogBigAllocationWithoutCheck(Int64 size [[maybe_unused]
         return;
 
     MemoryTrackerBlockerInThread blocker(VariableContext::Global);
-    LOG_TEST(getLogger("MemoryTracker"), "Too big allocation ({} bytes) without checking memory limits, "
-                                                   "it may lead to OOM. Stack trace: {}", size, StackTrace().toString());
+    LOG_TEST(
+        getLogger("MemoryTracker"),
+        "Too big allocation ({} bytes) without checking memory limits, "
+        "it may lead to OOM. Stack trace: {}",
+        size,
+        StackTrace().toString());
 #else
-    return;     /// Avoid trash logging in release builds
+    /// Avoid trash logging in release builds
 #endif
 }
 
@@ -228,6 +224,7 @@ AllocationTrace MemoryTracker::allocImpl(Int64 size, bool throw_if_memory_exceed
         {
             /// For global memory tracker always update memory usage.
             amount.fetch_add(size, std::memory_order_relaxed);
+            rss.fetch_add(size, std::memory_order_relaxed);
 
             auto metric_loaded = metric.load(std::memory_order_relaxed);
             if (metric_loaded != CurrentMetrics::end())
@@ -249,6 +246,7 @@ AllocationTrace MemoryTracker::allocImpl(Int64 size, bool throw_if_memory_exceed
       * So, we allow over-allocations.
       */
     Int64 will_be = size ? size + amount.fetch_add(size, std::memory_order_relaxed) : amount.load(std::memory_order_relaxed);
+    Int64 will_be_rss = size ? size + rss.fetch_add(size, std::memory_order_relaxed) : rss.load(std::memory_order_relaxed);
 
     auto metric_loaded = metric.load(std::memory_order_relaxed);
     if (metric_loaded != CurrentMetrics::end() && size)
@@ -275,6 +273,7 @@ AllocationTrace MemoryTracker::allocImpl(Int64 size, bool throw_if_memory_exceed
         {
             /// Revert
             amount.fetch_sub(size, std::memory_order_relaxed);
+            rss.fetch_sub(size, std::memory_order_relaxed);
 
             /// Prevent recursion. Exception::ctor -> std::string -> new[] -> MemoryTracker::alloc
             MemoryTrackerBlockerInThread untrack_lock(VariableContext::Global);
@@ -297,33 +296,8 @@ AllocationTrace MemoryTracker::allocImpl(Int64 size, bool throw_if_memory_exceed
         }
     }
 
-    Int64 limit_to_check = current_hard_limit;
-
-#if USE_JEMALLOC
-    if (level == VariableContext::Global && allow_use_jemalloc_memory.load(std::memory_order_relaxed))
-    {
-        /// Jemalloc arenas may keep some extra memory.
-        /// This memory was substucted from RSS to decrease memory drift.
-        /// In case memory is close to limit, try to pugre the arenas.
-        /// This is needed to avoid OOM, because some allocations are directly done with mmap.
-        Int64 current_free_memory_in_allocator_arenas = free_memory_in_allocator_arenas.load(std::memory_order_relaxed);
-
-        if (current_free_memory_in_allocator_arenas > 0 && current_hard_limit && current_free_memory_in_allocator_arenas + will_be > current_hard_limit)
-        {
-            if (free_memory_in_allocator_arenas.exchange(-current_free_memory_in_allocator_arenas) > 0)
-            {
-                Stopwatch watch;
-                mallctl("arena." STRINGIFY(MALLCTL_ARENAS_ALL) ".purge", nullptr, nullptr, nullptr, 0);
-                ProfileEvents::increment(ProfileEvents::MemoryAllocatorPurge);
-                ProfileEvents::increment(ProfileEvents::MemoryAllocatorPurgeTimeMicroseconds, watch.elapsedMicroseconds());
-            }
-        }
-
-        limit_to_check += abs(current_free_memory_in_allocator_arenas);
-    }
-#endif
-
-    if (unlikely(current_hard_limit && will_be > limit_to_check))
+    if (unlikely(
+            current_hard_limit && (will_be > current_hard_limit || (level == VariableContext::Global && will_be_rss > current_hard_limit))))
     {
         if (memoryTrackerCanThrow(level, false) && throw_if_memory_exceeded)
         {
@@ -335,6 +309,7 @@ AllocationTrace MemoryTracker::allocImpl(Int64 size, bool throw_if_memory_exceed
             {
                 /// Revert
                 amount.fetch_sub(size, std::memory_order_relaxed);
+                rss.fetch_sub(size, std::memory_order_relaxed);
 
                 /// Prevent recursion. Exception::ctor -> std::string -> new[] -> MemoryTracker::alloc
                 MemoryTrackerBlockerInThread untrack_lock(VariableContext::Global);
@@ -343,12 +318,13 @@ AllocationTrace MemoryTracker::allocImpl(Int64 size, bool throw_if_memory_exceed
                 throw DB::Exception(
                                     DB::ErrorCodes::MEMORY_LIMIT_EXCEEDED,
                                     "Memory limit{}{} exceeded: "
-                                    "would use {} (attempt to allocate chunk of {} bytes), maximum: {}."
+                                    "would use {} (attempt to allocate chunk of {} bytes), current RSS {}, maximum: {}."
                                     "{}{}",
                                     description ? " " : "",
                                     description ? description : "",
                                     formatReadableSizeWithBinarySuffix(will_be),
                                     size,
+                                    formatReadableSizeWithBinarySuffix(rss.load(std::memory_order_relaxed)),
                                     formatReadableSizeWithBinarySuffix(current_hard_limit),
                                     overcommit_result == OvercommitResult::NONE ? "" : " OvercommitTracker decision: ",
                                     toDescription(overcommit_result));
@@ -442,6 +418,7 @@ AllocationTrace MemoryTracker::free(Int64 size, double _sample_probability)
         {
             /// For global memory tracker always update memory usage.
             amount.fetch_sub(size, std::memory_order_relaxed);
+            rss.fetch_sub(size, std::memory_order_relaxed);
             auto metric_loaded = metric.load(std::memory_order_relaxed);
             if (metric_loaded != CurrentMetrics::end())
                 CurrentMetrics::sub(metric_loaded, size);
@@ -455,7 +432,12 @@ AllocationTrace MemoryTracker::free(Int64 size, double _sample_probability)
     }
 
     Int64 accounted_size = size;
-    if (level == VariableContext::Thread || level == VariableContext::Global)
+    if (level == VariableContext::Global)
+    {
+        amount.fetch_sub(accounted_size, std::memory_order_relaxed);
+        rss.fetch_sub(accounted_size, std::memory_order_relaxed);
+    }
+    else if (level == VariableContext::Thread)
     {
         /// Could become negative if memory allocated in this thread is freed in another one
         amount.fetch_sub(accounted_size, std::memory_order_relaxed);
@@ -529,21 +511,29 @@ void MemoryTracker::reset()
 }
 
 
-void MemoryTracker::setRSS(Int64 rss_, Int64 free_memory_in_allocator_arenas_)
+void MemoryTracker::updateRSS(Int64 rss_)
 {
-    Int64 new_amount = rss_;
+    total_memory_tracker.rss.store(rss_, std::memory_order_relaxed);
+}
+
+void MemoryTracker::updateAllocated(Int64 allocated_)
+{
+    Int64 new_amount = allocated_;
+    LOG_INFO(
+        getLogger("MemoryTracker"),
+        "Correcting the value of global memory tracker from {} to {}",
+        ReadableSize(total_memory_tracker.amount.load(std::memory_order_relaxed)),
+        ReadableSize(allocated_));
     total_memory_tracker.amount.store(new_amount, std::memory_order_relaxed);
-    free_memory_in_allocator_arenas.store(free_memory_in_allocator_arenas_, std::memory_order_relaxed);
 
     auto metric_loaded = total_memory_tracker.metric.load(std::memory_order_relaxed);
     if (metric_loaded != CurrentMetrics::end())
         CurrentMetrics::set(metric_loaded, new_amount);
 
     bool log_memory_usage = true;
-    total_memory_tracker.updatePeak(rss_, log_memory_usage);
+    total_memory_tracker.updatePeak(new_amount, log_memory_usage);
 }
 
-
 void MemoryTracker::setSoftLimit(Int64 value)
 {
     soft_limit.store(value, std::memory_order_relaxed);
diff --git a/src/Common/MemoryTracker.h b/src/Common/MemoryTracker.h
index fd32b631774..f15465a20c1 100644
--- a/src/Common/MemoryTracker.h
+++ b/src/Common/MemoryTracker.h
@@ -2,7 +2,6 @@
 
 #include <atomic>
 #include <chrono>
-#include <optional>
 #include <base/types.h>
 #include <Common/CurrentMetrics.h>
 #include <Common/VariableContext.h>
@@ -57,9 +56,8 @@ private:
     std::atomic<Int64> soft_limit {0};
     std::atomic<Int64> hard_limit {0};
     std::atomic<Int64> profiler_limit {0};
-    std::atomic_bool allow_use_jemalloc_memory {true};
 
-    static std::atomic<Int64> free_memory_in_allocator_arenas;
+    std::atomic<Int64> rss{0};
 
     Int64 profiler_step = 0;
 
@@ -122,6 +120,11 @@ public:
         return amount.load(std::memory_order_relaxed);
     }
 
+    Int64 getRSS() const
+    {
+        return rss.load(std::memory_order_relaxed);
+    }
+
     // Merges and mutations may pass memory ownership to other threads thus in the end of execution
     // MemoryTracker for background task may have a non-zero counter.
     // This method is intended to fix the counter inside of background_memory_tracker.
@@ -154,14 +157,6 @@ public:
     {
         return soft_limit.load(std::memory_order_relaxed);
     }
-    void setAllowUseJemallocMemory(bool value)
-    {
-        allow_use_jemalloc_memory.store(value, std::memory_order_relaxed);
-    }
-    bool getAllowUseJemallocMmemory() const
-    {
-        return allow_use_jemalloc_memory.load(std::memory_order_relaxed);
-    }
 
     /** Set limit if it was not set.
       * Otherwise, set limit to new value, if new value is greater than previous limit.
@@ -249,10 +244,9 @@ public:
     /// Reset the accumulated data.
     void reset();
 
-    /// Reset current counter to an RSS value.
-    /// Jemalloc may have pre-allocated arenas, they are accounted in RSS.
-    /// We can free this arenas in case of exception to avoid OOM.
-    static void setRSS(Int64 rss_, Int64 free_memory_in_allocator_arenas_);
+    /// update values based on external information (e.g. jemalloc's stat)
+    static void updateRSS(Int64 rss_);
+    static void updateAllocated(Int64 allocated_);
 
     /// Prints info about peak memory consumption into log.
     void logPeakMemoryUsage();
diff --git a/src/Common/MemoryWorker.cpp b/src/Common/MemoryWorker.cpp
new file mode 100644
index 00000000000..11f3bff348c
--- /dev/null
+++ b/src/Common/MemoryWorker.cpp
@@ -0,0 +1,333 @@
+#include <Common/MemoryWorker.h>
+
+#include <IO/ReadBufferFromFile.h>
+#include <IO/ReadBufferFromFileDescriptor.h>
+#include <IO/ReadHelpers.h>
+#include <base/cgroupsv2.h>
+#include <Common/Jemalloc.h>
+#include <Common/MemoryTracker.h>
+#include <Common/ProfileEvents.h>
+#include <Common/formatReadable.h>
+#include <Common/logger_useful.h>
+
+#include <fmt/ranges.h>
+
+#include <filesystem>
+#include <optional>
+
+namespace fs = std::filesystem;
+
+namespace ProfileEvents
+{
+    extern const Event MemoryAllocatorPurge;
+    extern const Event MemoryAllocatorPurgeTimeMicroseconds;
+    extern const Event MemoryWorkerRun;
+    extern const Event MemoryWorkerRunElapsedMicroseconds;
+}
+
+namespace DB
+{
+
+namespace ErrorCodes
+{
+    extern const int FILE_DOESNT_EXIST;
+    extern const int LOGICAL_ERROR;
+}
+
+#if defined(OS_LINUX)
+namespace
+{
+
+using Metrics = std::map<std::string, uint64_t>;
+
+/// Format is
+///   kernel 5
+///   rss 15
+///   [...]
+Metrics readAllMetricsFromStatFile(ReadBufferFromFile & buf)
+{
+    Metrics metrics;
+    while (!buf.eof())
+    {
+        std::string current_key;
+        readStringUntilWhitespace(current_key, buf);
+
+        assertChar(' ', buf);
+
+        uint64_t value = 0;
+        readIntText(value, buf);
+        assertChar('\n', buf);
+
+        auto [_, inserted] = metrics.emplace(std::move(current_key), value);
+        chassert(inserted, "Duplicate keys in stat file");
+    }
+    return metrics;
+}
+
+uint64_t readMetricFromStatFile(ReadBufferFromFile & buf, std::string_view key)
+{
+    while (!buf.eof())
+    {
+        std::string current_key;
+        readStringUntilWhitespace(current_key, buf);
+        if (current_key != key)
+        {
+            std::string dummy;
+            readStringUntilNewlineInto(dummy, buf);
+            buf.ignore();
+            continue;
+        }
+
+        assertChar(' ', buf);
+        uint64_t value = 0;
+        readIntText(value, buf);
+        return value;
+    }
+    LOG_ERROR(getLogger("CgroupsReader"), "Cannot find '{}' in '{}'", key, buf.getFileName());
+    return 0;
+}
+
+struct CgroupsV1Reader : ICgroupsReader
+{
+    explicit CgroupsV1Reader(const fs::path & stat_file_dir) : buf(stat_file_dir / "memory.stat") { }
+
+    uint64_t readMemoryUsage() override
+    {
+        std::lock_guard lock(mutex);
+        buf.rewind();
+        return readMetricFromStatFile(buf, "rss");
+    }
+
+    std::string dumpAllStats() override
+    {
+        std::lock_guard lock(mutex);
+        buf.rewind();
+        return fmt::format("{}", readAllMetricsFromStatFile(buf));
+    }
+
+private:
+    std::mutex mutex;
+    ReadBufferFromFile buf TSA_GUARDED_BY(mutex);
+};
+
+struct CgroupsV2Reader : ICgroupsReader
+{
+    explicit CgroupsV2Reader(const fs::path & stat_file_dir) : stat_buf(stat_file_dir / "memory.stat") { }
+
+    uint64_t readMemoryUsage() override
+    {
+        std::lock_guard lock(mutex);
+        stat_buf.rewind();
+        return readMetricFromStatFile(stat_buf, "anon");
+    }
+
+    std::string dumpAllStats() override
+    {
+        std::lock_guard lock(mutex);
+        stat_buf.rewind();
+        return fmt::format("{}", readAllMetricsFromStatFile(stat_buf));
+    }
+
+private:
+    std::mutex mutex;
+    ReadBufferFromFile stat_buf TSA_GUARDED_BY(mutex);
+};
+
+/// Caveats:
+/// - All of the logic in this file assumes that the current process is the only process in the
+///   containing cgroup (or more precisely: the only process with significant memory consumption).
+///   If this is not the case, then other processe's memory consumption may affect the internal
+///   memory tracker ...
+/// - Cgroups v1 and v2 allow nested cgroup hierarchies. As v1 is deprecated for over half a
+///   decade and will go away at some point, hierarchical detection is only implemented for v2.
+/// - I did not test what happens if a host has v1 and v2 simultaneously enabled. I believe such
+///   systems existed only for a short transition period.
+
+std::optional<std::string> getCgroupsV1Path()
+{
+    auto path = default_cgroups_mount / "memory/memory.stat";
+    if (!fs::exists(path))
+        return {};
+    return {default_cgroups_mount / "memory"};
+}
+
+std::pair<std::string, ICgroupsReader::CgroupsVersion> getCgroupsPath()
+{
+    auto v2_path = getCgroupsV2PathContainingFile("memory.current");
+    if (v2_path.has_value())
+        return {*v2_path, ICgroupsReader::CgroupsVersion::V2};
+
+    auto v1_path = getCgroupsV1Path();
+    if (v1_path.has_value())
+        return {*v1_path, ICgroupsReader::CgroupsVersion::V1};
+
+    throw Exception(ErrorCodes::FILE_DOESNT_EXIST, "Cannot find cgroups v1 or v2 current memory file");
+}
+
+}
+
+std::shared_ptr<ICgroupsReader> ICgroupsReader::createCgroupsReader(ICgroupsReader::CgroupsVersion version, const std::filesystem::path & cgroup_path)
+{
+    if (version == CgroupsVersion::V2)
+        return std::make_shared<CgroupsV2Reader>(cgroup_path);
+    else
+    {
+        chassert(version == CgroupsVersion::V1);
+        return std::make_shared<CgroupsV1Reader>(cgroup_path);
+    }
+}
+#endif
+
+namespace
+{
+
+std::string_view sourceToString(MemoryWorker::MemoryUsageSource source)
+{
+    switch (source)
+    {
+        case MemoryWorker::MemoryUsageSource::Cgroups: return "Cgroups";
+        case MemoryWorker::MemoryUsageSource::Jemalloc: return "Jemalloc";
+        case MemoryWorker::MemoryUsageSource::None: return "None";
+    }
+}
+
+}
+
+/// We try to pick the best possible supported source for reading memory usage.
+/// Supported sources in order of priority
+/// - reading from cgroups' pseudo-files (fastest and most accurate)
+/// - reading jemalloc's resident stat (doesn't take into account allocations that didn't use jemalloc)
+/// Also, different tick rates are used because not all options are equally fast
+MemoryWorker::MemoryWorker(uint64_t period_ms_)
+    : log(getLogger("MemoryWorker"))
+    , period_ms(period_ms_)
+{
+#if defined(OS_LINUX)
+    try
+    {
+        static constexpr uint64_t cgroups_memory_usage_tick_ms{50};
+
+        const auto [cgroup_path, version] = getCgroupsPath();
+        LOG_INFO(
+            getLogger("CgroupsReader"),
+            "Will create cgroup reader from '{}' (cgroups version: {})",
+            cgroup_path,
+            (version == ICgroupsReader::CgroupsVersion::V1) ? "v1" : "v2");
+
+        cgroups_reader = ICgroupsReader::createCgroupsReader(version, cgroup_path);
+        source = MemoryUsageSource::Cgroups;
+        if (period_ms == 0)
+            period_ms = cgroups_memory_usage_tick_ms;
+
+        return;
+    }
+    catch (...)
+    {
+        tryLogCurrentException(log, "Cannot use cgroups reader");
+    }
+#endif
+
+#if USE_JEMALLOC
+    static constexpr uint64_t jemalloc_memory_usage_tick_ms{100};
+
+    source = MemoryUsageSource::Jemalloc;
+    if (period_ms == 0)
+        period_ms = jemalloc_memory_usage_tick_ms;
+#endif
+}
+
+MemoryWorker::MemoryUsageSource MemoryWorker::getSource()
+{
+    return source;
+}
+
+void MemoryWorker::start()
+{
+    if (source == MemoryUsageSource::None)
+        return;
+
+    LOG_INFO(
+        getLogger("MemoryWorker"),
+        "Starting background memory thread with period of {}ms, using {} as source",
+        period_ms,
+        sourceToString(source));
+    background_thread = ThreadFromGlobalPool([this] { backgroundThread(); });
+}
+
+MemoryWorker::~MemoryWorker()
+{
+    {
+        std::unique_lock lock(mutex);
+        shutdown = true;
+    }
+    cv.notify_all();
+
+    if (background_thread.joinable())
+        background_thread.join();
+}
+
+uint64_t MemoryWorker::getMemoryUsage()
+{
+    switch (source)
+    {
+        case MemoryUsageSource::Cgroups:
+            return cgroups_reader != nullptr ? cgroups_reader->readMemoryUsage() : 0;
+        case MemoryUsageSource::Jemalloc:
+#if USE_JEMALLOC
+            return resident_mib.getValue();
+#else
+            return 0;
+#endif
+        case MemoryUsageSource::None:
+            throw DB::Exception(ErrorCodes::LOGICAL_ERROR, "Trying to fetch memory usage while no memory source can be used");
+    }
+}
+
+void MemoryWorker::backgroundThread()
+{
+    std::chrono::milliseconds chrono_period_ms{period_ms};
+    [[maybe_unused]] bool first_run = true;
+    std::unique_lock lock(mutex);
+    while (true)
+    {
+        cv.wait_for(lock, chrono_period_ms, [this] { return shutdown; });
+        if (shutdown)
+            return;
+
+        Stopwatch total_watch;
+
+#if USE_JEMALLOC
+        if (source == MemoryUsageSource::Jemalloc)
+            epoch_mib.setValue(0);
+#endif
+
+        Int64 resident = getMemoryUsage();
+        MemoryTracker::updateRSS(resident);
+
+#if USE_JEMALLOC
+        if (resident > total_memory_tracker.getHardLimit())
+        {
+            Stopwatch purge_watch;
+            purge_mib.run();
+            ProfileEvents::increment(ProfileEvents::MemoryAllocatorPurge);
+            ProfileEvents::increment(ProfileEvents::MemoryAllocatorPurgeTimeMicroseconds, purge_watch.elapsedMicroseconds());
+        }
+#endif
+
+#if USE_JEMALLOC
+        if (unlikely(first_run || total_memory_tracker.get() < 0))
+        {
+            if (source != MemoryUsageSource::Jemalloc)
+                epoch_mib.setValue(0);
+
+            MemoryTracker::updateAllocated(allocated_mib.getValue());
+        }
+#endif
+
+        ProfileEvents::increment(ProfileEvents::MemoryWorkerRun);
+        ProfileEvents::increment(ProfileEvents::MemoryWorkerRunElapsedMicroseconds, total_watch.elapsedMicroseconds());
+        first_run = false;
+    }
+}
+
+}
diff --git a/src/Common/MemoryWorker.h b/src/Common/MemoryWorker.h
new file mode 100644
index 00000000000..f4b0fed23ec
--- /dev/null
+++ b/src/Common/MemoryWorker.h
@@ -0,0 +1,84 @@
+#pragma once
+
+#include <Common/CgroupsMemoryUsageObserver.h>
+#include <Common/ThreadPool.h>
+#include <Common/Jemalloc.h>
+
+namespace DB
+{
+
+struct ICgroupsReader
+{
+    enum class CgroupsVersion : uint8_t
+    {
+        V1,
+        V2
+    };
+
+#if defined(OS_LINUX)
+    static std::shared_ptr<ICgroupsReader>
+    createCgroupsReader(ICgroupsReader::CgroupsVersion version, const std::filesystem::path & cgroup_path);
+#endif
+
+    virtual ~ICgroupsReader() = default;
+
+    virtual uint64_t readMemoryUsage() = 0;
+
+    virtual std::string dumpAllStats() = 0;
+};
+
+
+/// Correct MemoryTracker based on external information (e.g. Cgroups or stats.resident from jemalloc)
+/// The worker spawns a background thread which periodically reads current resident memory from the source,
+/// whose value is sent to global MemoryTracker.
+/// It can do additional things like purging jemalloc dirty pages if the current memory usage is higher than global hard limit.
+class MemoryWorker
+{
+public:
+    explicit MemoryWorker(uint64_t period_ms_);
+
+    enum class MemoryUsageSource : uint8_t
+    {
+        None,
+        Cgroups,
+        Jemalloc
+    };
+
+    MemoryUsageSource getSource();
+
+    void start();
+
+    ~MemoryWorker();
+private:
+    uint64_t getMemoryUsage();
+
+    void backgroundThread();
+
+    ThreadFromGlobalPool background_thread;
+
+    std::mutex mutex;
+    std::condition_variable cv;
+    bool shutdown = false;
+
+    LoggerPtr log;
+
+    uint64_t period_ms;
+
+    MemoryUsageSource source{MemoryUsageSource::None};
+
+    std::shared_ptr<ICgroupsReader> cgroups_reader;
+
+#if USE_JEMALLOC
+    JemallocMibCache<uint64_t> epoch_mib{"epoch"};
+    JemallocMibCache<size_t> resident_mib{"stats.resident"};
+    JemallocMibCache<size_t> allocated_mib{"stats.allocated"};
+
+#define STRINGIFY_HELPER(x) #x
+#define STRINGIFY(x) STRINGIFY_HELPER(x)
+    JemallocMibCache<size_t> purge_mib{"arena." STRINGIFY(MALLCTL_ARENAS_ALL) ".purge"};
+#undef STRINGIFY
+#undef STRINGIFY_HELPER
+#endif
+};
+
+}
diff --git a/src/Common/PoolWithFailoverBase.h b/src/Common/PoolWithFailoverBase.h
index 53a746c316e..989ffd888f8 100644
--- a/src/Common/PoolWithFailoverBase.h
+++ b/src/Common/PoolWithFailoverBase.h
@@ -122,12 +122,18 @@ public:
         return result.entry.isNull() || !result.is_usable || (skip_read_only_replicas && result.is_readonly);
     }
 
-    void checkTryResultIsValid(const TryResult & result, bool skip_read_only_replicas) const
+    TryResult getValidTryResult(const std::vector<TryResult> & results, bool skip_read_only_replicas) const
     {
+        if (results.empty())
+            throw DB::Exception(DB::ErrorCodes::ALL_CONNECTION_TRIES_FAILED, "Cannot get any valid connection because all connection tries failed");
+
+        auto result = results.front();
         if (isTryResultInvalid(result, skip_read_only_replicas))
             throw DB::Exception(DB::ErrorCodes::LOGICAL_ERROR,
                 "Got an invalid connection result: entry.isNull {}, is_usable {}, is_up_to_date {}, delay {}, is_readonly {}, skip_read_only_replicas {}",
                 result.entry.isNull(), result.is_usable, result.is_up_to_date, result.delay, result.is_readonly, skip_read_only_replicas);
+
+        return result;
     }
 
     size_t getPoolSize() const { return nested_pools.size(); }
diff --git a/src/Common/ProfileEvents.cpp b/src/Common/ProfileEvents.cpp
index af1b7fbeb4a..467dfe60cd7 100644
--- a/src/Common/ProfileEvents.cpp
+++ b/src/Common/ProfileEvents.cpp
@@ -827,6 +827,9 @@ The server successfully detected this situation and will download merged part fr
     M(GWPAsanAllocateSuccess, "Number of successful allocations done by GWPAsan") \
     M(GWPAsanAllocateFailed, "Number of failed allocations done by GWPAsan (i.e. filled pool)") \
     M(GWPAsanFree, "Number of free operations done by GWPAsan") \
+    \
+    M(MemoryWorkerRun, "Number of runs done by MemoryWorker in background") \
+    M(MemoryWorkerRunElapsedMicroseconds, "Total time spent by MemoryWorker for background work") \
 
 
 #ifdef APPLY_FOR_EXTERNAL_EVENTS
diff --git a/src/Common/StackTrace.cpp b/src/Common/StackTrace.cpp
index bd01b639913..3dce34803b2 100644
--- a/src/Common/StackTrace.cpp
+++ b/src/Common/StackTrace.cpp
@@ -67,10 +67,18 @@ std::string SigsegvErrorString(const siginfo_t & info, [[maybe_unused]] const uc
         = info.si_addr == nullptr ? "NULL pointer"s : (shouldShowAddress(info.si_addr) ? fmt::format("{}", info.si_addr) : ""s);
 
     const std::string_view access =
-#if defined(__x86_64__) && !defined(OS_FREEBSD) && !defined(OS_DARWIN) && !defined(__arm__) && !defined(__powerpc__)
-        (context.uc_mcontext.gregs[REG_ERR] & 0x02) ? "write" : "read";
+#if defined(__arm__)
+        "<not available on ARM>";
+#elif defined(__powerpc__)
+        "<not available on PowerPC>";
+#elif defined(OS_DARWIN)
+        "<not available on Darwin>";
+#elif defined(OS_FREEBSD)
+        "<not available on FreeBSD>";
+#elif !defined(__x86_64__)
+        "<not available>";
 #else
-        "";
+        (context.uc_mcontext.gregs[REG_ERR] & 0x02) ? "write" : "read";
 #endif
 
     std::string_view message;
diff --git a/src/Common/tests/gtest_cgroups_reader.cpp b/src/Common/tests/gtest_cgroups_reader.cpp
index 2de25bb42ce..e24b91a59b8 100644
--- a/src/Common/tests/gtest_cgroups_reader.cpp
+++ b/src/Common/tests/gtest_cgroups_reader.cpp
@@ -6,7 +6,7 @@
 #include <filesystem>
 
 #include <IO/WriteBufferFromFile.h>
-#include <Common/CgroupsMemoryUsageObserver.h>
+#include <Common/MemoryWorker.h>
 #include <Common/filesystemHelpers.h>
 
 using namespace DB;
@@ -126,7 +126,7 @@ const std::string EXPECTED[2]
        "\"workingset_restore_anon\": 0, \"workingset_restore_file\": 0, \"zswap\": 0, \"zswapped\": 0, \"zswpin\": 0, \"zswpout\": 0}"};
 
 
-class CgroupsMemoryUsageObserverFixture : public ::testing::TestWithParam<CgroupsMemoryUsageObserver::CgroupsVersion>
+class CgroupsMemoryUsageObserverFixture : public ::testing::TestWithParam<ICgroupsReader::CgroupsVersion>
 {
     void SetUp() override
     {
@@ -138,7 +138,7 @@ class CgroupsMemoryUsageObserverFixture : public ::testing::TestWithParam<Cgroup
         stat_file.write(SAMPLE_FILE[version].data(), SAMPLE_FILE[version].size());
         stat_file.sync();
 
-        if (GetParam() == CgroupsMemoryUsageObserver::CgroupsVersion::V2)
+        if (GetParam() == ICgroupsReader::CgroupsVersion::V2)
         {
             auto current_file = WriteBufferFromFile(tmp_dir + "/memory.current");
             current_file.write("29645422592", 11);
@@ -154,18 +154,18 @@ protected:
 TEST_P(CgroupsMemoryUsageObserverFixture, ReadMemoryUsageTest)
 {
     const auto version = GetParam();
-    auto reader = createCgroupsReader(version, tmp_dir);
+    auto reader = ICgroupsReader::createCgroupsReader(version, tmp_dir);
     ASSERT_EQ(
         reader->readMemoryUsage(),
-        version == CgroupsMemoryUsageObserver::CgroupsVersion::V1 ? /* rss from memory.stat */ 2232029184
-                                                                  : /* value from memory.current - inactive_file */ 20952338432);
+        version == ICgroupsReader::CgroupsVersion::V1 ? /* rss from memory.stat */ 2232029184
+                                                                  : /* anon from memory.stat */ 10429399040);
 }
 
 
 TEST_P(CgroupsMemoryUsageObserverFixture, DumpAllStatsTest)
 {
     const auto version = GetParam();
-    auto reader = createCgroupsReader(version, tmp_dir);
+    auto reader = ICgroupsReader::createCgroupsReader(version, tmp_dir);
     ASSERT_EQ(reader->dumpAllStats(), EXPECTED[static_cast<uint8_t>(version)]);
 }
 
@@ -173,6 +173,6 @@ TEST_P(CgroupsMemoryUsageObserverFixture, DumpAllStatsTest)
 INSTANTIATE_TEST_SUITE_P(
     CgroupsMemoryUsageObserverTests,
     CgroupsMemoryUsageObserverFixture,
-    ::testing::Values(CgroupsMemoryUsageObserver::CgroupsVersion::V1, CgroupsMemoryUsageObserver::CgroupsVersion::V2));
+    ::testing::Values(ICgroupsReader::CgroupsVersion::V1, ICgroupsReader::CgroupsVersion::V2));
 
 #endif
diff --git a/src/Compression/CompressedWriteBuffer.cpp b/src/Compression/CompressedWriteBuffer.cpp
index 83c9fbc9573..c3acfcb7da6 100644
--- a/src/Compression/CompressedWriteBuffer.cpp
+++ b/src/Compression/CompressedWriteBuffer.cpp
@@ -55,10 +55,29 @@ void CompressedWriteBuffer::nextImpl()
 
         out.write(compressed_buffer.data(), compressed_size);
     }
+
+    /// Increase buffer size for next data if adaptive buffer size is used and nextImpl was called because of end of buffer.
+    if (!available() && use_adaptive_buffer_size && memory.size() < adaptive_buffer_max_size)
+    {
+        memory.resize(std::min(memory.size() * 2, adaptive_buffer_max_size));
+        BufferBase::set(memory.data(), memory.size(), 0);
+    }
 }
 
-CompressedWriteBuffer::CompressedWriteBuffer(WriteBuffer & out_, CompressionCodecPtr codec_, size_t buf_size)
-    : BufferWithOwnMemory<WriteBuffer>(buf_size), out(out_), codec(std::move(codec_))
+void CompressedWriteBuffer::finalizeImpl()
+{
+    /// Don't try to resize buffer in nextImpl.
+    use_adaptive_buffer_size = false;
+    next();
+}
+
+CompressedWriteBuffer::CompressedWriteBuffer(
+    WriteBuffer & out_, CompressionCodecPtr codec_, size_t buf_size, bool use_adaptive_buffer_size_, size_t adaptive_buffer_initial_size)
+    : BufferWithOwnMemory<WriteBuffer>(use_adaptive_buffer_size_ ? adaptive_buffer_initial_size : buf_size)
+    , out(out_)
+    , codec(std::move(codec_))
+    , use_adaptive_buffer_size(use_adaptive_buffer_size_)
+    , adaptive_buffer_max_size(buf_size)
 {
 }
 
diff --git a/src/Compression/CompressedWriteBuffer.h b/src/Compression/CompressedWriteBuffer.h
index 6ae1fbee9cc..41596703bfe 100644
--- a/src/Compression/CompressedWriteBuffer.h
+++ b/src/Compression/CompressedWriteBuffer.h
@@ -19,7 +19,9 @@ public:
     explicit CompressedWriteBuffer(
         WriteBuffer & out_,
         CompressionCodecPtr codec_ = CompressionCodecFactory::instance().getDefaultCodec(),
-        size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE);
+        size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE,
+        bool use_adaptive_buffer_size_ = false,
+        size_t adaptive_buffer_initial_size = DBMS_DEFAULT_INITIAL_ADAPTIVE_BUFFER_SIZE);
 
     ~CompressedWriteBuffer() override;
 
@@ -45,10 +47,17 @@ public:
 
 private:
     void nextImpl() override;
+    void finalizeImpl() override;
 
     WriteBuffer & out;
     CompressionCodecPtr codec;
 
+    /// If true, the size of internal buffer will be exponentially increased up to
+    /// adaptive_buffer_max_size after each nextImpl call. It can be used to avoid
+    /// large buffer allocation when actual size of written data is small.
+    bool use_adaptive_buffer_size;
+    size_t adaptive_buffer_max_size;
+
     PODArray<char> compressed_buffer;
 };
 
diff --git a/src/Coordination/KeeperAsynchronousMetrics.cpp b/src/Coordination/KeeperAsynchronousMetrics.cpp
index 86166ffe31b..157858f3c44 100644
--- a/src/Coordination/KeeperAsynchronousMetrics.cpp
+++ b/src/Coordination/KeeperAsynchronousMetrics.cpp
@@ -114,8 +114,13 @@ void updateKeeperInformation(KeeperDispatcher & keeper_dispatcher, AsynchronousM
 }
 
 KeeperAsynchronousMetrics::KeeperAsynchronousMetrics(
-    ContextPtr context_, unsigned update_period_seconds, const ProtocolServerMetricsFunc & protocol_server_metrics_func_)
-    : AsynchronousMetrics(update_period_seconds, protocol_server_metrics_func_), context(std::move(context_))
+    ContextPtr context_,
+    unsigned update_period_seconds,
+    const ProtocolServerMetricsFunc & protocol_server_metrics_func_,
+    bool update_jemalloc_epoch_,
+    bool update_rss_)
+    : AsynchronousMetrics(update_period_seconds, protocol_server_metrics_func_, update_jemalloc_epoch_, update_rss_)
+    , context(std::move(context_))
 {
 }
 
diff --git a/src/Coordination/KeeperAsynchronousMetrics.h b/src/Coordination/KeeperAsynchronousMetrics.h
index ec0e60cbb6e..a2ab7cab756 100644
--- a/src/Coordination/KeeperAsynchronousMetrics.h
+++ b/src/Coordination/KeeperAsynchronousMetrics.h
@@ -13,9 +13,13 @@ class KeeperAsynchronousMetrics : public AsynchronousMetrics
 {
 public:
     KeeperAsynchronousMetrics(
-        ContextPtr context_, unsigned update_period_seconds, const ProtocolServerMetricsFunc & protocol_server_metrics_func_);
-    ~KeeperAsynchronousMetrics() override;
+        ContextPtr context_,
+        unsigned update_period_seconds,
+        const ProtocolServerMetricsFunc & protocol_server_metrics_func_,
+        bool update_jemalloc_epoch_,
+        bool update_rss_);
 
+    ~KeeperAsynchronousMetrics() override;
 private:
     ContextPtr context;
 
diff --git a/src/Coordination/KeeperDispatcher.cpp b/src/Coordination/KeeperDispatcher.cpp
index c91cdb53222..142bd3b7c71 100644
--- a/src/Coordination/KeeperDispatcher.cpp
+++ b/src/Coordination/KeeperDispatcher.cpp
@@ -154,7 +154,14 @@ void KeeperDispatcher::requestThread()
                 Int64 mem_soft_limit = keeper_context->getKeeperMemorySoftLimit();
                 if (configuration_and_settings->standalone_keeper && isExceedingMemorySoftLimit() && checkIfRequestIncreaseMem(request.request))
                 {
-                    LOG_WARNING(log, "Processing requests refused because of max_memory_usage_soft_limit {}, the total used memory is {}, request type is {}", ReadableSize(mem_soft_limit), ReadableSize(total_memory_tracker.get()), request.request->getOpNum());
+                    LOG_WARNING(
+                        log,
+                        "Processing requests refused because of max_memory_usage_soft_limit {}, the total allocated memory is {}, RSS is {}, request type "
+                        "is {}",
+                        ReadableSize(mem_soft_limit),
+                        ReadableSize(total_memory_tracker.get()),
+                        ReadableSize(total_memory_tracker.getRSS()),
+                        request.request->getOpNum());
                     addErrorResponses({request}, Coordination::Error::ZCONNECTIONLOSS);
                     continue;
                 }
diff --git a/src/Coordination/KeeperServer.cpp b/src/Coordination/KeeperServer.cpp
index f09ea56391a..918f24efb2c 100644
--- a/src/Coordination/KeeperServer.cpp
+++ b/src/Coordination/KeeperServer.cpp
@@ -602,7 +602,7 @@ bool KeeperServer::isLeaderAlive() const
 bool KeeperServer::isExceedingMemorySoftLimit() const
 {
     Int64 mem_soft_limit = keeper_context->getKeeperMemorySoftLimit();
-    return mem_soft_limit > 0 && total_memory_tracker.get() >= mem_soft_limit;
+    return mem_soft_limit > 0 && std::max(total_memory_tracker.get(), total_memory_tracker.getRSS()) >= mem_soft_limit;
 }
 
 /// TODO test whether taking failed peer in count
diff --git a/src/Core/Defines.h b/src/Core/Defines.h
index 6df335a9c8f..c6e65f34e90 100644
--- a/src/Core/Defines.h
+++ b/src/Core/Defines.h
@@ -20,6 +20,9 @@ static constexpr auto DBMS_DEFAULT_POLL_INTERVAL = 10;
 /// The size of the I/O buffer by default.
 static constexpr auto DBMS_DEFAULT_BUFFER_SIZE = 1048576ULL;
 
+/// The initial size of adaptive I/O buffer by default.
+static constexpr auto DBMS_DEFAULT_INITIAL_ADAPTIVE_BUFFER_SIZE = 16384ULL;
+
 static constexpr auto PADDING_FOR_SIMD = 64;
 
 /** Which blocks by default read the data (by number of rows).
@@ -40,7 +43,7 @@ static constexpr auto SHOW_CHARS_ON_SYNTAX_ERROR = ptrdiff_t(160);
 /// each period reduces the error counter by 2 times
 /// too short a period can cause errors to disappear immediately after creation.
 static constexpr auto DBMS_CONNECTION_POOL_WITH_FAILOVER_DEFAULT_DECREASE_ERROR_PERIOD = 60;
-/// replica error max cap, this is to prevent replica from accumulating too many errors and taking to long to recover.
+/// replica error max cap, this is to prevent replica from accumulating too many errors and taking too long to recover.
 static constexpr auto DBMS_CONNECTION_POOL_WITH_FAILOVER_MAX_ERROR_COUNT = 1000;
 
 /// The boundary on which the blocks for asynchronous file operations should be aligned.
diff --git a/src/Core/ServerSettings.h b/src/Core/ServerSettings.h
index 79173503f28..18ee096569a 100644
--- a/src/Core/ServerSettings.h
+++ b/src/Core/ServerSettings.h
@@ -148,6 +148,7 @@ namespace DB
     M(Bool, storage_metadata_write_full_object_key, false, "Write disk metadata files with VERSION_FULL_OBJECT_KEY format", 0) \
     M(UInt64, max_materialized_views_count_for_table, 0, "A limit on the number of materialized views attached to a table.", 0) \
     M(UInt32, max_database_replicated_create_table_thread_pool_size, 1, "The number of threads to create tables during replica recovery in DatabaseReplicated. Zero means number of threads equal number of cores.", 0) \
+    M(Bool, database_replicated_allow_detach_permanently, true, "Allow detaching tables permanently in Replicated databases", 0) \
     M(Bool, format_alter_operations_with_parentheses, false, "If enabled, each operation in alter queries will be surrounded with parentheses in formatted queries to make them less ambiguous.", 0) \
     M(String, default_replica_path, "/clickhouse/tables/{uuid}/{shard}", "The path to the table in ZooKeeper", 0) \
     M(String, default_replica_name, "{replica}", "The replica name in ZooKeeper", 0) \
@@ -169,6 +170,7 @@ namespace DB
     M(Bool, prepare_system_log_tables_on_startup, false, "If true, ClickHouse creates all configured `system.*_log` tables before the startup. It can be helpful if some startup scripts depend on these tables.", 0) \
     M(Double, gwp_asan_force_sample_probability, 0.0003, "Probability that an allocation from specific places will be sampled by GWP Asan (i.e. PODArray allocations)", 0) \
     M(UInt64, config_reload_interval_ms, 2000, "How often clickhouse will reload config and check for new changes", 0) \
+    M(UInt64, memory_worker_period_ms, 0, "Tick period of background memory worker which corrects memory tracker memory usages and cleans up unused pages during higher memory usage. If set to 0, default value will be used depending on the memory usage source", 0) \
     M(Bool, disable_insertion_and_mutation, false, "Disable all insert/alter/delete queries. This setting will be enabled if someone needs read-only nodes to prevent insertion and mutation affect reading performance.", 0)
 
 /// If you add a setting which can be updated at runtime, please update 'changeable_settings' map in StorageSystemServerSettings.cpp
diff --git a/src/Core/Settings.h b/src/Core/Settings.h
index 0e58562c745..3a2c7b12b13 100644
--- a/src/Core/Settings.h
+++ b/src/Core/Settings.h
@@ -710,7 +710,8 @@ class IColumn;
     M(UInt64, max_distributed_depth, 5, "Maximum distributed query depth", 0) \
     M(Bool, database_replicated_always_detach_permanently, false, "Execute DETACH TABLE as DETACH TABLE PERMANENTLY if database engine is Replicated", 0) \
     M(Bool, database_replicated_allow_only_replicated_engine, false, "Allow to create only Replicated tables in database with engine Replicated", 0) \
-    M(Bool, database_replicated_allow_replicated_engine_arguments, true, "Allow to create only Replicated tables in database with engine Replicated with explicit arguments", 0) \
+    M(UInt64, database_replicated_allow_replicated_engine_arguments, 0, "0 - Don't allow to explicitly specify ZooKeeper path and replica name for *MergeTree tables in Replicated databases. 1 - Allow. 2 - Allow, but ignore the specified path and use default one instead.", 0) \
+    M(UInt64, database_replicated_allow_explicit_uuid, 0, "0 - Don't allow to explicitly specify UUIDs for tables in Replicated databases. 1 - Allow. 2 - Allow, but ignore the specified UUID and generate a random one instead.", 0) \
     M(Bool, database_replicated_allow_heavy_create, false, "Allow long-running DDL queries (CREATE AS SELECT and POPULATE) in Replicated database engine. Note that it can block DDL queue for a long time.", 0) \
     M(Bool, cloud_mode, false, "Only available in ClickHouse Cloud", 0) \
     M(UInt64, cloud_mode_engine, 1, "Only available in ClickHouse Cloud", 0) \
@@ -922,6 +923,9 @@ class IColumn;
     M(Bool, implicit_transaction, false, "If enabled and not already inside a transaction, wraps the query inside a full transaction (begin + commit or rollback)", 0) \
     M(UInt64, grace_hash_join_initial_buckets, 1, "Initial number of grace hash join buckets", 0) \
     M(UInt64, grace_hash_join_max_buckets, 1024, "Limit on the number of grace hash join buckets", 0) \
+    M(Int32, join_to_sort_minimum_perkey_rows, 40, "The lower limit of per-key average rows in the right table to determine whether to rerange the right table by key in left or inner join. This setting ensures that the optimization is not applied for sparse table keys", 0) \
+    M(Int32, join_to_sort_maximum_table_rows, 10000, "The maximum number of rows in the right table to determine whether to rerange the right table by key in left or inner join.", 0) \
+    M(Bool, allow_experimental_join_right_table_sorting, false, "If it is set to true, and the conditions of `join_to_sort_minimum_perkey_rows` and `join_to_sort_maximum_table_rows` are met, rerange the right table by key to improve the performance in left or inner hash join.", 0) \
     M(Timezone, session_timezone, "", "This setting can be removed in the future due to potential caveats. It is experimental and is not suitable for production usage. The default timezone for current session or query. The server default timezone if empty.", 0) \
     M(Bool, use_hive_partitioning, false, "Allows to use hive partitioning for File, URL, S3, AzureBlobStorage and HDFS engines.", 0)\
     \
@@ -944,6 +948,7 @@ class IColumn;
     M(Bool, parallel_replicas_prefer_local_join, true, "If true, and JOIN can be executed with parallel replicas algorithm, and all storages of right JOIN part are *MergeTree, local JOIN will be used instead of GLOBAL JOIN.", 0) \
     M(UInt64, parallel_replicas_mark_segment_size, 128, "Parts virtually divided into segments to be distributed between replicas for parallel reading. This setting controls the size of these segments. Not recommended to change until you're absolutely sure in what you're doing", 0) \
     M(Bool, allow_archive_path_syntax, true, "File/S3 engines/table function will parse paths with '::' as '<archive> :: <file>' if archive has correct extension", 0) \
+    M(Bool, parallel_replicas_local_plan, false, "Build local plan for local replica", 0) \
     \
     M(Bool, allow_experimental_inverted_index, false, "If it is set to true, allow to use experimental inverted index.", 0) \
     M(Bool, allow_experimental_full_text_index, false, "If it is set to true, allow to use experimental full-text index.", 0) \
diff --git a/src/Core/SettingsChangesHistory.cpp b/src/Core/SettingsChangesHistory.cpp
index 258065dcfd4..fae8c25f5ed 100644
--- a/src/Core/SettingsChangesHistory.cpp
+++ b/src/Core/SettingsChangesHistory.cpp
@@ -76,12 +76,15 @@ static std::initializer_list<std::pair<ClickHouseVersion, SettingsChangesHistory
             {"create_if_not_exists", false, false, "New setting."},
             {"allow_materialized_view_with_bad_select", true, true, "Support (but not enable yet) stricter validation in CREATE MATERIALIZED VIEW"},
             {"output_format_always_quote_identifiers", false, false, "New setting."},
-            {"output_format_identifier_quoting_style", "Backticks", "Backticks", "New setting."}
+            {"output_format_identifier_quoting_style", "Backticks", "Backticks", "New setting."},
+            {"database_replicated_allow_replicated_engine_arguments", 1, 0, "Don't allow explicit arguments by default"},
+            {"database_replicated_allow_explicit_uuid", 0, 0, "Added a new setting to disallow explicitly specifying table UUID"},
+            {"parallel_replicas_local_plan", false, false, "Use local plan for local replica in a query with parallel replicas"},
         }
     },
     {"24.8",
         {
-            {"rows_before_aggregation", true, true, "Provide exact value for rows_before_aggregation statistic, represents the number of rows read before aggregation"},
+            {"rows_before_aggregation", false, false, "Provide exact value for rows_before_aggregation statistic, represents the number of rows read before aggregation"},
             {"restore_replace_external_table_functions_to_null", false, false, "New setting."},
             {"restore_replace_external_engines_to_null", false, false, "New setting."},
             {"input_format_json_max_depth", 1000000, 1000, "It was unlimited in previous versions, but that was unsafe."},
@@ -97,7 +100,10 @@ static std::initializer_list<std::pair<ClickHouseVersion, SettingsChangesHistory
             {"use_json_alias_for_old_object_type", true, false, "Use JSON type alias to create new JSON type"},
             {"type_json_skip_duplicated_paths", false, false, "Allow to skip duplicated paths during JSON parsing"},
             {"allow_experimental_vector_similarity_index", false, false, "Added new setting to allow experimental vector similarity indexes"},
-            {"input_format_try_infer_datetimes_only_datetime64", true, false, "Allow to infer DateTime instead of DateTime64 in data formats"}
+            {"input_format_try_infer_datetimes_only_datetime64", true, false, "Allow to infer DateTime instead of DateTime64 in data formats"},
+            {"join_to_sort_minimum_perkey_rows", 0, 40, "The lower limit of per-key average rows in the right table to determine whether to rerange the right table by key in left or inner join. This setting ensures that the optimization is not applied for sparse table keys"},
+            {"join_to_sort_maximum_table_rows", 0, 10000, "The maximum number of rows in the right table to determine whether to rerange the right table by key in left or inner join."},
+            {"allow_experimental_join_right_table_sorting", false, false, "If it is set to true, and the conditions of `join_to_sort_minimum_perkey_rows` and `join_to_sort_maximum_table_rows` are met, rerange the right table by key to improve the performance in left or inner hash join."},
         }
     },
     {"24.7",
diff --git a/src/DataTypes/Serializations/ISerialization.cpp b/src/DataTypes/Serializations/ISerialization.cpp
index 338edc3a144..dcf637c7d2b 100644
--- a/src/DataTypes/Serializations/ISerialization.cpp
+++ b/src/DataTypes/Serializations/ISerialization.cpp
@@ -420,6 +420,21 @@ bool ISerialization::isEphemeralSubcolumn(const DB::ISerialization::SubstreamPat
     return path[last_elem].type == Substream::VariantElementNullMap;
 }
 
+bool ISerialization::isDynamicSubcolumn(const DB::ISerialization::SubstreamPath & path, size_t prefix_len)
+{
+    if (prefix_len == 0 || prefix_len > path.size())
+        return false;
+
+    for (size_t i = 0; i != prefix_len; ++i)
+    {
+        if (path[i].type == SubstreamType::DynamicData || path[i].type == SubstreamType::DynamicStructure
+            || path[i].type == SubstreamType::ObjectData || path[i].type == SubstreamType::ObjectStructure)
+            return true;
+    }
+
+    return false;
+}
+
 ISerialization::SubstreamData ISerialization::createFromPath(const SubstreamPath & path, size_t prefix_len)
 {
     assert(prefix_len <= path.size());
diff --git a/src/DataTypes/Serializations/ISerialization.h b/src/DataTypes/Serializations/ISerialization.h
index 33575a07177..32f418e9132 100644
--- a/src/DataTypes/Serializations/ISerialization.h
+++ b/src/DataTypes/Serializations/ISerialization.h
@@ -457,6 +457,9 @@ public:
     /// for writing/reading data. For example, it's a null-map subcolumn of Variant type (it's always constructed from discriminators);.
     static bool isEphemeralSubcolumn(const SubstreamPath & path, size_t prefix_len);
 
+    /// Returns true if stream with specified path corresponds to dynamic subcolumn.
+    static bool isDynamicSubcolumn(const SubstreamPath & path, size_t prefix_len);
+
 protected:
     template <typename State, typename StatePtr>
     State * checkAndGetState(const StatePtr & state) const;
diff --git a/src/Databases/DatabaseReplicated.cpp b/src/Databases/DatabaseReplicated.cpp
index 8e3378bcc12..3d64c82ba7d 100644
--- a/src/Databases/DatabaseReplicated.cpp
+++ b/src/Databases/DatabaseReplicated.cpp
@@ -63,6 +63,7 @@ namespace ErrorCodes
     extern const int NO_ACTIVE_REPLICAS;
     extern const int CANNOT_GET_REPLICATED_DATABASE_SNAPSHOT;
     extern const int CANNOT_RESTORE_TABLE;
+    extern const int SUPPORT_IS_DISABLED;
 }
 
 static constexpr const char * REPLICATED_DATABASE_MARK = "DatabaseReplicated";
@@ -441,7 +442,8 @@ void DatabaseReplicated::tryConnectToZooKeeperAndInitDatabase(LoadingStrictnessL
         bool is_create_query = mode == LoadingStrictnessLevel::CREATE;
 
         String replica_host_id;
-        if (current_zookeeper->tryGet(replica_path, replica_host_id))
+        bool replica_exists_in_zk = current_zookeeper->tryGet(replica_path, replica_host_id);
+        if (replica_exists_in_zk)
         {
             if (replica_host_id == DROPPED_MARK && !is_create_query)
             {
@@ -454,7 +456,7 @@ void DatabaseReplicated::tryConnectToZooKeeperAndInitDatabase(LoadingStrictnessL
             String host_id = getHostID(getContext(), db_uuid, cluster_auth_info.cluster_secure_connection);
             String host_id_default = getHostID(getContext(), db_uuid, false);
 
-            if (is_create_query || (replica_host_id != host_id && replica_host_id != host_id_default))
+            if (replica_host_id != host_id && replica_host_id != host_id_default)
             {
                 throw Exception(
                     ErrorCodes::REPLICA_ALREADY_EXISTS,
@@ -484,13 +486,20 @@ void DatabaseReplicated::tryConnectToZooKeeperAndInitDatabase(LoadingStrictnessL
                 current_zookeeper->set(replica_path + "/replica_group", replica_group_name, -1);
                 createEmptyLogEntry(current_zookeeper);
             }
+
+            /// Needed to mark all the queries
+            /// in the range (max log ptr at replica ZooKeeper nodes creation, max log ptr after replica recovery] as successful.
+            String max_log_ptr_at_creation_str;
+            if (current_zookeeper->tryGet(replica_path + "/max_log_ptr_at_creation", max_log_ptr_at_creation_str))
+                max_log_ptr_at_creation = parse<UInt32>(max_log_ptr_at_creation_str);
         }
-        else if (is_create_query)
+
+        if (is_create_query)
         {
-            /// Create new replica. Throws if replica with the same name already exists
+            /// Create replica nodes in ZooKeeper. If newly initialized nodes already exist, reuse them.
             createReplicaNodesInZooKeeper(current_zookeeper);
         }
-        else
+        else if (!replica_exists_in_zk)
         {
             /// It's not CREATE query, but replica does not exist. Probably it was dropped.
             /// Do not create anything, continue as readonly.
@@ -606,37 +615,84 @@ void DatabaseReplicated::createReplicaNodesInZooKeeper(const zkutil::ZooKeeperPt
                         "already contains some data and it does not look like Replicated database path.", zookeeper_path);
 
     /// Write host name to replica_path, it will protect from multiple replicas with the same name
-    auto host_id = getHostID(getContext(), db_uuid, cluster_auth_info.cluster_secure_connection);
+    const auto host_id = getHostID(getContext(), db_uuid, cluster_auth_info.cluster_secure_connection);
+
+    const std::vector<String> check_paths = {
+        replica_path,
+        replica_path + "/replica_group",
+        replica_path + "/digest",
+    };
+    bool nodes_exist = true;
+    auto check_responses = current_zookeeper->tryGet(check_paths);
+    for (size_t i = 0; i < check_responses.size(); ++i)
+    {
+        const auto response = check_responses[i];
+
+        if (response.error == Coordination::Error::ZNONODE)
+        {
+            nodes_exist = false;
+            break;
+        } else if (response.error != Coordination::Error::ZOK)
+        {
+            throw zkutil::KeeperException::fromPath(response.error, check_paths[i]);
+        }
+    }
+
+    if (nodes_exist)
+    {
+        const std::vector<String> expected_data = {
+            host_id,
+            replica_group_name,
+            "0",
+        };
+        for (size_t i = 0; i != expected_data.size(); ++i)
+        {
+            if (check_responses[i].data != expected_data[i])
+            {
+                throw Exception(
+                    ErrorCodes::REPLICA_ALREADY_EXISTS,
+                    "Replica node {} in ZooKeeper already exists and contains unexpected value: {}",
+                    quoteString(check_paths[i]), quoteString(check_responses[i].data));
+            }
+        }
+
+        LOG_DEBUG(log, "Newly initialized replica nodes found in ZooKeeper, reusing them");
+        createEmptyLogEntry(current_zookeeper);
+        return;
+    }
 
     for (int attempts = 10; attempts > 0; --attempts)
     {
         Coordination::Stat stat;
-        String max_log_ptr_str = current_zookeeper->get(zookeeper_path + "/max_log_ptr", &stat);
+        const String max_log_ptr_str = current_zookeeper->get(zookeeper_path + "/max_log_ptr", &stat);
 
-        Coordination::Requests ops;
-        ops.emplace_back(zkutil::makeCreateRequest(replica_path, host_id, zkutil::CreateMode::Persistent));
-        ops.emplace_back(zkutil::makeCreateRequest(replica_path + "/log_ptr", "0", zkutil::CreateMode::Persistent));
-        ops.emplace_back(zkutil::makeCreateRequest(replica_path + "/digest", "0", zkutil::CreateMode::Persistent));
-        ops.emplace_back(zkutil::makeCreateRequest(replica_path + "/replica_group", replica_group_name, zkutil::CreateMode::Persistent));
-        /// In addition to creating the replica nodes, we record the max_log_ptr at the instant where
-        /// we declared ourself as an existing replica. We'll need this during recoverLostReplica to
-        /// notify other nodes that issued new queries while this node was recovering.
-        ops.emplace_back(zkutil::makeCheckRequest(zookeeper_path + "/max_log_ptr", stat.version));
+        const Coordination::Requests ops = {
+            zkutil::makeCreateRequest(replica_path, host_id, zkutil::CreateMode::Persistent),
+            zkutil::makeCreateRequest(replica_path + "/log_ptr", "0", zkutil::CreateMode::Persistent),
+            zkutil::makeCreateRequest(replica_path + "/digest", "0", zkutil::CreateMode::Persistent),
+            zkutil::makeCreateRequest(replica_path + "/replica_group", replica_group_name, zkutil::CreateMode::Persistent),
+
+            /// Previously, this method was not idempotent and max_log_ptr_at_creation could be stored in memory.
+            /// we need to store max_log_ptr_at_creation in ZooKeeper to make this method idempotent during replica creation.
+            zkutil::makeCreateRequest(replica_path + "/max_log_ptr_at_creation", max_log_ptr_str, zkutil::CreateMode::Persistent),
+            zkutil::makeCheckRequest(zookeeper_path + "/max_log_ptr", stat.version),
+        };
+
+        Coordination::Responses ops_responses;
+        const auto code = current_zookeeper->tryMulti(ops, ops_responses);
 
-        Coordination::Responses responses;
-        const auto code = current_zookeeper->tryMulti(ops, responses);
         if (code == Coordination::Error::ZOK)
         {
             max_log_ptr_at_creation = parse<UInt32>(max_log_ptr_str);
-            break;
+            createEmptyLogEntry(current_zookeeper);
+            return;
         }
-        else if (code == Coordination::Error::ZNODEEXISTS || attempts == 1)
+
+        if (attempts == 1)
         {
-            /// If its our last attempt, or if the replica already exists, fail immediately.
-            zkutil::KeeperMultiException::check(code, ops, responses);
+            zkutil::KeeperMultiException::check(code, ops, ops_responses);
         }
     }
-    createEmptyLogEntry(current_zookeeper);
 }
 
 void DatabaseReplicated::beforeLoadingMetadata(ContextMutablePtr context_, LoadingStrictnessLevel mode)
@@ -852,18 +908,6 @@ void DatabaseReplicated::checkTableEngine(const ASTCreateQuery & query, ASTStora
     bool maybe_replica_macros = info.expanded_other;
     bool enable_functional_tests_helper = getContext()->getConfigRef().has("_functional_tests_helper_database_replicated_replace_args_macros");
 
-    if (!enable_functional_tests_helper)
-    {
-        if (query_context->getSettingsRef().database_replicated_allow_replicated_engine_arguments)
-            LOG_WARNING(log, "It's not recommended to explicitly specify zookeeper_path and replica_name in ReplicatedMergeTree arguments");
-        else
-            throw Exception(ErrorCodes::INCORRECT_QUERY,
-                            "It's not allowed to specify explicit zookeeper_path and replica_name "
-                            "for ReplicatedMergeTree arguments in Replicated database. If you really want to "
-                            "specify them explicitly, enable setting "
-                            "database_replicated_allow_replicated_engine_arguments.");
-    }
-
     if (maybe_shard_macros && maybe_replica_macros)
         return;
 
@@ -876,7 +920,9 @@ void DatabaseReplicated::checkTableEngine(const ASTCreateQuery & query, ASTStora
         return;
     }
 
-    throw Exception(ErrorCodes::INCORRECT_QUERY,
+    /// We will replace it with default arguments if the setting is 2
+    if (query_context->getSettingsRef().database_replicated_allow_replicated_engine_arguments != 2)
+        throw Exception(ErrorCodes::INCORRECT_QUERY,
                     "Explicit zookeeper_path and replica_name are specified in ReplicatedMergeTree arguments. "
                     "If you really want to specify it explicitly, then you should use some macros "
                     "to distinguish different shards and replicas");
@@ -1145,6 +1191,9 @@ void DatabaseReplicated::recoverLostReplica(const ZooKeeperPtr & current_zookeep
         /// so we need to allow experimental features that can be used in a CREATE query
         enableAllExperimentalSettings(query_context);
 
+        query_context->setSetting("database_replicated_allow_explicit_uuid", 3);
+        query_context->setSetting("database_replicated_allow_replicated_engine_arguments", 3);
+
         auto txn = std::make_shared<ZooKeeperMetadataTransaction>(current_zookeeper, zookeeper_path, false, "");
         query_context->initZooKeeperMetadataTransaction(txn);
         return query_context;
@@ -1693,6 +1742,9 @@ void DatabaseReplicated::detachTablePermanently(ContextPtr local_context, const
 {
     waitDatabaseStarted();
 
+    if (!local_context->getServerSettings().database_replicated_allow_detach_permanently)
+        throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "Support for DETACH TABLE PERMANENTLY is disabled");
+
     auto txn = local_context->getZooKeeperMetadataTransaction();
     assert(!ddl_worker->isCurrentlyActive() || txn);
     if (txn && txn->isInitialQuery())
diff --git a/src/Disks/DiskLocal.cpp b/src/Disks/DiskLocal.cpp
index d1f0a928b1d..12a5b615234 100644
--- a/src/Disks/DiskLocal.cpp
+++ b/src/Disks/DiskLocal.cpp
@@ -339,7 +339,15 @@ DiskLocal::writeFile(const String & path, size_t buf_size, WriteMode mode, const
 {
     int flags = (mode == WriteMode::Append) ? (O_APPEND | O_CREAT | O_WRONLY) : -1;
     return std::make_unique<WriteBufferFromFile>(
-        fs::path(disk_path) / path, buf_size, flags, settings.local_throttler);
+        fs::path(disk_path) / path,
+        buf_size,
+        flags,
+        settings.local_throttler,
+        0666,
+        nullptr,
+        0,
+        settings.use_adaptive_write_buffer,
+        settings.adaptive_write_buffer_initial_size);
 }
 
 std::vector<String> DiskLocal::getBlobPath(const String & path) const
diff --git a/src/Disks/IO/WriteBufferFromAzureBlobStorage.cpp b/src/Disks/IO/WriteBufferFromAzureBlobStorage.cpp
index 29d3cc8ebd2..5315e331dbd 100644
--- a/src/Disks/IO/WriteBufferFromAzureBlobStorage.cpp
+++ b/src/Disks/IO/WriteBufferFromAzureBlobStorage.cpp
@@ -59,7 +59,7 @@ WriteBufferFromAzureBlobStorage::WriteBufferFromAzureBlobStorage(
     const WriteSettings & write_settings_,
     std::shared_ptr<const AzureBlobStorage::RequestSettings> settings_,
     ThreadPoolCallbackRunnerUnsafe<void> schedule_)
-    : WriteBufferFromFileBase(buf_size_, nullptr, 0)
+    : WriteBufferFromFileBase(std::min(buf_size_, static_cast<size_t>(DBMS_DEFAULT_BUFFER_SIZE)), nullptr, 0)
     , log(getLogger("WriteBufferFromAzureBlobStorage"))
     , buffer_allocation_policy(createBufferAllocationPolicy(*settings_))
     , max_single_part_upload_size(settings_->max_single_part_upload_size)
@@ -244,11 +244,21 @@ void WriteBufferFromAzureBlobStorage::allocateBuffer()
     buffer_allocation_policy->nextBuffer();
     chassert(0 == hidden_size);
 
-    auto size = buffer_allocation_policy->getBufferSize();
-
+    /// First buffer was already allocated in BufferWithOwnMemory constructor with buffer size provided in constructor.
+    /// It will be reallocated in subsequent nextImpl calls up to the desired buffer size from buffer_allocation_policy.
     if (buffer_allocation_policy->getBufferNumber() == 1)
-        size = std::min(size_t(DBMS_DEFAULT_BUFFER_SIZE), size);
+    {
+        /// Reduce memory size if initial size was larger then desired size from buffer_allocation_policy.
+        /// Usually it doesn't happen but we have it in unit tests.
+        if (memory.size() > buffer_allocation_policy->getBufferSize())
+        {
+            memory.resize(buffer_allocation_policy->getBufferSize());
+            WriteBuffer::set(memory.data(), memory.size());
+        }
+        return;
+    }
 
+    auto size = buffer_allocation_policy->getBufferSize();
     memory = Memory(size);
     WriteBuffer::set(memory.data(), memory.size());
 }
diff --git a/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp b/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp
index f85b5f45b37..fa48825e1a6 100644
--- a/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp
+++ b/src/Disks/ObjectStorages/AzureBlobStorage/AzureObjectStorage.cpp
@@ -289,7 +289,7 @@ std::unique_ptr<WriteBufferFromFileBase> AzureObjectStorage::writeObject( /// NO
     return std::make_unique<WriteBufferFromAzureBlobStorage>(
         client.get(),
         object.remote_path,
-        buf_size,
+        write_settings.use_adaptive_write_buffer ? write_settings.adaptive_write_buffer_initial_size : buf_size,
         patchSettings(write_settings),
         settings.get(),
         std::move(scheduler));
diff --git a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp
index 8de80971238..f26a3a8bd9d 100644
--- a/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp
+++ b/src/Disks/ObjectStorages/S3/S3ObjectStorage.cpp
@@ -282,7 +282,7 @@ std::unique_ptr<WriteBufferFromFileBase> S3ObjectStorage::writeObject( /// NOLIN
         client.get(),
         uri.bucket,
         object.remote_path,
-        buf_size,
+        write_settings.use_adaptive_write_buffer ? write_settings.adaptive_write_buffer_initial_size : buf_size,
         request_settings,
         std::move(blob_storage_log),
         attributes,
diff --git a/src/Formats/JSONUtils.cpp b/src/Formats/JSONUtils.cpp
index 9d898cd2470..123f2e4f608 100644
--- a/src/Formats/JSONUtils.cpp
+++ b/src/Formats/JSONUtils.cpp
@@ -483,6 +483,33 @@ namespace JSONUtils
         writeArrayEnd(out, 1);
     }
 
+
+    void writeCompactMetadata(const Names & names, const DataTypes & types, const FormatSettings & settings, WriteBuffer & out)
+    {
+        writeCompactArrayStart(out, 0, "meta");
+
+        for (size_t i = 0; i < names.size(); ++i)
+        {
+            writeCompactObjectStart(out);
+            writeTitle("name", out, 0, "");
+
+            /// The field names are pre-escaped to be put into JSON string literal.
+            writeChar('"', out);
+            writeString(names[i], out);
+            writeChar('"', out);
+
+            writeFieldCompactDelimiter(out);
+            writeTitle("type", out, 0, "");
+            writeJSONString(types[i]->getName(), out, settings);
+            writeCompactObjectEnd(out);
+
+            if (i + 1 < names.size())
+                writeFieldCompactDelimiter(out);
+        }
+
+        writeCompactArrayEnd(out);
+    }
+
     void writeAdditionalInfo(
         size_t rows,
         size_t rows_before_limit,
@@ -530,6 +557,45 @@ namespace JSONUtils
         }
     }
 
+    void writeCompactAdditionalInfo(
+        size_t rows,
+        size_t rows_before_limit,
+        bool applied_limit,
+        const Stopwatch & watch,
+        const Progress & progress,
+        bool write_statistics,
+        WriteBuffer & out)
+    {
+        writeCompactObjectStart(out);
+        writeCompactObjectStart(out, 0, "statistics");
+        writeTitle("rows", out, 0, "");
+        writeIntText(rows, out);
+
+        if (applied_limit)
+        {
+            writeFieldCompactDelimiter(out);
+            writeTitle("rows_before_limit_at_least", out, 0, "");
+            writeIntText(rows_before_limit, out);
+        }
+
+        if (write_statistics)
+        {
+            writeFieldCompactDelimiter(out);
+            writeTitle("elapsed", out, 0, "");
+            writeText(watch.elapsedSeconds(), out);
+            writeFieldCompactDelimiter(out);
+
+            writeTitle("rows_read", out, 0, "");
+            writeText(progress.read_rows.load(), out);
+            writeFieldCompactDelimiter(out);
+
+            writeTitle("bytes_read", out, 0, "");
+            writeText(progress.read_bytes.load(), out);
+        }
+        writeCompactObjectEnd(out);
+        writeCompactObjectEnd(out);
+    }
+
     void writeException(const String & exception_message, WriteBuffer & out, const FormatSettings & settings, size_t indent)
     {
         writeTitle("exception", out, indent, " ");
diff --git a/src/Formats/JSONUtils.h b/src/Formats/JSONUtils.h
index e2ac3467971..622703947b9 100644
--- a/src/Formats/JSONUtils.h
+++ b/src/Formats/JSONUtils.h
@@ -99,6 +99,7 @@ namespace JSONUtils
         WriteBuffer & out);
 
     void writeMetadata(const Names & names, const DataTypes & types, const FormatSettings & settings, WriteBuffer & out);
+    void writeCompactMetadata(const Names & names, const DataTypes & types, const FormatSettings & settings, WriteBuffer & out);
 
     void writeAdditionalInfo(
         size_t rows,
@@ -111,6 +112,15 @@ namespace JSONUtils
         bool write_statistics,
         WriteBuffer & out);
 
+    void writeCompactAdditionalInfo(
+        size_t rows,
+        size_t rows_before_limit,
+        bool applied_limit,
+        const Stopwatch & watch,
+        const Progress & progress,
+        bool write_statistics,
+        WriteBuffer & out);
+
     void writeException(const String & exception_message, WriteBuffer & out, const FormatSettings & settings, size_t indent = 0);
 
     void skipColon(ReadBuffer & in);
diff --git a/src/Formats/registerFormats.cpp b/src/Formats/registerFormats.cpp
index 57ca1bb49c8..770b747fafd 100644
--- a/src/Formats/registerFormats.cpp
+++ b/src/Formats/registerFormats.cpp
@@ -95,6 +95,7 @@ void registerOutputFormatMarkdown(FormatFactory & factory);
 void registerOutputFormatPostgreSQLWire(FormatFactory & factory);
 void registerOutputFormatPrometheus(FormatFactory & factory);
 void registerOutputFormatSQLInsert(FormatFactory & factory);
+void registerOutputFormatJSONCompactWithProgress(FormatFactory & factory);
 
 /// Input only formats.
 
