Merge branch 'master' into llvm-14

.clang-tidy

@@ -16,7 +16,6 @@ Checks: '-*,
     modernize-make-unique,
     modernize-raw-string-literal,
     modernize-redundant-void-arg,
-    modernize-replace-auto-ptr,
     modernize-replace-random-shuffle,
     modernize-use-bool-literals,
     modernize-use-nullptr,
@@ -145,6 +144,7 @@ Checks: '-*,
     clang-analyzer-cplusplus.SelfAssignment,
     clang-analyzer-deadcode.DeadStores,
     clang-analyzer-cplusplus.Move,
+    clang-analyzer-optin.cplusplus.UninitializedObject,
     clang-analyzer-optin.cplusplus.VirtualCall,
     clang-analyzer-security.insecureAPI.UncheckedReturn,
     clang-analyzer-security.insecureAPI.bcmp,
@@ -164,6 +164,8 @@ Checks: '-*,
     clang-analyzer-unix.cstring.NullArg,
 
     boost-use-to-string,
+
+    alpha.security.cert.env.InvalidPtr,
 '
 WarningsAsErrors: '*'
 

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,4 +1,4 @@
-Changelog category (leave one):
+### Changelog category (leave one):
 - New Feature
 - Improvement
 - Bug Fix (user-visible misbehaviour in official stable or prestable release)
@@ -9,7 +9,7 @@ Changelog category (leave one):
 - Not for changelog (changelog entry is not required)
 
 
-Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):
+### Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):
 ...
 
 

.github/workflows/backport_branches.yml

@@ -341,10 +341,15 @@ jobs:
     steps:
       - name: Set envs
         run: |
+          DEPENDENCIES=$(cat << 'EOF' | jq '. | length'
+          ${{ toJSON(needs) }}
+          EOF
+          )
+          echo "DEPENDENCIES=$DEPENDENCIES" >> "$GITHUB_ENV"
           cat >> "$GITHUB_ENV" << 'EOF'
-          TEMP_PATH=${{runner.temp}}/report_check
-          REPORTS_PATH=${{runner.temp}}/reports_dir
           CHECK_NAME=ClickHouse build check (actions)
+          REPORTS_PATH=${{runner.temp}}/reports_dir
+          TEMP_PATH=${{runner.temp}}/report_check
           EOF
       - name: Download json reports
         uses: actions/download-artifact@v2
@@ -360,7 +365,7 @@ jobs:
           sudo rm -fr "$TEMP_PATH"
           mkdir -p "$TEMP_PATH"
           cd "$GITHUB_WORKSPACE/tests/ci"
-          python3 build_report_check.py "$CHECK_NAME"
+          python3 build_report_check.py "$CHECK_NAME" "$DEPENDENCIES"
       - name: Cleanup
         if: always()
         run: |

.github/workflows/master.yml

@@ -947,6 +947,34 @@ jobs:
           docker rm -f "$(docker ps -a -q)" ||:
           sudo rm -fr "$TEMP_PATH" "$CACHES_PATH"
 ############################################################################################
+##################################### Docker images  #######################################
+############################################################################################
+  DockerServerImages:
+    needs:
+      - BuilderDebRelease
+      - BuilderDebAarch64
+    runs-on: [self-hosted, style-checker]
+    steps:
+      - name: Clear repository
+        run: |
+          sudo rm -fr "$GITHUB_WORKSPACE" && mkdir "$GITHUB_WORKSPACE"
+      - name: Check out repository code
+        uses: actions/checkout@v2
+        with:
+          fetch-depth: 0  # otherwise we will have no version info
+      - name: Check docker clickhouse/clickhouse-server building
+        run: |
+          cd "$GITHUB_WORKSPACE/tests/ci"
+          python3 docker_server.py --release-type head
+          python3 docker_server.py --release-type head --no-ubuntu \
+            --image-repo clickhouse/clickhouse-keeper --image-path docker/keeper
+      - name: Cleanup
+        if: always()
+        run: |
+          docker kill "$(docker ps -q)" ||:
+          docker rm -f "$(docker ps -a -q)" ||:
+          sudo rm -fr "$TEMP_PATH"
+############################################################################################
 ##################################### BUILD REPORTER #######################################
 ############################################################################################
   BuilderReport:
@@ -964,10 +992,16 @@ jobs:
     steps:
       - name: Set envs
         run: |
+          DEPENDENCIES=$(cat << 'EOF' | jq '. | length'
+          ${{ toJSON(needs) }}
+          EOF
+          )
+          echo "DEPENDENCIES=$DEPENDENCIES" >> "$GITHUB_ENV"
           cat >> "$GITHUB_ENV" << 'EOF'
-          TEMP_PATH=${{runner.temp}}/report_check
-          REPORTS_PATH=${{runner.temp}}/reports_dir
           CHECK_NAME=ClickHouse build check (actions)
+          REPORTS_PATH=${{runner.temp}}/reports_dir
+          REPORTS_PATH=${{runner.temp}}/reports_dir
+          TEMP_PATH=${{runner.temp}}/report_check
           EOF
       - name: Download json reports
         uses: actions/download-artifact@v2
@@ -983,7 +1017,7 @@ jobs:
           sudo rm -fr "$TEMP_PATH"
           mkdir -p "$TEMP_PATH"
           cd "$GITHUB_WORKSPACE/tests/ci"
-          python3 build_report_check.py "$CHECK_NAME"
+          python3 build_report_check.py "$CHECK_NAME" "$DEPENDENCIES"
       - name: Cleanup
         if: always()
         run: |

.github/workflows/nightly.yml

@@ -72,3 +72,53 @@ jobs:
         with:
           name: changed_images
           path: ${{ runner.temp }}/changed_images.json
+  BuilderCoverity:
+    needs: DockerHubPush
+    runs-on: [self-hosted, builder]
+    steps:
+      - name: Set envs
+        run: |
+          cat >> "$GITHUB_ENV" << 'EOF'
+          BUILD_NAME=coverity
+          CACHES_PATH=${{runner.temp}}/../ccaches
+          CHECK_NAME=ClickHouse build check (actions)
+          IMAGES_PATH=${{runner.temp}}/images_path
+          REPO_COPY=${{runner.temp}}/build_check/ClickHouse
+          TEMP_PATH=${{runner.temp}}/build_check
+          EOF
+          echo "COVERITY_TOKEN=${{ secrets.COVERITY_TOKEN }}" >> "$GITHUB_ENV"
+      - name: Download changed images
+        uses: actions/download-artifact@v2
+        with:
+          name: changed_images
+          path: ${{ env.IMAGES_PATH }}
+      - name: Clear repository
+        run: |
+          sudo rm -fr "$GITHUB_WORKSPACE" && mkdir "$GITHUB_WORKSPACE"
+      - name: Check out repository code
+        id: coverity-checkout
+        uses: actions/checkout@v2
+        with:
+          submodules: 'true'
+          fetch-depth: 0 # otherwise we will have no info about contributors
+      - name: Build
+        run: |
+          sudo rm -fr "$TEMP_PATH"
+          mkdir -p "$TEMP_PATH"
+          cp -r "$GITHUB_WORKSPACE" "$TEMP_PATH"
+          cd "$REPO_COPY/tests/ci" && python3 build_check.py "$CHECK_NAME" "$BUILD_NAME"
+      - name: Upload Coverity Analysis
+        if: ${{ success() || failure() }}
+        run: |
+          curl --form token="${COVERITY_TOKEN}" \
+            --form email='security+coverity@clickhouse.com' \
+            --form file="@$TEMP_PATH/$BUILD_NAME/clickhouse-scan.tgz" \
+            --form version="${GITHUB_REF#refs/heads/}-${GITHUB_SHA::6}" \
+            --form description="Nighly Scan: $(date +'%Y-%m-%dT%H:%M:%S')" \
+            https://scan.coverity.com/builds?project=ClickHouse%2FClickHouse
+      - name: Cleanup
+        if: always()
+        run: |
+          docker kill "$(docker ps -q)" ||:
+          docker rm -f "$(docker ps -a -q)" ||:
+          sudo rm -fr "$TEMP_PATH" "$CACHES_PATH"

.github/workflows/pull_request.yml

@@ -4,7 +4,7 @@ env:
   # Force the stdout and stderr streams to be unbuffered
   PYTHONUNBUFFERED: 1
 
-on: # yamllint disable-line rule:truthy
+on:  # yamllint disable-line rule:truthy
   pull_request:
     types:
       - synchronize
@@ -998,6 +998,34 @@ jobs:
           docker rm -f "$(docker ps -a -q)" ||:
           sudo rm -fr "$TEMP_PATH" "$CACHES_PATH"
 ############################################################################################
+##################################### Docker images  #######################################
+############################################################################################
+  DockerServerImages:
+    needs:
+      - BuilderDebRelease
+      - BuilderDebAarch64
+    runs-on: [self-hosted, style-checker]
+    steps:
+      - name: Clear repository
+        run: |
+          sudo rm -fr "$GITHUB_WORKSPACE" && mkdir "$GITHUB_WORKSPACE"
+      - name: Check out repository code
+        uses: actions/checkout@v2
+        with:
+          fetch-depth: 0  # otherwise we will have no version info
+      - name: Check docker clickhouse/clickhouse-server building
+        run: |
+          cd "$GITHUB_WORKSPACE/tests/ci"
+          python3 docker_server.py --release-type head --no-push
+          python3 docker_server.py --release-type head --no-push --no-ubuntu \
+            --image-repo clickhouse/clickhouse-keeper --image-path docker/keeper
+      - name: Cleanup
+        if: always()
+        run: |
+          docker kill "$(docker ps -q)" ||:
+          docker rm -f "$(docker ps -a -q)" ||:
+          sudo rm -fr "$TEMP_PATH"
+############################################################################################
 ##################################### BUILD REPORTER #######################################
 ############################################################################################
   BuilderReport:
@@ -1016,10 +1044,16 @@ jobs:
     steps:
       - name: Set envs
         run: |
+          DEPENDENCIES=$(cat << 'EOF' | jq '. | length'
+          ${{ toJSON(needs) }}
+          EOF
+          )
+          echo "DEPENDENCIES=$DEPENDENCIES" >> "$GITHUB_ENV"
           cat >> "$GITHUB_ENV" << 'EOF'
-          TEMP_PATH=${{runner.temp}}/report_check
-          REPORTS_PATH=${{runner.temp}}/reports_dir
           CHECK_NAME=ClickHouse build check (actions)
+          REPORTS_PATH=${{runner.temp}}/reports_dir
+          REPORTS_PATH=${{runner.temp}}/reports_dir
+          TEMP_PATH=${{runner.temp}}/report_check
           EOF
       - name: Download json reports
         uses: actions/download-artifact@v2
@@ -1035,7 +1069,7 @@ jobs:
           sudo rm -fr "$TEMP_PATH"
           mkdir -p "$TEMP_PATH"
           cd "$GITHUB_WORKSPACE/tests/ci"
-          python3 build_report_check.py "$CHECK_NAME"
+          python3 build_report_check.py "$CHECK_NAME" "$DEPENDENCIES"
       - name: Cleanup
         if: always()
         run: |
@@ -3138,6 +3172,7 @@ jobs:
     needs:
       - StyleCheck
       - DockerHubPush
+      - DockerServerImages
       - CheckLabels
       - BuilderReport
       - FastTest

.github/workflows/release.yml

@@ -36,3 +36,28 @@ jobs:
         overwrite: true
         tag: ${{ github.ref }}
         file_glob: true
+  ############################################################################################
+  ##################################### Docker images  #######################################
+  ############################################################################################
+  DockerServerImages:
+    runs-on: [self-hosted, style-checker]
+    steps:
+    - name: Clear repository
+      run: |
+        sudo rm -fr "$GITHUB_WORKSPACE" && mkdir "$GITHUB_WORKSPACE"
+    - name: Check out repository code
+      uses: actions/checkout@v2
+      with:
+        fetch-depth: 0  # otherwise we will have no version info
+    - name: Check docker clickhouse/clickhouse-server building
+      run: |
+        cd "$GITHUB_WORKSPACE/tests/ci"
+        python3 docker_server.py --release-type auto
+        python3 docker_server.py --release-type auto --no-ubuntu \
+          --image-repo clickhouse/clickhouse-keeper --image-path docker/keeper
+    - name: Cleanup
+      if: always()
+      run: |
+        docker kill "$(docker ps -q)" ||:
+        docker rm -f "$(docker ps -a -q)" ||:
+        sudo rm -fr "$TEMP_PATH"

.github/workflows/release_branches.yml

@@ -436,10 +436,16 @@ jobs:
     steps:
       - name: Set envs
         run: |
+          DEPENDENCIES=$(cat << 'EOF' | jq '. | length'
+          ${{ toJSON(needs) }}
+          EOF
+          )
+          echo "DEPENDENCIES=$DEPENDENCIES" >> "$GITHUB_ENV"
           cat >> "$GITHUB_ENV" << 'EOF'
-          TEMP_PATH=${{runner.temp}}/report_check
-          REPORTS_PATH=${{runner.temp}}/reports_dir
           CHECK_NAME=ClickHouse build check (actions)
+          REPORTS_PATH=${{runner.temp}}/reports_dir
+          REPORTS_PATH=${{runner.temp}}/reports_dir
+          TEMP_PATH=${{runner.temp}}/report_check
           EOF
       - name: Download json reports
         uses: actions/download-artifact@v2
@@ -455,7 +461,7 @@ jobs:
           sudo rm -fr "$TEMP_PATH"
           mkdir -p "$TEMP_PATH"
           cd "$GITHUB_WORKSPACE/tests/ci"
-          python3 build_report_check.py "$CHECK_NAME"
+          python3 build_report_check.py "$CHECK_NAME" "$DEPENDENCIES"
       - name: Cleanup
         if: always()
         run: |

CMakeLists.txt

@@ -294,14 +294,19 @@ include(cmake/cpu_features.cmake)
 # Enable it explicitly.
 set (COMPILER_FLAGS "${COMPILER_FLAGS} -fasynchronous-unwind-tables")
 
-# Reproducible builds
-# If turned `ON`, remap file source paths in debug info, predefined preprocessor macros and __builtin_FILE().
-option(ENABLE_BUILD_PATH_MAPPING "Enable remap file source paths in debug info, predefined preprocessor macros and __builtin_FILE(). It's to generate reproducible builds. See https://reproducible-builds.org/docs/build-path" ON)
+# Reproducible builds.
+if (CMAKE_BUILD_TYPE_UC STREQUAL "DEBUG")
+    set (ENABLE_BUILD_PATH_MAPPING_DEFAULT OFF)
+else ()
+    set (ENABLE_BUILD_PATH_MAPPING_DEFAULT ON)
+endif ()
+
+option (ENABLE_BUILD_PATH_MAPPING "Enable remapping of file source paths in debug info, predefined preprocessor macros, and __builtin_FILE(). It's used to generate reproducible builds. See https://reproducible-builds.org/docs/build-path" ${ENABLE_BUILD_PATH_MAPPING_DEFAULT})
 
 if (ENABLE_BUILD_PATH_MAPPING)
     set (COMPILER_FLAGS "${COMPILER_FLAGS} -ffile-prefix-map=${CMAKE_SOURCE_DIR}=.")
     set (CMAKE_ASM_FLAGS "${CMAKE_ASM_FLAGS} -ffile-prefix-map=${CMAKE_SOURCE_DIR}=.")
-endif()
+endif ()
 
 if (${CMAKE_VERSION} VERSION_LESS "3.12.4")
     # CMake < 3.12 doesn't support setting 20 as a C++ standard version.

base/loggers/Loggers.cpp

@@ -197,7 +197,6 @@ void Loggers::buildLoggers(Poco::Util::AbstractConfiguration & config, Poco::Log
 
         Poco::AutoPtr<OwnPatternFormatter> pf = new OwnPatternFormatter(color_enabled);
         Poco::AutoPtr<DB::OwnFormattingChannel> log = new DB::OwnFormattingChannel(pf, new Poco::ConsoleChannel);
-        logger.warning("Logging " + console_log_level_string + " to console");
         log->setLevel(console_log_level);
         split->addChannel(log, "console");
     }

contrib/arrow

@@ -1 +1 @@
-Subproject commit 1d9cc51daa4e7e9fc6926320ef73759818bd736e
+Subproject commit efdcd015cfdee1b6aa349c9ca227ca12c3d697f5

contrib/krb5-cmake/CMakeLists.txt

@@ -1,4 +1,4 @@
-set (ENABLE_KRB5_DEFAULT 1)
+set (ENABLE_KRB5_DEFAULT ${ENABLE_LIBRARIES})
 if (NOT CMAKE_SYSTEM_NAME MATCHES "Linux" AND NOT (CMAKE_SYSTEM_NAME MATCHES "Darwin" AND NOT CMAKE_CROSSCOMPILING))
     message (WARNING "krb5 disabled in non-Linux and non-native-Darwin environments")
     set (ENABLE_KRB5_DEFAULT 0)
@@ -16,6 +16,7 @@ if(NOT AWK_PROGRAM)
 endif()
 
 set(KRB5_SOURCE_DIR "${ClickHouse_SOURCE_DIR}/contrib/krb5/src")
+set(KRB5_ET_BIN_DIR "${CMAKE_CURRENT_BINARY_DIR}/include_private")
 
 set(ALL_SRCS
     "${KRB5_SOURCE_DIR}/util/et/et_name.c"
@@ -90,7 +91,6 @@ set(ALL_SRCS
     "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/get_tkt_flags.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/set_allowable_enctypes.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/k5sealiov.c"
-    "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/gssapi_err_krb5.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/canon_name.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/inq_cred.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/export_sec_context.c"
@@ -143,11 +143,12 @@ set(ALL_SRCS
     "${KRB5_SOURCE_DIR}/lib/gssapi/generic/util_buffer_set.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/generic/util_set.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/generic/util_token.c"
-    "${KRB5_SOURCE_DIR}/lib/gssapi/generic/gssapi_err_generic.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/generic/disp_major_status.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/generic/util_seqstate.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/generic/util_errmap.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/generic/rel_buffer.c"
+    "${KRB5_ET_BIN_DIR}/lib/gssapi/krb5/gssapi_err_krb5.c"
+    "${KRB5_ET_BIN_DIR}/lib/gssapi/generic/gssapi_err_generic.c"
 
     "${KRB5_SOURCE_DIR}/lib/gssapi/spnego/spnego_mech.c"
     "${KRB5_SOURCE_DIR}/lib/gssapi/spnego/negoex_util.c"
@@ -256,8 +257,8 @@ set(ALL_SRCS
     "${KRB5_SOURCE_DIR}/util/profile/prof_parse.c"
     "${KRB5_SOURCE_DIR}/util/profile/prof_get.c"
     "${KRB5_SOURCE_DIR}/util/profile/prof_set.c"
-    "${KRB5_SOURCE_DIR}/util/profile/prof_err.c"
     "${KRB5_SOURCE_DIR}/util/profile/prof_init.c"
+    "${KRB5_ET_BIN_DIR}/util/profile/prof_err.c"
     "${KRB5_SOURCE_DIR}/lib/krb5/krb/fwd_tgt.c"
     "${KRB5_SOURCE_DIR}/lib/krb5/krb/conv_creds.c"
     "${KRB5_SOURCE_DIR}/lib/krb5/krb/fast.c"
@@ -450,13 +451,12 @@ set(ALL_SRCS
 
 
 
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/k5e1_err.c"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/kdb5_err.c"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/asn1_err.c"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/krb5_err.c"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/krb524_err.c"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/kv5m_err.c"
-
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/k5e1_err.c"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/kdb5_err.c"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/asn1_err.c"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/krb5_err.c"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/krb524_err.c"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/kv5m_err.c"
 
 
     "${KRB5_SOURCE_DIR}/lib/krb5/rcache/rc_base.c"
@@ -473,7 +473,7 @@ set(ALL_SRCS
 )
 
 add_custom_command(
-    OUTPUT "${KRB5_SOURCE_DIR}/util/et/compile_et"
+    OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/compile_et"
     COMMAND /bin/sh
     ./config_script
     ./compile_et.sh
@@ -481,50 +481,17 @@ add_custom_command(
     ${AWK_PROGRAM}
     sed
     >
-    compile_et
+    ${CMAKE_CURRENT_BINARY_DIR}/compile_et
     DEPENDS "${KRB5_SOURCE_DIR}/util/et/compile_et.sh" "${KRB5_SOURCE_DIR}/util/et/config_script"
     WORKING_DIRECTORY "${KRB5_SOURCE_DIR}/util/et"
 )
 
-file(GLOB_RECURSE ET_FILES
-    "${KRB5_SOURCE_DIR}/*.et"
-)
-
-function(preprocess_et out_var)
-  set(result)
-  foreach(in_f ${ARGN})
-    string(REPLACE
-        .et
-        .c
-        F_C
-        ${in_f}
-    )
-    string(REPLACE
-        .et
-        .h
-        F_H
-        ${in_f}
-    )
-
-    get_filename_component(ET_PATH ${in_f} DIRECTORY)
-
-    add_custom_command(OUTPUT ${F_C} ${F_H}
-      COMMAND perl "${KRB5_SOURCE_DIR}/util/et/compile_et" -d "${KRB5_SOURCE_DIR}/util/et" ${in_f}
-      DEPENDS ${in_f} "${KRB5_SOURCE_DIR}/util/et/compile_et"
-      WORKING_DIRECTORY ${ET_PATH}
-      VERBATIM
-      )
-    list(APPEND result ${F_C})
-  endforeach()
-  set(${out_var} "${result}" PARENT_SCOPE)
-endfunction()
-
 add_custom_command(
-    OUTPUT "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/error_map.h"
+    OUTPUT "${KRB5_ET_BIN_DIR}/error_map.h"
     COMMAND perl
     -I../../../util
     ../../../util/gen-map.pl
-    -oerror_map.h
+    -o${KRB5_ET_BIN_DIR}/error_map.h
     NAME=gsserrmap
     KEY=OM_uint32
     VALUE=char*
@@ -536,22 +503,21 @@ add_custom_command(
 
 add_custom_target(
     ERROR_MAP_H
-    DEPENDS "${KRB5_SOURCE_DIR}/lib/gssapi/krb5/error_map.h"
+    DEPENDS "${KRB5_ET_BIN_DIR}/error_map.h"
     VERBATIM
 )
 
 add_custom_command(
-    OUTPUT "${KRB5_SOURCE_DIR}/lib/gssapi/generic/errmap.h"
-    COMMAND perl -w -I../../../util  ../../../util/gen.pl bimap errmap.h NAME=mecherrmap LEFT=OM_uint32 RIGHT=struct\ mecherror LEFTPRINT=print_OM_uint32 RIGHTPRINT=mecherror_print LEFTCMP=cmp_OM_uint32 RIGHTCMP=mecherror_cmp
+    OUTPUT "${KRB5_ET_BIN_DIR}/errmap.h"
+    COMMAND perl -w -I../../../util  ../../../util/gen.pl bimap ${KRB5_ET_BIN_DIR}/errmap.h NAME=mecherrmap LEFT=OM_uint32 RIGHT=struct\ mecherror LEFTPRINT=print_OM_uint32 RIGHTPRINT=mecherror_print LEFTCMP=cmp_OM_uint32 RIGHTCMP=mecherror_cmp
     WORKING_DIRECTORY "${KRB5_SOURCE_DIR}/lib/gssapi/generic"
 )
 
 add_custom_target(
     ERRMAP_H
-    DEPENDS "${KRB5_SOURCE_DIR}/lib/gssapi/generic/errmap.h"
+    DEPENDS "${KRB5_ET_BIN_DIR}/errmap.h"
     VERBATIM
 )
-
 add_custom_target(
     KRB_5_H
     DEPENDS "${CMAKE_CURRENT_BINARY_DIR}/include/krb5/krb5.h"
@@ -567,7 +533,40 @@ add_dependencies(
     KRB_5_H
 )
 