@@ -242,6 +243,7 @@ void registerFormats()
     registerOutputFormatCapnProto(factory);
     registerOutputFormatPrometheus(factory);
     registerOutputFormatSQLInsert(factory);
+    registerOutputFormatJSONCompactWithProgress(factory);
 
     registerInputFormatRegexp(factory);
     registerInputFormatJSONAsString(factory);
diff --git a/src/Functions/DateTimeTransforms.h b/src/Functions/DateTimeTransforms.h
index dfb4b76e5e2..d6d533f16ed 100644
--- a/src/Functions/DateTimeTransforms.h
+++ b/src/Functions/DateTimeTransforms.h
@@ -492,7 +492,7 @@ struct ToStartOfInterval<IntervalKind::Kind::Nanosecond>
     {
         throwDateTimeIsNotSupported(TO_START_OF_INTERVAL_NAME);
     }
-    static Int64 execute(Int64 t, Int64 nanoseconds, const DateLUTImpl &, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 nanoseconds, const DateLUTImpl &, Int64 scale_multiplier, Int64 /*origin*/ = 0)
     {
         if (scale_multiplier < 1000000000)
         {
@@ -527,7 +527,7 @@ struct ToStartOfInterval<IntervalKind::Kind::Microsecond>
     {
         throwDateTimeIsNotSupported(TO_START_OF_INTERVAL_NAME);
     }
-    static Int64 execute(Int64 t, Int64 microseconds, const DateLUTImpl &, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 microseconds, const DateLUTImpl &, Int64 scale_multiplier, Int64 /*origin*/ = 0)
     {
         if (scale_multiplier < 1000000)
         {
@@ -570,7 +570,7 @@ struct ToStartOfInterval<IntervalKind::Kind::Millisecond>
     {
         throwDateTimeIsNotSupported(TO_START_OF_INTERVAL_NAME);
     }
-    static Int64 execute(Int64 t, Int64 milliseconds, const DateLUTImpl &, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 milliseconds, const DateLUTImpl &, Int64 scale_multiplier, Int64 /*origin*/ = 0)
     {
         if (scale_multiplier < 1000)
         {
@@ -613,7 +613,7 @@ struct ToStartOfInterval<IntervalKind::Kind::Second>
     {
         return time_zone.toStartOfSecondInterval(t, seconds);
     }
-    static Int64 execute(Int64 t, Int64 seconds, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 seconds, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 /*origin*/ = 0)
     {
         return time_zone.toStartOfSecondInterval(t / scale_multiplier, seconds);
     }
@@ -634,7 +634,7 @@ struct ToStartOfInterval<IntervalKind::Kind::Minute>
     {
         return time_zone.toStartOfMinuteInterval(t, minutes);
     }
-    static Int64 execute(Int64 t, Int64 minutes, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 minutes, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 /*origin*/ = 0)
     {
         return time_zone.toStartOfMinuteInterval(t / scale_multiplier, minutes);
     }
@@ -655,7 +655,7 @@ struct ToStartOfInterval<IntervalKind::Kind::Hour>
     {
         return time_zone.toStartOfHourInterval(t, hours);
     }
-    static Int64 execute(Int64 t, Int64 hours, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 hours, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 /*origin*/ = 0)
     {
         return time_zone.toStartOfHourInterval(t / scale_multiplier, hours);
     }
@@ -676,7 +676,7 @@ struct ToStartOfInterval<IntervalKind::Kind::Day>
     {
         return static_cast<UInt32>(time_zone.toStartOfDayInterval(time_zone.toDayNum(t), days));
     }
-    static Int64 execute(Int64 t, Int64 days, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 days, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 /*origin*/ = 0)
     {
         return time_zone.toStartOfDayInterval(time_zone.toDayNum(t / scale_multiplier), days);
     }
@@ -697,9 +697,13 @@ struct ToStartOfInterval<IntervalKind::Kind::Week>
     {
         return time_zone.toStartOfWeekInterval(time_zone.toDayNum(t), weeks);
     }
-    static UInt16 execute(Int64 t, Int64 weeks, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 weeks, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 origin = 0)
     {
-        return time_zone.toStartOfWeekInterval(time_zone.toDayNum(t / scale_multiplier), weeks);
+        if (origin == 0)
+            return time_zone.toStartOfWeekInterval(time_zone.toDayNum(t / scale_multiplier), weeks);
+        else
+            return ToStartOfInterval<IntervalKind::Kind::Day>::execute(t, weeks * 7, time_zone, scale_multiplier, origin);
+
     }
 };
 
@@ -718,9 +722,23 @@ struct ToStartOfInterval<IntervalKind::Kind::Month>
     {
         return time_zone.toStartOfMonthInterval(time_zone.toDayNum(t), months);
     }
-    static UInt16 execute(Int64 t, Int64 months, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 months, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 origin = 0)
     {
-        return time_zone.toStartOfMonthInterval(time_zone.toDayNum(t / scale_multiplier), months);
+        const Int64 scaled_time = t / scale_multiplier;
+        if (origin == 0)
+            return time_zone.toStartOfMonthInterval(time_zone.toDayNum(scaled_time), months);
+        else
+        {
+            const Int64 scaled_origin = origin / scale_multiplier;
+            const Int64 days = time_zone.toDayOfMonth(scaled_time + scaled_origin) - time_zone.toDayOfMonth(scaled_origin);
+            Int64 months_to_add = time_zone.toMonth(scaled_time + scaled_origin) - time_zone.toMonth(scaled_origin);
+            const Int64 years = time_zone.toYear(scaled_time + scaled_origin) - time_zone.toYear(scaled_origin);
+            months_to_add = days < 0 ? months_to_add - 1 : months_to_add;
+            months_to_add += years * 12;
+            Int64 month_multiplier = (months_to_add / months) * months;
+
+            return (time_zone.addMonths(time_zone.toDate(scaled_origin), month_multiplier) - time_zone.toDate(scaled_origin));
+        }
     }
 };
 
@@ -739,9 +757,12 @@ struct ToStartOfInterval<IntervalKind::Kind::Quarter>
     {
         return time_zone.toStartOfQuarterInterval(time_zone.toDayNum(t), quarters);
     }
-    static UInt16 execute(Int64 t, Int64 quarters, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 quarters, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 origin = 0)
     {
-        return time_zone.toStartOfQuarterInterval(time_zone.toDayNum(t / scale_multiplier), quarters);
+        if (origin == 0)
+            return time_zone.toStartOfQuarterInterval(time_zone.toDayNum(t / scale_multiplier), quarters);
+        else
+            return ToStartOfInterval<IntervalKind::Kind::Month>::execute(t, quarters * 3, time_zone, scale_multiplier, origin);
     }
 };
 
@@ -760,9 +781,12 @@ struct ToStartOfInterval<IntervalKind::Kind::Year>
     {
         return time_zone.toStartOfYearInterval(time_zone.toDayNum(t), years);
     }
-    static UInt16 execute(Int64 t, Int64 years, const DateLUTImpl & time_zone, Int64 scale_multiplier)
+    static Int64 execute(Int64 t, Int64 years, const DateLUTImpl & time_zone, Int64 scale_multiplier, Int64 origin = 0)
     {
-        return time_zone.toStartOfYearInterval(time_zone.toDayNum(t / scale_multiplier), years);
+        if (origin == 0)
+            return time_zone.toStartOfYearInterval(time_zone.toDayNum(t / scale_multiplier), years);
+        else
+            return ToStartOfInterval<IntervalKind::Kind::Month>::execute(t, years * 12, time_zone, scale_multiplier, origin);
     }
 };
 
diff --git a/src/Functions/array/array.cpp b/src/Functions/array/array.cpp
index dfe589fb74f..d2aedd57f99 100644
--- a/src/Functions/array/array.cpp
+++ b/src/Functions/array/array.cpp
@@ -1,11 +1,15 @@
-#include <Functions/IFunction.h>
-#include <Functions/FunctionFactory.h>
+#include <Columns/ColumnArray.h>
+#include <Columns/ColumnDecimal.h>
+#include <Columns/ColumnFixedString.h>
+#include <Columns/ColumnNullable.h>
+#include <Columns/ColumnString.h>
+#include <Core/Settings.h>
 #include <DataTypes/DataTypeArray.h>
 #include <DataTypes/getLeastSupertype.h>
-#include <Columns/ColumnArray.h>
-#include <Core/Settings.h>
-#include <Interpreters/castColumn.h>
+#include <Functions/FunctionFactory.h>
+#include <Functions/IFunction.h>
 #include <Interpreters/Context.h>
+#include <Interpreters/castColumn.h>
 
 
 namespace DB
@@ -44,11 +48,13 @@ public:
 
     ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, size_t input_rows_count) const override
     {
-        size_t num_elements = arguments.size();
+        const size_t num_elements = arguments.size();
 
         if (num_elements == 0)
+        {
             /// We should return constant empty array.
             return result_type->createColumnConstWithDefaultValue(input_rows_count);
+        }
 
         const DataTypePtr & elem_type = static_cast<const DataTypeArray &>(*result_type).getNestedType();
 
@@ -60,7 +66,6 @@ public:
 
         Columns columns_holder(num_elements);
         ColumnRawPtrs column_ptrs(num_elements);
-
         for (size_t i = 0; i < num_elements; ++i)
         {
             const auto & arg = arguments[i];
@@ -77,35 +82,199 @@ public:
         }
 
         /// Create and fill the result array.
-
         auto out = ColumnArray::create(elem_type->createColumn());
         IColumn & out_data = out->getData();
         IColumn::Offsets & out_offsets = out->getOffsets();
 
-        out_data.reserve(input_rows_count * num_elements);
-        out_offsets.resize(input_rows_count);
-
+        /// Fill out_offsets
+        out_offsets.resize_exact(input_rows_count);
         IColumn::Offset current_offset = 0;
         for (size_t i = 0; i < input_rows_count; ++i)
         {
-            for (size_t j = 0; j < num_elements; ++j)
-                out_data.insertFrom(*column_ptrs[j], i);
-
             current_offset += num_elements;
             out_offsets[i] = current_offset;
         }
 
+        /// Fill out_data
+        out_data.reserve(input_rows_count * num_elements);
+        if (num_elements == 1)
+            out_data.insertRangeFrom(*column_ptrs[0], 0, input_rows_count);
+        else
+            execute(column_ptrs, out_data, input_rows_count);
         return out;
     }
 
 private:
+    bool execute(const ColumnRawPtrs & columns, IColumn & out_data, size_t input_rows_count) const
+    {
+        return executeNumber<UInt8>(columns, out_data, input_rows_count) || executeNumber<UInt16>(columns, out_data, input_rows_count)
+            || executeNumber<UInt32>(columns, out_data, input_rows_count) || executeNumber<UInt64>(columns, out_data, input_rows_count)
+            || executeNumber<UInt128>(columns, out_data, input_rows_count) || executeNumber<UInt256>(columns, out_data, input_rows_count)
+            || executeNumber<Int8>(columns, out_data, input_rows_count) || executeNumber<Int16>(columns, out_data, input_rows_count)
+            || executeNumber<Int32>(columns, out_data, input_rows_count) || executeNumber<Int64>(columns, out_data, input_rows_count)
+            || executeNumber<Int128>(columns, out_data, input_rows_count) || executeNumber<Int256>(columns, out_data, input_rows_count)
+            || executeNumber<Float32>(columns, out_data, input_rows_count) || executeNumber<Float64>(columns, out_data, input_rows_count)
+            || executeNumber<Decimal32>(columns, out_data, input_rows_count)
+            || executeNumber<Decimal64>(columns, out_data, input_rows_count)
+            || executeNumber<Decimal128>(columns, out_data, input_rows_count)
+            || executeNumber<Decimal256>(columns, out_data, input_rows_count)
+            || executeNumber<DateTime64>(columns, out_data, input_rows_count) || executeString(columns, out_data, input_rows_count)
+            || executeNullable(columns, out_data, input_rows_count) || executeTuple(columns, out_data, input_rows_count)
+            || executeFixedString(columns, out_data, input_rows_count) || executeGeneric(columns, out_data, input_rows_count);
+    }
+
+    template <typename T>
+    bool executeNumber(const ColumnRawPtrs & columns, IColumn & out_data, size_t input_rows_count) const
+    {
+        using Container = ColumnVectorOrDecimal<T>::Container;
+        std::vector<const Container *> containers(columns.size(), nullptr);
+        for (size_t i = 0; i < columns.size(); ++i)
+        {
+            const ColumnVectorOrDecimal<T> * concrete_column = checkAndGetColumn<ColumnVectorOrDecimal<T>>(columns[i]);
+            if (!concrete_column)
+                return false;
+
+            containers[i] = &concrete_column->getData();
+        }
+
+        ColumnVectorOrDecimal<T> & concrete_out_data = assert_cast<ColumnVectorOrDecimal<T> &>(out_data);
+        Container & out_container = concrete_out_data.getData();
+        out_container.resize_exact(columns.size() * input_rows_count);
+
+        for (size_t row_i = 0; row_i < input_rows_count; ++row_i)
+        {
+            const size_t base = row_i * columns.size();
+            for (size_t col_i = 0; col_i < columns.size(); ++col_i)
+                out_container[base + col_i] = (*containers[col_i])[row_i];
+        }
+        return true;
+    }
+
+    bool executeString(const ColumnRawPtrs & columns, IColumn & out_data, size_t input_rows_count) const
+    {
+        size_t total_bytes = 0;
+        std::vector<const ColumnString *> concrete_columns(columns.size(), nullptr);
+        for (size_t i = 0; i < columns.size(); ++i)
+        {
+            const ColumnString * concrete_column = checkAndGetColumn<ColumnString>(columns[i]);
+            if (!concrete_column)
+                return false;
+
+            total_bytes += concrete_column->getChars().size();
+            concrete_columns[i] = concrete_column;
+        }
+
+        ColumnString & concrete_out_data = assert_cast<ColumnString &>(out_data);
+        auto & out_chars = concrete_out_data.getChars();
+        auto & out_offsets = concrete_out_data.getOffsets();
+        out_chars.resize_exact(total_bytes);
+        out_offsets.resize_exact(input_rows_count * columns.size());
+
+        size_t cur_out_offset = 0;
+        for (size_t row_i = 0; row_i < input_rows_count; ++row_i)
+        {
+            const size_t base = row_i * columns.size();
+            for (size_t col_i = 0; col_i < columns.size(); ++col_i)
+            {
+                StringRef ref = concrete_columns[col_i]->getDataAt(row_i);
+                memcpySmallAllowReadWriteOverflow15(&out_chars[cur_out_offset], ref.data, ref.size);
+                out_chars[cur_out_offset + ref.size] = 0;
+
+                cur_out_offset += ref.size + 1;
+                out_offsets[base + col_i] = cur_out_offset;
+            }
+        }
+        return true;
+    }
+
+    bool executeFixedString(const ColumnRawPtrs & columns, IColumn & out_data, size_t input_rows_count) const
+    {
+        std::vector<const ColumnFixedString *> concrete_columns(columns.size(), nullptr);
+        for (size_t i = 0; i < columns.size(); ++i)
+        {
+            const ColumnFixedString * concrete_column = checkAndGetColumn<ColumnFixedString>(columns[i]);
+            if (!concrete_column)
+                return false;
+
+            concrete_columns[i] = concrete_column;
+        }
+
+        ColumnFixedString & concrete_out_data = assert_cast<ColumnFixedString &>(out_data);
+        auto & out_chars = concrete_out_data.getChars();
+
+        const size_t n = concrete_out_data.getN();
+        size_t total_bytes = n * columns.size() * input_rows_count;
+        out_chars.resize_exact(total_bytes);
+
+        size_t curr_out_offset = 0;
+        for (size_t row_i = 0; row_i < input_rows_count; ++row_i)
+        {
+            for (size_t col_i = 0; col_i < columns.size(); ++col_i)
+            {
+                StringRef ref = concrete_columns[col_i]->getDataAt(row_i);
+                memcpySmallAllowReadWriteOverflow15(&out_chars[curr_out_offset], ref.data, n);
+                curr_out_offset += n;
+            }
+        }
+        return true;
+    }
+
+    bool executeNullable(const ColumnRawPtrs & columns, IColumn & out_data, size_t input_rows_count) const
+    {
+        ColumnRawPtrs null_maps(columns.size(), nullptr);
+        ColumnRawPtrs nested_columns(columns.size(), nullptr);
+        for (size_t i = 0; i < columns.size(); ++i)
+        {
+            const ColumnNullable * concrete_column = checkAndGetColumn<ColumnNullable>(columns[i]);
+            if (!concrete_column)
+                return false;
+
+            null_maps[i] = &concrete_column->getNullMapColumn();
+            nested_columns[i] = &concrete_column->getNestedColumn();
+        }
+
+        ColumnNullable & concrete_out_data = assert_cast<ColumnNullable &>(out_data);
+        auto & out_null_map = concrete_out_data.getNullMapColumn();
+        auto & out_nested_column = concrete_out_data.getNestedColumn();
+        execute(null_maps, out_null_map, input_rows_count);
+        execute(nested_columns, out_nested_column, input_rows_count);
+        return true;
+    }
+
+    bool executeTuple(const ColumnRawPtrs & columns, IColumn & out_data, size_t input_rows_count) const
+    {
+        ColumnTuple * concrete_out_data = typeid_cast<ColumnTuple *>(&out_data);
+        if (!concrete_out_data)
+            return false;
+
+        const size_t tuple_size = concrete_out_data->tupleSize();
+        for (size_t i = 0; i < tuple_size; ++i)
+        {
+            ColumnRawPtrs elem_columns(columns.size(), nullptr);
+            for (size_t j = 0; j < columns.size(); ++j)
+            {
+                const ColumnTuple * concrete_column = assert_cast<const ColumnTuple *>(columns[j]);
+                elem_columns[j] = &concrete_column->getColumn(i);
+            }
+            execute(elem_columns, concrete_out_data->getColumn(i), input_rows_count);
+        }
+        return true;
+    }
+
+    bool executeGeneric(const ColumnRawPtrs & columns, IColumn & out_data, size_t input_rows_count) const
+    {
+        for (size_t i = 0; i < input_rows_count; ++i)
+            for (const auto * column : columns)
+                out_data.insertFrom(*column, i);
+        return true;
+    }
+
+
     String getName() const override
     {
         return name;
     }
 
-    bool addField(DataTypePtr type_res, const Field & f, Array & arr) const;
-
     bool use_variant_as_common_type = false;
 };
 
diff --git a/src/Functions/map.cpp b/src/Functions/map.cpp
index 534f7c0d8cd..14672cfd568 100644
--- a/src/Functions/map.cpp
+++ b/src/Functions/map.cpp
@@ -2,6 +2,8 @@
 #include <Columns/ColumnMap.h>
 #include <Columns/ColumnNullable.h>
 #include <Columns/ColumnsCommon.h>
+#include <Columns/ColumnsNumber.h>
+#include <Core/Settings.h>
 #include <DataTypes/DataTypeArray.h>
 #include <DataTypes/DataTypeMap.h>
 #include <DataTypes/DataTypeTuple.h>
@@ -13,7 +15,6 @@
 #include <Interpreters/Context.h>
 #include <Interpreters/castColumn.h>
 #include <Common/HashTable/HashSet.h>
-#include <Core/Settings.h>
 
 
 namespace DB
@@ -36,11 +37,18 @@ class FunctionMap : public IFunction
 public:
     static constexpr auto name = "map";
 
-    explicit FunctionMap(bool use_variant_as_common_type_) : use_variant_as_common_type(use_variant_as_common_type_) {}
+    explicit FunctionMap(ContextPtr context_)
+        : context(context_)
+        , use_variant_as_common_type(
+              context->getSettingsRef().allow_experimental_variant_type && context->getSettingsRef().use_variant_as_common_type)
+        , function_array(FunctionFactory::instance().get("array", context))
+        , function_map_from_arrays(FunctionFactory::instance().get("mapFromArrays", context))
+    {
+    }
 
     static FunctionPtr create(ContextPtr context)
     {
-        return std::make_shared<FunctionMap>(context->getSettingsRef().allow_experimental_variant_type && context->getSettingsRef().use_variant_as_common_type);
+        return std::make_shared<FunctionMap>(context);
     }
 
     String getName() const override
@@ -101,62 +109,38 @@ public:
     ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, size_t input_rows_count) const override
     {
         size_t num_elements = arguments.size();
-
         if (num_elements == 0)
             return result_type->createColumnConstWithDefaultValue(input_rows_count);
 
+        ColumnsWithTypeAndName key_args;
+        ColumnsWithTypeAndName value_args;
+        for (size_t i = 0; i < num_elements; i += 2)
+        {
+            key_args.emplace_back(arguments[i]);
+            value_args.emplace_back(arguments[i+1]);
+        }
+
         const auto & result_type_map = static_cast<const DataTypeMap &>(*result_type);
         const DataTypePtr & key_type = result_type_map.getKeyType();
         const DataTypePtr & value_type = result_type_map.getValueType();
+        const DataTypePtr & key_array_type = std::make_shared<DataTypeArray>(key_type);
+        const DataTypePtr & value_array_type = std::make_shared<DataTypeArray>(value_type);
 
-        Columns columns_holder(num_elements);
-        ColumnRawPtrs column_ptrs(num_elements);
+        /// key_array = array(args[0], args[2]...)
+        ColumnPtr key_array = function_array->build(key_args)->execute(key_args, key_array_type, input_rows_count);
+        /// value_array = array(args[1], args[3]...)
+        ColumnPtr value_array = function_array->build(value_args)->execute(value_args, value_array_type, input_rows_count);
 
-        for (size_t i = 0; i < num_elements; ++i)
-        {
-            const auto & arg = arguments[i];
-            const auto to_type = i % 2 == 0 ? key_type : value_type;
-
-            ColumnPtr preprocessed_column = castColumn(arg, to_type);
-            preprocessed_column = preprocessed_column->convertToFullColumnIfConst();
-
-            columns_holder[i] = std::move(preprocessed_column);
-            column_ptrs[i] = columns_holder[i].get();
-        }
-
-        /// Create and fill the result map.
-
-        MutableColumnPtr keys_data = key_type->createColumn();
-        MutableColumnPtr values_data = value_type->createColumn();
-        MutableColumnPtr offsets = DataTypeNumber<IColumn::Offset>().createColumn();
-
-        size_t total_elements = input_rows_count * num_elements / 2;
-        keys_data->reserve(total_elements);
-        values_data->reserve(total_elements);
-        offsets->reserve(input_rows_count);
-
-        IColumn::Offset current_offset = 0;
-        for (size_t i = 0; i < input_rows_count; ++i)
-        {
-            for (size_t j = 0; j < num_elements; j += 2)
-            {
-                keys_data->insertFrom(*column_ptrs[j], i);
-                values_data->insertFrom(*column_ptrs[j + 1], i);
-            }
-
-            current_offset += num_elements / 2;
-            offsets->insert(current_offset);
-        }
-
-        auto nested_column = ColumnArray::create(
-            ColumnTuple::create(Columns{std::move(keys_data), std::move(values_data)}),
-            std::move(offsets));
-
-        return ColumnMap::create(nested_column);
+        /// result = mapFromArrays(key_array, value_array)
+        ColumnsWithTypeAndName map_args{{key_array, key_array_type, ""}, {value_array, value_array_type, ""}};
+        return function_map_from_arrays->build(map_args)->execute(map_args, result_type, input_rows_count);
     }
 
 private:
+    ContextPtr context;
     bool use_variant_as_common_type = false;
+    FunctionOverloadResolverPtr function_array;
+    FunctionOverloadResolverPtr function_map_from_arrays;
 };
 
 /// mapFromArrays(keys, values) is a function that allows you to make key-value pair from a pair of arrays or maps
@@ -173,6 +157,7 @@ public:
     bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return false; }
     bool useDefaultImplementationForNulls() const override { return false; }
     bool useDefaultImplementationForConstants() const override { return true; }
+    bool useDefaultImplementationForLowCardinalityColumns() const override { return false; }
 
     DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override
     {
diff --git a/src/Functions/materialize.h b/src/Functions/materialize.h
index 41994509745..571391faba7 100644
--- a/src/Functions/materialize.h
+++ b/src/Functions/materialize.h
@@ -2,7 +2,7 @@
 #include <Functions/IFunction.h>
 #include <Functions/FunctionFactory.h>
 #include <Columns/ColumnLowCardinality.h>
-#include <DataTypes/DataTypeLowCardinality.h>
+#include <Columns/ColumnSparse.h>
 
 namespace DB
 {
@@ -18,11 +18,6 @@ public:
         return std::make_shared<FunctionMaterialize>();
     }
 
-    bool useDefaultImplementationForNulls() const override
-    {
-        return false;
-    }
-
     /// Get the function name.
     String getName() const override
     {
@@ -34,8 +29,16 @@ public:
         return true;
     }
 
+    bool useDefaultImplementationForNulls() const override { return false; }
+
+    bool useDefaultImplementationForNothing() const override { return false; }
+
+    bool useDefaultImplementationForConstants() const override { return false; }
+
     bool useDefaultImplementationForLowCardinalityColumns() const override { return false; }
 
+    bool useDefaultImplementationForSparseColumns() const override { return false; }
+
     bool isSuitableForConstantFolding() const override { return false; }
 
     bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return false; }
@@ -52,7 +55,7 @@ public:
 
     ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t /*input_rows_count*/) const override
     {
-        return arguments[0].column->convertToFullColumnIfConst();
+        return recursiveRemoveSparse(arguments[0].column->convertToFullColumnIfConst());
     }
 
     bool hasInformationAboutMonotonicity() const override { return true; }
diff --git a/src/Functions/sqid.cpp b/src/Functions/sqid.cpp
index 0e133590b84..074a34bd083 100644
--- a/src/Functions/sqid.cpp
+++ b/src/Functions/sqid.cpp
@@ -124,7 +124,7 @@ public:
                 std::string_view sqid = col_non_const->getDataAt(i).toView();
                 std::vector<UInt64> integers = sqids.decode(String(sqid));
                 res_nested_data.insert(integers.begin(), integers.end());
-                res_offsets_data.push_back(integers.size());
+                res_offsets_data.push_back(res_offsets_data.back() + integers.size());
             }
         }
         else
diff --git a/src/Functions/toStartOfInterval.cpp b/src/Functions/toStartOfInterval.cpp
index 709f5f86d80..6573fef6634 100644
--- a/src/Functions/toStartOfInterval.cpp
+++ b/src/Functions/toStartOfInterval.cpp
@@ -10,21 +10,31 @@
 #include <Functions/FunctionFactory.h>
 #include <Functions/IFunction.h>
 #include <IO/WriteHelpers.h>
+#include <algorithm>
 
 
 namespace DB
 {
 namespace ErrorCodes
 {
-    extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH;
+    extern const int ARGUMENT_OUT_OF_BOUND;
+    extern const int BAD_ARGUMENTS;
     extern const int ILLEGAL_COLUMN;
     extern const int ILLEGAL_TYPE_OF_ARGUMENT;
-    extern const int ARGUMENT_OUT_OF_BOUND;
+    extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH;
 }
 
 
 class FunctionToStartOfInterval : public IFunction
 {
+private:
+    enum class Overload
+    {
+        Default,    /// toStartOfInterval(time, interval) or toStartOfInterval(time, interval, timezone)
+        Origin      /// toStartOfInterval(time, interval, origin) or toStartOfInterval(time, interval, origin, timezone)
+    };
+    mutable Overload overload;
+
 public:
     static FunctionPtr create(ContextPtr) { return std::make_shared<FunctionToStartOfInterval>(); }
 
@@ -34,7 +44,7 @@ public:
     size_t getNumberOfArguments() const override { return 0; }
     bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return false; }
     bool useDefaultImplementationForConstants() const override { return true; }
-    ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {1, 2}; }
+    ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {1, 2, 3}; }
     bool hasInformationAboutMonotonicity() const override { return true; }
     Monotonicity getMonotonicityForRange(const IDataType &, const Field &, const Field &) const override { return { .is_monotonic = true, .is_always_monotonic = true }; }
 
@@ -72,6 +82,9 @@ public:
                     "Illegal type {} of 2nd argument of function {}, expected a time interval",
                     type_arg2->getName(), getName());
 
+            overload = Overload::Default;
+
+            /// Determine result type for default overload (no origin)
             switch (interval_type->getKind()) // NOLINT(bugprone-switch-missing-default-case)
             {
                 case IntervalKind::Kind::Nanosecond:
@@ -97,13 +110,49 @@ public:
         auto check_third_argument = [&]
         {
             const DataTypePtr & type_arg3 = arguments[2].type;
-            if (!isString(type_arg3))
-                throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT,
-                    "Illegal type {} of 3rd argument of function {}, expected a constant timezone string",
+            if (isString(type_arg3))
+            {
+                if (value_is_date && result_type == ResultType::Date)
+                    throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT,
+                        "A timezone argument of function {} with interval type {} is allowed only when the 1st argument has the type DateTime or DateTime64",
+                        getName(), interval_type->getKind().toString());
+            }
+            else if (isDateOrDate32OrDateTimeOrDateTime64(type_arg3))
+            {
+                overload = Overload::Origin;
+                const DataTypePtr & type_arg1 = arguments[0].type;
+                if (isDate(type_arg1) && isDate(type_arg3))
+                    result_type = ResultType::Date;
+                else if (isDate32(type_arg1) && isDate32(type_arg3))
+                    result_type = ResultType::Date32;
+                else if (isDateTime(type_arg1) && isDateTime(type_arg3))
+                    result_type = ResultType::DateTime;
+                else if (isDateTime64(type_arg1) && isDateTime64(type_arg3))
+                    result_type = ResultType::DateTime64;
+                else
+                    throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Datetime argument and origin argument for function {} must have the same type", getName());
+            }
+            else
+                throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type {} of 3rd argument of function {}. "
+                    "This argument is optional and must be a constant String with timezone name or a Date/Date32/DateTime/DateTime64 with a constant origin",
                     type_arg3->getName(), getName());
-            if (value_is_date && result_type == ResultType::Date) /// weird why this is && instead of || but too afraid to change it
+        };
+
+        auto check_fourth_argument = [&]
+        {
+            if (overload != Overload::Origin) /// sanity check
+                throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type {} of 3rd argument of function {}. "
+                    "The third argument must a Date/Date32/DateTime/DateTime64 with a constant origin",
+                    arguments[2].type->getName(), getName());
+
+            const DataTypePtr & type_arg4 = arguments[3].type;
+            if (!isString(type_arg4))
+                throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal type {} of 4th argument of function {}. "
+                    "This argument is optional and must be a constant String with timezone name",
+                    type_arg4->getName(), getName());
+            if (value_is_date && result_type == ResultType::Date)
                 throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT,
-                    "The timezone argument of function {} with interval type {} is allowed only when the 1st argument has type DateTime or DateTimt64",
+                    "A timezone argument of function {} with interval type {} is allowed only when the 1st argument has the type DateTime or DateTime64",
                     getName(), interval_type->getKind().toString());
         };
 
@@ -118,10 +167,17 @@ public:
             check_second_argument();
             check_third_argument();
         }
+        else if (arguments.size() == 4)
+        {
+            check_first_argument();
+            check_second_argument();
+            check_third_argument();
+            check_fourth_argument();
+        }
         else
         {
             throw Exception(ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH,
-                "Number of arguments for function {} doesn't match: passed {}, should be 2 or 3",
+                "Number of arguments for function {} doesn't match: passed {}, must be 2, 3 or 4",
                 getName(), arguments.size());
         }
 
@@ -132,10 +188,19 @@ public:
             case ResultType::Date32:
                 return std::make_shared<DataTypeDate32>();
             case ResultType::DateTime:
-                return std::make_shared<DataTypeDateTime>(extractTimeZoneNameFromFunctionArguments(arguments, 2, 0, false));
+            {
+                const size_t time_zone_arg_num = (overload == Overload::Default) ? 2 : 3;
+                return std::make_shared<DataTypeDateTime>(extractTimeZoneNameFromFunctionArguments(arguments, time_zone_arg_num, 0, false));
+            }
             case ResultType::DateTime64:
             {
                 UInt32 scale = 0;
+                if (isDateTime64(arguments[0].type) && overload == Overload::Origin)
+                {
+                    scale = assert_cast<const DataTypeDateTime64 &>(*arguments[0].type.get()).getScale();
+                    if (assert_cast<const DataTypeDateTime64 &>(*arguments[2].type.get()).getScale() != scale)
+                        throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Datetime argument and origin argument for function {} must have the same scale", getName());
+                }
                 if (interval_type->getKind() == IntervalKind::Kind::Nanosecond)
                     scale = 9;
                 else if (interval_type->getKind() == IntervalKind::Kind::Microsecond)
@@ -143,69 +208,103 @@ public:
                 else if (interval_type->getKind() == IntervalKind::Kind::Millisecond)
                     scale = 3;
 
-                return std::make_shared<DataTypeDateTime64>(scale, extractTimeZoneNameFromFunctionArguments(arguments, 2, 0, false));
+                const size_t time_zone_arg_num = (overload == Overload::Default) ? 2 : 3;
+                return std::make_shared<DataTypeDateTime64>(scale, extractTimeZoneNameFromFunctionArguments(arguments, time_zone_arg_num, 0, false));
             }
         }
 
         std::unreachable();
     }
 
-    ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, size_t input_rows_count) const override
+    ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, size_t /* input_rows_count */) const override
     {
         const auto & time_column = arguments[0];
         const auto & interval_column = arguments[1];
-        const auto & time_zone = extractTimeZoneFromFunctionArguments(arguments, 2, 0);
-        auto result_column = dispatchForTimeColumn(time_column, interval_column, result_type, time_zone, input_rows_count);
+
+        ColumnWithTypeAndName origin_column;
+        if (overload == Overload::Origin)
+            origin_column = arguments[2];
+
+        const size_t time_zone_arg_num = (overload == Overload::Default) ? 2 : 3;
+        const auto & time_zone = extractTimeZoneFromFunctionArguments(arguments, time_zone_arg_num, 0);
+
+        ColumnPtr result_column;
+        if (isDate(result_type))
+            result_column = dispatchForTimeColumn<DataTypeDate>(time_column, interval_column, origin_column, result_type, time_zone);
+        else if (isDate32(result_type))
+            result_column = dispatchForTimeColumn<DataTypeDate32>(time_column, interval_column, origin_column, result_type, time_zone);
+        else if (isDateTime(result_type))
+            result_column = dispatchForTimeColumn<DataTypeDateTime>(time_column, interval_column, origin_column, result_type, time_zone);
+        else if (isDateTime64(result_type))
+            result_column = dispatchForTimeColumn<DataTypeDateTime64>(time_column, interval_column, origin_column, result_type, time_zone);
         return result_column;
     }
 
 private:
+    template <typename ReturnType>
     ColumnPtr dispatchForTimeColumn(
-        const ColumnWithTypeAndName & time_column, const ColumnWithTypeAndName & interval_column,
-        const DataTypePtr & result_type, const DateLUTImpl & time_zone,
-        size_t input_rows_count) const
+        const ColumnWithTypeAndName & time_column, const ColumnWithTypeAndName & interval_column, const ColumnWithTypeAndName & origin_column, const DataTypePtr & result_type, const DateLUTImpl & time_zone) const
     {
         const auto & time_column_type = *time_column.type.get();
         const auto & time_column_col = *time_column.column.get();
 
-        if (isDateTime64(time_column_type))
-        {
-            const auto * time_column_vec = checkAndGetColumn<ColumnDateTime64>(&time_column_col);
-            auto scale = assert_cast<const DataTypeDateTime64 &>(time_column_type).getScale();
-
-            if (time_column_vec)
-                return dispatchForIntervalColumn(assert_cast<const DataTypeDateTime64 &>(time_column_type), *time_column_vec, interval_column, result_type, time_zone, input_rows_count, scale);
-        }
-        else if (isDateTime(time_column_type))
-        {
-            const auto * time_column_vec = checkAndGetColumn<ColumnDateTime>(&time_column_col);
-            if (time_column_vec)
-                return dispatchForIntervalColumn(assert_cast<const DataTypeDateTime &>(time_column_type), *time_column_vec, interval_column, result_type, time_zone, input_rows_count);
-        }
-        else if (isDate(time_column_type))
+        if (isDate(time_column_type))
         {
             const auto * time_column_vec = checkAndGetColumn<ColumnDate>(&time_column_col);
+
             if (time_column_vec)
-                return dispatchForIntervalColumn(assert_cast<const DataTypeDate &>(time_column_type), *time_column_vec, interval_column, result_type, time_zone, input_rows_count);
+                return dispatchForIntervalColumn<ReturnType, DataTypeDate, ColumnDate>(assert_cast<const DataTypeDate &>(time_column_type), *time_column_vec, interval_column, origin_column, result_type, time_zone);
         }
         else if (isDate32(time_column_type))
         {
             const auto * time_column_vec = checkAndGetColumn<ColumnDate32>(&time_column_col);
             if (time_column_vec)
-                return dispatchForIntervalColumn(assert_cast<const DataTypeDate32 &>(time_column_type), *time_column_vec, interval_column, result_type, time_zone, input_rows_count);
+                return dispatchForIntervalColumn<ReturnType, DataTypeDate32, ColumnDate32>(assert_cast<const DataTypeDate32 &>(time_column_type), *time_column_vec, interval_column, origin_column, result_type, time_zone);
+        }
+        else if (isDateTime(time_column_type))
+        {
+            const auto * time_column_vec = checkAndGetColumn<ColumnDateTime>(&time_column_col);
+            if (time_column_vec)
+                return dispatchForIntervalColumn<ReturnType, DataTypeDateTime, ColumnDateTime>(assert_cast<const DataTypeDateTime &>(time_column_type), *time_column_vec, interval_column, origin_column, result_type, time_zone);
+        }
+        else if (isDateTime64(time_column_type))
+        {
+            const auto * time_column_vec = checkAndGetColumn<ColumnDateTime64>(&time_column_col);
+            auto scale = assert_cast<const DataTypeDateTime64 &>(time_column_type).getScale();
+
+            if (time_column_vec)
+                return dispatchForIntervalColumn<ReturnType, DataTypeDateTime64, ColumnDateTime64>(assert_cast<const DataTypeDateTime64 &>(time_column_type), *time_column_vec, interval_column, origin_column, result_type, time_zone, scale);
         }
         throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal column for 1st argument of function {}, expected a Date, Date32, DateTime or DateTime64", getName());
     }
 
-    template <typename TimeDataType, typename TimeColumnType>
+    template <typename ReturnType, typename TimeDataType, typename TimeColumnType>
     ColumnPtr dispatchForIntervalColumn(
-        const TimeDataType & time_data_type, const TimeColumnType & time_column, const ColumnWithTypeAndName & interval_column,
-        const DataTypePtr & result_type, const DateLUTImpl & time_zone, size_t input_rows_count, UInt16 scale = 1) const
+        const TimeDataType & time_data_type, const TimeColumnType & time_column, const ColumnWithTypeAndName & interval_column, const ColumnWithTypeAndName & origin_column,
+        const DataTypePtr & result_type, const DateLUTImpl & time_zone, UInt16 scale = 1) const
     {
         const auto * interval_type = checkAndGetDataType<DataTypeInterval>(interval_column.type.get());
         if (!interval_type)
             throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Illegal column for 2nd argument of function {}, must be a time interval", getName());
 
+        switch (interval_type->getKind()) // NOLINT(bugprone-switch-missing-default-case)
+        {
+            case IntervalKind::Kind::Nanosecond:
+            case IntervalKind::Kind::Microsecond:
+            case IntervalKind::Kind::Millisecond:
+                if (isDateOrDate32(time_data_type) || isDateTime(time_data_type))
+                    throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal interval kind for argument data type {}", isDate(time_data_type) ? "Date" : "DateTime");
+                break;
+            case IntervalKind::Kind::Second:
+            case IntervalKind::Kind::Minute:
+            case IntervalKind::Kind::Hour:
+                if (isDateOrDate32(time_data_type))
+                    throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Illegal interval kind for argument data type Date");
+                break;
+            default:
+                break;
+        }
+
         const auto * interval_column_const_int64 = checkAndGetColumnConst<ColumnInt64>(interval_column.column.get());
         if (!interval_column_const_int64)
             throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Illegal column for 2nd argument of function {}, must be a const time interval", getName());
@@ -217,51 +316,102 @@ private:
         switch (interval_type->getKind()) // NOLINT(bugprone-switch-missing-default-case)
         {
             case IntervalKind::Kind::Nanosecond:
-                return execute<TimeDataType, TimeColumnType, DataTypeDateTime64, IntervalKind::Kind::Nanosecond>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Nanosecond>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Microsecond:
-                return execute<TimeDataType, TimeColumnType, DataTypeDateTime64, IntervalKind::Kind::Microsecond>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Microsecond>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Millisecond:
-                return execute<TimeDataType, TimeColumnType, DataTypeDateTime64, IntervalKind::Kind::Millisecond>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Millisecond>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Second:
-                return execute<TimeDataType, TimeColumnType, DataTypeDateTime, IntervalKind::Kind::Second>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Second>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Minute:
-                return execute<TimeDataType, TimeColumnType, DataTypeDateTime, IntervalKind::Kind::Minute>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Minute>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Hour:
-                return execute<TimeDataType, TimeColumnType, DataTypeDateTime, IntervalKind::Kind::Hour>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Hour>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Day:
-                return execute<TimeDataType, TimeColumnType, DataTypeDateTime, IntervalKind::Kind::Day>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Day>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Week:
-                return execute<TimeDataType, TimeColumnType, DataTypeDate, IntervalKind::Kind::Week>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Week>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Month:
-                return execute<TimeDataType, TimeColumnType, DataTypeDate, IntervalKind::Kind::Month>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Month>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Quarter:
-                return execute<TimeDataType, TimeColumnType, DataTypeDate, IntervalKind::Kind::Quarter>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Quarter>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
             case IntervalKind::Kind::Year:
-                return execute<TimeDataType, TimeColumnType, DataTypeDate, IntervalKind::Kind::Year>(time_data_type, time_column, num_units, result_type, time_zone, input_rows_count, scale);
+                return execute<ReturnType, TimeDataType, TimeColumnType, IntervalKind::Kind::Year>(time_data_type, time_column, num_units, origin_column, result_type, time_zone, scale);
         }
 
         std::unreachable();
     }
 
-    template <typename TimeDataType, typename TimeColumnType, typename ResultDataType, IntervalKind::Kind unit>
-    ColumnPtr execute(
-        const TimeDataType &, const TimeColumnType & time_column_type, Int64 num_units,
-        const DataTypePtr & result_type, const DateLUTImpl & time_zone, size_t input_rows_count, UInt16 scale) const
+    template <typename ResultDataType, typename TimeDataType, typename TimeColumnType, IntervalKind::Kind unit>
+    ColumnPtr execute(const TimeDataType &, const TimeColumnType & time_column_type, Int64 num_units, const ColumnWithTypeAndName & origin_column, const DataTypePtr & result_type, const DateLUTImpl & time_zone, UInt16 scale) const
     {
         using ResultColumnType = typename ResultDataType::ColumnType;
-        using ResultFieldType = typename ResultDataType::FieldType;
 
         const auto & time_data = time_column_type.getData();
+        size_t size = time_data.size();
 
         auto result_col = result_type->createColumn();
         auto * col_to = assert_cast<ResultColumnType *>(result_col.get());
         auto & result_data = col_to->getData();
-        result_data.resize(input_rows_count);
+        result_data.resize(size);
 
         Int64 scale_multiplier = DecimalUtils::scaleMultiplier<DateTime64>(scale);
 
-        for (size_t i = 0; i != input_rows_count; ++i)
-            result_data[i] = static_cast<ResultFieldType>(ToStartOfInterval<unit>::execute(time_data[i], num_units, time_zone, scale_multiplier));
+        if (origin_column.column) // Overload: Origin
+        {
+            const bool is_small_interval = (unit == IntervalKind::Kind::Nanosecond || unit == IntervalKind::Kind::Microsecond || unit == IntervalKind::Kind::Millisecond);
+            const bool is_result_date = isDateOrDate32(result_type);
+
+            Int64 result_scale = scale_multiplier;
+            Int64 origin_scale = 1;
+
+            if (isDateTime64(result_type)) /// We have origin scale only in case if arguments are DateTime64.
+                origin_scale = assert_cast<const DataTypeDateTime64 &>(*origin_column.type).getScaleMultiplier();
+            else if (!is_small_interval) /// In case of large interval and arguments are not DateTime64, we should not have scale in result.
+                result_scale = 1;
+
+            if (is_small_interval)
+                result_scale = assert_cast<const DataTypeDateTime64 &>(*result_type).getScaleMultiplier();
+
+            /// In case if we have a difference between time arguments and Interval, we need to calculate the difference between them
+            /// to get the right precision for the result. In case of large intervals, we should not have scale difference.
+            Int64 scale_diff = is_small_interval ? std::max(result_scale / origin_scale, origin_scale / result_scale) : 1;
+
+            static constexpr Int64 SECONDS_PER_DAY = 86'400;
+
+            UInt64 origin = origin_column.column->get64(0);
+            for (size_t i = 0; i != size; ++i)
+            {
+                UInt64 time_arg = time_data[i];
+                if (origin > static_cast<size_t>(time_arg))
+                    throw Exception(ErrorCodes::BAD_ARGUMENTS, "The origin must be before the end date / date with time");
+
+                if (is_result_date) /// All internal calculations of ToStartOfInterval<...> expect arguments to be seconds or milli-, micro-, nanoseconds.
+                {
+                    time_arg *= SECONDS_PER_DAY;
+                    origin *= SECONDS_PER_DAY;
+                }
+
+                Int64 offset = ToStartOfInterval<unit>::execute(time_arg - origin, num_units, time_zone, result_scale, origin);
+
+                /// In case if arguments are DateTime64 with large interval, we should apply scale on it.
+                offset *= (!is_small_interval) ? result_scale : 1;
+
+                if (is_result_date) /// Convert back to date after calculations.
+                {
+                    offset /= SECONDS_PER_DAY;
+                    origin /= SECONDS_PER_DAY;
+                }
+
+                result_data[i] = 0;
+                result_data[i] += (result_scale < origin_scale) ? (origin + offset) / scale_diff : (origin + offset) * scale_diff;
+            }
+        }
+        else // Overload: Default
+        {
+            for (size_t i = 0; i != size; ++i)
+                result_data[i] = static_cast<typename ResultDataType::FieldType>(ToStartOfInterval<unit>::execute(time_data[i], num_units, time_zone, scale_multiplier));
+        }
 
         return result_col;
     }
diff --git a/src/IO/S3/URI.cpp b/src/IO/S3/URI.cpp
index 9c80b377661..73bbba055d0 100644
--- a/src/IO/S3/URI.cpp
+++ b/src/IO/S3/URI.cpp
@@ -36,7 +36,7 @@ URI::URI(const std::string & uri_, bool allow_archive_path_syntax)
     /// Case when bucket name represented in domain name of S3 URL.
     /// E.g. (https://bucket-name.s3.region.amazonaws.com/key)
     /// https://docs.aws.amazon.com/AmazonS3/latest/dev/VirtualHosting.html#virtual-hosted-style-access
-    static const RE2 virtual_hosted_style_pattern(R"((.+)\.(s3express[\-a-z0-9]+|s3|cos|obs|oss|eos)([.\-][a-z0-9\-.:]+))");
+    static const RE2 virtual_hosted_style_pattern(R"(([^.]+)\.(s3express[\-a-z0-9]+|s3|cos|obs|.*oss[^\/]*|eos)([.\-][a-z0-9\-.:]+))");
 
     /// Case when AWS Private Link Interface is being used
     /// E.g. (bucket.vpce-07a1cd78f1bd55c5f-j3a3vg6w.s3.us-east-1.vpce.amazonaws.com/bucket-name/key)
diff --git a/src/IO/WriteBufferFromFile.cpp b/src/IO/WriteBufferFromFile.cpp
index f1825ce1e22..d68203029c1 100644
--- a/src/IO/WriteBufferFromFile.cpp
+++ b/src/IO/WriteBufferFromFile.cpp
@@ -32,8 +32,10 @@ WriteBufferFromFile::WriteBufferFromFile(
     ThrottlerPtr throttler_,
     mode_t mode,
     char * existing_memory,
-    size_t alignment)
-    : WriteBufferFromFileDescriptor(-1, buf_size, existing_memory, throttler_, alignment, file_name_)
+    size_t alignment,
+    bool use_adaptive_buffer_size_,
+    size_t adaptive_buffer_initial_size)
+    : WriteBufferFromFileDescriptor(-1, buf_size, existing_memory, throttler_, alignment, file_name_, use_adaptive_buffer_size_, adaptive_buffer_initial_size)
 {
     ProfileEvents::increment(ProfileEvents::FileOpen);
 
@@ -66,8 +68,10 @@ WriteBufferFromFile::WriteBufferFromFile(
     size_t buf_size,
     ThrottlerPtr throttler_,
     char * existing_memory,
-    size_t alignment)
-    : WriteBufferFromFileDescriptor(fd_, buf_size, existing_memory, throttler_, alignment, original_file_name)
+    size_t alignment,
+    bool use_adaptive_buffer_size_,
+    size_t adaptive_buffer_initial_size)
+    : WriteBufferFromFileDescriptor(fd_, buf_size, existing_memory, throttler_, alignment, original_file_name, use_adaptive_buffer_size_, adaptive_buffer_initial_size)
 {
     fd_ = -1;
 }
diff --git a/src/IO/WriteBufferFromFile.h b/src/IO/WriteBufferFromFile.h
index 57847d893af..c0fa7f0b233 100644
--- a/src/IO/WriteBufferFromFile.h
+++ b/src/IO/WriteBufferFromFile.h
@@ -36,7 +36,9 @@ public:
         ThrottlerPtr throttler_ = {},
         mode_t mode = 0666,
         char * existing_memory = nullptr,
-        size_t alignment = 0);
+        size_t alignment = 0,
+        bool use_adaptive_buffer_size_ = false,
+        size_t adaptive_buffer_initial_size = DBMS_DEFAULT_INITIAL_ADAPTIVE_BUFFER_SIZE);
 
     /// Use pre-opened file descriptor.
     explicit WriteBufferFromFile(
@@ -45,7 +47,9 @@ public:
         size_t buf_size = DBMS_DEFAULT_BUFFER_SIZE,
         ThrottlerPtr throttler_ = {},
         char * existing_memory = nullptr,
-        size_t alignment = 0);
+        size_t alignment = 0,
+        bool use_adaptive_buffer_size_ = false,
+        size_t adaptive_buffer_initial_size = DBMS_DEFAULT_INITIAL_ADAPTIVE_BUFFER_SIZE);
 
     ~WriteBufferFromFile() override;
 
diff --git a/src/IO/WriteBufferFromFileDescriptor.cpp b/src/IO/WriteBufferFromFileDescriptor.cpp
index f1207edc55b..b60a792e11c 100644
--- a/src/IO/WriteBufferFromFileDescriptor.cpp
+++ b/src/IO/WriteBufferFromFileDescriptor.cpp
@@ -83,6 +83,13 @@ void WriteBufferFromFileDescriptor::nextImpl()
 
     ProfileEvents::increment(ProfileEvents::DiskWriteElapsedMicroseconds, watch.elapsedMicroseconds());
     ProfileEvents::increment(ProfileEvents::WriteBufferFromFileDescriptorWriteBytes, bytes_written);
+
+    /// Increase buffer size for next data if adaptive buffer size is used and nextImpl was called because of end of buffer.
+    if (!available() && use_adaptive_buffer_size && memory.size() < adaptive_max_buffer_size)
+    {
+        memory.resize(std::min(memory.size() * 2, adaptive_max_buffer_size));
+        BufferBase::set(memory.data(), memory.size(), 0);
+    }
 }
 
 /// NOTE: This class can be used as a very low-level building block, for example
@@ -94,11 +101,15 @@ WriteBufferFromFileDescriptor::WriteBufferFromFileDescriptor(
     char * existing_memory,
     ThrottlerPtr throttler_,
     size_t alignment,
-    std::string file_name_)
-    : WriteBufferFromFileBase(buf_size, existing_memory, alignment)
+    std::string file_name_,
+    bool use_adaptive_buffer_size_,
+    size_t adaptive_buffer_initial_size)
+    : WriteBufferFromFileBase(use_adaptive_buffer_size_ ? adaptive_buffer_initial_size : buf_size, existing_memory, alignment)
     , fd(fd_)
     , throttler(throttler_)
     , file_name(std::move(file_name_))
+    , use_adaptive_buffer_size(use_adaptive_buffer_size_)
+    , adaptive_max_buffer_size(buf_size)
 {
 }
 
@@ -124,6 +135,7 @@ void WriteBufferFromFileDescriptor::finalizeImpl()
         return;
     }
 
+    use_adaptive_buffer_size = false;
     next();
 }
 
diff --git a/src/IO/WriteBufferFromFileDescriptor.h b/src/IO/WriteBufferFromFileDescriptor.h
index cb73b1e1d08..e893ecd80fb 100644
--- a/src/IO/WriteBufferFromFileDescriptor.h
+++ b/src/IO/WriteBufferFromFileDescriptor.h
@@ -18,7 +18,9 @@ public:
         char * existing_memory = nullptr,
         ThrottlerPtr throttler_ = {},
         size_t alignment = 0,
-        std::string file_name_ = "");
+        std::string file_name_ = "",
+        bool use_adaptive_buffer_size_ = false,
+        size_t adaptive_buffer_initial_size = DBMS_DEFAULT_INITIAL_ADAPTIVE_BUFFER_SIZE);
 
     /** Could be used before initialization if needed 'fd' was not passed to constructor.
       * It's not possible to change 'fd' during work.
@@ -56,6 +58,12 @@ protected:
     /// If file has name contains filename, otherwise contains string "(fd=...)"
     std::string file_name;
 
+    /// If true, the size of internal buffer will be exponentially increased up to
+    /// adaptive_buffer_max_size after each nextImpl call. It can be used to avoid
+    /// large buffer allocation when actual size of written data is small.
+    bool use_adaptive_buffer_size;
+    size_t adaptive_max_buffer_size;
+
     void finalizeImpl() override;
 };
 
diff --git a/src/IO/WriteBufferFromS3.cpp b/src/IO/WriteBufferFromS3.cpp
index 7a978e951a7..d03b486ec52 100644
--- a/src/IO/WriteBufferFromS3.cpp
+++ b/src/IO/WriteBufferFromS3.cpp
@@ -95,7 +95,7 @@ WriteBufferFromS3::WriteBufferFromS3(
     std::optional<std::map<String, String>> object_metadata_,
     ThreadPoolCallbackRunnerUnsafe<void> schedule_,
     const WriteSettings & write_settings_)
-    : WriteBufferFromFileBase(buf_size_, nullptr, 0)
+    : WriteBufferFromFileBase(std::min(buf_size_, static_cast<size_t>(DBMS_DEFAULT_BUFFER_SIZE)), nullptr, 0)
     , bucket(bucket_)
     , key(key_)
     , request_settings(request_settings_)
@@ -351,9 +351,17 @@ void WriteBufferFromS3::allocateBuffer()
     buffer_allocation_policy->nextBuffer();
     chassert(0 == hidden_size);
 
+    /// First buffer was already allocated in BufferWithOwnMemory constructor with provided in constructor buffer size.
+    /// It will be reallocated in subsequent nextImpl calls up to the desired buffer size from buffer_allocation_policy.
     if (buffer_allocation_policy->getBufferNumber() == 1)
     {
-        allocateFirstBuffer();
+        /// Reduce memory size if initial size was larger then desired size from buffer_allocation_policy.
+        /// Usually it doesn't happen but we have it in unit tests.
+        if (memory.size() > buffer_allocation_policy->getBufferSize())
+        {
+            memory.resize(buffer_allocation_policy->getBufferSize());
+            WriteBuffer::set(memory.data(), memory.size());
+        }
         return;
     }
 
@@ -361,14 +369,6 @@ void WriteBufferFromS3::allocateBuffer()
     WriteBuffer::set(memory.data(), memory.size());
 }
 
-void WriteBufferFromS3::allocateFirstBuffer()
-{
-    const auto max_first_buffer = buffer_allocation_policy->getBufferSize();
-    const auto size = std::min(size_t(DBMS_DEFAULT_BUFFER_SIZE), max_first_buffer);
-    memory = Memory(size);
-    WriteBuffer::set(memory.data(), memory.size());
-}
-
 void WriteBufferFromS3::setFakeBufferWhenPreFinalized()
 {
     WriteBuffer::set(fake_buffer_when_prefinalized, sizeof(fake_buffer_when_prefinalized));
diff --git a/src/IO/WriteBufferFromS3.h b/src/IO/WriteBufferFromS3.h
index b026da607c5..604f036fcb8 100644
--- a/src/IO/WriteBufferFromS3.h
+++ b/src/IO/WriteBufferFromS3.h
@@ -64,7 +64,6 @@ private:
     void reallocateFirstBuffer();
     void detachBuffer();
     void allocateBuffer();
-    void allocateFirstBuffer();
     void setFakeBufferWhenPreFinalized();
 
     S3::UploadPartRequest getUploadRequest(size_t part_number, PartData & data);
diff --git a/src/IO/WriteBufferValidUTF8.cpp b/src/IO/WriteBufferValidUTF8.cpp
index d611befac37..2a86f8c2801 100644
--- a/src/IO/WriteBufferValidUTF8.cpp
+++ b/src/IO/WriteBufferValidUTF8.cpp
@@ -54,7 +54,7 @@ inline void WriteBufferValidUTF8::putReplacement()
 }
 
 
-inline void WriteBufferValidUTF8::putValid(char *data, size_t len)
+inline void WriteBufferValidUTF8::putValid(const char *data, size_t len)
 {
     if (len == 0)
         return;
@@ -149,9 +149,34 @@ void WriteBufferValidUTF8::finalizeImpl()
     /// Write all complete sequences from buffer.
     nextImpl();
 