-preprocess_et(processed_et_files ${ET_FILES})
+#
+# Generate error tables
+#
+function(preprocess_et et_path)
+    string(REPLACE .et .c F_C ${et_path})
+    string(REPLACE .et .h F_H ${et_path})
+    get_filename_component(et_dir ${et_path} DIRECTORY)
+    get_filename_component(et_name ${et_path} NAME_WLE)
+
+    add_custom_command(OUTPUT ${F_C} ${F_H} ${KRB5_ET_BIN_DIR}/${et_name}.h
+        COMMAND perl "${CMAKE_CURRENT_BINARY_DIR}/compile_et" -d "${KRB5_SOURCE_DIR}/util/et" ${et_path}
+        # for #include w/o path (via -iquote)
+        COMMAND ${CMAKE_COMMAND} -E create_symlink ${F_H} ${KRB5_ET_BIN_DIR}/${et_name}.h
+        DEPENDS ${et_path} "${CMAKE_CURRENT_BINARY_DIR}/compile_et"
+        WORKING_DIRECTORY ${et_dir}
+        VERBATIM
+    )
+endfunction()
+
+function(generate_error_tables)
+    file(GLOB_RECURSE ET_FILES "${KRB5_SOURCE_DIR}/*.et")
+    foreach(et_path ${ET_FILES})
+        string(REPLACE ${KRB5_SOURCE_DIR} ${KRB5_ET_BIN_DIR} et_bin_path ${et_path})
+        string(REPLACE / _ et_target_name ${et_path})
+        get_filename_component(et_bin_dir ${et_bin_path} DIRECTORY)
+        add_custom_command(OUTPUT ${et_bin_path}
+            COMMAND ${CMAKE_COMMAND} -E make_directory ${et_bin_dir}
+            COMMAND ${CMAKE_COMMAND} -E copy_if_different ${et_path} ${et_bin_path}
+            VERBATIM
+        )
+        preprocess_et(${et_bin_path})
+    endforeach()
+endfunction()
+generate_error_tables()
 
 if(CMAKE_SYSTEM_NAME MATCHES "Darwin")
     add_custom_command(
@@ -634,12 +633,12 @@ file(MAKE_DIRECTORY
 
 SET(KRBHDEP
     "${KRB5_SOURCE_DIR}/include/krb5/krb5.hin"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/krb5_err.h"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/k5e1_err.h"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/kdb5_err.h"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/kv5m_err.h"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/krb524_err.h"
-    "${KRB5_SOURCE_DIR}/lib/krb5/error_tables/asn1_err.h"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/krb5_err.h"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/k5e1_err.h"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/kdb5_err.h"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/kv5m_err.h"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/krb524_err.h"
+    "${KRB5_ET_BIN_DIR}/lib/krb5/error_tables/asn1_err.h"
 )
 
 # cmake < 3.18 does not have 'cat' command
@@ -656,6 +655,11 @@ target_include_directories(_krb5 SYSTEM BEFORE PUBLIC
     "${CMAKE_CURRENT_BINARY_DIR}/include"
 )
 
+target_compile_options(_krb5 PRIVATE
+    # For '#include "file.h"'
+    -iquote "${CMAKE_CURRENT_BINARY_DIR}/include_private"
+)
+
 target_include_directories(_krb5 PRIVATE
     "${CMAKE_CURRENT_BINARY_DIR}/include_private" # For autoconf.h and other generated headers.
     ${KRB5_SOURCE_DIR}

contrib/llvm-cmake/CMakeLists.txt

@@ -1,12 +1,9 @@
-# During cross-compilation in our CI we have to use llvm-tblgen and other building tools
-# tools to be build for host architecture and everything else for target architecture (e.g. AArch64)
-# Possible workaround is to use llvm-tblgen from some package...
-# But lets just enable LLVM for native builds
-if (CMAKE_CROSSCOMPILING OR SANITIZE STREQUAL "undefined")
-    set (ENABLE_EMBEDDED_COMPILER_DEFAULT OFF)
+if (APPLE OR NOT ARCH_AMD64 OR SANITIZE STREQUAL "undefined")
+   set (ENABLE_EMBEDDED_COMPILER_DEFAULT OFF)
 else()
-    set (ENABLE_EMBEDDED_COMPILER_DEFAULT ON)
+   set (ENABLE_EMBEDDED_COMPILER_DEFAULT ON)
 endif()
+
 option (ENABLE_EMBEDDED_COMPILER "Enable support for 'compile_expressions' option for query execution" ${ENABLE_EMBEDDED_COMPILER_DEFAULT})
 
 if (NOT ENABLE_EMBEDDED_COMPILER)

contrib/poco

@@ -1 +1 @@
-Subproject commit 520a90e02e3e5cb90afeae1846d161dbc508a6f1
+Subproject commit 008b16469471d55b176db181756c94e3f14dd2dc

contrib/replxx

@@ -1 +1 @@
-Subproject commit 6f0b6f151ae2a044625ae93acd19ca365fcea64d
+Subproject commit 3fd0e3c9364a589447453d9906d854ebd8d385c5

contrib/unixodbc

@@ -1 +1 @@
-Subproject commit b0ad30f7f6289c12b76f04bfb9d466374bb32168
+Subproject commit a2cd5395e8c7f7390025ec93af5bfebef3fb5fcd

debian/clickhouse-server.service

@@ -20,7 +20,7 @@ ExecStart=/usr/bin/clickhouse-server --config=/etc/clickhouse-server/config.xml
 EnvironmentFile=-/etc/default/clickhouse
 LimitCORE=infinity
 LimitNOFILE=500000
-CapabilityBoundingSet=CAP_NET_ADMIN CAP_IPC_LOCK CAP_SYS_NICE
+CapabilityBoundingSet=CAP_NET_ADMIN CAP_IPC_LOCK CAP_SYS_NICE CAP_NET_BIND_SERVICE
 
 [Install]
 # ClickHouse should not start from the rescue shell (rescue.target).

docker/docs/builder/Dockerfile

@@ -1,4 +1,3 @@
-# rebuild in #33610
 # docker build -t clickhouse/docs-builder .
 FROM ubuntu:20.04
 

docker/keeper/Dockerfile

@@ -0,0 +1,74 @@
+FROM ubuntu:20.04 AS glibc-donor
+
+ARG TARGETARCH
+RUN arch=${TARGETARCH:-amd64} \
+    && case $arch in \
+        amd64) rarch=x86_64 ;; \
+        arm64) rarch=aarch64 ;; \
+    esac \
+    && ln -s "${rarch}-linux-gnu" /lib/linux-gnu
+
+
+FROM alpine
+
+ENV LANG=en_US.UTF-8 \
+    LANGUAGE=en_US:en \
+    LC_ALL=en_US.UTF-8 \
+    TZ=UTC \
+    CLICKHOUSE_CONFIG=/etc/clickhouse-server/config.xml
+
+COPY --from=glibc-donor /lib/linux-gnu/libc.so.6 /lib/linux-gnu/libdl.so.2 /lib/linux-gnu/libm.so.6 /lib/linux-gnu/libpthread.so.0 /lib/linux-gnu/librt.so.1 /lib/linux-gnu/libnss_dns.so.2 /lib/linux-gnu/libnss_files.so.2 /lib/linux-gnu/libresolv.so.2 /lib/linux-gnu/ld-2.31.so /lib/
+COPY --from=glibc-donor /etc/nsswitch.conf /etc/
+COPY entrypoint.sh /entrypoint.sh
+
+ARG TARGETARCH
+RUN arch=${TARGETARCH:-amd64} \
+    && case $arch in \
+        amd64) mkdir -p /lib64 && ln -sf /lib/ld-2.31.so /lib64/ld-linux-x86-64.so.2 ;; \
+        arm64) ln -sf /lib/ld-2.31.so /lib/ld-linux-aarch64.so.1 ;; \
+    esac
+
+ARG REPOSITORY="https://s3.amazonaws.com/clickhouse-builds/22.4/31c367d3cd3aefd316778601ff6565119fe36682/package_release"
+ARG VERSION="22.4.1.917"
+ARG PACKAGES="clickhouse-keeper"
+
+# user/group precreated explicitly with fixed uid/gid on purpose.
+# It is especially important for rootless containers: in that case entrypoint
+# can't do chown and owners of mounted volumes should be configured externally.
+# We do that in advance at the begining of Dockerfile before any packages will be
+# installed to prevent picking those uid / gid by some unrelated software.
+# The same uid / gid (101) is used both for alpine and ubuntu.
+
+
+ARG TARGETARCH
+RUN arch=${TARGETARCH:-amd64} \
+    && for package in ${PACKAGES}; do \
+        { \
+            { echo "Get ${REPOSITORY}/${package}-${VERSION}-${arch}.tgz" \
+                && wget -c -q "${REPOSITORY}/${package}-${VERSION}-${arch}.tgz" -O "/tmp/${package}-${VERSION}-${arch}.tgz" \
+                && tar xvzf "/tmp/${package}-${VERSION}-${arch}.tgz" --strip-components=1 -C / ; \
+            } || \
+            { echo "Fallback to ${REPOSITORY}/${package}-${VERSION}.tgz" \
+                && wget -c -q "${REPOSITORY}/${package}-${VERSION}.tgz" -O "/tmp/${package}-${VERSION}.tgz" \
+                && tar xvzf "/tmp/${package}-${VERSION}.tgz" --strip-components=2 -C / ; \
+            } ; \
+        } || exit 1 \
+    ; done \
+    && rm /tmp/*.tgz /install -r \
+    && addgroup -S -g 101 clickhouse \
+    && adduser -S -h /var/lib/clickhouse -s /bin/bash -G clickhouse -g "ClickHouse keeper" -u 101 clickhouse \
+    && mkdir -p /var/lib/clickhouse /var/log/clickhouse-keeper /etc/clickhouse-keeper \
+    && chown clickhouse:clickhouse /var/lib/clickhouse \
+    && chown root:clickhouse /var/log/clickhouse-keeper \
+    && chmod +x /entrypoint.sh \
+    && apk add --no-cache su-exec bash tzdata \
+    && cp /usr/share/zoneinfo/UTC /etc/localtime \
+    && echo "UTC" > /etc/timezone \
+    && chmod ugo+Xrw -R /var/lib/clickhouse /var/log/clickhouse-keeper /etc/clickhouse-keeper
+
+
+EXPOSE 2181 10181 44444
+
+VOLUME /var/lib/clickhouse /var/log/clickhouse-keeper /etc/clickhouse-keeper
+
+ENTRYPOINT ["/entrypoint.sh"]

docker/keeper/Dockerfile.alpine

@@ -0,0 +1 @@
+Dockerfile

docker/keeper/entrypoint.sh

@@ -0,0 +1,93 @@
+#!/bin/bash
+
+set +x
+set -eo pipefail
+shopt -s nullglob
+
+DO_CHOWN=1
+if [ "${CLICKHOUSE_DO_NOT_CHOWN:-0}" = "1" ]; then
+    DO_CHOWN=0
+fi
+
+CLICKHOUSE_UID="${CLICKHOUSE_UID:-"$(id -u clickhouse)"}"
+CLICKHOUSE_GID="${CLICKHOUSE_GID:-"$(id -g clickhouse)"}"
+
+# support --user
+if [ "$(id -u)" = "0" ]; then
+    USER=$CLICKHOUSE_UID
+    GROUP=$CLICKHOUSE_GID
+    if command -v gosu &> /dev/null; then
+        gosu="gosu $USER:$GROUP"
+    elif command -v su-exec &> /dev/null; then
+        gosu="su-exec $USER:$GROUP"
+    else
+        echo "No gosu/su-exec detected!"
+        exit 1
+    fi
+else
+    USER="$(id -u)"
+    GROUP="$(id -g)"
+    gosu=""
+    DO_CHOWN=0
+fi
+
+KEEPER_CONFIG="${KEEPER_CONFIG:-/etc/clickhouse-keeper/config.yaml}"
+
+if [ -f "$KEEPER_CONFIG" ] && ! $gosu test -f "$KEEPER_CONFIG" -a -r "$KEEPER_CONFIG"; then
+    echo "Configuration file '$KEEPER_CONFIG' isn't readable by user with id '$USER'"
+    exit 1
+fi
+
+DATA_DIR="${CLICKHOUSE_DATA_DIR:-/var/lib/clickhouse}"
+LOG_DIR="${LOG_DIR:-/var/log/clickhouse-keeper}"
+LOG_PATH="${LOG_DIR}/clickhouse-keeper.log"
+ERROR_LOG_PATH="${LOG_DIR}/clickhouse-keeper.err.log"
+COORDINATION_LOG_DIR="${DATA_DIR}/coordination/log"
+COORDINATION_SNAPSHOT_DIR="${DATA_DIR}/coordination/snapshots"
+CLICKHOUSE_WATCHDOG_ENABLE=${CLICKHOUSE_WATCHDOG_ENABLE:-0}
+
+for dir in "$DATA_DIR" \
+  "$LOG_DIR" \
+  "$TMP_DIR" \
+  "$COORDINATION_LOG_DIR" \
+  "$COORDINATION_SNAPSHOT_DIR"
+do
+    # check if variable not empty
+    [ -z "$dir" ] && continue
+    # ensure directories exist
+    if ! mkdir -p "$dir"; then
+        echo "Couldn't create necessary directory: $dir"
+        exit 1
+    fi
+
+    if [ "$DO_CHOWN" = "1" ]; then
+        # ensure proper directories permissions
+        # but skip it for if directory already has proper premissions, cause recursive chown may be slow
+        if [ "$(stat -c %u "$dir")" != "$USER" ] || [ "$(stat -c %g "$dir")" != "$GROUP" ]; then
+            chown -R "$USER:$GROUP" "$dir"
+        fi
+    elif ! $gosu test -d "$dir" -a -w "$dir" -a -r "$dir"; then
+        echo "Necessary directory '$dir' isn't accessible by user with id '$USER'"
+        exit 1
+    fi
+done
+
+# if no args passed to `docker run` or first argument start with `--`, then the user is passing clickhouse-server arguments
+if [[ $# -lt 1 ]] || [[ "$1" == "--"* ]]; then
+    # Watchdog is launched by default, but does not send SIGINT to the main process,
+    # so the container can't be finished by ctrl+c
+    export CLICKHOUSE_WATCHDOG_ENABLE
+
+    cd /var/lib/clickhouse
+
+    # There is a config file. It is already tested with gosu (if it is readably by keeper user)
+    if [ -f "$KEEPER_CONFIG" ]; then
+        exec $gosu /usr/bin/clickhouse-keeper --config-file="$KEEPER_CONFIG" --log-file="$LOG_PATH" --errorlog-file="$ERROR_LOG_PATH" "$@"
+    fi
+
+    # There is no config file. Will use embedded one
+    exec $gosu /usr/bin/clickhouse-keeper --log-file="$LOG_PATH" --errorlog-file="$ERROR_LOG_PATH" "$@"
+fi
+
+# Otherwise, we assume the user want to run his own process, for example a `bash` shell to explore this image
+exec "$@"

docker/packager/binary/build.sh

@@ -25,13 +25,23 @@ read -ra CMAKE_FLAGS <<< "${CMAKE_FLAGS:-}"
 env
 cmake --debug-trycompile --verbose=1 -DCMAKE_VERBOSE_MAKEFILE=1 -LA "-DCMAKE_BUILD_TYPE=$BUILD_TYPE" "-DSANITIZE=$SANITIZER" -DENABLE_CHECK_HEAVY_BUILDS=1 "${CMAKE_FLAGS[@]}" ..
 
+if [ "coverity" == "$COMBINED_OUTPUT" ]
+then
+    mkdir -p /opt/cov-analysis
+
+    wget --post-data "token=$COVERITY_TOKEN&project=ClickHouse%2FClickHouse" -qO- https://scan.coverity.com/download/linux64 | tar xz -C /opt/cov-analysis --strip-components 1
+    export PATH=$PATH:/opt/cov-analysis/bin
+    cov-configure --config ./coverity.config --template --comptype clangcc --compiler "$CC"
+    SCAN_WRAPPER="cov-build --config ./coverity.config --dir cov-int"
+fi
+
 cache_status
 # clear cache stats
 ccache --zero-stats ||:
 
 # No quotes because I want it to expand to nothing if empty.
-# shellcheck disable=SC2086
-ninja $NINJA_FLAGS clickhouse-bundle
+# shellcheck disable=SC2086 # No quotes because I want it to expand to nothing if empty.
+$SCAN_WRAPPER ninja $NINJA_FLAGS clickhouse-bundle
 
 cache_status
 
@@ -91,6 +101,12 @@ then
     mv "$COMBINED_OUTPUT.tgz" /output
 fi
 
+if [ "coverity" == "$COMBINED_OUTPUT" ]
+then
+    tar -cv -I pigz -f "coverity-scan.tgz" cov-int
+    mv "coverity-scan.tgz" /output
+fi
+
 # Also build fuzzers if any sanitizer specified
 # if [ -n "$SANITIZER" ]
 # then

docker/packager/packager

@@ -176,6 +176,9 @@ def parse_env_variables(
     if package_type == "performance":
         result.append("COMBINED_OUTPUT=performance")
         cmake_flags.append("-DENABLE_TESTS=0")
+    elif package_type == "coverity":
+        result.append("COMBINED_OUTPUT=coverity")
+        result.append('COVERITY_TOKEN="$COVERITY_TOKEN"')
     elif split_binary:
         result.append("COMBINED_OUTPUT=shared_build")
 
@@ -262,9 +265,8 @@ if __name__ == "__main__":
     # and configs to be used for performance test.
     parser.add_argument(
         "--package-type",
-        choices=("deb", "binary", "performance"),
+        choices=["deb", "binary", "performance", "coverity"],
         required=True,
-        help="a build type",
     )
     parser.add_argument(
         "--clickhouse-repo-path",
@@ -322,7 +324,11 @@ if __name__ == "__main__":
     if not os.path.isabs(args.output_dir):
         args.output_dir = os.path.abspath(os.path.join(os.getcwd(), args.output_dir))
 
-    image_type = "binary" if args.package_type == "performance" else args.package_type
+    image_type = (
+        "binary"
+        if args.package_type in ("performance", "coverity")
+        else args.package_type
+    )
     image_name = "clickhouse/binary-builder"
 
     if not os.path.isabs(args.clickhouse_repo_path):

docker/server/.gitignore

@@ -1,2 +0,0 @@
-alpine-root/*
-tgz-packages/*

docker/server/Dockerfile

@@ -1,122 +0,0 @@
-FROM ubuntu:20.04
-
-# ARG for quick switch to a given ubuntu mirror
-ARG apt_archive="http://archive.ubuntu.com"
-RUN sed -i "s|http://archive.ubuntu.com|$apt_archive|g" /etc/apt/sources.list
-
-ARG repository="deb https://packages.clickhouse.com/deb stable main"
-ARG version=22.1.1.*
-
-# set non-empty deb_location_url url to create a docker image
-# from debs created by CI build, for example:
-# docker build . --network host --build-arg version="21.4.1.6282" --build-arg deb_location_url="https://clickhouse-builds.s3.yandex.net/21852/069cfbff388b3d478d1a16dc7060b48073f5d522/clickhouse_build_check/clang-11_relwithdebuginfo_none_bundled_unsplitted_disable_False_deb/" -t filimonovq/clickhouse-server:pr21852
-ARG deb_location_url=""
-
-# set non-empty single_binary_location_url to create docker image
-# from a single binary url (useful for non-standard builds - with sanitizers, for arm64).
-# for example (run on aarch64 server):
-# docker build . --network host --build-arg single_binary_location_url="https://builds.clickhouse.com/master/aarch64/clickhouse" -t altinity/clickhouse-server:master-testing-arm
-# note: clickhouse-odbc-bridge is not supported there.
-ARG single_binary_location_url=""
-
-# see https://github.com/moby/moby/issues/4032#issuecomment-192327844
-ARG DEBIAN_FRONTEND=noninteractive
-
-# user/group precreated explicitly with fixed uid/gid on purpose.
-# It is especially important for rootless containers: in that case entrypoint
-# can't do chown and owners of mounted volumes should be configured externally.
-# We do that in advance at the begining of Dockerfile before any packages will be
-# installed to prevent picking those uid / gid by some unrelated software.
-# The same uid / gid (101) is used both for alpine and ubuntu.
-
-# To drop privileges, we need 'su' command, that simply changes uid and gid.
-# In fact, the 'su' command from Linux is not so simple, due to inherent vulnerability in Linux:
-# https://ruderich.org/simon/notes/su-sudo-from-root-tty-hijacking
-# It has to mitigate this drawback of Linux, and to do this, 'su' command is creating it's own pseudo-terminal
-# and forwarding commands. Due to some ridiculous curcumstances, it does not work in Docker (or it does)
-# and for these reasons people are using alternatives to the 'su' command in Docker,
-# that don't mess with the terminal, don't care about closing the opened files, etc...
-# but can only be safe to drop privileges inside Docker.
-# The question - what implementation of 'su' command to use.
-# It should be a simple script doing about just two syscalls.
-# Some people tend to use 'gosu' tool that is written in Go.
-# It is not used for several reasons:
-# 1. Dependency on some foreign code in yet another programming language - does not sound alright.
-# 2. Anselmo D. Adams suggested not to use it due to false positive alarms in some undisclosed security scanners.
-
-COPY su-exec.c /su-exec.c
-
-RUN groupadd -r clickhouse --gid=101 \
-    && useradd -r -g clickhouse --uid=101 --home-dir=/var/lib/clickhouse --shell=/bin/bash clickhouse \
-    && apt-get update \
-    && apt-get install --yes --no-install-recommends \
-        apt-transport-https \
-        ca-certificates \
-        dirmngr \
-        gnupg \
-        locales \
-        wget \
-        tzdata \
-    && mkdir -p /etc/apt/sources.list.d \
-    && apt-key adv --keyserver keyserver.ubuntu.com --recv 8919F6BD2B48D754 \
-    && echo $repository > /etc/apt/sources.list.d/clickhouse.list \
-    && if [ -n "$deb_location_url" ]; then \
-            echo "installing from custom url with deb packages: $deb_location_url" \
-            rm -rf /tmp/clickhouse_debs \
-            && mkdir -p /tmp/clickhouse_debs \
-            && wget --progress=bar:force:noscroll "${deb_location_url}/clickhouse-common-static_${version}_amd64.deb" -P /tmp/clickhouse_debs \
-            && wget --progress=bar:force:noscroll "${deb_location_url}/clickhouse-client_${version}_all.deb" -P /tmp/clickhouse_debs \
-            && wget --progress=bar:force:noscroll "${deb_location_url}/clickhouse-server_${version}_all.deb" -P /tmp/clickhouse_debs \
-            && dpkg -i /tmp/clickhouse_debs/*.deb ; \
-       elif [ -n "$single_binary_location_url" ]; then \
-            echo "installing from single binary url: $single_binary_location_url" \
-            && rm -rf /tmp/clickhouse_binary \
-            && mkdir -p /tmp/clickhouse_binary \
-            && wget --progress=bar:force:noscroll "$single_binary_location_url" -O /tmp/clickhouse_binary/clickhouse \
-            && chmod +x /tmp/clickhouse_binary/clickhouse \
-            && /tmp/clickhouse_binary/clickhouse install --user "clickhouse" --group "clickhouse" ; \
-       else \
-           echo "installing from repository: $repository" \
-           && apt-get update \
-           && apt-get --yes -o "Dpkg::Options::=--force-confdef" -o "Dpkg::Options::=--force-confold" upgrade \
-           && apt-get install --allow-unauthenticated --yes --no-install-recommends \
-                clickhouse-common-static=$version \
-                clickhouse-client=$version \
-                clickhouse-server=$version ; \
-       fi \
-    && apt-get install -y --no-install-recommends tcc libc-dev && \
-        tcc /su-exec.c -o /bin/su-exec && \
-        chown root:root /bin/su-exec && \
-        chmod 0755 /bin/su-exec && \
-        rm /su-exec.c && \
-        apt-get purge -y --auto-remove tcc libc-dev libc-dev-bin libc6-dev linux-libc-dev \
-    && clickhouse-local -q 'SELECT * FROM system.build_options' \
-    && rm -rf \
-        /var/lib/apt/lists/* \
-        /var/cache/debconf \
-        /tmp/* \
-    && apt-get clean \
-    && mkdir -p /var/lib/clickhouse /var/log/clickhouse-server /etc/clickhouse-server /etc/clickhouse-client \
-    && chmod ugo+Xrw -R /var/lib/clickhouse /var/log/clickhouse-server /etc/clickhouse-server /etc/clickhouse-client
-
-# we need to allow "others" access to clickhouse folder, because docker container
-# can be started with arbitrary uid (openshift usecase)
-
-RUN locale-gen en_US.UTF-8
-ENV LANG en_US.UTF-8
-ENV LANGUAGE en_US:en
-ENV LC_ALL en_US.UTF-8
-ENV TZ UTC
-
-RUN mkdir /docker-entrypoint-initdb.d
-
-COPY docker_related_config.xml /etc/clickhouse-server/config.d/
-COPY entrypoint.sh /entrypoint.sh
-RUN chmod +x /entrypoint.sh
-
-EXPOSE 9000 8123 9009
-VOLUME /var/lib/clickhouse
-
-ENV CLICKHOUSE_CONFIG /etc/clickhouse-server/config.xml
-
-ENTRYPOINT ["/entrypoint.sh"]

docker/server/Dockerfile

@@ -0,0 +1 @@
+Dockerfile.ubuntu

docker/server/Dockerfile.alpine

@@ -1,3 +1,14 @@
+FROM ubuntu:20.04 AS glibc-donor
+ARG TARGETARCH
+
+RUN arch=${TARGETARCH:-amd64} \
+    && case $arch in \
+        amd64) rarch=x86_64 ;; \
+        arm64) rarch=aarch64 ;; \
+    esac \
+    && ln -s "${rarch}-linux-gnu" /lib/linux-gnu
+
+
 FROM alpine
 
 ENV LANG=en_US.UTF-8 \
@@ -6,7 +17,24 @@ ENV LANG=en_US.UTF-8 \
     TZ=UTC \
     CLICKHOUSE_CONFIG=/etc/clickhouse-server/config.xml
 
-COPY alpine-root/ /
+COPY --from=glibc-donor /lib/linux-gnu/libc.so.6 /lib/linux-gnu/libdl.so.2 /lib/linux-gnu/libm.so.6 /lib/linux-gnu/libpthread.so.0 /lib/linux-gnu/librt.so.1 /lib/linux-gnu/libnss_dns.so.2 /lib/linux-gnu/libnss_files.so.2 /lib/linux-gnu/libresolv.so.2 /lib/linux-gnu/ld-2.31.so /lib/
+COPY --from=glibc-donor /etc/nsswitch.conf /etc/
+COPY docker_related_config.xml /etc/clickhouse-server/config.d/
+COPY entrypoint.sh /entrypoint.sh
+
+ARG TARGETARCH
+
+RUN arch=${TARGETARCH:-amd64} \
+    && case $arch in \
+        amd64) mkdir -p /lib64 && ln -sf /lib/ld-2.31.so /lib64/ld-linux-x86-64.so.2 ;; \
+        arm64) ln -sf /lib/ld-2.31.so /lib/ld-linux-aarch64.so.1 ;; \
+    esac
+
+# lts / testing / prestable / etc
+ARG REPO_CHANNEL="stable"
+ARG REPOSITORY="https://packages.clickhouse.com/tgz/${REPO_CHANNEL}"
+ARG VERSION="20.9.3.45"
+ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static"
 
 # user/group precreated explicitly with fixed uid/gid on purpose.
 # It is especially important for rootless containers: in that case entrypoint
@@ -15,9 +43,23 @@ COPY alpine-root/ /
 # installed to prevent picking those uid / gid by some unrelated software.
 # The same uid / gid (101) is used both for alpine and ubuntu.
 
-RUN addgroup -S -g 101 clickhouse \
+RUN arch=${TARGETARCH:-amd64} \
+    && for package in ${PACKAGES}; do \
+        { \
+            { echo "Get ${REPOSITORY}/${package}-${VERSION}-${arch}.tgz" \
+                && wget -c -q "${REPOSITORY}/${package}-${VERSION}-${arch}.tgz" -O "/tmp/${package}-${VERSION}-${arch}.tgz" \
+                && tar xvzf "/tmp/${package}-${VERSION}-${arch}.tgz" --strip-components=1 -C / ; \
+            } || \
+            { echo "Fallback to ${REPOSITORY}/${package}-${VERSION}.tgz" \
+                && wget -c -q "${REPOSITORY}/${package}-${VERSION}.tgz" -O "/tmp/${package}-${VERSION}.tgz" \
+                && tar xvzf "/tmp/${package}-${VERSION}.tgz" --strip-components=2 -C / ; \
+            } ; \
+        } || exit 1 \
+    ; done \
+    && rm /tmp/*.tgz /install -r \
+    && addgroup -S -g 101 clickhouse \
     && adduser -S -h /var/lib/clickhouse -s /bin/bash -G clickhouse -g "ClickHouse server" -u 101 clickhouse \
-    && mkdir -p /var/lib/clickhouse /var/log/clickhouse-server /etc/clickhouse-server /etc/clickhouse-client \
+    && mkdir -p /var/lib/clickhouse /var/log/clickhouse-server /etc/clickhouse-server/config.d /etc/clickhouse-server/users.d /etc/clickhouse-client /docker-entrypoint-initdb.d \
     && chown clickhouse:clickhouse /var/lib/clickhouse \
     && chown root:clickhouse /var/log/clickhouse-server \
     && chmod +x /entrypoint.sh \

docker/server/Dockerfile.ubuntu

@@ -0,0 +1,129 @@
+FROM ubuntu:20.04
+
+# see https://github.com/moby/moby/issues/4032#issuecomment-192327844
+ARG DEBIAN_FRONTEND=noninteractive
+
+COPY su-exec.c /su-exec.c
+
+# ARG for quick switch to a given ubuntu mirror
+ARG apt_archive="http://archive.ubuntu.com"
+RUN sed -i "s|http://archive.ubuntu.com|${apt_archive}|g" /etc/apt/sources.list \
+    && groupadd -r clickhouse --gid=101 \
+    && useradd -r -g clickhouse --uid=101 --home-dir=/var/lib/clickhouse --shell=/bin/bash clickhouse \
+    && apt-get update \
+    && apt-get install --yes --no-install-recommends \
+        apt-transport-https \
+        ca-certificates \
+        dirmngr \
+        gnupg \
+        locales \
+        wget \
+        tzdata \
+    && apt-get install -y --no-install-recommends tcc libc-dev && \
+        tcc /su-exec.c -o /bin/su-exec && \
+        chown root:root /bin/su-exec && \
+        chmod 0755 /bin/su-exec && \
+        rm /su-exec.c && \
+        apt-get purge -y --auto-remove tcc libc-dev libc-dev-bin libc6-dev linux-libc-dev \
+    && apt-get clean
+
+ARG REPO_CHANNEL="stable"
+ARG REPOSITORY="deb https://packages.clickhouse.com/deb ${REPO_CHANNEL} main"
+ARG VERSION=22.1.1.*
+ARG PACKAGES="clickhouse-client clickhouse-server clickhouse-common-static"
+
+# set non-empty deb_location_url url to create a docker image
+# from debs created by CI build, for example:
+# docker build . --network host --build-arg version="21.4.1.6282" --build-arg deb_location_url="https://clickhouse-builds.s3.yandex.net/21852/069cfbff388b3d478d1a16dc7060b48073f5d522/clickhouse_build_check/clang-11_relwithdebuginfo_none_bundled_unsplitted_disable_False_deb/" -t filimonovq/clickhouse-server:pr21852
+ARG deb_location_url=""
+
+# set non-empty single_binary_location_url to create docker image
+# from a single binary url (useful for non-standard builds - with sanitizers, for arm64).
+# for example (run on aarch64 server):
+# docker build . --network host --build-arg single_binary_location_url="https://builds.clickhouse.com/master/aarch64/clickhouse" -t altinity/clickhouse-server:master-testing-arm
+# note: clickhouse-odbc-bridge is not supported there.
+ARG single_binary_location_url=""
+
+# user/group precreated explicitly with fixed uid/gid on purpose.
+# It is especially important for rootless containers: in that case entrypoint
+# can't do chown and owners of mounted volumes should be configured externally.
+# We do that in advance at the begining of Dockerfile before any packages will be
+# installed to prevent picking those uid / gid by some unrelated software.
+# The same uid / gid (101) is used both for alpine and ubuntu.
+
+# To drop privileges, we need 'su' command, that simply changes uid and gid.
+# In fact, the 'su' command from Linux is not so simple, due to inherent vulnerability in Linux:
+# https://ruderich.org/simon/notes/su-sudo-from-root-tty-hijacking
+# It has to mitigate this drawback of Linux, and to do this, 'su' command is creating it's own pseudo-terminal
+# and forwarding commands. Due to some ridiculous curcumstances, it does not work in Docker (or it does)
+# and for these reasons people are using alternatives to the 'su' command in Docker,
+# that don't mess with the terminal, don't care about closing the opened files, etc...
+# but can only be safe to drop privileges inside Docker.
+# The question - what implementation of 'su' command to use.
+# It should be a simple script doing about just two syscalls.
+# Some people tend to use 'gosu' tool that is written in Go.
+# It is not used for several reasons:
+# 1. Dependency on some foreign code in yet another programming language - does not sound alright.
+# 2. Anselmo D. Adams suggested not to use it due to false positive alarms in some undisclosed security scanners.
+
+ARG TARGETARCH
+
+RUN arch=${TARGETARCH:-amd64} \
+    && if [ -n "${deb_location_url}" ]; then \
+        echo "installing from custom url with deb packages: ${deb_location_url}" \
+        rm -rf /tmp/clickhouse_debs \
+        && mkdir -p /tmp/clickhouse_debs \
+        && for package in ${PACKAGES}; do \
+            { wget --progress=bar:force:noscroll "${deb_location_url}/${package}_${VERSION}_${arch}.deb" -P /tmp/clickhouse_debs || \
+                wget --progress=bar:force:noscroll "${deb_location_url}/${package}_${VERSION}_all.deb" -P /tmp/clickhouse_debs ; } \
+            || exit 1 \
+        ; done \
+        && dpkg -i /tmp/clickhouse_debs/*.deb ; \
+    elif [ -n "${single_binary_location_url}" ]; then \
+        echo "installing from single binary url: ${single_binary_location_url}" \
+        && rm -rf /tmp/clickhouse_binary \
+        && mkdir -p /tmp/clickhouse_binary \
+        && wget --progress=bar:force:noscroll "${single_binary_location_url}" -O /tmp/clickhouse_binary/clickhouse \
+        && chmod +x /tmp/clickhouse_binary/clickhouse \
+        && /tmp/clickhouse_binary/clickhouse install --user "clickhouse" --group "clickhouse" ; \
+    else \
+        mkdir -p /etc/apt/sources.list.d \
+        && apt-key adv --keyserver keyserver.ubuntu.com --recv 8919F6BD2B48D754 \
+        && echo ${REPOSITORY} > /etc/apt/sources.list.d/clickhouse.list \
+        && echo "installing from repository: ${REPOSITORY}" \
+        && apt-get update \
+        && apt-get --yes -o "Dpkg::Options::=--force-confdef" -o "Dpkg::Options::=--force-confold" upgrade \
+        && for package in ${PACKAGES}; do \
+            packages="${packages} ${package}=${VERSION}" \
+        ; done \
+        && apt-get install --allow-unauthenticated --yes --no-install-recommends ${packages} || exit 1 \
+    ; fi \
+    && clickhouse-local -q 'SELECT * FROM system.build_options' \
+    && rm -rf \
+        /var/lib/apt/lists/* \
+        /var/cache/debconf \
+        /tmp/* \
+    && mkdir -p /var/lib/clickhouse /var/log/clickhouse-server /etc/clickhouse-server /etc/clickhouse-client \
+    && chmod ugo+Xrw -R /var/lib/clickhouse /var/log/clickhouse-server /etc/clickhouse-server /etc/clickhouse-client
+
+# we need to allow "others" access to clickhouse folder, because docker container
+# can be started with arbitrary uid (openshift usecase)
+
+RUN locale-gen en_US.UTF-8
+ENV LANG en_US.UTF-8
+ENV LANGUAGE en_US:en
+ENV LC_ALL en_US.UTF-8
+ENV TZ UTC
+
+RUN mkdir /docker-entrypoint-initdb.d
+
+COPY docker_related_config.xml /etc/clickhouse-server/config.d/
+COPY entrypoint.sh /entrypoint.sh
+RUN chmod +x /entrypoint.sh
+
+EXPOSE 9000 8123 9009
+VOLUME /var/lib/clickhouse
+
+ENV CLICKHOUSE_CONFIG /etc/clickhouse-server/config.xml
+
+ENTRYPOINT ["/entrypoint.sh"]

docker/server/alpine-build.sh

@@ -1,63 +0,0 @@
-#!/bin/bash
-set -x
-
-REPO_CHANNEL="${REPO_CHANNEL:-stable}" # lts / testing / prestable / etc
-REPO_URL="${REPO_URL:-"https://repo.yandex.ru/clickhouse/tgz/${REPO_CHANNEL}"}"
-VERSION="${VERSION:-20.9.3.45}"
-DOCKER_IMAGE="${DOCKER_IMAGE:-clickhouse/clickhouse-server}"
-
-# where original files live
-DOCKER_BUILD_FOLDER="${BASH_SOURCE%/*}"
-
-# we will create root for our image here
-CONTAINER_ROOT_FOLDER="${DOCKER_BUILD_FOLDER}/alpine-root"
-
-# clean up the root from old runs, it's reconstructed each time
-rm -rf "$CONTAINER_ROOT_FOLDER"
-mkdir -p "$CONTAINER_ROOT_FOLDER"
-
-# where to put downloaded tgz
-TGZ_PACKAGES_FOLDER="${DOCKER_BUILD_FOLDER}/tgz-packages"
-mkdir -p "$TGZ_PACKAGES_FOLDER"
-
-PACKAGES=( "clickhouse-client" "clickhouse-server" "clickhouse-common-static" )
-
-# download tars from the repo
-for package in "${PACKAGES[@]}"
-do
-    wget -c -q --show-progress "${REPO_URL}/${package}-${VERSION}.tgz" -O "${TGZ_PACKAGES_FOLDER}/${package}-${VERSION}.tgz"
-done
-
-# unpack tars
-for package in "${PACKAGES[@]}"
-do
-    tar xvzf "${TGZ_PACKAGES_FOLDER}/${package}-${VERSION}.tgz" --strip-components=2 -C "$CONTAINER_ROOT_FOLDER"
-done
-
-# prepare few more folders
-mkdir -p "${CONTAINER_ROOT_FOLDER}/etc/clickhouse-server/users.d" \
-         "${CONTAINER_ROOT_FOLDER}/etc/clickhouse-server/config.d" \
-         "${CONTAINER_ROOT_FOLDER}/var/log/clickhouse-server" \
-         "${CONTAINER_ROOT_FOLDER}/var/lib/clickhouse" \
-         "${CONTAINER_ROOT_FOLDER}/docker-entrypoint-initdb.d" \
-         "${CONTAINER_ROOT_FOLDER}/lib64"
-
-cp "${DOCKER_BUILD_FOLDER}/docker_related_config.xml" "${CONTAINER_ROOT_FOLDER}/etc/clickhouse-server/config.d/"
-cp "${DOCKER_BUILD_FOLDER}/entrypoint.sh"             "${CONTAINER_ROOT_FOLDER}/entrypoint.sh"
-
-## get glibc components from ubuntu 20.04 and put them to expected place
-docker pull ubuntu:20.04
-ubuntu20image=$(docker create --rm ubuntu:20.04)
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/libc.so.6       "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/libdl.so.2      "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/libm.so.6       "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/libpthread.so.0 "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/librt.so.1      "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/libnss_dns.so.2 "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/libnss_files.so.2 "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib/x86_64-linux-gnu/libresolv.so.2  "${CONTAINER_ROOT_FOLDER}/lib"
-docker cp -L "${ubuntu20image}":/lib64/ld-linux-x86-64.so.2           "${CONTAINER_ROOT_FOLDER}/lib64"
-docker cp -L "${ubuntu20image}":/etc/nsswitch.conf                    "${CONTAINER_ROOT_FOLDER}/etc"
-
-docker build "$DOCKER_BUILD_FOLDER" -f Dockerfile.alpine -t "${DOCKER_IMAGE}:${VERSION}-alpine" --pull
-rm -rf "$CONTAINER_ROOT_FOLDER"

docker/server/local.Dockerfile

@@ -1,47 +0,0 @@
-# Since right now we can't set volumes to the docker during build, we split building container in stages:
-# 1. build base container
-# 2. run base conatiner with mounted volumes
-# 3. commit container as image
-# 4. build final container atop that image
-# Middle steps are performed by the bash script.
-
-FROM ubuntu:18.04 as clickhouse-server-base
-ARG gosu_ver=1.14
-
-VOLUME /packages/
-
-# update to allow installing dependencies of clickhouse automatically
-RUN apt update; \
-    DEBIAN_FRONTEND=noninteractive \
-    apt install -y locales;
-
-ADD https://github.com/tianon/gosu/releases/download/${gosu_ver}/gosu-amd64 /bin/gosu
-
-RUN locale-gen en_US.UTF-8
-ENV LANG en_US.UTF-8
-ENV LANGUAGE en_US:en
-ENV LC_ALL en_US.UTF-8
-
-# installing via apt to simulate real-world scenario, where user installs deb package and all it's dependecies automatically.
-CMD DEBIAN_FRONTEND=noninteractive \
-    apt install -y \
-        /packages/clickhouse-common-static_*.deb \
-        /packages/clickhouse-server_*.deb ;
-
-FROM clickhouse-server-base:postinstall as clickhouse-server
-
-RUN mkdir /docker-entrypoint-initdb.d
-
-COPY docker_related_config.xml /etc/clickhouse-server/config.d/
-COPY entrypoint.sh /entrypoint.sh
-
-RUN chmod +x \
-    /entrypoint.sh \
-    /bin/gosu
-
-EXPOSE 9000 8123 9009
-VOLUME /var/lib/clickhouse
-
-ENV CLICKHOUSE_CONFIG /etc/clickhouse-server/config.xml
-
-ENTRYPOINT ["/entrypoint.sh"]

docker/test/fuzzer/run-fuzzer.sh

@@ -226,7 +226,6 @@ quit
         --receive_data_timeout_ms=10000 \
         --stacktrace \
         --query-fuzzer-runs=1000 \
-        --testmode \
         --queries-file $(ls -1 ch/tests/queries/0_stateless/*.sql | sort -R) \
         $NEW_TESTS_OPT \
         > >(tail -n 100000 > fuzzer.log) \

docker/test/integration/mysql_js_client/Dockerfile

@@ -1,8 +1,10 @@
 # docker build -t clickhouse/mysql-js-client .
 # MySQL JavaScript client docker container
 
-FROM node:8
+FROM node:16.14.2
+
+WORKDIR /usr/app
 
 RUN npm install mysql
 
-COPY ./test.js test.js
+COPY ./test.js ./test.js

docker/test/stateful/run.sh

@@ -96,7 +96,7 @@ else
     clickhouse-client --query "RENAME TABLE datasets.hits_v1 TO test.hits"
     clickhouse-client --query "RENAME TABLE datasets.visits_v1 TO test.visits"
     clickhouse-client --query "CREATE TABLE test.hits_s3  (WatchID UInt64, JavaEnable UInt8, Title String, GoodEvent Int16, EventTime DateTime, EventDate Date, CounterID UInt32, ClientIP UInt32, ClientIP6 FixedString(16), RegionID UInt32, UserID UInt64, CounterClass Int8, OS UInt8, UserAgent UInt8, URL String, Referer String, URLDomain String, RefererDomain String, Refresh UInt8, IsRobot UInt8, RefererCategories Array(UInt16), URLCategories Array(UInt16), URLRegions Array(UInt32), RefererRegions Array(UInt32), ResolutionWidth UInt16, ResolutionHeight UInt16, ResolutionDepth UInt8, FlashMajor UInt8, FlashMinor UInt8, FlashMinor2 String, NetMajor UInt8, NetMinor UInt8, UserAgentMajor UInt16, UserAgentMinor FixedString(2), CookieEnable UInt8, JavascriptEnable UInt8, IsMobile UInt8, MobilePhone UInt8, MobilePhoneModel String, Params String, IPNetworkID UInt32, TraficSourceID Int8, SearchEngineID UInt16, SearchPhrase String, AdvEngineID UInt8, IsArtifical UInt8, WindowClientWidth UInt16, WindowClientHeight UInt16, ClientTimeZone Int16, ClientEventTime DateTime, SilverlightVersion1 UInt8, SilverlightVersion2 UInt8, SilverlightVersion3 UInt32, SilverlightVersion4 UInt16, PageCharset String, CodeVersion UInt32, IsLink UInt8, IsDownload UInt8, IsNotBounce UInt8, FUniqID UInt64, HID UInt32, IsOldCounter UInt8, IsEvent UInt8, IsParameter UInt8, DontCountHits UInt8, WithHash UInt8, HitColor FixedString(1), UTCEventTime DateTime, Age UInt8, Sex UInt8, Income UInt8, Interests UInt16, Robotness UInt8, GeneralInterests Array(UInt16), RemoteIP UInt32, RemoteIP6 FixedString(16), WindowName Int32, OpenerName Int32, HistoryLength Int16, BrowserLanguage FixedString(2), BrowserCountry FixedString(2), SocialNetwork String, SocialAction String, HTTPError UInt16, SendTiming Int32, DNSTiming Int32, ConnectTiming Int32, ResponseStartTiming Int32, ResponseEndTiming Int32, FetchTiming Int32, RedirectTiming Int32, DOMInteractiveTiming Int32, DOMContentLoadedTiming Int32, DOMCompleteTiming Int32, LoadEventStartTiming Int32, LoadEventEndTiming Int32, NSToDOMContentLoadedTiming Int32, FirstPaintTiming Int32, RedirectCount Int8, SocialSourceNetworkID UInt8, SocialSourcePage String, ParamPrice Int64, ParamOrderID String, ParamCurrency FixedString(3), ParamCurrencyID UInt16, GoalsReached Array(UInt32), OpenstatServiceName String, OpenstatCampaignID String, OpenstatAdID String, OpenstatSourceID String, UTMSource String, UTMMedium String, UTMCampaign String, UTMContent String, UTMTerm String, FromTag String, HasGCLID UInt8, RefererHash UInt64, URLHash UInt64, CLID UInt32, YCLID UInt64, ShareService String, ShareURL String, ShareTitle String, ParsedParams Nested(Key1 String, Key2 String, Key3 String, Key4 String, Key5 String, ValueDouble Float64), IslandID FixedString(16), RequestNum UInt32, RequestTry UInt8) ENGINE = MergeTree() PARTITION BY toYYYYMM(EventDate) ORDER BY (CounterID, EventDate, intHash32(UserID)) SAMPLE BY intHash32(UserID) SETTINGS index_granularity = 8192, storage_policy='s3_cache'"
-    clickhouse-client --query "INSERT INTO test.hits_s3 SELECT * FROM test.hits"
+    clickhouse-client --query "INSERT INTO test.hits_s3 SELECT * FROM test.hits SETTINGS enable_filesystem_cache_on_write_operations=0"
 fi
 
 clickhouse-client --query "SHOW TABLES FROM test"

docker/test/stress/run.sh

@@ -106,17 +106,6 @@ function stop()
 
 function start()
 {
-    # Rename existing log file - it will be more convenient to read separate files for separate server runs.
-    if [ -f '/var/log/clickhouse-server/clickhouse-server.log' ]
-    then
-        log_file_counter=1
-        while [ -f "/var/log/clickhouse-server/clickhouse-server.log.${log_file_counter}" ]
-        do
-            log_file_counter=$((log_file_counter + 1))
-        done
-        mv '/var/log/clickhouse-server/clickhouse-server.log' "/var/log/clickhouse-server/clickhouse-server.log.${log_file_counter}"
-    fi
-
     counter=0
     until clickhouse-client --query "SELECT 1"
     do
@@ -190,6 +179,8 @@ clickhouse-client --query "ATTACH DATABASE IF NOT EXISTS datasets ENGINE = Ordin
 clickhouse-client --query "CREATE DATABASE IF NOT EXISTS test"
 
 stop
+mv /var/log/clickhouse-server/clickhouse-server.log /var/log/clickhouse-server/clickhouse-server.initial.log
+
 start
 
 clickhouse-client --query "SHOW TABLES FROM datasets"
@@ -205,6 +196,8 @@ clickhouse-client --query "SHOW TABLES FROM test"
     || echo -e 'Test script failed\tFAIL' >> /test_output/test_results.tsv
 
 stop
+mv /var/log/clickhouse-server/clickhouse-server.log /var/log/clickhouse-server/clickhouse-server.stress.log
+
 start
 
 clickhouse-client --query "SELECT 'Server successfully started', 'OK'" >> /test_output/test_results.tsv \
@@ -263,10 +256,12 @@ mkdir previous_release_package_folder
 clickhouse-client --query="SELECT version()" | ./download_previous_release && echo -e 'Download script exit code\tOK' >> /test_output/test_results.tsv \
     || echo -e 'Download script failed\tFAIL' >> /test_output/test_results.tsv
 