-    /// If unfinished sequence at end, then write replacement.
+    /// Handle remaining bytes if we have an incomplete sequence
     if (working_buffer.begin() != memory.data())
-        putReplacement();
+    {
+        const char * p = memory.data();
+
+        while (p < pos)
+        {
+            UInt8 len = length_of_utf8_sequence[static_cast<const unsigned char>(*p)];
+            if (p + len > pos)
+            {
+                /// Incomplete sequence. Skip one byte.
+                putReplacement();
+                ++p;
+            }
+            else if (Poco::UTF8Encoding::isLegal(reinterpret_cast<const unsigned char *>(p), len))
+            {
+                /// Valid sequence
+                putValid(p, len);
+                p += len;
+            }
+            else
+            {
+                /// Invalid sequence, skip first byte.
+                putReplacement();
+                ++p;
+            }
+        }
+    }
 }
 
 }
diff --git a/src/IO/WriteBufferValidUTF8.h b/src/IO/WriteBufferValidUTF8.h
index daaf0427f88..a398b8ded01 100644
--- a/src/IO/WriteBufferValidUTF8.h
+++ b/src/IO/WriteBufferValidUTF8.h
@@ -26,7 +26,7 @@ public:
 
 private:
     void putReplacement();
-    void putValid(char * data, size_t len);
+    void putValid(const char * data, size_t len);
 
     void nextImpl() override;
     void finalizeImpl() override;
diff --git a/src/IO/WriteSettings.h b/src/IO/WriteSettings.h
index cdc75e8c0e9..6b3d04f4e5c 100644
--- a/src/IO/WriteSettings.h
+++ b/src/IO/WriteSettings.h
@@ -24,6 +24,9 @@ struct WriteSettings
     bool s3_allow_parallel_part_upload = true;
     bool azure_allow_parallel_part_upload = true;
 
+    bool use_adaptive_write_buffer = false;
+    size_t adaptive_write_buffer_initial_size = 16 * 1024;
+
     bool operator==(const WriteSettings & other) const = default;
 };
 
diff --git a/src/IO/tests/gtest_s3_uri.cpp b/src/IO/tests/gtest_s3_uri.cpp
index c0bf7fcb28a..abe80db7ba5 100644
--- a/src/IO/tests/gtest_s3_uri.cpp
+++ b/src/IO/tests/gtest_s3_uri.cpp
@@ -204,6 +204,14 @@ TEST(S3UriTest, validPatterns)
         ASSERT_EQ("", uri.version_id);
         ASSERT_EQ(true, uri.is_virtual_hosted_style);
     }
+    {
+        S3::URI uri("https://bucket-test.cn-beijing-internal.oss-data-acc.aliyuncs.com/cc-2zeh496zqm0g6e09g");
+        ASSERT_EQ("https://cn-beijing-internal.oss-data-acc.aliyuncs.com", uri.endpoint);
+        ASSERT_EQ("bucket-test", uri.bucket);
+        ASSERT_EQ("cc-2zeh496zqm0g6e09g", uri.key);
+        ASSERT_EQ("", uri.version_id);
+        ASSERT_EQ(true, uri.is_virtual_hosted_style);
+    }
 }
 
 TEST(S3UriTest, versionIdChecks)
diff --git a/src/Interpreters/Aggregator.cpp b/src/Interpreters/Aggregator.cpp
index d1aa8a0fff0..aff06c490c5 100644
--- a/src/Interpreters/Aggregator.cpp
+++ b/src/Interpreters/Aggregator.cpp
@@ -2371,7 +2371,7 @@ void NO_INLINE Aggregator::mergeDataNullKey(
 
 template <typename Method, typename Table>
 void NO_INLINE Aggregator::mergeDataImpl(
-    Table & table_dst, Table & table_src, Arena * arena, bool use_compiled_functions [[maybe_unused]], bool prefetch) const
+    Table & table_dst, Table & table_src, Arena * arena, bool use_compiled_functions [[maybe_unused]], bool prefetch, ThreadPool & thread_pool, std::atomic<bool> & is_cancelled) const
 {
     if constexpr (Method::low_cardinality_optimization || Method::one_key_nullable_optimization)
         mergeDataNullKey<Method, Table>(table_dst, table_src, arena);
@@ -2410,7 +2410,7 @@ void NO_INLINE Aggregator::mergeDataImpl(
         {
             if (!is_aggregate_function_compiled[i])
                 aggregate_functions[i]->mergeAndDestroyBatch(
-                    dst_places.data(), src_places.data(), dst_places.size(), offsets_of_aggregate_states[i], arena);
+                    dst_places.data(), src_places.data(), dst_places.size(), offsets_of_aggregate_states[i], thread_pool, is_cancelled, arena);
         }
 
         return;
@@ -2420,7 +2420,7 @@ void NO_INLINE Aggregator::mergeDataImpl(
     for (size_t i = 0; i < params.aggregates_size; ++i)
     {
         aggregate_functions[i]->mergeAndDestroyBatch(
-            dst_places.data(), src_places.data(), dst_places.size(), offsets_of_aggregate_states[i], arena);
+            dst_places.data(), src_places.data(), dst_places.size(), offsets_of_aggregate_states[i], thread_pool, is_cancelled, arena);
     }
 }
 
@@ -2535,8 +2535,10 @@ void NO_INLINE Aggregator::mergeWithoutKeyDataImpl(
 
 template <typename Method>
 void NO_INLINE Aggregator::mergeSingleLevelDataImpl(
-    ManyAggregatedDataVariants & non_empty_data) const
+    ManyAggregatedDataVariants & non_empty_data, std::atomic<bool> & is_cancelled) const
 {
+    ThreadPool thread_pool{CurrentMetrics::AggregatorThreads, CurrentMetrics::AggregatorThreadsActive, CurrentMetrics::AggregatorThreadsScheduled, params.max_threads};
+
     AggregatedDataVariantsPtr & res = non_empty_data[0];
     bool no_more_keys = false;
 
@@ -2557,13 +2559,13 @@ void NO_INLINE Aggregator::mergeSingleLevelDataImpl(
             if (compiled_aggregate_functions_holder)
             {
                 mergeDataImpl<Method>(
-                    getDataVariant<Method>(*res).data, getDataVariant<Method>(current).data, res->aggregates_pool, true, prefetch);
+                    getDataVariant<Method>(*res).data, getDataVariant<Method>(current).data, res->aggregates_pool, true, prefetch, thread_pool, is_cancelled);
             }
             else
 #endif
             {
                 mergeDataImpl<Method>(
-                    getDataVariant<Method>(*res).data, getDataVariant<Method>(current).data, res->aggregates_pool, false, prefetch);
+                    getDataVariant<Method>(*res).data, getDataVariant<Method>(current).data, res->aggregates_pool, false, prefetch, thread_pool, is_cancelled);
             }
         }
         else if (res->without_key)
@@ -2589,7 +2591,7 @@ void NO_INLINE Aggregator::mergeSingleLevelDataImpl(
 
 #define M(NAME) \
     template void NO_INLINE Aggregator::mergeSingleLevelDataImpl<decltype(AggregatedDataVariants::NAME)::element_type>( \
-        ManyAggregatedDataVariants & non_empty_data) const;
+        ManyAggregatedDataVariants & non_empty_data, std::atomic<bool> & is_cancelled) const;
     APPLY_FOR_VARIANTS_SINGLE_LEVEL(M)
 #undef M
 
@@ -2597,6 +2599,8 @@ template <typename Method>
 void NO_INLINE Aggregator::mergeBucketImpl(
     ManyAggregatedDataVariants & data, Int32 bucket, Arena * arena, std::atomic<bool> & is_cancelled) const
 {
+    ThreadPool thread_pool{CurrentMetrics::AggregatorThreads, CurrentMetrics::AggregatorThreadsActive, CurrentMetrics::AggregatorThreadsScheduled, params.max_threads};
+
     /// We merge all aggregation results to the first.
     AggregatedDataVariantsPtr & res = data[0];
 
@@ -2613,7 +2617,7 @@ void NO_INLINE Aggregator::mergeBucketImpl(
         if (compiled_aggregate_functions_holder)
         {
             mergeDataImpl<Method>(
-                getDataVariant<Method>(*res).data.impls[bucket], getDataVariant<Method>(current).data.impls[bucket], arena, true, prefetch);
+                getDataVariant<Method>(*res).data.impls[bucket], getDataVariant<Method>(current).data.impls[bucket], arena, true, prefetch, thread_pool, is_cancelled);
         }
         else
 #endif
@@ -2623,7 +2627,9 @@ void NO_INLINE Aggregator::mergeBucketImpl(
                 getDataVariant<Method>(current).data.impls[bucket],
                 arena,
                 false,
-                prefetch);
+                prefetch,
+                thread_pool,
+                is_cancelled);
         }
     }
 }
diff --git a/src/Interpreters/Aggregator.h b/src/Interpreters/Aggregator.h
index 2cb04fc7c51..4de0a640219 100644
--- a/src/Interpreters/Aggregator.h
+++ b/src/Interpreters/Aggregator.h
@@ -467,7 +467,7 @@ private:
 
     /// Merge data from hash table `src` into `dst`.
     template <typename Method, typename Table>
-    void mergeDataImpl(Table & table_dst, Table & table_src, Arena * arena, bool use_compiled_functions, bool prefetch) const;
+    void mergeDataImpl(Table & table_dst, Table & table_src, Arena * arena, bool use_compiled_functions, bool prefetch, ThreadPool & thread_pool, std::atomic<bool> & is_cancelled) const;
 
     /// Merge data from hash table `src` into `dst`, but only for keys that already exist in dst. In other cases, merge the data into `overflows`.
     template <typename Method, typename Table>
@@ -490,7 +490,7 @@ private:
 
     template <typename Method>
     void mergeSingleLevelDataImpl(
-        ManyAggregatedDataVariants & non_empty_data) const;
+        ManyAggregatedDataVariants & non_empty_data, std::atomic<bool> & is_cancelled) const;
 
     template <bool return_single_block>
     using ConvertToBlockRes = std::conditional_t<return_single_block, Block, BlocksList>;
diff --git a/src/Interpreters/ClusterProxy/executeQuery.cpp b/src/Interpreters/ClusterProxy/executeQuery.cpp
index d04a73e384e..771c6a89caa 100644
--- a/src/Interpreters/ClusterProxy/executeQuery.cpp
+++ b/src/Interpreters/ClusterProxy/executeQuery.cpp
@@ -28,6 +28,7 @@
 #include <Storages/StorageReplicatedMergeTree.h>
 #include <Storages/StorageSnapshot.h>
 #include <Storages/buildQueryTreeForShard.h>
+#include <Processors/QueryPlan/ParallelReplicasLocalPlan.h>
 #include <Storages/getStructureOfRemoteTable.h>
 
 namespace DB
@@ -38,6 +39,7 @@ namespace ErrorCodes
     extern const int TOO_LARGE_DISTRIBUTED_DEPTH;
     extern const int LOGICAL_ERROR;
     extern const int UNEXPECTED_CLUSTER;
+    extern const int INCONSISTENT_CLUSTER_DEFINITION;
 }
 
 namespace ClusterProxy
@@ -439,7 +441,8 @@ void executeQueryWithParallelReplicas(
     QueryProcessingStage::Enum processed_stage,
     const ASTPtr & query_ast,
     ContextPtr context,
-    std::shared_ptr<const StorageLimitsList> storage_limits)
+    std::shared_ptr<const StorageLimitsList> storage_limits,
+    QueryPlanStepPtr analyzed_read_from_merge_tree)
 {
     auto logger = getLogger("executeQueryWithParallelReplicas");
     LOG_DEBUG(logger, "Executing read from {}, header {}, query ({}), stage {} with parallel replicas",
@@ -485,12 +488,12 @@ void executeQueryWithParallelReplicas(
         shard_num = column->getUInt(0);
     }
 
-    const auto shard_count = not_optimized_cluster->getShardCount();
     ClusterPtr new_cluster = not_optimized_cluster;
     /// if got valid shard_num from query initiator, then parallel replicas scope is the specified shard
     /// shards are numbered in order of appearance in the cluster config
     if (shard_num > 0)
     {
+        const auto shard_count = not_optimized_cluster->getShardCount();
         if (shard_num > shard_count)
             throw Exception(
                 ErrorCodes::LOGICAL_ERROR,
@@ -516,21 +519,147 @@ void executeQueryWithParallelReplicas(
                 "`cluster_for_parallel_replicas` setting refers to cluster with several shards. Expected a cluster with one shard");
     }
 
-    auto external_tables = new_context->getExternalTables();
-    auto read_from_remote = std::make_unique<ReadFromParallelRemoteReplicasStep>(
-        query_ast,
-        new_cluster,
-        storage_id,
-        header,
-        processed_stage,
-        new_context,
-        getThrottler(new_context),
-        std::move(scalars),
-        std::move(external_tables),
-        getLogger("ReadFromParallelRemoteReplicasStep"),
-        std::move(storage_limits));
+    const auto & shard = new_cluster->getShardsInfo().at(0);
+    size_t max_replicas_to_use = settings.max_parallel_replicas;
+    if (max_replicas_to_use > shard.getAllNodeCount())
+    {
+        LOG_INFO(
+            getLogger("ReadFromParallelRemoteReplicasStep"),
+            "The number of replicas requested ({}) is bigger than the real number available in the cluster ({}). "
+            "Will use the latter number to execute the query.",
+            settings.max_parallel_replicas,
+            shard.getAllNodeCount());
+        max_replicas_to_use = shard.getAllNodeCount();
+    }
 
-    query_plan.addStep(std::move(read_from_remote));
+    auto coordinator = std::make_shared<ParallelReplicasReadingCoordinator>(max_replicas_to_use, settings.parallel_replicas_mark_segment_size);
+
+    auto external_tables = new_context->getExternalTables();
+
+    std::vector<ConnectionPoolWithFailover::Base::ShuffledPool> shuffled_pool;
+    if (max_replicas_to_use < shard.getAllNodeCount())
+    {
+        // will be shuffled according to `load_balancing` setting
+        shuffled_pool = shard.pool->getShuffledPools(settings);
+    }
+    else
+    {
+        /// If all replicas in cluster are used for query execution,
+        /// try to preserve replicas order as in cluster definition.
+        /// It's important for data locality during query execution
+        /// independently of the query initiator
+        auto priority_func = [](size_t i) { return Priority{static_cast<Int64>(i)}; };
+        shuffled_pool = shard.pool->getShuffledPools(settings, priority_func);
+    }
+
+    std::vector<ConnectionPoolPtr> pools_to_use;
+    pools_to_use.reserve(shuffled_pool.size());
+    for (auto & pool : shuffled_pool)
+        pools_to_use.emplace_back(std::move(pool.pool));
+
+    /// do not build local plan for distributed queries for now (address it later)
+    if (settings.allow_experimental_analyzer && settings.parallel_replicas_local_plan && !shard_num)
+    {
+        /// find local replica index in pool
+        std::optional<size_t> local_replica_index;
+        for (size_t i = 0, s = pools_to_use.size(); i < s; ++i)
+        {
+            const auto & hostname = pools_to_use[i]->getHost();
+            const auto found = std::find_if(
+                begin(shard.local_addresses),
+                end(shard.local_addresses),
+                [&hostname](const Cluster::Address & local_addr) { return hostname == local_addr.host_name; });
+            if (found != shard.local_addresses.end())
+            {
+                local_replica_index = i;
+                break;
+            }
+        }
+        if (!local_replica_index)
+            throw Exception(
+                ErrorCodes::INCONSISTENT_CLUSTER_DEFINITION,
+                "Local replica is not found in '{}' cluster definition, see 'cluster_for_parallel_replicas' setting",
+                new_cluster->getName());
+
+        // resize the pool but keep local replicas in it (and update its index)
+        chassert(max_replicas_to_use <= pools_to_use.size());
+        if (local_replica_index >= max_replicas_to_use)
+        {
+            std::swap(pools_to_use[max_replicas_to_use - 1], pools_to_use[local_replica_index.value()]);
+            local_replica_index = max_replicas_to_use - 1;
+        }
+        pools_to_use.resize(max_replicas_to_use);
+
+        auto [local_plan, with_parallel_replicas] = createLocalPlanForParallelReplicas(
+            query_ast,
+            header,
+            new_context,
+            processed_stage,
+            coordinator,
+            std::move(analyzed_read_from_merge_tree),
+            local_replica_index.value());
+
+        if (!with_parallel_replicas)
+        {
+            query_plan = std::move(*local_plan);
+            return;
+        }
+
+        LOG_DEBUG(logger, "Local replica got replica number {}", local_replica_index.value());
+
+        auto read_from_remote = std::make_unique<ReadFromParallelRemoteReplicasStep>(
+            query_ast,
+            new_cluster,
+            storage_id,
+            coordinator,
+            header,
+            processed_stage,
+            new_context,
+            getThrottler(new_context),
+            std::move(scalars),
+            std::move(external_tables),
+            getLogger("ReadFromParallelRemoteReplicasStep"),
+            std::move(storage_limits),
+            std::move(pools_to_use),
+            local_replica_index);
+
+        auto remote_plan = std::make_unique<QueryPlan>();
+        remote_plan->addStep(std::move(read_from_remote));
+
+        DataStreams input_streams;
+        input_streams.reserve(2);
+        input_streams.emplace_back(local_plan->getCurrentDataStream());
+        input_streams.emplace_back(remote_plan->getCurrentDataStream());
+
+        std::vector<QueryPlanPtr> plans;
+        plans.emplace_back(std::move(local_plan));
+        plans.emplace_back(std::move(remote_plan));
+
+        auto union_step = std::make_unique<UnionStep>(std::move(input_streams));
+        query_plan.unitePlans(std::move(union_step), std::move(plans));
+    }
+    else
+    {
+        chassert(max_replicas_to_use <= pools_to_use.size());
+        pools_to_use.resize(max_replicas_to_use);
+
+        auto read_from_remote = std::make_unique<ReadFromParallelRemoteReplicasStep>(
+            query_ast,
+            new_cluster,
+            storage_id,
+            std::move(coordinator),
+            header,
+            processed_stage,
+            new_context,
+            getThrottler(new_context),
+            std::move(scalars),
+            std::move(external_tables),
+            getLogger("ReadFromParallelRemoteReplicasStep"),
+            std::move(storage_limits),
+            std::move(pools_to_use));
+
+        query_plan.addStep(std::move(read_from_remote));
+    }
 }
 
 void executeQueryWithParallelReplicas(
@@ -540,7 +669,8 @@ void executeQueryWithParallelReplicas(
     const QueryTreeNodePtr & query_tree,
     const PlannerContextPtr & planner_context,
     ContextPtr context,
-    std::shared_ptr<const StorageLimitsList> storage_limits)
+    std::shared_ptr<const StorageLimitsList> storage_limits,
+    QueryPlanStepPtr analyzed_read_from_merge_tree)
 {
     QueryTreeNodePtr modified_query_tree = query_tree->clone();
     rewriteJoinToGlobalJoin(modified_query_tree, context);
@@ -550,7 +680,8 @@ void executeQueryWithParallelReplicas(
         = InterpreterSelectQueryAnalyzer::getSampleBlock(modified_query_tree, context, SelectQueryOptions(processed_stage).analyze());
     auto modified_query_ast = queryNodeToDistributedSelectQuery(modified_query_tree);
 
-    executeQueryWithParallelReplicas(query_plan, storage_id, header, processed_stage, modified_query_ast, context, storage_limits);
+    executeQueryWithParallelReplicas(
+        query_plan, storage_id, header, processed_stage, modified_query_ast, context, storage_limits, std::move(analyzed_read_from_merge_tree));
 }
 
 void executeQueryWithParallelReplicas(
@@ -646,6 +777,49 @@ void executeQueryWithParallelReplicasCustomKey(
         context, query_info.query, storage_id.getDatabaseName(), storage_id.getTableName(), /*table_function_ptr=*/nullptr);
     executeQueryWithParallelReplicasCustomKey(query_plan, storage_id, query_info, columns, snapshot, processed_stage, header, context);
 }
+
+bool canUseParallelReplicasOnInitiator(const ContextPtr & context)
+{
+    if (!context->canUseParallelReplicasOnInitiator())
+        return false;
+
+    auto cluster = context->getClusterForParallelReplicas();
+    if (cluster->getShardCount() == 1)
+        return cluster->getShardsInfo()[0].getAllNodeCount() > 1;
+
+    /// parallel replicas with distributed table
+    auto scalars = context->hasQueryContext() ? context->getQueryContext()->getScalars() : Scalars{};
+    UInt64 shard_num = 0; /// shard_num is 1-based, so 0 - no shard specified
+    const auto it = scalars.find("_shard_num");
+    if (it != scalars.end())
+    {
+        const Block & block = it->second;
+        const auto & column = block.safeGetByPosition(0).column;
+        shard_num = column->getUInt(0);
+    }
+    if (shard_num > 0)
+    {
+        const auto shard_count = cluster->getShardCount();
+        if (shard_num > shard_count)
+            throw Exception(
+                ErrorCodes::LOGICAL_ERROR,
+                "Shard number is greater than shard count: shard_num={} shard_count={} cluster={}",
+                shard_num,
+                shard_count,
+                cluster->getName());
+
+        return cluster->getShardsInfo().at(shard_num - 1).getAllNodeCount() > 1;
+    }
+
+    if (cluster->getShardCount() > 1)
+        throw DB::Exception(
+            ErrorCodes::UNEXPECTED_CLUSTER,
+            "`cluster_for_parallel_replicas` setting refers to cluster with {} shards. Expected a cluster with one shard",
+            cluster->getShardCount());
+
+    return false;
+}
+
 }
 
 }
diff --git a/src/Interpreters/ClusterProxy/executeQuery.h b/src/Interpreters/ClusterProxy/executeQuery.h
index c22fcd24f03..2a21f3e8255 100644
--- a/src/Interpreters/ClusterProxy/executeQuery.h
+++ b/src/Interpreters/ClusterProxy/executeQuery.h
@@ -35,6 +35,9 @@ using QueryTreeNodePtr = std::shared_ptr<IQueryTreeNode>;
 class PlannerContext;
 using PlannerContextPtr = std::shared_ptr<PlannerContext>;
 
+class IQueryPlanStep;
+using QueryPlanStepPtr = std::unique_ptr<IQueryPlanStep>;
+
 namespace ClusterProxy
 {
 
@@ -55,6 +58,8 @@ using AdditionalShardFilterGenerator = std::function<ASTPtr(uint64_t)>;
 AdditionalShardFilterGenerator
 getShardFilterGeneratorForCustomKey(const Cluster & cluster, ContextPtr context, const ColumnsDescription & columns);
 
+bool canUseParallelReplicasOnInitiator(const ContextPtr & context);
+
 /// Execute a distributed query, creating a query plan, from which the query pipeline can be built.
 /// `stream_factory` object encapsulates the logic of creating plans for a different type of query
 /// (currently SELECT, DESCRIBE).
@@ -81,7 +86,8 @@ void executeQueryWithParallelReplicas(
     QueryProcessingStage::Enum processed_stage,
     const ASTPtr & query_ast,
     ContextPtr context,
-    std::shared_ptr<const StorageLimitsList> storage_limits);
+    std::shared_ptr<const StorageLimitsList> storage_limits,
+    QueryPlanStepPtr read_from_merge_tree = nullptr);
 
 void executeQueryWithParallelReplicas(
     QueryPlan & query_plan,
@@ -98,7 +104,8 @@ void executeQueryWithParallelReplicas(
     const QueryTreeNodePtr & query_tree,
     const PlannerContextPtr & planner_context,
     ContextPtr context,
-    std::shared_ptr<const StorageLimitsList> storage_limits);
+    std::shared_ptr<const StorageLimitsList> storage_limits,
+    QueryPlanStepPtr read_from_merge_tree);
 
 void executeQueryWithParallelReplicasCustomKey(
     QueryPlan & query_plan,
diff --git a/src/Interpreters/Context.cpp b/src/Interpreters/Context.cpp
index 373cc91ebcb..7adfb42fb51 100644
--- a/src/Interpreters/Context.cpp
+++ b/src/Interpreters/Context.cpp
@@ -5,6 +5,7 @@
 #include <Poco/UUID.h>
 #include <Poco/Util/Application.h>
 #include <Common/AsyncLoader.h>
+#include <Common/CgroupsMemoryUsageObserver.h>
 #include <Common/PoolId.h>
 #include <Common/SensitiveDataMasker.h>
 #include <Common/Macros.h>
diff --git a/src/Interpreters/Context.h b/src/Interpreters/Context.h
index fb5337158ba..858b4a78430 100644
--- a/src/Interpreters/Context.h
+++ b/src/Interpreters/Context.h
@@ -152,6 +152,7 @@ class ServerType;
 template <class Queue>
 class MergeTreeBackgroundExecutor;
 class AsyncLoader;
+struct ICgroupsReader;
 
 struct TemporaryTableHolder;
 using TemporaryTablesMapping = std::map<String, std::shared_ptr<TemporaryTableHolder>>;
diff --git a/src/Interpreters/HashJoin/AddedColumns.cpp b/src/Interpreters/HashJoin/AddedColumns.cpp
index 21cb6e401ed..2e1ecb0da72 100644
--- a/src/Interpreters/HashJoin/AddedColumns.cpp
+++ b/src/Interpreters/HashJoin/AddedColumns.cpp
@@ -34,6 +34,23 @@ void AddedColumns<true>::buildOutput()
     {
         if (join_data_avg_perkey_rows < output_by_row_list_threshold)
             buildOutputFromBlocks<true>();
+        else if (join_data_sorted)
+        {
+            for (size_t i = 0; i < this->size(); ++i)
+            {
+                auto & col = columns[i];
+                for (auto row_ref_i : lazy_output.row_refs)
+                {
+                    if (row_ref_i)
+                    {
+                        const RowRefList * row_ref_list = reinterpret_cast<const RowRefList *>(row_ref_i);
+                        col->insertRangeFrom(*row_ref_list->block->getByPosition(right_indexes[i]).column, row_ref_list->row_num, row_ref_list->rows);
+                    }
+                    else
+                        type_name[i].type->insertDefaultInto(*col);
+                }
+            }
+        }
         else
         {
             for (size_t i = 0; i < this->size(); ++i)
diff --git a/src/Interpreters/HashJoin/AddedColumns.h b/src/Interpreters/HashJoin/AddedColumns.h
index f1b95a63be6..4603d493329 100644
--- a/src/Interpreters/HashJoin/AddedColumns.h
+++ b/src/Interpreters/HashJoin/AddedColumns.h
@@ -66,6 +66,9 @@ public:
         , join_on_keys(join_on_keys_)
         , additional_filter_expression(additional_filter_expression_)
         , rows_to_add(left_block.rows())
+        , join_data_avg_perkey_rows(join.getJoinedData()->avgPerKeyRows())
+        , output_by_row_list_threshold(join.getTableJoin().outputByRowListPerkeyRowsThreshold())
+        , join_data_sorted(join.getJoinedData()->sorted)
         , is_join_get(is_join_get_)
     {
         size_t num_columns_to_add = block_with_columns_to_add.columns();
@@ -113,8 +116,6 @@ public:
             if (columns[j]->isNullable() && !saved_column->isNullable())
                 nullable_column_ptrs[j] = typeid_cast<ColumnNullable *>(columns[j].get());
         }
-        join_data_avg_perkey_rows = join.getJoinedData()->avgPerKeyRows();
-        output_by_row_list_threshold = join.getTableJoin().outputByRowListPerkeyRowsThreshold();
     }
 
     size_t size() const { return columns.size(); }
@@ -149,6 +150,7 @@ public:
     bool output_by_row_list = false;
     size_t join_data_avg_perkey_rows = 0;
     size_t output_by_row_list_threshold = 0;
+    bool join_data_sorted = false;
     IColumn::Filter filter;
 
     void reserve(bool need_replicate)
diff --git a/src/Interpreters/HashJoin/HashJoin.cpp b/src/Interpreters/HashJoin/HashJoin.cpp
index 9c07a71e614..1b8b45b94ea 100644
--- a/src/Interpreters/HashJoin/HashJoin.cpp
+++ b/src/Interpreters/HashJoin/HashJoin.cpp
@@ -649,7 +649,6 @@ bool HashJoin::addBlockToJoin(const Block & source_block_, bool check_limits)
     }
     data->keys_to_join = total_rows;
     shrinkStoredBlocksToFit(total_bytes);
-
     return table_join->sizeLimits().check(total_rows, total_bytes, "JOIN", ErrorCodes::SET_SIZE_LIMIT_EXCEEDED);
 }
 
@@ -1361,4 +1360,96 @@ bool HashJoin::needUsedFlagsForPerRightTableRow(std::shared_ptr<TableJoin> table
     return false;
 }
 
+template <JoinKind KIND, typename Map, JoinStrictness STRICTNESS>
+void HashJoin::tryRerangeRightTableDataImpl(Map & map [[maybe_unused]])
+{
+    constexpr JoinFeatures<KIND, STRICTNESS, Map> join_features;
+    if constexpr (!join_features.is_all_join || (!join_features.left && !join_features.inner))
+        throw Exception(ErrorCodes::LOGICAL_ERROR, "Only left or inner join table can be reranged.");
+    else
+    {
+        auto merge_rows_into_one_block = [&](BlocksList & blocks, RowRefList & rows_ref)
+        {
+            auto it = rows_ref.begin();
+            if (it.ok())
+            {
+                if (blocks.empty() || blocks.back().rows() >= DEFAULT_BLOCK_SIZE)
+                    blocks.emplace_back(it->block->cloneEmpty());
+            }
+            else
+            {
+                return;
+            }
+            auto & block = blocks.back();
+            size_t start_row = block.rows();
+            for (; it.ok(); ++it)
+            {
+                for (size_t i = 0; i < block.columns(); ++i)
+                {
+                    auto & col = block.getByPosition(i).column->assumeMutableRef();
+                    col.insertFrom(*it->block->getByPosition(i).column, it->row_num);
+                }
+            }
+            if (block.rows() > start_row)
+            {
+                RowRefList new_rows_ref(&block, start_row, block.rows() - start_row);
+                rows_ref = std::move(new_rows_ref);
+            }
+        };
+
+        auto visit_rows_map = [&](BlocksList & blocks, MapsAll & rows_map)
+        {
+            switch (data->type)
+            {
+        #define M(TYPE) \
+                case Type::TYPE: \
+                {\
+                    rows_map.TYPE->forEachMapped([&](RowRefList & rows_ref) { merge_rows_into_one_block(blocks, rows_ref); }); \
+                    break; \
+                }
+                APPLY_FOR_JOIN_VARIANTS(M)
+        #undef M
+                default:
+                    break;
+            }
+        };
+        BlocksList sorted_blocks;
+        visit_rows_map(sorted_blocks, map);
+        data->blocks.swap(sorted_blocks);
+    }
+}
+
+void HashJoin::tryRerangeRightTableData()
+{
+    if (!table_join->allowJoinSorting() || table_join->getMixedJoinExpression() || !isInnerOrLeft(kind) || strictness != JoinStrictness::All)
+        return;
+
+    /// We should not rerange the right table on such conditions:
+    /// 1. the right table is already reranged by key or it is empty.
+    /// 2. the join clauses size is greater than 1, like `...join on a.key1=b.key1 or a.key2=b.key2`, we can not rerange the right table on different set of keys.
+    /// 3. the number of right table rows exceed the threshold, which may result in a significant cost for reranging and lead to performance degradation.
+    /// 4. the keys of right table is very sparse, which may result in insignificant performance improvement after reranging by key.
+    if (!data || data->sorted || data->blocks.empty() || data->maps.size() > 1 || data->rows_to_join > table_join->sortRightMaximumTableRows() ||  data->avgPerKeyRows() < table_join->sortRightMinimumPerkeyRows())
+        return;
+
+    if (data->keys_to_join == 0)
+        data->keys_to_join = getTotalRowCount();
+
+    /// If the there is no columns to add, means no columns to output, then the rerange would not improve performance by using column's `insertRangeFrom`
+    /// to replace column's `insertFrom` to make the output.
+    if (sample_block_with_columns_to_add.columns() == 0)
+    {
+        LOG_DEBUG(log, "The joined right table total rows :{}, total keys :{}", data->rows_to_join, data->keys_to_join);
+        return;
+    }
+    [[maybe_unused]] bool result = joinDispatch(
+        kind,
+        strictness,
+        data->maps.front(),
+        /*prefer_use_maps_all*/ false,
+        [&](auto kind_, auto strictness_, auto & map_) { tryRerangeRightTableDataImpl<kind_, decltype(map_), strictness_>(map_); });
+    chassert(result);
+    data->sorted = true;
+}
+
 }
diff --git a/src/Interpreters/HashJoin/HashJoin.h b/src/Interpreters/HashJoin/HashJoin.h
index d645b8e9273..230343691ea 100644
--- a/src/Interpreters/HashJoin/HashJoin.h
+++ b/src/Interpreters/HashJoin/HashJoin.h
@@ -345,11 +345,12 @@ public:
 
         size_t blocks_allocated_size = 0;
         size_t blocks_nullmaps_allocated_size = 0;
-
         /// Number of rows of right table to join
         size_t rows_to_join = 0;
         /// Number of keys of right table to join
         size_t keys_to_join = 0;
+        /// Whether the right table reranged by key
+        bool sorted = false;
 
         size_t avgPerKeyRows() const
         {
@@ -465,6 +466,10 @@ private:
 
     void validateAdditionalFilterExpression(std::shared_ptr<ExpressionActions> additional_filter_expression);
     bool needUsedFlagsForPerRightTableRow(std::shared_ptr<TableJoin> table_join_) const;
+
+    void tryRerangeRightTableData() override;
+    template <JoinKind KIND, typename Map, JoinStrictness STRICTNESS>
+    void tryRerangeRightTableDataImpl(Map & map);
 };
 
 }
diff --git a/src/Interpreters/HashJoin/HashJoinMethods.h b/src/Interpreters/HashJoin/HashJoinMethods.h
index 97ad57d26ea..c5b54a62f36 100644
--- a/src/Interpreters/HashJoin/HashJoinMethods.h
+++ b/src/Interpreters/HashJoin/HashJoinMethods.h
@@ -83,7 +83,6 @@ public:
         const Block & block_with_columns_to_add,
         const MapsTemplateVector & maps_,
         bool is_join_get = false);
-
 private:
     template <typename KeyGetter, bool is_asof_join>
     static KeyGetter createKeyGetter(const ColumnRawPtrs & key_columns, const Sizes & key_sizes);
@@ -199,4 +198,3 @@ extern template class HashJoinMethods<JoinKind::Full, JoinStrictness::Semi, Hash
 extern template class HashJoinMethods<JoinKind::Full, JoinStrictness::Anti, HashJoin::MapsOne>;
 extern template class HashJoinMethods<JoinKind::Full, JoinStrictness::Asof, HashJoin::MapsAsof>;
 }
-
diff --git a/src/Interpreters/IJoin.h b/src/Interpreters/IJoin.h
index 7374348da50..8f648de2538 100644
--- a/src/Interpreters/IJoin.h
+++ b/src/Interpreters/IJoin.h
@@ -115,6 +115,7 @@ public:
     /// Peek next stream of delayed joined blocks.
     virtual IBlocksStreamPtr getDelayedBlocks() { return nullptr; }
     virtual bool hasDelayedBlocks() const { return false; }
+    virtual void tryRerangeRightTableData() {}
 
     virtual IBlocksStreamPtr
         getNonJoinedBlocks(const Block & left_sample_block, const Block & result_sample_block, UInt64 max_block_size) const = 0;
diff --git a/src/Interpreters/InterpreterCreateQuery.cpp b/src/Interpreters/InterpreterCreateQuery.cpp
index d83c148ebad..c1f9b4637f8 100644
--- a/src/Interpreters/InterpreterCreateQuery.cpp
+++ b/src/Interpreters/InterpreterCreateQuery.cpp
@@ -228,8 +228,8 @@ BlockIO InterpreterCreateQuery::createDatabase(ASTCreateQuery & create)
 
         metadata_path = metadata_path / "store" / DatabaseCatalog::getPathForUUID(create.uuid);
 
-        if (!create.attach && fs::exists(metadata_path))
-            throw Exception(ErrorCodes::DATABASE_ALREADY_EXISTS, "Metadata directory {} already exists", metadata_path.string());
+        if (!create.attach && fs::exists(metadata_path) && !fs::is_empty(metadata_path))
+            throw Exception(ErrorCodes::DATABASE_ALREADY_EXISTS, "Metadata directory {} already exists and is not empty", metadata_path.string());
     }
     else if (create.storage->engine->name == "MaterializeMySQL"
         || create.storage->engine->name == "MaterializedMySQL")
@@ -329,6 +329,9 @@ BlockIO InterpreterCreateQuery::createDatabase(ASTCreateQuery & create)
         writeChar('\n', statement_buf);
         String statement = statement_buf.str();
 
+        /// Needed to make database creation retriable if it fails after the file is created
+        fs::remove(metadata_file_tmp_path);
+
         /// Exclusive flag guarantees, that database is not created right now in another thread.
         WriteBufferFromFile out(metadata_file_tmp_path, statement.size(), O_WRONLY | O_CREAT | O_EXCL);
         writeString(statement, out);
@@ -350,13 +353,6 @@ BlockIO InterpreterCreateQuery::createDatabase(ASTCreateQuery & create)
         DatabaseCatalog::instance().attachDatabase(database_name, database);
         added = true;
 
-        if (need_write_metadata)
-        {
-            /// Prevents from overwriting metadata of detached database
-            renameNoReplace(metadata_file_tmp_path, metadata_file_path);
-            renamed = true;
-        }
-
         if (!load_database_without_tables)
         {
             /// We use global context here, because storages lifetime is bigger than query context lifetime
@@ -368,6 +364,13 @@ BlockIO InterpreterCreateQuery::createDatabase(ASTCreateQuery & create)
             /// Only then prioritize, schedule and wait all the startup tasks
             waitLoad(currentPoolOr(TablesLoaderForegroundPoolId), startup_tasks);
         }
+
+        if (need_write_metadata)
+        {
+            /// Prevents from overwriting metadata of detached database
+            renameNoReplace(metadata_file_tmp_path, metadata_file_path);
+            renamed = true;
+        }
     }
     catch (...)
     {
@@ -1226,6 +1229,27 @@ void InterpreterCreateQuery::assertOrSetUUID(ASTCreateQuery & create, const Data
     bool from_path = create.attach_from_path.has_value();
     bool is_on_cluster = getContext()->getClientInfo().query_kind == ClientInfo::QueryKind::SECONDARY_QUERY;
 
+    if (database->getEngineName() == "Replicated" && create.uuid != UUIDHelpers::Nil && !is_replicated_database_internal && !is_on_cluster && !create.attach)
+    {
+        if (getContext()->getSettingsRef().database_replicated_allow_explicit_uuid == 0)
+        {
+            throw Exception(ErrorCodes::BAD_ARGUMENTS, "It's not allowed to explicitly specify UUIDs for tables in Replicated databases, "
+                                                       "see database_replicated_allow_explicit_uuid");
+        }
+        else if (getContext()->getSettingsRef().database_replicated_allow_explicit_uuid == 1)
+        {
+            LOG_WARNING(&Poco::Logger::get("InterpreterCreateQuery"), "It's not recommended to explicitly specify UUIDs for tables in Replicated databases");
+        }
+        else if (getContext()->getSettingsRef().database_replicated_allow_explicit_uuid == 2)
+        {
+            UUID old_uuid = create.uuid;
+            create.uuid = UUIDHelpers::Nil;
+            create.generateRandomUUIDs();
+            LOG_WARNING(&Poco::Logger::get("InterpreterCreateQuery"), "Replaced a user-provided UUID ({}) with a random one ({}) "
+                                                                   "to make sure it's unique", old_uuid, create.uuid);
+        }
+    }
+
     if (is_replicated_database_internal && !internal)
     {
         if (create.uuid == UUIDHelpers::Nil)
diff --git a/src/Interpreters/MutationsInterpreter.cpp b/src/Interpreters/MutationsInterpreter.cpp
index 11df92d071f..916dee01431 100644
--- a/src/Interpreters/MutationsInterpreter.cpp
+++ b/src/Interpreters/MutationsInterpreter.cpp
@@ -1224,9 +1224,18 @@ void MutationsInterpreter::Source::read(
 
         createReadFromPartStep(
             MergeTreeSequentialSourceType::Mutation,
-            plan, *data, storage_snapshot,
-            part, alter_conversions, required_columns,
-            apply_deleted_mask_, std::move(filter), context_,
+            plan,
+            *data,
+            storage_snapshot,
+            part,
+            alter_conversions,
+            required_columns,
+            nullptr,
+            apply_deleted_mask_,
+            std::move(filter),
+            false,
+            false,
+            context_,
             getLogger("MutationsInterpreter"));
     }
     else
diff --git a/src/Interpreters/ProcessList.cpp b/src/Interpreters/ProcessList.cpp
index 6cb50b310ad..f8a808f6c68 100644
--- a/src/Interpreters/ProcessList.cpp
+++ b/src/Interpreters/ProcessList.cpp
@@ -258,7 +258,7 @@ ProcessList::insert(const String & query_, const IAST * ast, ContextMutablePtr q
                 query_context,
                 query_,
                 client_info,
-                priorities.insert(static_cast<int>(settings.priority)),
+                priorities.insert(settings.priority),
                 std::move(thread_group),
                 query_kind,
                 settings,
diff --git a/src/Interpreters/QueryPriorities.h b/src/Interpreters/QueryPriorities.h
index 9e18e7bcff3..7601c7ba6eb 100644
--- a/src/Interpreters/QueryPriorities.h
+++ b/src/Interpreters/QueryPriorities.h
@@ -31,7 +31,7 @@ namespace DB
 class QueryPriorities
 {
 public:
-    using Priority = int;
+    using Priority = size_t;
 
 private:
     friend struct Handle;
diff --git a/src/Interpreters/RowRefs.h b/src/Interpreters/RowRefs.h
index 7c98c47dd11..f8ac68191d6 100644
--- a/src/Interpreters/RowRefs.h
+++ b/src/Interpreters/RowRefs.h
@@ -123,6 +123,7 @@ struct RowRefList : RowRef
 
     RowRefList() {} /// NOLINT
     RowRefList(const Block * block_, size_t row_num_) : RowRef(block_, row_num_), rows(1) {}
+    RowRefList(const Block * block_, size_t row_start_, size_t rows_) : RowRef(block_, row_start_), rows(static_cast<SizeT>(rows_)) {}
 
     ForwardIterator begin() const { return ForwardIterator(this); }
 
diff --git a/src/Interpreters/ServerAsynchronousMetrics.cpp b/src/Interpreters/ServerAsynchronousMetrics.cpp
index 872a9f864df..079029695c9 100644
--- a/src/Interpreters/ServerAsynchronousMetrics.cpp
+++ b/src/Interpreters/ServerAsynchronousMetrics.cpp
@@ -55,9 +55,11 @@ ServerAsynchronousMetrics::ServerAsynchronousMetrics(
     ContextPtr global_context_,
     unsigned update_period_seconds,
     unsigned heavy_metrics_update_period_seconds,
-    const ProtocolServerMetricsFunc & protocol_server_metrics_func_)
+    const ProtocolServerMetricsFunc & protocol_server_metrics_func_,
+    bool update_jemalloc_epoch_,
+    bool update_rss_)
     : WithContext(global_context_)
-    , AsynchronousMetrics(update_period_seconds, protocol_server_metrics_func_)
+    , AsynchronousMetrics(update_period_seconds, protocol_server_metrics_func_, update_jemalloc_epoch_, update_rss_)
     , heavy_metric_update_period(heavy_metrics_update_period_seconds)
 {
     /// sanity check
diff --git a/src/Interpreters/ServerAsynchronousMetrics.h b/src/Interpreters/ServerAsynchronousMetrics.h
index e3c83dc748e..5fab419a32b 100644
--- a/src/Interpreters/ServerAsynchronousMetrics.h
+++ b/src/Interpreters/ServerAsynchronousMetrics.h
@@ -14,7 +14,10 @@ public:
         ContextPtr global_context_,
         unsigned update_period_seconds,
         unsigned heavy_metrics_update_period_seconds,
-        const ProtocolServerMetricsFunc & protocol_server_metrics_func_);
+        const ProtocolServerMetricsFunc & protocol_server_metrics_func_,
+        bool update_jemalloc_epoch_,
+        bool update_rss_);
+
     ~ServerAsynchronousMetrics() override;
 
 private:
diff --git a/src/Interpreters/Squashing.cpp b/src/Interpreters/Squashing.cpp
index 95b76c60063..c656a1a797b 100644
--- a/src/Interpreters/Squashing.cpp
+++ b/src/Interpreters/Squashing.cpp
@@ -45,7 +45,7 @@ Chunk Squashing::squash(Chunk && input_chunk)
 
 Chunk Squashing::add(Chunk && input_chunk)
 {
-    if (!input_chunk)
+    if (!input_chunk || input_chunk.getNumRows() == 0)
         return {};
 
     /// Just read block is already enough.
diff --git a/src/Interpreters/TableJoin.cpp b/src/Interpreters/TableJoin.cpp
index 138085f0710..59a0374051f 100644
--- a/src/Interpreters/TableJoin.cpp
+++ b/src/Interpreters/TableJoin.cpp
@@ -116,6 +116,9 @@ TableJoin::TableJoin(const Settings & settings, VolumePtr tmp_volume_, Temporary
     , max_files_to_merge(settings.join_on_disk_max_files_to_merge)
     , temporary_files_codec(settings.temporary_files_codec)
     , output_by_rowlist_perkey_rows_threshold(settings.join_output_by_rowlist_perkey_rows_threshold)
+    , sort_right_minimum_perkey_rows(settings.join_to_sort_minimum_perkey_rows)
+    , sort_right_maximum_table_rows(settings.join_to_sort_maximum_table_rows)
+    , allow_join_sorting(settings.allow_experimental_join_right_table_sorting)
     , max_memory_usage(settings.max_memory_usage)
     , tmp_volume(tmp_volume_)
     , tmp_data(tmp_data_)
diff --git a/src/Interpreters/TableJoin.h b/src/Interpreters/TableJoin.h
index 4d626084d81..e1bae55a4ed 100644
--- a/src/Interpreters/TableJoin.h
+++ b/src/Interpreters/TableJoin.h
@@ -149,6 +149,9 @@ private:
     const size_t max_files_to_merge = 0;
     const String temporary_files_codec = "LZ4";
     const size_t output_by_rowlist_perkey_rows_threshold = 0;
+    const size_t sort_right_minimum_perkey_rows = 0;
+    const size_t sort_right_maximum_table_rows = 0;
+    const bool allow_join_sorting = false;
 
     /// Value if setting max_memory_usage for query, can be used when max_bytes_in_join is not specified.
     size_t max_memory_usage = 0;
@@ -297,6 +300,9 @@ public:
     }
 
     size_t outputByRowListPerkeyRowsThreshold() const { return output_by_rowlist_perkey_rows_threshold; }
+    size_t sortRightMinimumPerkeyRows() const { return sort_right_minimum_perkey_rows; }
+    size_t sortRightMaximumTableRows() const { return sort_right_maximum_table_rows; }
+    bool allowJoinSorting() const { return allow_join_sorting; }
     size_t defaultMaxBytes() const { return default_max_bytes; }
     size_t maxJoinedBlockRows() const { return max_joined_block_rows; }
     size_t maxRowsInRightBlock() const { return partial_merge_join_rows_in_right_blocks; }
diff --git a/src/Interpreters/convertFieldToType.cpp b/src/Interpreters/convertFieldToType.cpp
index 7e1b4e2fb0e..737353095b8 100644
--- a/src/Interpreters/convertFieldToType.cpp
+++ b/src/Interpreters/convertFieldToType.cpp
@@ -164,7 +164,7 @@ Field convertDecimalType(const Field & from, const To & type)
 }
 
 
-Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const IDataType * from_type_hint)
+Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const IDataType * from_type_hint, const FormatSettings & format_settings)
 {
     if (from_type_hint && from_type_hint->equals(type))
     {
@@ -359,7 +359,7 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID
             Array res(src_arr_size);
             for (size_t i = 0; i < src_arr_size; ++i)
             {
-                res[i] = convertFieldToType(src_arr[i], element_type);
+                res[i] = convertFieldToType(src_arr[i], element_type, nullptr, format_settings);
                 if (res[i].isNull() && !canContainNull(element_type))
                 {
                     // See the comment for Tuples below.
@@ -387,7 +387,7 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID
             for (size_t i = 0; i < dst_tuple_size; ++i)
             {
                 const auto & element_type = *(type_tuple->getElements()[i]);
-                res[i] = convertFieldToType(src_tuple[i], element_type);
+                res[i] = convertFieldToType(src_tuple[i], element_type, nullptr, format_settings);
                 if (res[i].isNull() && !canContainNull(element_type))
                 {
                     /*
@@ -435,12 +435,12 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID
 
                 Tuple updated_entry(2);
 
-                updated_entry[0] = convertFieldToType(key, key_type);
+                updated_entry[0] = convertFieldToType(key, key_type, nullptr, format_settings);
 
                 if (updated_entry[0].isNull() && !canContainNull(key_type))
                     have_unconvertible_element = true;
 
-                updated_entry[1] = convertFieldToType(value, value_type);
+                updated_entry[1] = convertFieldToType(value, value_type, nullptr, format_settings);
                 if (updated_entry[1].isNull() && !canContainNull(value_type))
                     have_unconvertible_element = true;
 
@@ -551,7 +551,7 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID
         ReadBufferFromString in_buffer(src.safeGet<String>());
         try
         {
-            type_to_parse->getDefaultSerialization()->deserializeWholeText(*col, in_buffer, FormatSettings{});
+            type_to_parse->getDefaultSerialization()->deserializeWholeText(*col, in_buffer, format_settings);
         }
         catch (Exception & e)
         {
@@ -563,7 +563,7 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID
         }
 
         Field parsed = (*col)[0];
-        return convertFieldToType(parsed, type, from_type_hint);
+        return convertFieldToType(parsed, type, from_type_hint, format_settings);
     }
 
     throw Exception(ErrorCodes::TYPE_MISMATCH, "Type mismatch in IN or VALUES section. Expected: {}. Got: {}",
@@ -573,7 +573,7 @@ Field convertFieldToTypeImpl(const Field & src, const IDataType & type, const ID
 }
 
 
-Field convertFieldToType(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint)
+Field convertFieldToType(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint, const FormatSettings & format_settings)
 {
     if (from_value.isNull())
         return from_value;
@@ -582,7 +582,7 @@ Field convertFieldToType(const Field & from_value, const IDataType & to_type, co
         return from_value;
 
     if (const auto * low_cardinality_type = typeid_cast<const DataTypeLowCardinality *>(&to_type))
-        return convertFieldToType(from_value, *low_cardinality_type->getDictionaryType(), from_type_hint);
+        return convertFieldToType(from_value, *low_cardinality_type->getDictionaryType(), from_type_hint, format_settings);
     else if (const auto * nullable_type = typeid_cast<const DataTypeNullable *>(&to_type))
     {
         const IDataType & nested_type = *nullable_type->getNestedType();
@@ -593,20 +593,20 @@ Field convertFieldToType(const Field & from_value, const IDataType & to_type, co
 
         if (from_type_hint && from_type_hint->equals(nested_type))
             return from_value;
-        return convertFieldToTypeImpl(from_value, nested_type, from_type_hint);
+        return convertFieldToTypeImpl(from_value, nested_type, from_type_hint, format_settings);
     }
     else
-        return convertFieldToTypeImpl(from_value, to_type, from_type_hint);
+        return convertFieldToTypeImpl(from_value, to_type, from_type_hint, format_settings);
 }
 
 
-Field convertFieldToTypeOrThrow(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint)
+Field convertFieldToTypeOrThrow(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint, const FormatSettings & format_settings)
 {
     bool is_null = from_value.isNull();
     if (is_null && !canContainNull(to_type))
         throw Exception(ErrorCodes::TYPE_MISMATCH, "Cannot convert NULL to {}", to_type.getName());
 
-    Field converted = convertFieldToType(from_value, to_type, from_type_hint);
+    Field converted = convertFieldToType(from_value, to_type, from_type_hint, format_settings);
 
     if (!is_null && converted.isNull())
         throw Exception(ErrorCodes::ARGUMENT_OUT_OF_BOUND,
@@ -626,9 +626,9 @@ static bool decimalEqualsFloat(Field field, Float64 float_value)
     return decimal_to_float == float_value;
 }
 
-std::optional<Field> convertFieldToTypeStrict(const Field & from_value, const IDataType & from_type, const IDataType & to_type)
+std::optional<Field> convertFieldToTypeStrict(const Field & from_value, const IDataType & from_type, const IDataType & to_type, const FormatSettings & format_settings)
 {
-    Field result_value = convertFieldToType(from_value, to_type, &from_type);
+    Field result_value = convertFieldToType(from_value, to_type, &from_type, format_settings);
 
     if (Field::isDecimal(from_value.getType()) && Field::isDecimal(result_value.getType()))
     {
diff --git a/src/Interpreters/convertFieldToType.h b/src/Interpreters/convertFieldToType.h
index 4aa09f8619e..c3c6271a157 100644
--- a/src/Interpreters/convertFieldToType.h
+++ b/src/Interpreters/convertFieldToType.h
@@ -1,6 +1,7 @@
 #pragma once
 
 #include <Core/Field.h>
+#include <Formats/FormatSettings.h>
 
 
 namespace DB
@@ -15,13 +16,13 @@ class IDataType;
   * Checks for the compatibility of types, checks values fall in the range of valid values of the type, makes type conversion.
   * If the value does not fall into the range - returns Null.
   */
-Field convertFieldToType(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint = nullptr);
+Field convertFieldToType(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint = nullptr, const FormatSettings & format_settings = {});
 
 /// Does the same, but throws ARGUMENT_OUT_OF_BOUND if value does not fall into the range.
-Field convertFieldToTypeOrThrow(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint = nullptr);
+Field convertFieldToTypeOrThrow(const Field & from_value, const IDataType & to_type, const IDataType * from_type_hint = nullptr, const FormatSettings & format_settings = {});
 
 /// Applies stricter rules than convertFieldToType, doesn't allow loss of precision converting to Decimal.
 /// Returns `Field` if the conversion was successful and the result is equal to the original value, otherwise returns nullopt.
-std::optional<Field> convertFieldToTypeStrict(const Field & from_value, const IDataType & from_type, const IDataType & to_type);
+std::optional<Field> convertFieldToTypeStrict(const Field & from_value, const IDataType & from_type, const IDataType & to_type, const FormatSettings & format_settings = {});
 
 }
diff --git a/src/Planner/PlannerJoinTree.cpp b/src/Planner/PlannerJoinTree.cpp
index bc31af32a20..76583c82583 100644
--- a/src/Planner/PlannerJoinTree.cpp
+++ b/src/Planner/PlannerJoinTree.cpp
@@ -893,31 +893,6 @@ JoinTreeQueryPlan buildQueryPlanForTableExpression(QueryTreeNodePtr table_expres
                 /// query_plan can be empty if there is nothing to read
                 if (query_plan.isInitialized() && parallel_replicas_enabled_for_storage(storage, settings))
                 {
-                    // (1) find read step
-                    QueryPlan::Node * node = query_plan.getRootNode();
-                    ReadFromMergeTree * reading = nullptr;
-                    while (node)
-                    {
-                        reading = typeid_cast<ReadFromMergeTree *>(node->step.get());
-                        if (reading)
-                            break;
-
-                        QueryPlan::Node * prev_node = node;
-                        if (!node->children.empty())
-                        {
-                            chassert(node->children.size() == 1);
-                            node = node->children.at(0);
-                        }
-                        else
-                        {
-                            throw Exception(
-                                ErrorCodes::LOGICAL_ERROR,
-                                "Step is expected to be ReadFromMergeTree but it's {}",
-                                prev_node->step->getName());
-                        }
-                    }
-
-                    chassert(reading);
                     if (query_context->canUseParallelReplicasCustomKey() && query_context->getClientInfo().distributed_depth == 0)
                     {
                         if (auto cluster = query_context->getClusterForParallelReplicas();
@@ -940,21 +915,50 @@ JoinTreeQueryPlan buildQueryPlanForTableExpression(QueryTreeNodePtr table_expres
                             query_plan = std::move(query_plan_parallel_replicas);
                         }
                     }
-                    else if (query_context->canUseParallelReplicasOnInitiator())
+                    else if (ClusterProxy::canUseParallelReplicasOnInitiator(query_context))
                     {
+                        // (1) find read step
+                        QueryPlan::Node * node = query_plan.getRootNode();
+                        ReadFromMergeTree * reading = nullptr;
+                        while (node)
+                        {
+                            reading = typeid_cast<ReadFromMergeTree *>(node->step.get());
+                            if (reading)
+                                break;
+
+                            QueryPlan::Node * prev_node = node;
+                            if (!node->children.empty())
+                            {
+                                chassert(node->children.size() == 1);
+                                node = node->children.at(0);
+                            }
+                            else
+                            {
+                                throw Exception(
+                                    ErrorCodes::LOGICAL_ERROR,
+                                    "Step is expected to be ReadFromMergeTree but it's {}",
+                                    prev_node->step->getName());
+                            }
+                        }
+
+                        chassert(reading);
+
                         // (2) if it's ReadFromMergeTree - run index analysis and check number of rows to read
                         if (settings.parallel_replicas_min_number_of_rows_per_replica > 0)
                         {
                             auto result_ptr = reading->selectRangesToRead();
-
                             UInt64 rows_to_read = result_ptr->selected_rows;
+
+                            reading->setAnalyzedResult(std::move(result_ptr));
+
                             if (table_expression_query_info.trivial_limit > 0 && table_expression_query_info.trivial_limit < rows_to_read)
                                 rows_to_read = table_expression_query_info.trivial_limit;
 
                             if (max_block_size_limited && (max_block_size_limited < rows_to_read))
                                 rows_to_read = max_block_size_limited;
 
-                            const size_t number_of_replicas_to_use = rows_to_read / settings.parallel_replicas_min_number_of_rows_per_replica;
+                            const size_t number_of_replicas_to_use
+                                = rows_to_read / settings.parallel_replicas_min_number_of_rows_per_replica;
                             LOG_TRACE(
                                 getLogger("Planner"),
                                 "Estimated {} rows to read. It is enough work for {} parallel replicas",
@@ -980,6 +984,7 @@ JoinTreeQueryPlan buildQueryPlanForTableExpression(QueryTreeNodePtr table_expres
                         {
                             from_stage = QueryProcessingStage::WithMergeableState;
                             QueryPlan query_plan_parallel_replicas;
+                            QueryPlanStepPtr reading_step = std::move(node->step);
                             ClusterProxy::executeQueryWithParallelReplicas(
                                 query_plan_parallel_replicas,
                                 storage->getStorageID(),
@@ -987,9 +992,24 @@ JoinTreeQueryPlan buildQueryPlanForTableExpression(QueryTreeNodePtr table_expres
                                 table_expression_query_info.query_tree,
                                 table_expression_query_info.planner_context,
                                 query_context,
-                                table_expression_query_info.storage_limits);
+                                table_expression_query_info.storage_limits,
+                                std::move(reading_step));
                             query_plan = std::move(query_plan_parallel_replicas);
                         }
+                        else
+                        {
+                            QueryPlan query_plan_no_parallel_replicas;
+                            storage->read(
+                                query_plan_no_parallel_replicas,
+                                columns_names,
+                                storage_snapshot,
+                                table_expression_query_info,
+                                query_context,
+                                from_stage,
+                                max_block_size,
+                                max_streams);
+                            query_plan = std::move(query_plan_no_parallel_replicas);
+                        }
                     }
                 }
 
diff --git a/src/Planner/findParallelReplicasQuery.cpp b/src/Planner/findParallelReplicasQuery.cpp
index 39edb1e6516..25481d06670 100644
--- a/src/Planner/findParallelReplicasQuery.cpp
+++ b/src/Planner/findParallelReplicasQuery.cpp
@@ -52,7 +52,13 @@ std::stack<const QueryNode *> getSupportingParallelReplicasQuery(const IQueryTre
                 const auto & storage = table_node.getStorage();
                 /// Here we check StorageDummy as well, to support a query tree with replaced storages.
                 if (std::dynamic_pointer_cast<MergeTreeData>(storage) || typeid_cast<const StorageDummy *>(storage.get()))
+                {
+                    /// parallel replicas is not supported with FINAL
+                    if (table_node.getTableExpressionModifiers() && table_node.getTableExpressionModifiers()->hasFinal())
+                        return {};
+
                     return res;
+                }
 
                 return {};
             }
diff --git a/src/Processors/Formats/Impl/JSONCompactWithProgressRowOutputFormat.cpp b/src/Processors/Formats/Impl/JSONCompactWithProgressRowOutputFormat.cpp
new file mode 100644
index 00000000000..e90864ecdf3
--- /dev/null
+++ b/src/Processors/Formats/Impl/JSONCompactWithProgressRowOutputFormat.cpp
@@ -0,0 +1,154 @@
+#include <Formats/FormatFactory.h>
+#include <Formats/JSONUtils.h>
+#include <Processors/Formats/Impl/JSONCompactWithProgressRowOutputFormat.h>
+
+#include <IO/WriteHelpers.h>
+
+
+namespace DB
+{
+
+JSONCompactWithProgressRowOutputFormat::JSONCompactWithProgressRowOutputFormat(
+    WriteBuffer & out_, const Block & header, const FormatSettings & settings_, bool yield_strings_)
+    : JSONRowOutputFormat(out_, header, settings_, yield_strings_)
+{
+}
+
+void JSONCompactWithProgressRowOutputFormat::writePrefix()
+{
+    JSONUtils::writeCompactObjectStart(*ostr);
+    JSONUtils::writeCompactMetadata(names, types, settings, *ostr);
+    JSONUtils::writeCompactObjectEnd(*ostr);
+    writeCString("\n", *ostr);
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeField(const IColumn & column, const ISerialization & serialization, size_t row_num)
+{
+    JSONUtils::writeFieldFromColumn(column, serialization, row_num, yield_strings, settings, *ostr);
+    ++field_number;
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeFieldDelimiter()
+{
+    JSONUtils::writeFieldCompactDelimiter(*ostr);
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeRowStartDelimiter()
+{
+    if (has_progress)
+        writeProgress();
+    writeCString("{\"data\":", *ostr);
+    JSONUtils::writeCompactArrayStart(*ostr);
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeRowEndDelimiter()
+{
+    JSONUtils::writeCompactArrayEnd(*ostr);
+    writeCString("}\n", *ostr);
+    field_number = 0;
+    ++row_count;
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeRowBetweenDelimiter()
+{
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeBeforeTotals()
+{
+    JSONUtils::writeCompactObjectStart(*ostr);
+    JSONUtils::writeCompactArrayStart(*ostr, 0, "totals");
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeTotals(const Columns & columns, size_t row_num)
+{
+    JSONUtils::writeCompactColumns(columns, serializations, row_num, yield_strings, settings, *ostr);
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeAfterTotals()
+{
+    JSONUtils::writeCompactArrayEnd(*ostr);
+    JSONUtils::writeCompactObjectEnd(*ostr);
+    writeCString("\n", *ostr);
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeExtremesElement(const char * title, const Columns & columns, size_t row_num)
+{
+    JSONUtils::writeCompactArrayStart(*ostr, 2, title);
+    JSONUtils::writeCompactColumns(columns, serializations, row_num, yield_strings, settings, *ostr);
+    JSONUtils::writeCompactArrayEnd(*ostr);
+}
+
+void JSONCompactWithProgressRowOutputFormat::onProgress(const Progress & value)
+{
+    statistics.progress.incrementPiecewiseAtomically(value);
+    String progress_line;
+    WriteBufferFromString buf(progress_line);
+    writeCString("{\"progress\":", buf);
+    statistics.progress.writeJSON(buf);
+    writeCString("}\n", buf);
+    buf.finalize();
+    std::lock_guard lock(progress_lines_mutex);
+    progress_lines.emplace_back(std::move(progress_line));
+    has_progress = true;
+}
+
+
+void JSONCompactWithProgressRowOutputFormat::flush()
+{
+    if (has_progress)
+        writeProgress();
+    JSONRowOutputFormat::flush();
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeSuffix()
+{
+    if (has_progress)
+        writeProgress();
+}
+
+void JSONCompactWithProgressRowOutputFormat::writeProgress()
+{
+    std::lock_guard lock(progress_lines_mutex);
+    for (const auto & progress_line : progress_lines)
+        writeString(progress_line, *ostr);
+    progress_lines.clear();
+    has_progress = false;
+}
+
+void JSONCompactWithProgressRowOutputFormat::finalizeImpl()
+{
+    if (exception_message.empty())
+    {
+        JSONUtils::writeCompactAdditionalInfo(
+            row_count,
+            statistics.rows_before_limit,
+            statistics.applied_limit,
+            statistics.watch,
+            statistics.progress,
+            settings.write_statistics,
+            *ostr);
+    }
+    else
+    {
+        JSONUtils::writeCompactObjectStart(*ostr);
+        JSONUtils::writeException(exception_message, *ostr, settings, 0);
+        JSONUtils::writeCompactObjectEnd(*ostr);
+    }
+    writeCString("\n", *ostr);
+    ostr->next();
+}
+
+void registerOutputFormatJSONCompactWithProgress(FormatFactory & factory)
+{
+    factory.registerOutputFormat(
+        "JSONCompactWithProgress",
+        [](WriteBuffer & buf, const Block & sample, const FormatSettings & format_settings)
+        { return std::make_shared<JSONCompactWithProgressRowOutputFormat>(buf, sample, format_settings, false); });
+
+    factory.registerOutputFormat(
+        "JSONCompactWithProgressStrings",
+        [](WriteBuffer & buf, const Block & sample, const FormatSettings & format_settings)
+        { return std::make_shared<JSONCompactWithProgressRowOutputFormat>(buf, sample, format_settings, true); });
+}
+
+}
diff --git a/src/Processors/Formats/Impl/JSONCompactWithProgressRowOutputFormat.h b/src/Processors/Formats/Impl/JSONCompactWithProgressRowOutputFormat.h
new file mode 100644
index 00000000000..1c21914d8cb
--- /dev/null
+++ b/src/Processors/Formats/Impl/JSONCompactWithProgressRowOutputFormat.h
@@ -0,0 +1,50 @@
+#pragma once
+
+#include <Core/Block.h>
+#include <IO/WriteBuffer.h>
+#include <IO/WriteBufferValidUTF8.h>
+#include <Processors/Formats/Impl/JSONRowOutputFormat.h>
+
+
+namespace DB
+{
+
+struct FormatSettings;
+
+class JSONCompactWithProgressRowOutputFormat final : public JSONRowOutputFormat
+{
+public:
+    JSONCompactWithProgressRowOutputFormat(WriteBuffer & out_, const Block & header, const FormatSettings & settings_, bool yield_strings_);
+
+    String getName() const override { return "JSONCompactWithProgressRowOutputFormat"; }
+
+    void onProgress(const Progress & value) override;
+    void flush() override;
+
+private:
+    void writeField(const IColumn & column, const ISerialization & serialization, size_t row_num) override;
+    void writeFieldDelimiter() override;
+    void writeRowStartDelimiter() override;
+    void writeRowEndDelimiter() override;
+    void writeRowBetweenDelimiter() override;
+    bool supportTotals() const override { return true; }
+    bool supportExtremes() const override { return true; }
+    void writeBeforeTotals() override;
+    void writeAfterTotals() override;
+    void writeExtremesElement(const char * title, const Columns & columns, size_t row_num) override;
+    void writeTotals(const Columns & columns, size_t row_num) override;
+
+    void writeProgress();
+    void writePrefix() override;
+    void writeSuffix() override;
+    void finalizeImpl() override;
+
+
+    std::vector<String> progress_lines;
+    std::mutex progress_lines_mutex;
+    /// To not lock mutex and check progress_lines every row,
+    /// we will use atomic flag that progress_lines is not empty.
+    std::atomic_bool has_progress = false;
+};
+
+}
diff --git a/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp b/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp
index 10d0e051665..16b88d0b8dc 100644
--- a/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp
+++ b/src/Processors/Formats/Impl/ValuesBlockInputFormat.cpp
@@ -542,7 +542,7 @@ bool ValuesBlockInputFormat::parseExpression(IColumn & column, size_t column_idx
     if (format_settings.null_as_default)
         tryToReplaceNullFieldsInComplexTypesWithDefaultValues(expression_value, type);
 