+stop
+mv /var/log/clickhouse-server/clickhouse-server.log /var/log/clickhouse-server/clickhouse-server.clean.log
+
 if [ "$(ls -A previous_release_package_folder/clickhouse-common-static_*.deb && ls -A previous_release_package_folder/clickhouse-server_*.deb)" ]
 then
     echo -e "Successfully downloaded previous release packets\tOK" >> /test_output/test_results.tsv
-    stop
 
     # Uninstall current packages
     dpkg --remove clickhouse-client
@@ -289,7 +284,7 @@ then
     install_packages package_folder
 
     mkdir tmp_stress_output
-    
+
     ./stress --backward-compatibility-check --output-folder tmp_stress_output --global-time-limit=1200 \
         && echo -e 'Backward compatibility check: Test script exit code\tOK' >> /test_output/test_results.tsv \
         || echo -e 'Backward compatibility check: Test script failed\tFAIL' >> /test_output/test_results.tsv
@@ -297,8 +292,9 @@ then
 
     clickhouse-client --query="SELECT 'Tables count:', count() FROM system.tables"
 
-    stop    
-    
+    stop
+    mv /var/log/clickhouse-server/clickhouse-server.log /var/log/clickhouse-server/clickhouse-server.backward.stress.log
+
     # Start new server
     configure
     start 500
@@ -310,8 +306,9 @@ then
 
     # Let the server run for a while before checking log.
     sleep 60
-    
+
     stop
+    mv /var/log/clickhouse-server/clickhouse-server.log /var/log/clickhouse-server/clickhouse-server.backward.clean.log
 
     # Error messages (we should ignore some errors)
     echo "Check for Error messages in server log:"
@@ -332,8 +329,8 @@ then
                -e "Code: 1000, e.code() = 111, Connection refused" \
                -e "UNFINISHED" \
                -e "Renaming unexpected part" \
-        /var/log/clickhouse-server/clickhouse-server.log | zgrep -Fa "<Error>" > /test_output/bc_check_error_messages.txt \
-        && echo -e 'Backward compatibility check: Error message in clickhouse-server.log (see bc_check_error_messages.txt)\tFAIL' >> /test_output/test_results.tsv \
+        /var/log/clickhouse-server/clickhouse-server.backward.*.log | zgrep -Fa "<Error>" > /test_output/bc_check_error_messages.txt \
+        && echo -e 'Backward compatibility check: Error message in clickhouse-server.log (see bc_check_error_messages.txt)\tOK' >> /test_output/test_results.tsv \
         || echo -e 'Backward compatibility check: No Error messages in clickhouse-server.log\tOK' >> /test_output/test_results.tsv
 
     # Remove file bc_check_error_messages.txt if it's empty
@@ -348,13 +345,13 @@ then
     rm -f /test_output/tmp
 
     # OOM
-    zgrep -Fa " <Fatal> Application: Child process was terminated by signal 9" /var/log/clickhouse-server/clickhouse-server.log > /dev/null \
-        && echo -e 'Backward compatibility check: OOM killer (or signal 9) in clickhouse-server.log\tFAIL' >> /test_output/test_results.tsv \
+    zgrep -Fa " <Fatal> Application: Child process was terminated by signal 9" /var/log/clickhouse-server/clickhouse-server.backward.*.log > /dev/null \
+        && echo -e 'Backward compatibility check: OOM killer (or signal 9) in clickhouse-server.log\tOK' >> /test_output/test_results.tsv \
         || echo -e 'Backward compatibility check: No OOM messages in clickhouse-server.log\tOK' >> /test_output/test_results.tsv
 
     # Logical errors
     echo "Check for Logical errors in server log:"
-    zgrep -Fa -A20 "Code: 49, e.displayText() = DB::Exception:" /var/log/clickhouse-server/clickhouse-server.log > /test_output/bc_check_logical_errors.txt \
+    zgrep -Fa -A20 "Code: 49, e.displayText() = DB::Exception:" /var/log/clickhouse-server/clickhouse-server.backward.*.log > /test_output/bc_check_logical_errors.txt \
         && echo -e 'Backward compatibility check: Logical error thrown (see clickhouse-server.log or bc_check_logical_errors.txt)\tFAIL' >> /test_output/test_results.tsv \
         || echo -e 'Backward compatibility check: No logical errors\tOK' >> /test_output/test_results.tsv
 
@@ -362,19 +359,18 @@ then
     [ -s /test_output/bc_check_logical_errors.txt ] || rm /test_output/bc_check_logical_errors.txt
 
     # Crash
-    zgrep -Fa "########################################" /var/log/clickhouse-server/clickhouse-server.log > /dev/null \
+    zgrep -Fa "########################################" /var/log/clickhouse-server/clickhouse-server.backward.*.log > /dev/null \
         && echo -e 'Backward compatibility check: Killed by signal (in clickhouse-server.log)\tFAIL' >> /test_output/test_results.tsv \
         || echo -e 'Backward compatibility check: Not crashed\tOK' >> /test_output/test_results.tsv
 
     # It also checks for crash without stacktrace (printed by watchdog)
     echo "Check for Fatal message in server log:"
-    zgrep -Fa " <Fatal> " /var/log/clickhouse-server/clickhouse-server.log > /test_output/bc_check_fatal_messages.txt \
-        && echo -e 'Backward compatibility check: Fatal message in clickhouse-server.log (see bc_check_fatal_messages.txt)\tFAIL' >> /test_output/test_results.tsv \
+    zgrep -Fa " <Fatal> " /var/log/clickhouse-server/clickhouse-server.backward.*.log > /test_output/bc_check_fatal_messages.txt \
+        && echo -e 'Backward compatibility check: Fatal message in clickhouse-server.log (see bc_check_fatal_messages.txt)\tOK' >> /test_output/test_results.tsv \
         || echo -e 'Backward compatibility check: No fatal messages in clickhouse-server.log\tOK' >> /test_output/test_results.tsv
 
     # Remove file bc_check_fatal_messages.txt if it's empty
     [ -s /test_output/bc_check_fatal_messages.txt ] || rm /test_output/bc_check_fatal_messages.txt
-
 else
     echo -e "Backward compatibility check: Failed to download previous release packets\tFAIL" >> /test_output/test_results.tsv
 fi

docker/test/stress/stress

@@ -10,7 +10,7 @@ import logging
 import time
 
 
-def get_options(i):
+def get_options(i, backward_compatibility_check):
     options = []
     client_options = []
     if 0 < i:
@@ -19,7 +19,7 @@ def get_options(i):
     if i % 3 == 1:
         options.append("--db-engine=Ordinary")
 
-    if i % 3 == 2:
+    if i % 3 == 2 and not backward_compatibility_check:
         options.append('''--db-engine="Replicated('/test/db/test_{}', 's1', 'r1')"'''.format(i))
         client_options.append('allow_experimental_database_replicated=1')
 
@@ -57,7 +57,7 @@ def run_func_test(cmd, output_prefix, num_processes, skip_tests_option, global_t
     pipes = []
     for i in range(0, len(output_paths)):
         f = open(output_paths[i], 'w')
-        full_command = "{} {} {} {} {}".format(cmd, get_options(i), global_time_limit_option, skip_tests_option, backward_compatibility_check_option)
+        full_command = "{} {} {} {} {}".format(cmd, get_options(i, backward_compatibility_check), global_time_limit_option, skip_tests_option, backward_compatibility_check_option)
         logging.info("Run func tests '%s'", full_command)
         p = Popen(full_command, shell=True, stdout=f, stderr=f)
         pipes.append(p)

docker/test/test_runner.sh

@@ -1,86 +0,0 @@
-#!/bin/sh
-
-set -e -x
-
-# Not sure why shellcheck complains that rc is not assigned before it is referenced.
-# shellcheck disable=SC2154
-trap 'rc=$?; echo EXITED WITH: $rc; exit $rc' EXIT
-
-# CLI option to prevent rebuilding images, just re-run tests with images leftover from previuos time
-readonly NO_REBUILD_FLAG="--no-rebuild"
-
-readonly CLICKHOUSE_DOCKER_DIR="$(realpath "${1}")"
-readonly CLICKHOUSE_PACKAGES_ARG="${2}"
-CLICKHOUSE_SERVER_IMAGE="${3}"
-
-if [ "${CLICKHOUSE_PACKAGES_ARG}" != "${NO_REBUILD_FLAG}" ]; then
-    readonly CLICKHOUSE_PACKAGES_DIR="$(realpath "${2}")" # or --no-rebuild
-fi
-
-
-# In order to allow packages directory to be anywhere, and to reduce amount of context sent to the docker daemon,
-# all images are built in multiple stages:
-# 1. build base image, install dependencies
-# 2. run image with volume mounted, install what needed from those volumes
-# 3. tag container as image
-# 4. [optional] build another image atop of tagged.
-
-# TODO: optionally mount most recent clickhouse-test and queries directory from local machine
-
-if [ "${CLICKHOUSE_PACKAGES_ARG}" != "${NO_REBUILD_FLAG}" ]; then
-    docker build --network=host \
-        -f "${CLICKHOUSE_DOCKER_DIR}/test/stateless/clickhouse-statelest-test-runner.Dockerfile" \
-        --target clickhouse-test-runner-base \
-        -t clickhouse-test-runner-base:preinstall \
-        "${CLICKHOUSE_DOCKER_DIR}/test/stateless"
-
-    docker rm -f clickhouse-test-runner-installing-packages || true
-    docker run  --network=host \
-        -v "${CLICKHOUSE_PACKAGES_DIR}:/packages" \
-        --name clickhouse-test-runner-installing-packages \
-        clickhouse-test-runner-base:preinstall
-    docker commit clickhouse-test-runner-installing-packages clickhouse-statelest-test-runner:local
-    docker rm -f clickhouse-test-runner-installing-packages || true
-fi
-
-# # Create a bind-volume to the clickhouse-test script file
-# docker volume create --driver local --opt type=none --opt device=/home/enmk/proj/ClickHouse_master/tests/clickhouse-test --opt o=bind clickhouse-test-script-volume
-# docker volume create --driver local --opt type=none --opt device=/home/enmk/proj/ClickHouse_master/tests/queries --opt o=bind clickhouse-test-queries-dir-volume
-
-# Build server image (optional) from local packages
-if [ -z "${CLICKHOUSE_SERVER_IMAGE}" ]; then
-    CLICKHOUSE_SERVER_IMAGE="clickhouse/server:local"
-
-    if [ "${CLICKHOUSE_PACKAGES_ARG}" != "${NO_REBUILD_FLAG}" ]; then
-        docker build --network=host \
-            -f "${CLICKHOUSE_DOCKER_DIR}/server/local.Dockerfile" \
-            --target clickhouse-server-base \
-            -t clickhouse-server-base:preinstall \
-            "${CLICKHOUSE_DOCKER_DIR}/server"
-
-        docker rm -f clickhouse_server_base_installing_server || true
-        docker run  --network=host -v "${CLICKHOUSE_PACKAGES_DIR}:/packages" \
-            --name clickhouse_server_base_installing_server \
-            clickhouse-server-base:preinstall
-        docker commit clickhouse_server_base_installing_server clickhouse-server-base:postinstall
-
-        docker build --network=host \
-            -f "${CLICKHOUSE_DOCKER_DIR}/server/local.Dockerfile" \
-            --target clickhouse-server \
-            -t "${CLICKHOUSE_SERVER_IMAGE}" \
-            "${CLICKHOUSE_DOCKER_DIR}/server"
-    fi
-fi
-
-docker rm -f test-runner || true
-docker-compose down
-CLICKHOUSE_SERVER_IMAGE="${CLICKHOUSE_SERVER_IMAGE}" \
-    docker-compose -f "${CLICKHOUSE_DOCKER_DIR}/test/test_runner_docker_compose.yaml" \
-    create \
-    --build --force-recreate
-
-CLICKHOUSE_SERVER_IMAGE="${CLICKHOUSE_SERVER_IMAGE}" \
-    docker-compose -f "${CLICKHOUSE_DOCKER_DIR}/test/test_runner_docker_compose.yaml" \
-    run \
-    --name test-runner \
-    test-runner

docker/test/test_runner_docker_compose.yaml

@@ -1,34 +0,0 @@
-version: "2"
-
-services:
-  clickhouse-server:
-    image: ${CLICKHOUSE_SERVER_IMAGE}
-    expose:
-      - "8123" # HTTP
-      - "9000" # TCP
-      - "9009" # HTTP-interserver
-    restart: "no"
-
-  test-runner:
-    image: clickhouse-statelest-test-runner:local
-
-    restart: "no"
-    depends_on:
-      - clickhouse-server
-    environment:
-      # these are used by clickhouse-test to point clickhouse-client to the right server
-      - CLICKHOUSE_HOST=clickhouse-server
-      - CLICKHOUSE_PORT=9009
-      - CLICKHOUSE_TEST_HOST_EXPOSED_PORT=51234
-    expose:
-      # port for any test to serve data to clickhouse-server on rare occasion (like URL-engine tables in 00646),
-      # should match value of CLICKHOUSE_TEST_HOST_EXPOSED_PORT above
-      - "51234"
-
-    # NOTE: Dev-mode: mount newest versions of the queries and clickhouse-test script into container.
-    # volumes:
-    #   - /home/enmk/proj/ClickHouse_master/tests/queries:/usr/share/clickhouse-test/queries:ro
-    #   - /home/enmk/proj/ClickHouse_master/tests/clickhouse-test:/usr/bin/clickhouse-test:ro
-
-    # String-form instead of list-form to allow multiple arguments in "${CLICKHOUSE_TEST_ARGS}"
-    entrypoint: "clickhouse-test ${CLICKHOUSE_TEST_ARGS}"

docs/en/introduction/adopters.md

@@ -43,7 +43,7 @@ toc_title: Adopters
 | <a href="https://city-mobil.ru" class="favicon">Citymobil</a> | Taxi | Analytics | — | — | [Blog Post in Russian, March 2020](https://habr.com/en/company/citymobil/blog/490660/) |
 | <a href="https://cloudflare.com" class="favicon">Cloudflare</a> | CDN | Traffic analysis | 36 servers | — | [Blog post, May 2017](https://blog.cloudflare.com/how-cloudflare-analyzes-1m-dns-queries-per-second/), [Blog post, March 2018](https://blog.cloudflare.com/http-analytics-for-6m-requests-per-second-using-clickhouse/) |
 | <a href="https://corporate.comcast.com/" class="favicon">Comcast</a> | Media | CDN Traffic Analysis | — | — | [ApacheCon 2019 Talk](https://www.youtube.com/watch?v=e9TZ6gFDjNg) |
-| <a href="https://contentsquare.com" class="favicon">ContentSquare</a> | Web analytics | Main product | — | — | [Blog post in French, November 2018](http://souslecapot.net/2018/11/21/patrick-chatain-vp-engineering-chez-contentsquare-penser-davantage-amelioration-continue-que-revolution-constante/) |
+| <a href="https://contentsquare.com" class="favicon">Contentsquare</a> | Web analytics | Main product | — | — | [Blog post in French, November 2018](http://souslecapot.net/2018/11/21/patrick-chatain-vp-engineering-chez-contentsquare-penser-davantage-amelioration-continue-que-revolution-constante/) |
 | <a href="https://coru.net/" class="favicon">Corunet</a> | Analytics | Main product | — | — | [Slides in English, April 2019](https://github.com/ClickHouse/clickhouse-presentations/blob/master/meetup21/predictive_models.pdf) |
 | <a href="https://www.creditx.com" class="favicon">CraiditX 氪信</a> | Finance AI | Analysis | — | — | [Slides in English, November 2019](https://github.com/ClickHouse/clickhouse-presentations/blob/master/meetup33/udf.pptx) |
 | <a href="https://crazypanda.ru/en/" class="favicon">Crazypanda</a> | Games | | — | — | Live session on ClickHouse meetup |
@@ -158,6 +158,7 @@ toc_title: Adopters
 | <a href="https://www.staffcop.ru/" class="favicon">Staffcop</a> | Information Security | Main Product | — | — | [Official website, Documentation](https://www.staffcop.ru/sce43) |
 | <a href="https://www.suning.com/" class="favicon">Suning</a> | E-Commerce | User behaviour analytics | — | — | [Blog article](https://www.sohu.com/a/434152235_411876) |
 | <a href="https://superwall.me/" class="favicon">Superwall</a> | Monetization Tooling | Main product | — | — | [Word of mouth, Jan 2022](https://github.com/ClickHouse/ClickHouse/pull/33573) |
+| <a href="https://swetrix.com" class="favicon">Swetrix</a> | Analytics | Main Product | — | — | [Source code](https://github.com/swetrix/swetrix-api) |
 | <a href="https://www.teralytics.net/" class="favicon">Teralytics</a> | Mobility | Analytics | — | — | [Tech blog](https://www.teralytics.net/knowledge-hub/visualizing-mobility-data-the-scalability-challenge) |
 | <a href="https://www.tencent.com" class="favicon">Tencent</a> | Big Data | Data processing | — | — | [Slides in Chinese, October 2018](https://github.com/ClickHouse/clickhouse-presentations/blob/master/meetup19/5.%20ClickHouse大数据集群应用_李俊飞腾讯网媒事业部.pdf) |
 | <a href="https://www.tencent.com" class="favicon">Tencent</a> | Messaging | Logging | — | — | [Talk in Chinese, November 2019](https://youtu.be/T-iVQRuw-QY?t=5050) |

docs/en/operations/server-configuration-parameters/settings.md

@@ -1467,6 +1467,18 @@ The update is performed asynchronously, in a separate system thread.
 
 -   [background_schedule_pool_size](../../operations/settings/settings.md#background_schedule_pool_size)
 
+
+## dns_max_consecutive_failures {#server-settings-dns-max-consecutive-failures}
+
+The number of consecutive failures accepted when updating a DNS cache entry before it is dropped.
+Use `0` to disable cache dropping (entries will only be cleaned by `SYSTEM DROP DNS CACHE`)
+
+**Default value**: 5.
+
+**See also**
+
+-   [`SYSTEM DROP DNS CACHE`](../../sql-reference/statements/system.md#query_language-system-drop-dns-cache)
+
 ## distributed_ddl {#server-settings-distributed_ddl}
 
 Manage executing [distributed ddl queries](../../sql-reference/distributed-ddl.md)  (CREATE, DROP, ALTER, RENAME) on cluster.

docs/en/operations/settings/settings.md

@@ -519,6 +519,33 @@ Possible values:
 
 Default value: `1`.
 
+## allow_settings_after_format_in_insert {#allow_settings_after_format_in_insert}
+
+Control whether `SETTINGS` after `FORMAT` in `INSERT` queries is allowed or not. It is not recommended to use this, since this may interpret part of `SETTINGS` as values.
+
+Example:
+
+```sql
+INSERT INTO FUNCTION null('foo String') SETTINGS max_threads=1 VALUES ('bar');
+```
+
+But the following query will work only with `allow_settings_after_format_in_insert`:
+
+```sql
+SET allow_settings_after_format_in_insert=1;
+INSERT INTO FUNCTION null('foo String') VALUES ('bar') SETTINGS max_threads=1;
+```
+
+Possible values:
+
+-   0 — Disallow.
+-   1 — Allow.
+
+Default value: `0`.
+
+!!! note "Warning"
+    Use this setting only for backward compatibility if your use cases depend on old syntax.
+
 ## input_format_skip_unknown_fields {#settings-input-format-skip-unknown-fields}
 
 Enables or disables skipping insertion of extra data.
@@ -1062,6 +1089,15 @@ Result:
 └─────────────┴───────────┘
 ```
 
+## log_processors_profiles {#settings-log_processors_profiles}
+
+Write time that processor spent during execution/waiting for data to `system.processors_profile_log` table.
+
+See also:
+
+-   [`system.processors_profile_log`](../../operations/system-tables/processors_profile_log.md#system-processors_profile_log)
+-   [`EXPLAIN PIPELINE`](../../sql-reference/statements/explain.md#explain-pipeline)
+
 ## max_insert_block_size {#settings-max_insert_block_size}
 
 The size of blocks (in a count of rows) to form for insertion into a table.

docs/en/operations/system-tables/processors_profile_log.md

@@ -0,0 +1,75 @@
+# system.processors_profile_log {#system-processors_profile_log}
+
+This table contains profiling on processors level (that you can find in [`EXPLAIN PIPELINE`](../../sql-reference/statements/explain.md#explain-pipeline)).
+
+Columns:
+
+-   `event_date` ([Date](../../sql-reference/data-types/date.md)) — The date when the event happened.
+-   `event_time` ([DateTime64](../../sql-reference/data-types/datetime64.md)) — The date and time when the event happened.
+-   `id` ([UInt64](../../sql-reference/data-types/int-uint.md)) — ID of processor
+-   `parent_ids` ([Array(UInt64)](../../sql-reference/data-types/array.md)) — Parent processors IDs
+-   `query_id` ([String](../../sql-reference/data-types/string.md)) — ID of the query
+-   `name` ([LowCardinality(String)](../../sql-reference/data-types/lowcardinality.md)) — Name of the processor.
+-   `elapsed_us` ([UInt64](../../sql-reference/data-types/int-uint.md)) — Number of microseconds this processor was executed.
+-   `input_wait_elapsed_us` ([UInt64](../../sql-reference/data-types/int-uint.md)) — Number of microseconds this processor was waiting for data (from other processor).
+-   `output_wait_elapsed_us` ([UInt64](../../sql-reference/data-types/int-uint.md)) — Number of microseconds this processor was waiting because output port was full.
+
+**Example**
+
+Query:
+
+``` sql
+EXPLAIN PIPELINE
+SELECT sleep(1)
+
+┌─explain─────────────────────────┐
+│ (Expression)                    │
+│ ExpressionTransform             │
+│   (SettingQuotaAndLimits)       │
+│     (ReadFromStorage)           │
+│     SourceFromSingleChunk 0 → 1 │
+└─────────────────────────────────┘
+
+SELECT sleep(1)
+SETTINGS log_processors_profiles = 1
+
+Query id: feb5ed16-1c24-4227-aa54-78c02b3b27d4
+
+┌─sleep(1)─┐
+│        0 │
+└──────────┘
+
+1 rows in set. Elapsed: 1.018 sec.
+
+SELECT
+    name,
+    elapsed_us,
+    input_wait_elapsed_us,
+    output_wait_elapsed_us
+FROM system.processors_profile_log
+WHERE query_id = 'feb5ed16-1c24-4227-aa54-78c02b3b27d4'
+ORDER BY name ASC
+```
+
+Result:
+
+``` text
+┌─name────────────────────┬─elapsed_us─┬─input_wait_elapsed_us─┬─output_wait_elapsed_us─┐
+│ ExpressionTransform     │    1000497 │                  2823 │                    197 │
+│ LazyOutputFormat        │         36 │               1002188 │                      0 │
+│ LimitsCheckingTransform │         10 │               1002994 │                    106 │
+│ NullSource              │          5 │               1002074 │                      0 │
+│ NullSource              │          1 │               1002084 │                      0 │
+│ SourceFromSingleChunk   │         45 │                  4736 │                1000819 │
+└─────────────────────────┴────────────┴───────────────────────┴────────────────────────┘
+```
+
+Here you can see:
+
+-  `ExpressionTransform` was executing `sleep(1)` function, so it `work` will takes 1e6, and so `elapsed_us` > 1e6.
+-  `SourceFromSingleChunk` need to wait, because `ExpressionTransform` does not accept any data during execution of `sleep(1)`, so it will be in `PortFull` state for 1e6 us, and so `output_wait_elapsed_us` > 1e6.
+-  `LimitsCheckingTransform`/`NullSource`/`LazyOutputFormat` need to wait until `ExpressionTransform` will execute `sleep(1)` to process the result, so `input_wait_elapsed_us` > 1e6.
+
+**See Also**
+
+-   [`EXPLAIN PIPELINE`](../../sql-reference/statements/explain.md#explain-pipeline)

docs/en/sql-reference/functions/date-time-functions.md

@@ -393,6 +393,13 @@ This is a generalization of other functions named `toStartOf*`. For example,
 `toStartOfInterval(t, INTERVAL 1 day)` returns the same as `toStartOfDay(t)`,
 `toStartOfInterval(t, INTERVAL 15 minute)` returns the same as `toStartOfFifteenMinutes(t)` etc.
 
+## toLastDayOfMonth {#toLastDayOfMonth}
+
+Rounds up a date or date with time to the last day of the month.
+Returns the date.
+
+Alias: `LAST_DAY`.
+
 ## toTime {#totime}
 
 Converts a date with time to a certain fixed date, while preserving the time.

docs/en/sql-reference/functions/index.md

@@ -77,7 +77,7 @@ A function configuration contains the following settings:
 -   `argument` - argument description with the `type`, and optional `name` of an argument. Each argument is described in a separate setting. Specifying name is necessary if argument names are part of serialization for user defined function format like [Native](../../interfaces/formats.md#native) or [JSONEachRow](../../interfaces/formats.md#jsoneachrow). Default argument name value is `c` + argument_number.
 -   `format` - a [format](../../interfaces/formats.md) in which arguments are passed to the command.
 -   `return_type` - the type of a returned value.
--   `return_name` - name of retuned value. Specifying return name is necessary if return name is part of serialization for user defined function format like [Native](../../interfaces/formats.md#native) or [JSONEachRow](../../interfaces/formats.md#jsoneachrow). Optional. Default value is `result`.
+-   `return_name` - name of returned value. Specifying return name is necessary if return name is part of serialization for user defined function format like [Native](../../interfaces/formats.md#native) or [JSONEachRow](../../interfaces/formats.md#jsoneachrow). Optional. Default value is `result`.
 -   `type` - an executable type. If `type` is set to `executable` then single command is started. If it is set to `executable_pool` then a pool of commands is created.
 -   `max_command_execution_time` - maximum execution time in seconds for processing block of data. This setting is valid for `executable_pool` commands only. Optional. Default value is `10`.
 -   `command_termination_timeout` - time in seconds during which a command should finish after its pipe is closed. After that time `SIGTERM` is sent to the process executing the command. Optional. Default value is `10`.

docs/en/sql-reference/functions/other-functions.md

@@ -2499,3 +2499,41 @@ Result:
 │                      286 │
 └──────────────────────────┘
 ```
+
+## getTypeSerializationStreams {#getTypeSerializationStreams}
+
+return the serialization streams of data type.
+
+**Syntax**
+``` sql
+getTypeSerializationStreams(type_name)
+
+getTypeSerializationStreams(column)
+```
+
+**Arguments**
+- `type_name` - Name of data type to get its serialization paths. [String](../../sql-reference/data-types/string.md#string).
+- `column`       - any column which has a data type
+
+**Returned value**
+- List of serialization streams;
+
+Type: [Array](../../sql-reference/data-types/array.md)([String](../../sql-reference/data-types/string.md)).
+
+
+
+**Example**
+
+Query:
+
+``` sql
+SELECT getTypeSerializationStreams('Array(Array(Int8))')
+```
+
+Result:
+
+``` text
+┌───────────────────────getTypeSerializationStreams('Array(Array(Int8))')─────────────────────────────┐
+│ ['{ArraySizes}','{ArrayElements, ArraySizes}','{ArrayElements, ArrayElements, Regular}']            │
+└─────────────────────────────────────────────────────────────────────────────────────────────────────┘
+```

docs/en/sql-reference/statements/create/table.md

@@ -114,9 +114,9 @@ In addition, this column is not substituted when using an asterisk in a SELECT q
 
 ### EPHEMERAL {#ephemeral}
 
-`EPHEMERAL expr`
+`EPHEMERAL [expr]`
 
-Ephemeral column. Such a column isn't stored in the table and cannot be SELECTed, but can be referenced in the defaults of CREATE statement.
+Ephemeral column. Such a column isn't stored in the table and cannot be SELECTed, but can be referenced in the defaults of CREATE statement. If `expr` is omitted type for column is required.
 INSERT without list of columns will skip such column, so SELECT/INSERT invariant is preserved -  the dump obtained using `SELECT *` can be inserted back into the table using INSERT without specifying the list of columns.
 
 ### ALIAS {#alias}

docs/ja/introduction/adopters.md

@@ -27,7 +27,7 @@ toc_title: "\u30A2\u30C0\u30D7\u30BF\u30FC"
 | <a href="http://cisco.com/" class="favicon">Cisco</a>                                          | ネットワーク                 | トラフィック分析     | —                                                 | —                                                                        | [ライトニングトーク2019](https://youtu.be/-hI1vDR2oPY?t=5057)                                                                                                                                                     |
 | <a href="https://www.citadelsecurities.com/" class="favicon">Citadel Securities</a>            | 金融                         | —                    | —                                                 | —                                                                        | [2019年の貢献](https://github.com/ClickHouse/ClickHouse/pull/4774)                                                                                                                                                |
 | <a href="https://city-mobil.ru" class="favicon">シティモービル</a>                             | タクシー                     | 分析                 | —                                                 | —                                                                        | [ロシア語でのブログ投稿,月2020](https://habr.com/en/company/citymobil/blog/490660/)                                                                                                                               |
-| <a href="https://contentsquare.com" class="favicon">ContentSquare</a>                          | ウェブ分析                   | 主な製品             | —                                                 | —                                                                        | [フランス語でのブログ投稿,November2018](http://souslecapot.net/2018/11/21/patrick-chatain-vp-engineering-chez-contentsquare-penser-davantage-amelioration-continue-que-revolution-constante/)                     |
+| <a href="https://contentsquare.com" class="favicon">Contentsquare</a>                          | ウェブ分析                   | 主な製品             | —                                                 | —                                                                        | [フランス語でのブログ投稿,November2018](http://souslecapot.net/2018/11/21/patrick-chatain-vp-engineering-chez-contentsquare-penser-davantage-amelioration-continue-que-revolution-constante/)                     |
 | <a href="https://cloudflare.com" class="favicon">Cloudflare</a>                                | CDN                          | トラフィック分析     | 36台のサーバー                                    | —                                                                        | [ブログ投稿,月2017](https://blog.cloudflare.com/how-cloudflare-analyzes-1m-dns-queries-per-second/), [ブログ投稿,月2018](https://blog.cloudflare.com/http-analytics-for-6m-requests-per-second-using-clickhouse/) |
 | <a href="https://coru.net/" class="favicon">コルネット</a>                                     | 分析                         | 主な製品             | —                                                 | —                                                                        | [2019年英語スライド](https://github.com/ClickHouse/clickhouse-presentations/blob/master/meetup21/predictive_models.pdf)                                                                                           |
 | <a href="https://www.creditx.com" class="favicon">CraiditX 氪信</a>                            | ファイナンスAI               | 分析                 | —                                                 | —                                                                        | [2019年のスライド](https://github.com/ClickHouse/clickhouse-presentations/blob/master/meetup33/udf.pptx)                                                                                                          |

docs/ru/sql-reference/statements/create/table.md

@@ -110,9 +110,9 @@ SELECT x, toTypeName(x) FROM t1;
 
 ### EPHEMERAL {#ephemeral}
 
-`EPHEMERAL expr`
+`EPHEMERAL [expr]`
 
-Эфемерное выражение. Такой столбец не хранится в таблице и не может быть получен в запросе SELECT, но на него можно ссылаться в выражениях по умолчанию запроса CREATE.
+Эфемерное выражение. Такой столбец не хранится в таблице и не может быть получен в запросе SELECT, но на него можно ссылаться в выражениях по умолчанию запроса CREATE. Если значение по умолчанию `expr` не указано, то тип колонки должен быть специфицирован.
 INSERT без списка столбцов игнорирует этот столбец, таким образом сохраняется инвариант - т.е. дамп, полученный путём `SELECT *`, можно вставить обратно в таблицу INSERT-ом без указания списка столбцов.
 
 ### ALIAS {#alias}

docs/ru/sql-reference/table-functions/postgresql.md

@@ -126,7 +126,7 @@ CREATE TABLE pg_table_schema_with_dots (a UInt32)
 
 **См. также**
 
--   [Движок таблиц PostgreSQL](../../sql-reference/table-functions/postgresql.md)
+-   [Движок таблиц PostgreSQL](../../engines/table-engines/integrations/postgresql.md)
 -   [Использование PostgreSQL как источника данных для внешнего словаря](../../sql-reference/dictionaries/external-dictionaries/external-dicts-dict-sources.md#dicts-external_dicts_dict_sources-postgresql)
 
 [Оригинальная статья](https://clickhouse.com/docs/ru/sql-reference/table-functions/postgresql/) <!--hide-->

docs/tools/requirements.txt

@@ -16,7 +16,7 @@ jsmin==3.0.0
 livereload==2.6.3
 Markdown==3.3.2
 MarkupSafe==2.1.0
-mkdocs==1.1.2
+mkdocs==1.3.0
 mkdocs-htmlproofer-plugin==0.0.3
 mkdocs-macros-plugin==0.4.20
 nltk==3.7

packages/clickhouse-server.service

@@ -20,7 +20,7 @@ ExecStart=/usr/bin/clickhouse-server --config=/etc/clickhouse-server/config.xml
 EnvironmentFile=-/etc/default/clickhouse
 LimitCORE=infinity
 LimitNOFILE=500000
-CapabilityBoundingSet=CAP_NET_ADMIN CAP_IPC_LOCK CAP_SYS_NICE
+CapabilityBoundingSet=CAP_NET_ADMIN CAP_IPC_LOCK CAP_SYS_NICE CAP_NET_BIND_SERVICE
 
 [Install]
 # ClickHouse should not start from the rescue shell (rescue.target).

programs/client/Client.cpp

@@ -163,10 +163,24 @@ void Client::initialize(Poco::Util::Application & self)
 
     configReadClient(config(), home_path);
 
+    /** getenv is thread-safe in Linux glibc and in all sane libc implementations.
+      * But the standard does not guarantee that subsequent calls will not rewrite the value by returned pointer.
+      *
+      * man getenv:
+      *
+      * As typically implemented, getenv() returns a pointer to a string within the environment list.
+      * The caller must take care not to modify this string, since that would change the environment of
+      * the process.
+      *
+      * The implementation of getenv() is not required to be reentrant. The string pointed to by the return value of getenv()
+      * may be statically allocated, and can be modified by a subsequent call to getenv(), putenv(3), setenv(3), or unsetenv(3).
+      */
+
     const char * env_user = getenv("CLICKHOUSE_USER");
-    const char * env_password = getenv("CLICKHOUSE_PASSWORD");
     if (env_user)
         config().setString("user", env_user);
+
+    const char * env_password = getenv("CLICKHOUSE_PASSWORD");
     if (env_password)
         config().setString("password", env_password);
 
@@ -810,7 +824,7 @@ void Client::addOptions(OptionsDescription & options_description)
         ("quota_key", po::value<std::string>(), "A string to differentiate quotas when the user have keyed quotas configured on server")
 
         ("max_client_network_bandwidth", po::value<int>(), "the maximum speed of data exchange over the network for the client in bytes per second.")
-        ("compression", po::value<bool>(), "enable or disable compression")
+        ("compression", po::value<bool>(), "enable or disable compression (enabled by default for remote communication and disabled for localhost communication).")
 
         ("query-fuzzer-runs", po::value<int>()->default_value(0), "After executing every SELECT query, do random mutations in it and run again specified number of times. This is used for testing to discover unexpected corner cases.")
         ("interleave-queries-file", po::value<std::vector<std::string>>()->multitoken(),
@@ -1005,6 +1019,7 @@ void Client::processConfig()
             global_context->setCurrentQueryId(query_id);
     }
     print_stack_trace = config().getBool("stacktrace", false);
+    logging_initialized = true;
 
     if (config().has("multiquery"))
         is_multiquery = true;

programs/format/Format.cpp

@@ -54,6 +54,7 @@ int mainEntryClickHouseFormat(int argc, char ** argv)
         ("multiquery,n", "allow multiple queries in the same file")
         ("obfuscate", "obfuscate instead of formatting")
         ("backslash", "add a backslash at the end of each line of the formatted query")
+        ("allow_settings_after_format_in_insert", "Allow SETTINGS after FORMAT, but note, that this is not always safe")
         ("seed", po::value<std::string>(), "seed (arbitrary string) that determines the result of obfuscation")
     ;
 
@@ -83,6 +84,7 @@ int mainEntryClickHouseFormat(int argc, char ** argv)
         bool multiple = options.count("multiquery");
         bool obfuscate = options.count("obfuscate");
         bool backslash = options.count("backslash");
+        bool allow_settings_after_format_in_insert = options.count("allow_settings_after_format_in_insert");
 
         if (quiet && (hilite || oneline || obfuscate))
         {
@@ -154,7 +156,7 @@ int mainEntryClickHouseFormat(int argc, char ** argv)
             const char * pos = query.data();
             const char * end = pos + query.size();
 
-            ParserQuery parser(end);
+            ParserQuery parser(end, allow_settings_after_format_in_insert);
             do
             {
                 ASTPtr res = parseQueryAndMovePosition(

programs/local/LocalServer.cpp

@@ -434,6 +434,14 @@ catch (...)
     return getCurrentExceptionCode();
 }
 
+void LocalServer::updateLoggerLevel(const String & logs_level)
+{
+    if (!logging_initialized)
+        return;
+
+    config().setString("logger.level", logs_level);
+    updateLevels(config(), logger());
+}
 
 void LocalServer::processConfig()
 {
@@ -460,30 +468,31 @@ void LocalServer::processConfig()
     auto logging = (config().has("logger.console")
                     || config().has("logger.level")
                     || config().has("log-level")
+                    || config().has("send_logs_level")
                     || config().has("logger.log"));
 
-    auto file_logging = config().has("server_logs_file");
-    if (is_interactive && logging && !file_logging)
-        throw Exception("For interactive mode logging is allowed only with --server_logs_file option",
-                        ErrorCodes::BAD_ARGUMENTS);
+    auto level = config().getString("log-level", "trace");
 
-    if (file_logging)
+    if (config().has("server_logs_file"))
     {
-        auto level = Poco::Logger::parseLevel(config().getString("log-level", "trace"));
-        Poco::Logger::root().setLevel(level);
+        auto poco_logs_level = Poco::Logger::parseLevel(level);
+        Poco::Logger::root().setLevel(poco_logs_level);
         Poco::Logger::root().setChannel(Poco::AutoPtr<Poco::SimpleFileChannel>(new Poco::SimpleFileChannel(server_logs_file)));
+        logging_initialized = true;
     }
-    else if (logging)
+    else if (logging || is_interactive)
     {
-        // force enable logging
         config().setString("logger", "logger");
-        // sensitive data rules are not used here
+        auto log_level_default = is_interactive && !logging ? "none" : level;
+        config().setString("logger.level", config().getString("log-level", config().getString("send_logs_level", log_level_default)));
         buildLoggers(config(), logger(), "clickhouse-local");
+        logging_initialized = true;
     }
     else
     {
         Poco::Logger::root().setLevel("none");
         Poco::Logger::root().setChannel(Poco::AutoPtr<Poco::NullChannel>(new Poco::NullChannel()));
+        logging_initialized = false;
     }
 
     shared_context = Context::createShared();
@@ -713,6 +722,8 @@ void LocalServer::processOptions(const OptionsDescription &, const CommandLineOp
         config().setString("logger.log", options["logger.log"].as<std::string>());
     if (options.count("logger.level"))
         config().setString("logger.level", options["logger.level"].as<std::string>());
+    if (options.count("send_logs_level"))
+        config().setString("send_logs_level", options["send_logs_level"].as<std::string>());
 }
 
 }

programs/local/LocalServer.h

@@ -46,6 +46,8 @@ protected:
 
     void processConfig() override;
 
+    void updateLoggerLevel(const String & logs_level) override;
+
 private:
     /** Composes CREATE subquery based on passed arguments (--structure --file --table and --input-format)
       * This query will be executed first, before queries passed through --query argument

programs/server/Server.cpp

@@ -1503,7 +1503,8 @@ int Server::main(const std::vector<std::string> & /*args*/)
     else
     {
         /// Initialize a watcher periodically updating DNS cache
-        dns_cache_updater = std::make_unique<DNSCacheUpdater>(global_context, config().getInt("dns_cache_update_period", 15));
+        dns_cache_updater = std::make_unique<DNSCacheUpdater>(
+            global_context, config().getInt("dns_cache_update_period", 15), config().getUInt("dns_max_consecutive_failures", 5));
     }
 
 #if defined(OS_LINUX)

programs/server/config.xml

@@ -148,13 +148,13 @@
     <!-- <interserver_https_port>9010</interserver_https_port> -->
 
     <!-- Hostname that is used by other replicas to request this server.
-         If not specified, than it is determined analogous to 'hostname -f' command.
+         If not specified, then it is determined analogous to 'hostname -f' command.
          This setting could be used to switch replication to another network interface
          (the server may be connected to multiple networks via multiple addresses)
       -->
 
     <!--
-    <interserver_http_host>example.yandex.ru</interserver_http_host>
+    <interserver_http_host>example.clickhouse.com</interserver_http_host>
     -->
 
     <!-- You can specify credentials for authenthication between replicas.
@@ -765,14 +765,14 @@
     -->
     <!--<remote_url_allow_hosts>-->
         <!-- Host should be specified exactly as in URL. The name is checked before DNS resolution.
-            Example: "yandex.ru", "yandex.ru." and "www.yandex.ru" are different hosts.
+            Example: "clickhouse.com", "clickhouse.com." and "www.clickhouse.com" are different hosts.
                     If port is explicitly specified in URL, the host:port is checked as a whole.
                     If host specified here without port, any port with this host allowed.
-                    "yandex.ru" -> "yandex.ru:443", "yandex.ru:80" etc. is allowed, but "yandex.ru:80" -> only "yandex.ru:80" is allowed.
+                    "clickhouse.com" -> "clickhouse.com:443", "clickhouse.com:80" etc. is allowed, but "clickhouse.com:80" -> only "clickhouse.com:80" is allowed.
             If the host is specified as IP address, it is checked as specified in URL. Example: "[2a02:6b8:a::a]".
             If there are redirects and support for redirects is enabled, every redirect (the Location field) is checked.
             Host should be specified using the host xml tag:
-                    <host>yandex.ru</host>
+                    <host>clickhouse.com</host>
         -->
 
         <!-- Regular expression can be specified. RE2 engine is used for regexps.
@@ -1030,25 +1030,26 @@
         <flush_interval_milliseconds>1000</flush_interval_milliseconds>
     </crash_log>
 
-    <!-- Session log. Stores user log in (successful or not) and log out events. -->
-    <session_log>
+    <!-- Session log. Stores user log in (successful or not) and log out events.
+
+        Note: session log has known security issues and should not be used in production.
+    -->
+    <!-- <session_log>
         <database>system</database>
         <table>session_log</table>
 
         <partition_by>toYYYYMM(event_date)</partition_by>
         <flush_interval_milliseconds>7500</flush_interval_milliseconds>
-    </session_log>
+    </session_log> -->
 
-    <!-- Parameters for embedded dictionaries, used in Yandex.Metrica.
-         See https://clickhouse.com/docs/en/dicts/internal_dicts/
-    -->
-
-    <!-- Path to file with region hierarchy. -->
-    <!-- <path_to_regions_hierarchy_file>/opt/geo/regions_hierarchy.txt</path_to_regions_hierarchy_file> -->
-
-    <!-- Path to directory with files containing names of regions -->
-    <!-- <path_to_regions_names_files>/opt/geo/</path_to_regions_names_files> -->
+    <!-- Profiling on Processors level. -->
+    <processors_profile_log>
+        <database>system</database>
+        <table>processors_profile_log</table>
 
+        <partition_by>toYYYYMM(event_date)</partition_by>
+        <flush_interval_milliseconds>7500</flush_interval_milliseconds>
+    </processors_profile_log>
 
     <!-- <top_level_domains_path>/var/lib/clickhouse/top_level_domains/</top_level_domains_path> -->
     <!-- Custom TLD lists.

programs/server/config.yaml.example

@@ -103,7 +103,7 @@ interserver_http_port: 9009
 # If not specified, than it is determined analogous to 'hostname -f' command.
 # This setting could be used to switch replication to another network interface
 # (the server may be connected to multiple networks via multiple addresses)
-# interserver_http_host: example.yandex.ru
+# interserver_http_host: example.clickhouse.com
 
 # You can specify credentials for authenthication between replicas.
 # This is required when interserver_https_port is accessible from untrusted networks,
@@ -592,10 +592,10 @@ remote_servers:
 # remote_url_allow_hosts:
 
 # Host should be specified exactly as in URL. The name is checked before DNS resolution.
-# Example: "yandex.ru", "yandex.ru." and "www.yandex.ru" are different hosts.
+# Example: "clickhouse.com", "clickhouse.com." and "www.clickhouse.com" are different hosts.
 # If port is explicitly specified in URL, the host:port is checked as a whole.
 # If host specified here without port, any port with this host allowed.
-# "yandex.ru" -> "yandex.ru:443", "yandex.ru:80" etc. is allowed, but "yandex.ru:80" -> only "yandex.ru:80" is allowed.
+# "clickhouse.com" -> "clickhouse.com:443", "clickhouse.com:80" etc. is allowed, but "clickhouse.com:80" -> only "clickhouse.com:80" is allowed.
 # If the host is specified as IP address, it is checked as specified in URL. Example: "[2a02:6b8:a::a]".
 # If there are redirects and support for redirects is enabled, every redirect (the Location field) is checked.
 