-    Field value = convertFieldToType(expression_value, type, value_raw.second.get());
+    Field value = convertFieldToType(expression_value, type, value_raw.second.get(), format_settings);
 
     /// Check that we are indeed allowed to insert a NULL.
     if (value.isNull() && !type.isNullable() && !type.isLowCardinalityNullable())
diff --git a/src/Processors/QueryPlan/ConvertingActions.cpp b/src/Processors/QueryPlan/ConvertingActions.cpp
new file mode 100644
index 00000000000..b9703ef59cd
--- /dev/null
+++ b/src/Processors/QueryPlan/ConvertingActions.cpp
@@ -0,0 +1,32 @@
+#include <Processors/QueryPlan/QueryPlan.h>
+#include <Interpreters/ActionsDAG.h>
+#include <Processors/QueryPlan/ExpressionStep.h>
+
+namespace DB
+{
+
+void addConvertingActions(QueryPlan & plan, const Block & header, bool has_missing_objects)
+{
+    if (blocksHaveEqualStructure(plan.getCurrentDataStream().header, header))
+        return;
+
+    auto mode = has_missing_objects ? ActionsDAG::MatchColumnsMode::Position : ActionsDAG::MatchColumnsMode::Name;
+
+    auto get_converting_dag = [mode](const Block & block_, const Block & header_)
+    {
+        /// Convert header structure to expected.
+        /// Also we ignore constants from result and replace it with constants from header.
+        /// It is needed for functions like `now64()` or `randConstant()` because their values may be different.
+        return ActionsDAG::makeConvertingActions(
+            block_.getColumnsWithTypeAndName(),
+            header_.getColumnsWithTypeAndName(),
+            mode,
+            true);
+    };
+
+    auto convert_actions_dag = get_converting_dag(plan.getCurrentDataStream().header, header);
+    auto converting = std::make_unique<ExpressionStep>(plan.getCurrentDataStream(), std::move(convert_actions_dag));
+    plan.addStep(std::move(converting));
+}
+
+}
diff --git a/src/Processors/QueryPlan/ConvertingActions.h b/src/Processors/QueryPlan/ConvertingActions.h
new file mode 100644
index 00000000000..6bdf9b8af9a
--- /dev/null
+++ b/src/Processors/QueryPlan/ConvertingActions.h
@@ -0,0 +1,9 @@
+#pragma once
+
+namespace DB
+{
+class QueryPlan;
+class Block;
+
+void addConvertingActions(QueryPlan & plan, const Block & header, bool has_missing_objects);
+}
diff --git a/src/Processors/QueryPlan/DistributedCreateLocalPlan.cpp b/src/Processors/QueryPlan/DistributedCreateLocalPlan.cpp
index d8624a1c99b..eb699858bdf 100644
--- a/src/Processors/QueryPlan/DistributedCreateLocalPlan.cpp
+++ b/src/Processors/QueryPlan/DistributedCreateLocalPlan.cpp
@@ -2,43 +2,13 @@
 
 #include <Common/checkStackSize.h>
 #include <Core/Settings.h>
-#include <Interpreters/ActionsDAG.h>
 #include <Interpreters/InterpreterSelectQuery.h>
 #include <Interpreters/InterpreterSelectQueryAnalyzer.h>
-#include <Processors/QueryPlan/ExpressionStep.h>
+#include <Processors/QueryPlan/ConvertingActions.h>
 
 namespace DB
 {
 
-namespace
-{
-
-void addConvertingActions(QueryPlan & plan, const Block & header, bool has_missing_objects)
-{
-    if (blocksHaveEqualStructure(plan.getCurrentDataStream().header, header))
-        return;
-
-    auto mode = has_missing_objects ? ActionsDAG::MatchColumnsMode::Position : ActionsDAG::MatchColumnsMode::Name;
-
-    auto get_converting_dag = [mode](const Block & block_, const Block & header_)
-    {
-        /// Convert header structure to expected.
-        /// Also we ignore constants from result and replace it with constants from header.
-        /// It is needed for functions like `now64()` or `randConstant()` because their values may be different.
-        return ActionsDAG::makeConvertingActions(
-            block_.getColumnsWithTypeAndName(),
-            header_.getColumnsWithTypeAndName(),
-            mode,
-            true);
-    };
-
-    auto convert_actions_dag = get_converting_dag(plan.getCurrentDataStream().header, header);
-    auto converting = std::make_unique<ExpressionStep>(plan.getCurrentDataStream(), std::move(convert_actions_dag));
-    plan.addStep(std::move(converting));
-}
-
-}
-
 std::unique_ptr<QueryPlan> createLocalPlan(
     const ASTPtr & query_ast,
     const Block & header,
diff --git a/src/Processors/QueryPlan/Optimizations/optimizeReadInOrder.cpp b/src/Processors/QueryPlan/Optimizations/optimizeReadInOrder.cpp
index 9dc4992cddd..ac7fcdcf83f 100644
--- a/src/Processors/QueryPlan/Optimizations/optimizeReadInOrder.cpp
+++ b/src/Processors/QueryPlan/Optimizations/optimizeReadInOrder.cpp
@@ -18,6 +18,7 @@
 #include <Processors/QueryPlan/Optimizations/Optimizations.h>
 #include <Processors/QueryPlan/Optimizations/actionsDAGUtils.h>
 #include <Processors/QueryPlan/ReadFromMergeTree.h>
+#include <Processors/QueryPlan/ReadFromRemote.h>
 #include <Processors/QueryPlan/SortingStep.h>
 #include <Processors/QueryPlan/TotalsHavingStep.h>
 #include <Processors/QueryPlan/UnionStep.h>
@@ -899,6 +900,18 @@ AggregationInputOrder buildInputOrderInfo(AggregatingStep & aggregating, QueryPl
     return {};
 }
 
+static bool readingFromParallelReplicas(const QueryPlan::Node * node)
+{
+    IQueryPlanStep * step = node->step.get();
+    while (!node->children.empty())
+    {
+        step = node->children.front()->step.get();
+        node = node->children.front();
+    }
+
+    return typeid_cast<const ReadFromParallelRemoteReplicasStep *>(step);
+}
+
 void optimizeReadInOrder(QueryPlan::Node & node, QueryPlan::Nodes & nodes)
 {
     if (node.children.size() != 1)
@@ -924,6 +937,16 @@ void optimizeReadInOrder(QueryPlan::Node & node, QueryPlan::Nodes & nodes)
         std::vector<InputOrderInfoPtr> infos;
         infos.reserve(node.children.size());
 
+        for (const auto * child : union_node->children)
+        {
+            /// in case of parallel replicas
+            /// avoid applying read-in-order optimization for local replica
+            /// since it will lead to different parallel replicas modes
+            /// between local and remote nodes
+            if (readingFromParallelReplicas(child))
+                return;
+        }
+
         for (auto * child : union_node->children)
         {
             infos.push_back(buildInputOrderInfo(*sorting, *child, steps_to_update));
diff --git a/src/Processors/QueryPlan/ParallelReplicasLocalPlan.cpp b/src/Processors/QueryPlan/ParallelReplicasLocalPlan.cpp
new file mode 100644
index 00000000000..050044edd3a
--- /dev/null
+++ b/src/Processors/QueryPlan/ParallelReplicasLocalPlan.cpp
@@ -0,0 +1,97 @@
+#include <Processors/QueryPlan/ParallelReplicasLocalPlan.h>
+
+#include <Common/checkStackSize.h>
+#include <Interpreters/ActionsDAG.h>
+#include <Interpreters/Context.h>
+#include <Interpreters/InterpreterSelectQueryAnalyzer.h>
+#include <Interpreters/StorageID.h>
+#include <Parsers/ASTFunction.h>
+#include <Processors/QueryPlan/ConvertingActions.h>
+#include <Processors/QueryPlan/ExpressionStep.h>
+#include <Processors/QueryPlan/ISourceStep.h>
+#include <Processors/QueryPlan/ReadFromMergeTree.h>
+#include <Processors/Sources/NullSource.h>
+#include <Processors/Transforms/ExpressionTransform.h>
+#include <Processors/Transforms/FilterTransform.h>
+#include <QueryPipeline/Pipe.h>
+#include <QueryPipeline/QueryPipelineBuilder.h>
+#include <Storages/MergeTree/MergeTreeDataSelectExecutor.h>
+#include <Storages/MergeTree/RequestResponse.h>
+
+namespace DB
+{
+
+std::pair<std::unique_ptr<QueryPlan>, bool> createLocalPlanForParallelReplicas(
+    const ASTPtr & query_ast,
+    const Block & header,
+    ContextPtr context,
+    QueryProcessingStage::Enum processed_stage,
+    ParallelReplicasReadingCoordinatorPtr coordinator,
+    QueryPlanStepPtr analyzed_read_from_merge_tree,
+    size_t replica_number)
+{
+    checkStackSize();
+
+    auto query_plan = std::make_unique<QueryPlan>();
+    auto new_context = Context::createCopy(context);
+
+    /// Do not push down limit to local plan, as it will break `rows_before_limit_at_least` counter.
+    if (processed_stage == QueryProcessingStage::WithMergeableStateAfterAggregationAndLimit)
+        processed_stage = QueryProcessingStage::WithMergeableStateAfterAggregation;
+
+    /// Do not apply AST optimizations, because query
+    /// is already optimized and some optimizations
+    /// can be applied only for non-distributed tables
+    /// and we can produce query, inconsistent with remote plans.
+    auto select_query_options = SelectQueryOptions(processed_stage).ignoreASTOptimizations();
+
+    /// For Analyzer, identifier in GROUP BY/ORDER BY/LIMIT BY lists has been resolved to
+    /// ConstantNode in QueryTree if it is an alias of a constant, so we should not replace
+    /// ConstantNode with ProjectionNode again(https://github.com/ClickHouse/ClickHouse/issues/62289).
+    new_context->setSetting("enable_positional_arguments", Field(false));
+    new_context->setSetting("allow_experimental_parallel_reading_from_replicas", Field(0));
+    auto interpreter = InterpreterSelectQueryAnalyzer(query_ast, new_context, select_query_options);
+    query_plan = std::make_unique<QueryPlan>(std::move(interpreter).extractQueryPlan());
+
+    QueryPlan::Node * node = query_plan->getRootNode();
+    ReadFromMergeTree * reading = nullptr;
+    while (node)
+    {
+        reading = typeid_cast<ReadFromMergeTree *>(node->step.get());
+        if (reading)
+            break;
+
+        if (!node->children.empty())
+            node = node->children.at(0);
+        else
+            node = nullptr;
+    }
+
+    if (!reading)
+        /// it can happened if merge tree table is empty, - it'll be replaced with ReadFromPreparedSource
+        return {std::move(query_plan), false};
+
+    ReadFromMergeTree::AnalysisResultPtr analyzed_result_ptr;
+    if (analyzed_read_from_merge_tree.get())
+    {
+        auto * analyzed_merge_tree = typeid_cast<ReadFromMergeTree *>(analyzed_read_from_merge_tree.get());
+        if (analyzed_merge_tree)
+            analyzed_result_ptr = analyzed_merge_tree->getAnalyzedResult();
+    }
+
+    MergeTreeAllRangesCallback all_ranges_cb = [coordinator](InitialAllRangesAnnouncement announcement)
+    { coordinator->handleInitialAllRangesAnnouncement(std::move(announcement)); };
+
+    MergeTreeReadTaskCallback read_task_cb = [coordinator](ParallelReadRequest req) -> std::optional<ParallelReadResponse>
+    { return coordinator->handleRequest(std::move(req)); };
+
+    auto read_from_merge_tree_parallel_replicas = reading->createLocalParallelReplicasReadingStep(
+        analyzed_result_ptr, std::move(all_ranges_cb), std::move(read_task_cb), replica_number);
+    node->step = std::move(read_from_merge_tree_parallel_replicas);
+
+    addConvertingActions(*query_plan, header, /*has_missing_objects=*/false);
+
+    return {std::move(query_plan), true};
+}
+
+}
diff --git a/src/Processors/QueryPlan/ParallelReplicasLocalPlan.h b/src/Processors/QueryPlan/ParallelReplicasLocalPlan.h
new file mode 100644
index 00000000000..ab0bbeaeeff
--- /dev/null
+++ b/src/Processors/QueryPlan/ParallelReplicasLocalPlan.h
@@ -0,0 +1,19 @@
+#pragma once
+
+#include <Core/QueryProcessingStage.h>
+#include <Parsers/IAST_fwd.h>
+#include <Processors/QueryPlan/QueryPlan.h>
+#include <Storages/MergeTree/ParallelReplicasReadingCoordinator.h>
+
+namespace DB
+{
+
+std::pair<std::unique_ptr<QueryPlan>, bool> createLocalPlanForParallelReplicas(
+    const ASTPtr & query_ast,
+    const Block & header,
+    ContextPtr context,
+    QueryProcessingStage::Enum processed_stage,
+    ParallelReplicasReadingCoordinatorPtr coordinator,
+    QueryPlanStepPtr read_from_merge_tree,
+    size_t replica_number);
+}
diff --git a/src/Processors/QueryPlan/PartsSplitter.cpp b/src/Processors/QueryPlan/PartsSplitter.cpp
index 44d2703e973..aaa1f53b5ee 100644
--- a/src/Processors/QueryPlan/PartsSplitter.cpp
+++ b/src/Processors/QueryPlan/PartsSplitter.cpp
@@ -50,6 +50,9 @@ bool isSafePrimaryDataKeyType(const IDataType & data_type)
         case TypeIndex::Float64:
         case TypeIndex::Nullable:
         case TypeIndex::ObjectDeprecated:
+        case TypeIndex::Object:
+        case TypeIndex::Variant:
+        case TypeIndex::Dynamic:
             return false;
         case TypeIndex::Array:
         {
@@ -76,16 +79,6 @@ bool isSafePrimaryDataKeyType(const IDataType & data_type)
             const auto & data_type_map = static_cast<const DataTypeMap &>(data_type);
             return isSafePrimaryDataKeyType(*data_type_map.getKeyType()) && isSafePrimaryDataKeyType(*data_type_map.getValueType());
         }
-        case TypeIndex::Variant:
-        {
-            const auto & data_type_variant = static_cast<const DataTypeVariant &>(data_type);
-            const auto & data_type_variant_elements = data_type_variant.getVariants();
-            for (const auto & data_type_variant_element : data_type_variant_elements)
-                if (!isSafePrimaryDataKeyType(*data_type_variant_element))
-                    return false;
-
-            return false;
-        }
         default:
         {
             break;
diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.cpp b/src/Processors/QueryPlan/ReadFromMergeTree.cpp
index 339d97126f4..218f0a61a48 100644
--- a/src/Processors/QueryPlan/ReadFromMergeTree.cpp
+++ b/src/Processors/QueryPlan/ReadFromMergeTree.cpp
@@ -277,7 +277,10 @@ ReadFromMergeTree::ReadFromMergeTree(
     std::shared_ptr<PartitionIdToMaxBlock> max_block_numbers_to_read_,
     LoggerPtr log_,
     AnalysisResultPtr analyzed_result_ptr_,
-    bool enable_parallel_reading)
+    bool enable_parallel_reading_,
+    std::optional<MergeTreeAllRangesCallback> all_ranges_callback_,
+    std::optional<MergeTreeReadTaskCallback> read_task_callback_,
+    std::optional<size_t> number_of_current_replica_)
     : SourceStepWithFilter(DataStream{.header = MergeTreeSelectProcessor::transformHeader(
         storage_snapshot_->getSampleBlockForColumns(all_column_names_),
         query_info_.prewhere_info)}, all_column_names_, query_info_, storage_snapshot_, context_)
@@ -295,13 +298,21 @@ ReadFromMergeTree::ReadFromMergeTree(
     , max_block_numbers_to_read(std::move(max_block_numbers_to_read_))
     , log(std::move(log_))
     , analyzed_result_ptr(analyzed_result_ptr_)
-    , is_parallel_reading_from_replicas(enable_parallel_reading)
+    , is_parallel_reading_from_replicas(enable_parallel_reading_)
     , enable_remove_parts_from_snapshot_optimization(query_info_.merge_tree_enable_remove_parts_from_snapshot_optimization)
+    , number_of_current_replica(number_of_current_replica_)
 {
     if (is_parallel_reading_from_replicas)
     {
-        all_ranges_callback = context->getMergeTreeAllRangesCallback();
-        read_task_callback = context->getMergeTreeReadTaskCallback();
+        if (all_ranges_callback_.has_value())
+            all_ranges_callback = all_ranges_callback_.value();
+        else
+            all_ranges_callback = context->getMergeTreeAllRangesCallback();
+
+        if (read_task_callback_.has_value())
+            read_task_callback = read_task_callback_.value();
+        else
+            read_task_callback = context->getMergeTreeReadTaskCallback();
     }
 
     const auto & settings = context->getSettingsRef();
@@ -335,11 +346,33 @@ ReadFromMergeTree::ReadFromMergeTree(
         enable_vertical_final);
 }
 
+std::unique_ptr<ReadFromMergeTree> ReadFromMergeTree::createLocalParallelReplicasReadingStep(
+    AnalysisResultPtr analyzed_result_ptr_,
+    MergeTreeAllRangesCallback all_ranges_callback_,
+    MergeTreeReadTaskCallback read_task_callback_,
+    size_t replica_number)
+{
+    const bool enable_parallel_reading = true;
+    return std::make_unique<ReadFromMergeTree>(
+        prepared_parts,
+        mutations_snapshot,
+        all_column_names,
+        data,
+        getQueryInfo(),
+        getStorageSnapshot(),
+        getContext(),
+        block_size.max_block_size_rows,
+        requested_num_streams,
+        max_block_numbers_to_read,
+        log,
+        std::move(analyzed_result_ptr_),
+        enable_parallel_reading,
+        all_ranges_callback_,
+        read_task_callback_,
+        replica_number);
+}
 
-Pipe ReadFromMergeTree::readFromPoolParallelReplicas(
-    RangesInDataParts parts_with_range,
-    Names required_columns,
-    PoolSettings pool_settings)
+Pipe ReadFromMergeTree::readFromPoolParallelReplicas(RangesInDataParts parts_with_range, Names required_columns, PoolSettings pool_settings)
 {
     const auto & client_info = context->getClientInfo();
 
@@ -347,7 +380,7 @@ Pipe ReadFromMergeTree::readFromPoolParallelReplicas(
     {
         .all_callback = all_ranges_callback.value(),
         .callback = read_task_callback.value(),
-        .number_of_current_replica = client_info.number_of_current_replica,
+        .number_of_current_replica = number_of_current_replica.value_or(client_info.number_of_current_replica),
     };
 
     /// We have a special logic for local replica. It has to read less data, because in some cases it should
@@ -529,7 +562,7 @@ Pipe ReadFromMergeTree::readInOrder(
         {
             .all_callback = all_ranges_callback.value(),
             .callback = read_task_callback.value(),
-            .number_of_current_replica = client_info.number_of_current_replica,
+            .number_of_current_replica = number_of_current_replica.value_or(client_info.number_of_current_replica),
         };
 
         auto multiplier = context->getSettingsRef().parallel_replicas_single_task_marks_count_multiplier;
@@ -584,11 +617,12 @@ Pipe ReadFromMergeTree::readInOrder(
             context);
     }
 
-    /// Actually it means that parallel reading from replicas enabled
-    /// and we have to collaborate with initiator.
-    /// In this case we won't set approximate rows, because it will be accounted multiple times.
-    const auto in_order_limit = query_info.input_order_info ? query_info.input_order_info->limit : 0;
-    const bool set_total_rows_approx = !is_parallel_reading_from_replicas;
+    /// If parallel replicas enabled, set total rows in progress here only on initiator with local plan
+    /// Otherwise rows will counted multiple times
+    const UInt64 in_order_limit = query_info.input_order_info ? query_info.input_order_info->limit : 0;
+    const bool parallel_replicas_local_plan_for_initiator = is_parallel_reading_from_replicas
+        && context->getSettingsRef().parallel_replicas_local_plan && context->canUseParallelReplicasOnInitiator();
+    const bool set_total_rows_approx = !is_parallel_reading_from_replicas || parallel_replicas_local_plan_for_initiator;
 
     Pipes pipes;
     for (size_t i = 0; i < parts_with_ranges.size(); ++i)
@@ -1422,11 +1456,8 @@ static void buildIndexes(
 
     const auto & settings = context->getSettingsRef();
 
-    indexes.emplace(ReadFromMergeTree::Indexes{{
-        filter_actions_dag,
-        context,
-        primary_key_column_names,
-        primary_key.expression}, {}, {}, {}, {}, false, {}});
+    indexes.emplace(
+        ReadFromMergeTree::Indexes{KeyCondition{filter_actions_dag, context, primary_key_column_names, primary_key.expression}});
 
     if (metadata_snapshot->hasPartitionKey())
     {
@@ -1978,6 +2009,33 @@ void ReadFromMergeTree::initializePipeline(QueryPipelineBuilder & pipeline, cons
 {
     auto result = getAnalysisResult();
 
+    if (is_parallel_reading_from_replicas && context->canUseParallelReplicasOnInitiator()
+        && context->getSettingsRef().parallel_replicas_local_plan)
+    {
+        CoordinationMode mode = CoordinationMode::Default;
+        switch (result.read_type)
+        {
+            case ReadFromMergeTree::ReadType::Default:
+                mode = CoordinationMode::Default;
+                break;
+            case ReadFromMergeTree::ReadType::InOrder:
+                mode = CoordinationMode::WithOrder;
+                break;
+            case ReadFromMergeTree::ReadType::InReverseOrder:
+                mode = CoordinationMode::ReverseOrder;
+                break;
+            case ReadFromMergeTree::ReadType::ParallelReplicas:
+                throw Exception(ErrorCodes::LOGICAL_ERROR, "Read type can't be ParallelReplicas on initiator");
+        }
+
+        chassert(number_of_current_replica.has_value());
+        chassert(all_ranges_callback.has_value());
+
+        /// initialize working set from local replica
+        all_ranges_callback.value()(
+            InitialAllRangesAnnouncement(mode, result.parts_with_ranges.getDescriptions(), number_of_current_replica.value()));
+    }
+
     if (enable_remove_parts_from_snapshot_optimization)
     {
         /// Do not keep data parts in snapshot.
diff --git a/src/Processors/QueryPlan/ReadFromMergeTree.h b/src/Processors/QueryPlan/ReadFromMergeTree.h
index 0bdf0427dd0..60707c406b3 100644
--- a/src/Processors/QueryPlan/ReadFromMergeTree.h
+++ b/src/Processors/QueryPlan/ReadFromMergeTree.h
@@ -121,7 +121,16 @@ public:
         std::shared_ptr<PartitionIdToMaxBlock> max_block_numbers_to_read_,
         LoggerPtr log_,
         AnalysisResultPtr analyzed_result_ptr_,
-        bool enable_parallel_reading);
+        bool enable_parallel_reading_,
+        std::optional<MergeTreeAllRangesCallback> all_ranges_callback_ = std::nullopt,
+        std::optional<MergeTreeReadTaskCallback> read_task_callback_ = std::nullopt,
+        std::optional<size_t> number_of_current_replica_ = std::nullopt);
+
+    std::unique_ptr<ReadFromMergeTree> createLocalParallelReplicasReadingStep(
+        AnalysisResultPtr analyzed_result_ptr_,
+        MergeTreeAllRangesCallback all_ranges_callback_,
+        MergeTreeReadTaskCallback read_task_callback_,
+        size_t replica_number);
 
     static constexpr auto name = "ReadFromMergeTree";
     String getName() const override { return name; }
@@ -143,6 +152,11 @@ public:
 
     struct Indexes
     {
+        explicit Indexes(KeyCondition key_condition_)
+            : key_condition(std::move(key_condition_))
+            , use_skip_indexes(false)
+        {}
+
         KeyCondition key_condition;
         std::optional<PartitionPruner> partition_pruner;
         std::optional<KeyCondition> minmax_idx_condition;
@@ -268,6 +282,7 @@ private:
     std::optional<MergeTreeReadTaskCallback> read_task_callback;
     bool enable_vertical_final = false;
     bool enable_remove_parts_from_snapshot_optimization = true;
+    std::optional<size_t> number_of_current_replica;
 };
 
 }
diff --git a/src/Processors/QueryPlan/ReadFromRemote.cpp b/src/Processors/QueryPlan/ReadFromRemote.cpp
index cf11052cd59..3df46eb1987 100644
--- a/src/Processors/QueryPlan/ReadFromRemote.cpp
+++ b/src/Processors/QueryPlan/ReadFromRemote.cpp
@@ -21,8 +21,8 @@
 #include <Client/ConnectionPoolWithFailover.h>
 #include <QueryPipeline/QueryPipelineBuilder.h>
 #include <Parsers/ASTFunction.h>
-#include <Storages/MergeTree/ParallelReplicasReadingCoordinator.h>
-#include <boost/algorithm/string/join.hpp>
+
+#include <fmt/format.h>
 
 namespace DB
 {
@@ -362,6 +362,7 @@ ReadFromParallelRemoteReplicasStep::ReadFromParallelRemoteReplicasStep(
     ASTPtr query_ast_,
     ClusterPtr cluster_,
     const StorageID & storage_id_,
+    ParallelReplicasReadingCoordinatorPtr coordinator_,
     Block header_,
     QueryProcessingStage::Enum stage_,
     ContextMutablePtr context_,
@@ -369,11 +370,14 @@ ReadFromParallelRemoteReplicasStep::ReadFromParallelRemoteReplicasStep(
     Scalars scalars_,
     Tables external_tables_,
     LoggerPtr log_,
-    std::shared_ptr<const StorageLimitsList> storage_limits_)
+    std::shared_ptr<const StorageLimitsList> storage_limits_,
+    std::vector<ConnectionPoolPtr> pools_to_use_,
+    std::optional<size_t> exclude_pool_index_)
     : ISourceStep(DataStream{.header = std::move(header_)})
     , cluster(cluster_)
     , query_ast(query_ast_)
     , storage_id(storage_id_)
+    , coordinator(std::move(coordinator_))
     , stage(std::move(stage_))
     , context(context_)
     , throttler(throttler_)
@@ -381,16 +385,24 @@ ReadFromParallelRemoteReplicasStep::ReadFromParallelRemoteReplicasStep(
     , external_tables{external_tables_}
     , storage_limits(std::move(storage_limits_))
     , log(log_)
+    , pools_to_use(std::move(pools_to_use_))
+    , exclude_pool_index(exclude_pool_index_)
 {
     chassert(cluster->getShardCount() == 1);
 
-    std::vector<String> description;
-    description.push_back(fmt::format("query: {}", formattedAST(query_ast)));
+    std::vector<String> replicas;
+    replicas.reserve(pools_to_use.size());
 
-    for (const auto & pool : cluster->getShardsInfo().front().per_replica_pools)
-        description.push_back(fmt::format("Replica: {}", pool->getHost()));
+    for (size_t i = 0, l = pools_to_use.size(); i < l; ++i)
+    {
+        if (exclude_pool_index.has_value() && i == exclude_pool_index)
+            continue;
 
-    setStepDescription(boost::algorithm::join(description, ", "));
+        replicas.push_back(pools_to_use[i]->getAddress());
+    }
+
+    auto description = fmt::format("Query: {} Replicas: {}", formattedAST(query_ast), fmt::join(replicas, ", "));
+    setStepDescription(std::move(description));
 }
 
 void ReadFromParallelRemoteReplicasStep::enforceSorting(SortDescription output_sort_description)
@@ -406,48 +418,29 @@ void ReadFromParallelRemoteReplicasStep::enforceAggregationInOrder()
 void ReadFromParallelRemoteReplicasStep::initializePipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &)
 {
     Pipes pipes;
-    const Settings & current_settings = context->getSettingsRef();
-    auto timeouts = ConnectionTimeouts::getTCPTimeoutsWithFailover(current_settings);
 
-    const auto & shard = cluster->getShardsInfo().at(0);
-    size_t max_replicas_to_use = current_settings.max_parallel_replicas;
-    if (max_replicas_to_use > shard.getAllNodeCount())
+    std::vector<std::string_view> addresses;
+    addresses.reserve(pools_to_use.size());
+    for (size_t i = 0, l = pools_to_use.size(); i < l; ++i)
     {
-        LOG_INFO(
-            getLogger("ReadFromParallelRemoteReplicasStep"),
-            "The number of replicas requested ({}) is bigger than the real number available in the cluster ({}). "
-            "Will use the latter number to execute the query.",
-            current_settings.max_parallel_replicas,
-            shard.getAllNodeCount());
-        max_replicas_to_use = shard.getAllNodeCount();
+        if (exclude_pool_index.has_value() && i == exclude_pool_index)
+            continue;
+
+        addresses.emplace_back(pools_to_use[i]->getAddress());
     }
+    LOG_DEBUG(getLogger("ReadFromParallelRemoteReplicasStep"), "Addresses to use: {}", fmt::join(addresses, ", "));
 
-    std::vector<ConnectionPoolWithFailover::Base::ShuffledPool> shuffled_pool;
-    if (max_replicas_to_use < shard.getAllNodeCount())
+    for (size_t i = 0, l = pools_to_use.size(); i < l; ++i)
     {
-        shuffled_pool = shard.pool->getShuffledPools(current_settings);
-        shuffled_pool.resize(max_replicas_to_use);
-    }
-    else
-    {
-        /// try to preserve replicas order if all replicas in cluster are used for query execution
-        /// it's important for data locality during query execution
-        auto priority_func = [](size_t i) { return Priority{static_cast<Int64>(i)}; };
-        shuffled_pool = shard.pool->getShuffledPools(current_settings, priority_func);
-    }
+        if (exclude_pool_index.has_value() && i == exclude_pool_index)
+            continue;
 
-    coordinator
-        = std::make_shared<ParallelReplicasReadingCoordinator>(max_replicas_to_use, current_settings.parallel_replicas_mark_segment_size);
-
-    for (size_t i=0; i < max_replicas_to_use; ++i)
-    {
-        IConnections::ReplicaInfo replica_info
-        {
+        IConnections::ReplicaInfo replica_info{
             /// we should use this number specifically because efficiency of data distribution by consistent hash depends on it.
             .number_of_current_replica = i,
         };
 
-        addPipeForSingeReplica(pipes, shuffled_pool[i].pool, replica_info);
+        addPipeForSingeReplica(pipes, pools_to_use[i], replica_info);
     }
 
     auto pipe = Pipe::unitePipes(std::move(pipes));
diff --git a/src/Processors/QueryPlan/ReadFromRemote.h b/src/Processors/QueryPlan/ReadFromRemote.h
index 1adb26b2915..74389c8f9eb 100644
--- a/src/Processors/QueryPlan/ReadFromRemote.h
+++ b/src/Processors/QueryPlan/ReadFromRemote.h
@@ -70,6 +70,7 @@ public:
         ASTPtr query_ast_,
         ClusterPtr cluster_,
         const StorageID & storage_id_,
+        ParallelReplicasReadingCoordinatorPtr coordinator_,
         Block header_,
         QueryProcessingStage::Enum stage_,
         ContextMutablePtr context_,
@@ -77,7 +78,9 @@ public:
         Scalars scalars_,
         Tables external_tables_,
         LoggerPtr log_,
-        std::shared_ptr<const StorageLimitsList> storage_limits_);
+        std::shared_ptr<const StorageLimitsList> storage_limits_,
+        std::vector<ConnectionPoolPtr> pools_to_use,
+        std::optional<size_t> exclude_pool_index_ = std::nullopt);
 
     String getName() const override { return "ReadFromRemoteParallelReplicas"; }
 
@@ -100,6 +103,8 @@ private:
     Tables external_tables;
     std::shared_ptr<const StorageLimitsList> storage_limits;
     LoggerPtr log;
+    std::vector<ConnectionPoolPtr> pools_to_use;
+    std::optional<size_t> exclude_pool_index;
 };
 
 }
diff --git a/src/Processors/Transforms/AggregatingTransform.cpp b/src/Processors/Transforms/AggregatingTransform.cpp
index c9ada32b839..2e21df0f387 100644
--- a/src/Processors/Transforms/AggregatingTransform.cpp
+++ b/src/Processors/Transforms/AggregatingTransform.cpp
@@ -486,7 +486,7 @@ private:
 
     #define M(NAME) \
                 else if (first->type == AggregatedDataVariants::Type::NAME) \
-                    params->aggregator.mergeSingleLevelDataImpl<decltype(first->NAME)::element_type>(*data);
+                    params->aggregator.mergeSingleLevelDataImpl<decltype(first->NAME)::element_type>(*data, shared_data->is_cancelled);
         if (false) {} // NOLINT
         APPLY_FOR_VARIANTS_SINGLE_LEVEL(M)
     #undef M
diff --git a/src/Processors/Transforms/JoiningTransform.cpp b/src/Processors/Transforms/JoiningTransform.cpp
index ca204bcb482..f2fb6327129 100644
--- a/src/Processors/Transforms/JoiningTransform.cpp
+++ b/src/Processors/Transforms/JoiningTransform.cpp
@@ -299,13 +299,17 @@ IProcessor::Status FillingRightJoinSideTransform::prepare()
 
 void FillingRightJoinSideTransform::work()
 {
-    auto block = inputs.front().getHeader().cloneWithColumns(chunk.detachColumns());
+    auto & input = inputs.front();
+    auto block = input.getHeader().cloneWithColumns(chunk.detachColumns());
 
     if (for_totals)
         join->setTotals(block);
     else
         stop_reading = !join->addBlockToJoin(block);
 
+    if (input.isFinished())
+        join->tryRerangeRightTableData();
+
     set_totals = for_totals;
 }
 
diff --git a/src/Storages/Distributed/DistributedAsyncInsertBatch.cpp b/src/Storages/Distributed/DistributedAsyncInsertBatch.cpp
index 2db2bdf3981..625c64128e7 100644
--- a/src/Storages/Distributed/DistributedAsyncInsertBatch.cpp
+++ b/src/Storages/Distributed/DistributedAsyncInsertBatch.cpp
@@ -242,8 +242,7 @@ void DistributedAsyncInsertBatch::sendBatch(const SettingsChanges & settings_cha
 
                 auto timeouts = ConnectionTimeouts::getTCPTimeoutsWithFailover(insert_settings);
                 auto results = parent.pool->getManyCheckedForInsert(timeouts, insert_settings, PoolMode::GET_ONE, parent.storage.remote_storage.getQualifiedName());
-                auto result = results.front();
-                parent.pool->checkTryResultIsValid(result, insert_settings.distributed_insert_skip_read_only_replicas);
+                auto result = parent.pool->getValidTryResult(results, insert_settings.distributed_insert_skip_read_only_replicas);
                 connection = std::move(result.entry);
                 compression_expected = connection->getCompression() == Protocol::Compression::Enable;
 
@@ -302,8 +301,7 @@ void DistributedAsyncInsertBatch::sendSeparateFiles(const SettingsChanges & sett
 
             auto timeouts = ConnectionTimeouts::getTCPTimeoutsWithFailover(insert_settings);
             auto results = parent.pool->getManyCheckedForInsert(timeouts, insert_settings, PoolMode::GET_ONE, parent.storage.remote_storage.getQualifiedName());
-            auto result = results.front();
-            parent.pool->checkTryResultIsValid(result, insert_settings.distributed_insert_skip_read_only_replicas);
+            auto result = parent.pool->getValidTryResult(results, insert_settings.distributed_insert_skip_read_only_replicas);
             auto connection = std::move(result.entry);
             bool compression_expected = connection->getCompression() == Protocol::Compression::Enable;
 
diff --git a/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.cpp b/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.cpp
index 2400de4c07c..7f368102dfd 100644
--- a/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.cpp
+++ b/src/Storages/Distributed/DistributedAsyncInsertDirectoryQueue.cpp
@@ -415,8 +415,7 @@ void DistributedAsyncInsertDirectoryQueue::processFile(std::string & file_path,
 
         auto timeouts = ConnectionTimeouts::getTCPTimeoutsWithFailover(insert_settings);
         auto results = pool->getManyCheckedForInsert(timeouts, insert_settings, PoolMode::GET_ONE, storage.remote_storage.getQualifiedName());
-        auto result = results.front();
-        pool->checkTryResultIsValid(result, insert_settings.distributed_insert_skip_read_only_replicas);
+        auto result = pool->getValidTryResult(results, insert_settings.distributed_insert_skip_read_only_replicas);
         auto connection = std::move(result.entry);
 
         LOG_DEBUG(log, "Sending `{}` to {} ({} rows, {} bytes)",
diff --git a/src/Storages/Distributed/DistributedSink.cpp b/src/Storages/Distributed/DistributedSink.cpp
index 39f75fd7d56..f01ea10065c 100644
--- a/src/Storages/Distributed/DistributedSink.cpp
+++ b/src/Storages/Distributed/DistributedSink.cpp
@@ -377,8 +377,7 @@ DistributedSink::runWritingJob(JobReplica & job, const Block & current_block, si
                     /// NOTE: INSERT will also take into account max_replica_delay_for_distributed_queries
                     /// (anyway fallback_to_stale_replicas_for_distributed_queries=true by default)
                     auto results = shard_info.pool->getManyCheckedForInsert(timeouts, settings, PoolMode::GET_ONE, storage.remote_storage.getQualifiedName());
-                    auto result = results.front();
-                    shard_info.pool->checkTryResultIsValid(result, settings.distributed_insert_skip_read_only_replicas);
+                    auto result = shard_info.pool->getValidTryResult(results, settings.distributed_insert_skip_read_only_replicas);
                     job.connection_entry = std::move(result.entry);
                 }
                 else
diff --git a/src/Storages/MergeTree/MergeTask.cpp b/src/Storages/MergeTree/MergeTask.cpp
index 7365b9334aa..5c993504245 100644
--- a/src/Storages/MergeTree/MergeTask.cpp
+++ b/src/Storages/MergeTree/MergeTask.cpp
@@ -38,6 +38,11 @@
 #include <Processors/Transforms/DistinctSortedTransform.h>
 #include <Processors/Transforms/DistinctTransform.h>
 #include <Processors/QueryPlan/CreatingSetsStep.h>
+#include <Processors/QueryPlan/DistinctStep.h>
+#include <Processors/QueryPlan/Optimizations/QueryPlanOptimizationSettings.h>
+#include <Processors/QueryPlan/QueryPlan.h>
+#include <Processors/QueryPlan/UnionStep.h>
+#include <Processors/QueryPlan/ExpressionStep.h>
 #include <Interpreters/PreparedSets.h>
 #include <Interpreters/MergeTreeTransaction.h>
 #include <QueryPipeline/QueryPipelineBuilder.h>
@@ -190,7 +195,7 @@ void MergeTask::ExecuteAndFinalizeHorizontalPart::extractMergingAndGatheringColu
     }
 }
 
-bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare()
+bool MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() const
 {
     ProfileEvents::increment(ProfileEvents::Merge);
 
@@ -664,7 +669,7 @@ void MergeTask::ExecuteAndFinalizeHorizontalPart::constructTaskForProjectionPart
 }
 
 
-bool MergeTask::ExecuteAndFinalizeHorizontalPart::executeMergeProjections() // NOLINT
+bool MergeTask::ExecuteAndFinalizeHorizontalPart::executeMergeProjections() const
 {
     /// In case if there are no projections we didn't construct a task
     if (!ctx->merge_projection_parts_task_ptr)
@@ -683,7 +688,7 @@ bool MergeTask::ExecuteAndFinalizeHorizontalPart::executeMergeProjections() // N
     return true;
 }
 
-bool MergeTask::ExecuteAndFinalizeHorizontalPart::executeImpl()
+bool MergeTask::ExecuteAndFinalizeHorizontalPart::executeImpl() const
 {
     Stopwatch watch(CLOCK_MONOTONIC_COARSE);
     UInt64 step_time_ms = global_ctx->data->getSettings()->background_task_preferred_step_execution_time_ms.totalMilliseconds();
@@ -812,33 +817,154 @@ bool MergeTask::VerticalMergeStage::prepareVerticalMergeForAllColumns() const
     ctx->use_prefetch = all_parts_on_remote_disks && global_ctx->data->getSettings()->vertical_merge_remote_filesystem_prefetch;
 
     if (ctx->use_prefetch && ctx->it_name_and_type != global_ctx->gathering_columns.end())
-        ctx->prepared_pipe = createPipeForReadingOneColumn(ctx->it_name_and_type->name);
+        ctx->prepared_pipeline = createPipelineForReadingOneColumn(ctx->it_name_and_type->name);
 
     return false;
 }
 
-Pipe MergeTask::VerticalMergeStage::createPipeForReadingOneColumn(const String & column_name) const
+/// Gathers values from all parts for one column using rows sources temporary file
+class ColumnGathererStep : public ITransformingStep
 {
-    Pipes pipes;
+public:
+    ColumnGathererStep(
+        const DataStream & input_stream_,
+        CompressedReadBufferFromFile * rows_sources_read_buf_,
+        UInt64 merge_block_size_rows_,
+        UInt64 merge_block_size_bytes_,
+        bool is_result_sparse_)
+        : ITransformingStep(input_stream_, input_stream_.header, getTraits())
+        , rows_sources_read_buf(rows_sources_read_buf_)
+        , merge_block_size_rows(merge_block_size_rows_)
+        , merge_block_size_bytes(merge_block_size_bytes_)
+        , is_result_sparse(is_result_sparse_)
+    {}
+
+    String getName() const override { return "ColumnGatherer"; }
+
+    void transformPipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings & /*pipelineSettings*/) override
+    {
+        const auto &header = pipeline.getHeader();
+        const auto input_streams_count = pipeline.getNumStreams();
+
+        rows_sources_read_buf->seek(0, 0);
+
+        auto transform = std::make_unique<ColumnGathererTransform>(
+            header,
+            input_streams_count,
+            *rows_sources_read_buf,
+            merge_block_size_rows,
+            merge_block_size_bytes,
+            is_result_sparse);
+
+        pipeline.addTransform(std::move(transform));
+    }
+
+    void updateOutputStream() override
+    {
+        output_stream = createOutputStream(input_streams.front(), input_streams.front().header, getDataStreamTraits());
+    }
+
+private:
+    static Traits getTraits()
+    {
+        return ITransformingStep::Traits
+        {
+            {
+                .returns_single_stream = true,
+                .preserves_number_of_streams = true,
+                .preserves_sorting = true,
+            },
+            {
+                .preserves_number_of_rows = false,
+            }
+        };
+    }
+
+    MergeTreeData::MergingParams merging_params{};
+    CompressedReadBufferFromFile * rows_sources_read_buf;
+    const UInt64 merge_block_size_rows;
+    const UInt64 merge_block_size_bytes;
+    const bool is_result_sparse;
+};
+
+MergeTask::VerticalMergeRuntimeContext::PreparedColumnPipeline MergeTask::VerticalMergeStage::createPipelineForReadingOneColumn(const String & column_name) const
+{
+    /// Read from all parts
+    std::vector<QueryPlanPtr> plans;
     for (size_t part_num = 0; part_num < global_ctx->future_part->parts.size(); ++part_num)
     {
-        Pipe pipe = createMergeTreeSequentialSource(
+        auto plan_for_part = std::make_unique<QueryPlan>();
+        createReadFromPartStep(
             MergeTreeSequentialSourceType::Merge,
+            *plan_for_part,
             *global_ctx->data,
             global_ctx->storage_snapshot,
             global_ctx->future_part->parts[part_num],
             global_ctx->alter_conversions[part_num],
             Names{column_name},
-            /*mark_ranges=*/ {},
             global_ctx->input_rows_filtered,
             /*apply_deleted_mask=*/ true,
+            std::nullopt,
             ctx->read_with_direct_io,
-            ctx->use_prefetch);
+            ctx->use_prefetch,
+            global_ctx->context,
+            getLogger("VerticalMergeStage"));
 
-        pipes.emplace_back(std::move(pipe));
+        plans.emplace_back(std::move(plan_for_part));
     }
 
-    return Pipe::unitePipes(std::move(pipes));
+    QueryPlan merge_column_query_plan;
+
+    /// Union of all parts streams
+    {
+        DataStreams input_streams;
+        input_streams.reserve(plans.size());
+        for (auto & plan : plans)
+            input_streams.emplace_back(plan->getCurrentDataStream());
+
+        auto union_step = std::make_unique<UnionStep>(std::move(input_streams));
+        merge_column_query_plan.unitePlans(std::move(union_step), std::move(plans));
+    }
+
+    /// Add column gatherer step
+    {
+        bool is_result_sparse = global_ctx->new_data_part->getSerialization(column_name)->getKind() == ISerialization::Kind::SPARSE;
+        const auto data_settings = global_ctx->data->getSettings();
+        auto merge_step = std::make_unique<ColumnGathererStep>(
+            merge_column_query_plan.getCurrentDataStream(),
+            ctx->rows_sources_read_buf.get(), //global_ctx->rows_sources_temporary_file_name,
+            data_settings->merge_max_block_size,
+            data_settings->merge_max_block_size_bytes,
+            is_result_sparse);
+        merge_step->setStepDescription("Gather column");
+        merge_column_query_plan.addStep(std::move(merge_step));
+    }
+
+    /// Add expression step for indexes
+    MergeTreeIndices indexes_to_recalc;
+    IndicesDescription indexes_to_recalc_description;
+    {
+        auto indexes_it = global_ctx->skip_indexes_by_column.find(column_name);
+
+        if (indexes_it != global_ctx->skip_indexes_by_column.end())
+        {
+            indexes_to_recalc_description = indexes_it->second;
+            indexes_to_recalc = MergeTreeIndexFactory::instance().getMany(indexes_it->second);
+
+            auto indices_expression_dag = indexes_it->second.getSingleExpressionForIndices(global_ctx->metadata_snapshot->getColumns(), global_ctx->data->getContext())->getActionsDAG().clone();
+            indices_expression_dag.addMaterializingOutputActions(); /// Const columns cannot be written without materialization.
+            auto calculate_indices_expression_step = std::make_unique<ExpressionStep>(
+                merge_column_query_plan.getCurrentDataStream(),
+                std::move(indices_expression_dag));
+            merge_column_query_plan.addStep(std::move(calculate_indices_expression_step));
+        }
+    }
+
+    auto pipeline_settings = BuildQueryPipelineSettings::fromContext(global_ctx->context);
+    auto optimization_settings =    QueryPlanOptimizationSettings::fromContext(global_ctx->context);
+    auto builder = merge_column_query_plan.buildQueryPipeline(optimization_settings, pipeline_settings);
+
+    return {QueryPipelineBuilder::getPipeline(std::move(*builder)), std::move(indexes_to_recalc)};
 }
 
 void MergeTask::VerticalMergeStage::prepareVerticalMergeForOneColumn() const
@@ -848,50 +974,22 @@ void MergeTask::VerticalMergeStage::prepareVerticalMergeForOneColumn() const
     ctx->progress_before = global_ctx->merge_list_element_ptr->progress.load(std::memory_order_relaxed);
     global_ctx->column_progress = std::make_unique<MergeStageProgress>(ctx->progress_before, ctx->column_sizes->columnWeight(column_name));
 
-    Pipe pipe;
-    if (ctx->prepared_pipe)
+    VerticalMergeRuntimeContext::PreparedColumnPipeline column_pipepline;
+    if (ctx->prepared_pipeline)
     {
-        pipe = std::move(*ctx->prepared_pipe);
+        column_pipepline = std::move(*ctx->prepared_pipeline);
 
+        /// Prepare next column pipeline to initiate prefetching
         auto next_column_it = std::next(ctx->it_name_and_type);
         if (next_column_it != global_ctx->gathering_columns.end())
-            ctx->prepared_pipe = createPipeForReadingOneColumn(next_column_it->name);
+            ctx->prepared_pipeline = createPipelineForReadingOneColumn(next_column_it->name);
     }
     else
     {
-        pipe = createPipeForReadingOneColumn(column_name);
+        column_pipepline = createPipelineForReadingOneColumn(column_name);
     }
 
-    ctx->rows_sources_read_buf->seek(0, 0);
-    bool is_result_sparse = global_ctx->new_data_part->getSerialization(column_name)->getKind() == ISerialization::Kind::SPARSE;
-
-    const auto data_settings = global_ctx->data->getSettings();
-    auto transform = std::make_unique<ColumnGathererTransform>(
-        pipe.getHeader(),
-        pipe.numOutputPorts(),
-        *ctx->rows_sources_read_buf,
-        data_settings->merge_max_block_size,
-        data_settings->merge_max_block_size_bytes,
-        is_result_sparse);
-
-    pipe.addTransform(std::move(transform));
-
-    MergeTreeIndices indexes_to_recalc;
-    auto indexes_it = global_ctx->skip_indexes_by_column.find(column_name);
-
-    if (indexes_it != global_ctx->skip_indexes_by_column.end())
-    {
-        indexes_to_recalc = MergeTreeIndexFactory::instance().getMany(indexes_it->second);
-
-        pipe.addTransform(std::make_shared<ExpressionTransform>(
-            pipe.getHeader(),
-            indexes_it->second.getSingleExpressionForIndices(global_ctx->metadata_snapshot->getColumns(),
-            global_ctx->data->getContext())));
-
-        pipe.addTransform(std::make_shared<MaterializingTransform>(pipe.getHeader()));
-    }
-
-    ctx->column_parts_pipeline = QueryPipeline(std::move(pipe));
+    ctx->column_parts_pipeline = std::move(column_pipepline.pipeline);
 
     /// Dereference unique_ptr
     ctx->column_parts_pipeline.setProgressCallback(MergeProgressCallback(
@@ -909,12 +1007,16 @@ void MergeTask::VerticalMergeStage::prepareVerticalMergeForOneColumn() const
         global_ctx->metadata_snapshot,
         columns_list,
         ctx->compression_codec,
-        indexes_to_recalc,
+        column_pipepline.indexes_to_recalc,
         getStatisticsForColumns(columns_list, global_ctx->metadata_snapshot),
         &global_ctx->written_offset_columns,
         global_ctx->to->getIndexGranularity());
 
     ctx->column_elems_written = 0;
+
+    /// rows_sources_read_buf is reused for each column so we need to rewind it explicitly each time
+    /// This sharing also prevents from from running multiple merge of individual columns in parallel.
+    ctx->rows_sources_read_buf->seek(0, 0);
 }
 
 
@@ -1219,12 +1321,200 @@ bool MergeTask::execute()
 }
 
 
-void MergeTask::ExecuteAndFinalizeHorizontalPart::createMergedStream()
+/// Apply merge strategy (Ordinary, Colapsing, Aggregating, etc) to the stream
+class MergePartsStep : public ITransformingStep
+{
+public:
+    MergePartsStep(
+        const DataStream & input_stream_,
+        const SortDescription & sort_description_,
+        const Names partition_key_columns_,
+        const MergeTreeData::MergingParams & merging_params_,
+        WriteBuffer * rows_sources_write_buf_,
+        UInt64 merge_block_size_rows_,
+        UInt64 merge_block_size_bytes_,
+        bool blocks_are_granules_size_,
+        bool cleanup_,
+        time_t time_of_merge_)
+        : ITransformingStep(input_stream_, input_stream_.header, getTraits())
+        , sort_description(sort_description_)
+        , partition_key_columns(partition_key_columns_)
+        , merging_params(merging_params_)
+        , rows_sources_write_buf(rows_sources_write_buf_)
+        , merge_block_size_rows(merge_block_size_rows_)
+        , merge_block_size_bytes(merge_block_size_bytes_)
+        , blocks_are_granules_size(blocks_are_granules_size_)
+        , cleanup(cleanup_)
+        , time_of_merge(time_of_merge_)
+    {}
+
+    String getName() const override { return "MergeParts"; }
+
+    void transformPipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings & /*pipelineSettings*/) override
+    {
+        /// The order of the streams is important: when the key is matched, the elements go in the order of the source stream number.
+        /// In the merged part, the lines with the same key must be in the ascending order of the identifier of original part,
+        ///  that is going in insertion order.
+        ProcessorPtr merged_transform;
+
+        const auto &header = pipeline.getHeader();
+        const auto input_streams_count = pipeline.getNumStreams();
+
+        switch (merging_params.mode)
+        {
+            case MergeTreeData::MergingParams::Ordinary:
+                merged_transform = std::make_shared<MergingSortedTransform>(
+                    header,
+                    input_streams_count,
+                    sort_description,
+                    merge_block_size_rows,
+                    merge_block_size_bytes,
+                    SortingQueueStrategy::Default,
+                    /* limit_= */0,
+                    /* always_read_till_end_= */false,
+                    rows_sources_write_buf,
+                    blocks_are_granules_size);
+                break;
+
+            case MergeTreeData::MergingParams::Collapsing:
+                merged_transform = std::make_shared<CollapsingSortedTransform>(
+                    header, input_streams_count, sort_description, merging_params.sign_column, false,
+                    merge_block_size_rows, merge_block_size_bytes, rows_sources_write_buf, blocks_are_granules_size);
+                break;
+
+            case MergeTreeData::MergingParams::Summing:
+                merged_transform = std::make_shared<SummingSortedTransform>(
+                    header, input_streams_count, sort_description, merging_params.columns_to_sum, partition_key_columns, merge_block_size_rows, merge_block_size_bytes);
+                break;
+
+            case MergeTreeData::MergingParams::Aggregating:
+                merged_transform = std::make_shared<AggregatingSortedTransform>(header, input_streams_count, sort_description, merge_block_size_rows, merge_block_size_bytes);
+                break;
+
+            case MergeTreeData::MergingParams::Replacing:
+                merged_transform = std::make_shared<ReplacingSortedTransform>(
+                    header, input_streams_count, sort_description, merging_params.is_deleted_column, merging_params.version_column,
+                    merge_block_size_rows, merge_block_size_bytes, rows_sources_write_buf, blocks_are_granules_size,
+                    cleanup);
+                break;
+
+            case MergeTreeData::MergingParams::Graphite:
+                merged_transform = std::make_shared<GraphiteRollupSortedTransform>(
+                    header, input_streams_count, sort_description, merge_block_size_rows, merge_block_size_bytes,
+                    merging_params.graphite_params, time_of_merge);
+                break;
+
+            case MergeTreeData::MergingParams::VersionedCollapsing:
+                merged_transform = std::make_shared<VersionedCollapsingTransform>(
+                    header, input_streams_count, sort_description, merging_params.sign_column,
+                    merge_block_size_rows, merge_block_size_bytes, rows_sources_write_buf, blocks_are_granules_size);
+                break;
+        }
+
+        pipeline.addTransform(std::move(merged_transform));
+
+#ifndef NDEBUG
+        if (!sort_description.empty())
+        {
+            pipeline.addSimpleTransform([&](const Block & header_)
+            {
+                auto transform = std::make_shared<CheckSortedTransform>(header_, sort_description);
+                return transform;
+            });
+        }
+#endif
+    }
+
+    void updateOutputStream() override
+    {
+        output_stream = createOutputStream(input_streams.front(), input_streams.front().header, getDataStreamTraits());
+    }
+
+private:
+    static Traits getTraits()
+    {
+        return ITransformingStep::Traits
+        {
+            {
+                .returns_single_stream = true,
+                .preserves_number_of_streams = true,
+                .preserves_sorting = true,
+            },
+            {
+                .preserves_number_of_rows = false,
+            }
+        };
+    }
+
+    const SortDescription sort_description;
+    const Names partition_key_columns;
+    const MergeTreeData::MergingParams merging_params{};
+    WriteBuffer * rows_sources_write_buf;
+    const UInt64 merge_block_size_rows;
+    const UInt64 merge_block_size_bytes;
+    const bool blocks_are_granules_size;
+    const bool cleanup{false};
+    const time_t time_of_merge{0};
+};
+
+class TTLStep : public ITransformingStep
+{
+public:
+    TTLStep(
+        const DataStream & input_stream_,
+        const ContextPtr & context_,
+        const MergeTreeData & storage_,
+        const StorageMetadataPtr & metadata_snapshot_,
+        const MergeTreeData::MutableDataPartPtr & data_part_,
+        time_t current_time,
+        bool force_)
+        : ITransformingStep(input_stream_, input_stream_.header, getTraits())
+    {
+        transform = std::make_shared<TTLTransform>(context_, input_stream_.header, storage_, metadata_snapshot_, data_part_, current_time, force_);
+        subqueries_for_sets = transform->getSubqueries();
+    }
+
+    String getName() const override { return "TTL"; }
+
+    PreparedSets::Subqueries getSubqueries() { return std::move(subqueries_for_sets); }
+
+    void transformPipeline(QueryPipelineBuilder & pipeline, const BuildQueryPipelineSettings &) override
+    {
+        pipeline.addTransform(transform);
+    }
+
+    void updateOutputStream() override
+    {
+        output_stream = createOutputStream(input_streams.front(), input_streams.front().header, getDataStreamTraits());
+    }
+
+private:
+    static Traits getTraits()
+    {
+        return ITransformingStep::Traits
+        {
+            {
+                .returns_single_stream = true,
+                .preserves_number_of_streams = true,
+                .preserves_sorting = true,
+            },
+            {
+                .preserves_number_of_rows = false,
+            }
+        };
+    }
+
+    std::shared_ptr<TTLTransform> transform;
+    PreparedSets::Subqueries subqueries_for_sets;
+};
+
+
+void MergeTask::ExecuteAndFinalizeHorizontalPart::createMergedStream() const
 {
     /** Read from all parts, merge and write into a new one.
       * In passing, we calculate expression for sorting.
       */
-    Pipes pipes;
+
     global_ctx->watch_prev_elapsed = 0;
 
     /// We count total amount of bytes in parts
@@ -1251,143 +1541,91 @@ void MergeTask::ExecuteAndFinalizeHorizontalPart::createMergedStream()
     global_ctx->horizontal_stage_progress = std::make_unique<MergeStageProgress>(
         ctx->column_sizes ? ctx->column_sizes->keyColumnsWeight() : 1.0);
 
+    /// Read from all parts
+    std::vector<QueryPlanPtr> plans;
     for (size_t i = 0; i < global_ctx->future_part->parts.size(); ++i)
     {
-        Pipe pipe = createMergeTreeSequentialSource(
+        if (global_ctx->future_part->parts[i]->getMarksCount() == 0)
+            LOG_TRACE(ctx->log, "Part {} is empty", global_ctx->future_part->parts[i]->name);
+
+        auto plan_for_part = std::make_unique<QueryPlan>();
+        createReadFromPartStep(
             MergeTreeSequentialSourceType::Merge,
+            *plan_for_part,
             *global_ctx->data,
             global_ctx->storage_snapshot,
             global_ctx->future_part->parts[i],
             global_ctx->alter_conversions[i],
             global_ctx->merging_columns.getNames(),
-            /*mark_ranges=*/ {},
             global_ctx->input_rows_filtered,
             /*apply_deleted_mask=*/ true,
+            /*filter=*/ std::nullopt,
             ctx->read_with_direct_io,
-            /*prefetch=*/ false);
+            /*prefetch=*/ false,
+            global_ctx->context,
+            ctx->log);
 
-        if (global_ctx->metadata_snapshot->hasSortingKey())
-        {
-            pipe.addSimpleTransform([this](const Block & header)
-            {
-                return std::make_shared<ExpressionTransform>(header, global_ctx->metadata_snapshot->getSortingKey().expression);
-            });
-        }
-
-        pipes.emplace_back(std::move(pipe));
+        plans.emplace_back(std::move(plan_for_part));
     }
 
+    QueryPlan merge_parts_query_plan;
 
-    Names sort_columns = global_ctx->metadata_snapshot->getSortingKeyColumns();
-    SortDescription sort_description;
-    sort_description.compile_sort_description = global_ctx->data->getContext()->getSettingsRef().compile_sort_description;
-    sort_description.min_count_to_compile_sort_description = global_ctx->data->getContext()->getSettingsRef().min_count_to_compile_sort_description;
-
-    size_t sort_columns_size = sort_columns.size();
-    sort_description.reserve(sort_columns_size);
-
-    Names partition_key_columns = global_ctx->metadata_snapshot->getPartitionKey().column_names;
-
-    Block header = pipes.at(0).getHeader();
-    for (size_t i = 0; i < sort_columns_size; ++i)
-        sort_description.emplace_back(sort_columns[i], 1, 1);
-
-#ifndef NDEBUG
-    if (!sort_description.empty())
+    /// Union of all parts streams
     {
-        for (size_t i = 0; i < pipes.size(); ++i)
-        {
-            auto & pipe = pipes[i];
-            pipe.addSimpleTransform([&](const Block & header_)
-            {
-                auto transform = std::make_shared<CheckSortedTransform>(header_, sort_description);
-                transform->setDescription(global_ctx->future_part->parts[i]->name);
-                return transform;
-            });
-        }
+        DataStreams input_streams;
+        input_streams.reserve(plans.size());
+        for (auto & plan : plans)
+            input_streams.emplace_back(plan->getCurrentDataStream());
+
+        auto union_step = std::make_unique<UnionStep>(std::move(input_streams));
+        merge_parts_query_plan.unitePlans(std::move(union_step), std::move(plans));
     }
-#endif
 
-    /// The order of the streams is important: when the key is matched, the elements go in the order of the source stream number.
-    /// In the merged part, the lines with the same key must be in the ascending order of the identifier of original part,
-    ///  that is going in insertion order.
-    ProcessorPtr merged_transform;
-
-    /// If merge is vertical we cannot calculate it
-    ctx->blocks_are_granules_size = (global_ctx->chosen_merge_algorithm == MergeAlgorithm::Vertical);
-
-    /// There is no sense to have the block size bigger than one granule for merge operations.
-    const UInt64 merge_block_size_rows = data_settings->merge_max_block_size;
-    const UInt64 merge_block_size_bytes = data_settings->merge_max_block_size_bytes;
-
-    switch (ctx->merging_params.mode)
+    if (global_ctx->metadata_snapshot->hasSortingKey())
     {
-        case MergeTreeData::MergingParams::Ordinary:
-            merged_transform = std::make_shared<MergingSortedTransform>(
-                header,
-                pipes.size(),
-                sort_description,
-                merge_block_size_rows,
-                merge_block_size_bytes,
-                SortingQueueStrategy::Default,
-                /* limit_= */0,
-                /* always_read_till_end_= */false,
-                ctx->rows_sources_write_buf.get(),
-                ctx->blocks_are_granules_size);
-            break;
-
-        case MergeTreeData::MergingParams::Collapsing:
-            merged_transform = std::make_shared<CollapsingSortedTransform>(
-                header, pipes.size(), sort_description, ctx->merging_params.sign_column, false,
-                merge_block_size_rows, merge_block_size_bytes, ctx->rows_sources_write_buf.get(), ctx->blocks_are_granules_size);
-            break;
-
-        case MergeTreeData::MergingParams::Summing:
-            merged_transform = std::make_shared<SummingSortedTransform>(
-                header, pipes.size(), sort_description, ctx->merging_params.columns_to_sum, partition_key_columns, merge_block_size_rows, merge_block_size_bytes);
-            break;
-
-        case MergeTreeData::MergingParams::Aggregating:
-            merged_transform = std::make_shared<AggregatingSortedTransform>(header, pipes.size(), sort_description, merge_block_size_rows, merge_block_size_bytes);
-            break;
-
-        case MergeTreeData::MergingParams::Replacing:
-            if (global_ctx->cleanup && !data_settings->allow_experimental_replacing_merge_with_cleanup)
-                throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "Experimental merges with CLEANUP are not allowed");
-
-            merged_transform = std::make_shared<ReplacingSortedTransform>(
-                header, pipes.size(), sort_description, ctx->merging_params.is_deleted_column, ctx->merging_params.version_column,
-                merge_block_size_rows, merge_block_size_bytes, ctx->rows_sources_write_buf.get(), ctx->blocks_are_granules_size,
-                global_ctx->cleanup);
-            break;
-
-        case MergeTreeData::MergingParams::Graphite:
-            merged_transform = std::make_shared<GraphiteRollupSortedTransform>(
-                header, pipes.size(), sort_description, merge_block_size_rows, merge_block_size_bytes,
-                ctx->merging_params.graphite_params, global_ctx->time_of_merge);
-            break;
-
-        case MergeTreeData::MergingParams::VersionedCollapsing:
-            merged_transform = std::make_shared<VersionedCollapsingTransform>(
-                header, pipes.size(), sort_description, ctx->merging_params.sign_column,
-                merge_block_size_rows, merge_block_size_bytes, ctx->rows_sources_write_buf.get(), ctx->blocks_are_granules_size);
-            break;
+        /// Calculate sorting key expressions so that they are available for merge sorting.
+        auto sorting_key_expression_dag = global_ctx->metadata_snapshot->getSortingKey().expression->getActionsDAG().clone();
+        auto calculate_sorting_key_expression_step = std::make_unique<ExpressionStep>(
+            merge_parts_query_plan.getCurrentDataStream(),
+            std::move(sorting_key_expression_dag));
+        merge_parts_query_plan.addStep(std::move(calculate_sorting_key_expression_step));
     }
 
-    auto builder = std::make_unique<QueryPipelineBuilder>();
-    builder->init(Pipe::unitePipes(std::move(pipes)));
-    builder->addTransform(std::move(merged_transform));
-
-#ifndef NDEBUG
-    if (!sort_description.empty())
+    /// Merge
     {
-        builder->addSimpleTransform([&](const Block & header_)
-        {
-            auto transform = std::make_shared<CheckSortedTransform>(header_, sort_description);
-            return transform;
-        });
+        Names sort_columns = global_ctx->metadata_snapshot->getSortingKeyColumns();
+        SortDescription sort_description;
+        sort_description.compile_sort_description = global_ctx->data->getContext()->getSettingsRef().compile_sort_description;
+        sort_description.min_count_to_compile_sort_description = global_ctx->data->getContext()->getSettingsRef().min_count_to_compile_sort_description;
+
+        size_t sort_columns_size = sort_columns.size();
+        sort_description.reserve(sort_columns_size);
+
+        Names partition_key_columns = global_ctx->metadata_snapshot->getPartitionKey().column_names;
+
+        for (size_t i = 0; i < sort_columns_size; ++i)
+            sort_description.emplace_back(sort_columns[i], 1, 1);
+
+        /// If merge is vertical we cannot calculate it
+        ctx->blocks_are_granules_size = (global_ctx->chosen_merge_algorithm == MergeAlgorithm::Vertical);
+
+        if (global_ctx->cleanup && !data_settings->allow_experimental_replacing_merge_with_cleanup)
+            throw Exception(ErrorCodes::SUPPORT_IS_DISABLED, "Experimental merges with CLEANUP are not allowed");
+
+        auto merge_step = std::make_unique<MergePartsStep>(
+            merge_parts_query_plan.getCurrentDataStream(),
+            sort_description,
+            partition_key_columns,
+            ctx->merging_params,
+            ctx->rows_sources_write_buf.get(),
+            data_settings->merge_max_block_size,
+            data_settings->merge_max_block_size_bytes,
+            ctx->blocks_are_granules_size,
+            global_ctx->cleanup,
+            global_ctx->time_of_merge);
+        merge_step->setStepDescription("Merge sorted parts");
+        merge_parts_query_plan.addStep(std::move(merge_step));
     }
-#endif
 
     if (global_ctx->deduplicate)
     {
@@ -1406,37 +1644,50 @@ void MergeTask::ExecuteAndFinalizeHorizontalPart::createMergedStream()
             }
         }
 
-        if (DistinctSortedTransform::isApplicable(header, sort_description, global_ctx->deduplicate_by_columns))
-            builder->addTransform(std::make_shared<DistinctSortedTransform>(
-                builder->getHeader(), sort_description, SizeLimits(), 0 /*limit_hint*/, global_ctx->deduplicate_by_columns));
-        else
-            builder->addTransform(std::make_shared<DistinctTransform>(
-                builder->getHeader(), SizeLimits(), 0 /*limit_hint*/, global_ctx->deduplicate_by_columns));
+        auto deduplication_step = std::make_unique<DistinctStep>(
+            merge_parts_query_plan.getCurrentDataStream(),
+            SizeLimits(), 0 /*limit_hint*/,
+            global_ctx->deduplicate_by_columns,
+            false /*pre_distinct*/,
+            true /*optimize_distinct_in_order TODO: looks like it should be enabled*/);
+        deduplication_step->setStepDescription("Deduplication step");
+        merge_parts_query_plan.addStep(std::move(deduplication_step));
     }
 
     PreparedSets::Subqueries subqueries;
 
+    /// TTL step
     if (ctx->need_remove_expired_values)
     {
-        auto transform = std::make_shared<TTLTransform>(global_ctx->context, builder->getHeader(), *global_ctx->data, global_ctx->metadata_snapshot, global_ctx->new_data_part, global_ctx->time_of_merge, ctx->force_ttl);
-        subqueries = transform->getSubqueries();
-        builder->addTransform(std::move(transform));
+        auto ttl_step = std::make_unique<TTLStep>(
+            merge_parts_query_plan.getCurrentDataStream(), global_ctx->context, *global_ctx->data, global_ctx->metadata_snapshot, global_ctx->new_data_part, global_ctx->time_of_merge, ctx->force_ttl);
+        subqueries = ttl_step->getSubqueries();
+        ttl_step->setStepDescription("TTL step");
+        merge_parts_query_plan.addStep(std::move(ttl_step));
     }
 
+    /// Secondary indices expressions
     if (!global_ctx->merging_skip_indexes.empty())
     {
-        builder->addTransform(std::make_shared<ExpressionTransform>(
-            builder->getHeader(),
-            global_ctx->merging_skip_indexes.getSingleExpressionForIndices(global_ctx->metadata_snapshot->getColumns(),
-            global_ctx->data->getContext())));
-
-        builder->addTransform(std::make_shared<MaterializingTransform>(builder->getHeader()));
+        auto indices_expression_dag = global_ctx->merging_skip_indexes.getSingleExpressionForIndices(global_ctx->metadata_snapshot->getColumns(), global_ctx->data->getContext())->getActionsDAG().clone();
+        indices_expression_dag.addMaterializingOutputActions(); /// Const columns cannot be written without materialization.
+        auto calculate_indices_expression_step = std::make_unique<ExpressionStep>(
+            merge_parts_query_plan.getCurrentDataStream(),
+            std::move(indices_expression_dag));
+        merge_parts_query_plan.addStep(std::move(calculate_indices_expression_step));
     }
 
     if (!subqueries.empty())
-        builder = addCreatingSetsTransform(std::move(builder), std::move(subqueries), global_ctx->context);
+        addCreatingSetsStep(merge_parts_query_plan, std::move(subqueries), global_ctx->context);
+
+    {
+        auto pipeline_settings = BuildQueryPipelineSettings::fromContext(global_ctx->context);
+        auto optimization_settings =    QueryPlanOptimizationSettings::fromContext(global_ctx->context);
+        auto builder = merge_parts_query_plan.buildQueryPipeline(optimization_settings, pipeline_settings);
+
+        global_ctx->merged_pipeline = QueryPipelineBuilder::getPipeline(std::move(*builder));
+    }
 
-    global_ctx->merged_pipeline = QueryPipelineBuilder::getPipeline(std::move(*builder));
     /// Dereference unique_ptr and pass horizontal_stage_progress by reference
     global_ctx->merged_pipeline.setProgressCallback(MergeProgressCallback(global_ctx->merge_list_element_ptr, global_ctx->watch_prev_elapsed, *global_ctx->horizontal_stage_progress));
     /// Is calculated inside MergeProgressCallback.
diff --git a/src/Storages/MergeTree/MergeTask.h b/src/Storages/MergeTree/MergeTask.h
index 9450fda7b08..84f00d521fd 100644
--- a/src/Storages/MergeTree/MergeTask.h
+++ b/src/Storages/MergeTree/MergeTask.h
@@ -269,12 +269,12 @@ private:
     {
         bool execute() override;
 
-        bool prepare();
-        bool executeImpl();
+        bool prepare() const;
+        bool executeImpl() const;
         void finalize() const;
 
         /// NOTE: Using pointer-to-member instead of std::function and lambda makes stacktraces much more concise and readable
-        using ExecuteAndFinalizeHorizontalPartSubtasks = std::array<bool(ExecuteAndFinalizeHorizontalPart::*)(), 3>;
+        using ExecuteAndFinalizeHorizontalPartSubtasks = std::array<bool(ExecuteAndFinalizeHorizontalPart::*)()const, 3>;
 
         const ExecuteAndFinalizeHorizontalPartSubtasks subtasks
         {
@@ -289,10 +289,10 @@ private:
         void calculateProjections(const Block & block) const;
         void finalizeProjections() const;
         void constructTaskForProjectionPartsMerge() const;
-        bool executeMergeProjections();
+        bool executeMergeProjections() const;
 
         MergeAlgorithm chooseMergeAlgorithm() const;
-        void createMergedStream();
+        void createMergedStream() const;
         void extractMergingAndGatheringColumns() const;
 
         void setRuntimeContext(StageRuntimeContextPtr local, StageRuntimeContextPtr global) override
@@ -334,7 +334,16 @@ private:
 
         Float64 progress_before = 0;
         std::unique_ptr<MergedColumnOnlyOutputStream> column_to{nullptr};
-        std::optional<Pipe> prepared_pipe;
+
+        /// Used for prefetching. Right before starting merge of a column we create a pipeline for the next column
+        /// and it initiates prefetching of the first range of that column.
+        struct PreparedColumnPipeline
+        {
+            QueryPipeline pipeline;
+            MergeTreeIndices indexes_to_recalc;
+        };
+
+        std::optional<PreparedColumnPipeline> prepared_pipeline;
         size_t max_delayed_streams = 0;
         bool use_prefetch = false;
         std::list<std::unique_ptr<MergedColumnOnlyOutputStream>> delayed_streams;
@@ -379,7 +388,7 @@ private:
         bool executeVerticalMergeForOneColumn() const;
         void finalizeVerticalMergeForOneColumn() const;
 
-        Pipe createPipeForReadingOneColumn(const String & column_name) const;
+        VerticalMergeRuntimeContext::PreparedColumnPipeline createPipelineForReadingOneColumn(const String & column_name) const;
 
         VerticalMergeRuntimeContextPtr ctx;
         GlobalRuntimeContextPtr global_ctx;
diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp
index b0e70e94b73..9bfc87135d9 100644
--- a/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp
+++ b/src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp
@@ -85,11 +85,11 @@ MergeTreeDataPartWriterOnDisk::Stream<false>::Stream(
     marks_file_extension{marks_file_extension_},
     plain_file(data_part_storage->writeFile(data_path_ + data_file_extension, max_compress_block_size_, query_write_settings)),
     plain_hashing(*plain_file),
-    compressor(plain_hashing, compression_codec_, max_compress_block_size_),
+    compressor(plain_hashing, compression_codec_, max_compress_block_size_, query_write_settings.use_adaptive_write_buffer, query_write_settings.adaptive_write_buffer_initial_size),
     compressed_hashing(compressor),
     marks_file(data_part_storage->writeFile(marks_path_ + marks_file_extension, 4096, query_write_settings)),
     marks_hashing(*marks_file),
-    marks_compressor(marks_hashing, marks_compression_codec_, marks_compress_block_size_),
+    marks_compressor(marks_hashing, marks_compression_codec_, marks_compress_block_size_, query_write_settings.use_adaptive_write_buffer, query_write_settings.adaptive_write_buffer_initial_size),
     marks_compressed_hashing(marks_compressor),
     compress_marks(MarkType(marks_file_extension).compressed)
 {
@@ -108,7 +108,7 @@ MergeTreeDataPartWriterOnDisk::Stream<true>::Stream(
     data_file_extension{data_file_extension_},
     plain_file(data_part_storage->writeFile(data_path_ + data_file_extension, max_compress_block_size_, query_write_settings)),
     plain_hashing(*plain_file),
-    compressor(plain_hashing, compression_codec_, max_compress_block_size_),
+    compressor(plain_hashing, compression_codec_, max_compress_block_size_, query_write_settings.use_adaptive_write_buffer, query_write_settings.adaptive_write_buffer_initial_size),
     compressed_hashing(compressor),
     compress_marks(false)
 {
diff --git a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp
index 8b6735e0fe2..f050accd7a1 100644
--- a/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp
+++ b/src/Storages/MergeTree/MergeTreeDataPartWriterWide.cpp
@@ -177,6 +177,10 @@ void MergeTreeDataPartWriterWide::addStreams(
         if (!max_compress_block_size)
             max_compress_block_size = settings.max_compress_block_size;
 
+        WriteSettings query_write_settings = settings.query_write_settings;
+        query_write_settings.use_adaptive_write_buffer = settings.use_adaptive_write_buffer_for_dynamic_subcolumns && ISerialization::isDynamicSubcolumn(substream_path, substream_path.size());
+        query_write_settings.adaptive_write_buffer_initial_size = settings.adaptive_write_buffer_initial_size;
+
         column_streams[stream_name] = std::make_unique<Stream<false>>(
             stream_name,
             data_part_storage,
@@ -186,7 +190,7 @@ void MergeTreeDataPartWriterWide::addStreams(
             max_compress_block_size,
             marks_compression_codec,
             settings.marks_compress_block_size,
-            settings.query_write_settings);
+            query_write_settings);
 
         full_name_to_stream_name.emplace(full_stream_name, stream_name);
         stream_name_to_full_name.emplace(stream_name, full_stream_name);
diff --git a/src/Storages/MergeTree/MergeTreeIOSettings.cpp b/src/Storages/MergeTree/MergeTreeIOSettings.cpp
index 24cb25afe47..19365a90a14 100644
--- a/src/Storages/MergeTree/MergeTreeIOSettings.cpp
+++ b/src/Storages/MergeTree/MergeTreeIOSettings.cpp
@@ -30,6 +30,8 @@ MergeTreeWriterSettings::MergeTreeWriterSettings(
     , low_cardinality_max_dictionary_size(global_settings.low_cardinality_max_dictionary_size)
     , low_cardinality_use_single_dictionary_for_part(global_settings.low_cardinality_use_single_dictionary_for_part != 0)
     , use_compact_variant_discriminators_serialization(storage_settings->use_compact_variant_discriminators_serialization)
+    , use_adaptive_write_buffer_for_dynamic_subcolumns(storage_settings->use_adaptive_write_buffer_for_dynamic_subcolumns)
+    , adaptive_write_buffer_initial_size(storage_settings->adaptive_write_buffer_initial_size)
 {
 }
 
diff --git a/src/Storages/MergeTree/MergeTreeIOSettings.h b/src/Storages/MergeTree/MergeTreeIOSettings.h
index 47b174b2e29..fcc72815d8f 100644
--- a/src/Storages/MergeTree/MergeTreeIOSettings.h
+++ b/src/Storages/MergeTree/MergeTreeIOSettings.h
@@ -80,6 +80,8 @@ struct MergeTreeWriterSettings
     size_t low_cardinality_max_dictionary_size;
     bool low_cardinality_use_single_dictionary_for_part;
     bool use_compact_variant_discriminators_serialization;
+    bool use_adaptive_write_buffer_for_dynamic_subcolumns;
+    size_t adaptive_write_buffer_initial_size;
 };
 
 }
diff --git a/src/Storages/MergeTree/MergeTreeSequentialSource.cpp b/src/Storages/MergeTree/MergeTreeSequentialSource.cpp
index a120716d180..78ba02aa7ac 100644
--- a/src/Storages/MergeTree/MergeTreeSequentialSource.cpp
+++ b/src/Storages/MergeTree/MergeTreeSequentialSource.cpp
@@ -354,8 +354,11 @@ public:
         MergeTreeData::DataPartPtr data_part_,
         AlterConversionsPtr alter_conversions_,
         Names columns_to_read_,
+        std::shared_ptr<std::atomic<size_t>> filtered_rows_count_,
         bool apply_deleted_mask_,
         std::optional<ActionsDAG> filter_,
+        bool read_with_direct_io_,
+        bool prefetch_,
         ContextPtr context_,
         LoggerPtr log_)
         : ISourceStep(DataStream{.header = storage_snapshot_->getSampleBlockForColumns(columns_to_read_)})
@@ -365,8 +368,11 @@ public:
         , data_part(std::move(data_part_))
         , alter_conversions(std::move(alter_conversions_))
         , columns_to_read(std::move(columns_to_read_))
+        , filtered_rows_count(std::move(filtered_rows_count_))
         , apply_deleted_mask(apply_deleted_mask_)
         , filter(std::move(filter_))
+        , read_with_direct_io(read_with_direct_io_)
+        , prefetch(prefetch_)
         , context(std::move(context_))
         , log(log_)
     {
@@ -410,25 +416,28 @@ public:
             alter_conversions,
             columns_to_read,
             std::move(mark_ranges),
-            /*filtered_rows_count=*/ nullptr,
+            filtered_rows_count,
             apply_deleted_mask,
-            /*read_with_direct_io=*/ false,
-            /*prefetch=*/ false);
+            read_with_direct_io,
+            prefetch);
 
         pipeline.init(Pipe(std::move(source)));
     }
 
 private:
-    MergeTreeSequentialSourceType type;
+    const MergeTreeSequentialSourceType type;
     const MergeTreeData & storage;
-    StorageSnapshotPtr storage_snapshot;
-    MergeTreeData::DataPartPtr data_part;
-    AlterConversionsPtr alter_conversions;
-    Names columns_to_read;
-    bool apply_deleted_mask;
-    std::optional<ActionsDAG> filter;
-    ContextPtr context;
-    LoggerPtr log;
+    const StorageSnapshotPtr storage_snapshot;
+    const MergeTreeData::DataPartPtr data_part;
+    const AlterConversionsPtr alter_conversions;
+    const Names columns_to_read;
+    const std::shared_ptr<std::atomic<size_t>> filtered_rows_count;
+    const bool apply_deleted_mask;
+    const std::optional<ActionsDAG> filter;
+    const bool read_with_direct_io;
+    const bool prefetch;
+    const ContextPtr context;
+    const LoggerPtr log;
 };
 
 void createReadFromPartStep(
@@ -439,16 +448,28 @@ void createReadFromPartStep(
     MergeTreeData::DataPartPtr data_part,
     AlterConversionsPtr alter_conversions,
     Names columns_to_read,
+    std::shared_ptr<std::atomic<size_t>> filtered_rows_count,
     bool apply_deleted_mask,
     std::optional<ActionsDAG> filter,
+    bool read_with_direct_io,
+    bool prefetch,
     ContextPtr context,
     LoggerPtr log)
 {
-    auto reading = std::make_unique<ReadFromPart>(type,
-        storage, storage_snapshot,
-        std::move(data_part), std::move(alter_conversions),
-        std::move(columns_to_read), apply_deleted_mask,
-        std::move(filter), std::move(context), log);
+    auto reading = std::make_unique<ReadFromPart>(
+        type,
+        storage,
+        storage_snapshot,
+        std::move(data_part),
+        std::move(alter_conversions),
+        std::move(columns_to_read),
+        filtered_rows_count,
+        apply_deleted_mask,
+        std::move(filter),
+        read_with_direct_io,
+        prefetch,
+        std::move(context),
+        log);
 
     plan.addStep(std::move(reading));
 }
diff --git a/src/Storages/MergeTree/MergeTreeSequentialSource.h b/src/Storages/MergeTree/MergeTreeSequentialSource.h
index 5b7c80385f6..d2ed1394dbd 100644
--- a/src/Storages/MergeTree/MergeTreeSequentialSource.h
+++ b/src/Storages/MergeTree/MergeTreeSequentialSource.h
@@ -39,8 +39,11 @@ void createReadFromPartStep(
     MergeTreeData::DataPartPtr data_part,
     AlterConversionsPtr alter_conversions,
     Names columns_to_read,
+    std::shared_ptr<std::atomic<size_t>> filtered_rows_count,
     bool apply_deleted_mask,
     std::optional<ActionsDAG> filter,
+    bool read_with_direct_io,
+    bool prefetch,
     ContextPtr context,
     LoggerPtr log);
 
diff --git a/src/Storages/MergeTree/MergeTreeSettings.h b/src/Storages/MergeTree/MergeTreeSettings.h
index 0769b60dc6b..dcb18155114 100644
--- a/src/Storages/MergeTree/MergeTreeSettings.h
+++ b/src/Storages/MergeTree/MergeTreeSettings.h
@@ -99,6 +99,8 @@ struct Settings;
     M(Bool, add_implicit_sign_column_constraint_for_collapsing_engine, false, "If true, add implicit constraint for sign column for CollapsingMergeTree engine.", 0) \
     M(Milliseconds, sleep_before_commit_local_part_in_replicated_table_ms, 0, "For testing. Do not change it.", 0) \
     M(Bool, optimize_row_order, false, "Allow reshuffling of rows during part inserts and merges to improve the compressibility of the new part", 0) \
+    M(Bool, use_adaptive_write_buffer_for_dynamic_subcolumns, true, "Allow to use adaptive writer buffers during writing dynamic subcolumns to reduce memory usage", 0) \
+    M(UInt64, adaptive_write_buffer_initial_size, 16 * 1024, "Initial size of an adaptive write buffer", 0) \
     \
     /* Part removal settings. */ \
     M(UInt64, simultaneous_parts_removal_limit, 0, "Maximum number of parts to remove during one CleanupThread iteration (0 means unlimited).", 0) \
diff --git a/src/Storages/MergeTree/ParallelReplicasReadingCoordinator.cpp b/src/Storages/MergeTree/ParallelReplicasReadingCoordinator.cpp
index ee47fe3549a..8abf735b49f 100644
--- a/src/Storages/MergeTree/ParallelReplicasReadingCoordinator.cpp
+++ b/src/Storages/MergeTree/ParallelReplicasReadingCoordinator.cpp
@@ -185,7 +185,8 @@ public:
     void handleInitialAllRangesAnnouncement(InitialAllRangesAnnouncement announcement)
     {
         if (++sent_initial_requests > replicas_count)
-            throw Exception(ErrorCodes::LOGICAL_ERROR, "Initiator received more initial requests than there are replicas");
+            throw Exception(
+                ErrorCodes::LOGICAL_ERROR, "Initiator received more initial requests than there are replicas: replica_num={}", announcement.replica_num);
 
         doHandleInitialAllRangesAnnouncement(std::move(announcement));
     }
@@ -433,9 +434,9 @@ void DefaultCoordinator::setProgressCallback()
 
 void DefaultCoordinator::doHandleInitialAllRangesAnnouncement(InitialAllRangesAnnouncement announcement)
 {
-    const auto replica_num = announcement.replica_num;
+    LOG_DEBUG(log, "Initial request: {}", announcement.describe());
 
-    LOG_DEBUG(log, "Initial request from replica {}: {}", announcement.replica_num, announcement.describe());
+    const auto replica_num = announcement.replica_num;
 
     initializeReadingState(std::move(announcement));
 
@@ -839,6 +840,7 @@ public:
 
     Parts all_parts_to_read;
     size_t total_rows_to_read = 0;
+    bool state_initialized{false};
 
     LoggerPtr log = getLogger(fmt::format("{}{}", magic_enum::enum_name(mode), "Coordinator"));
 };
@@ -858,7 +860,9 @@ void InOrderCoordinator<mode>::markReplicaAsUnavailable(size_t replica_number)
 template <CoordinationMode mode>
 void InOrderCoordinator<mode>::doHandleInitialAllRangesAnnouncement(InitialAllRangesAnnouncement announcement)
 {
-    LOG_TRACE(log, "Received an announcement {}", announcement.describe());
+    LOG_TRACE(log, "Received an announcement : {}", announcement.describe());
+
+    ++stats[announcement.replica_num].number_of_requests;
 
     size_t new_rows_to_read = 0;
 
@@ -868,13 +872,16 @@ void InOrderCoordinator<mode>::doHandleInitialAllRangesAnnouncement(InitialAllRa
         auto the_same_it = std::find_if(all_parts_to_read.begin(), all_parts_to_read.end(),
             [&part] (const Part & other) { return other.description.info == part.info; });
 
-        /// We have the same part - add the info about presence on current replica to it
+        /// We have the same part - add the info about presence on the corresponding replica to it
         if (the_same_it != all_parts_to_read.end())
         {
             the_same_it->replicas.insert(announcement.replica_num);
             continue;
         }
 
+        if (state_initialized)
+            continue;
+
         auto covering_or_the_same_it = std::find_if(all_parts_to_read.begin(), all_parts_to_read.end(),
             [&part] (const Part & other) { return other.description.info.contains(part.info) ||  part.info.contains(other.description.info); });
 
@@ -889,9 +896,10 @@ void InOrderCoordinator<mode>::doHandleInitialAllRangesAnnouncement(InitialAllRa
         std::sort(ranges.begin(), ranges.end());
     }
 
-    ++stats[announcement.replica_num].number_of_requests;
+    state_initialized = true;
 
-    if (new_rows_to_read > 0)
+    // progress_callback is not set when local plan is used for initiator
+    if (progress_callback && new_rows_to_read > 0)
     {
         Progress progress;
         progress.total_rows_to_read = new_rows_to_read;
@@ -911,7 +919,7 @@ ParallelReadResponse InOrderCoordinator<mode>::handleRequest(ParallelReadRequest
             "Replica {} decided to read in {} mode, not in {}. This is a bug",
             request.replica_num, magic_enum::enum_name(request.mode), magic_enum::enum_name(mode));
 
-    LOG_TRACE(log, "Got request from replica {}, data {}", request.replica_num, request.describe());
+    LOG_TRACE(log, "Got read request: {}", request.describe());
 
     ParallelReadResponse response;
     response.description = request.description;
@@ -925,8 +933,15 @@ ParallelReadResponse InOrderCoordinator<mode>::handleRequest(ParallelReadRequest
         if (global_part_it == all_parts_to_read.end())
             continue;
 
+        if (global_part_it->replicas.empty())
+            throw Exception(
+                ErrorCodes::LOGICAL_ERROR,
+                "Part {} requested by replica {} is not registered in working set",
+                part.info.getPartNameV1(),
+                request.replica_num);
+
         if (!global_part_it->replicas.contains(request.replica_num))
-            throw Exception(ErrorCodes::LOGICAL_ERROR, "Part {} doesn't exist on replica {} according to the global state", part.info.getPartNameV1(), request.replica_num);
+            continue;
 
         size_t current_mark_size = 0;
 
@@ -1057,6 +1072,7 @@ void ParallelReplicasReadingCoordinator::initialize(CoordinationMode mode)
             break;
     }
 
+    // progress_callback is not set when local plan is used for initiator
     if (progress_callback)
         pimpl->setProgressCallback(std::move(progress_callback));
 
diff --git a/src/Storages/MergeTree/RequestResponse.cpp b/src/Storages/MergeTree/RequestResponse.cpp
index 2ce0e20dcd2..bcdeb443a0b 100644
--- a/src/Storages/MergeTree/RequestResponse.cpp
+++ b/src/Storages/MergeTree/RequestResponse.cpp
@@ -44,9 +44,7 @@ void ParallelReadRequest::serialize(WriteBuffer & out) const
 
 String ParallelReadRequest::describe() const
 {
-    String result;
-    result += fmt::format("replica_num: {} \n", replica_num);
-    result += fmt::format("min_num_of_marks: {} \n", min_number_of_marks);
+    String result = fmt::format("replica_num {}, min_num_of_marks {}, ", replica_num, min_number_of_marks);
     result += description.describe();
     return result;
 }
@@ -131,10 +129,7 @@ void InitialAllRangesAnnouncement::serialize(WriteBuffer & out) const
 
 String InitialAllRangesAnnouncement::describe()
 {
-    String result;
-    result += description.describe();
-    result += fmt::format("----------\nReceived from {} replica\n", replica_num);
-    return result;
+    return fmt::format("replica {}, mode {}, {}", replica_num, mode, description.describe());
 }
 
 InitialAllRangesAnnouncement InitialAllRangesAnnouncement::deserialize(ReadBuffer & in)
diff --git a/src/Storages/MergeTree/extractZooKeeperPathFromReplicatedTableDef.h b/src/Storages/MergeTree/extractZooKeeperPathFromReplicatedTableDef.h
index 5ef5e1db62e..212dc048868 100644
--- a/src/Storages/MergeTree/extractZooKeeperPathFromReplicatedTableDef.h
+++ b/src/Storages/MergeTree/extractZooKeeperPathFromReplicatedTableDef.h
@@ -14,6 +14,6 @@ using ContextPtr = std::shared_ptr<const Context>;
 /// Extracts a zookeeper path from a specified CREATE TABLE query.
 /// The function checks the table engine and if it is Replicated*MergeTree then it takes the first argument and expands macros in it.
 /// Returns std::nullopt if the specified CREATE query doesn't describe a Replicated table or its arguments can't be evaluated.
-std::optional<String> extractZooKeeperPathFromReplicatedTableDef(const ASTCreateQuery & create_query, const ContextPtr & context);
+std::optional<String> extractZooKeeperPathFromReplicatedTableDef(const ASTCreateQuery & create_query, const ContextPtr & local_context);
 
 }
diff --git a/src/Storages/MergeTree/registerStorageMergeTree.cpp b/src/Storages/MergeTree/registerStorageMergeTree.cpp
index 9a65d590453..18ed7df9b5d 100644
--- a/src/Storages/MergeTree/registerStorageMergeTree.cpp
+++ b/src/Storages/MergeTree/registerStorageMergeTree.cpp
@@ -12,6 +12,7 @@
 #include <Common/Macros.h>
 #include <Common/OptimizedRegularExpression.h>
 #include <Common/typeid_cast.h>
+#include <Common/logger_useful.h>
 
 #include <Parsers/ASTFunction.h>
 #include <Parsers/ASTIdentifier.h>
@@ -189,7 +190,7 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
     const String & engine_name,
     ASTs & engine_args,
     LoadingStrictnessLevel mode,
-    const ContextPtr & context,
+    const ContextPtr & local_context,
     String & zookeeper_path,
     String & replica_name,
     RenamingRestrictions & renaming_restrictions)
@@ -206,11 +207,11 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
     {
         /// Allow expressions in engine arguments.
         /// In new syntax argument can be literal or identifier or array/tuple of identifiers.
-        evaluateEngineArgs(engine_args, context);
+        evaluateEngineArgs(engine_args, local_context);
     }
 
-    bool is_on_cluster = context->getClientInfo().query_kind == ClientInfo::QueryKind::SECONDARY_QUERY;
-    bool is_replicated_database = context->getClientInfo().query_kind == ClientInfo::QueryKind::SECONDARY_QUERY &&
+    bool is_on_cluster = local_context->getClientInfo().query_kind == ClientInfo::QueryKind::SECONDARY_QUERY;
+    bool is_replicated_database = local_context->getClientInfo().query_kind == ClientInfo::QueryKind::SECONDARY_QUERY &&
         DatabaseCatalog::instance().getDatabase(table_id.database_name)->getEngineName() == "Replicated";
 
     /// Allow implicit {uuid} macros only for zookeeper_path in ON CLUSTER queries
@@ -230,10 +231,10 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
             /// We did unfold it in previous versions to make moving table from Atomic to Ordinary database work correctly,
             /// but now it's not allowed (and it was the only reason to unfold {uuid} macro).
             info.table_id.uuid = UUIDHelpers::Nil;
-            zookeeper_path = context->getMacros()->expand(zookeeper_path, info);
+            zookeeper_path = local_context->getMacros()->expand(zookeeper_path, info);
 
             info.level = 0;
-            replica_name = context->getMacros()->expand(replica_name, info);
+            replica_name = local_context->getMacros()->expand(replica_name, info);
         }
 
         ast_zk_path->value = zookeeper_path;
@@ -251,11 +252,11 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
         }
         if (!allow_uuid_macro)
             info.table_id.uuid = UUIDHelpers::Nil;
-        zookeeper_path = context->getMacros()->expand(zookeeper_path, info);
+        zookeeper_path = local_context->getMacros()->expand(zookeeper_path, info);
 
         info.level = 0;
         info.table_id.uuid = UUIDHelpers::Nil;
-        replica_name = context->getMacros()->expand(replica_name, info);
+        replica_name = local_context->getMacros()->expand(replica_name, info);
 
         /// We do not allow renaming table with these macros in metadata, because zookeeper_path will be broken after RENAME TABLE.
         /// NOTE: it may happen if table was created by older version of ClickHouse (< 20.10) and macros was not unfolded on table creation
@@ -272,9 +273,24 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
 
     bool has_arguments = (arg_num + 2 <= arg_cnt);
     bool has_valid_arguments = has_arguments && engine_args[arg_num]->as<ASTLiteral>() && engine_args[arg_num + 1]->as<ASTLiteral>();
+    const auto & server_settings = local_context->getServerSettings();
 
     if (has_valid_arguments)
     {
+        if (is_replicated_database && local_context->getSettingsRef().database_replicated_allow_replicated_engine_arguments == 0)
+        {
+            throw Exception(ErrorCodes::BAD_ARGUMENTS,
+                            "It's not allowed to specify explicit zookeeper_path and replica_name "
+                            "for ReplicatedMergeTree arguments in Replicated database. If you really want to "
+                            "specify them explicitly, enable setting "
+                            "database_replicated_allow_replicated_engine_arguments.");
+        }
+        else if (is_replicated_database && local_context->getSettingsRef().database_replicated_allow_replicated_engine_arguments == 1)
+        {
+            LOG_WARNING(&Poco::Logger::get("registerStorageMergeTree"), "It's not recommended to explicitly specify "
+                                                            "zookeeper_path and replica_name in ReplicatedMergeTree arguments");
+        }
+
         /// Get path and name from engine arguments
         auto * ast_zk_path = engine_args[arg_num]->as<ASTLiteral>();
         if (ast_zk_path && ast_zk_path->value.getType() == Field::Types::String)
@@ -288,6 +304,15 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
         else
             throw Exception(ErrorCodes::BAD_ARGUMENTS, "Replica name must be a string literal{}", verbose_help_message);
 
+
+        if (is_replicated_database && local_context->getSettingsRef().database_replicated_allow_replicated_engine_arguments == 2)
+        {
+            LOG_WARNING(&Poco::Logger::get("registerStorageMergeTree"), "Replacing user-provided ZooKeeper path and replica name ({}, {}) "
+                                                                     "with default arguments", zookeeper_path, replica_name);
+            engine_args[arg_num]->as<ASTLiteral>()->value = zookeeper_path = server_settings.default_replica_path;
+            engine_args[arg_num + 1]->as<ASTLiteral>()->value = replica_name = server_settings.default_replica_name;
+        }
+
         expand_macro(ast_zk_path, ast_replica_name);
     }
     else if (is_extended_storage_def
@@ -297,7 +322,6 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
     {
         /// Try use default values if arguments are not specified.
         /// Note: {uuid} macro works for ON CLUSTER queries when database engine is Atomic.
-        const auto & server_settings = context->getServerSettings();
         zookeeper_path = server_settings.default_replica_path;
         /// TODO maybe use hostname if {replica} is not defined?
         replica_name = server_settings.default_replica_name;
@@ -322,7 +346,7 @@ static void extractZooKeeperPathAndReplicaNameFromEngineArgs(
 }
 
 /// Extracts a zookeeper path from a specified CREATE TABLE query.
-std::optional<String> extractZooKeeperPathFromReplicatedTableDef(const ASTCreateQuery & query, const ContextPtr & context)
+std::optional<String> extractZooKeeperPathFromReplicatedTableDef(const ASTCreateQuery & query, const ContextPtr & local_context)
 {
     if (!query.storage || !query.storage->engine)
         return {};
@@ -346,7 +370,7 @@ std::optional<String> extractZooKeeperPathFromReplicatedTableDef(const ASTCreate
 
     try
     {
-        extractZooKeeperPathAndReplicaNameFromEngineArgs(query, table_id, engine_name, engine_args, mode, context,
+        extractZooKeeperPathAndReplicaNameFromEngineArgs(query, table_id, engine_name, engine_args, mode, local_context,
                                                          zookeeper_path, replica_name, renaming_restrictions);
     }
     catch (Exception & e)
diff --git a/src/Storages/ObjectStorage/DataLakes/DeltaLakeMetadata.cpp b/src/Storages/ObjectStorage/DataLakes/DeltaLakeMetadata.cpp
index 667a925d11e..f04e868ee5a 100644
--- a/src/Storages/ObjectStorage/DataLakes/DeltaLakeMetadata.cpp
+++ b/src/Storages/ObjectStorage/DataLakes/DeltaLakeMetadata.cpp
@@ -207,18 +207,28 @@ struct DeltaLakeMetadataImpl
             Poco::Dynamic::Var json = parser.parse(json_str);
             Poco::JSON::Object::Ptr object = json.extract<Poco::JSON::Object::Ptr>();
 
+            if (!object)
+                throw Exception(ErrorCodes::LOGICAL_ERROR, "Failed to parse metadata file");
+
+#ifdef ABORT_ON_LOGICAL_ERROR
             std::ostringstream oss; // STYLE_CHECK_ALLOW_STD_STRING_STREAM
             object->stringify(oss);
             LOG_TEST(log, "Metadata: {}", oss.str());
+#endif
 
             if (object->has("metaData"))
             {
                 const auto metadata_object = object->get("metaData").extract<Poco::JSON::Object::Ptr>();
+                if (!metadata_object)
+                    throw Exception(ErrorCodes::LOGICAL_ERROR, "Failed to extract `metaData` field");
+
                 const auto schema_object = metadata_object->getValue<String>("schemaString");
 
                 Poco::JSON::Parser p;
                 Poco::Dynamic::Var fields_json = parser.parse(schema_object);
                 const Poco::JSON::Object::Ptr & fields_object = fields_json.extract<Poco::JSON::Object::Ptr>();
+                if (!fields_object)
+                    throw Exception(ErrorCodes::LOGICAL_ERROR, "Failed to extract `fields` field");
 
                 auto current_schema = parseMetadata(fields_object);
                 if (file_schema.empty())
@@ -237,6 +247,9 @@ struct DeltaLakeMetadataImpl
             if (object->has("add"))
             {
                 auto add_object = object->get("add").extract<Poco::JSON::Object::Ptr>();
+                if (!add_object)
+                    throw Exception(ErrorCodes::LOGICAL_ERROR, "Failed to extract `add` field");
+
                 auto path = add_object->getValue<String>("path");
                 result.insert(fs::path(configuration->getPath()) / path);
 
@@ -247,6 +260,9 @@ struct DeltaLakeMetadataImpl
                     if (add_object->has("partitionValues"))
                     {
                         auto partition_values = add_object->get("partitionValues").extract<Poco::JSON::Object::Ptr>();
+                        if (!partition_values)
+                            throw Exception(ErrorCodes::LOGICAL_ERROR, "Failed to extract `partitionValues` field");
+
                         if (partition_values->size())
                         {
                             auto & current_partition_columns = file_partition_columns[filename];
@@ -274,7 +290,11 @@ struct DeltaLakeMetadataImpl
             }
             else if (object->has("remove"))
             {
-                auto path = object->get("remove").extract<Poco::JSON::Object::Ptr>()->getValue<String>("path");
+                auto remove_object = object->get("remove").extract<Poco::JSON::Object::Ptr>();
+                if (!remove_object)
+                    throw Exception(ErrorCodes::LOGICAL_ERROR, "Failed to extract `remove` field");
+
+                auto path = remove_object->getValue<String>("path");
                 result.erase(fs::path(configuration->getPath()) / path);
             }
         }
diff --git a/src/Storages/ObjectStorageQueue/StorageObjectStorageQueue.cpp b/src/Storages/ObjectStorageQueue/StorageObjectStorageQueue.cpp
index 9452ce81e9e..c1ef37e1a48 100644
--- a/src/Storages/ObjectStorageQueue/StorageObjectStorageQueue.cpp
+++ b/src/Storages/ObjectStorageQueue/StorageObjectStorageQueue.cpp
@@ -64,9 +64,7 @@ namespace
 
     void checkAndAdjustSettings(
         ObjectStorageQueueSettings & queue_settings,
-        ASTStorage * engine_args,
-        bool is_attach,
-        const LoggerPtr & log)
+        bool is_attach)
     {
         if (!is_attach && !queue_settings.mode.changed)
         {
@@ -85,16 +83,6 @@ namespace
                             "Setting `cleanup_interval_min_ms` ({}) must be less or equal to `cleanup_interval_max_ms` ({})",
                             queue_settings.cleanup_interval_min_ms, queue_settings.cleanup_interval_max_ms);
         }
-
-        if (!is_attach && !queue_settings.processing_threads_num.changed)
-        {
-            queue_settings.processing_threads_num = std::max<uint32_t>(getNumberOfPhysicalCPUCores(), 16);
-            engine_args->settings->as<ASTSetQuery>()->changes.insertSetting(
-                "processing_threads_num",
-                queue_settings.processing_threads_num.value);
-
-            LOG_TRACE(log, "Set `processing_threads_num` to {}", queue_settings.processing_threads_num);
-        }
     }
 
     std::shared_ptr<ObjectStorageQueueLog> getQueueLog(const ObjectStoragePtr & storage, const ContextPtr & context, const ObjectStorageQueueSettings & table_settings)
@@ -130,7 +118,7 @@ StorageObjectStorageQueue::StorageObjectStorageQueue(
     const String & comment,
     ContextPtr context_,
     std::optional<FormatSettings> format_settings_,
-    ASTStorage * engine_args,
+    ASTStorage * /* engine_args */,
     LoadingStrictnessLevel mode)
     : IStorage(table_id_)
     , WithContext(context_)
@@ -154,7 +142,7 @@ StorageObjectStorageQueue::StorageObjectStorageQueue(
         throw Exception(ErrorCodes::BAD_QUERY_PARAMETER, "ObjectStorageQueue url must either end with '/' or contain globs");
     }
 
-    checkAndAdjustSettings(*queue_settings, engine_args, mode > LoadingStrictnessLevel::CREATE, log);
+    checkAndAdjustSettings(*queue_settings, mode > LoadingStrictnessLevel::CREATE);
 
     object_storage = configuration->createObjectStorage(context_, /* is_readonly */true);
     FormatFactory::instance().checkFormatName(configuration->format);
diff --git a/src/Storages/Statistics/Statistics.cpp b/src/Storages/Statistics/Statistics.cpp
index 795963bd55d..0557530515f 100644
--- a/src/Storages/Statistics/Statistics.cpp
+++ b/src/Storages/Statistics/Statistics.cpp
@@ -248,7 +248,7 @@ ColumnStatisticsPtr MergeTreeStatisticsFactory::get(const ColumnDescription & co
     {
         auto it = creators.find(type);
         if (it == creators.end())
-            throw Exception(ErrorCodes::INCORRECT_QUERY, "Unknown statistic type '{}'. Available types: 'count_min', 'minmax', 'tdigest' and 'uniq'", type);
+            throw Exception(ErrorCodes::INCORRECT_QUERY, "Unknown statistic type '{}'. Available types: 'countmin', 'minmax', 'tdigest' and 'uniq'", type);
         auto stat_ptr = (it->second)(desc, column_desc.type);
         column_stat->stats[type] = stat_ptr;
     }
diff --git a/src/Storages/Statistics/StatisticsCountMinSketch.cpp b/src/Storages/Statistics/StatisticsCountMinSketch.cpp
index 1a2459c230d..f477181ec2d 100644
--- a/src/Storages/Statistics/StatisticsCountMinSketch.cpp
+++ b/src/Storages/Statistics/StatisticsCountMinSketch.cpp
@@ -49,7 +49,7 @@ Float64 StatisticsCountMinSketch::estimateEqual(const Field & val) const
     if (isStringOrFixedString(data_type))
         return sketch.get_estimate(val.safeGet<String>());
 
-    throw Exception(ErrorCodes::LOGICAL_ERROR, "Statistics 'count_min' does not support estimate data type of {}", data_type->getName());
+    throw Exception(ErrorCodes::LOGICAL_ERROR, "Statistics 'countmin' does not support estimate data type of {}", data_type->getName());
 }
 
 void StatisticsCountMinSketch::update(const ColumnPtr & column)
@@ -88,7 +88,7 @@ void countMinSketchStatisticsValidator(const SingleStatisticsDescription & /*des
     DataTypePtr inner_data_type = removeNullable(data_type);
     inner_data_type = removeLowCardinalityAndNullable(inner_data_type);
     if (!inner_data_type->isValueRepresentedByNumber() && !isStringOrFixedString(inner_data_type))
-        throw Exception(ErrorCodes::ILLEGAL_STATISTICS, "Statistics of type 'count_min' does not support type {}", data_type->getName());
+        throw Exception(ErrorCodes::ILLEGAL_STATISTICS, "Statistics of type 'countmin' does not support type {}", data_type->getName());
 }
 
 StatisticsPtr countMinSketchStatisticsCreator(const SingleStatisticsDescription & description, const DataTypePtr & data_type)
diff --git a/src/Storages/StatisticsDescription.cpp b/src/Storages/StatisticsDescription.cpp
index 4e0d901d0c7..ac7fa8898de 100644
--- a/src/Storages/StatisticsDescription.cpp
+++ b/src/Storages/StatisticsDescription.cpp
@@ -48,11 +48,11 @@ static StatisticsType stringToStatisticsType(String type)
         return StatisticsType::TDigest;
     if (type == "uniq")
         return StatisticsType::Uniq;
-    if (type == "count_min")
+    if (type == "countmin")
         return StatisticsType::CountMinSketch;
     if (type == "minmax")
         return StatisticsType::MinMax;
-    throw Exception(ErrorCodes::INCORRECT_QUERY, "Unknown statistics type: {}. Supported statistics types are 'count_min', 'minmax', 'tdigest' and 'uniq'.", type);
+    throw Exception(ErrorCodes::INCORRECT_QUERY, "Unknown statistics type: {}. Supported statistics types are 'countmin', 'minmax', 'tdigest' and 'uniq'.", type);
 }
 
 String SingleStatisticsDescription::getTypeName() const
@@ -64,11 +64,11 @@ String SingleStatisticsDescription::getTypeName() const
         case StatisticsType::Uniq:
             return "Uniq";
         case StatisticsType::CountMinSketch:
-            return "count_min";
+            return "countmin";
         case StatisticsType::MinMax:
             return "minmax";
         default:
-            throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown statistics type: {}. Supported statistics types are 'count_min', 'minmax', 'tdigest' and 'uniq'.", type);
+            throw Exception(ErrorCodes::LOGICAL_ERROR, "Unknown statistics type: {}. Supported statistics types are 'countmin', 'minmax', 'tdigest' and 'uniq'.", type);
     }
 }
 
diff --git a/src/Storages/System/StorageSystemServerSettings.cpp b/src/Storages/System/StorageSystemServerSettings.cpp
index d242b6de4ec..ee99c472620 100644
--- a/src/Storages/System/StorageSystemServerSettings.cpp
+++ b/src/Storages/System/StorageSystemServerSettings.cpp
@@ -63,7 +63,6 @@ void StorageSystemServerSettings::fillData(MutableColumns & res_columns, Context
     /// current setting values, one needs to ask the components directly.
     std::unordered_map<String, std::pair<String, ChangeableWithoutRestart>> changeable_settings = {
         {"max_server_memory_usage", {std::to_string(total_memory_tracker.getHardLimit()), ChangeableWithoutRestart::Yes}},
-        {"allow_use_jemalloc_memory", {std::to_string(total_memory_tracker.getAllowUseJemallocMmemory()), ChangeableWithoutRestart::Yes}},
 
         {"max_table_size_to_drop", {std::to_string(context->getMaxTableSizeToDrop()), ChangeableWithoutRestart::Yes}},
         {"max_partition_size_to_drop", {std::to_string(context->getMaxPartitionSizeToDrop()), ChangeableWithoutRestart::Yes}},
diff --git a/tests/ci/ci_config.py b/tests/ci/ci_config.py
index 436fc5b6bb2..3587405d95f 100644
--- a/tests/ci/ci_config.py
+++ b/tests/ci/ci_config.py
@@ -473,7 +473,7 @@ class CI:
         JobNames.STATELESS_TEST_FLAKY_ASAN: CommonJobConfigs.STATELESS_TEST.with_properties(
             required_builds=[BuildNames.PACKAGE_ASAN],
             pr_only=True,
-            timeout=3600,
+            timeout=3 * 3600,
             # TODO: approach with reference job names does not work because digest may not be calculated if job skipped in wf
             # reference_job_name=JobNames.STATELESS_TEST_RELEASE,
         ),
diff --git a/tests/ci/pr_info.py b/tests/ci/pr_info.py
index 5c051b093e0..d3c4ffb1e68 100644
--- a/tests/ci/pr_info.py
+++ b/tests/ci/pr_info.py
@@ -9,6 +9,7 @@ from urllib.parse import quote
 from unidiff import PatchSet  # type: ignore
 
 from build_download_helper import get_gh_api
+from ci_config import Labels
 from env_helper import (
     GITHUB_EVENT_PATH,
     GITHUB_REPOSITORY,
@@ -16,7 +17,6 @@ from env_helper import (
     GITHUB_SERVER_URL,
     GITHUB_UPSTREAM_REPOSITORY,
 )
-from ci_config import Labels
 from get_robot_token import get_best_robot_token
 from github_helper import GitHub
 
@@ -459,16 +459,18 @@ class PRInfo:
         sync_repo = gh.get_repo(GITHUB_REPOSITORY)
         sync_pr = sync_repo.get_pull(self.number)
         # Find the commit that is in both repos, upstream and cloud
-        sync_commits = sync_pr.get_commits().reversed
-        upstream_commits = upstream_pr.get_commits().reversed
+        # Do not ever use `reversed` here, otherwise the list of commits is not full
+        sync_commits = list(sync_pr.get_commits())
+        upstream_commits = list(upstream_pr.get_commits())
         # Github objects are compared by _url attribute. We can't compare them directly and
         # should compare commits by SHA1
         upstream_shas = [c.sha for c in upstream_commits]
         logging.info("Commits in upstream PR:\n %s", ", ".join(upstream_shas))
         sync_shas = [c.sha for c in sync_commits]
-        logging.info("Commits in sync PR:\n %s", ", ".join(reversed(sync_shas)))
+        logging.info("Commits in sync PR:\n %s", ", ".join(sync_shas))
 
-        # find latest synced commit
+        # find latest synced commit, search from the latest
+        upstream_commits.reverse()
         last_synced_upstream_commit = None
         for commit in upstream_commits:
             if commit.sha in sync_shas:
diff --git a/tests/clickhouse-test b/tests/clickhouse-test
index 6cb7d026417..75a180f356b 100755
--- a/tests/clickhouse-test
+++ b/tests/clickhouse-test
@@ -916,6 +916,7 @@ class SettingsRandomizer:
         "min_external_table_block_size_bytes": lambda: random.choice([0, 1, 100000000]),
         "max_parsing_threads": lambda: random.choice([0, 1, 10]),
         "optimize_functions_to_subcolumns": lambda: random.randint(0, 1),
+        "parallel_replicas_local_plan": lambda: random.randint(0, 1),
     }
 
     @staticmethod
diff --git a/tests/config/config.d/clusters.xml b/tests/config/config.d/clusters.xml
index af26565d7e6..ff42b5828e3 100644
--- a/tests/config/config.d/clusters.xml
+++ b/tests/config/config.d/clusters.xml
@@ -115,10 +115,6 @@
         <parallel_replicas>
             <shard>
                 <internal_replication>false</internal_replication>
-                <replica>
-                    <host>127.0.0.1</host>
-                    <port>9000</port>
-                </replica>
                 <replica>
                     <host>127.0.0.2</host>
                     <port>9000</port>
@@ -147,6 +143,10 @@
                     <host>127.0.0.8</host>
                     <port>9000</port>
                 </replica>
+                <replica>
+                    <host>127.0.0.1</host>
+                    <port>9000</port>
+                </replica>
                 <replica>
                     <host>127.0.0.9</host>
                     <port>9000</port>
diff --git a/tests/config/users.d/database_replicated.xml b/tests/config/users.d/database_replicated.xml
index c049c3559fc..1c2cf2ac22b 100644
--- a/tests/config/users.d/database_replicated.xml
+++ b/tests/config/users.d/database_replicated.xml
@@ -6,6 +6,7 @@
             <distributed_ddl_task_timeout>120</distributed_ddl_task_timeout>
             <database_replicated_always_detach_permanently>1</database_replicated_always_detach_permanently>
             <database_replicated_enforce_synchronous_settings>1</database_replicated_enforce_synchronous_settings>
+            <database_replicated_allow_replicated_engine_arguments>3</database_replicated_allow_replicated_engine_arguments>
         </default>
     </profiles>
 </clickhouse>
diff --git a/tests/integration/test_disk_over_web_server/test.py b/tests/integration/test_disk_over_web_server/test.py
index 891ee8f00f5..ec0bef23731 100644
--- a/tests/integration/test_disk_over_web_server/test.py
+++ b/tests/integration/test_disk_over_web_server/test.py
@@ -311,7 +311,8 @@ def test_replicated_database(cluster):
         SETTINGS storage_policy = 'web';
     """.format(
             uuids[0]
-        )
+        ),
+        settings={"database_replicated_allow_explicit_uuid": 3},
     )
 
     node2 = cluster.instances["node2"]
diff --git a/tests/integration/test_disks_app_func/test.py b/tests/integration/test_disks_app_func/test.py
index 56ea5c8846a..a4b2399e117 100644
--- a/tests/integration/test_disks_app_func/test.py
+++ b/tests/integration/test_disks_app_func/test.py
@@ -13,8 +13,20 @@ def started_cluster():
             main_configs=["config.xml"],
             with_minio=True,
         )
-
         cluster.start()
+
+        # local disk requires its `path` directory to exist.
+        # the two paths below belong to `test1` and `test2` disks
+        node = cluster.instances["disks_app_test"]
+        for path in ["path1", "path2"]:
+            node.exec_in_container(
+                [
+                    "bash",
+                    "-c",
+                    f"mkdir -p /var/lib/clickhouse/{path}",
+                ]
+            )
+
         yield cluster
 
     finally:
diff --git a/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config2.xml b/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config2.xml
index 25ececea3e8..e71b93379d0 100644
--- a/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config2.xml
+++ b/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config2.xml
@@ -16,7 +16,7 @@
             <value>az-zoo2</value>
             <enable_auto_detection_on_cloud>1</enable_auto_detection_on_cloud>
         </availability_zone>
-        <max_memory_usage_soft_limit>20000000</max_memory_usage_soft_limit>
+        <max_memory_usage_soft_limit>200000000</max_memory_usage_soft_limit>
 
         <coordination_settings>
             <operation_timeout_ms>10000</operation_timeout_ms>
diff --git a/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config3.xml b/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config3.xml
index 81e343b77c9..cf4a4686f2c 100644
--- a/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config3.xml
+++ b/tests/integration/test_keeper_memory_soft_limit/configs/keeper_config3.xml
@@ -13,7 +13,7 @@
         <tcp_port>2181</tcp_port>
         <server_id>3</server_id>
 
-        <max_memory_usage_soft_limit>20000000</max_memory_usage_soft_limit>
+        <max_memory_usage_soft_limit>200000000</max_memory_usage_soft_limit>
 
         <coordination_settings>
             <operation_timeout_ms>10000</operation_timeout_ms>
diff --git a/tests/integration/test_memory_limit/test.py b/tests/integration/test_memory_limit/test.py
index 6d6745711da..db68a38c1b1 100644
--- a/tests/integration/test_memory_limit/test.py
+++ b/tests/integration/test_memory_limit/test.py
@@ -13,7 +13,6 @@ node = cluster.add_instance(
         "configs/async_metrics_no.xml",
     ],
     mem_limit="4g",
-    env_variables={"MALLOC_CONF": "dirty_decay_ms:0"},
 )
 
 
diff --git a/tests/integration/test_parallel_replicas_no_replicas/test.py b/tests/integration/test_parallel_replicas_no_replicas/test.py
index 62d4b005d94..5ae4cc7d882 100644
--- a/tests/integration/test_parallel_replicas_no_replicas/test.py
+++ b/tests/integration/test_parallel_replicas_no_replicas/test.py
@@ -49,5 +49,6 @@ def test_skip_all_replicas(
                 "max_parallel_replicas": max_parallel_replicas,
                 "cluster_for_parallel_replicas": cluster_name,
                 "skip_unavailable_shards": skip_unavailable_shards,
+                "parallel_replicas_local_plan": 0,
             },
         )
diff --git a/tests/integration/test_replicated_database/configs/settings.xml b/tests/integration/test_replicated_database/configs/settings.xml
index c637fe8eead..41799c5bed2 100644
--- a/tests/integration/test_replicated_database/configs/settings.xml
+++ b/tests/integration/test_replicated_database/configs/settings.xml
@@ -5,6 +5,8 @@
             <allow_experimental_alter_materialized_view_structure>1</allow_experimental_alter_materialized_view_structure>
             <allow_experimental_object_type>0</allow_experimental_object_type>
             <allow_suspicious_codecs>0</allow_suspicious_codecs>
+            <database_replicated_allow_replicated_engine_arguments>3</database_replicated_allow_replicated_engine_arguments>
+            <database_replicated_allow_explicit_uuid>3</database_replicated_allow_explicit_uuid>
         </default>
     </profiles>
     <users>
diff --git a/tests/integration/test_replicated_database/configs/settings2.xml b/tests/integration/test_replicated_database/configs/settings2.xml
index dad5740a8ae..fb7f0c8d4d3 100644
--- a/tests/integration/test_replicated_database/configs/settings2.xml
+++ b/tests/integration/test_replicated_database/configs/settings2.xml
@@ -5,6 +5,8 @@
             <allow_experimental_alter_materialized_view_structure>1</allow_experimental_alter_materialized_view_structure>
             <allow_experimental_object_type>0</allow_experimental_object_type>
             <allow_suspicious_codecs>0</allow_suspicious_codecs>
+            <database_replicated_allow_replicated_engine_arguments>3</database_replicated_allow_replicated_engine_arguments>
+            <database_replicated_allow_explicit_uuid>3</database_replicated_allow_explicit_uuid>
 
             <throw_on_unsupported_query_inside_transaction>0</throw_on_unsupported_query_inside_transaction>
         </default>
diff --git a/tests/integration/test_replicated_database_alter_modify_order_by/configs/settings.xml b/tests/integration/test_replicated_database_alter_modify_order_by/configs/settings.xml
index 16caee9ba20..a5e45ead44e 100644
--- a/tests/integration/test_replicated_database_alter_modify_order_by/configs/settings.xml
+++ b/tests/integration/test_replicated_database_alter_modify_order_by/configs/settings.xml
@@ -1,6 +1,8 @@
 <clickhouse>
     <profiles>
         <default>
+            <database_replicated_allow_replicated_engine_arguments>3</database_replicated_allow_replicated_engine_arguments>
+            <database_replicated_allow_explicit_uuid>3</database_replicated_allow_explicit_uuid>
         </default>
     </profiles>
     <users>
diff --git a/tests/queries/0_stateless/00115_shard_in_incomplete_result.sh b/tests/queries/0_stateless/00115_shard_in_incomplete_result.sh
index 4916721764c..1efe529ac24 100755
--- a/tests/queries/0_stateless/00115_shard_in_incomplete_result.sh
+++ b/tests/queries/0_stateless/00115_shard_in_incomplete_result.sh
@@ -15,6 +15,6 @@ $CLICKHOUSE_CLIENT --query="
     INSERT INTO users VALUES (1321770221388956068);
 ";
 
-for _ in {1..10}; do seq 1 10 | sed "s/.*/SELECT count() FROM (SELECT * FROM remote('127.0.0.{2,3}', ${CLICKHOUSE_DATABASE}, users) WHERE UserID IN (SELECT arrayJoin([1468013291393583084, 1321770221388956068])));/" | $CLICKHOUSE_CLIENT -n | grep -vE '^4$' && echo 'Fail!' && break; echo -n '.'; done; echo
+for _ in {1..10}; do seq 1 10 | sed "s/.*/SELECT count() FROM (SELECT * FROM remote('127.0.0.{2,3}', ${CLICKHOUSE_DATABASE}, users) WHERE UserID IN (SELECT arrayJoin([1468013291393583084, 1321770221388956068])));/" | $CLICKHOUSE_CLIENT | grep -vE '^4$' && echo 'Fail!' && break; echo -n '.'; done; echo
 
 $CLICKHOUSE_CLIENT --query="DROP TABLE users;";
diff --git a/tests/queries/0_stateless/00133_long_shard_memory_tracker_and_exception_safety.sh b/tests/queries/0_stateless/00133_long_shard_memory_tracker_and_exception_safety.sh
index d57efaa1f0e..f0c0354ab33 100755
--- a/tests/queries/0_stateless/00133_long_shard_memory_tracker_and_exception_safety.sh
+++ b/tests/queries/0_stateless/00133_long_shard_memory_tracker_and_exception_safety.sh
@@ -5,7 +5,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-$CLICKHOUSE_CLIENT -n --query="
+$CLICKHOUSE_CLIENT --query="
     DROP TABLE IF EXISTS numbers_100k;
     CREATE VIEW numbers_100k AS SELECT * FROM system.numbers LIMIT 100000;
 ";
diff --git a/tests/queries/0_stateless/00385_storage_file_and_clickhouse-local_app_long.sh b/tests/queries/0_stateless/00385_storage_file_and_clickhouse-local_app_long.sh
index b59aae83e81..b6cc270994f 100755
--- a/tests/queries/0_stateless/00385_storage_file_and_clickhouse-local_app_long.sh
+++ b/tests/queries/0_stateless/00385_storage_file_and_clickhouse-local_app_long.sh
@@ -48,13 +48,13 @@ pack_unpack_compare "SELECT name, is_aggregate FROM system.functions" "name Stri
 echo
 # Check settings are passed correctly
 ${CLICKHOUSE_LOCAL} --max_rows_in_distinct=33 -q "SELECT name, value FROM system.settings WHERE name = 'max_rows_in_distinct'"
-${CLICKHOUSE_LOCAL} -n -q "SET max_rows_in_distinct=33; SELECT name, value FROM system.settings WHERE name = 'max_rows_in_distinct'"
+${CLICKHOUSE_LOCAL} -q "SET max_rows_in_distinct=33; SELECT name, value FROM system.settings WHERE name = 'max_rows_in_distinct'"
 ${CLICKHOUSE_LOCAL} --max_bytes_before_external_group_by=1 --max_block_size=10 -q "SELECT sum(ignore(*)) FROM (SELECT number, count() FROM numbers(1000) GROUP BY number)"
 echo
 # Check exta options, we expect zero exit code and no stderr output
-(${CLICKHOUSE_LOCAL} --ignore-error -n --echo -q "SELECT nothing_to_do();SELECT 42;" 2>/dev/null || echo "Wrong RC")
+(${CLICKHOUSE_LOCAL} --ignore-error --echo -q "SELECT nothing_to_do();SELECT 42;" 2>/dev/null || echo "Wrong RC")
 echo
-${CLICKHOUSE_LOCAL} -n -q "CREATE TABLE sophisticated_default
+${CLICKHOUSE_LOCAL} -q "CREATE TABLE sophisticated_default
 (
     a UInt8 DEFAULT 3,
     b UInt8 ALIAS a + 5,
diff --git a/tests/queries/0_stateless/00505_secure.sh b/tests/queries/0_stateless/00505_secure.sh
index eaa50bce6b1..eed0d3bf5c6 100755
--- a/tests/queries/0_stateless/00505_secure.sh
+++ b/tests/queries/0_stateless/00505_secure.sh
@@ -23,7 +23,7 @@ $CLICKHOUSE_CLIENT_SECURE -q "SELECT 4;"
 