@@ -803,16 +803,6 @@ crash_log:
     partition_by: ''
     flush_interval_milliseconds: 1000
 
-# Parameters for embedded dictionaries, used in Yandex.Metrica.
-# See https://clickhouse.com/docs/en/dicts/internal_dicts/
-
-# Path to file with region hierarchy.
-# path_to_regions_hierarchy_file: /opt/geo/regions_hierarchy.txt
-
-# Path to directory with files containing names of regions
-# path_to_regions_names_files: /opt/geo/
-
-
 # top_level_domains_path: /var/lib/clickhouse/top_level_domains/
 # Custom TLD lists.
 # Format: name: /path/to/file

programs/server/play.html

@@ -266,12 +266,25 @@
             color: var(--null-color);
         }
 
+        @keyframes hourglass-animation {
+            0% {
+                transform: rotate(-180deg);
+            }
+            50% {
+                transform: rotate(-180deg);
+            }
+            100% {
+                transform: none;
+            }
+        }
+
         #hourglass
         {
             display: none;
-            padding-left: 1rem;
+            margin-left: 1rem;
             font-size: 110%;
             color: #888;
+            animation: hourglass-animation 1s linear infinite;
         }
 
         #check-mark
@@ -457,7 +470,7 @@
         }
 
         document.getElementById('check-mark').style.display = 'none';
-        document.getElementById('hourglass').style.display = 'inline';
+        document.getElementById('hourglass').style.display = 'inline-block';
 
         xhr.send(query);
     }

programs/server/users.xml

@@ -79,9 +79,9 @@
                  Each element of list has one of the following forms:
                  <ip> IP-address or network mask. Examples: 213.180.204.3 or 10.0.0.1/8 or 10.0.0.1/255.255.255.0
                      2a02:6b8::3 or 2a02:6b8::3/64 or 2a02:6b8::3/ffff:ffff:ffff:ffff::.
-                 <host> Hostname. Example: server01.yandex.ru.
+                 <host> Hostname. Example: server01.clickhouse.com.
                      To check access, DNS query is performed, and all received addresses compared to peer address.
-                 <host_regexp> Regular expression for host names. Example, ^server\d\d-\d\d-\d\.yandex\.ru$
+                 <host_regexp> Regular expression for host names. Example, ^server\d\d-\d\d-\d\.clickhouse\.com$
                      To check access, DNS PTR query is performed for peer address and then regexp is applied.
                      Then, for result of PTR query, another DNS query is performed and all received addresses compared to peer address.
                      Strongly recommended that regexp is ends with $

programs/server/users.yaml.example

@@ -70,9 +70,9 @@ users:
         # Each element of list has one of the following forms:
         # ip: IP-address or network mask. Examples: 213.180.204.3 or 10.0.0.1/8 or 10.0.0.1/255.255.255.0
         # 2a02:6b8::3 or 2a02:6b8::3/64 or 2a02:6b8::3/ffff:ffff:ffff:ffff::.
-        # host: Hostname. Example: server01.yandex.ru.
+        # host: Hostname. Example: server01.clickhouse.com.
         # To check access, DNS query is performed, and all received addresses compared to peer address.
-        # host_regexp: Regular expression for host names. Example, ^server\d\d-\d\d-\d\.yandex\.ru$
+        # host_regexp: Regular expression for host names. Example, ^server\d\d-\d\d-\d\.clickhouse\.com$
         # To check access, DNS PTR query is performed for peer address and then regexp is applied.
         # Then, for result of PTR query, another DNS query is performed and all received addresses compared to peer address.
         # Strongly recommended that regexp is ends with $ and take all expression in ''

src/Access/Common/AccessType.h

@@ -182,6 +182,7 @@ enum class AccessType
     M(JDBC, "", GLOBAL, SOURCES) \
     M(HDFS, "", GLOBAL, SOURCES) \
     M(S3, "", GLOBAL, SOURCES) \
+    M(HIVE, "", GLOBAL, SOURCES) \
     M(SOURCES, "", GROUP, ALL) \
     \
     M(ALL, "ALL PRIVILEGES", GROUP, NONE) /* full access */ \

src/Access/Common/QuotaDefs.cpp

@@ -107,6 +107,11 @@ const QuotaTypeInfo & QuotaTypeInfo::get(QuotaType type)
             static const auto info = make_info("EXECUTION_TIME", 1000000000 /* execution_time is stored in nanoseconds */);
             return info;
         }
+        case QuotaType::WRITTEN_BYTES:
+        {
+            static const auto info = make_info("WRITTEN_BYTES", 1);
+            return info;
+        }
         case QuotaType::MAX: break;
     }
     throw Exception("Unexpected quota type: " + std::to_string(static_cast<int>(type)), ErrorCodes::LOGICAL_ERROR);

src/Access/Common/QuotaDefs.h

@@ -20,6 +20,7 @@ enum class QuotaType
     READ_ROWS,      /// Number of rows read from tables.
     READ_BYTES,     /// Number of bytes read from tables.
     EXECUTION_TIME, /// Total amount of query execution time in nanoseconds.
+    WRITTEN_BYTES,  /// Number of bytes written to tables.
 
     MAX
 };

src/Access/EnabledQuota.cpp

@@ -13,7 +13,7 @@ namespace DB
 {
 namespace ErrorCodes
 {
-    extern const int QUOTA_EXPIRED;
+    extern const int QUOTA_EXCEEDED;
 }
 
 
@@ -33,7 +33,7 @@ struct EnabledQuota::Impl
             "Quota for user " + backQuote(user_name) + " for " + to_string(duration) + " has been exceeded: "
                 + type_info.valueToStringWithName(used) + "/" + type_info.valueToString(max) + ". "
                 + "Interval will end at " + to_string(end_of_interval) + ". " + "Name of quota template: " + backQuote(quota_name),
-            ErrorCodes::QUOTA_EXPIRED);
+            ErrorCodes::QUOTA_EXCEEDED);
     }
 
 

src/CMakeLists.txt

@@ -49,6 +49,18 @@ if (COMPILER_GCC)
     add_definitions ("-fno-tree-loop-distribute-patterns")
 endif ()
 
+# ClickHouse developers may use platform-dependent code under some macro (e.g. `#ifdef ENABLE_MULTITARGET`).
+# If turned ON, this option defines such macro.
+# See `src/Common/TargetSpecific.h`
+option(ENABLE_MULTITARGET_CODE "Enable platform-dependent code" ON)
+
+if (ENABLE_MULTITARGET_CODE)
+    add_definitions(-DENABLE_MULTITARGET_CODE=1)
+else()
+    add_definitions(-DENABLE_MULTITARGET_CODE=0)
+endif()
+
+
 add_subdirectory (Access)
 add_subdirectory (Backups)
 add_subdirectory (Columns)

src/Client/ClientBase.cpp

@@ -240,6 +240,14 @@ void interruptSignalHandler(int signum)
 }
 
 
+/// To cancel the query on local format error.
+class LocalFormatError : public DB::Exception
+{
+public:
+    using Exception::Exception;
+};
+
+
 ClientBase::~ClientBase() = default;
 ClientBase::ClientBase() = default;
 
@@ -268,7 +276,7 @@ void ClientBase::setupSignalHandler()
 
 ASTPtr ClientBase::parseQuery(const char *& pos, const char * end, bool allow_multi_statements) const
 {
-    ParserQuery parser(end);
+    ParserQuery parser(end, global_context->getSettings().allow_settings_after_format_in_insert);
     ASTPtr res;
 
     const auto & settings = global_context->getSettingsRef();
@@ -442,6 +450,7 @@ void ClientBase::onProfileInfo(const ProfileInfo & profile_info)
 
 
 void ClientBase::initBlockOutputStream(const Block & block, ASTPtr parsed_query)
+try
 {
     if (!output_format)
     {
@@ -530,6 +539,10 @@ void ClientBase::initBlockOutputStream(const Block & block, ASTPtr parsed_query)
         output_format->setAutoFlush();
     }
 }
+catch (...)
+{
+    throw LocalFormatError(getCurrentExceptionMessage(print_stack_trace), getCurrentExceptionCode());
+}
 
 
 void ClientBase::initLogsOutputStream()
@@ -721,6 +734,9 @@ void ClientBase::receiveResult(ASTPtr parsed_query)
         = std::max(min_poll_interval, std::min<size_t>(receive_timeout.totalMicroseconds(), default_poll_interval));
 
     bool break_on_timeout = connection->getConnectionType() != IServerConnection::Type::LOCAL;
+
+    std::exception_ptr local_format_error;
+
     while (true)
     {
         Stopwatch receive_watch(CLOCK_MONOTONIC_COARSE);
@@ -769,10 +785,21 @@ void ClientBase::receiveResult(ASTPtr parsed_query)
                 break;
         }
 
-        if (!receiveAndProcessPacket(parsed_query, cancelled))
-            break;
+        try
+        {
+            if (!receiveAndProcessPacket(parsed_query, cancelled))
+                break;
+        }
+        catch (const LocalFormatError &)
+        {
+            local_format_error = std::current_exception();
+            connection->sendCancel();
+        }
     }
 
+    if (local_format_error)
+        std::rethrow_exception(local_format_error);
+
     if (cancelled && is_interactive)
         std::cout << "Query was cancelled." << std::endl;
 }
@@ -1298,6 +1325,13 @@ void ClientBase::processParsedSingleQuery(const String & full_query, const Strin
         }
     }
 
+    if (const auto * set_query = parsed_query->as<ASTSetQuery>())
+    {
+        const auto * logs_level_field = set_query->changes.tryGet(std::string_view{"send_logs_level"});
+        if (logs_level_field)
+            updateLoggerLevel(logs_level_field->safeGet<String>());
+    }
+
     processed_rows = 0;
     written_first_block = false;
     progress_indication.resetProgress();
@@ -1494,24 +1528,19 @@ MultiQueryProcessingStage ClientBase::analyzeMultiQueryText(
 
 bool ClientBase::executeMultiQuery(const String & all_queries_text)
 {
-    // It makes sense not to base any control flow on this, so that it is
-    // the same in tests and in normal usage. The only difference is that in
-    // normal mode we ignore the test hints.
-    const bool test_mode = config().has("testmode");
-    if (test_mode)
-    {
-        /// disable logs if expects errors
-        TestHint test_hint(test_mode, all_queries_text);
-        if (test_hint.clientError() || test_hint.serverError())
-            processTextAsSingleQuery("SET send_logs_level = 'fatal'");
-    }
-
     bool echo_query = echo_queries;
 
     /// Test tags are started with "--" so they are interpreted as comments anyway.
     /// But if the echo is enabled we have to remove the test tags from `all_queries_text`
     /// because we don't want test tags to be echoed.
-    size_t test_tags_length = test_mode ? getTestTagsLength(all_queries_text) : 0;
+    {
+        /// disable logs if expects errors
+        TestHint test_hint(all_queries_text);
+        if (test_hint.clientError() || test_hint.serverError())
+            processTextAsSingleQuery("SET send_logs_level = 'fatal'");
+    }
+
+    size_t test_tags_length = getTestTagsLength(all_queries_text);
 
     /// Several queries separated by ';'.
     /// INSERT data is ended by the end of line, not ';'.
@@ -1548,7 +1577,7 @@ bool ClientBase::executeMultiQuery(const String & all_queries_text)
                 // Try to find test hint for syntax error. We don't know where
                 // the query ends because we failed to parse it, so we consume
                 // the entire line.
-                TestHint hint(test_mode, String(this_query_begin, this_query_end - this_query_begin));
+                TestHint hint(String(this_query_begin, this_query_end - this_query_begin));
                 if (hint.serverError())
                 {
                     // Syntax errors are considered as client errors
@@ -1586,7 +1615,7 @@ bool ClientBase::executeMultiQuery(const String & all_queries_text)
                 // Look for the hint in the text of query + insert data + trailing
                 // comments, e.g. insert into t format CSV 'a' -- { serverError 123 }.
                 // Use the updated query boundaries we just calculated.
-                TestHint test_hint(test_mode, full_query);
+                TestHint test_hint(full_query);
 
                 // Echo all queries if asked; makes for a more readable reference file.
                 echo_query = test_hint.echoQueries().value_or(echo_query);
@@ -2187,8 +2216,6 @@ void ClientBase::init(int argc, char ** argv)
         ("suggestion_limit", po::value<int>()->default_value(10000),
             "Suggestion limit for how many databases, tables and columns to fetch.")
 
-        ("testmode,T", "enable test hints in comments")
-
         ("format,f", po::value<std::string>(), "default output format")
         ("vertical,E", "vertical output format, same as --format=Vertical or FORMAT Vertical or \\G at end of command")
         ("highlight", po::value<bool>()->default_value(true), "enable or disable basic syntax highlight in interactive command line")
@@ -2294,8 +2321,6 @@ void ClientBase::init(int argc, char ** argv)
         config().setBool("interactive", true);
     if (options.count("pager"))
         config().setString("pager", options["pager"].as<std::string>());
-    if (options.count("testmode"))
-        config().setBool("testmode", true);
 
     if (options.count("log-level"))
         Poco::Logger::root().setLevel(options["log-level"].as<std::string>());

src/Client/ClientBase.h

@@ -95,6 +95,7 @@ protected:
         std::optional<ProgramOptionsDescription> hosts_and_ports_description;
     };
 
+    virtual void updateLoggerLevel(const String &) {}
     virtual void printHelpMessage(const OptionsDescription & options_description) = 0;
     virtual void addOptions(OptionsDescription & options_description) = 0;
     virtual void processOptions(const OptionsDescription & options_description,
@@ -265,6 +266,8 @@ protected:
     bool allow_repeated_settings = false;
 
     bool cancelled = false;
+
+    bool logging_initialized = false;
 };
 
 }

src/Client/TestHint.cpp

@@ -32,12 +32,9 @@ int parseErrorCode(DB::ReadBufferFromString & in)
 namespace DB
 {
 
-TestHint::TestHint(bool enabled_, const String & query_)
+TestHint::TestHint(const String & query_)
     : query(query_)
 {
-    if (!enabled_)
-        return;
-
     // Don't parse error hints in leading comments, because it feels weird.
     // Leading 'echo' hint is OK.
     bool is_leading_hint = true;

src/Client/TestHint.h

@@ -7,7 +7,7 @@
 namespace DB
 {
 
-/// Checks expected server and client error codes in --testmode.
+/// Checks expected server and client error codes.
 ///
 /// The following comment hints are supported:
 ///
@@ -25,12 +25,12 @@ namespace DB
 ///
 /// Examples:
 ///
-/// - echo 'select / -- { clientError 62 }' | clickhouse-client --testmode -nm
+/// - echo 'select / -- { clientError 62 }' | clickhouse-client -nm
 ///
 //    Here the client parses the query but it is incorrect, so it expects
 ///   SYNTAX_ERROR (62).
 ///
-/// - echo 'select foo -- { serverError 47 }' | clickhouse-client --testmode -nm
+/// - echo 'select foo -- { serverError 47 }' | clickhouse-client -nm
 ///
 ///   But here the query is correct, but there is no such column "foo", so it
 ///   is UNKNOWN_IDENTIFIER server error.
@@ -43,7 +43,7 @@ namespace DB
 class TestHint
 {
 public:
-    TestHint(bool enabled_, const String & query_);
+    TestHint(const String & query_);
 
     int serverError() const { return server_error; }
     int clientError() const { return client_error; }

src/Common/ColumnsHashingImpl.h

@@ -125,7 +125,7 @@ class FindResultImpl : public FindResultImplBase, public FindResultImplOffsetBas
 
 public:
     FindResultImpl()
-        : FindResultImplBase(false), FindResultImplOffsetBase<need_offset>(0)
+        : FindResultImplBase(false), FindResultImplOffsetBase<need_offset>(0) // NOLINT(clang-analyzer-optin.cplusplus.UninitializedObject)  intentionally allow uninitialized value here
     {}
 
     FindResultImpl(Mapped * value_, bool found_, size_t off)

src/Common/CompactArray.h

@@ -214,6 +214,9 @@ private:
 
         /// offset in bits to the next to the rightmost bit at that byte; or zero if the rightmost bit is the rightmost bit in that byte.
         offset_r = (l + content_width) % 8;
+
+        content_l = nullptr;
+        content_r = nullptr;
     }
 
     UInt8 ALWAYS_INLINE read(UInt8 value_l) const

src/Common/CurrentMetrics.cpp

@@ -81,6 +81,14 @@
     M(ActiveSyncDrainedConnections, "Number of active connections drained synchronously.") \
     M(AsynchronousReadWait, "Number of threads waiting for asynchronous read.") \
     M(PendingAsyncInsert, "Number of asynchronous inserts that are waiting for flush.") \
+    M(KafkaConsumers, "Number of active Kafka consumers") \
+    M(KafkaConsumersWithAssignment, "Number of active Kafka consumers which have some partitions assigned.") \
+    M(KafkaProducers, "Number of active Kafka producer created") \
+    M(KafkaLibrdkafkaThreads, "Number of active librdkafka threads") \
+    M(KafkaBackgroundReads, "Number of background reads currently working (populating materialized views from Kafka)") \
+    M(KafkaConsumersInUse, "Number of consumers which are currently used by direct or background reads") \
+    M(KafkaWrites, "Number of currently running inserts to Kafka") \
+    M(KafkaAssignedPartitions, "Number of partitions Kafka tables currently assigned to") \
 
 namespace CurrentMetrics
 {

src/Common/CurrentThread.h

@@ -91,6 +91,7 @@ public:
     struct QueryScope
     {
         explicit QueryScope(ContextMutablePtr query_context);
+        explicit QueryScope(ContextPtr query_context);
         ~QueryScope();
 
         void logPeakMemoryUsage();

src/Common/DNSResolver.cpp

@@ -118,12 +118,15 @@ static DNSResolver::IPAddresses resolveIPAddressImpl(const std::string & host)
     }
     catch (const Poco::Net::DNSException & e)
     {
-        LOG_ERROR(&Poco::Logger::get("DNSResolver"), "Cannot resolve host ({}), error {}: {}.", host, e.code(), e.message());
+        LOG_ERROR(&Poco::Logger::get("DNSResolver"), "Cannot resolve host ({}), error {}: {}.", host, e.code(), e.name());
         addresses.clear();
     }
 
     if (addresses.empty())
+    {
+        ProfileEvents::increment(ProfileEvents::DNSError);
         throw Exception("Not found address of host: " + host, ErrorCodes::DNS_ERROR);
+    }
 
     return addresses;
 }
@@ -142,6 +145,9 @@ static String reverseResolveImpl(const Poco::Net::IPAddress & address)
 
 struct DNSResolver::Impl
 {
+    using HostWithConsecutiveFailures = std::unordered_map<String, UInt32>;
+    using AddressWithConsecutiveFailures = std::unordered_map<Poco::Net::IPAddress, UInt32>;
+
     CachedFn<&resolveIPAddressImpl> cache_host;
     CachedFn<&reverseResolveImpl> cache_address;
 
@@ -152,12 +158,12 @@ struct DNSResolver::Impl
     std::optional<String> host_name;
 
     /// Store hosts, which was asked to resolve from last update of DNS cache.
-    NameSet new_hosts;
-    std::unordered_set<Poco::Net::IPAddress> new_addresses;
+    HostWithConsecutiveFailures new_hosts;
+    AddressWithConsecutiveFailures new_addresses;
 
     /// Store all hosts, which was whenever asked to resolve
-    NameSet known_hosts;
-    std::unordered_set<Poco::Net::IPAddress> known_addresses;
+    HostWithConsecutiveFailures known_hosts;
+    AddressWithConsecutiveFailures known_addresses;
 
     /// If disabled, will not make cache lookups, will resolve addresses manually on each call
     std::atomic<bool> disable_cache{false};
@@ -246,38 +252,68 @@ String DNSResolver::getHostName()
 static const String & cacheElemToString(const String & str) { return str; }
 static String cacheElemToString(const Poco::Net::IPAddress & addr) { return addr.toString(); }
 
-template<typename UpdateF, typename ElemsT>
-bool DNSResolver::updateCacheImpl(UpdateF && update_func, ElemsT && elems, const String & log_msg)
+template <typename UpdateF, typename ElemsT>
+bool DNSResolver::updateCacheImpl(
+    UpdateF && update_func,
+    ElemsT && elems,
+    UInt32 max_consecutive_failures,
+    const String & notfound_log_msg,
+    const String & dropped_log_msg)
 {
     bool updated = false;
     String lost_elems;
-    for (const auto & elem : elems)
+    using iterators = typename std::remove_reference_t<decltype(elems)>::iterator;
+    std::vector<iterators> elements_to_drop;
+    for (auto it = elems.begin(); it != elems.end(); it++)
     {
         try
         {
-            updated |= (this->*update_func)(elem);
+            updated |= (this->*update_func)(it->first);
+            it->second = 0;
         }
-        catch (const Poco::Net::NetException &)
+        catch (const DB::Exception & e)
         {
-            ProfileEvents::increment(ProfileEvents::DNSError);
-
+            if (e.code() != ErrorCodes::DNS_ERROR)
+            {
+                tryLogCurrentException(log, __PRETTY_FUNCTION__);
+                continue;
+            }
             if (!lost_elems.empty())
                 lost_elems += ", ";
-            lost_elems += cacheElemToString(elem);
+            lost_elems += cacheElemToString(it->first);
+            if (max_consecutive_failures)
+            {
+                it->second++;
+                if (it->second >= max_consecutive_failures)
+                    elements_to_drop.emplace_back(it);
+            }
         }
         catch (...)
         {
-            tryLogCurrentException(__PRETTY_FUNCTION__);
+            tryLogCurrentException(log, __PRETTY_FUNCTION__);
         }
     }
 
     if (!lost_elems.empty())
-        LOG_INFO(log, fmt::runtime(log_msg), lost_elems);
+        LOG_INFO(log, fmt::runtime(notfound_log_msg), lost_elems);
+    if (elements_to_drop.size())
+    {
+        updated = true;
+        String deleted_elements;
+        for (auto it : elements_to_drop)
+        {
+            if (!deleted_elements.empty())
+                deleted_elements += ", ";
+            deleted_elements += cacheElemToString(it->first);
+            elems.erase(it);
+        }
+        LOG_INFO(log, fmt::runtime(dropped_log_msg), deleted_elements);
+    }
 
     return updated;
 }
 
-bool DNSResolver::updateCache()
+bool DNSResolver::updateCache(UInt32 max_consecutive_failures)
 {
     LOG_DEBUG(log, "Updating DNS cache");
 
@@ -301,8 +337,14 @@ bool DNSResolver::updateCache()
     /// DROP DNS CACHE will wait on update_mutex (possibly while holding drop_mutex)
     std::lock_guard lock(impl->update_mutex);
 
-    bool hosts_updated = updateCacheImpl(&DNSResolver::updateHost, impl->known_hosts, "Cached hosts not found: {}");
-    updateCacheImpl(&DNSResolver::updateAddress, impl->known_addresses, "Cached addresses not found: {}");
+    bool hosts_updated = updateCacheImpl(
+        &DNSResolver::updateHost, impl->known_hosts, max_consecutive_failures, "Cached hosts not found: {}", "Cached hosts dropped: {}");
+    updateCacheImpl(
+        &DNSResolver::updateAddress,
+        impl->known_addresses,
+        max_consecutive_failures,
+        "Cached addresses not found: {}",
+        "Cached addresses dropped: {}");
 