 # TODO: can test only on unchanged port. Possible solutions: generate config or pass shard port via command line
 if [[ "$CLICKHOUSE_PORT_TCP_SECURE" = "$CLICKHOUSE_PORT_TCP_SECURE" ]]; then
-    cat "$CURDIR"/00505_distributed_secure.data | $CLICKHOUSE_CLIENT_SECURE -n -m
+    cat "$CURDIR"/00505_distributed_secure.data | $CLICKHOUSE_CLIENT_SECURE -m
 else
     tail -n 13 "$CURDIR"/00505_secure.reference
 fi
diff --git a/tests/queries/0_stateless/00510_materizlized_view_and_deduplication_zookeeper.sql b/tests/queries/0_stateless/00510_materizlized_view_and_deduplication_zookeeper.sql
index 51e6a513608..329f6ad2248 100644
--- a/tests/queries/0_stateless/00510_materizlized_view_and_deduplication_zookeeper.sql
+++ b/tests/queries/0_stateless/00510_materizlized_view_and_deduplication_zookeeper.sql
@@ -6,6 +6,8 @@ DROP TABLE IF EXISTS without_deduplication;
 DROP TABLE IF EXISTS with_deduplication_mv;
 DROP TABLE IF EXISTS without_deduplication_mv;
 
+SET database_replicated_allow_explicit_uuid=3;
+SET database_replicated_allow_replicated_engine_arguments=3;
 CREATE TABLE with_deduplication(x UInt32)
     ENGINE ReplicatedMergeTree('/clickhouse/tables/{database}/test_00510/with_deduplication', 'r1') ORDER BY x;
 CREATE TABLE without_deduplication(x UInt32)
diff --git a/tests/queries/0_stateless/00531_client_ignore_error.sh b/tests/queries/0_stateless/00531_client_ignore_error.sh
index d3215e1beac..553cb9fa897 100755
--- a/tests/queries/0_stateless/00531_client_ignore_error.sh
+++ b/tests/queries/0_stateless/00531_client_ignore_error.sh
@@ -5,8 +5,8 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-echo "SELECT 1; SELECT 2; SELECT CAST(); SELECT ';'; SELECT 3;SELECT CAST();SELECT 4;" | $CLICKHOUSE_CLIENT -n --ignore-error 2>/dev/null
-echo "SELECT CAST();" | $CLICKHOUSE_CLIENT -n --ignore-error 2>/dev/null
-echo "SELECT 5;" | $CLICKHOUSE_CLIENT -n --ignore-error
+echo "SELECT 1; SELECT 2; SELECT CAST(); SELECT ';'; SELECT 3;SELECT CAST();SELECT 4;" | $CLICKHOUSE_CLIENT --ignore-error 2>/dev/null
+echo "SELECT CAST();" | $CLICKHOUSE_CLIENT --ignore-error 2>/dev/null
+echo "SELECT 5;" | $CLICKHOUSE_CLIENT --ignore-error
 
 #$CLICKHOUSE_CLIENT -q "SELECT 'Still alive'"
diff --git a/tests/queries/0_stateless/00534_client_ignore_error.sh b/tests/queries/0_stateless/00534_client_ignore_error.sh
index d3215e1beac..553cb9fa897 100755
--- a/tests/queries/0_stateless/00534_client_ignore_error.sh
+++ b/tests/queries/0_stateless/00534_client_ignore_error.sh
@@ -5,8 +5,8 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-echo "SELECT 1; SELECT 2; SELECT CAST(); SELECT ';'; SELECT 3;SELECT CAST();SELECT 4;" | $CLICKHOUSE_CLIENT -n --ignore-error 2>/dev/null
-echo "SELECT CAST();" | $CLICKHOUSE_CLIENT -n --ignore-error 2>/dev/null
-echo "SELECT 5;" | $CLICKHOUSE_CLIENT -n --ignore-error
+echo "SELECT 1; SELECT 2; SELECT CAST(); SELECT ';'; SELECT 3;SELECT CAST();SELECT 4;" | $CLICKHOUSE_CLIENT --ignore-error 2>/dev/null
+echo "SELECT CAST();" | $CLICKHOUSE_CLIENT --ignore-error 2>/dev/null
+echo "SELECT 5;" | $CLICKHOUSE_CLIENT --ignore-error
 
 #$CLICKHOUSE_CLIENT -q "SELECT 'Still alive'"
diff --git a/tests/queries/0_stateless/00609_mv_index_in_in.sql b/tests/queries/0_stateless/00609_mv_index_in_in.sql
index bd9f35350c1..848938780c2 100644
--- a/tests/queries/0_stateless/00609_mv_index_in_in.sql
+++ b/tests/queries/0_stateless/00609_mv_index_in_in.sql
@@ -1,4 +1,4 @@
--- Tags: no-ordinary-database
+-- Tags: no-ordinary-database, no-parallel
 
 DROP TABLE IF EXISTS test_00609;
 DROP TABLE IF EXISTS test_mv_00609;
@@ -6,6 +6,7 @@ DROP TABLE IF EXISTS test_mv_00609;
 create table test_00609 (a Int8) engine=Memory;
 
 insert into test_00609 values (1);
+set database_replicated_allow_explicit_uuid=3;
 set allow_deprecated_syntax_for_merge_tree=1;
 create materialized view test_mv_00609 uuid '00000609-1000-4000-8000-000000000001' Engine=MergeTree(date, (a), 8192) populate as select a, toDate('2000-01-01') date from test_00609;
 
diff --git a/tests/queries/0_stateless/00686_client_exit_code.sh b/tests/queries/0_stateless/00686_client_exit_code.sh
index 9586ddbd0a3..e348f93f30d 100755
--- a/tests/queries/0_stateless/00686_client_exit_code.sh
+++ b/tests/queries/0_stateless/00686_client_exit_code.sh
@@ -8,5 +8,5 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=./mergetree_mutations.lib
 . "$CURDIR"/mergetree_mutations.lib
 
-echo "INSERT INTO test FORMAT CSV" | ${CLICKHOUSE_CLIENT} -n 2>/dev/null
+echo "INSERT INTO test FORMAT CSV" | ${CLICKHOUSE_CLIENT} 2>/dev/null
 echo $?
diff --git a/tests/queries/0_stateless/00705_drop_create_merge_tree.sh b/tests/queries/0_stateless/00705_drop_create_merge_tree.sh
index ea8b9d02e49..fd002668696 100755
--- a/tests/queries/0_stateless/00705_drop_create_merge_tree.sh
+++ b/tests/queries/0_stateless/00705_drop_create_merge_tree.sh
@@ -5,8 +5,8 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-yes 'CREATE TABLE IF NOT EXISTS table (x UInt8) ENGINE = MergeTree ORDER BY tuple();' | head -n 1000 | $CLICKHOUSE_CLIENT --multiquery &
-yes 'DROP TABLE IF EXISTS table;' | head -n 1000 | $CLICKHOUSE_CLIENT --multiquery &
+yes 'CREATE TABLE IF NOT EXISTS table (x UInt8) ENGINE = MergeTree ORDER BY tuple();' | head -n 1000 | $CLICKHOUSE_CLIENT &
+yes 'DROP TABLE IF EXISTS table;' | head -n 1000 | $CLICKHOUSE_CLIENT &
 wait
 
 ${CLICKHOUSE_CLIENT} --query "DROP TABLE IF EXISTS table"
diff --git a/tests/queries/0_stateless/00731_long_merge_tree_select_opened_files.sh b/tests/queries/0_stateless/00731_long_merge_tree_select_opened_files.sh
index 5a4fd901f8d..285fd3945f9 100755
--- a/tests/queries/0_stateless/00731_long_merge_tree_select_opened_files.sh
+++ b/tests/queries/0_stateless/00731_long_merge_tree_select_opened_files.sh
@@ -12,7 +12,7 @@ settings="--log_queries=1 --log_query_threads=1 --log_profile_events=1 --log_que
 
 # Test insert logging on each block and checkPacket() method
 
-$CLICKHOUSE_CLIENT $settings -n -q "
+$CLICKHOUSE_CLIENT $settings -q "
 DROP TABLE IF EXISTS merge_tree_table;
 CREATE TABLE merge_tree_table (id UInt64, date Date, uid UInt32) ENGINE = MergeTree(date, id, 8192);"
 
diff --git a/tests/queries/0_stateless/00738_lock_for_inner_table.sh b/tests/queries/0_stateless/00738_lock_for_inner_table.sh
index b62a639d8f4..9bc84dd1063 100755
--- a/tests/queries/0_stateless/00738_lock_for_inner_table.sh
+++ b/tests/queries/0_stateless/00738_lock_for_inner_table.sh
@@ -7,6 +7,8 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
+CLICKHOUSE_CLIENT="${CLICKHOUSE_CLIENT} --database_replicated_allow_explicit_uuid 3"
+
 # there are some issues with Atomic database, let's generate it uniq
 # otherwise flaky check will not pass.
 uuid=$(${CLICKHOUSE_CLIENT} --query "SELECT reinterpretAsUUID(currentDatabase())")
diff --git a/tests/queries/0_stateless/00838_system_tables_drop_table_race.sh b/tests/queries/0_stateless/00838_system_tables_drop_table_race.sh
index da4d3b05987..a29d2e5bc71 100755
--- a/tests/queries/0_stateless/00838_system_tables_drop_table_race.sh
+++ b/tests/queries/0_stateless/00838_system_tables_drop_table_race.sh
@@ -9,7 +9,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 
 $CLICKHOUSE_CLIENT -q "DROP TABLE IF EXISTS table"
 
-seq 1 100 | sed -r -e "s/.+/CREATE TABLE table (x UInt8) ENGINE = MergeTree ORDER BY x; DROP TABLE table;/" | $CLICKHOUSE_CLIENT -n &
-seq 1 100 | sed -r -e "s/.+/SELECT * FROM system.tables WHERE database = '${CLICKHOUSE_DATABASE}' LIMIT 1000000, 1;/" | $CLICKHOUSE_CLIENT -n 2>/dev/null &
+seq 1 100 | sed -r -e "s/.+/CREATE TABLE table (x UInt8) ENGINE = MergeTree ORDER BY x; DROP TABLE table;/" | $CLICKHOUSE_CLIENT &
+seq 1 100 | sed -r -e "s/.+/SELECT * FROM system.tables WHERE database = '${CLICKHOUSE_DATABASE}' LIMIT 1000000, 1;/" | $CLICKHOUSE_CLIENT 2>/dev/null &
 
 wait
diff --git a/tests/queries/0_stateless/00840_long_concurrent_select_and_drop_deadlock.sh b/tests/queries/0_stateless/00840_long_concurrent_select_and_drop_deadlock.sh
index f0bc52ee356..cb774116356 100755
--- a/tests/queries/0_stateless/00840_long_concurrent_select_and_drop_deadlock.sh
+++ b/tests/queries/0_stateless/00840_long_concurrent_select_and_drop_deadlock.sh
@@ -27,7 +27,7 @@ function thread_drop_create()
     while [ $SECONDS -lt "$TIMELIMIT" ] && [ $it -lt 100 ];
     do
         it=$((it+1))
-        $CLICKHOUSE_CLIENT -nm -q "
+        $CLICKHOUSE_CLIENT -m -q "
             drop table if exists view_00840;
             create view view_00840 as select count(*),database,table from system.columns group by database,table;
         "
diff --git a/tests/queries/0_stateless/00900_long_parquet.sh b/tests/queries/0_stateless/00900_long_parquet.sh
index 07d2f24e446..86a0d013078 100755
--- a/tests/queries/0_stateless/00900_long_parquet.sh
+++ b/tests/queries/0_stateless/00900_long_parquet.sh
@@ -8,11 +8,11 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 . "$CUR_DIR"/../shell_config.sh
 
 
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     DROP TABLE IF EXISTS contributors;
     CREATE TABLE contributors (name String) ENGINE = Memory;"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM system.contributors ORDER BY name DESC FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO contributors FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     -- random results
     SELECT * FROM contributors LIMIT 10 FORMAT Null;
     DROP TABLE contributors;
@@ -21,30 +21,30 @@ ${CLICKHOUSE_CLIENT} -n --query="
     CREATE TABLE parquet_numbers (number UInt64) ENGINE = Memory;"
 # less than default block size (65k)
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM system.numbers LIMIT 10000 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_numbers FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_numbers ORDER BY number DESC LIMIT 10;
     TRUNCATE TABLE parquet_numbers;"
 
 # More than default block size
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM system.numbers LIMIT 100000 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_numbers FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_numbers ORDER BY number DESC LIMIT 10;
     TRUNCATE TABLE parquet_numbers;"
 
 ${CLICKHOUSE_CLIENT} --max_block_size=2 --query="SELECT * FROM system.numbers LIMIT 3 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_numbers FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_numbers ORDER BY number DESC LIMIT 10;
 
     TRUNCATE TABLE parquet_numbers;"
 ${CLICKHOUSE_CLIENT} --max_block_size=1 --query="SELECT * FROM system.numbers LIMIT 1000 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_numbers FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_numbers ORDER BY number DESC LIMIT 10;
     DROP TABLE parquet_numbers;
 
     DROP TABLE IF EXISTS parquet_events;
     CREATE TABLE parquet_events (event String, value UInt64, description String) ENGINE = Memory;"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM system.events FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_events FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT event, description FROM parquet_events WHERE event IN ('ContextLock', 'Query') ORDER BY event;
     DROP TABLE parquet_events;
 
@@ -78,7 +78,7 @@ ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_types2 ORDER BY int8 FORMAT
 echo diff:
 diff "${CLICKHOUSE_TMP}"/parquet_all_types_1.dump "${CLICKHOUSE_TMP}"/parquet_all_types_2.dump
 
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     TRUNCATE TABLE parquet_types2;
     INSERT INTO parquet_types3 values (       79,          81,          82,            83,          84,            85,          86,            87,              88,              89,         'str01',                  'fstr1', '2003-03-04', '2004-05-06', toDateTime64('2004-05-06 07:08:09.012', 9));"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_types3 ORDER BY int8 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_types2 FORMAT Parquet"
@@ -88,7 +88,7 @@ ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_types4 values (       80,
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_types4 ORDER BY int8 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_types2 FORMAT Parquet"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_types1 ORDER BY int8 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_types4 FORMAT Parquet"
 
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT 'dest:';
     SELECT * FROM parquet_types2 ORDER BY int8;
     SELECT 'min:';
@@ -106,7 +106,7 @@ ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_types5 ORDER BY int8 FORMAT
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_types5 ORDER BY int8 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_types6 FORMAT Parquet"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_types1 ORDER BY int8 FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_types6 FORMAT Parquet"
 echo dest from null:
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_types6 ORDER BY int8;
 
     DROP TABLE parquet_types5;
@@ -126,7 +126,7 @@ ${CLICKHOUSE_CLIENT} -n --query="
     INSERT INTO parquet_arrays VALUES (2, [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], []);"
 
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_arrays FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_arrays FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_arrays ORDER BY id;
 
     DROP TABLE parquet_arrays;
@@ -135,7 +135,7 @@ ${CLICKHOUSE_CLIENT} -n --query="
     CREATE TABLE parquet_nullable_arrays (id UInt32, a1 Array(Nullable(UInt32)), a2 Array(Nullable(String)), a3 Array(Nullable(Decimal(4, 2)))) engine=Memory();
     INSERT INTO parquet_nullable_arrays VALUES (1, [1, Null, 2], [Null, 'Some string', Null], [0.001, Null, 42.42]), (2, [Null], [Null], [Null]), (3, [], [], []);"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_nullable_arrays FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_nullable_arrays FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_nullable_arrays ORDER BY id;
     DROP TABLE parquet_nullable_arrays;
 
@@ -143,7 +143,7 @@ ${CLICKHOUSE_CLIENT} -n --query="
     CREATE TABLE parquet_nested_arrays (a1 Array(Array(Array(UInt32))), a2 Array(Array(Array(String))), a3 Array(Array(Nullable(UInt32))), a4 Array(Array(Nullable(String)))) engine=Memory();
     INSERT INTO parquet_nested_arrays VALUES ([[[1,2,3], [1,2,3]], [[1,2,3]], [[], [1,2,3]]], [[['Some string', 'Some string'], []], [['Some string']], [[]]], [[Null, 1, 2], [Null], [1, 2], []], [['Some string', Null, 'Some string'], [Null], []]);"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_nested_arrays FORMAT Parquet" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_nested_arrays FORMAT Parquet"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_nested_arrays;
     DROP TABLE parquet_nested_arrays;
 
@@ -151,6 +151,6 @@ ${CLICKHOUSE_CLIENT} -n --query="
     CREATE TABLE parquet_decimal (d1 Decimal32(4), d2 Decimal64(8), d3 Decimal128(16), d4 Decimal256(32)) ENGINE = Memory;
     INSERT INTO TABLE parquet_decimal VALUES (0.123, 0.123123123, 0.123123123123, 0.123123123123123123);"
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM parquet_decimal FORMAT Arrow" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO parquet_decimal FORMAT Arrow"
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM parquet_decimal;
     DROP TABLE parquet_decimal;"
diff --git a/tests/queries/0_stateless/00900_long_parquet_decimal.sh b/tests/queries/0_stateless/00900_long_parquet_decimal.sh
index a819dcbcdc3..14e8fdcc038 100755
--- a/tests/queries/0_stateless/00900_long_parquet_decimal.sh
+++ b/tests/queries/0_stateless/00900_long_parquet_decimal.sh
@@ -7,7 +7,7 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CUR_DIR"/../shell_config.sh
 
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     DROP TABLE IF EXISTS decimal;
     DROP TABLE IF EXISTS decimal2;
 
@@ -26,7 +26,7 @@ ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d, e, f, g
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal2 ORDER BY a, b, c, d, e, f, g, h, i, j;" > "${CLICKHOUSE_TMP}"/parquet_decimal0_2.dump
 echo diff0:
 diff "${CLICKHOUSE_TMP}"/parquet_decimal0_1.dump "${CLICKHOUSE_TMP}"/parquet_decimal0_2.dump
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     DROP TABLE IF EXISTS decimal;
     DROP TABLE IF EXISTS decimal2;
 
@@ -61,7 +61,7 @@ ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d, e, f, g
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal2 ORDER BY a, b, c, d, e, f, g, h, i, j;" > "${CLICKHOUSE_TMP}"/parquet_decimal1_2.dump
 echo diff1:
 diff "${CLICKHOUSE_TMP}"/parquet_decimal1_1.dump "${CLICKHOUSE_TMP}"/parquet_decimal1_2.dump
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     DROP TABLE IF EXISTS decimal;
     DROP TABLE IF EXISTS decimal2;
 
@@ -75,7 +75,7 @@ ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d, e, f, g
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal2 ORDER BY a, b, c, d, e, f, g, h, i, j;" > "${CLICKHOUSE_TMP}"/parquet_decimal2_2.dump
 echo diff2:
 diff "${CLICKHOUSE_TMP}"/parquet_decimal2_1.dump "${CLICKHOUSE_TMP}"/parquet_decimal2_2.dump
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     DROP TABLE IF EXISTS decimal;
     DROP TABLE IF EXISTS decimal2;
 
@@ -86,7 +86,7 @@ ${CLICKHOUSE_CLIENT} -n --query="
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d  FORMAT Parquet;" > "${CLICKHOUSE_TMP}"/parquet_decimal3_1.parquet
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d FORMAT Parquet;" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO decimal2 FORMAT Parquet" 2> /dev/null
 echo nothing:
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM decimal2 ORDER BY a, b, c, d;
     TRUNCATE TABLE decimal2;
 
@@ -94,7 +94,7 @@ ${CLICKHOUSE_CLIENT} -n --query="
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d FORMAT Parquet;" > "${CLICKHOUSE_TMP}"/parquet_decimal3_2.parquet
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d FORMAT Parquet;" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO decimal2 FORMAT Parquet"
 echo nulls:
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT * FROM decimal2 ORDER BY a, b, c, d;
     TRUNCATE TABLE decimal2;
 
@@ -104,7 +104,7 @@ ${CLICKHOUSE_CLIENT} -n --query="
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d FORMAT Parquet;" > "${CLICKHOUSE_TMP}"/parquet_decimal3_3.parquet
 ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal ORDER BY a, b, c, d FORMAT Parquet;" | ${CLICKHOUSE_CLIENT} --query="INSERT INTO decimal2 FORMAT Parquet"
 
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     SELECT 'full orig:';
     SELECT * FROM decimal ORDER BY a, b, c, d;
     SELECT 'full inserted:';
@@ -115,6 +115,6 @@ ${CLICKHOUSE_CLIENT} --query="SELECT * FROM decimal2 ORDER BY a, b, c, d;" > "${
 
 echo diff3:
 diff "${CLICKHOUSE_TMP}"/parquet_decimal3_1.dump "${CLICKHOUSE_TMP}"/parquet_decimal3_2.dump
-${CLICKHOUSE_CLIENT} -n --query="
+${CLICKHOUSE_CLIENT} --query="
     DROP TABLE IF EXISTS decimal;
     DROP TABLE IF EXISTS decimal2;"
diff --git a/tests/queries/0_stateless/00921_datetime64_compatibility_long.sh b/tests/queries/0_stateless/00921_datetime64_compatibility_long.sh
index d310a2c3612..152d5a847b7 100755
--- a/tests/queries/0_stateless/00921_datetime64_compatibility_long.sh
+++ b/tests/queries/0_stateless/00921_datetime64_compatibility_long.sh
@@ -13,5 +13,5 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # ${CURDIR}/00921_datetime64_compatibility.python
 
 python3 "${CURDIR}"/00921_datetime64_compatibility_long.python \
-    | ${CLICKHOUSE_CLIENT} --ignore-error -nm --calculate_text_stack_trace 0 --log-level 'error' 2>&1 \
+    | ${CLICKHOUSE_CLIENT} --ignore-error -m --calculate_text_stack_trace 0 --log-level 'error' 2>&1 \
     | grep -v -e 'Received exception .*$' -e '^(query: ' | sed 's/^\(Code: [0-9]\+\).*$/\1/g'
diff --git a/tests/queries/0_stateless/00975_indices_mutation_replicated_zookeeper_long.sh b/tests/queries/0_stateless/00975_indices_mutation_replicated_zookeeper_long.sh
index 686dd7f6df0..1d35daf9f2e 100755
--- a/tests/queries/0_stateless/00975_indices_mutation_replicated_zookeeper_long.sh
+++ b/tests/queries/0_stateless/00975_indices_mutation_replicated_zookeeper_long.sh
@@ -11,7 +11,7 @@ $CLICKHOUSE_CLIENT --query="DROP TABLE IF EXISTS indices_mutaions1;"
 $CLICKHOUSE_CLIENT --query="DROP TABLE IF EXISTS indices_mutaions2;"
 
 
-$CLICKHOUSE_CLIENT -n --query="
+$CLICKHOUSE_CLIENT --query="
 CREATE TABLE indices_mutaions1
 (
     u64 UInt64,
diff --git a/tests/queries/0_stateless/00991_system_parts_race_condition_long.sh b/tests/queries/0_stateless/00991_system_parts_race_condition_long.sh
index 8243c6bde62..df330b82c80 100755
--- a/tests/queries/0_stateless/00991_system_parts_race_condition_long.sh
+++ b/tests/queries/0_stateless/00991_system_parts_race_condition_long.sh
@@ -22,7 +22,7 @@ function thread1()
 
 function thread2()
 {
-    while true; do $CLICKHOUSE_CLIENT -n --query "ALTER TABLE alter_table ADD COLUMN h String '0'; ALTER TABLE alter_table MODIFY COLUMN h UInt64; ALTER TABLE alter_table DROP COLUMN h;"; done
+    while true; do $CLICKHOUSE_CLIENT --query "ALTER TABLE alter_table ADD COLUMN h String '0'; ALTER TABLE alter_table MODIFY COLUMN h UInt64; ALTER TABLE alter_table DROP COLUMN h;"; done
 }
 
 function thread3()
diff --git a/tests/queries/0_stateless/00992_system_parts_race_condition_zookeeper_long.sh b/tests/queries/0_stateless/00992_system_parts_race_condition_zookeeper_long.sh
index 02a739ece4a..d5c0248e2b3 100755
--- a/tests/queries/0_stateless/00992_system_parts_race_condition_zookeeper_long.sh
+++ b/tests/queries/0_stateless/00992_system_parts_race_condition_zookeeper_long.sh
@@ -9,7 +9,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 
 set -e
 
-$CLICKHOUSE_CLIENT -n -q "
+$CLICKHOUSE_CLIENT -q "
     DROP TABLE IF EXISTS alter_table0;
     DROP TABLE IF EXISTS alter_table1;
 
@@ -31,7 +31,7 @@ function thread1()
 
 function thread2()
 {
-    while true; do $CLICKHOUSE_CLIENT -n --query "ALTER TABLE alter_table0 ADD COLUMN h String DEFAULT '0'; ALTER TABLE alter_table0 MODIFY COLUMN h UInt64; ALTER TABLE alter_table0 DROP COLUMN h;"; done
+    while true; do $CLICKHOUSE_CLIENT --query "ALTER TABLE alter_table0 ADD COLUMN h String DEFAULT '0'; ALTER TABLE alter_table0 MODIFY COLUMN h UInt64; ALTER TABLE alter_table0 DROP COLUMN h;"; done
 }
 
 function thread3()
@@ -87,6 +87,6 @@ check_replication_consistency "alter_table" "count(), sum(a), sum(b), round(sum(
 
 $CLICKHOUSE_CLIENT -q "SELECT table, lost_part_count FROM system.replicas WHERE database=currentDatabase() AND lost_part_count!=0";
 
-$CLICKHOUSE_CLIENT -n -q "DROP TABLE alter_table0;" 2> >(grep -F -v 'is already started to be removing by another replica right now') &
-$CLICKHOUSE_CLIENT -n -q "DROP TABLE alter_table1;" 2> >(grep -F -v 'is already started to be removing by another replica right now') &
+$CLICKHOUSE_CLIENT -q "DROP TABLE alter_table0;" 2> >(grep -F -v 'is already started to be removing by another replica right now') &
+$CLICKHOUSE_CLIENT -q "DROP TABLE alter_table1;" 2> >(grep -F -v 'is already started to be removing by another replica right now') &
 wait
diff --git a/tests/queries/0_stateless/01014_lazy_database_basic.sh b/tests/queries/0_stateless/01014_lazy_database_basic.sh
index ea7603b2519..55d18a7c527 100755
--- a/tests/queries/0_stateless/01014_lazy_database_basic.sh
+++ b/tests/queries/0_stateless/01014_lazy_database_basic.sh
@@ -5,9 +5,9 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-${CLICKHOUSE_CLIENT} -n -q "DROP DATABASE IF EXISTS testlazy"
+${CLICKHOUSE_CLIENT} -q "DROP DATABASE IF EXISTS testlazy"
 
-${CLICKHOUSE_CLIENT} -n -q "
+${CLICKHOUSE_CLIENT} -q "
     CREATE DATABASE testlazy ENGINE = Lazy(1);
     CREATE TABLE testlazy.log (a UInt64, b UInt64) ENGINE = Log;
     CREATE TABLE testlazy.slog (a UInt64, b UInt64) ENGINE = StripeLog;
@@ -30,7 +30,7 @@ ${CLICKHOUSE_CLIENT} -q "
 
 sleep 1.5
 
-${CLICKHOUSE_CLIENT} -n -q "
+${CLICKHOUSE_CLIENT} -q "
     SELECT * FROM testlazy.log LIMIT 0; -- drop testlazy.log from cache
     RENAME TABLE testlazy.log TO testlazy.log2;
     SELECT database, name FROM system.tables WHERE database = 'testlazy';
@@ -44,7 +44,7 @@ ${CLICKHOUSE_CLIENT} -q "
 
 sleep 1.5
 
-${CLICKHOUSE_CLIENT} -n -q "
+${CLICKHOUSE_CLIENT} -q "
     INSERT INTO testlazy.log2 VALUES (1, 1);
     INSERT INTO testlazy.slog VALUES (2, 2);
     INSERT INTO testlazy.tlog VALUES (3, 3);
@@ -55,14 +55,14 @@ ${CLICKHOUSE_CLIENT} -n -q "
 
 sleep 1.5
 
-${CLICKHOUSE_CLIENT} -n -q "
+${CLICKHOUSE_CLIENT} -q "
     SELECT * FROM testlazy.log2 LIMIT 0; -- drop testlazy.log2 from cache
     DROP TABLE testlazy.log2;
 "
 
 sleep 1.5
 
-${CLICKHOUSE_CLIENT} -n -q "
+${CLICKHOUSE_CLIENT} -q "
     SELECT * FROM testlazy.slog;
     SELECT * FROM testlazy.tlog;
 "
diff --git a/tests/queries/0_stateless/01014_lazy_database_concurrent_recreate_reattach_and_show_tables.sh b/tests/queries/0_stateless/01014_lazy_database_concurrent_recreate_reattach_and_show_tables.sh
index e4b3a31b13f..ff2c0b8821e 100755
--- a/tests/queries/0_stateless/01014_lazy_database_concurrent_recreate_reattach_and_show_tables.sh
+++ b/tests/queries/0_stateless/01014_lazy_database_concurrent_recreate_reattach_and_show_tables.sh
@@ -83,7 +83,7 @@ export -f recreate_lazy_func4;
 export -f show_tables_func;
 
 
-${CLICKHOUSE_CLIENT} -n -q "
+${CLICKHOUSE_CLIENT} -q "
     DROP DATABASE IF EXISTS $CURR_DATABASE;
     CREATE DATABASE $CURR_DATABASE ENGINE = Lazy(1);
 "
diff --git a/tests/queries/0_stateless/01018_ddl_dictionaries_concurrent_requrests.sh b/tests/queries/0_stateless/01018_ddl_dictionaries_concurrent_requrests.sh
index 872b0a7c1a1..82221eb06e9 100755
--- a/tests/queries/0_stateless/01018_ddl_dictionaries_concurrent_requrests.sh
+++ b/tests/queries/0_stateless/01018_ddl_dictionaries_concurrent_requrests.sh
@@ -7,7 +7,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 
 set -e
 
-$CLICKHOUSE_CLIENT -n -q "
+$CLICKHOUSE_CLIENT -q "
     DROP DATABASE IF EXISTS database_for_dict;
     DROP TABLE IF EXISTS table_for_dict1;
     DROP TABLE IF EXISTS table_for_dict2;
@@ -44,7 +44,7 @@ function thread3()
 
 function thread4()
 {
-    while true; do $CLICKHOUSE_CLIENT -n -q "
+    while true; do $CLICKHOUSE_CLIENT -q "
         SELECT * FROM database_for_dict.dict1 FORMAT Null;
         SELECT * FROM database_for_dict.dict2 FORMAT Null;
     " ||: ; done
@@ -52,7 +52,7 @@ function thread4()
 
 function thread5()
 {
-    while true; do $CLICKHOUSE_CLIENT -n -q "
+    while true; do $CLICKHOUSE_CLIENT -q "
         SELECT dictGetString('database_for_dict.dict1', 'value_column', toUInt64(number)) from numbers(1000) FROM FORMAT Null;
         SELECT dictGetString('database_for_dict.dict2', 'value_column', toUInt64(number)) from numbers(1000) FROM FORMAT Null;
     " ||: ; done
@@ -117,7 +117,7 @@ $CLICKHOUSE_CLIENT -q "SELECT 'Still alive'"
 $CLICKHOUSE_CLIENT -q "ATTACH DICTIONARY IF NOT EXISTS database_for_dict.dict1"
 $CLICKHOUSE_CLIENT -q "ATTACH DICTIONARY IF NOT EXISTS database_for_dict.dict2"
 
-$CLICKHOUSE_CLIENT -n -q "
+$CLICKHOUSE_CLIENT -q "
     DROP DATABASE database_for_dict;
     DROP TABLE table_for_dict1;
     DROP TABLE table_for_dict2;
diff --git a/tests/queries/0_stateless/01019_alter_materialized_view_atomic.sh b/tests/queries/0_stateless/01019_alter_materialized_view_atomic.sh
index 4bd21fcee02..eb12a76eb62 100755
--- a/tests/queries/0_stateless/01019_alter_materialized_view_atomic.sh
+++ b/tests/queries/0_stateless/01019_alter_materialized_view_atomic.sh
@@ -7,7 +7,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-$CLICKHOUSE_CLIENT --multiquery <<EOF
+$CLICKHOUSE_CLIENT <<EOF
 DROP TABLE IF EXISTS src;
 DROP TABLE IF EXISTS mv;
 
diff --git a/tests/queries/0_stateless/01035_avg_weighted_long.sh b/tests/queries/0_stateless/01035_avg_weighted_long.sh
index 92c86981b37..6c3356324ed 100755
--- a/tests/queries/0_stateless/01035_avg_weighted_long.sh
+++ b/tests/queries/0_stateless/01035_avg_weighted_long.sh
@@ -28,7 +28,7 @@ exttypes=("Int128" "Int256" "UInt256")
             echo "SELECT avgWeighted(to${left}(1), to${right}(2));"
         done
     done
-) | $CLICKHOUSE_CLIENT_BINARY -nm
+) | $CLICKHOUSE_CLIENT_BINARY -m
 
 ${CLICKHOUSE_CLIENT} --server_logs_file=/dev/null --query="SELECT avgWeighted(['string'], toFloat64(0))" 2>&1 \
   | grep -c 'Code: 43. DB::Exception: .* DB::Exception:.* Types .* are non-conforming as arguments for aggregate function avgWeighted'
diff --git a/tests/queries/0_stateless/01053_ssd_dictionary.sh b/tests/queries/0_stateless/01053_ssd_dictionary.sh
index fdd40940ab5..7dd3fa4657a 100755
--- a/tests/queries/0_stateless/01053_ssd_dictionary.sh
+++ b/tests/queries/0_stateless/01053_ssd_dictionary.sh
@@ -6,7 +6,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-$CLICKHOUSE_CLIENT --allow_deprecated_database_ordinary=1 -n --query="
+$CLICKHOUSE_CLIENT --allow_deprecated_database_ordinary=1 --query="
   DROP DATABASE IF EXISTS 01053_db;
 
   CREATE DATABASE 01053_db;
diff --git a/tests/queries/0_stateless/01079_bad_alters_zookeeper_long.sh b/tests/queries/0_stateless/01079_bad_alters_zookeeper_long.sh
index 39e65af039b..22f8e5269bd 100755
--- a/tests/queries/0_stateless/01079_bad_alters_zookeeper_long.sh
+++ b/tests/queries/0_stateless/01079_bad_alters_zookeeper_long.sh
@@ -7,7 +7,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 
 $CLICKHOUSE_CLIENT --query "DROP TABLE IF EXISTS table_for_bad_alters";
 
-$CLICKHOUSE_CLIENT -n --query "CREATE TABLE table_for_bad_alters (
+$CLICKHOUSE_CLIENT --query "CREATE TABLE table_for_bad_alters (
     key UInt64,
     value1 UInt8,
     value2 String
diff --git a/tests/queries/0_stateless/01111_create_drop_replicated_db_stress.sh b/tests/queries/0_stateless/01111_create_drop_replicated_db_stress.sh
index 619b6e91d11..9101b9faa3d 100755
--- a/tests/queries/0_stateless/01111_create_drop_replicated_db_stress.sh
+++ b/tests/queries/0_stateless/01111_create_drop_replicated_db_stress.sh
@@ -30,7 +30,7 @@ function drop_db()
         database=$($CLICKHOUSE_CLIENT -q "select name from system.databases where name like '${CLICKHOUSE_DATABASE}%' order by rand() limit 1")
         if [[ "$database" == "$CLICKHOUSE_DATABASE" ]]; then continue; fi
         if [ -z "$database" ]; then continue; fi
-        $CLICKHOUSE_CLIENT -n --query \
+        $CLICKHOUSE_CLIENT --query \
         "drop database if exists $database" 2>&1| grep -Fa "Exception: "
         sleep 0.$RANDOM
     done
diff --git a/tests/queries/0_stateless/01114_database_atomic.sh b/tests/queries/0_stateless/01114_database_atomic.sh
index 5eebb558575..5fe85136d05 100755
--- a/tests/queries/0_stateless/01114_database_atomic.sh
+++ b/tests/queries/0_stateless/01114_database_atomic.sh
@@ -31,7 +31,7 @@ $CLICKHOUSE_CLIENT -q "SELECT name,
                               splitByChar('/', metadata_path)[-2] as uuid_path, ((splitByChar('/', metadata_path)[-3] as metadata) = substr(uuid_path, 1, 3)) OR metadata='metadata'
                               FROM system.databases WHERE name LIKE '${CLICKHOUSE_DATABASE}_%'" | sed "s/$uuid_db_1/00001114-1000-4000-8000-000000000001/g" | sed "s/$uuid_db_2/00001114-1000-4000-8000-000000000002/g"
 
-$CLICKHOUSE_CLIENT -nm -q "
+$CLICKHOUSE_CLIENT -m -q "
 CREATE TABLE ${DATABASE_1}.mt_tmp (n UInt64) ENGINE=MergeTree() ORDER BY tuple();
 INSERT INTO ${DATABASE_1}.mt_tmp SELECT * FROM numbers(100);
 CREATE TABLE ${DATABASE_3}.mt (n UInt64) ENGINE=MergeTree() ORDER BY tuple() PARTITION BY (n % 5);
@@ -65,7 +65,7 @@ while [[ $($CLICKHOUSE_CLIENT -q "SELECT count() FROM system.processes WHERE que
     sleep 0.1
 done
 
-$CLICKHOUSE_CLIENT -nm -q "
+$CLICKHOUSE_CLIENT -m -q "
 RENAME TABLE ${DATABASE_1}.mt TO ${DATABASE_1}.mt_tmp;
 RENAME TABLE ${DATABASE_1}.mt_tmp TO ${DATABASE_2}.mt_tmp;
 EXCHANGE TABLES ${DATABASE_2}.mt AND ${DATABASE_2}.mt_tmp;
@@ -79,7 +79,7 @@ uuid_mt1=$($CLICKHOUSE_CLIENT -q "SELECT uuid FROM system.tables WHERE database=
 $CLICKHOUSE_CLIENT --show_table_uuid_in_table_create_query_if_not_nil=1 -q "SHOW CREATE TABLE ${DATABASE_1}.mt" | sed "s/$uuid_mt1/00001114-0000-4000-8000-000000000001/g"
 $CLICKHOUSE_CLIENT --show_table_uuid_in_table_create_query_if_not_nil=1 -q "SHOW CREATE TABLE ${DATABASE_2}.mt" | sed "s/$explicit_uuid/00001114-0000-4000-8000-000000000002/g"
 
-$CLICKHOUSE_CLIENT -nm -q "
+$CLICKHOUSE_CLIENT -m -q "
 DROP TABLE ${DATABASE_1}.mt SETTINGS database_atomic_wait_for_drop_and_detach_synchronously=0;
 CREATE TABLE ${DATABASE_1}.mt (s String) ENGINE=Log();
 INSERT INTO ${DATABASE_1}.mt SELECT 's' || toString(number) FROM numbers(5);
diff --git a/tests/queries/0_stateless/01119_session_log.sh b/tests/queries/0_stateless/01119_session_log.sh
index 2d17b545276..61bb7cf3ea8 100755
--- a/tests/queries/0_stateless/01119_session_log.sh
+++ b/tests/queries/0_stateless/01119_session_log.sh
@@ -14,7 +14,7 @@ and interface in ('HTTP', 'TCP', 'TCP_Interserver')
 and (user != 'default' or (a=1 and b=1)) -- FIXME: we should not write uninitialized address and port (but we do sometimes)
 and event_time >= now() - interval 5 minute"
 
-$CLICKHOUSE_CLIENT -nm -q "
+$CLICKHOUSE_CLIENT -m -q "
 select * from remote('127.0.0.2', system, one, 'default', '');
 select * from remote('127.0.0.2', system, one, 'default', 'wrong password'); -- { serverError AUTHENTICATION_FAILED }
 select * from remote('127.0.0.2', system, one, 'nonexistsnt_user_1119', ''); -- { serverError AUTHENTICATION_FAILED }
diff --git a/tests/queries/0_stateless/01148_zookeeper_path_macros_unfolding.sql b/tests/queries/0_stateless/01148_zookeeper_path_macros_unfolding.sql
index a585ef1c324..c689542e4c3 100644
--- a/tests/queries/0_stateless/01148_zookeeper_path_macros_unfolding.sql
+++ b/tests/queries/0_stateless/01148_zookeeper_path_macros_unfolding.sql
@@ -7,6 +7,8 @@ DROP TABLE IF EXISTS rmt1;
 DROP TABLE IF EXISTS rmt2;
 DROP TABLE IF EXISTS rmt3;
 
+SET database_replicated_allow_replicated_engine_arguments=1;
+
 CREATE TABLE rmt (n UInt64, s String) ENGINE = ReplicatedMergeTree('/clickhouse/test_01148/{shard}/{database}/{table}', '{replica}') ORDER BY n;
 SHOW CREATE TABLE rmt;
 RENAME TABLE rmt TO rmt1;
diff --git a/tests/queries/0_stateless/01153_attach_mv_uuid.sql b/tests/queries/0_stateless/01153_attach_mv_uuid.sql
index 00cce8a1de4..0ef16449096 100644
--- a/tests/queries/0_stateless/01153_attach_mv_uuid.sql
+++ b/tests/queries/0_stateless/01153_attach_mv_uuid.sql
@@ -14,6 +14,8 @@ INSERT INTO src VALUES (3), (4);
 SELECT * FROM mv ORDER BY n;
 DROP TABLE mv SYNC;
 
+SET database_replicated_allow_explicit_uuid=3;
+
 SET show_table_uuid_in_table_create_query_if_not_nil=1;
 CREATE TABLE ".inner_id.e15f3ab5-6cae-4df3-b879-f40deafd82c2" (n Int32, n2 Int64) ENGINE = MergeTree PARTITION BY n % 10 ORDER BY n;
 ATTACH MATERIALIZED VIEW mv UUID 'e15f3ab5-6cae-4df3-b879-f40deafd82c2' (n Int32, n2 Int64) ENGINE = MergeTree PARTITION BY n % 10 ORDER BY n AS SELECT n, n * n AS n2 FROM src;
diff --git a/tests/queries/0_stateless/01213_alter_rename_column_zookeeper_long.sh b/tests/queries/0_stateless/01213_alter_rename_column_zookeeper_long.sh
index a255c1db30e..8afb0c18462 100755
--- a/tests/queries/0_stateless/01213_alter_rename_column_zookeeper_long.sh
+++ b/tests/queries/0_stateless/01213_alter_rename_column_zookeeper_long.sh
@@ -7,7 +7,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 
 $CLICKHOUSE_CLIENT --query "DROP TABLE IF EXISTS table_for_rename_replicated"
 
-$CLICKHOUSE_CLIENT -n --query "
+$CLICKHOUSE_CLIENT --query "
 CREATE TABLE table_for_rename_replicated
 (
   date Date,
diff --git a/tests/queries/0_stateless/01238_http_memory_tracking.sh b/tests/queries/0_stateless/01238_http_memory_tracking.sh
index ce1310cf302..f88c8fb47c6 100755
--- a/tests/queries/0_stateless/01238_http_memory_tracking.sh
+++ b/tests/queries/0_stateless/01238_http_memory_tracking.sh
@@ -14,7 +14,7 @@ ${CLICKHOUSE_CLIENT} --format Null -q "CREATE USER $MISTER_USER"
 
 # This is needed to keep at least one running query for user for the time of test.
 # (1k http queries takes ~1 second, let's run for 5x more to avoid flaps)
-${CLICKHOUSE_CLIENT} --user ${MISTER_USER} --function_sleep_max_microseconds_per_block 5000000 --format Null -n <<<'SELECT sleepEachRow(1) FROM numbers(5)' &
+${CLICKHOUSE_CLIENT} --user ${MISTER_USER} --function_sleep_max_microseconds_per_block 5000000 --format Null <<<'SELECT sleepEachRow(1) FROM numbers(5)' &
 
 # ignore "yes: standard output: Broken pipe"
 yes 'SELECT 1' 2>/dev/null | {
diff --git a/tests/queries/0_stateless/01280_ssd_complex_key_dictionary.sh b/tests/queries/0_stateless/01280_ssd_complex_key_dictionary.sh
index 60a65b9a253..c1ec812875c 100755
--- a/tests/queries/0_stateless/01280_ssd_complex_key_dictionary.sh
+++ b/tests/queries/0_stateless/01280_ssd_complex_key_dictionary.sh
@@ -5,7 +5,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-$CLICKHOUSE_CLIENT -n --query="
+$CLICKHOUSE_CLIENT --query="
     DROP DATABASE IF EXISTS 01280_db;
     CREATE DATABASE 01280_db;
     DROP TABLE IF EXISTS 01280_db.table_for_dict;
@@ -39,9 +39,9 @@ $CLICKHOUSE_CLIENT -n --query="
     LIFETIME(MIN 1000 MAX 2000)
     LAYOUT(COMPLEX_KEY_SSD_CACHE(FILE_SIZE 8192 PATH '$USER_FILES_PATH/0d'));"
 
-$CLICKHOUSE_CLIENT -nq "SELECT dictHas('01280_db.ssd_dict', 'a', tuple('1')); -- { serverError 43 }"
+$CLICKHOUSE_CLIENT -q "SELECT dictHas('01280_db.ssd_dict', 'a', tuple('1')); -- { serverError 43 }"
 
-$CLICKHOUSE_CLIENT -n --query="
+$CLICKHOUSE_CLIENT --query="
     SELECT 'TEST_SMALL';
     SELECT 'VALUE FROM RAM BUFFER';
     SELECT dictGetUInt64('01280_db.ssd_dict', 'a', tuple('1', toInt32(3)));
@@ -63,9 +63,9 @@ $CLICKHOUSE_CLIENT -n --query="
     SELECT dictGetInt32('01280_db.ssd_dict', 'b', tuple('10', toInt32(-20)));
     SELECT dictGetString('01280_db.ssd_dict', 'c', tuple('10', toInt32(-20)));"
 
-$CLICKHOUSE_CLIENT -nq "SELECT dictGetUInt64('01280_db.ssd_dict', 'a', tuple(toInt32(3))); -- { serverError 53 }"
+$CLICKHOUSE_CLIENT -q "SELECT dictGetUInt64('01280_db.ssd_dict', 'a', tuple(toInt32(3))); -- { serverError 53 }"
 
-$CLICKHOUSE_CLIENT -n --query="DROP DICTIONARY 01280_db.ssd_dict;
+$CLICKHOUSE_CLIENT --query="DROP DICTIONARY 01280_db.ssd_dict;
     DROP TABLE IF EXISTS 01280_db.keys_table;
     CREATE TABLE 01280_db.keys_table
     (
@@ -122,4 +122,4 @@ $CLICKHOUSE_CLIENT -n --query="DROP DICTIONARY 01280_db.ssd_dict;
     DROP DICTIONARY IF EXISTS database_for_dict.ssd_dict;
     DROP TABLE IF EXISTS database_for_dict.keys_table;"
 
-$CLICKHOUSE_CLIENT -n --query="DROP DATABASE IF EXISTS 01280_db;"
+$CLICKHOUSE_CLIENT --query="DROP DATABASE IF EXISTS 01280_db;"
diff --git a/tests/queries/0_stateless/01294_lazy_database_concurrent_recreate_reattach_and_show_tables_long.sh b/tests/queries/0_stateless/01294_lazy_database_concurrent_recreate_reattach_and_show_tables_long.sh
index 21f46a34514..00619f13173 100755
--- a/tests/queries/0_stateless/01294_lazy_database_concurrent_recreate_reattach_and_show_tables_long.sh
+++ b/tests/queries/0_stateless/01294_lazy_database_concurrent_recreate_reattach_and_show_tables_long.sh
@@ -85,7 +85,7 @@ export -f recreate_lazy_func4;
 export -f test_func;
 
 
-${CLICKHOUSE_CLIENT} -n -q "
+${CLICKHOUSE_CLIENT} -q "
     DROP DATABASE IF EXISTS $CURR_DATABASE;
     CREATE DATABASE $CURR_DATABASE ENGINE = Lazy(1);
 "
diff --git a/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh b/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh
index 1d2d4516b9c..6ff6644f11e 100755
--- a/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh
+++ b/tests/queries/0_stateless/01305_replica_create_drop_zookeeper.sh
@@ -10,7 +10,7 @@ set -e
 function thread()
 {
     while true; do
-        $CLICKHOUSE_CLIENT -n -q "DROP TABLE IF EXISTS test_table_$1 SYNC;
+        $CLICKHOUSE_CLIENT -q "DROP TABLE IF EXISTS test_table_$1 SYNC;
             CREATE TABLE test_table_$1 (a UInt8) ENGINE = ReplicatedMergeTree('/clickhouse/tables/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/alter_table', 'r_$1') ORDER BY tuple();" 2>&1 |
                 grep -vP '(^$)|(^Received exception from server)|(^\d+\. )|because the last replica of the table was dropped right now|is already started to be removing by another replica right now| were removed by another replica|Removing leftovers from table|Another replica was suddenly created|was created by another server at the same moment|was suddenly removed|some other replicas were created at the same time|^\(query: '
     done
diff --git a/tests/queries/0_stateless/01320_create_sync_race_condition_zookeeper.sh b/tests/queries/0_stateless/01320_create_sync_race_condition_zookeeper.sh
index 1d5f5d54853..cc96a37a0ce 100755
--- a/tests/queries/0_stateless/01320_create_sync_race_condition_zookeeper.sh
+++ b/tests/queries/0_stateless/01320_create_sync_race_condition_zookeeper.sh
@@ -17,7 +17,7 @@ function thread1()
 {
     local TIMELIMIT=$((SECONDS+$1))
     while [ $SECONDS -lt "$TIMELIMIT" ]; do
-        $CLICKHOUSE_CLIENT -n --query "CREATE TABLE test_01320.r (x UInt64) ENGINE = ReplicatedMergeTree('/test/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/table', 'r') ORDER BY x; DROP TABLE test_01320.r;"
+        $CLICKHOUSE_CLIENT --query "CREATE TABLE test_01320.r (x UInt64) ENGINE = ReplicatedMergeTree('/test/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/table', 'r') ORDER BY x; DROP TABLE test_01320.r;"
     done
 }
 
diff --git a/tests/queries/0_stateless/01395_limit_more_cases.sh b/tests/queries/0_stateless/01395_limit_more_cases.sh
index 9709bd74f26..6be8a91f0c7 100755
--- a/tests/queries/0_stateless/01395_limit_more_cases.sh
+++ b/tests/queries/0_stateless/01395_limit_more_cases.sh
@@ -20,4 +20,4 @@ for OFFSET in {0..15}; do
             FROM (SELECT * FROM numbers($SIZE) LIMIT $OFFSET, $LIMIT);
         "
     done
-done | $CLICKHOUSE_CLIENT -n --max_block_size 5
+done | $CLICKHOUSE_CLIENT --max_block_size 5
diff --git a/tests/queries/0_stateless/01395_limit_more_cases_random.sh b/tests/queries/0_stateless/01395_limit_more_cases_random.sh
index c2f6b060aab..bb942906e63 100755
--- a/tests/queries/0_stateless/01395_limit_more_cases_random.sh
+++ b/tests/queries/0_stateless/01395_limit_more_cases_random.sh
@@ -19,4 +19,4 @@ for _ in $(seq $ITERATIONS); do
                 throwIf((c != 0 OR first != 0 OR last != 0) AND (c != last - first + 1))
             FROM (SELECT * FROM numbers($SIZE) LIMIT $OFFSET, $LIMIT);
         "
-done | $CLICKHOUSE_CLIENT -n --max_block_size $(($RANDOM % 20 + 1)) | uniq
+done | $CLICKHOUSE_CLIENT --max_block_size $(($RANDOM % 20 + 1)) | uniq
diff --git a/tests/queries/0_stateless/01412_cache_dictionary_race.sh b/tests/queries/0_stateless/01412_cache_dictionary_race.sh
index 36295ca01ea..b0c73cf742f 100755
--- a/tests/queries/0_stateless/01412_cache_dictionary_race.sh
+++ b/tests/queries/0_stateless/01412_cache_dictionary_race.sh
@@ -10,7 +10,7 @@ $CLICKHOUSE_CLIENT --query "DROP DATABASE IF EXISTS ordinary_db"
 
 $CLICKHOUSE_CLIENT --query "CREATE DATABASE ordinary_db"
 
-$CLICKHOUSE_CLIENT -n -q "
+$CLICKHOUSE_CLIENT -q "
 
 CREATE DICTIONARY ordinary_db.dict1
 (
@@ -35,7 +35,7 @@ function dict_get_thread()
 function drop_create_table_thread()
 {
     while true; do
-        $CLICKHOUSE_CLIENT -n --query "CREATE TABLE ordinary_db.table_for_dict_real (
+        $CLICKHOUSE_CLIENT --query "CREATE TABLE ordinary_db.table_for_dict_real (
             key_column UInt64,
             second_column UInt8,
             third_column String
diff --git a/tests/queries/0_stateless/01454_storagememory_data_race_challenge.sh b/tests/queries/0_stateless/01454_storagememory_data_race_challenge.sh
index ec9c5134059..950afea9086 100755
--- a/tests/queries/0_stateless/01454_storagememory_data_race_challenge.sh
+++ b/tests/queries/0_stateless/01454_storagememory_data_race_challenge.sh
@@ -23,7 +23,7 @@ function f {
 function g {
   local TIMELIMIT=$((SECONDS+$1))
   for _ in $(seq 1 100); do
-    $CLICKHOUSE_CLIENT -n -q "
+    $CLICKHOUSE_CLIENT -q "
         INSERT INTO mem SELECT number FROM numbers(1000000);
         INSERT INTO mem SELECT number FROM numbers(1000000);
         INSERT INTO mem SELECT number FROM numbers(1000000);
diff --git a/tests/queries/0_stateless/01509_check_many_parallel_quorum_inserts_long.sh b/tests/queries/0_stateless/01509_check_many_parallel_quorum_inserts_long.sh
index 22cd6fb8127..0e6ab287146 100755
--- a/tests/queries/0_stateless/01509_check_many_parallel_quorum_inserts_long.sh
+++ b/tests/queries/0_stateless/01509_check_many_parallel_quorum_inserts_long.sh
@@ -11,7 +11,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 NUM_REPLICAS=6
 
 for i in $(seq 1 $NUM_REPLICAS); do
-    $CLICKHOUSE_CLIENT -n -q "
+    $CLICKHOUSE_CLIENT -q "
         DROP TABLE IF EXISTS r$i SYNC;
         CREATE TABLE r$i (x UInt64) ENGINE = ReplicatedMergeTree('/clickhouse/tables/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/parallel_quorum_many', 'r$i') ORDER BY x;
     "
@@ -39,12 +39,12 @@ done
 wait
 
 for i in $(seq 1 $NUM_REPLICAS); do
-    $CLICKHOUSE_CLIENT -n -q "
+    $CLICKHOUSE_CLIENT -q "
         SYSTEM SYNC REPLICA r$i;
         SELECT count(), min(x), max(x), sum(x) FROM r$i;
     "
 done
 
 for i in $(seq 1 $NUM_REPLICAS); do
-    $CLICKHOUSE_CLIENT -n -q "DROP TABLE IF EXISTS r$i SYNC;"
+    $CLICKHOUSE_CLIENT -q "DROP TABLE IF EXISTS r$i SYNC;"
 done
diff --git a/tests/queries/0_stateless/01509_check_parallel_quorum_inserts_long.sh b/tests/queries/0_stateless/01509_check_parallel_quorum_inserts_long.sh
index 1589f17c752..3f4210f9bb0 100755
--- a/tests/queries/0_stateless/01509_check_parallel_quorum_inserts_long.sh
+++ b/tests/queries/0_stateless/01509_check_parallel_quorum_inserts_long.sh
@@ -12,13 +12,13 @@ NUM_REPLICAS=2
 NUM_INSERTS=5
 
 for i in $(seq 1 $NUM_REPLICAS); do
-    $CLICKHOUSE_CLIENT -n -q "
+    $CLICKHOUSE_CLIENT -q "
         DROP TABLE IF EXISTS r$i;
         CREATE TABLE r$i (x UInt64) ENGINE = ReplicatedMergeTree('/clickhouse/tables/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/parallel_quorum', 'r$i') ORDER BY x;
     "
 done
 
-$CLICKHOUSE_CLIENT -n -q "SYSTEM STOP REPLICATION QUEUES r2;"
+$CLICKHOUSE_CLIENT -q "SYSTEM STOP REPLICATION QUEUES r2;"
 
 function thread {
     $CLICKHOUSE_CLIENT --insert_quorum 2 --insert_quorum_parallel 1 --query "INSERT INTO r1 SELECT $1"
@@ -28,12 +28,12 @@ for i in $(seq 1 $NUM_INSERTS); do
     thread $i &
 done
 
-$CLICKHOUSE_CLIENT -n -q "SYSTEM START REPLICATION QUEUES r2;"
+$CLICKHOUSE_CLIENT -q "SYSTEM START REPLICATION QUEUES r2;"
 
 wait
 
 for i in $(seq 1 $NUM_REPLICAS); do
-    $CLICKHOUSE_CLIENT -n -q "
+    $CLICKHOUSE_CLIENT -q "
         SELECT count(), min(x), max(x), sum(x) FROM r$i;
         DROP TABLE IF EXISTS r$i;
 "
diff --git a/tests/queries/0_stateless/01563_distributed_query_finish.sh b/tests/queries/0_stateless/01563_distributed_query_finish.sh
index e3c5928f108..50d194b1f15 100755
--- a/tests/queries/0_stateless/01563_distributed_query_finish.sh
+++ b/tests/queries/0_stateless/01563_distributed_query_finish.sh
@@ -9,7 +9,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
-$CLICKHOUSE_CLIENT -nm <<EOL
+$CLICKHOUSE_CLIENT -m <<EOL
 drop table if exists dist_01247;
 drop table if exists data_01247;
 
@@ -29,7 +29,7 @@ for ((i = 0; i < 100; ++i)); do
         "--optimize_distributed_group_by_sharding_key=1"
         "--prefer_localhost_replica=0"
     )
-    $CLICKHOUSE_CLIENT "${opts[@]}" --format CSV -nm -q "select count(), * from dist_01247 group by number order by number limit 1 format Null"
+    $CLICKHOUSE_CLIENT "${opts[@]}" --format CSV -m -q "select count(), * from dist_01247 group by number order by number limit 1 format Null"
 
     # expect zero new network errors
     network_errors_after=$($CLICKHOUSE_CLIENT -q "SELECT value FROM system.errors WHERE name = 'NETWORK_ERROR'")
diff --git a/tests/queries/0_stateless/01603_read_with_backoff_bug.sql b/tests/queries/0_stateless/01603_read_with_backoff_bug.sql
index 8a6fa9b7845..212a18b5799 100644
--- a/tests/queries/0_stateless/01603_read_with_backoff_bug.sql
+++ b/tests/queries/0_stateless/01603_read_with_backoff_bug.sql
@@ -11,7 +11,7 @@ create table t (x UInt64, s String) engine = MergeTree order by x SETTINGS index
 INSERT INTO t SELECT
     number,
     if(number < (8129 * 1024), arrayStringConcat(arrayMap(x -> toString(x), range(number % 128)), ' '), '')
-FROM numbers_mt((8129 * 1024) * 3) settings max_insert_threads=8, max_rows_to_read=0;
+FROM numbers_mt((8129 * 1024) * 3) settings max_insert_threads=8, max_rows_to_read=0, max_memory_usage='10Gi';
 
 -- optimize table t final;
 
diff --git a/tests/queries/0_stateless/01732_race_condition_storage_join_long.sh b/tests/queries/0_stateless/01732_race_condition_storage_join_long.sh
index 5fc41890a18..dbfb3518a61 100755
--- a/tests/queries/0_stateless/01732_race_condition_storage_join_long.sh
+++ b/tests/queries/0_stateless/01732_race_condition_storage_join_long.sh
@@ -11,14 +11,14 @@ set -o pipefail
 echo "
 	DROP TABLE IF EXISTS storage_join_race;
 	CREATE TABLE storage_join_race (x UInt64, y UInt64) Engine = Join(ALL, FULL, x);
-" | $CLICKHOUSE_CLIENT -n
+" | $CLICKHOUSE_CLIENT
 
 function read_thread_big()
 {
     while true; do
         echo "
             SELECT * FROM ( SELECT number AS x FROM numbers(100000) ) AS t1 ALL FULL JOIN storage_join_race USING (x) FORMAT Null;
-        " | $CLICKHOUSE_CLIENT -n
+        " | $CLICKHOUSE_CLIENT
     done
 }
 
@@ -27,7 +27,7 @@ function read_thread_small()
     while true; do
         echo "
             SELECT * FROM ( SELECT number AS x FROM numbers(10) ) AS t1 ALL FULL JOIN storage_join_race USING (x) FORMAT Null;
-        " | $CLICKHOUSE_CLIENT -n
+        " | $CLICKHOUSE_CLIENT
     done
 }
 
@@ -36,7 +36,7 @@ function read_thread_select()
     while true; do
         echo "
             SELECT * FROM storage_join_race FORMAT Null;
-        " | $CLICKHOUSE_CLIENT -n
+        " | $CLICKHOUSE_CLIENT
     done
 }
 
@@ -56,7 +56,7 @@ echo "
     INSERT INTO storage_join_race
         SELECT number AS x, sleepEachRow(0.1) + number AS y FROM numbers ($TIMEOUT * 10)
         SETTINGS function_sleep_max_microseconds_per_block = 100000000, max_block_size = 10;
-" | $CLICKHOUSE_CLIENT -n
+" | $CLICKHOUSE_CLIENT
 
 wait
 
diff --git a/tests/queries/0_stateless/01780_column_sparse_materialize.reference b/tests/queries/0_stateless/01780_column_sparse_materialize.reference
new file mode 100644
index 00000000000..660cfabff33
--- /dev/null
+++ b/tests/queries/0_stateless/01780_column_sparse_materialize.reference
@@ -0,0 +1,50 @@
+-- { echoOn }
+
+SELECT dumpColumnStructure(id) FROM sparse_t;
+UInt64, Sparse(size = 2, UInt64(size = 2), UInt64(size = 1))
+UInt64, Sparse(size = 2, UInt64(size = 2), UInt64(size = 1))
+SELECT dumpColumnStructure(materialize(id)) FROM sparse_t;
+UInt64, UInt64(size = 2)
+UInt64, UInt64(size = 2)
+SELECT dumpColumnStructure(u) FROM sparse_t;
+UInt64, Sparse(size = 2, UInt64(size = 1), UInt64(size = 0))
+UInt64, Sparse(size = 2, UInt64(size = 1), UInt64(size = 0))
+SELECT dumpColumnStructure(materialize(u)) FROM sparse_t;
+UInt64, UInt64(size = 2)
+UInt64, UInt64(size = 2)
+SELECT dumpColumnStructure(s) FROM sparse_t;
+String, Sparse(size = 2, String(size = 2), UInt64(size = 1))
+String, Sparse(size = 2, String(size = 2), UInt64(size = 1))
+SELECT dumpColumnStructure(materialize(s)) FROM sparse_t;
+String, String(size = 2)
+String, String(size = 2)
+SELECT dumpColumnStructure(arr1) FROM sparse_t;
+Array(String), Array(size = 2, UInt64(size = 2), String(size = 1))
+Array(String), Array(size = 2, UInt64(size = 2), String(size = 1))
+SELECT dumpColumnStructure(materialize(arr1)) FROM sparse_t;
+Array(String), Array(size = 2, UInt64(size = 2), String(size = 1))
+Array(String), Array(size = 2, UInt64(size = 2), String(size = 1))
+SELECT dumpColumnStructure(arr2) FROM sparse_t;
+Array(UInt64), Array(size = 2, UInt64(size = 2), UInt64(size = 1))
+Array(UInt64), Array(size = 2, UInt64(size = 2), UInt64(size = 1))
+SELECT dumpColumnStructure(materialize(arr2)) FROM sparse_t;
+Array(UInt64), Array(size = 2, UInt64(size = 2), UInt64(size = 1))
+Array(UInt64), Array(size = 2, UInt64(size = 2), UInt64(size = 1))
+SELECT dumpColumnStructure(t) FROM sparse_t;
+Tuple(a UInt64, s String), Tuple(size = 2, Sparse(size = 2, UInt64(size = 1), UInt64(size = 0)), Sparse(size = 2, String(size = 1), UInt64(size = 0)))
+Tuple(a UInt64, s String), Tuple(size = 2, Sparse(size = 2, UInt64(size = 1), UInt64(size = 0)), Sparse(size = 2, String(size = 1), UInt64(size = 0)))
+SELECT dumpColumnStructure(materialize(t)) FROM sparse_t;
+Tuple(a UInt64, s String), Tuple(size = 2, UInt64(size = 2), String(size = 2))
+Tuple(a UInt64, s String), Tuple(size = 2, UInt64(size = 2), String(size = 2))
+SELECT dumpColumnStructure(t.a) FROM sparse_t;
+UInt64, Sparse(size = 2, UInt64(size = 1), UInt64(size = 0))
+UInt64, Sparse(size = 2, UInt64(size = 1), UInt64(size = 0))
+SELECT dumpColumnStructure(materialize(t.a)) FROM sparse_t;
+UInt64, UInt64(size = 2)
+UInt64, UInt64(size = 2)
+SELECT dumpColumnStructure(t.s) FROM sparse_t;
+String, Sparse(size = 2, String(size = 1), UInt64(size = 0))
+String, Sparse(size = 2, String(size = 1), UInt64(size = 0))
+SELECT dumpColumnStructure(materialize(t.s)) FROM sparse_t;
+String, String(size = 2)
+String, String(size = 2)
diff --git a/tests/queries/0_stateless/01780_column_sparse_materialize.sql b/tests/queries/0_stateless/01780_column_sparse_materialize.sql
new file mode 100644
index 00000000000..a53ea140f0f
--- /dev/null
+++ b/tests/queries/0_stateless/01780_column_sparse_materialize.sql
@@ -0,0 +1,52 @@
+DROP TABLE IF EXISTS sparse_t;
+
+CREATE TABLE sparse_t (
+    id UInt64,
+    u UInt64,
+    s String,
+    arr1 Array(String),
+    arr2 Array(UInt64),
+    t Tuple(a UInt64, s String))
+ENGINE = MergeTree ORDER BY tuple()
+SETTINGS ratio_of_defaults_for_sparse_serialization = 0.1;
+
+INSERT INTO sparse_t SELECT
+    number,
+    if (number % 2 = 0, number, 0),
+    if (number % 2 = 0, toString(number), ''),
+    if (number % 2 = 0, [''], []),
+    if (number % 2 = 0, [0], []),
+    (if (number % 2 = 0, number, 0), '')
+FROM numbers(2);
+
+-- { echoOn }
+
+SELECT dumpColumnStructure(id) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(id)) FROM sparse_t;
+
+SELECT dumpColumnStructure(u) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(u)) FROM sparse_t;
+
+SELECT dumpColumnStructure(s) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(s)) FROM sparse_t;
+
+SELECT dumpColumnStructure(arr1) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(arr1)) FROM sparse_t;
+
+SELECT dumpColumnStructure(arr2) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(arr2)) FROM sparse_t;
+
+SELECT dumpColumnStructure(t) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(t)) FROM sparse_t;
+
+SELECT dumpColumnStructure(t.a) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(t.a)) FROM sparse_t;
+
+SELECT dumpColumnStructure(t.s) FROM sparse_t;
+SELECT dumpColumnStructure(materialize(t.s)) FROM sparse_t;
+
+-- { echoOff }
+
+
+DROP TABLE IF EXISTS sparse_t
+;
diff --git a/tests/queries/0_stateless/01802_test_postgresql_protocol_with_row_policy.sh b/tests/queries/0_stateless/01802_test_postgresql_protocol_with_row_policy.sh
index 4713b15bb5b..4cf487274d4 100755
--- a/tests/queries/0_stateless/01802_test_postgresql_protocol_with_row_policy.sh
+++ b/tests/queries/0_stateless/01802_test_postgresql_protocol_with_row_policy.sh
@@ -17,7 +17,7 @@ INSERT INTO db01802.postgresql SELECT number FROM numbers(10);
 
 SELECT 'before row policy';
 SELECT * FROM db01802.postgresql;
-" | $CLICKHOUSE_CLIENT -n
+" | $CLICKHOUSE_CLIENT
 
 
 echo "
@@ -28,7 +28,7 @@ CREATE ROW POLICY IF NOT EXISTS test_policy ON db01802.postgresql FOR SELECT USI
 
 SELECT '';
 SELECT 'after row policy with no password';
-" | $CLICKHOUSE_CLIENT -n
+" | $CLICKHOUSE_CLIENT
 
 psql --host localhost --port ${CLICKHOUSE_PORT_POSTGRESQL} db01802 --user postgresql_user -c "SELECT * FROM postgresql;"
 
@@ -40,7 +40,7 @@ GRANT SELECT(val) ON db01802.postgresql TO postgresql_user;
 CREATE ROW POLICY IF NOT EXISTS test_policy ON db01802.postgresql FOR SELECT USING val = 2 TO postgresql_user;
 
 SELECT 'after row policy with plaintext_password';
-" | $CLICKHOUSE_CLIENT -n
+" | $CLICKHOUSE_CLIENT
 
 psql "postgresql://postgresql_user:qwerty@localhost:${CLICKHOUSE_PORT_POSTGRESQL}/db01802" -c "SELECT * FROM postgresql;"
 
diff --git a/tests/queries/0_stateless/01810_max_part_removal_threads_long.sh b/tests/queries/0_stateless/01810_max_part_removal_threads_long.sh
index c38fc505fa8..4196598e461 100755
--- a/tests/queries/0_stateless/01810_max_part_removal_threads_long.sh
+++ b/tests/queries/0_stateless/01810_max_part_removal_threads_long.sh
@@ -17,10 +17,10 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # The number of threads removing data parts should be between 1 and 129.
 # Because max_parts_cleaning_thread_pool_size is 128 by default
 
-$CLICKHOUSE_CLIENT --allow_deprecated_database_ordinary=1 -nm -q "create database ordinary_$CLICKHOUSE_DATABASE engine=Ordinary"
+$CLICKHOUSE_CLIENT --allow_deprecated_database_ordinary=1 -m -q "create database ordinary_$CLICKHOUSE_DATABASE engine=Ordinary"
 
 # MergeTree
-$CLICKHOUSE_CLIENT -nm -q """
+$CLICKHOUSE_CLIENT -m -q """
     use ordinary_$CLICKHOUSE_DATABASE;
     drop table if exists data_01810;
 
@@ -47,7 +47,7 @@ $CLICKHOUSE_CLIENT -nm -q """
 """
 
 # ReplicatedMergeTree
-$CLICKHOUSE_CLIENT -nm -q """
+$CLICKHOUSE_CLIENT -m -q """
     use ordinary_$CLICKHOUSE_DATABASE;
     drop table if exists rep_data_01810;
 
@@ -76,4 +76,4 @@ $CLICKHOUSE_CLIENT -nm -q """
     format Null;
 """
 
-$CLICKHOUSE_CLIENT -nm -q "drop database ordinary_$CLICKHOUSE_DATABASE"
+$CLICKHOUSE_CLIENT -m -q "drop database ordinary_$CLICKHOUSE_DATABASE"
diff --git a/tests/queries/0_stateless/01889_postgresql_protocol_null_fields.sh b/tests/queries/0_stateless/01889_postgresql_protocol_null_fields.sh
index e9467d77451..c0f200ed20d 100755
--- a/tests/queries/0_stateless/01889_postgresql_protocol_null_fields.sh
+++ b/tests/queries/0_stateless/01889_postgresql_protocol_null_fields.sh
@@ -9,6 +9,6 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 echo "
 DROP USER IF EXISTS postgresql_user;
 CREATE USER postgresql_user HOST IP '127.0.0.1' IDENTIFIED WITH no_password;
-" | $CLICKHOUSE_CLIENT -n
+" | $CLICKHOUSE_CLIENT
 
 psql --host localhost --port ${CLICKHOUSE_PORT_POSTGRESQL} ${CLICKHOUSE_DATABASE} --user postgresql_user -c "SELECT NULL;"
diff --git a/tests/queries/0_stateless/01900_kill_mutation_parallel_long.sh b/tests/queries/0_stateless/01900_kill_mutation_parallel_long.sh
index 30dff6db0c4..81a674be3c4 100755
--- a/tests/queries/0_stateless/01900_kill_mutation_parallel_long.sh
+++ b/tests/queries/0_stateless/01900_kill_mutation_parallel_long.sh
@@ -12,7 +12,7 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CUR_DIR"/../shell_config.sh
 
-$CLICKHOUSE_CLIENT -nm -q "
+$CLICKHOUSE_CLIENT -m -q "
     drop table if exists data_01900_1;
     drop table if exists data_01900_2;
 
@@ -27,18 +27,18 @@ $CLICKHOUSE_CLIENT -nm -q "
 # so 100 mutations will be scheduled and killed later.
 for i in {1..100}; do
     echo "alter table data_01900_1 update s = 'foo_$i' where 1;"
-done | $CLICKHOUSE_CLIENT -nm
+done | $CLICKHOUSE_CLIENT -m
 
 # but these mutations should not be killed.
 (
     for i in {1..100}; do
         echo "alter table data_01900_2 update s = 'bar_$i' where 1;"
-    done | $CLICKHOUSE_CLIENT -nm --mutations_sync=1
+    done | $CLICKHOUSE_CLIENT -m --mutations_sync=1
 ) &
-$CLICKHOUSE_CLIENT --format Null -nm -q "kill mutation where table = 'data_01900_1' and database = '$CLICKHOUSE_DATABASE';"
+$CLICKHOUSE_CLIENT --format Null -m -q "kill mutation where table = 'data_01900_1' and database = '$CLICKHOUSE_DATABASE';"
 wait
 
-$CLICKHOUSE_CLIENT -nm -q "select * from data_01900_2"
+$CLICKHOUSE_CLIENT -m -q "select * from data_01900_2"
 
 $CLICKHOUSE_CLIENT -q "drop table data_01900_1"
 $CLICKHOUSE_CLIENT -q "drop table data_01900_2"
diff --git a/tests/queries/0_stateless/01921_concurrent_ttl_and_normal_merges_zookeeper_long.sh b/tests/queries/0_stateless/01921_concurrent_ttl_and_normal_merges_zookeeper_long.sh
index 0dfde4997a7..bc8b01fa9f5 100755
--- a/tests/queries/0_stateless/01921_concurrent_ttl_and_normal_merges_zookeeper_long.sh
+++ b/tests/queries/0_stateless/01921_concurrent_ttl_and_normal_merges_zookeeper_long.sh
@@ -18,7 +18,7 @@ done
 wait
 
 for i in $(seq 1 $NUM_REPLICAS); do
-    $CLICKHOUSE_CLIENT -n --query "CREATE TABLE ttl_table$i(
+    $CLICKHOUSE_CLIENT --query "CREATE TABLE ttl_table$i(
         key DateTime
     )
     ENGINE ReplicatedMergeTree('/test/01921_concurrent_ttl_and_normal_merges/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/ttl_table', '$i')
diff --git a/tests/queries/0_stateless/02404_memory_bound_merging.reference b/tests/queries/0_stateless/02404_memory_bound_merging.reference
index 0409c48f846..e02684c3b95 100644
--- a/tests/queries/0_stateless/02404_memory_bound_merging.reference
+++ b/tests/queries/0_stateless/02404_memory_bound_merging.reference
@@ -100,7 +100,7 @@ select a, count() from dist_t_different_dbs group by a, b order by a limit 5 off
 504	2000
 1000000
 -- { echoOn } --
-explain pipeline select a from pr_t group by a order by a limit 5 offset 500;
+explain pipeline select a from pr_t group by a order by a limit 5 offset 500 settings parallel_replicas_local_plan=0;
 (Expression)
 ExpressionTransform
   (Limit)
@@ -119,6 +119,31 @@ ExpressionTransform
                     Resize 1 → 4
                       GroupingAggregatedTransform 3 → 1
                         (ReadFromRemoteParallelReplicas)
+explain pipeline select a from pr_t group by a order by a limit 5 offset 500 settings allow_experimental_analyzer=1, parallel_replicas_local_plan=1;
+(Expression)
+ExpressionTransform
+  (Limit)
+  Limit
+    (Sorting)
+    MergingSortedTransform 4 → 1
+      MergeSortingTransform × 4
+        LimitsCheckingTransform × 4
+          PartialSortingTransform × 4
+            (Expression)
+            ExpressionTransform × 4
+              (MergingAggregated)
+              MergingAggregatedBucketTransform × 4
+                Resize 1 → 4
+                  FinishAggregatingInOrderTransform 3 → 1
+                    (Union)
+                      (Aggregating)
+                      FinalizeAggregatedTransform
+                        AggregatingInOrderTransform
+                          (Expression)
+                          ExpressionTransform
+                            (ReadFromMergeTree)
+                            MergeTreeSelect(pool: ReadPoolParallelReplicasInOrder, algorithm: InOrder) 0 → 1
+                      (ReadFromRemoteParallelReplicas)
 select a, count() from pr_t group by a order by a limit 5 offset 500;
 500	1000
 501	1000
diff --git a/tests/queries/0_stateless/02404_memory_bound_merging.sql b/tests/queries/0_stateless/02404_memory_bound_merging.sql
index a2de19dff8a..112640a2e2c 100644
--- a/tests/queries/0_stateless/02404_memory_bound_merging.sql
+++ b/tests/queries/0_stateless/02404_memory_bound_merging.sql
@@ -66,14 +66,14 @@ insert into pr_t select number % 1000, number % 1000 from numbers_mt(1e6);
 set allow_experimental_parallel_reading_from_replicas = 1;
 set parallel_replicas_for_non_replicated_merge_tree = 1;
 set max_parallel_replicas = 3;
-set use_hedged_requests = 0;
 set cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
 set distributed_aggregation_memory_efficient=1;
 
 select count() from pr_t;
 
 -- { echoOn } --
-explain pipeline select a from pr_t group by a order by a limit 5 offset 500;
+explain pipeline select a from pr_t group by a order by a limit 5 offset 500 settings parallel_replicas_local_plan=0;
+explain pipeline select a from pr_t group by a order by a limit 5 offset 500 settings allow_experimental_analyzer=1, parallel_replicas_local_plan=1;
 
 select a, count() from pr_t group by a order by a limit 5 offset 500;
 select a, count() from pr_t group by a, b order by a limit 5 offset 500;
diff --git a/tests/queries/0_stateless/02514_database_replicated_no_arguments_for_rmt.sh b/tests/queries/0_stateless/02514_database_replicated_no_arguments_for_rmt.sh
index c1aa24943c1..096a0fcc0b0 100755
--- a/tests/queries/0_stateless/02514_database_replicated_no_arguments_for_rmt.sh
+++ b/tests/queries/0_stateless/02514_database_replicated_no_arguments_for_rmt.sh
@@ -15,7 +15,7 @@ ${CLICKHOUSE_CLIENT} -q "GRANT CREATE TABLE ON ${CLICKHOUSE_DATABASE}_db.* TO us
 ${CLICKHOUSE_CLIENT} -q "GRANT TABLE ENGINE ON ReplicatedMergeTree TO user_${CLICKHOUSE_DATABASE}"
 ${CLICKHOUSE_CLIENT} -q "CREATE DATABASE ${CLICKHOUSE_DATABASE}_db engine = Replicated('/clickhouse/databases/${CLICKHOUSE_TEST_ZOOKEEPER_PREFIX}/${CLICKHOUSE_DATABASE}_db', '{shard}', '{replica}')"
 ${CLICKHOUSE_CLIENT} --distributed_ddl_output_mode=none --user "user_${CLICKHOUSE_DATABASE}" --query "CREATE TABLE ${CLICKHOUSE_DATABASE}_db.tab_rmt_ok (x UInt32) engine = ReplicatedMergeTree order by x;"
-${CLICKHOUSE_CLIENT} --distributed_ddl_output_mode=none --user "user_${CLICKHOUSE_DATABASE}" --query "CREATE TABLE ${CLICKHOUSE_DATABASE}_db.tab_rmt_fail (x UInt32) engine = ReplicatedMergeTree('/clickhouse/tables/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/root/{shard}', '{replica}') order by x; -- { serverError 80 }"
+${CLICKHOUSE_CLIENT} --distributed_ddl_output_mode=none --user "user_${CLICKHOUSE_DATABASE}" --query "CREATE TABLE ${CLICKHOUSE_DATABASE}_db.tab_rmt_fail (x UInt32) engine = ReplicatedMergeTree('/clickhouse/tables/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/root/{shard}', '{replica}') order by x; -- { serverError 36 }"
 ${CLICKHOUSE_CLIENT} --query "DROP DATABASE ${CLICKHOUSE_DATABASE}_db"
 ${CLICKHOUSE_CLIENT} -q "DROP USER user_${CLICKHOUSE_DATABASE}"
 
diff --git a/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.reference b/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.reference
index 125fa524c4a..682a43aab5f 100644
--- a/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.reference
+++ b/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.reference
@@ -31,7 +31,7 @@
 29	2j&S)ba?XG	QuQj	17163829389637435056
 3	UlI+1		14144472852965836438
 =============== QUERIES EXECUTED BY PARALLEL INNER QUERY ALONE ===============
-0	3	SELECT `__table1`.`key` AS `key`, `__table1`.`value1` AS `value1`, `__table1`.`value2` AS `value2`, toUInt64(min(`__table1`.`time`)) AS `start_ts` FROM `default`.`join_inner_table` AS `__table1` PREWHERE (`__table1`.`id` = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (`__table1`.`number` > _CAST(1610517366120, \'UInt64\')) GROUP BY `__table1`.`key`, `__table1`.`value1`, `__table1`.`value2` ORDER BY `__table1`.`key` ASC, `__table1`.`value1` ASC, `__table1`.`value2` ASC LIMIT _CAST(10, \'UInt64\')
+0	2	SELECT `__table1`.`key` AS `key`, `__table1`.`value1` AS `value1`, `__table1`.`value2` AS `value2`, toUInt64(min(`__table1`.`time`)) AS `start_ts` FROM `default`.`join_inner_table` AS `__table1` PREWHERE (`__table1`.`id` = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (`__table1`.`number` > _CAST(1610517366120, \'UInt64\')) GROUP BY `__table1`.`key`, `__table1`.`value1`, `__table1`.`value2` ORDER BY `__table1`.`key` ASC, `__table1`.`value1` ASC, `__table1`.`value2` ASC LIMIT _CAST(10, \'UInt64\')
 0	3	SELECT `key`, `value1`, `value2`, toUInt64(min(`time`)) AS `start_ts` FROM `default`.`join_inner_table` PREWHERE (`id` = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (`number` > toUInt64(\'1610517366120\')) GROUP BY `key`, `value1`, `value2` ORDER BY `key` ASC, `value1` ASC, `value2` ASC LIMIT 10
 1	1	-- Parallel inner query alone\nSELECT\n    key,\n    value1,\n    value2,\n    toUInt64(min(time)) AS start_ts\nFROM join_inner_table\nPREWHERE (id = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (number > toUInt64(\'1610517366120\'))\nGROUP BY key, value1, value2\nORDER BY key, value1, value2\nLIMIT 10\nSETTINGS allow_experimental_parallel_reading_from_replicas = 1, enable_analyzer=0;
 1	1	-- Parallel inner query alone\nSELECT\n    key,\n    value1,\n    value2,\n    toUInt64(min(time)) AS start_ts\nFROM join_inner_table\nPREWHERE (id = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (number > toUInt64(\'1610517366120\'))\nGROUP BY key, value1, value2\nORDER BY key, value1, value2\nLIMIT 10\nSETTINGS allow_experimental_parallel_reading_from_replicas = 1, enable_analyzer=1;
@@ -58,7 +58,7 @@ U	c	10
 UlI+1		10
 bX?}ix [	Ny]2 G	10
 t<iT	X48q:Z]t0	10
-0	3	SELECT `__table2`.`value1` AS `value1`, `__table2`.`value2` AS `value2`, count() AS `count` FROM `default`.`join_outer_table` AS `__table1` ALL INNER JOIN (SELECT `__table3`.`key` AS `key`, `__table3`.`value1` AS `value1`, `__table3`.`value2` AS `value2` FROM `default`.`join_inner_table` AS `__table3` PREWHERE (`__table3`.`id` = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (`__table3`.`number` > _CAST(1610517366120, \'UInt64\')) GROUP BY `__table3`.`key`, `__table3`.`value1`, `__table3`.`value2`) AS `__table2` USING (`key`) GROUP BY `__table1`.`key`, `__table2`.`value1`, `__table2`.`value2`
+0	2	SELECT `__table2`.`value1` AS `value1`, `__table2`.`value2` AS `value2`, count() AS `count` FROM `default`.`join_outer_table` AS `__table1` ALL INNER JOIN (SELECT `__table3`.`key` AS `key`, `__table3`.`value1` AS `value1`, `__table3`.`value2` AS `value2` FROM `default`.`join_inner_table` AS `__table3` PREWHERE (`__table3`.`id` = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (`__table3`.`number` > _CAST(1610517366120, \'UInt64\')) GROUP BY `__table3`.`key`, `__table3`.`value1`, `__table3`.`value2`) AS `__table2` USING (`key`) GROUP BY `__table1`.`key`, `__table2`.`value1`, `__table2`.`value2`
 0	3	SELECT `key`, `value1`, `value2` FROM `default`.`join_inner_table` PREWHERE (`id` = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (`number` > toUInt64(\'1610517366120\')) GROUP BY `key`, `value1`, `value2`
 0	3	SELECT `value1`, `value2`, count() AS `count` FROM `default`.`join_outer_table` ALL INNER JOIN `_data_` USING (`key`) GROUP BY `key`, `value1`, `value2`
 1	1	-- Parallel full query\nSELECT\n    value1,\n    value2,\n    avg(count) AS avg\nFROM\n    (\n        SELECT\n            key,\n            value1,\n            value2,\n            count() AS count\n        FROM join_outer_table\n        INNER JOIN\n        (\n            SELECT\n                key,\n                value1,\n                value2,\n                toUInt64(min(time)) AS start_ts\n            FROM join_inner_table\n            PREWHERE (id = \'833c9e22-c245-4eb5-8745-117a9a1f26b1\') AND (number > toUInt64(\'1610517366120\'))\n            GROUP BY key, value1, value2\n        ) USING (key)\n        GROUP BY key, value1, value2\n        )\nGROUP BY value1, value2\nORDER BY value1, value2\nSETTINGS allow_experimental_parallel_reading_from_replicas = 1, enable_analyzer=0;
diff --git a/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.sql b/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.sql
index 8121d60a05b..299f1cbbc32 100644
--- a/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.sql
+++ b/tests/queries/0_stateless/02731_parallel_replicas_join_subquery.sql
@@ -22,8 +22,8 @@ SELECT
 LIMIT 100;
 
 SET max_parallel_replicas = 3;
-SET prefer_localhost_replica = 1;
 SET cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
+SET parallel_replicas_local_plan = 1;
 SET joined_subquery_requires_alias = 0;
 
 SELECT '=============== INNER QUERY (NO PARALLEL) ===============';
diff --git a/tests/queries/0_stateless/02769_parallel_replicas_unavailable_shards.sql b/tests/queries/0_stateless/02769_parallel_replicas_unavailable_shards.sql
index 1a75e000349..7dbc389b55c 100644
--- a/tests/queries/0_stateless/02769_parallel_replicas_unavailable_shards.sql
+++ b/tests/queries/0_stateless/02769_parallel_replicas_unavailable_shards.sql
@@ -4,7 +4,9 @@ INSERT INTO test_parallel_replicas_unavailable_shards SELECT * FROM numbers(10);
 
 SET allow_experimental_parallel_reading_from_replicas=2, max_parallel_replicas=11, cluster_for_parallel_replicas='parallel_replicas', parallel_replicas_for_non_replicated_merge_tree=1;
 SET send_logs_level='error';
-SELECT count() FROM test_parallel_replicas_unavailable_shards WHERE NOT ignore(*) SETTINGS log_comment = '02769_7b513191-5082-4073-8568-53b86a49da79';
+-- with local plan for initiator, the query can be executed fast on initator, we can simply not come to the point where unavailable replica can be detected
+-- therefore disable local plan for now
+SELECT count() FROM test_parallel_replicas_unavailable_shards WHERE NOT ignore(*) SETTINGS log_comment = '02769_7b513191-5082-4073-8568-53b86a49da79', parallel_replicas_local_plan=0;
 
 SYSTEM FLUSH LOGS;
 
diff --git a/tests/queries/0_stateless/02771_parallel_replicas_analyzer.reference b/tests/queries/0_stateless/02771_parallel_replicas_analyzer.reference
index f60f1e0a376..3c54d5977f3 100644
--- a/tests/queries/0_stateless/02771_parallel_replicas_analyzer.reference
+++ b/tests/queries/0_stateless/02771_parallel_replicas_analyzer.reference
@@ -9,4 +9,4 @@
 7885388429666205427
 8124171311239967992
 1	1	-- Simple query with analyzer and pure parallel replicas\nSELECT number\nFROM join_inner_table__fuzz_146_replicated\n    SETTINGS\n    enable_analyzer = 1,\n    max_parallel_replicas = 2,\n    cluster_for_parallel_replicas = \'test_cluster_one_shard_three_replicas_localhost\',\n    allow_experimental_parallel_reading_from_replicas = 1;
-0	2	SELECT `__table1`.`number` AS `number` FROM `default`.`join_inner_table__fuzz_146_replicated` AS `__table1`
+0	1	SELECT `__table1`.`number` AS `number` FROM `default`.`join_inner_table__fuzz_146_replicated` AS `__table1`
diff --git a/tests/queries/0_stateless/02771_parallel_replicas_analyzer.sql b/tests/queries/0_stateless/02771_parallel_replicas_analyzer.sql
index e60049f2756..54295ffda17 100644
--- a/tests/queries/0_stateless/02771_parallel_replicas_analyzer.sql
+++ b/tests/queries/0_stateless/02771_parallel_replicas_analyzer.sql
@@ -1,5 +1,5 @@
 -- Tags: zookeeper
-
+DROP TABLE IF EXISTS join_inner_table__fuzz_146_replicated;
 CREATE TABLE join_inner_table__fuzz_146_replicated
 (
     `id` UUID,
@@ -17,6 +17,8 @@ INSERT INTO join_inner_table__fuzz_146_replicated
     SELECT CAST('833c9e22-c245-4eb5-8745-117a9a1f26b1', 'UUID') AS id, CAST(rowNumberInAllBlocks(), 'String') AS key, *
     FROM generateRandom('number Int64, value1 String, value2 String, time Int64', 1, 10, 2) LIMIT 10;
 
+SET parallel_replicas_local_plan = 1;
+
 -- Simple query with analyzer and pure parallel replicas
 SELECT number
 FROM join_inner_table__fuzz_146_replicated
@@ -49,3 +51,5 @@ WHERE
       )
 GROUP BY is_initial_query, query
 ORDER BY is_initial_query DESC, c, query;
+
+DROP TABLE join_inner_table__fuzz_146_replicated;
diff --git a/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision.sh b/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision.sh
index 8a3b34e5cfa..ef41a5d6277 100755
--- a/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision.sh
+++ b/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision.sh
@@ -6,16 +6,17 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 CLICKHOUSE_CLIENT_TRACE=${CLICKHOUSE_CLIENT/"--send_logs_level=${CLICKHOUSE_CLIENT_SERVER_LOGS_LEVEL}"/"--send_logs_level=trace"}
 
 function were_parallel_replicas_used () {
-    # Not using current_database = '$CLICKHOUSE_DATABASE' as nested parallel queries aren't run with it
     $CLICKHOUSE_CLIENT --query "
         SELECT
             initial_query_id,
-            concat('Used parallel replicas: ', (countIf(initial_query_id != query_id) != 0)::bool::String) as used
+            concat('Used parallel replicas: ', (ProfileEvents['ParallelReplicasUsedCount'] > 0)::bool::String) as used
         FROM system.query_log
     WHERE event_date >= yesterday()
       AND initial_query_id LIKE '$1%'
-    GROUP BY initial_query_id
-    ORDER BY min(event_time_microseconds) ASC
+      AND query_id = initial_query_id
+      AND type = 'QueryFinish'
+      AND current_database = '$CLICKHOUSE_DATABASE'
+    ORDER BY event_time_microseconds ASC
     FORMAT TSV"
 }
 
@@ -48,11 +49,11 @@ function run_query_with_pure_parallel_replicas () {
         --query "$3" \
         --query_id "${1}_pure" \
         --max_parallel_replicas 3 \
-        --prefer_localhost_replica 1 \
         --cluster_for_parallel_replicas "parallel_replicas" \
         --allow_experimental_parallel_reading_from_replicas 1 \
         --parallel_replicas_for_non_replicated_merge_tree 1 \
         --parallel_replicas_min_number_of_rows_per_replica "$2" \
+        --max_threads 5 \
     |& grep "It is enough work for" | awk '{ print substr($7, 2, length($7) - 2) "\t" $20 " estimated parallel replicas" }'
 }
 
diff --git a/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.reference b/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.reference
index 521e3e2edbc..595b35db610 100644
--- a/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.reference
+++ b/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.reference
@@ -4,8 +4,8 @@
 02784_automatic_parallel_replicas_join-default_simple_join_1M_pure	10 estimated parallel replicas
 02784_automatic_parallel_replicas_join-default_simple_join_300k_pure	3 estimated parallel replicas
 02784_automatic_parallel_replicas_join-default_simple_join_300k_pure	33 estimated parallel replicas
-02784_automatic_parallel_replicas_join-default_simple_join_0_pure	Distinct parallel subqueries: 2	Used parallel replicas: true
-02784_automatic_parallel_replicas_join-default_simple_join_10M_pure	Distinct parallel subqueries: 0	Used parallel replicas: false
-02784_automatic_parallel_replicas_join-default_simple_join_5M_pure	Distinct parallel subqueries: 1	Used parallel replicas: true
-02784_automatic_parallel_replicas_join-default_simple_join_1M_pure	Distinct parallel subqueries: 1	Used parallel replicas: true
-02784_automatic_parallel_replicas_join-default_simple_join_300k_pure	Distinct parallel subqueries: 2	Used parallel replicas: true
+02784_automatic_parallel_replicas_join-default_simple_join_0_pure	Used parallel replicas: true
+02784_automatic_parallel_replicas_join-default_simple_join_10M_pure	Used parallel replicas: false
+02784_automatic_parallel_replicas_join-default_simple_join_5M_pure	Used parallel replicas: true
+02784_automatic_parallel_replicas_join-default_simple_join_1M_pure	Used parallel replicas: true
+02784_automatic_parallel_replicas_join-default_simple_join_300k_pure	Used parallel replicas: true
diff --git a/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.sh b/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.sh
index 801cd22b79f..238d63e6980 100755
--- a/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.sh
+++ b/tests/queries/0_stateless/02784_parallel_replicas_automatic_decision_join.sh
@@ -9,17 +9,17 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 CLICKHOUSE_CLIENT_TRACE=${CLICKHOUSE_CLIENT/"--send_logs_level=${CLICKHOUSE_CLIENT_SERVER_LOGS_LEVEL}"/"--send_logs_level=trace"}
 
 function were_parallel_replicas_used () {
-    # Not using current_database = '$CLICKHOUSE_DATABASE' as nested parallel queries aren't run with it
     $CLICKHOUSE_CLIENT --query "
         SELECT
             initial_query_id,
-            concat('Distinct parallel subqueries: ' , countDistinctIf(query, initial_query_id != query_id)::String) as subqueries_parallelized,
-            concat('Used parallel replicas: ', (countIf(initial_query_id != query_id) != 0)::bool::String) as used
+            concat('Used parallel replicas: ', (ProfileEvents['ParallelReplicasUsedCount'] > 0)::bool::String) as used
         FROM system.query_log
     WHERE event_date >= yesterday()
       AND initial_query_id LIKE '$1%'
-    GROUP BY initial_query_id
-    ORDER BY min(event_time_microseconds) ASC
+      AND query_id = initial_query_id
+      AND type = 'QueryFinish'
+      AND current_database = '$CLICKHOUSE_DATABASE'
+    ORDER BY event_time_microseconds ASC
     FORMAT TSV"
 }
 
diff --git a/tests/queries/0_stateless/02841_parallel_replicas_summary.sh b/tests/queries/0_stateless/02841_parallel_replicas_summary.sh
index c82d2c8b0c0..01e2883c547 100755
--- a/tests/queries/0_stateless/02841_parallel_replicas_summary.sh
+++ b/tests/queries/0_stateless/02841_parallel_replicas_summary.sh
@@ -27,6 +27,7 @@ $CLICKHOUSE_CLIENT --query "CREATE TABLE replicas_summary (n Int64) ENGINE = Mer
 
 query_id_base="02841_summary_$CLICKHOUSE_DATABASE"
 
+# TODO: rethink the test, for now temporary disable parallel_replicas_local_plan
 echo "
     SELECT *
     FROM replicas_summary
@@ -36,7 +37,8 @@ echo "
         cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost',
         allow_experimental_parallel_reading_from_replicas = 2,
         parallel_replicas_for_non_replicated_merge_tree = 1,
-        interactive_delay=0
+        interactive_delay=0,
+        parallel_replicas_local_plan=0
     "\
     | ${CLICKHOUSE_CURL} -sS "${CLICKHOUSE_URL}&wait_end_of_query=1&query_id=${query_id_base}_interactive_0" --data-binary @- -vvv 2>&1 \
     | grep "Summary" | grep -cv '"read_rows":"0"'
@@ -50,7 +52,8 @@ echo "
         cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost',
         allow_experimental_parallel_reading_from_replicas = 2,
         parallel_replicas_for_non_replicated_merge_tree = 1,
-        interactive_delay=99999999999
+        interactive_delay=99999999999,
+        parallel_replicas_local_plan=0
     "\
     | ${CLICKHOUSE_CURL} -sS "${CLICKHOUSE_URL}&wait_end_of_query=1&query_id=${query_id_base}_interactive_high" --data-binary @- -vvv 2>&1 \
     | grep "Summary" | grep -cv '"read_rows":"0"'
diff --git a/tests/queries/0_stateless/02858_explicit_uuid_and_zk_path.reference b/tests/queries/0_stateless/02858_explicit_uuid_and_zk_path.reference
new file mode 100644
index 00000000000..874494fb061
--- /dev/null
+++ b/tests/queries/0_stateless/02858_explicit_uuid_and_zk_path.reference
@@ -0,0 +1,10 @@
+2
+3
+m1
+m2
+rmt1
+rmt2
+02858000-1000-4000-8000-000000000
+0
+CREATE TABLE default.rmt1\n(\n    `n` Int32\n)\nENGINE = ReplicatedMergeTree(\'/test/02858_explicit_uuid_and_zk_path_default/rmt/{shard}\', \'_{replica}\')\nORDER BY n\nSETTINGS index_granularity = 8192
+CREATE TABLE default.rmt2\n(\n    `n` Int32\n)\nENGINE = ReplicatedMergeTree(\'/clickhouse/tables/{uuid}/{shard}\', \'{replica}\')\nORDER BY n\nSETTINGS index_granularity = 8192
diff --git a/tests/queries/0_stateless/02858_explicit_uuid_and_zk_path.sh b/tests/queries/0_stateless/02858_explicit_uuid_and_zk_path.sh
new file mode 100755
index 00000000000..81a9cef02ff
--- /dev/null
+++ b/tests/queries/0_stateless/02858_explicit_uuid_and_zk_path.sh
@@ -0,0 +1,44 @@
+#!/usr/bin/env bash
+
+CLICKHOUSE_CLIENT_SERVER_LOGS_LEVEL=error
+
+CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+# shellcheck source=../shell_config.sh
+. "$CURDIR"/../shell_config.sh
+
+
+db=$CLICKHOUSE_DATABASE
+if [[ $($CLICKHOUSE_CLIENT -q "SELECT engine = 'Replicated' FROM system.databases WHERE name='$CLICKHOUSE_DATABASE'") != 1 ]]; then
+  $CLICKHOUSE_CLIENT -q "CREATE DATABASE rdb_$CLICKHOUSE_DATABASE ENGINE=Replicated('/test/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/rdb', '1', '1')"
+  db="rdb_$CLICKHOUSE_DATABASE"
+fi
+
+$CLICKHOUSE_CLIENT --distributed_ddl_output_mode=none --database_replicated_allow_explicit_uuid=0 -q "CREATE TABLE $db.m0
+UUID '02858000-1000-4000-8000-000000000000' (n int) ENGINE=Memory" 2>&1| grep -Fac "database_replicated_allow_explicit_uuid"
+
+$CLICKHOUSE_CLIENT --distributed_ddl_output_mode=none --database_replicated_allow_explicit_uuid=1 -q "CREATE TABLE $db.m1
+UUID '02858000-1000-4000-8000-000000000$(($RANDOM % 10))$(($RANDOM % 10))$(($RANDOM % 10))' (n int) ENGINE=Memory"
+
+$CLICKHOUSE_CLIENT --distributed_ddl_output_mode=none --database_replicated_allow_explicit_uuid=2 -q "CREATE TABLE $db.m2
+UUID '02858000-1000-4000-8000-000000000002' (n int) ENGINE=Memory"
+
+
+$CLICKHOUSE_CLIENT --distributed_ddl_output_mode=none --database_replicated_allow_replicated_engine_arguments=0 -q "CREATE TABLE $db.rmt0 (n int)
+ENGINE=ReplicatedMergeTree('/test/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/rmt/{shard}', '_{replica}') ORDER BY n" 2>&1| grep -Fac "database_replicated_allow_replicated_engine_arguments"
+
+$CLICKHOUSE_CLIENT --distributed_ddl_output_mode=none --database_replicated_allow_replicated_engine_arguments=1 -q "CREATE TABLE $db.rmt1 (n int)
+ENGINE=ReplicatedMergeTree('/test/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/rmt/{shard}', '_{replica}') ORDER BY n"
+
+$CLICKHOUSE_CLIENT --distributed_ddl_output_mode=none --database_replicated_allow_replicated_engine_arguments=2 -q "CREATE TABLE $db.rmt2 (n int)
+ENGINE=ReplicatedMergeTree('/test/$CLICKHOUSE_TEST_ZOOKEEPER_PREFIX/rmt/{shard}', '_{replica}') ORDER BY n"
+
+
+$CLICKHOUSE_CLIENT -q "SELECT name FROM system.tables WHERE database='$db' ORDER BY name"
+
+$CLICKHOUSE_CLIENT -q "SELECT substring(toString(uuid) as s, 1, length(s) - 3) FROM system.tables WHERE database='$db' and name='m1'"
+$CLICKHOUSE_CLIENT -q "SELECT toString(uuid) LIKE '02858000%' FROM system.tables WHERE database='$db' and name='m2'"
+
+$CLICKHOUSE_CLIENT -q "SHOW CREATE $db.rmt1" | sed "s/$db/default/g"
+$CLICKHOUSE_CLIENT -q "SHOW CREATE $db.rmt2" | sed "s/$db/default/g"
+
+$CLICKHOUSE_CLIENT -q "DROP DATABASE IF EXISTS rdb_$CLICKHOUSE_DATABASE"
diff --git a/tests/queries/0_stateless/02864_statistics_create_materialize_drop.reference b/tests/queries/0_stateless/02864_statistics_create_materialize_drop.reference
index 4ca2c5e5f9b..1843964377d 100644
--- a/tests/queries/0_stateless/02864_statistics_create_materialize_drop.reference
+++ b/tests/queries/0_stateless/02864_statistics_create_materialize_drop.reference
@@ -1,5 +1,5 @@
 Test create statistics:
-CREATE TABLE default.tab\n(\n    `a` LowCardinality(Int64) STATISTICS(tdigest, uniq, count_min, minmax),\n    `b` LowCardinality(Nullable(String)) STATISTICS(uniq, count_min),\n    `c` LowCardinality(Nullable(Int64)) STATISTICS(tdigest, uniq, count_min, minmax),\n    `d` DateTime STATISTICS(tdigest, uniq, count_min, minmax),\n    `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS index_granularity = 8192
+CREATE TABLE default.tab\n(\n    `a` LowCardinality(Int64) STATISTICS(tdigest, uniq, countmin, minmax),\n    `b` LowCardinality(Nullable(String)) STATISTICS(uniq, countmin),\n    `c` LowCardinality(Nullable(Int64)) STATISTICS(tdigest, uniq, countmin, minmax),\n    `d` DateTime STATISTICS(tdigest, uniq, countmin, minmax),\n    `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS index_granularity = 8192
 Test materialize and drop statistics:
-CREATE TABLE default.tab\n(\n    `a` LowCardinality(Int64),\n    `b` LowCardinality(Nullable(String)) STATISTICS(uniq, count_min),\n    `c` LowCardinality(Nullable(Int64)),\n    `d` DateTime,\n    `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS index_granularity = 8192
+CREATE TABLE default.tab\n(\n    `a` LowCardinality(Int64),\n    `b` LowCardinality(Nullable(String)) STATISTICS(uniq, countmin),\n    `c` LowCardinality(Nullable(Int64)),\n    `d` DateTime,\n    `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS index_granularity = 8192
 CREATE TABLE default.tab\n(\n    `a` LowCardinality(Int64),\n    `b` LowCardinality(Nullable(String)),\n    `c` LowCardinality(Nullable(Int64)),\n    `d` DateTime,\n    `pk` String\n)\nENGINE = MergeTree\nORDER BY pk\nSETTINGS index_granularity = 8192
diff --git a/tests/queries/0_stateless/02864_statistics_create_materialize_drop.sql b/tests/queries/0_stateless/02864_statistics_create_materialize_drop.sql
index 90a57c99624..249e3c84a51 100644
--- a/tests/queries/0_stateless/02864_statistics_create_materialize_drop.sql
+++ b/tests/queries/0_stateless/02864_statistics_create_materialize_drop.sql
@@ -12,10 +12,10 @@ SELECT 'Test create statistics:';
 
 CREATE TABLE tab
 (
-    a LowCardinality(Int64) STATISTICS(count_min, minmax, tdigest, uniq),
-    b LowCardinality(Nullable(String)) STATISTICS(count_min, uniq),
-    c LowCardinality(Nullable(Int64)) STATISTICS(count_min, minmax, tdigest, uniq),
-    d DateTime STATISTICS(count_min, minmax, tdigest, uniq),
+    a LowCardinality(Int64) STATISTICS(countmin, minmax, tdigest, uniq),
+    b LowCardinality(Nullable(String)) STATISTICS(countmin, uniq),
+    c LowCardinality(Nullable(Int64)) STATISTICS(countmin, minmax, tdigest, uniq),
+    d DateTime STATISTICS(countmin, minmax, tdigest, uniq),
     pk String,
 ) Engine = MergeTree() ORDER BY pk;
 
@@ -25,7 +25,7 @@ SHOW CREATE TABLE tab;
 
 SELECT 'Test materialize and drop statistics:';
 ALTER TABLE tab DROP STATISTICS a, b, c, d;
-ALTER TABLE tab ADD STATISTICS b TYPE count_min, uniq;
+ALTER TABLE tab ADD STATISTICS b TYPE countmin, uniq;
 ALTER TABLE tab MATERIALIZE STATISTICS b;
 SHOW CREATE TABLE tab;
 
diff --git a/tests/queries/0_stateless/02864_statistics_ddl.sql b/tests/queries/0_stateless/02864_statistics_ddl.sql
index bcaaa9e7b61..5b2c5cebc1d 100644
--- a/tests/queries/0_stateless/02864_statistics_ddl.sql
+++ b/tests/queries/0_stateless/02864_statistics_ddl.sql
@@ -1,5 +1,5 @@
 -- Tags: no-fasttest
--- no-fasttest: 'count_min' sketches need a 3rd party library
+-- no-fasttest: 'countmin' sketches need a 3rd party library
 
 -- Tests that DDL statements which create / drop / materialize statistics
 
@@ -71,29 +71,29 @@ CREATE TABLE tab (col Map(UInt64, UInt64) STATISTICS(uniq)) Engine = MergeTree()
 CREATE TABLE tab (col UUID STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
 CREATE TABLE tab (col IPv6 STATISTICS(uniq)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
 
---   count_min requires data_type.isValueRepresentedByInteger or data_type = (Fixed)String
+--   countmin requires data_type.isValueRepresentedByInteger or data_type = (Fixed)String
 --     These types work:
-CREATE TABLE tab (col UInt8 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col UInt256 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col Float32 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col Decimal32(3) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col Date STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col Date32 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col DateTime STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col DateTime64 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col Enum('hello', 'world') STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col IPv4 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col Nullable(UInt8) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col LowCardinality(UInt8) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col LowCardinality(Nullable(UInt8)) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col String STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
-CREATE TABLE tab (col FixedString(1) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col UInt8 STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col UInt256 STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col Float32 STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col Decimal32(3) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col Date STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col Date32 STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col DateTime STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col DateTime64 STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col Enum('hello', 'world') STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col IPv4 STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col Nullable(UInt8) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col LowCardinality(UInt8) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col LowCardinality(Nullable(UInt8)) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col String STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
+CREATE TABLE tab (col FixedString(1) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); DROP TABLE tab;
 --     These types don't work:
-CREATE TABLE tab (col Array(Float64) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
-CREATE TABLE tab (col Tuple(Float64, Float64) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
-CREATE TABLE tab (col Map(UInt64, UInt64) STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
-CREATE TABLE tab (col UUID STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
-CREATE TABLE tab (col IPv6 STATISTICS(count_min)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
+CREATE TABLE tab (col Array(Float64) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
+CREATE TABLE tab (col Tuple(Float64, Float64) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
+CREATE TABLE tab (col Map(UInt64, UInt64) STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
+CREATE TABLE tab (col UUID STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
+CREATE TABLE tab (col IPv6 STATISTICS(countmin)) Engine = MergeTree() ORDER BY tuple(); -- { serverError ILLEGAL_STATISTICS }
 
 --   minmax requires data_type.isValueRepresentedByInteger
 --     These types work:
@@ -187,17 +187,17 @@ ALTER TABLE tab MODIFY STATISTICS a TYPE tdigest; -- { serverError ILLEGAL_STATI
 --   uniq
 --     Works:
 ALTER TABLE tab ADD STATISTICS f64 TYPE uniq; ALTER TABLE tab DROP STATISTICS f64;
-ALTER TABLE tab MODIFY STATISTICS f64 TYPE count_min; ALTER TABLE tab DROP STATISTICS f64;
+ALTER TABLE tab MODIFY STATISTICS f64 TYPE countmin; ALTER TABLE tab DROP STATISTICS f64;
 --     Doesn't work:
 ALTER TABLE tab ADD STATISTICS a TYPE uniq; -- { serverError ILLEGAL_STATISTICS }
 ALTER TABLE tab MODIFY STATISTICS a TYPE uniq; -- { serverError ILLEGAL_STATISTICS }
---   count_min
+--   countmin
 --     Works:
-ALTER TABLE tab ADD STATISTICS f64 TYPE count_min; ALTER TABLE tab DROP STATISTICS f64;
-ALTER TABLE tab MODIFY STATISTICS f64 TYPE count_min; ALTER TABLE tab DROP STATISTICS f64;
+ALTER TABLE tab ADD STATISTICS f64 TYPE countmin; ALTER TABLE tab DROP STATISTICS f64;
+ALTER TABLE tab MODIFY STATISTICS f64 TYPE countmin; ALTER TABLE tab DROP STATISTICS f64;
 --     Doesn't work:
-ALTER TABLE tab ADD STATISTICS a TYPE count_min; -- { serverError ILLEGAL_STATISTICS }
-ALTER TABLE tab MODIFY STATISTICS a TYPE count_min; -- { serverError ILLEGAL_STATISTICS }
+ALTER TABLE tab ADD STATISTICS a TYPE countmin; -- { serverError ILLEGAL_STATISTICS }
+ALTER TABLE tab MODIFY STATISTICS a TYPE countmin; -- { serverError ILLEGAL_STATISTICS }
 --   minmax
 --     Works:
 ALTER TABLE tab ADD STATISTICS f64 TYPE minmax; ALTER TABLE tab DROP STATISTICS f64;
diff --git a/tests/queries/0_stateless/02864_statistics_predicates.sql b/tests/queries/0_stateless/02864_statistics_predicates.sql
index 473a7bc95ad..d7afba12c1d 100644
--- a/tests/queries/0_stateless/02864_statistics_predicates.sql
+++ b/tests/queries/0_stateless/02864_statistics_predicates.sql
@@ -1,5 +1,5 @@
 -- Tags: no-fasttest
--- no-fasttest: 'count_min' sketches need a 3rd party library
+-- no-fasttest: 'countmin' sketches need a 3rd party library
 
 -- Tests the cross product of all predicates with all right-hand sides on all data types and all statistics types.
 
@@ -13,27 +13,27 @@ CREATE TABLE tab
     u64                 UInt64,
     u64_tdigest         UInt64 STATISTICS(tdigest),
     u64_minmax          UInt64 STATISTICS(minmax),
-    u64_count_min       UInt64 STATISTICS(count_min),
+    u64_countmin       UInt64 STATISTICS(countmin),
     u64_uniq            UInt64 STATISTICS(uniq),
     f64                 Float64,
     f64_tdigest         Float64 STATISTICS(tdigest),
     f64_minmax          Float64 STATISTICS(minmax),
-    f64_count_min       Float64 STATISTICS(count_min),
+    f64_countmin       Float64 STATISTICS(countmin),
     f64_uniq            Float64 STATISTICS(uniq),
     dt                  DateTime,
     dt_tdigest          DateTime STATISTICS(tdigest),
     dt_minmax           DateTime STATISTICS(minmax),
-    dt_count_min        DateTime STATISTICS(count_min),
+    dt_countmin        DateTime STATISTICS(countmin),
     dt_uniq             DateTime STATISTICS(uniq),
     b                   Bool,
     b_tdigest           Bool STATISTICS(tdigest),
     b_minmax            Bool STATISTICS(minmax),
-    b_count_min         Bool STATISTICS(count_min),
+    b_countmin         Bool STATISTICS(countmin),
     b_uniq              Bool STATISTICS(uniq),
     s                   String,
  -- s_tdigest           String STATISTICS(tdigest), -- not supported by tdigest
  -- s_minmax            String STATISTICS(minmax), -- not supported by minmax
-    s_count_min         String STATISTICS(count_min),
+    s_countmin         String STATISTICS(countmin),
     s_uniq              String STATISTICS(uniq)
 ) Engine = MergeTree() ORDER BY tuple()
 SETTINGS min_bytes_for_wide_part = 0;
@@ -72,25 +72,25 @@ SELECT 'u64 and =';
 SELECT count(*) FROM tab WHERE u64 = 7;
 SELECT count(*) FROM tab WHERE u64_tdigest = 7;
 SELECT count(*) FROM tab WHERE u64_minmax = 7;
-SELECT count(*) FROM tab WHERE u64_count_min = 7;
+SELECT count(*) FROM tab WHERE u64_countmin = 7;
 SELECT count(*) FROM tab WHERE u64_uniq = 7;
 
 SELECT count(*) FROM tab WHERE u64 = 7.7;
 SELECT count(*) FROM tab WHERE u64_tdigest = 7.7;
 SELECT count(*) FROM tab WHERE u64_minmax = 7.7;
-SELECT count(*) FROM tab WHERE u64_count_min = 7.7;
+SELECT count(*) FROM tab WHERE u64_countmin = 7.7;
 SELECT count(*) FROM tab WHERE u64_uniq = 7.7;
 
 SELECT count(*) FROM tab WHERE u64 = '7';
 SELECT count(*) FROM tab WHERE u64_tdigest = '7';
 SELECT count(*) FROM tab WHERE u64_minmax = '7';
-SELECT count(*) FROM tab WHERE u64_count_min = '7';
+SELECT count(*) FROM tab WHERE u64_countmin = '7';
 SELECT count(*) FROM tab WHERE u64_uniq = '7';
 
 SELECT count(*) FROM tab WHERE u64 = '7.7'; -- { serverError TYPE_MISMATCH }
 SELECT count(*) FROM tab WHERE u64_tdigest = '7.7'; -- { serverError TYPE_MISMATCH }
 SELECT count(*) FROM tab WHERE u64_minmax = '7.7'; -- { serverError TYPE_MISMATCH }
-SELECT count(*) FROM tab WHERE u64_count_min = '7.7'; -- { serverError TYPE_MISMATCH }
+SELECT count(*) FROM tab WHERE u64_countmin = '7.7'; -- { serverError TYPE_MISMATCH }
 SELECT count(*) FROM tab WHERE u64_uniq = '7.7'; -- { serverError TYPE_MISMATCH }
 
 SELECT 'u64 and <';
@@ -98,25 +98,25 @@ SELECT 'u64 and <';
 SELECT count(*) FROM tab WHERE u64 < 7;
 SELECT count(*) FROM tab WHERE u64_tdigest < 7;
 SELECT count(*) FROM tab WHERE u64_minmax < 7;
-SELECT count(*) FROM tab WHERE u64_count_min < 7;
+SELECT count(*) FROM tab WHERE u64_countmin < 7;
 SELECT count(*) FROM tab WHERE u64_uniq < 7;
 
 SELECT count(*) FROM tab WHERE u64 < 7.7;
 SELECT count(*) FROM tab WHERE u64_tdigest < 7.7;
 SELECT count(*) FROM tab WHERE u64_minmax < 7.7;
-SELECT count(*) FROM tab WHERE u64_count_min < 7.7;
+SELECT count(*) FROM tab WHERE u64_countmin < 7.7;
 SELECT count(*) FROM tab WHERE u64_uniq < 7.7;
 
 SELECT count(*) FROM tab WHERE u64 < '7';
 SELECT count(*) FROM tab WHERE u64_tdigest < '7';
 SELECT count(*) FROM tab WHERE u64_minmax < '7';
-SELECT count(*) FROM tab WHERE u64_count_min < '7';
+SELECT count(*) FROM tab WHERE u64_countmin < '7';
 SELECT count(*) FROM tab WHERE u64_uniq < '7';
 
 SELECT count(*) FROM tab WHERE u64 < '7.7'; -- { serverError TYPE_MISMATCH }
 SELECT count(*) FROM tab WHERE u64_tdigest < '7.7'; -- { serverError TYPE_MISMATCH }
 SELECT count(*) FROM tab WHERE u64_minmax < '7.7'; -- { serverError TYPE_MISMATCH }
-SELECT count(*) FROM tab WHERE u64_count_min < '7.7'; -- { serverError TYPE_MISMATCH }
+SELECT count(*) FROM tab WHERE u64_countmin < '7.7'; -- { serverError TYPE_MISMATCH }
 SELECT count(*) FROM tab WHERE u64_uniq < '7.7'; -- { serverError TYPE_MISMATCH }
 
 -- f64 ----------------------------------------------------
@@ -126,25 +126,25 @@ SELECT 'f64 and =';
 SELECT count(*) FROM tab WHERE f64 = 7;
 SELECT count(*) FROM tab WHERE f64_tdigest = 7;
 SELECT count(*) FROM tab WHERE f64_minmax = 7;
-SELECT count(*) FROM tab WHERE f64_count_min = 7;
+SELECT count(*) FROM tab WHERE f64_countmin = 7;
 SELECT count(*) FROM tab WHERE f64_uniq = 7;
 
 SELECT count(*) FROM tab WHERE f64 = 7.7;
 SELECT count(*) FROM tab WHERE f64_tdigest = 7.7;
 SELECT count(*) FROM tab WHERE f64_minmax = 7.7;
-SELECT count(*) FROM tab WHERE f64_count_min = 7.7;
+SELECT count(*) FROM tab WHERE f64_countmin = 7.7;
 SELECT count(*) FROM tab WHERE f64_uniq = 7.7;
 
 SELECT count(*) FROM tab WHERE f64 = '7';
 SELECT count(*) FROM tab WHERE f64_tdigest = '7';
 SELECT count(*) FROM tab WHERE f64_minmax = '7';
-SELECT count(*) FROM tab WHERE f64_count_min = '7';
+SELECT count(*) FROM tab WHERE f64_countmin = '7';
 SELECT count(*) FROM tab WHERE f64_uniq = '7';
 
 SELECT count(*) FROM tab WHERE f64 = '7.7';
 SELECT count(*) FROM tab WHERE f64_tdigest = '7.7';
 SELECT count(*) FROM tab WHERE f64_minmax = '7.7';
-SELECT count(*) FROM tab WHERE f64_count_min = '7.7';
+SELECT count(*) FROM tab WHERE f64_countmin = '7.7';
 SELECT count(*) FROM tab WHERE f64_uniq = '7.7';
 
 SELECT 'f64 and <';
@@ -152,25 +152,25 @@ SELECT 'f64 and <';
 SELECT count(*) FROM tab WHERE f64 < 7;
 SELECT count(*) FROM tab WHERE f64_tdigest < 7;
 SELECT count(*) FROM tab WHERE f64_minmax < 7;
-SELECT count(*) FROM tab WHERE f64_count_min < 7;
+SELECT count(*) FROM tab WHERE f64_countmin < 7;
 SELECT count(*) FROM tab WHERE f64_uniq < 7;
 
 SELECT count(*) FROM tab WHERE f64 < 7.7;
 SELECT count(*) FROM tab WHERE f64_tdigest < 7.7;
 SELECT count(*) FROM tab WHERE f64_minmax < 7.7;
-SELECT count(*) FROM tab WHERE f64_count_min < 7.7;
+SELECT count(*) FROM tab WHERE f64_countmin < 7.7;
 SELECT count(*) FROM tab WHERE f64_uniq < 7.7;
 
 SELECT count(*) FROM tab WHERE f64 < '7';
 SELECT count(*) FROM tab WHERE f64_tdigest < '7';
 SELECT count(*) FROM tab WHERE f64_minmax < '7';
-SELECT count(*) FROM tab WHERE f64_count_min < '7';
+SELECT count(*) FROM tab WHERE f64_countmin < '7';
 SELECT count(*) FROM tab WHERE f64_uniq < '7';
 
 SELECT count(*) FROM tab WHERE f64 < '7.7';
 SELECT count(*) FROM tab WHERE f64_tdigest < '7.7';
 SELECT count(*) FROM tab WHERE f64_minmax < '7.7';
-SELECT count(*) FROM tab WHERE f64_count_min < '7.7';
+SELECT count(*) FROM tab WHERE f64_countmin < '7.7';
 SELECT count(*) FROM tab WHERE f64_uniq < '7.7';
 
 -- dt ----------------------------------------------------
@@ -180,13 +180,13 @@ SELECT 'dt and =';
 SELECT count(*) FROM tab WHERE dt = '2024-08-08 11:12:13';
 SELECT count(*) FROM tab WHERE dt_tdigest = '2024-08-08 11:12:13';
 SELECT count(*) FROM tab WHERE dt_minmax = '2024-08-08 11:12:13';
-SELECT count(*) FROM tab WHERE dt_count_min = '2024-08-08 11:12:13';
+SELECT count(*) FROM tab WHERE dt_countmin = '2024-08-08 11:12:13';
 SELECT count(*) FROM tab WHERE dt_uniq = '2024-08-08 11:12:13';
 
 SELECT count(*) FROM tab WHERE dt = 7;
 SELECT count(*) FROM tab WHERE dt_tdigest = 7;
 SELECT count(*) FROM tab WHERE dt_minmax = 7;
-SELECT count(*) FROM tab WHERE dt_count_min = 7;
+SELECT count(*) FROM tab WHERE dt_countmin = 7;
 SELECT count(*) FROM tab WHERE dt_uniq = 7;
 
 SELECT 'dt and <';
@@ -194,13 +194,13 @@ SELECT 'dt and <';
 SELECT count(*) FROM tab WHERE dt < '2024-08-08 11:12:13';
 SELECT count(*) FROM tab WHERE dt_tdigest < '2024-08-08 11:12:13';
 SELECT count(*) FROM tab WHERE dt_minmax < '2024-08-08 11:12:13';
-SELECT count(*) FROM tab WHERE dt_count_min < '2024-08-08 11:12:13';
+SELECT count(*) FROM tab WHERE dt_countmin < '2024-08-08 11:12:13';
 SELECT count(*) FROM tab WHERE dt_uniq < '2024-08-08 11:12:13';
 
 SELECT count(*) FROM tab WHERE dt < 7;
 SELECT count(*) FROM tab WHERE dt_tdigest < 7;
 SELECT count(*) FROM tab WHERE dt_minmax < 7;
-SELECT count(*) FROM tab WHERE dt_count_min < 7;
+SELECT count(*) FROM tab WHERE dt_countmin < 7;
 SELECT count(*) FROM tab WHERE dt_uniq < 7;
 
 -- b ----------------------------------------------------
@@ -210,25 +210,25 @@ SELECT 'b and =';
 SELECT count(*) FROM tab WHERE b = true;
 SELECT count(*) FROM tab WHERE b_tdigest = true;
 SELECT count(*) FROM tab WHERE b_minmax = true;
-SELECT count(*) FROM tab WHERE b_count_min = true;
+SELECT count(*) FROM tab WHERE b_countmin = true;
 SELECT count(*) FROM tab WHERE b_uniq = true;
 
 SELECT count(*) FROM tab WHERE b = 'true';
 SELECT count(*) FROM tab WHERE b_tdigest = 'true';
 SELECT count(*) FROM tab WHERE b_minmax = 'true';
-SELECT count(*) FROM tab WHERE b_count_min = 'true';
+SELECT count(*) FROM tab WHERE b_countmin = 'true';
 SELECT count(*) FROM tab WHERE b_uniq = 'true';
 
 SELECT count(*) FROM tab WHERE b = 1;
 SELECT count(*) FROM tab WHERE b_tdigest = 1;
 SELECT count(*) FROM tab WHERE b_minmax = 1;
-SELECT count(*) FROM tab WHERE b_count_min = 1;
+SELECT count(*) FROM tab WHERE b_countmin = 1;
 SELECT count(*) FROM tab WHERE b_uniq = 1;
 
 SELECT count(*) FROM tab WHERE b = 1.1;
 SELECT count(*) FROM tab WHERE b_tdigest = 1.1;
 SELECT count(*) FROM tab WHERE b_minmax = 1.1;
-SELECT count(*) FROM tab WHERE b_count_min = 1.1;
+SELECT count(*) FROM tab WHERE b_countmin = 1.1;
 SELECT count(*) FROM tab WHERE b_uniq = 1.1;
 
 -- s ----------------------------------------------------
@@ -238,13 +238,13 @@ SELECT 's and =';
 SELECT count(*) FROM tab WHERE s = 7; -- { serverError NO_COMMON_TYPE }
 -- SELECT count(*) FROM tab WHERE s_tdigest = 7; -- not supported
 -- SELECT count(*) FROM tab WHERE s_minmax = 7; -- not supported
-SELECT count(*) FROM tab WHERE s_count_min = 7; -- { serverError NO_COMMON_TYPE }
+SELECT count(*) FROM tab WHERE s_countmin = 7; -- { serverError NO_COMMON_TYPE }
 SELECT count(*) FROM tab WHERE s_uniq = 7; -- { serverError NO_COMMON_TYPE }
 
 SELECT count(*) FROM tab WHERE s = '7';
 -- SELECT count(*) FROM tab WHERE s_tdigest = '7'; -- not supported
 -- SELECT count(*) FROM tab WHERE s_minmax = '7'; -- not supported
-SELECT count(*) FROM tab WHERE s_count_min = '7';
+SELECT count(*) FROM tab WHERE s_countmin = '7';
 SELECT count(*) FROM tab WHERE s_uniq = '7';
 
 DROP TABLE tab;
diff --git a/tests/queries/0_stateless/02888_replicated_merge_tree_creation.sh b/tests/queries/0_stateless/02888_replicated_merge_tree_creation.sh
index b9603e75d2e..466f0d01a7f 100755
--- a/tests/queries/0_stateless/02888_replicated_merge_tree_creation.sh
+++ b/tests/queries/0_stateless/02888_replicated_merge_tree_creation.sh
@@ -5,6 +5,7 @@ CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CUR_DIR"/../shell_config.sh
 
+CLICKHOUSE_CLIENT="${CLICKHOUSE_CLIENT} --database_replicated_allow_explicit_uuid 3 --database_replicated_allow_replicated_engine_arguments 3"
 
 ${CLICKHOUSE_CLIENT} -q "DROP TABLE IF EXISTS test_exception_replicated SYNC"
 
diff --git a/tests/queries/0_stateless/02898_parallel_replicas_progress_bar.sql b/tests/queries/0_stateless/02898_parallel_replicas_progress_bar.sql
index d8bfec12b3a..9348ea1dc32 100644
--- a/tests/queries/0_stateless/02898_parallel_replicas_progress_bar.sql
+++ b/tests/queries/0_stateless/02898_parallel_replicas_progress_bar.sql
@@ -26,7 +26,8 @@ WHERE query_id in (select query_id from system.query_log where current_database
     AND message LIKE '%Total rows to read: 3000%' SETTINGS allow_experimental_parallel_reading_from_replicas=0;
 
 -- reading in order coordinator
-SELECT k, sipHash64(v) FROM t1 order by k limit 5 offset 998 SETTINGS optimize_read_in_order=1, log_comment='02898_inorder_190aed82-2423-413b-ad4c-24dcca50f65b';
+-- disable parallel_replicas_local_plan since the test relay on traces which only present in case of no local plan
+SELECT k, sipHash64(v) FROM t1 order by k limit 5 offset 998 SETTINGS optimize_read_in_order=1, parallel_replicas_local_plan=0, log_comment='02898_inorder_190aed82-2423-413b-ad4c-24dcca50f65b';
 