     LOG_DEBUG(log, "Updated DNS cache");
     return hosts_updated;
@@ -326,13 +368,15 @@ bool DNSResolver::updateAddress(const Poco::Net::IPAddress & address)
 void DNSResolver::addToNewHosts(const String & host)
 {
     std::lock_guard lock(impl->drop_mutex);
-    impl->new_hosts.insert(host);
+    UInt8 consecutive_failures = 0;
+    impl->new_hosts.insert({host, consecutive_failures});
 }
 
 void DNSResolver::addToNewAddresses(const Poco::Net::IPAddress & address)
 {
     std::lock_guard lock(impl->drop_mutex);
-    impl->new_addresses.insert(address);
+    UInt8 consecutive_failures = 0;
+    impl->new_addresses.insert({address, consecutive_failures});
 }
 
 DNSResolver::~DNSResolver() = default;

src/Common/DNSResolver.h

@@ -47,14 +47,20 @@ public:
     void dropCache();
 
     /// Updates all known hosts in cache.
-    /// Returns true if IP of any host has been changed.
-    bool updateCache();
+    /// Returns true if IP of any host has been changed or an element was dropped (too many failures)
+    bool updateCache(UInt32 max_consecutive_failures);
 
     ~DNSResolver();
 
 private:
-    template<typename UpdateF, typename ElemsT>
-    bool updateCacheImpl(UpdateF && update_func, ElemsT && elems, const String & log_msg);
+    template <typename UpdateF, typename ElemsT>
+
+    bool updateCacheImpl(
+        UpdateF && update_func,
+        ElemsT && elems,
+        UInt32 max_consecutive_failures,
+        const String & notfound_log_msg,
+        const String & dropped_log_msg);
 
     DNSResolver();
 

src/Common/DateLUTImpl.h

@@ -360,6 +360,27 @@ public:
             return toDayNum(LUTIndex(i - (lut[i].day_of_month - 1)));
     }
 
+    /// Round up to last day of month.
+    template <typename DateOrTime>
+    inline Time toLastDayOfMonth(DateOrTime v) const
+    {
+        const LUTIndex i = toLUTIndex(v);
+        if constexpr (std::is_unsigned_v<DateOrTime> || std::is_same_v<DateOrTime, DayNum>)
+            return lut_saturated[i - lut[i].day_of_month + lut[i].days_in_month].date;
+        else
+            return lut[i - lut[i].day_of_month + lut[i].days_in_month].date;
+    }
+
+    template <typename DateOrTime>
+    inline auto toLastDayNumOfMonth(DateOrTime v) const
+    {
+        const LUTIndex i = toLUTIndex(v);
+        if constexpr (std::is_unsigned_v<DateOrTime> || std::is_same_v<DateOrTime, DayNum>)
+            return toDayNum(LUTIndexWithSaturation(i - lut[i].day_of_month + lut[i].days_in_month));
+        else
+            return toDayNum(LUTIndex(i - lut[i].day_of_month + lut[i].days_in_month));
+    }
+
     /// Round down to start of quarter.
     template <typename DateOrTime>
     inline auto toFirstDayNumOfQuarter(DateOrTime v) const

src/Common/ErrorCodes.cpp

@@ -208,7 +208,7 @@
     M(198, DNS_ERROR) \
     M(199, UNKNOWN_QUOTA) \
     M(200, QUOTA_DOESNT_ALLOW_KEYS) \
-    M(201, QUOTA_EXPIRED) \
+    M(201, QUOTA_EXCEEDED) \
     M(202, TOO_MANY_SIMULTANEOUS_QUERIES) \
     M(203, NO_FREE_CONNECTION) \
     M(204, CANNOT_FSYNC) \

src/Common/FileCache.cpp

@@ -3,6 +3,7 @@
 #include <Common/randomSeed.h>
 #include <Common/SipHash.h>
 #include <Common/hex.h>
+#include <Common/FileCacheSettings.h>
 #include <IO/ReadHelpers.h>
 #include <IO/WriteBufferFromFile.h>
 #include <IO/ReadSettings.h>
@@ -31,13 +32,11 @@ namespace
 
 IFileCache::IFileCache(
     const String & cache_base_path_,
-    size_t max_size_,
-    size_t max_element_size_,
-    size_t max_file_segment_size_)
+    const FileCacheSettings & cache_settings_)
     : cache_base_path(cache_base_path_)
-    , max_size(max_size_)
-    , max_element_size(max_element_size_)
-    , max_file_segment_size(max_file_segment_size_)
+    , max_size(cache_settings_.max_size)
+    , max_element_size(cache_settings_.max_elements)
+    , max_file_segment_size(cache_settings_.max_file_segment_size)
 {
 }
 
@@ -58,7 +57,7 @@ String IFileCache::getPathInLocalCache(const Key & key)
     return fs::path(cache_base_path) / key_str.substr(0, 3) / key_str;
 }
 
-bool IFileCache::shouldBypassCache()
+bool IFileCache::isReadOnly()
 {
     return !CurrentThread::isInitialized()
         || !CurrentThread::get().getQueryContext()
@@ -71,8 +70,8 @@ void IFileCache::assertInitialized() const
         throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Cache not initialized");
 }
 
-LRUFileCache::LRUFileCache(const String & cache_base_path_, size_t max_size_, size_t max_element_size_, size_t max_file_segment_size_)
-    : IFileCache(cache_base_path_, max_size_, max_element_size_, max_file_segment_size_)
+LRUFileCache::LRUFileCache(const String & cache_base_path_, const FileCacheSettings & cache_settings_)
+    : IFileCache(cache_base_path_, cache_settings_)
     , log(&Poco::Logger::get("LRUFileCache"))
 {
 }
@@ -205,8 +204,8 @@ FileSegments LRUFileCache::getImpl(
     return result;
 }
 
-FileSegments LRUFileCache::splitRangeIntoEmptyCells(
-    const Key & key, size_t offset, size_t size, std::lock_guard<std::mutex> & cache_lock)
+FileSegments LRUFileCache::splitRangeIntoCells(
+    const Key & key, size_t offset, size_t size, FileSegment::State state, std::lock_guard<std::mutex> & cache_lock)
 {
     assert(size > 0);
 
@@ -222,9 +221,10 @@ FileSegments LRUFileCache::splitRangeIntoEmptyCells(
         current_cell_size = std::min(remaining_size, max_file_segment_size);
         remaining_size -= current_cell_size;
 
-        auto * cell = addCell(key, current_pos, current_cell_size, FileSegment::State::EMPTY, cache_lock);
+        auto * cell = addCell(key, current_pos, current_cell_size, state, cache_lock);
         if (cell)
             file_segments.push_back(cell->file_segment);
+        assert(cell);
 
         current_pos += current_cell_size;
     }
@@ -241,12 +241,16 @@ FileSegmentsHolder LRUFileCache::getOrSet(const Key & key, size_t offset, size_t
 
     std::lock_guard cache_lock(mutex);
 
+#ifndef NDEBUG
+    assertCacheCorrectness(key, cache_lock);
+#endif
+
     /// Get all segments which intersect with the given range.
     auto file_segments = getImpl(key, range, cache_lock);
 
     if (file_segments.empty())
     {
-        file_segments = splitRangeIntoEmptyCells(key, offset, size, cache_lock);
+        file_segments = splitRangeIntoCells(key, offset, size, FileSegment::State::EMPTY, cache_lock);
     }
     else
     {
@@ -291,7 +295,7 @@ FileSegmentsHolder LRUFileCache::getOrSet(const Key & key, size_t offset, size_t
             assert(current_pos < segment_range.left);
 
             auto hole_size = segment_range.left - current_pos;
-            file_segments.splice(it, splitRangeIntoEmptyCells(key, current_pos, hole_size, cache_lock));
+            file_segments.splice(it, splitRangeIntoCells(key, current_pos, hole_size, FileSegment::State::EMPTY, cache_lock));
 
             current_pos = segment_range.right + 1;
             ++it;
@@ -305,7 +309,7 @@ FileSegmentsHolder LRUFileCache::getOrSet(const Key & key, size_t offset, size_t
             /// segmentN
 
             auto hole_size = range.right - current_pos + 1;
-            file_segments.splice(file_segments.end(), splitRangeIntoEmptyCells(key, current_pos, hole_size, cache_lock));
+            file_segments.splice(file_segments.end(), splitRangeIntoCells(key, current_pos, hole_size, FileSegment::State::EMPTY, cache_lock));
         }
     }
 
@@ -315,7 +319,7 @@ FileSegmentsHolder LRUFileCache::getOrSet(const Key & key, size_t offset, size_t
 
 LRUFileCache::FileSegmentCell * LRUFileCache::addCell(
     const Key & key, size_t offset, size_t size, FileSegment::State state,
-    std::lock_guard<std::mutex> & /* cache_lock */)
+    std::lock_guard<std::mutex> & cache_lock)
 {
     /// Create a file segment cell and put it in `files` map by [key][offset].
 
@@ -323,8 +327,10 @@ LRUFileCache::FileSegmentCell * LRUFileCache::addCell(
         return nullptr; /// Empty files are not cached.
 
     if (files[key].contains(offset))
-        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
-            "Cache already exists for key: `{}`, offset: {}, size: {}", keyToStr(key), offset, size);
+        throw Exception(
+            ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
+            "Cache already exists for key: `{}`, offset: {}, size: {}.\nCurrent cache structure: {}",
+            keyToStr(key), offset, size, dumpStructureImpl(key, cache_lock));
 
     auto file_segment = std::make_shared<FileSegment>(offset, size, key, this, state);
     FileSegmentCell cell(std::move(file_segment), queue);
@@ -340,12 +346,29 @@ LRUFileCache::FileSegmentCell * LRUFileCache::addCell(
 
     auto [it, inserted] = offsets.insert({offset, std::move(cell)});
     if (!inserted)
-        throw Exception(ErrorCodes::LOGICAL_ERROR,
-            "Failed to insert into cache key: `{}`, offset: {}, size: {}", keyToStr(key), offset, size);
+        throw Exception(
+            ErrorCodes::LOGICAL_ERROR,
+            "Failed to insert into cache key: `{}`, offset: {}, size: {}",
+            keyToStr(key), offset, size);
 
     return &(it->second);
 }
 
+FileSegmentsHolder LRUFileCache::setDownloading(const Key & key, size_t offset, size_t size)
+{
+    std::lock_guard cache_lock(mutex);
+
+    auto * cell = getCell(key, offset, cache_lock);
+    if (cell)
+        throw Exception(
+            ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
+            "Cache cell already exists for key `{}` and offset {}",
+            keyToStr(key), offset);
+
+    auto file_segments = splitRangeIntoCells(key, offset, size, FileSegment::State::DOWNLOADING, cache_lock);
+    return FileSegmentsHolder(std::move(file_segments));
+}
+
 bool LRUFileCache::tryReserve(
     const Key & key_, size_t offset_, size_t size, std::lock_guard<std::mutex> & cache_lock)
 {
@@ -364,7 +387,8 @@ bool LRUFileCache::tryReserve(
 
     auto is_overflow = [&]
     {
-        return (current_size + size - removed_size > max_size)
+        /// max_size == 0 means unlimited cache size, max_element_size means unlimited number of cache elements.
+        return (max_size != 0 && current_size + size - removed_size > max_size)
             || (max_element_size != 0 && queue_size > max_element_size);
     };
 
@@ -476,6 +500,30 @@ void LRUFileCache::remove(const Key & key)
         fs::remove(key_path);
 }
 
+void LRUFileCache::tryRemoveAll()
+{
+    /// Try remove all cached files by cache_base_path.
+    /// Only releasable file segments are evicted.
+
+    std::lock_guard cache_lock(mutex);
+
+    for (auto it = queue.begin(); it != queue.end();)
+    {
+        auto & [key, offset] = *it++;
+
+        auto * cell = getCell(key, offset, cache_lock);
+        if (cell->releasable())
+        {
+            auto file_segment = cell->file_segment;
+            if (file_segment)
+            {
+                std::lock_guard<std::mutex> segment_lock(file_segment->mutex);
+                remove(file_segment->key(), file_segment->offset(), cache_lock, segment_lock);
+            }
+        }
+    }
+}
+
 void LRUFileCache::remove(
     Key key, size_t offset,
     std::lock_guard<std::mutex> & cache_lock, std::lock_guard<std::mutex> & /* segment_lock */)
@@ -523,8 +571,8 @@ void LRUFileCache::loadCacheInfoIntoMemory()
     std::lock_guard cache_lock(mutex);
 
     Key key;
-    UInt64 offset;
-    size_t size;
+    UInt64 offset = 0;
+    size_t size = 0;
     std::vector<FileSegmentCell *> cells;
 
     /// cache_base_path / key_prefix / key / offset
@@ -660,6 +708,38 @@ bool LRUFileCache::isLastFileSegmentHolder(
     return cell->file_segment.use_count() == 2;
 }
 
+FileSegments LRUFileCache::getSnapshot() const
+{
+    std::lock_guard cache_lock(mutex);
+
+    FileSegments file_segments;
+
+    for (const auto & [key, cells_by_offset] : files)
+    {
+        for (const auto & [offset, cell] : cells_by_offset)
+            file_segments.push_back(FileSegment::getSnapshot(cell.file_segment, cache_lock));
+    }
+
+    return file_segments;
+}
+
+std::vector<String> LRUFileCache::tryGetCachePaths(const Key & key)
+{
+    std::lock_guard cache_lock(mutex);
+
+    std::vector<String> cache_paths;
+
+    const auto & cells_by_offset = files[key];
+
+    for (const auto & [offset, cell] : cells_by_offset)
+    {
+        if (cell.file_segment->state() == FileSegment::State::DOWNLOADED)
+            cache_paths.push_back(getPathInLocalCache(key, offset));
+    }
+
+    return cache_paths;
+}
+
 LRUFileCache::FileSegmentCell::FileSegmentCell(FileSegmentPtr file_segment_, LRUQueue & queue_)
     : file_segment(file_segment_)
 {
@@ -677,32 +757,43 @@ LRUFileCache::FileSegmentCell::FileSegmentCell(FileSegmentPtr file_segment_, LRU
             break;
         }
         case FileSegment::State::EMPTY:
+        case FileSegment::State::DOWNLOADING:
         {
             break;
         }
         default:
             throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
-                            "Can create cell with either DOWNLOADED or EMPTY state, got: {}",
+                            "Can create cell with either EMPTY, DOWNLOADED, DOWNLOADING state, got: {}",
                             FileSegment::stateToString(file_segment->download_state));
     }
 }
 
-String LRUFileCache::dumpStructure(const Key & key_)
+String LRUFileCache::dumpStructure(const Key & key)
 {
     std::lock_guard cache_lock(mutex);
+    return dumpStructureImpl(key, cache_lock);
+}
 
+String LRUFileCache::dumpStructureImpl(const Key & key, std::lock_guard<std::mutex> & /* cache_lock */)
+{
     WriteBufferFromOwnString result;
-    for (auto it = queue.begin(); it != queue.end(); ++it)
-    {
-        auto [key, offset] = *it;
-        if (key == key_)
-        {
-            auto * cell = getCell(key, offset, cache_lock);
-            result << (it != queue.begin() ? ", " : "") << cell->file_segment->range().toString();
-            result << "(state: " << cell->file_segment->download_state << ")";
-        }
-    }
+    const auto & cells_by_offset = files[key];
+
+    for (const auto & [offset, cell] : cells_by_offset)
+        result << cell.file_segment->getInfoForLog() << "\n";
+
     return result.str();
 }
 
+void LRUFileCache::assertCacheCorrectness(const Key & key, std::lock_guard<std::mutex> & /* cache_lock */)
+{
+    const auto & cells_by_offset = files[key];
+
+    for (const auto & [_, cell] : cells_by_offset)
+    {
+        const auto & file_segment = cell.file_segment;
+        file_segment->assertCorrectness();
+    }
+}
+
 }

src/Common/FileCache.h

@@ -25,6 +25,7 @@ namespace DB
 class IFileCache : private boost::noncopyable
 {
 friend class FileSegment;
+friend struct FileSegmentsHolder;
 
 public:
     using Key = UInt128;
@@ -32,9 +33,7 @@ public:
 
     IFileCache(
         const String & cache_base_path_,
-        size_t max_size_,
-        size_t max_element_size_,
-        size_t max_file_segment_size_);
+        const FileCacheSettings & cache_settings_);
 
     virtual ~IFileCache() = default;
 
@@ -43,7 +42,9 @@ public:
 
     virtual void remove(const Key & key) = 0;
 
-    static bool shouldBypassCache();
+    virtual void tryRemoveAll() = 0;
+
+    static bool isReadOnly();
 
     /// Cache capacity in bytes.
     size_t capacity() const { return max_size; }
@@ -54,6 +55,10 @@ public:
 
     String getPathInLocalCache(const Key & key);
 
+    const String & getBasePath() const { return cache_base_path; }
+
+    virtual std::vector<String> tryGetCachePaths(const Key & key) = 0;
+
     /**
      * Given an `offset` and `size` representing [offset, offset + size) bytes interval,
      * return list of cached non-overlapping non-empty
@@ -67,6 +72,10 @@ public:
      */
     virtual FileSegmentsHolder getOrSet(const Key & key, size_t offset, size_t size) = 0;
 
+    virtual FileSegmentsHolder setDownloading(const Key & key, size_t offset, size_t size) = 0;
+
+    virtual FileSegments getSnapshot() const = 0;
+
     /// For debug.
     virtual String dumpStructure(const Key & key) = 0;
 
@@ -111,16 +120,22 @@ class LRUFileCache final : public IFileCache
 public:
     LRUFileCache(
         const String & cache_base_path_,
-        size_t max_size_,
-        size_t max_element_size_ = REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_ELEMENTS,
-        size_t max_file_segment_size_ = REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_FILE_SEGMENT_SIZE);
+        const FileCacheSettings & cache_settings_);
 
     FileSegmentsHolder getOrSet(const Key & key, size_t offset, size_t size) override;
 
+    FileSegments getSnapshot() const override;
+
+    FileSegmentsHolder setDownloading(const Key & key, size_t offset, size_t size) override;
+
     void initialize() override;
 
     void remove(const Key & key) override;
 
+    void tryRemoveAll() override;
+
+    std::vector<String> tryGetCachePaths(const Key & key) override;
+
 private:
     using FileKeyAndOffset = std::pair<Key, size_t>;
     using LRUQueue = std::list<FileKeyAndOffset>;
@@ -193,8 +208,10 @@ private:
 
     void loadCacheInfoIntoMemory();
 
-    FileSegments splitRangeIntoEmptyCells(
-        const Key & key, size_t offset, size_t size, std::lock_guard<std::mutex> & cache_lock);
+    FileSegments splitRangeIntoCells(
+        const Key & key, size_t offset, size_t size, FileSegment::State state, std::lock_guard<std::mutex> & cache_lock);
+
+    String dumpStructureImpl(const Key & key_, std::lock_guard<std::mutex> & cache_lock);
 
 public:
     struct Stat
@@ -208,6 +225,7 @@ public:
     Stat getStat();
 
     String dumpStructure(const Key & key_) override;
+    void assertCacheCorrectness(const Key & key, std::lock_guard<std::mutex> & cache_lock);
 };
 
 }

src/Common/FileCacheFactory.cpp

@@ -15,28 +15,53 @@ FileCacheFactory & FileCacheFactory::instance()
     return ret;
 }
 
-FileCachePtr FileCacheFactory::getImpl(const std::string & cache_base_path, std::lock_guard<std::mutex> &)
+FileCacheFactory::CacheByBasePath FileCacheFactory::getAll()
+{
+    std::lock_guard lock(mutex);
+    return caches;
+}
+
+const FileCacheSettings & FileCacheFactory::getSettings(const std::string & cache_base_path)
+{
+    std::lock_guard lock(mutex);
+
+    auto * cache_data = getImpl(cache_base_path, lock);
+    if (cache_data)
+        return cache_data->settings;
+
+    throw Exception(ErrorCodes::BAD_ARGUMENTS, "No cache found by path: {}", cache_base_path);
+}
+
+FileCacheFactory::CacheData * FileCacheFactory::getImpl(const std::string & cache_base_path, std::lock_guard<std::mutex> &)
 {
     auto it = caches.find(cache_base_path);
     if (it == caches.end())
         return nullptr;
-    return it->second;
+    return &it->second;
+}
+
+FileCachePtr FileCacheFactory::get(const std::string & cache_base_path)
+{
+    std::lock_guard lock(mutex);
+
+    auto * cache_data = getImpl(cache_base_path, lock);
+    if (cache_data)
+        return cache_data->cache;
+
+    throw Exception(ErrorCodes::BAD_ARGUMENTS, "No cache found by path: {}", cache_base_path);
 }
 
 FileCachePtr FileCacheFactory::getOrCreate(
-    const std::string & cache_base_path, size_t max_size, size_t max_elements_size, size_t max_file_segment_size)
+    const std::string & cache_base_path, const FileCacheSettings & file_cache_settings)
 {
     std::lock_guard lock(mutex);
-    auto cache = getImpl(cache_base_path, lock);
-    if (cache)
-    {
-        if (cache->capacity() != max_size)
-            throw Exception(ErrorCodes::BAD_ARGUMENTS, "Cache with path `{}` already exists, but has different max size", cache_base_path);
-        return cache;
-    }
 
-    cache = std::make_shared<LRUFileCache>(cache_base_path, max_size, max_elements_size, max_file_segment_size);
-    caches.emplace(cache_base_path, cache);
+    auto * cache_data = getImpl(cache_base_path, lock);
+    if (cache_data)
+        return cache_data->cache;
+
+    auto cache = std::make_shared<LRUFileCache>(cache_base_path, file_cache_settings);
+    caches.emplace(cache_base_path, CacheData(cache, file_cache_settings));
     return cache;
 }
 

src/Common/FileCacheFactory.h

@@ -1,6 +1,7 @@
 #pragma once
 
 #include <Common/FileCache_fwd.h>
+#include <Common/FileCacheSettings.h>
 
 #include <boost/noncopyable.hpp>
 #include <unordered_map>
@@ -14,16 +15,32 @@ namespace DB
  */
 class FileCacheFactory final : private boost::noncopyable
 {
+    struct CacheData
+    {
+        FileCachePtr cache;
+        FileCacheSettings settings;
+
+        CacheData(FileCachePtr cache_, const FileCacheSettings & settings_) : cache(cache_), settings(settings_) {}
+    };
+
+    using CacheByBasePath = std::unordered_map<std::string, CacheData>;
+
 public:
     static FileCacheFactory & instance();
 
-    FileCachePtr getOrCreate(const std::string & cache_base_path, size_t max_size, size_t max_elements_size, size_t max_file_segment_size);
+    FileCachePtr getOrCreate(const std::string & cache_base_path, const FileCacheSettings & file_cache_settings);
+
+    FileCachePtr get(const std::string & cache_base_path);
+
+    CacheByBasePath getAll();
+
+    const FileCacheSettings & getSettings(const std::string & cache_base_path);
 
 private:
-    FileCachePtr getImpl(const std::string & cache_base_path, std::lock_guard<std::mutex> &);
+    CacheData * getImpl(const std::string & cache_base_path, std::lock_guard<std::mutex> &);
 
     std::mutex mutex;
-    std::unordered_map<std::string, FileCachePtr> caches;
+    CacheByBasePath caches;
 };
 