 SYSTEM FLUSH LOGS;
 SELECT count() > 0 FROM system.text_log
diff --git a/tests/queries/0_stateless/02916_to_start_of_interval_with_origin.reference b/tests/queries/0_stateless/02916_to_start_of_interval_with_origin.reference
new file mode 100644
index 00000000000..f0afdf03963
--- /dev/null
+++ b/tests/queries/0_stateless/02916_to_start_of_interval_with_origin.reference
@@ -0,0 +1,59 @@
+-- Negative tests
+Time and origin as Date
+2023-02-01
+2023-08-01
+2023-10-08
+2023-10-08
+2023-10-09
+Time and origin as Date32
+2023-02-01
+2023-08-01
+2023-10-08
+2023-10-08
+2023-10-09
+Time and origin as DateTime
+2023-02-01 09:08:07
+2023-08-01 09:08:07
+2023-10-08 09:08:07
+2023-10-08 09:08:07
+2023-10-09 09:08:07
+2023-10-09 10:10:07
+2023-10-09 10:11:07
+2023-10-09 10:11:12
+Time and origin as DateTime64(9)
+2023-02-01 09:08:07.123456789
+2023-08-01 09:08:07.123456789
+2023-09-10 09:08:07.123456789
+2023-10-08 09:08:07.123456789
+2023-10-09 09:08:07.123456789
+2023-10-09 10:10:07.123456789
+2023-10-09 10:11:11.123456789
+2023-10-09 10:11:12.123456789
+2023-10-09 10:11:12.987
+2023-10-09 10:11:12.987654
+2023-10-09 10:11:12.987654321
+Time and origin as DateTime64(3)
+2023-02-01 09:08:07.123
+2023-08-01 09:08:07.123
+2023-10-08 09:08:07.123
+2023-10-08 09:08:07.123
+2023-10-09 09:08:07.123
+2023-10-09 10:10:07.123
+2023-10-09 10:11:11.123
+2023-10-09 10:11:12.123
+2023-10-09 10:11:12.987
+2023-10-09 10:11:12.987000
+2023-10-09 10:11:12.987000000
+Non-const arguments
+2023-03-01 16:55:00.00
+2023-02-01 16:55:00.00
+2023-03-01 16:55:00.00
+2023-02-01 16:55:00.00
+2023-03-01 16:55:00.00
+2023-03-01 16:55:00
+2023-02-01 16:55:00
+2023-03-01 16:55:00
+2023-02-01 16:55:00
+2023-03-01 16:55:00
+2023-01-02 15:44:30
+2023-02-01 16:44:30.00
diff --git a/tests/queries/0_stateless/02916_to_start_of_interval_with_origin.sql b/tests/queries/0_stateless/02916_to_start_of_interval_with_origin.sql
new file mode 100644
index 00000000000..b03ccae31d9
--- /dev/null
+++ b/tests/queries/0_stateless/02916_to_start_of_interval_with_origin.sql
@@ -0,0 +1,95 @@
+set session_timezone = 'UTC';
+
+SELECT '-- Negative tests';
+
+-- time and origin arguments must have the same type
+SELECT toStartOfInterval(toDate('2023-01-02 14:45:50'), toIntervalSecond(5), toDate32('2023-01-02 14:44:30')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate('2023-01-02 14:45:50'), toIntervalMillisecond(12), toDateTime('2023-01-02 14:44:30')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate32('2023-01-02 14:45:50'), toIntervalHour(5), toDate('2023-01-02 14:44:30')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDateTime('2023-01-02 14:45:50'), toIntervalMinute(1), toDateTime64('2023-01-02 14:44:30', 2)); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDateTime64('2023-01-02 14:45:50', 2), toIntervalMinute(1), toDate('2023-01-02 14:44:30')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+
+-- the origin must be before the time
+SELECT toStartOfInterval(toDateTime('2023-01-02 14:42:50'), toIntervalMinute(1), toDateTime('2023-01-02 14:44:30')); -- { serverError BAD_ARGUMENTS }
+
+-- the origin must be constant
+SELECT toStartOfInterval(toDateTime('2023-01-02 14:45:50'), toIntervalMinute(1), number % 2 == 0 ? toDateTime('2023-02-01 15:55:00') : toDateTime('2023-01-01 15:55:00')) from numbers(1); -- { serverError ILLEGAL_COLUMN }
+SELECT toStartOfInterval(toDateTime('2023-01-02 14:45:50'), toIntervalHour(1), materialize(toDateTime('2023-01-02 14:44:30')), 'Europe/Amsterdam'); -- { serverError ILLEGAL_COLUMN }
+
+-- with 4 arguments, the 3rd one must not be a string or an integer
+SELECT toStartOfInterval(toDateTime('2023-01-02 14:45:50'), toIntervalYear(1), 'Europe/Amsterdam', 'Europe/Amsterdam'); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDateTime('2023-01-02 14:45:50'), toIntervalYear(1), 5, 'Europe/Amsterdam'); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+
+-- too many arguments
+SELECT toStartOfInterval(toDateTime('2023-01-02 14:45:50'), toIntervalYear(1), toDateTime('2020-01-02 14:44:30'), 'Europe/Amsterdam', 5); -- { serverError NUMBER_OF_ARGUMENTS_DOESNT_MATCH }
+
+SELECT 'Time and origin as Date';
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalYear(1), toDate('2022-02-01'));
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalQuarter(1), toDate('2022-02-01'));
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalMonth(1), toDate('2023-09-08'));
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalWeek(1), toDate('2023-10-01'));
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalDay(1), toDate('2023-10-08'));
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalHour(1), toDate('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalMinute(1), toDate('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalSecond(1), toDate('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalMillisecond(1), toDate('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalMicrosecond(1), toDate('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate('2023-10-09'), toIntervalNanosecond(1), toDate('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+
+SELECT 'Time and origin as Date32';
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalYear(1), toDate32('2022-02-01'));
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalQuarter(1), toDate32('2022-02-01'));
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalMonth(1), toDate32('2023-09-08'));
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalWeek(1), toDate32('2023-10-01'));
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalDay(1), toDate32('2023-10-08'));
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalHour(1), toDate32('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalMinute(1), toDate32('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalSecond(1), toDate32('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalMillisecond(1), toDate32('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalMicrosecond(1), toDate32('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDate32('2023-10-09'), toIntervalNanosecond(1), toDate32('2023-10-09')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+
+SELECT 'Time and origin as DateTime';
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalYear(1), toDateTime('2022-02-01 09:08:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalQuarter(1), toDateTime('2022-02-01 09:08:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalMonth(1), toDateTime('2023-09-08 09:08:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalWeek(1), toDateTime('2023-10-01 09:08:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalDay(1), toDateTime('2023-10-08 09:08:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalHour(1), toDateTime('2023-10-09 09:10:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalMinute(1), toDateTime('2023-10-09 09:10:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalSecond(1), toDateTime('2023-10-09 09:10:07'));
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalMillisecond(1), toDateTime('2023-10-09 10:11:12')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalMicrosecond(1), toDateTime('2023-10-09 10:11:12')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+SELECT toStartOfInterval(toDateTime('2023-10-09 10:11:12'), toIntervalNanosecond(1), toDateTime('2023-10-09 10:11:12')); -- { serverError ILLEGAL_TYPE_OF_ARGUMENT }
+
+SELECT 'Time and origin as DateTime64(9)';
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalYear(1), toDateTime64('2022-02-01 09:08:07.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalQuarter(1), toDateTime64('2022-02-01 09:08:07.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalMonth(1), toDateTime64('2023-09-10 09:08:07.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalWeek(1), toDateTime64('2023-10-01 09:08:07.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalDay(1), toDateTime64('2023-10-08 09:08:07.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalHour(1), toDateTime64('2023-10-09 09:10:07.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalMinute(1), toDateTime64('2023-10-09 09:10:11.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalSecond(1), toDateTime64('2023-10-09 10:11:10.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalMillisecond(1), toDateTime64('2023-10-09 10:11:12.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalMicrosecond(1), toDateTime64('2023-10-09 10:11:12.123456789', 9));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987654321', 9), toIntervalNanosecond(1), toDateTime64('2023-10-09 10:11:12.123456789', 9));
+
+SELECT 'Time and origin as DateTime64(3)';
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalYear(1), toDateTime64('2022-02-01 09:08:07.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalQuarter(1), toDateTime64('2022-02-01 09:08:07.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalMonth(1), toDateTime64('2023-09-08 09:08:07.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalWeek(1), toDateTime64('2023-10-01 09:08:07.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalDay(1), toDateTime64('2023-10-08 09:08:07.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalHour(1), toDateTime64('2023-10-09 09:10:07.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalMinute(1), toDateTime64('2023-10-09 10:10:11.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalSecond(1), toDateTime64('2023-10-09 10:11:10.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalMillisecond(1), toDateTime64('2023-10-09 10:11:12.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalMicrosecond(1), toDateTime64('2023-10-09 10:11:12.123', 3));
+SELECT toStartOfInterval(toDateTime64('2023-10-09 10:11:12.987', 3), toIntervalNanosecond(1), toDateTime64('2023-10-09 10:11:12.123', 3));
+
+SELECT 'Non-const arguments';
+SELECT toStartOfInterval(number % 2 == 0 ? toDateTime64('2023-03-01 15:55:00', 2) : toDateTime64('2023-02-01 15:55:00', 2), toIntervalMinute(1), toDateTime64('2023-01-01 13:55:00', 2), 'Europe/Amsterdam') from numbers(5);
+SELECT toStartOfInterval(number % 2 == 0 ? toDateTime('2023-03-01 15:55:00') : toDateTime('2023-02-01 15:55:00'), toIntervalHour(1), toDateTime('2023-01-01 13:55:00'), 'Europe/Amsterdam') from numbers(5);
+SELECT toStartOfInterval(materialize(toDateTime('2023-01-02 14:45:50')), toIntervalHour(1), toDateTime('2023-01-02 14:44:30'), 'Europe/Amsterdam');
+SELECT toStartOfInterval(materialize(toDateTime64('2023-02-01 15:45:50', 2)), toIntervalHour(1), toDateTime64('2023-01-02 14:44:30', 2), 'Europe/Amsterdam');
diff --git a/tests/queries/0_stateless/02933_sqid.reference b/tests/queries/0_stateless/02933_sqid.reference
index a559bacb0ac..4597e2347e3 100644
--- a/tests/queries/0_stateless/02933_sqid.reference
+++ b/tests/queries/0_stateless/02933_sqid.reference
@@ -13,5 +13,6 @@ Td1EnWQo	[1,2,3,4]
 XMbT
 -- invalid sqid
 []
+-- bug 69450
 -- alias
 XMbT
diff --git a/tests/queries/0_stateless/02933_sqid.sql b/tests/queries/0_stateless/02933_sqid.sql
index 81d4b2bc35c..822fe33df51 100644
--- a/tests/queries/0_stateless/02933_sqid.sql
+++ b/tests/queries/0_stateless/02933_sqid.sql
@@ -25,5 +25,12 @@ SELECT sqidEncode(toNullable(materialize(1)), toLowCardinality(materialize(2)));
 SELECT '-- invalid sqid';
 SELECT sqidDecode('invalid sqid');
 
+SELECT '-- bug 69450';
+DROP TABLE IF EXISTS tab;
+CREATE TABLE tab (id String) ENGINE = MergeTree ORDER BY id;
+INSERT INTO tab SELECT * FROM generateRandom() LIMIT 1000000;
+SELECT sqidDecode(id) FROM tab FORMAT Null;
+DROP TABLE tab;
+
 SELECT '-- alias';
 SELECT sqid(1, 2);
diff --git a/tests/queries/0_stateless/02950_parallel_replicas_used_count.reference b/tests/queries/0_stateless/02950_parallel_replicas_used_count.reference
index 21b7b527b7a..c1265b7ca14 100644
--- a/tests/queries/0_stateless/02950_parallel_replicas_used_count.reference
+++ b/tests/queries/0_stateless/02950_parallel_replicas_used_count.reference
@@ -1,4 +1,4 @@
-100	4950
+10000	49995000
 1
 89
 90
@@ -6,3 +6,9 @@
 92
 93
 1
+93
+92
+91
+90
+89
+1
diff --git a/tests/queries/0_stateless/02950_parallel_replicas_used_count.sql b/tests/queries/0_stateless/02950_parallel_replicas_used_count.sql
index 22f55acd365..4396ca60e0e 100644
--- a/tests/queries/0_stateless/02950_parallel_replicas_used_count.sql
+++ b/tests/queries/0_stateless/02950_parallel_replicas_used_count.sql
@@ -2,11 +2,12 @@ DROP TABLE IF EXISTS test;
 
 CREATE TABLE test (k UInt64, v String)
 ENGINE = MergeTree
-ORDER BY k;
+ORDER BY k
+SETTINGS index_granularity=1;
 
-INSERT INTO test SELECT number, toString(number) FROM numbers(100);
+INSERT INTO test SELECT number, toString(number) FROM numbers(10_000);
 
-SET allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 3, prefer_localhost_replica = 0, parallel_replicas_for_non_replicated_merge_tree=1, cluster_for_parallel_replicas='test_cluster_one_shard_three_replicas_localhost';
+SET allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 3, parallel_replicas_for_non_replicated_merge_tree=1, cluster_for_parallel_replicas='test_cluster_one_shard_three_replicas_localhost';
 
 -- default coordinator
 SELECT count(), sum(k)
@@ -14,12 +15,18 @@ FROM test
 SETTINGS log_comment = '02950_parallel_replicas_used_replicas_count';
 
 SYSTEM FLUSH LOGS;
-SELECT ProfileEvents['ParallelReplicasUsedCount'] FROM system.query_log WHERE type = 'QueryFinish' AND query_id IN (SELECT query_id FROM system.query_log WHERE current_database = currentDatabase() AND log_comment = '02950_parallel_replicas_used_replicas_count' AND type = 'QueryFinish' AND initial_query_id = query_id)  SETTINGS allow_experimental_parallel_reading_from_replicas=0;
+SELECT ProfileEvents['ParallelReplicasUsedCount'] > 0 FROM system.query_log WHERE type = 'QueryFinish' AND query_id IN (SELECT query_id FROM system.query_log WHERE current_database = currentDatabase() AND log_comment = '02950_parallel_replicas_used_replicas_count' AND type = 'QueryFinish' AND initial_query_id = query_id)  SETTINGS allow_experimental_parallel_reading_from_replicas=0;
 
 -- In order coordinator
-SELECT k FROM test order by k limit 5 offset 89 SETTINGS optimize_read_in_order=1, log_comment='02950_parallel_replicas_used_replicas_count_2';
+SELECT k FROM test order by k limit 5 offset 89 SETTINGS optimize_read_in_order=1, log_comment='02950_parallel_replicas_used_replicas_count_2', merge_tree_min_rows_for_concurrent_read=1, max_threads=1;
 
 SYSTEM FLUSH LOGS;
-SELECT ProfileEvents['ParallelReplicasUsedCount'] FROM system.query_log WHERE type = 'QueryFinish' AND query_id IN (SELECT query_id FROM system.query_log WHERE current_database = currentDatabase() AND log_comment = '02950_parallel_replicas_used_replicas_count_2' AND type = 'QueryFinish' AND initial_query_id = query_id)  SETTINGS allow_experimental_parallel_reading_from_replicas=0;
+SELECT ProfileEvents['ParallelReplicasUsedCount'] > 0 FROM system.query_log WHERE type = 'QueryFinish' AND query_id IN (SELECT query_id FROM system.query_log WHERE current_database = currentDatabase() AND log_comment = '02950_parallel_replicas_used_replicas_count_2' AND type = 'QueryFinish' AND initial_query_id = query_id)  SETTINGS allow_experimental_parallel_reading_from_replicas=0;
+
+-- In reverse order coordinator
+SELECT k FROM test order by k desc limit 5 offset 9906 SETTINGS optimize_read_in_order=1, log_comment='02950_parallel_replicas_used_replicas_count_3', merge_tree_min_rows_for_concurrent_read=1, max_threads=1;
+
+SYSTEM FLUSH LOGS;
+SELECT ProfileEvents['ParallelReplicasUsedCount'] > 0 FROM system.query_log WHERE type = 'QueryFinish' AND query_id IN (SELECT query_id FROM system.query_log WHERE current_database = currentDatabase() AND log_comment = '02950_parallel_replicas_used_replicas_count_3' AND type = 'QueryFinish' AND initial_query_id = query_id)  SETTINGS allow_experimental_parallel_reading_from_replicas=0;
 
 DROP TABLE test;
diff --git a/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.reference b/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.reference
index c0485b817c4..87e78d7d116 100644
--- a/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.reference
+++ b/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.reference
@@ -11,10 +11,8 @@ nested join with analyzer and parallel replicas, both global
 420336	420336	420336	-140112
 420378	420378	420378	-140126
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
 <Debug> DefaultCoordinator: Coordination done
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
 <Debug> DefaultCoordinator: Coordination done
 
 nested join with analyzer and parallel replicas, global + local
@@ -29,10 +27,8 @@ nested join with analyzer and parallel replicas, global + local
 420336	420336	420336	-140112
 420378	420378	420378	-140126
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` ALL INNER JOIN (SELECT `__table3`.`number` * 7 AS `key` FROM numbers(100000.) AS `__table3`) AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` ALL INNER JOIN (SELECT `__table3`.`number` * 7 AS `key` FROM numbers(100000.) AS `__table3`) AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
 <Debug> DefaultCoordinator: Coordination done
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
 <Debug> DefaultCoordinator: Coordination done
 
 nested join with analyzer and parallel replicas, both local, both full sorting merge join
@@ -47,10 +43,8 @@ nested join with analyzer and parallel replicas, both local, both full sorting m
 420336	420336	420336	-140112
 420378	420378	420378	-140126
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
 <Debug> WithOrderCoordinator: Coordination done
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
 <Debug> WithOrderCoordinator: Coordination done
 
 nested join with analyzer and parallel replicas, both local, both full sorting and hash join
@@ -65,10 +59,8 @@ nested join with analyzer and parallel replicas, both local, both full sorting a
 420336	420336	420336	-140112
 420378	420378	420378	-140126
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
 <Debug> DefaultCoordinator: Coordination done
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
 <Debug> WithOrderCoordinator: Coordination done
 
 nested join with analyzer and parallel replicas, both local, both full sorting and hash join
@@ -83,8 +75,6 @@ nested join with analyzer and parallel replicas, both local, both full sorting a
 420336	420336	420336	-140112
 420378	420378	420378	-140126
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value` FROM `default`.`num_2` AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table2` ON `__table1`.`key` = `__table2`.`key` (stage: WithMergeableState)
 <Debug> WithOrderCoordinator: Coordination done
 SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
-SELECT `__table1`.`key` AS `key`, `__table1`.`value` AS `value`, `__table3`.`key` AS `r.key`, `__table3`.`value` AS `r.value` FROM (SELECT `__table2`.`key` AS `key`, `__table2`.`value` AS `value` FROM `default`.`num_1` AS `__table2`) AS `__table1` GLOBAL ALL INNER JOIN `_data_` AS `__table3` ON `__table1`.`key` = `__table3`.`key` ORDER BY `__table1`.`key` ASC LIMIT _CAST(10000, 'UInt64'), _CAST(10, 'UInt64') (stage: WithMergeableState)
 <Debug> DefaultCoordinator: Coordination done
diff --git a/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.sh b/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.sh
index c4b44ce11c5..a595e363ef4 100755
--- a/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.sh
+++ b/tests/queries/0_stateless/02967_parallel_replicas_join_algo_and_analyzer_3.sh
@@ -17,6 +17,8 @@ insert into num_1 select number * 2, toString(number * 2) from numbers(1e7);
 insert into num_2 select number * 3, -number from numbers(1.5e6);
 "
 
+PARALLEL_REPLICAS_SETTINGS="allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', parallel_replicas_local_plan=1"
+
 ##############
 echo
 echo "nested join with analyzer and parallel replicas, both global"
@@ -26,18 +28,14 @@ select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings parallel_replicas_prefer_local_join=0) r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1,
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', parallel_replicas_prefer_local_join=0"
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0"
 
 $CLICKHOUSE_CLIENT --max_rows_in_set_to_optimize_join 0 -q "
 select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings parallel_replicas_prefer_local_join=0) r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, send_logs_level='trace',
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', parallel_replicas_prefer_local_join=0" 2>&1 |
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, send_logs_level='trace', parallel_replicas_prefer_local_join=0" 2>&1 |
 grep "executeQuery\|<Debug>.*Coordinator: Coordination done" |
 grep -o "SELECT.*WithMergeableState)\|<Debug>.*Coordinator: Coordination done" |
 sed -re 's/_data_[[:digit:]]+_[[:digit:]]+/_data_/g'
@@ -51,18 +49,14 @@ select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings parallel_replicas_prefer_local_join=1) r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1,
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', parallel_replicas_prefer_local_join=0"
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0"
 
 $CLICKHOUSE_CLIENT --max_rows_in_set_to_optimize_join 0 -q "
 select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings parallel_replicas_prefer_local_join=1) r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, send_logs_level='trace',
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', parallel_replicas_prefer_local_join=0" 2>&1 |
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, send_logs_level='trace', parallel_replicas_prefer_local_join=0" 2>&1 |
 grep "executeQuery\|<Debug>.*Coordinator: Coordination done" |
 grep -o "SELECT.*WithMergeableState)\|<Debug>.*Coordinator: Coordination done" |
 sed -re 's/_data_[[:digit:]]+_[[:digit:]]+/_data_/g'
@@ -77,18 +71,14 @@ select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings join_algorithm='full_sorting_merge') r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, parallel_replicas_prefer_local_join=0,
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', join_algorithm='full_sorting_merge'"
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0, join_algorithm='full_sorting_merge'"
 
 $CLICKHOUSE_CLIENT --max_rows_in_set_to_optimize_join 0 -q "
 select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings join_algorithm='full_sorting_merge') r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, parallel_replicas_prefer_local_join=0, send_logs_level='trace',
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', join_algorithm='full_sorting_merge'" 2>&1 |
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0, send_logs_level='trace', join_algorithm='full_sorting_merge'" 2>&1 |
 grep "executeQuery\|<Debug>.*Coordinator: Coordination done" |
 grep -o "SELECT.*WithMergeableState)\|<Debug>.*Coordinator: Coordination done" |
 sed -re 's/_data_[[:digit:]]+_[[:digit:]]+/_data_/g'
@@ -102,18 +92,14 @@ select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings join_algorithm='hash') r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, parallel_replicas_prefer_local_join=0,
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', join_algorithm='full_sorting_merge'"
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0, join_algorithm='full_sorting_merge'"
 
 $CLICKHOUSE_CLIENT --max_rows_in_set_to_optimize_join 0 -q "
 select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings join_algorithm='hash') r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, parallel_replicas_prefer_local_join=0, send_logs_level='trace',
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', join_algorithm='full_sorting_merge'" 2>&1 |
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0, send_logs_level='trace', join_algorithm='full_sorting_merge'" 2>&1 |
 grep "executeQuery\|<Debug>.*Coordinator: Coordination done" |
 grep -o "SELECT.*WithMergeableState)\|<Debug>.*Coordinator: Coordination done" |
 sed -re 's/_data_[[:digit:]]+_[[:digit:]]+/_data_/g'
@@ -127,18 +113,14 @@ select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings join_algorithm='full_sorting_merge') r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, parallel_replicas_prefer_local_join=0,
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', join_algorithm='hash'"
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0, join_algorithm='hash'"
 
 $CLICKHOUSE_CLIENT --max_rows_in_set_to_optimize_join 0 -q "
 select * from (select key, value from num_1) l
 inner join (select key, value from num_2 inner join
   (select number * 7 as key from numbers(1e5)) as nn on num_2.key = nn.key settings join_algorithm='full_sorting_merge') r
 on l.key = r.key order by l.key limit 10 offset 10000
-SETTINGS enable_analyzer=1, parallel_replicas_prefer_local_join=0, send_logs_level='trace',
-allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1,
-cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', join_algorithm='hash'" 2>&1 |
+SETTINGS enable_analyzer=1, $PARALLEL_REPLICAS_SETTINGS, parallel_replicas_prefer_local_join=0, send_logs_level='trace', join_algorithm='hash'" 2>&1 |
 grep "executeQuery\|<Debug>.*Coordinator: Coordination done" |
 grep -o "SELECT.*WithMergeableState)\|<Debug>.*Coordinator: Coordination done" |
 sed -re 's/_data_[[:digit:]]+_[[:digit:]]+/_data_/g'
diff --git a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference
index 36f02b2f764..ba67acd7d0b 100644
--- a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference
+++ b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.reference
@@ -19,8 +19,20 @@ select x, y, r.y, z, rr.z, a from (select l.x, l.y, r.y, r.z as z from (select x
 14	14	14	14	0	0
 15	15	0	0	0	0
 explain description=0 select x, y, r.y, z, rr.z, a from (select l.x, l.y, r.y, r.z as z from (select x, y from tab1 where x != 2) l any left join (select y, z from tab2 where y != 4) r on l.y = r.y) ll any left join (select z, a from tab3 where z != 8) rr on ll.z = rr.z SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', enable_analyzer=1;
-Expression
-  ReadFromRemoteParallelReplicas
+Union
+  Expression
+    Join
+      Expression
+        Join
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromMemoryStorage
+      Expression
+        ReadFromMemoryStorage
+  Expression
+    ReadFromRemoteParallelReplicas
 --
 -- The same query with cte;
 with sub1 as (select x, y from tab1 where x != 2),
@@ -55,8 +67,22 @@ select * from sub5 order by x
 SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', enable_analyzer=1;
 Expression
   Sorting
-    Expression
-      ReadFromRemoteParallelReplicas
+    Union
+      Expression
+        Sorting
+          Expression
+            Join
+              Expression
+                Join
+                  Expression
+                    Expression
+                      ReadFromMergeTree
+                  Expression
+                    ReadFromMemoryStorage
+              Expression
+                ReadFromMemoryStorage
+      Expression
+        ReadFromRemoteParallelReplicas
 --
 -- GROUP BY should work up to WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -80,8 +106,22 @@ Expression
   Sorting
     Expression
       MergingAggregated
-        Expression
-          ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Aggregating
+              Expression
+                Join
+                  Expression
+                    Join
+                      Expression
+                        Expression
+                          ReadFromMergeTree
+                      Expression
+                        ReadFromMemoryStorage
+                  Expression
+                    ReadFromMemoryStorage
+          Expression
+            ReadFromRemoteParallelReplicas
 --
 -- ORDER BY in sub3 : sub1 is fully pushed, sub3 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -118,10 +158,22 @@ Expression
   Sorting
     Expression
       Join
-        Expression
-          ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Join
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromMemoryStorage
+          Expression
+            ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
 --
 -- ORDER BY in sub1 : sub1 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2 order by y),
@@ -160,12 +212,24 @@ Expression
       Join
         Expression
           Join
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+        Union
+          Expression
             Expression
-              ReadFromRemoteParallelReplicas
-            Expression
-              ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
 --
 -- RIGHT JOIN in sub3: sub3 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -202,12 +266,24 @@ Expression
   Join
     Expression
       Join
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
+    Union
+      Expression
         Expression
-          ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
-    Expression
-      ReadFromRemoteParallelReplicas
+          ReadFromMergeTree
+      Expression
+        ReadFromRemoteParallelReplicas
 --
 -- RIGHT JOIN in sub5: sub5 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -242,14 +318,26 @@ Expression
   Sorting
     Expression
       Join
-        Expression
-          ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
         Expression
           Join
-            Expression
-              ReadFromRemoteParallelReplicas
-            Expression
-              ReadFromRemoteParallelReplicas
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
 --
 -- Subqueries for IN allowed
 with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)),
@@ -284,8 +372,28 @@ select * from sub5 order by x
 SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', enable_analyzer=1;
 Expression
   Sorting
-    Expression
-      ReadFromRemoteParallelReplicas
+    Union
+      Expression
+        Sorting
+          Expression
+            Join
+              Expression
+                Join
+                  Expression
+                    CreatingSets
+                      Expression
+                        Expression
+                          ReadFromMergeTree
+                      CreatingSet
+                        Expression
+                          Filter
+                            ReadFromSystemNumbers
+                  Expression
+                    ReadFromMemoryStorage
+              Expression
+                ReadFromMemoryStorage
+      Expression
+        ReadFromRemoteParallelReplicas
 --
 -- Subqueries for IN are not allowed
 with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)),
@@ -333,10 +441,18 @@ Expression
                   Expression
                     Filter
                       ReadFromSystemNumbers
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+        Union
+          Expression
             Expression
-              ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
 set parallel_replicas_prefer_local_join = 1;
 -- A query with only INNER/LEFT joins is fully send to replicas. JOIN is executed in GLOBAL mode.
 select x, y, r.y, z, rr.z, a from (select l.x, l.y, r.y, r.z as z from (select x, y from tab1 where x != 2) l any left join (select y, z from tab2 where y != 4) r on l.y = r.y) ll any left join (select z, a from tab3 where z != 8) rr on ll.z = rr.z order by x SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', enable_analyzer=1;
@@ -356,8 +472,22 @@ select x, y, r.y, z, rr.z, a from (select l.x, l.y, r.y, r.z as z from (select x
 14	14	14	14	0	0
 15	15	0	0	0	0
 explain description=0 select x, y, r.y, z, rr.z, a from (select l.x, l.y, r.y, r.z as z from (select x, y from tab1 where x != 2) l any left join (select y, z from tab2 where y != 4) r on l.y = r.y) ll any left join (select z, a from tab3 where z != 8) rr on ll.z = rr.z SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', enable_analyzer=1;
-Expression
-  ReadFromRemoteParallelReplicas
+Union
+  Expression
+    Join
+      Expression
+        Join
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            Expression
+              ReadFromMergeTree
+      Expression
+        Expression
+          ReadFromMergeTree
+  Expression
+    ReadFromRemoteParallelReplicas
 --
 -- The same query with cte;
 with sub1 as (select x, y from tab1 where x != 2),
@@ -392,8 +522,24 @@ select * from sub5 order by x
 SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', enable_analyzer=1;
 Expression
   Sorting
-    Expression
-      ReadFromRemoteParallelReplicas
+    Union
+      Expression
+        Sorting
+          Expression
+            Join
+              Expression
+                Join
+                  Expression
+                    Expression
+                      ReadFromMergeTree
+                  Expression
+                    Expression
+                      ReadFromMergeTree
+              Expression
+                Expression
+                  ReadFromMergeTree
+      Expression
+        ReadFromRemoteParallelReplicas
 --
 -- GROUP BY should work up to WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -417,8 +563,24 @@ Expression
   Sorting
     Expression
       MergingAggregated
-        Expression
-          ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Aggregating
+              Expression
+                Join
+                  Expression
+                    Join
+                      Expression
+                        Expression
+                          ReadFromMergeTree
+                      Expression
+                        Expression
+                          ReadFromMergeTree
+                  Expression
+                    Expression
+                      ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
 --
 -- ORDER BY in sub3 : sub1 is fully pushed, sub3 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -455,10 +617,23 @@ Expression
   Sorting
     Expression
       Join
-        Expression
-          ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Join
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                Expression
+                  ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
 --
 -- ORDER BY in sub1 : sub1 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2 order by y),
@@ -497,12 +672,24 @@ Expression
       Join
         Expression
           Join
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+        Union
+          Expression
             Expression
-              ReadFromRemoteParallelReplicas
-            Expression
-              ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
 --
 -- RIGHT JOIN in sub3: sub3 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -539,12 +726,24 @@ Expression
   Join
     Expression
       Join
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
+    Union
+      Expression
         Expression
-          ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
-    Expression
-      ReadFromRemoteParallelReplicas
+          ReadFromMergeTree
+      Expression
+        ReadFromRemoteParallelReplicas
 --
 -- RIGHT JOIN in sub5: sub5 -> WithMergableStage
 with sub1 as (select x, y from tab1 where x != 2),
@@ -579,14 +778,26 @@ Expression
   Sorting
     Expression
       Join
-        Expression
-          ReadFromRemoteParallelReplicas
+        Union
+          Expression
+            Expression
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
         Expression
           Join
-            Expression
-              ReadFromRemoteParallelReplicas
-            Expression
-              ReadFromRemoteParallelReplicas
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
 --
 -- Subqueries for IN allowed
 with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)),
@@ -621,8 +832,30 @@ select * from sub5 order by x
 SETTINGS allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 2, parallel_replicas_for_non_replicated_merge_tree = 1, prefer_localhost_replica = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', enable_analyzer=1;
 Expression
   Sorting
-    Expression
-      ReadFromRemoteParallelReplicas
+    Union
+      Expression
+        Sorting
+          Expression
+            Join
+              Expression
+                Join
+                  Expression
+                    CreatingSets
+                      Expression
+                        Expression
+                          ReadFromMergeTree
+                      CreatingSet
+                        Expression
+                          Filter
+                            ReadFromSystemNumbers
+                  Expression
+                    Expression
+                      ReadFromMergeTree
+              Expression
+                Expression
+                  ReadFromMergeTree
+      Expression
+        ReadFromRemoteParallelReplicas
 --
 -- Subqueries for IN are not allowed
 with sub1 as (select x, y from tab1 where x in (select number from numbers(16) where number != 2)),
@@ -670,7 +903,15 @@ Expression
                   Expression
                     Filter
                       ReadFromSystemNumbers
+            Union
+              Expression
+                Expression
+                  ReadFromMergeTree
+              Expression
+                ReadFromRemoteParallelReplicas
+        Union
+          Expression
             Expression
-              ReadFromRemoteParallelReplicas
-        Expression
-          ReadFromRemoteParallelReplicas
+              ReadFromMergeTree
+          Expression
+            ReadFromRemoteParallelReplicas
diff --git a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2 b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2
index 775663768fe..6f90f07c7a5 100644
--- a/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2
+++ b/tests/queries/0_stateless/02967_parallel_replicas_joins_and_analyzer.sql.j2
@@ -10,6 +10,8 @@ insert into tab1 select number, number, number from numbers(16);
 insert into tab2 select number * 2, number * 2 from numbers(8);
 insert into tab3 select number * 4, number * 4 from numbers(4);
 
+set parallel_replicas_local_plan=1;
+
 {% for use_global_in in [0, 1] -%}
 
 -- { echoOn }
diff --git a/tests/queries/0_stateless/03006_parallel_replicas_prewhere.reference b/tests/queries/0_stateless/03006_parallel_replicas_prewhere.reference
index e69de29bb2d..8b8d0e5d565 100644
--- a/tests/queries/0_stateless/03006_parallel_replicas_prewhere.reference
+++ b/tests/queries/0_stateless/03006_parallel_replicas_prewhere.reference
@@ -0,0 +1 @@
+default	url_na_log	1	130000	130
diff --git a/tests/queries/0_stateless/03006_parallel_replicas_prewhere.sql b/tests/queries/0_stateless/03006_parallel_replicas_prewhere.sql
index 4b84646c034..de43302690a 100644
--- a/tests/queries/0_stateless/03006_parallel_replicas_prewhere.sql
+++ b/tests/queries/0_stateless/03006_parallel_replicas_prewhere.sql
@@ -21,11 +21,21 @@ SELECT
 FROM numbers(130000)
 SETTINGS max_insert_block_size = 200000;
 
+SET max_block_size = 1048576, max_threads = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', max_parallel_replicas = 3, parallel_replicas_min_number_of_rows_per_replica=10000;
+
 EXPLAIN ESTIMATE
 SELECT count()
 FROM url_na_log
 PREWHERE (DateVisit >= toFixedString('2022-08-10', 10)) AND (DateVisit <= '2022-08-20')
-SETTINGS max_block_size = 1048576, max_threads = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost', max_parallel_replicas = 3, parallel_replicas_min_number_of_rows_per_replica=10000;
+SETTINGS parallel_replicas_local_plan=0;
+
+-- here parallel replicas uses local snapshot as working set
+-- so, the estimation can be done
+EXPLAIN ESTIMATE
+SELECT count()
+FROM url_na_log
+PREWHERE (DateVisit >= toFixedString('2022-08-10', 10)) AND (DateVisit <= '2022-08-20')
+SETTINGS allow_experimental_analyzer=1, parallel_replicas_local_plan=1;
 
 DROP POLICY url_na_log_policy0 ON url_na_log;
 DROP TABLE url_na_log;
diff --git a/tests/queries/0_stateless/03174_json_compact_with_progress.reference b/tests/queries/0_stateless/03174_json_compact_with_progress.reference
new file mode 100644
index 00000000000..cdbe7cfcb3e
--- /dev/null
+++ b/tests/queries/0_stateless/03174_json_compact_with_progress.reference
@@ -0,0 +1,15 @@
+1
+{"meta": [{"name":"value", "type":"UInt8"}, {"name":"name", "type":"String"}]}
+{"data":[1, "a"]}
+{"data":[2, "b"]}
+{"data":[3, "c"]}
+{"statistics": {"rows":3, "elapsed":ELAPSED, "rows_read":3, "bytes_read":33}}
+2
+{"meta": [{"name":"name", "type":"String"}, {"name":"c", "type":"UInt64"}]}
+{"data":["a", "1"]}
+{"data":["b", "1"]}
+{"data":["c", "1"]}
+{"totals": ["", "3"]}
+{"statistics": {"rows":3, "elapsed":ELAPSED, "rows_read":3, "bytes_read":30}}
+3
+Value passed to 'throwIf' function is non-zero:
diff --git a/tests/queries/0_stateless/03174_json_compact_with_progress.sh b/tests/queries/0_stateless/03174_json_compact_with_progress.sh
new file mode 100755
index 00000000000..b15dc7cfdb2
--- /dev/null
+++ b/tests/queries/0_stateless/03174_json_compact_with_progress.sh
@@ -0,0 +1,23 @@
+#!/usr/bin/env bash
+
+CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+# shellcheck source=../shell_config.sh
+. "$CUR_DIR"/../shell_config.sh
+
+$CLICKHOUSE_CLIENT -q "DROP TABLE IF EXISTS test_table;"
+
+$CLICKHOUSE_CLIENT -q "SELECT 1;"
+# Check JSONCompactWithProgress Output 
+$CLICKHOUSE_CLIENT -q "CREATE TABLE test_table (value UInt8, name String) ENGINE = MergeTree() ORDER BY value;"
+$CLICKHOUSE_CLIENT -q "INSERT INTO test_table VALUES (1, 'a'), (2, 'b'), (3, 'c');"
+$CLICKHOUSE_CLIENT -q "SELECT * FROM test_table FORMAT JSONCompactWithProgress settings max_block_size=2;" |  grep -v --text "progress" | sed -E 's/"elapsed":[0-9]+\.[0-9]+/"elapsed":ELAPSED/g'
+
+$CLICKHOUSE_CLIENT -q "SELECT 2;"
+# Check Totals 
+$CLICKHOUSE_CLIENT -q "SELECT name, count() AS c FROM test_table GROUP BY name WITH TOTALS ORDER BY name FORMAT JSONCompactWithProgress settings max_block_size=2;" | grep -v --text "progress" | sed -E 's/"elapsed":[0-9]+\.[0-9]+/"elapsed":ELAPSED/g'
+
+$CLICKHOUSE_CLIENT -q "SELECT 3;"
+# Check exceptions
+${CLICKHOUSE_CURL} -sS "$CLICKHOUSE_URL" -d "SELECT throwIf(number = 15), 1::Int64 as a, '\"' from numbers(100) format JSONCompactWithProgress settings output_format_json_quote_64bit_integers=1, max_block_size=10" | grep "exception" | cut -c42-88
+
+$CLICKHOUSE_CLIENT -q "DROP TABLE IF EXISTS test_table;"
diff --git a/tests/queries/0_stateless/03175_sparse_and_skip_index.reference b/tests/queries/0_stateless/03175_sparse_and_skip_index.reference
new file mode 100644
index 00000000000..619e98a152a
--- /dev/null
+++ b/tests/queries/0_stateless/03175_sparse_and_skip_index.reference
@@ -0,0 +1,4 @@
+key	Sparse
+value	Sparse
+1000
+1
diff --git a/tests/queries/0_stateless/03175_sparse_and_skip_index.sql b/tests/queries/0_stateless/03175_sparse_and_skip_index.sql
new file mode 100644
index 00000000000..4de6d1ac6df
--- /dev/null
+++ b/tests/queries/0_stateless/03175_sparse_and_skip_index.sql
@@ -0,0 +1,45 @@
+DROP TABLE IF EXISTS t_bloom_filter;
+CREATE TABLE t_bloom_filter(
+  key UInt64,
+  value UInt64,
+
+  INDEX key_bf key TYPE bloom_filter(0.01) GRANULARITY 2147483648, -- bloom filter on sorting key column
+  INDEX value_bf value TYPE bloom_filter(0.01) GRANULARITY 2147483648  -- bloom filter on no-sorting column
+) ENGINE=MergeTree ORDER BY key
+SETTINGS
+    -- settings to trigger sparse serialization and vertical merge
+    ratio_of_defaults_for_sparse_serialization = 0.0
+    ,vertical_merge_algorithm_min_rows_to_activate = 1
+    ,vertical_merge_algorithm_min_columns_to_activate = 1
+    ,allow_vertical_merges_from_compact_to_wide_parts = 1
+    ,min_bytes_for_wide_part=0
+;
+
+SYSTEM STOP MERGES t_bloom_filter;
+
+-- Create at least one part
+INSERT INTO t_bloom_filter
+SELECT
+    number % 100 as key, -- 100 unique keys
+    rand() % 100 as value -- 100 unique values
+FROM numbers(50_000);
+
+-- And another part
+INSERT INTO t_bloom_filter
+SELECT
+    number % 100 as key, -- 100 unique keys
+    rand() % 100 as value -- 100 unique values
+FROM numbers(50_000, 50_000);
+
+SYSTEM START MERGES t_bloom_filter;
+
+-- Merge everything into a single part
+OPTIMIZE TABLE t_bloom_filter FINAL;
+
+-- Check sparse serialization
+SELECT column, serialization_kind FROM system.parts_columns WHERE database = currentDatabase() AND table = 't_bloom_filter' AND active ORDER BY column;
+
+SELECT COUNT() FROM t_bloom_filter WHERE key = 1;
+
+-- Check bloom filter non-zero size
+SELECT COUNT() FROM system.parts WHERE database = currentDatabase() AND table = 't_bloom_filter' AND secondary_indices_uncompressed_bytes > 200 AND active;
diff --git a/tests/queries/0_stateless/03221_incomplete_utf8_sequence.reference b/tests/queries/0_stateless/03221_incomplete_utf8_sequence.reference
new file mode 100644
index 00000000000..4577427251d
--- /dev/null
+++ b/tests/queries/0_stateless/03221_incomplete_utf8_sequence.reference
@@ -0,0 +1,16 @@
+{
+	"meta":
+	[
+		{
+			"name": "unhex('f0')",
+			"type": "String"
+		}
+	],
+
+	"data":
+	[
+		["�"]
+	],
+
+	"rows": 1
+}
diff --git a/tests/queries/0_stateless/03221_incomplete_utf8_sequence.sql b/tests/queries/0_stateless/03221_incomplete_utf8_sequence.sql
new file mode 100644
index 00000000000..ee4f25f3b4a
--- /dev/null
+++ b/tests/queries/0_stateless/03221_incomplete_utf8_sequence.sql
@@ -0,0 +1,2 @@
+SET output_format_write_statistics = 0;
+SELECT unhex('f0') FORMAT JSONCompact;
diff --git a/tests/queries/0_stateless/03222_parallel_replicas_final_in_subquery.reference b/tests/queries/0_stateless/03222_parallel_replicas_final_in_subquery.reference
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/queries/0_stateless/03222_parallel_replicas_final_in_subquery.sql b/tests/queries/0_stateless/03222_parallel_replicas_final_in_subquery.sql
new file mode 100644
index 00000000000..6c2e05c12a9
--- /dev/null
+++ b/tests/queries/0_stateless/03222_parallel_replicas_final_in_subquery.sql
@@ -0,0 +1,21 @@
+DROP TABLE IF EXISTS test_00808;
+
+CREATE TABLE test_00808
+(
+    `date` Date,
+    `id` Int8,
+    `name` String,
+    `value` Int64,
+    `sign` Int8
+)
+ENGINE = CollapsingMergeTree(sign)
+ORDER BY (id, date);
+
+INSERT INTO test_00808 VALUES('2000-01-01', 1, 'test string 1', 1, 1);
+INSERT INTO test_00808 VALUES('2000-01-01', 2, 'test string 2', 2, 1);
+
+SET allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 3, parallel_replicas_for_non_replicated_merge_tree=1, cluster_for_parallel_replicas='test_cluster_one_shard_three_replicas_localhost';
+
+SELECT * FROM (SELECT * FROM test_00808 FINAL) WHERE id = 1; -- { serverError SUPPORT_IS_DISABLED }
+
+DROP TABLE test_00808;
diff --git a/tests/queries/0_stateless/03222_pr_asan_index_granularity.reference b/tests/queries/0_stateless/03222_pr_asan_index_granularity.reference
new file mode 100644
index 00000000000..662880055bd
--- /dev/null
+++ b/tests/queries/0_stateless/03222_pr_asan_index_granularity.reference
@@ -0,0 +1,100 @@
+0	18	9899
+0	18	9898
+0	18	9897
+0	18	9896
+0	18	9895
+0	18	9894
+0	18	9893
+0	18	9892
+0	18	9891
+0	18	9890
+0	18	9889
+0	18	9888
+0	18	9887
+0	18	9886
+0	18	9885
+0	18	9884
+0	18	9883
+0	18	9882
+0	18	9881
+0	18	9880
+0	18	9879
+0	18	9878
+0	18	9877
+0	18	9876
+0	18	9875
+0	18	9874
+0	18	9873
+0	18	9872
+0	18	9871
+0	18	9870
+0	18	9869
+0	18	9868
+0	18	9867
+0	18	9866
+0	18	9865
+0	18	9864
+0	18	9863
+0	18	9862
+0	18	9861
+0	18	9860
+0	18	9859
+0	18	9858
+0	18	9857
+0	18	9856
+0	18	9855
+0	18	9854
+0	18	9853
+0	18	9852
+0	18	9851
+0	18	9850
+0	18	9849
+0	18	9848
+0	18	9847
+0	18	9846
+0	18	9845
+0	18	9844
+0	18	9843
+0	18	9842
+0	18	9841
+0	18	9840
+0	18	9839
+0	18	9838
+0	18	9837
+0	18	9836
+0	18	9835
+0	18	9834
+0	18	9833
+0	18	9832
+0	18	9831
+0	18	9830
+0	18	9829
+0	18	9828
+0	18	9827
+0	18	9826
+0	18	9825
+0	18	9824
+0	18	9823
+0	18	9822
+0	18	9821
+0	18	9820
+0	18	9819
+0	18	9818
+0	18	9817
+0	18	9816
+0	18	9815
+0	18	9814
+0	18	9813
+0	18	9812
+0	18	9811
+0	18	9810
+0	18	9809
+0	18	9808
+0	18	9807
+0	18	9806
+0	18	9805
+0	18	9804
+0	18	9803
+0	18	9802
+0	18	9801
+0	18	9800
diff --git a/tests/queries/0_stateless/03222_pr_asan_index_granularity.sql b/tests/queries/0_stateless/03222_pr_asan_index_granularity.sql
new file mode 100644
index 00000000000..b7f37dd2856
--- /dev/null
+++ b/tests/queries/0_stateless/03222_pr_asan_index_granularity.sql
@@ -0,0 +1,14 @@
+DROP TABLE IF EXISTS test;
+
+CREATE TABLE test (k UInt64, v String)
+ENGINE = MergeTree
+ORDER BY k
+SETTINGS index_granularity=1;
+
+INSERT INTO test SELECT number, toString(number) FROM numbers(10_000);
+
+SET allow_experimental_parallel_reading_from_replicas = 2, max_parallel_replicas = 3, parallel_replicas_for_non_replicated_merge_tree=1, cluster_for_parallel_replicas='test_cluster_one_shard_three_replicas_localhost';
+
+SELECT 0, materialize(18), k FROM test PREWHERE toNullable(toNullable(11)) WHERE toNullable(11) ORDER BY k DESC NULLS LAST LIMIT 100, 100 SETTINGS optimize_read_in_order = 1, merge_tree_min_rows_for_concurrent_read = 9223372036854775806, max_threads = 1;
+
+-- DROP TABLE test;
diff --git a/tests/queries/0_stateless/03228_join_to_rerange_right_table.reference b/tests/queries/0_stateless/03228_join_to_rerange_right_table.reference
new file mode 100644
index 00000000000..b62923296e5
--- /dev/null
+++ b/tests/queries/0_stateless/03228_join_to_rerange_right_table.reference
@@ -0,0 +1,2 @@
+9
+9
diff --git a/tests/queries/0_stateless/03228_join_to_rerange_right_table.sql b/tests/queries/0_stateless/03228_join_to_rerange_right_table.sql
new file mode 100644
index 00000000000..f3ee0f0b933
--- /dev/null
+++ b/tests/queries/0_stateless/03228_join_to_rerange_right_table.sql
@@ -0,0 +1,14 @@
+drop table if exists test_left;
+drop table if exists test_right;
+
+CREATE TABLE test_left (a Int64, b String, c LowCardinality(String)) ENGINE = MergeTree() ORDER BY a;
+CREATE TABLE test_right (a Int64, b String, c LowCardinality(String)) ENGINE = MergeTree() ORDER BY a;
+
+INSERT INTO test_left SELECT number % 10000, number % 10000, number % 10000 FROM numbers(100000);
+INSERT INTO test_right SELECT number % 10 , number % 10, number % 10 FROM numbers(10000);
+
+SELECT MAX(test_right.a) FROM test_left INNER JOIN test_right on test_left.b = test_right.b SETTINGS allow_experimental_join_right_table_sorting=true;
+SELECT MAX(test_right.a) FROM test_left LEFT JOIN test_right on test_left.b = test_right.b SETTINGS allow_experimental_join_right_table_sorting=true;
+
+drop table test_left;
+drop table test_right;
diff --git a/tests/queries/0_stateless/03228_pr_subquery_view_order_by.reference b/tests/queries/0_stateless/03228_pr_subquery_view_order_by.reference
new file mode 100644
index 00000000000..aad720e27cf
--- /dev/null
+++ b/tests/queries/0_stateless/03228_pr_subquery_view_order_by.reference
@@ -0,0 +1,20 @@
+300
+299
+298
+297
+296
+295
+294
+293
+292
+291
+290
+289
+288
+287
+286
+285
+284
+283
+282
+281
diff --git a/tests/queries/0_stateless/03228_pr_subquery_view_order_by.sql b/tests/queries/0_stateless/03228_pr_subquery_view_order_by.sql
new file mode 100644
index 00000000000..804a97f737f
--- /dev/null
+++ b/tests/queries/0_stateless/03228_pr_subquery_view_order_by.sql
@@ -0,0 +1,18 @@
+DROP TABLE IF EXISTS view1;
+DROP TABLE IF EXISTS table1;
+CREATE TABLE table1 (number UInt64) ENGINE=MergeTree ORDER BY number SETTINGS index_granularity=1;
+INSERT INTO table1 SELECT number FROM numbers(1, 300);
+CREATE VIEW view1 AS SELECT number FROM table1;
+
+SELECT *
+FROM
+(
+    SELECT *
+    FROM view1
+)
+ORDER BY number DESC
+LIMIT 20
+SETTINGS cluster_for_parallel_replicas = 'parallel_replicas', allow_experimental_parallel_reading_from_replicas = 1, max_parallel_replicas = 3, parallel_replicas_for_non_replicated_merge_tree = 1, parallel_replicas_local_plan = 1;
+
+DROP TABLE view1;
+DROP TABLE table1;
diff --git a/tests/queries/0_stateless/03231_dynamic_not_safe_primary_key.reference b/tests/queries/0_stateless/03231_dynamic_not_safe_primary_key.reference
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/queries/0_stateless/03231_dynamic_not_safe_primary_key.sql b/tests/queries/0_stateless/03231_dynamic_not_safe_primary_key.sql
new file mode 100644
index 00000000000..f207581f482
--- /dev/null
+++ b/tests/queries/0_stateless/03231_dynamic_not_safe_primary_key.sql
@@ -0,0 +1,10 @@
+SET allow_experimental_dynamic_type = 1;
+DROP TABLE IF EXISTS t0;
+DROP TABLE IF EXISTS t1;
+CREATE TABLE t0 (c0 Int) ENGINE = AggregatingMergeTree() ORDER BY (c0);
+CREATE TABLE t1 (c0 Array(Dynamic), c1 Int) ENGINE = MergeTree() ORDER BY (c0);
+INSERT INTO t1 (c0, c1) VALUES ([18446717433683171873], 13623876564923702671), ([-4], 6111684076076982207);
+SELECT 1 FROM t0 FINAL JOIN t1 ON TRUE;
+DROP TABLE t0;
+DROP TABLE t1;
+
diff --git a/tests/queries/0_stateless/03231_values_respect_format_settings_in_fields_conversion.reference b/tests/queries/0_stateless/03231_values_respect_format_settings_in_fields_conversion.reference
new file mode 100644
index 00000000000..8dbf92d6590
--- /dev/null
+++ b/tests/queries/0_stateless/03231_values_respect_format_settings_in_fields_conversion.reference
@@ -0,0 +1 @@
+{'Hello':'2020-01-01 00:00:00'}
diff --git a/tests/queries/0_stateless/03231_values_respect_format_settings_in_fields_conversion.sql b/tests/queries/0_stateless/03231_values_respect_format_settings_in_fields_conversion.sql
new file mode 100644
index 00000000000..484a16bb22f
--- /dev/null
+++ b/tests/queries/0_stateless/03231_values_respect_format_settings_in_fields_conversion.sql
@@ -0,0 +1,7 @@
+drop table if exists test;
+create table test (map Map(String, DateTime)) engine=Memory;
+set date_time_input_format='best_effort';
+insert into test values (map('Hello', '01/01/2020'));
+select * from test;
+drop table test;
+
diff --git a/tests/queries/0_stateless/03232_pr_not_ready_set.sql b/tests/queries/0_stateless/03232_pr_not_ready_set.sql
index 1a724085903..3b2d5d28cfb 100644
--- a/tests/queries/0_stateless/03232_pr_not_ready_set.sql
+++ b/tests/queries/0_stateless/03232_pr_not_ready_set.sql
@@ -1,3 +1,4 @@
+SYSTEM FLUSH LOGS;
 SELECT
     is_initial_query,
     count() AS c,
diff --git a/tests/queries/0_stateless/03235_groupArray_string_consistency.reference b/tests/queries/0_stateless/03235_groupArray_string_consistency.reference
new file mode 100644
index 00000000000..d00491fd7e5
--- /dev/null
+++ b/tests/queries/0_stateless/03235_groupArray_string_consistency.reference
@@ -0,0 +1 @@
+1
diff --git a/tests/queries/0_stateless/03235_groupArray_string_consistency.sql b/tests/queries/0_stateless/03235_groupArray_string_consistency.sql
new file mode 100644
index 00000000000..618ec6f839b
--- /dev/null
+++ b/tests/queries/0_stateless/03235_groupArray_string_consistency.sql
@@ -0,0 +1,10 @@
+CREATE TABLE t (st FixedString(54)) ENGINE=MergeTree ORDER BY ();
+
+INSERT INTO t VALUES 
+('abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRTUVWXYZ'),
+('\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0'),
+('IIIIIIIIII\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0');
+
+WITH (SELECT groupConcat(',')(st) FROM t) AS a,
+     (SELECT groupConcat(',')(st :: String) FROM t) AS b
+SELECT equals(a, b);
diff --git a/tests/queries/0_stateless/03236_squashing_high_memory.reference b/tests/queries/0_stateless/03236_squashing_high_memory.reference
new file mode 100644
index 00000000000..e69de29bb2d
diff --git a/tests/queries/0_stateless/03236_squashing_high_memory.sql b/tests/queries/0_stateless/03236_squashing_high_memory.sql
new file mode 100644
index 00000000000..f6e5dbdef03
--- /dev/null
+++ b/tests/queries/0_stateless/03236_squashing_high_memory.sql
@@ -0,0 +1,27 @@
+-- Tags: no-fasttest, no-asan, no-tsan, no-msan, no-ubsan
+-- reason: test requires too many rows to read
+
+SET max_rows_to_read = '501G';
+
+DROP TABLE IF EXISTS id_values;
+
+DROP TABLE IF EXISTS test_table;
+
+CREATE TABLE id_values ENGINE MergeTree ORDER BY id1 AS
+    SELECT arrayJoin(range(500000)) AS id1, arrayJoin(range(1000)) AS id2;
+
+SET max_memory_usage = '1G';
+
+CREATE TABLE test_table ENGINE MergeTree ORDER BY id AS
+SELECT id_values.id1             AS id,
+    string_values.string_val1 AS string_val1,
+    string_values.string_val2 AS string_val2
+FROM id_values
+        JOIN (SELECT arrayJoin(range(10)) AS id1,
+                    'qwe'                AS string_val1,
+                    'asd'                AS string_val2) AS string_values
+            ON id_values.id1 = string_values.id1
+    SETTINGS join_algorithm = 'hash';
+
+DROP TABLE IF EXISTS id_values;
+DROP TABLE IF EXISTS test_table;
diff --git a/tests/queries/1_stateful/00177_memory_bound_merging.reference b/tests/queries/1_stateful/00177_memory_bound_merging.reference
index 5689152d60a..4c7505d1123 100644
--- a/tests/queries/1_stateful/00177_memory_bound_merging.reference
+++ b/tests/queries/1_stateful/00177_memory_bound_merging.reference
@@ -10,6 +10,8 @@ http://auto.ru/chatay-baranta_bound-in-thankYou=ru/tver/zhanny	2014-03-18	http:/
 http://auto.ru/chatay-baranta_bound-in-thankYou=ru/tver/zhanny	2014-03-19	http://auto.ru/chatay-baranta_bound-in-thankYou=ru/tver/zhanny
 http://auto.ru/chatay-baranta_bound-in-thankYou=ru/tver/zhanny	2014-03-20	http://auto.ru/chatay-baranta_bound-in-thankYou=ru/tver/zhanny
 1
-SortingAggregatedTransform
 MergingAggregatedBucketTransform
-GroupingAggregatedTransform
+FinishAggregatingInOrderTransform
+FinalizeAggregatedTransform
+AggregatingInOrderTransform
+MergeTreeSelect
diff --git a/tests/queries/1_stateful/00177_memory_bound_merging.sh b/tests/queries/1_stateful/00177_memory_bound_merging.sh
index d5cd1a05cd8..3744d89f354 100755
--- a/tests/queries/1_stateful/00177_memory_bound_merging.sh
+++ b/tests/queries/1_stateful/00177_memory_bound_merging.sh
@@ -9,7 +9,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 
 check_replicas_read_in_order() {
     # NOTE: lack of "current_database = '$CLICKHOUSE_DATABASE'" filter is made on purpose
-    $CLICKHOUSE_CLIENT -nq "
+    $CLICKHOUSE_CLIENT -q "
         SYSTEM FLUSH LOGS;
 
         SELECT COUNT() > 0
@@ -22,7 +22,7 @@ check_replicas_read_in_order() {
 # at some point we had a bug in this logic (see https://github.com/ClickHouse/ClickHouse/pull/45892#issue-1566140414)
 test1() {
     query_id="query_id_memory_bound_merging_$RANDOM$RANDOM"
-    $CLICKHOUSE_CLIENT --query_id="$query_id" -nq "
+    $CLICKHOUSE_CLIENT --query_id="$query_id" -q "
         SET cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
 
         SELECT URL, EventDate, max(URL)
@@ -39,7 +39,7 @@ test1() {
 # at some point we had a bug in this logic (see https://github.com/ClickHouse/ClickHouse/pull/45892#issue-1566140414)
 test2() {
     query_id="query_id_memory_bound_merging_$RANDOM$RANDOM"
-    $CLICKHOUSE_CLIENT --query_id="$query_id" -nq "
+    $CLICKHOUSE_CLIENT --query_id="$query_id" -q "
         SET cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
 
         SELECT URL, EventDate, max(URL)
@@ -53,9 +53,9 @@ test2() {
 }
 
 test3() {
-    $CLICKHOUSE_CLIENT -nq "
+    $CLICKHOUSE_CLIENT -q "
         SET cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
-        SET max_threads = 16, prefer_localhost_replica = 1, read_in_order_two_level_merge_threshold = 1000, query_plan_aggregation_in_order = 1, distributed_aggregation_memory_efficient = 1;
+        SET max_threads = 16, read_in_order_two_level_merge_threshold = 1000, query_plan_aggregation_in_order = 1, distributed_aggregation_memory_efficient = 1;
 
         SELECT replaceRegexpOne(explain, '^ *(\w+).*', '\\1')
         FROM (
@@ -64,7 +64,7 @@ test3() {
             FROM test.hits
             WHERE CounterID = 1704509 AND UserID = 4322253409885123546
             GROUP BY URL, EventDate
-            SETTINGS optimize_aggregation_in_order = 1, enable_memory_bound_merging_of_aggregation_results = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, max_parallel_replicas = 3
+            SETTINGS optimize_aggregation_in_order = 1, enable_memory_bound_merging_of_aggregation_results = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, max_parallel_replicas = 3, parallel_replicas_local_plan=1
         )
         WHERE explain LIKE '%Aggr%Transform%' OR explain LIKE '%InOrder%'"
 }
diff --git a/tests/queries/1_stateful/00180_no_seek_avoiding_when_reading_from_cache.sh b/tests/queries/1_stateful/00180_no_seek_avoiding_when_reading_from_cache.sh
index f9fea2c1dad..bf44f2d7ce7 100755
--- a/tests/queries/1_stateful/00180_no_seek_avoiding_when_reading_from_cache.sh
+++ b/tests/queries/1_stateful/00180_no_seek_avoiding_when_reading_from_cache.sh
@@ -23,7 +23,7 @@ $CLICKHOUSE_CLIENT -q "SELECT * FROM hits_s3_sampled WHERE URL LIKE '%google%' O
 query_id=02906_read_from_cache_$RANDOM
 $CLICKHOUSE_CLIENT --query_id ${query_id} -q "SELECT * FROM hits_s3_sampled WHERE URL LIKE '%google%' ORDER BY EventTime LIMIT 10 FORMAT Null SETTINGS filesystem_cache_reserve_space_wait_lock_timeout_milliseconds=2000"
 
-$CLICKHOUSE_CLIENT -nq "
+$CLICKHOUSE_CLIENT -q "
   SYSTEM FLUSH LOGS;
 
   -- AsynchronousReaderIgnoredBytes = 0: no seek-avoiding happened
diff --git a/utils/check-style/aspell-ignore/en/aspell-dict.txt b/utils/check-style/aspell-ignore/en/aspell-dict.txt
index d10db5f0d3d..3467f21c812 100644
--- a/utils/check-style/aspell-ignore/en/aspell-dict.txt
+++ b/utils/check-style/aspell-ignore/en/aspell-dict.txt
@@ -24,6 +24,7 @@ Aggregatefunction
 AggregatingMergeTree
 AggregatorThreads
 AggregatorThreadsActive
+AzureQueue
 Akka
 AlertManager
 Alexey
@@ -120,6 +121,7 @@ CMPLNT
 CMake
 CMakeLists
 CODECS
+CountMin
 COVID
 CPUFrequencyMHz
 CPUs
@@ -151,6 +153,7 @@ ChannelID
 Cidr
 Ciphertext
 CityHash
+CKibana
 Clangd
 ClickBench
 ClickCat
@@ -247,6 +250,7 @@ DoubleDelta
 Doxygen
 Durre
 ECMA
+ElasticSearch
 ETag
 Ecto
 EdgeAngle
@@ -421,6 +425,7 @@ JSONCompactStringsEachRowWithNames
 JSONCompactStringsEachRowWithNamesAndTypes
 JSONDynamicPaths
 JSONDynamicPathsWithTypes
+JSONCompactWithProgress
 JSONEachRow
 JSONEachRowWithProgress
 JSONExtract
@@ -471,6 +476,7 @@ KeeperMap
 KeeperOutstandingRequests
 Kerberos
 Khanna
+Kibana
 KittenHouse
 Klickhouse
 Kolmogorov
@@ -975,6 +981,7 @@ ThreadPoolRemoteFSReaderThreads
 ThreadPoolRemoteFSReaderThreadsActive
 ThreadsActive
 ThreadsInOvercommitTracker
+TimescaleDB's
 TimeSeries
 Timeunit
 TinyLog
@@ -1367,6 +1374,7 @@ cipherList
 ciphertext
 cityHash
 cityhash
+ckibana
 ckman
 clangd
 cli
@@ -1915,6 +1923,7 @@ jsoncompactstrings
 jsoncompactstringseachrow
 jsoncompactstringseachrowwithnames
 jsoncompactstringseachrowwithnamesandtypes
+jsoncompactwithprogress
 jsoneachrow
 jsoneachrowwithprogress
 jsonobjecteachrow
@@ -2583,6 +2592,7 @@ sqlinsert
 sqlite
 sqrt
 src
+dest
 srcReplicas
 sshkey
 stackoverflow
diff --git a/utils/check-style/experimental_settings_ignore.txt b/utils/check-style/experimental_settings_ignore.txt
index 94c46cf562e..3eda9821799 100644
--- a/utils/check-style/experimental_settings_ignore.txt
+++ b/utils/check-style/experimental_settings_ignore.txt
@@ -31,6 +31,7 @@ allow_experimental_statistics
 allow_experimental_time_series_table
 allow_experimental_undrop_table_query
 allow_experimental_usearch_index
+allow_experimental_join_right_table_sorting
 allow_get_client_http_header
 allow_introspection_functions
 allow_materialized_view_with_bad_select