 }

src/Common/FileCacheSettings.cpp

@@ -0,0 +1,16 @@
+#include "FileCacheSettings.h"
+
+#include <Poco/Util/AbstractConfiguration.h>
+
+namespace DB
+{
+
+void FileCacheSettings::loadFromConfig(const Poco::Util::AbstractConfiguration & config, const std::string & config_prefix)
+{
+    max_size = config.getUInt64(config_prefix + ".data_cache_max_size", REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_CACHE_SIZE);
+    max_elements = config.getUInt64(config_prefix + ".data_cache_max_elements", REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_ELEMENTS);
+    max_file_segment_size = config.getUInt64(config_prefix + ".max_file_segment_size", REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_FILE_SEGMENT_SIZE);
+    cache_on_write_operations = config.getUInt64(config_prefix + ".cache_on_write_operations", false);
+}
+
+}

src/Common/FileCacheSettings.h

@@ -0,0 +1,20 @@
+#pragma once
+
+#include <Common/FileCache_fwd.h>
+
+namespace Poco { namespace Util { class AbstractConfiguration; } }
+
+namespace DB
+{
+
+struct FileCacheSettings
+{
+    size_t max_size = 0;
+    size_t max_elements = REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_ELEMENTS;
+    size_t max_file_segment_size = REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_FILE_SEGMENT_SIZE;
+    bool cache_on_write_operations = false;
+
+    void loadFromConfig(const Poco::Util::AbstractConfiguration & config, const std::string & config_prefix);
+};
+
+}

src/Common/FileCache_fwd.h

@@ -4,10 +4,13 @@
 namespace DB
 {
 
+static constexpr int REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_CACHE_SIZE = 1024 * 1024 * 1024;
 static constexpr int REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_FILE_SEGMENT_SIZE = 100 * 1024 * 1024;
 static constexpr int REMOTE_FS_OBJECTS_CACHE_DEFAULT_MAX_ELEMENTS = 1024 * 1024;
 
 class IFileCache;
 using FileCachePtr = std::shared_ptr<IFileCache>;
 
+struct FileCacheSettings;
+
 }

src/Common/FileSegment.cpp

@@ -31,10 +31,34 @@ FileSegment::FileSegment(
     , log(&Poco::Logger::get("FileSegment"))
 #endif
 {
-    if (download_state == State::DOWNLOADED)
-        reserved_size = downloaded_size = size_;
-    else if (download_state != State::EMPTY)
-        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Can create cell with either DOWNLOADED or EMPTY state");
+    /// On creation, file segment state can be EMPTY, DOWNLOADED, DOWNLOADING.
+    switch (download_state)
+    {
+        /// EMPTY is used when file segment is not in cache and
+        /// someone will _potentially_ want to download it (after calling getOrSetDownloader()).
+        case (State::EMPTY):
+        {
+            break;
+        }
+        /// DOWNLOADED is used either on initial cache metadata load into memory on server startup
+        /// or on reduceSizeToDownloaded() -- when file segment object is updated.
+        case (State::DOWNLOADED):
+        {
+            reserved_size = downloaded_size = size_;
+            break;
+        }
+        /// DOWNLOADING is used only for write-through caching (e.g. getOrSetDownloader() is not
+        /// needed, downloader is set on file segment creation).
+        case (State::DOWNLOADING):
+        {
+            downloader_id = getCallerId();
+            break;
+        }
+        default:
+        {
+            throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Can create cell with either EMPTY, DOWNLOADED, DOWNLOADING state");
+        }
+    }
 }
 
 FileSegment::State FileSegment::state() const
@@ -49,6 +73,12 @@ size_t FileSegment::getDownloadOffset() const
     return range().left + getDownloadedSize(segment_lock);
 }
 
+size_t FileSegment::getDownloadedSize() const
+{
+    std::lock_guard segment_lock(mutex);
+    return getDownloadedSize(segment_lock);
+}
+
 size_t FileSegment::getDownloadedSize(std::lock_guard<std::mutex> & /* segment_lock */) const
 {
     if (download_state == State::DOWNLOADED)
@@ -60,24 +90,15 @@ size_t FileSegment::getDownloadedSize(std::lock_guard<std::mutex> & /* segment_l
 
 String FileSegment::getCallerId()
 {
-    return getCallerIdImpl(false);
+    return getCallerIdImpl();
 }
 
-String FileSegment::getCallerIdImpl(bool allow_non_strict_checking)
+String FileSegment::getCallerIdImpl()
 {
-    if (IFileCache::shouldBypassCache())
-    {
-        /// getCallerId() can be called from completeImpl(), which can be called from complete().
-        /// complete() is called from destructor of CachedReadBufferFromRemoteFS when there is no query id anymore.
-        /// Allow non strict checking in this case. This works correctly as if getCallerIdImpl() is called from destructor,
-        /// then we know that caller is not a downloader, because downloader is reset each nextImpl() call either
-        /// manually or via SCOPE_EXIT.
-
-        if (allow_non_strict_checking)
-            return "None";
-
-        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Cannot use cache without query id");
-    }
+    if (!CurrentThread::isInitialized()
+        || !CurrentThread::get().getQueryContext()
+        || CurrentThread::getQueryId().size == 0)
+        return "None:" + toString(getThreadId());
 
     return CurrentThread::getQueryId().toString() + ":" + toString(getThreadId());
 }
@@ -136,7 +157,6 @@ String FileSegment::getDownloader() const
 bool FileSegment::isDownloader() const
 {
     std::lock_guard segment_lock(mutex);
-    LOG_TEST(log, "Checking for current downloader. Caller: {}, downloader: {}, current state: {}", getCallerId(), downloader_id, stateToString(download_state));
     return getCallerId() == downloader_id;
 }
 
@@ -159,7 +179,18 @@ void FileSegment::setRemoteFileReader(RemoteFileReaderPtr remote_file_reader_)
     remote_file_reader = remote_file_reader_;
 }
 
-void FileSegment::write(const char * from, size_t size)
+void FileSegment::resetRemoteFileReader()
+{
+    if (!isDownloader())
+        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Only downloader can use remote filesystem file reader");
+
+    if (!remote_file_reader)
+        throw Exception(ErrorCodes::LOGICAL_ERROR, "Remote file reader does not exist");
+
+    remote_file_reader.reset();
+}
+
+void FileSegment::write(const char * from, size_t size, size_t offset_)
 {
     if (!size)
         throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Writing zero size is not allowed");
@@ -174,8 +205,24 @@ void FileSegment::write(const char * from, size_t size)
                         "Only downloader can do the downloading. (CallerId: {}, DownloaderId: {})",
                         getCallerId(), downloader_id);
 
+    if (downloaded_size == range().size())
+        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
+                        "Attempt to write {} bytes to offset: {}, but current file segment is already fully downloaded",
+                        size, offset_);
+
+    auto download_offset = range().left + downloaded_size;
+    if (offset_ != download_offset)
+        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
+                        "Attempt to write {} bytes to offset: {}, but current download offset is {}",
+                        size, offset_, download_offset);
+
     if (!cache_writer)
     {
+        if (downloaded_size > 0)
+            throw Exception(ErrorCodes::LOGICAL_ERROR,
+                            "Cache writer was finalized (downloaded size: {}, state: {})",
+                            downloaded_size, stateToString(download_state));
+
         auto download_path = cache->getPathInLocalCache(key(), offset());
         cache_writer = std::make_unique<WriteBufferFromFile>(download_path);
     }
@@ -190,19 +237,91 @@ void FileSegment::write(const char * from, size_t size)
 
         downloaded_size += size;
     }
-    catch (...)
+    catch (Exception & e)
     {
         std::lock_guard segment_lock(mutex);
 
-        LOG_ERROR(log, "Failed to write to cache. File segment info: {}", getInfoForLogImpl(segment_lock));
+        wrapWithCacheInfo(e, "while writing into cache", segment_lock);
 
-        download_state = State::PARTIALLY_DOWNLOADED_NO_CONTINUATION;
+        setDownloadFailed(segment_lock);
 
-        cache_writer->finalize();
-        cache_writer.reset();
+        cv.notify_all();
 
         throw;
     }
+
+    assert(getDownloadOffset() == offset_ + size);
+}
+
+void FileSegment::writeInMemory(const char * from, size_t size)
+{
+    if (!size)
+        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Attempt to write zero size cache file");
+
+    if (availableSize() < size)
+        throw Exception(
+            ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
+            "Not enough space is reserved. Available: {}, expected: {}", availableSize(), size);
+
+    std::lock_guard segment_lock(mutex);
+
+    if (cache_writer)
+        throw Exception(ErrorCodes::LOGICAL_ERROR, "Cache writer already initialized");
+
+    auto download_path = cache->getPathInLocalCache(key(), offset());
+    cache_writer = std::make_unique<WriteBufferFromFile>(download_path, size + 1);
+
+    try
+    {
+        cache_writer->write(from, size);
+    }
+    catch (Exception & e)
+    {
+        wrapWithCacheInfo(e, "while writing into cache", segment_lock);
+
+        setDownloadFailed(segment_lock);
+
+        cv.notify_all();
+
+        throw;
+    }
+}
+
+size_t FileSegment::finalizeWrite()
+{
+    std::lock_guard segment_lock(mutex);
+
+    if (!cache_writer)
+        throw Exception(ErrorCodes::LOGICAL_ERROR, "Cache writer not initialized");
+
+    size_t size = cache_writer->offset();
+
+    if (size == 0)
+        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR, "Writing size is not allowed");
+
+    try
+    {
+        cache_writer->next();
+    }
+    catch (Exception & e)
+    {
+        wrapWithCacheInfo(e, "while writing into cache", segment_lock);
+
+        setDownloadFailed(segment_lock);
+
+        cv.notify_all();
+
+        throw;
+    }
+
+    downloaded_size += size;
+
+    if (downloaded_size != range().size())
+        throw Exception(ErrorCodes::LOGICAL_ERROR, "Expected downloaded size to equal file segment size ({} == {})", downloaded_size, range().size());
+
+    setDownloaded(segment_lock);
+
+    return size;
 }
 
 FileSegment::State FileSegment::wait()
@@ -269,8 +388,21 @@ void FileSegment::setDownloaded(std::lock_guard<std::mutex> & /* segment_lock */
 {
     download_state = State::DOWNLOADED;
     is_downloaded = true;
+    downloader_id.clear();
+
+    if (cache_writer)
+    {
+        cache_writer->finalize();
+        cache_writer.reset();
+        remote_file_reader.reset();
+    }
+}
+
+void FileSegment::setDownloadFailed(std::lock_guard<std::mutex> & /* segment_lock */)
+{
+    download_state = State::PARTIALLY_DOWNLOADED_NO_CONTINUATION;
+    downloader_id.clear();
 
-    assert(cache_writer);
     if (cache_writer)
     {
         cache_writer->finalize();
@@ -299,107 +431,125 @@ void FileSegment::completeBatchAndResetDownloader()
 
 void FileSegment::complete(State state)
 {
-    {
-        std::lock_guard segment_lock(mutex);
-
-        bool is_downloader = downloader_id == getCallerId();
-        if (!is_downloader)
-        {
-            cv.notify_all();
-            throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
-                            "File segment can be completed only by downloader or downloader's FileSegmentsHodler");
-        }
-
-        if (state != State::DOWNLOADED
-            && state != State::PARTIALLY_DOWNLOADED
-            && state != State::PARTIALLY_DOWNLOADED_NO_CONTINUATION)
-        {
-            cv.notify_all();
-            throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
-                            "Cannot complete file segment with state: {}", stateToString(state));
-        }
-
-        download_state = state;
-    }
-
-    completeImpl();
-    cv.notify_all();
-}
-
-void FileSegment::complete()
-{
-    {
-        std::lock_guard segment_lock(mutex);
-
-        if (download_state == State::SKIP_CACHE || detached)
-            return;
-
-        if (download_state != State::DOWNLOADED && getDownloadedSize(segment_lock) == range().size())
-            setDownloaded(segment_lock);
-
-        if (download_state == State::DOWNLOADING || download_state == State::EMPTY)
-            download_state = State::PARTIALLY_DOWNLOADED;
-    }
-
-    completeImpl(true);
-    cv.notify_all();
-}
-
-void FileSegment::completeImpl(bool allow_non_strict_checking)
-{
-    /// cache lock is always taken before segment lock.
     std::lock_guard cache_lock(cache->mutex);
     std::lock_guard segment_lock(mutex);
 
-    bool download_can_continue = false;
-
-    if (download_state == State::PARTIALLY_DOWNLOADED
-                || download_state == State::PARTIALLY_DOWNLOADED_NO_CONTINUATION)
+    bool is_downloader = downloader_id == getCallerId();
+    if (!is_downloader)
     {
-        bool is_last_holder = cache->isLastFileSegmentHolder(key(), offset(), cache_lock, segment_lock);
-        download_can_continue = !is_last_holder && download_state == State::PARTIALLY_DOWNLOADED;
+        cv.notify_all();
+        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
+                        "File segment can be completed only by downloader or downloader's FileSegmentsHodler");
+    }
 
-        if (!download_can_continue)
+    if (state != State::DOWNLOADED
+        && state != State::PARTIALLY_DOWNLOADED
+        && state != State::PARTIALLY_DOWNLOADED_NO_CONTINUATION)
+    {
+        cv.notify_all();
+        throw Exception(ErrorCodes::REMOTE_FS_OBJECT_CACHE_ERROR,
+                        "Cannot complete file segment with state: {}", stateToString(state));
+    }
+
+    download_state = state;
+
+    try
+    {
+        completeImpl(cache_lock, segment_lock);
+    }
+    catch (...)
+    {
+        if (!downloader_id.empty() && downloader_id == getCallerIdImpl())
+            downloader_id.clear();
+
+        cv.notify_all();
+        throw;
+    }
+
+    cv.notify_all();
+}
+
+void FileSegment::complete(std::lock_guard<std::mutex> & cache_lock)
+{
+    std::lock_guard segment_lock(mutex);
+
+    if (download_state == State::SKIP_CACHE || detached)
+        return;
+
+    if (download_state != State::DOWNLOADED && getDownloadedSize(segment_lock) == range().size())
+        setDownloaded(segment_lock);
+
+    if (download_state == State::DOWNLOADING || download_state == State::EMPTY)
+    {
+        /// Segment state can be changed from DOWNLOADING or EMPTY only if the caller is the
+        /// downloader or the only owner of the segment.
+
+        bool can_update_segment_state = downloader_id == getCallerIdImpl()
+            || cache->isLastFileSegmentHolder(key(), offset(), cache_lock, segment_lock);
+
+        if (can_update_segment_state)
+            download_state = State::PARTIALLY_DOWNLOADED;
+    }
+
+    try
+    {
+        completeImpl(cache_lock, segment_lock);
+    }
+    catch (...)
+    {
+        if (!downloader_id.empty() && downloader_id == getCallerIdImpl())
+            downloader_id.clear();
+
+        cv.notify_all();
+        throw;
+    }
+
+    cv.notify_all();
+}
+
+void FileSegment::completeImpl(std::lock_guard<std::mutex> & cache_lock, std::lock_guard<std::mutex> & segment_lock)
+{
+    bool is_last_holder = cache->isLastFileSegmentHolder(key(), offset(), cache_lock, segment_lock);
+
+    if (is_last_holder
+        && (download_state == State::PARTIALLY_DOWNLOADED || download_state == State::PARTIALLY_DOWNLOADED_NO_CONTINUATION))
+    {
+        size_t current_downloaded_size = getDownloadedSize(segment_lock);
+        if (current_downloaded_size == 0)
         {
-            size_t current_downloaded_size = getDownloadedSize(segment_lock);
-            if (current_downloaded_size == 0)
-            {
-                download_state = State::SKIP_CACHE;
-                LOG_TEST(log, "Remove cell {} (nothing downloaded)", range().toString());
-                cache->remove(key(), offset(), cache_lock, segment_lock);
+            download_state = State::SKIP_CACHE;
+            LOG_TEST(log, "Remove cell {} (nothing downloaded)", range().toString());
+            cache->remove(key(), offset(), cache_lock, segment_lock);
+        }
+        else
+        {
+            /**
+            * Only last holder of current file segment can resize the cell,
+            * because there is an invariant that file segments returned to users
+            * in FileSegmentsHolder represent a contiguous range, so we can resize
+            * it only when nobody needs it.
+            */
+            LOG_TEST(log, "Resize cell {} to downloaded: {}", range().toString(), current_downloaded_size);
+            cache->reduceSizeToDownloaded(key(), offset(), cache_lock, segment_lock);
+        }
 
-                detached = true;
-            }
-            else if (is_last_holder)
-            {
-                /**
-                * Only last holder of current file segment can resize the cell,
-                * because there is an invariant that file segments returned to users
-                * in FileSegmentsHolder represent a contiguous range, so we can resize
-                * it only when nobody needs it.
-                */
-                LOG_TEST(log, "Resize cell {} to downloaded: {}", range().toString(), current_downloaded_size);
-                cache->reduceSizeToDownloaded(key(), offset(), cache_lock, segment_lock);
+        detached = true;
 
-                detached = true;
-            }
+        if (cache_writer)
+        {
+            cache_writer->finalize();
+            cache_writer.reset();
+            remote_file_reader.reset();
         }
     }
 
-    if (!downloader_id.empty() && downloader_id == getCallerIdImpl(allow_non_strict_checking))
+    if (!downloader_id.empty() && (downloader_id == getCallerIdImpl() || is_last_holder))
     {
         LOG_TEST(log, "Clearing downloader id: {}, current state: {}", downloader_id, stateToString(download_state));
         downloader_id.clear();
     }
 
-    if (!download_can_continue && cache_writer)
-    {
-        cache_writer->finalize();
-        cache_writer.reset();
-        remote_file_reader.reset();
-    }
-
-    assert(download_state != FileSegment::State::DOWNLOADED || std::filesystem::file_size(cache->getPathInLocalCache(key(), offset())) > 0);
+    assertCorrectnessImpl(segment_lock);
 }
 
 String FileSegment::getInfoForLog() const
@@ -420,6 +570,11 @@ String FileSegment::getInfoForLogImpl(std::lock_guard<std::mutex> & segment_lock
     return info.str();
 }
 
+void FileSegment::wrapWithCacheInfo(Exception & e, const String & message, std::lock_guard<std::mutex> & segment_lock) const
+{
+    e.addMessage(fmt::format("{}, current cache state: {}", message, getInfoForLogImpl(segment_lock)));
+}
+
 String FileSegment::stateToString(FileSegment::State state)
 {
     switch (state)
@@ -440,6 +595,70 @@ String FileSegment::stateToString(FileSegment::State state)
     __builtin_unreachable();
 }
 
+void FileSegment::assertCorrectness() const
+{
+    std::lock_guard segment_lock(mutex);
+    assertCorrectnessImpl(segment_lock);
+}
+
+void FileSegment::assertCorrectnessImpl(std::lock_guard<std::mutex> & /* segment_lock */) const
+{
+    assert(downloader_id.empty() == (download_state != FileSegment::State::DOWNLOADING));
+    assert(!downloader_id.empty() == (download_state == FileSegment::State::DOWNLOADING));
+    assert(download_state != FileSegment::State::DOWNLOADED || std::filesystem::file_size(cache->getPathInLocalCache(key(), offset())) > 0);
+}
+
+FileSegmentPtr FileSegment::getSnapshot(const FileSegmentPtr & file_segment, std::lock_guard<std::mutex> & /* cache_lock */)
+{
+    auto snapshot = std::make_shared<FileSegment>(
+        file_segment->offset(),
+        file_segment->range().size(),
+        file_segment->key(),
+        nullptr,
+        State::EMPTY);
+
+    snapshot->hits_count = file_segment->getHitsCount();
+    snapshot->ref_count = file_segment.use_count();
+    snapshot->downloaded_size = file_segment->getDownloadedSize();
+    snapshot->download_state = file_segment->state();
+
+    return snapshot;
+}
+
+FileSegmentsHolder::~FileSegmentsHolder()
+{
+    /// In CacheableReadBufferFromRemoteFS file segment's downloader removes file segments from
+    /// FileSegmentsHolder right after calling file_segment->complete(), so on destruction here
+    /// remain only uncompleted file segments.
+
+    IFileCache * cache = nullptr;
+
+    for (auto file_segment_it = file_segments.begin(); file_segment_it != file_segments.end();)
+    {
+        auto current_file_segment_it = file_segment_it;
+        auto & file_segment = *current_file_segment_it;
+
+        if (!cache)
+            cache = file_segment->cache;
+
+        try
+        {
+            /// File segment pointer must be reset right after calling complete() and
+            /// under the same mutex, because complete() checks for segment pointers.
+            std::lock_guard cache_lock(cache->mutex);
+
+            file_segment->complete(cache_lock);
+
+            file_segment_it = file_segments.erase(current_file_segment_it);
+        }
+        catch (...)
+        {
+            tryLogCurrentException(__PRETTY_FUNCTION__);
+            assert(false);
+        }
+    }
+}
+
 String FileSegmentsHolder::toString()
 {
     String ranges;

src/Common/FileSegment.h

@@ -95,12 +95,23 @@ public:
 
     bool reserve(size_t size);
 
-    void write(const char * from, size_t size);
+    void write(const char * from, size_t size, size_t offset_);
+
+    /**
+     * writeInMemory and finalizeWrite are used together to write a single file with delay.
+     * Both can be called only once, one after another. Used for writing cache via threadpool
+     * on wrote operations. TODO: this solution is temporary, until adding a separate cache layer.
+     */
+    void writeInMemory(const char * from, size_t size);
+
+    size_t finalizeWrite();
 
     RemoteFileReaderPtr getRemoteFileReader();
 
     void setRemoteFileReader(RemoteFileReaderPtr remote_file_reader_);
 
+    void resetRemoteFileReader();
+
     String getOrSetDownloader();
 
     String getDownloader() const;
@@ -115,22 +126,51 @@ public:
 
     size_t getDownloadOffset() const;
 
+    size_t getDownloadedSize() const;
+
     void completeBatchAndResetDownloader();
 
     void complete(State state);
 
     String getInfoForLog() const;
 
+    size_t getHitsCount() const { return hits_count; }
+
+    size_t getRefCount() const { return ref_count; }
+
+    void incrementHitsCount() { ++hits_count; }
+
+    void assertCorrectness() const;
+
+    static FileSegmentPtr getSnapshot(const FileSegmentPtr & file_segment, std::lock_guard<std::mutex> & cache_lock);
+
 private:
     size_t availableSize() const { return reserved_size - downloaded_size; }
-    bool lastFileSegmentHolder() const;
-    void complete();
-    void completeImpl(bool allow_non_strict_checking = false);
-    void setDownloaded(std::lock_guard<std::mutex> & segment_lock);
-    static String getCallerIdImpl(bool allow_non_strict_checking = false);
-    void resetDownloaderImpl(std::lock_guard<std::mutex> & segment_lock);
+
     size_t getDownloadedSize(std::lock_guard<std::mutex> & segment_lock) const;
     String getInfoForLogImpl(std::lock_guard<std::mutex> & segment_lock) const;
+    void assertCorrectnessImpl(std::lock_guard<std::mutex> & segment_lock) const;
+
+    void setDownloaded(std::lock_guard<std::mutex> & segment_lock);
+    void setDownloadFailed(std::lock_guard<std::mutex> & segment_lock);
+
+    void wrapWithCacheInfo(Exception & e, const String & message, std::lock_guard<std::mutex> & segment_lock) const;
+
+    bool lastFileSegmentHolder() const;
+
+    /// complete() without any completion state is called from destructor of
+    /// FileSegmentsHolder. complete() might check if the caller of the method
+    /// is the last alive holder of the segment. Therefore, complete() and destruction
+    /// of the file segment pointer must be done under the same cache mutex.
+    void complete(std::lock_guard<std::mutex> & cache_lock);
+
+    void completeImpl(
+        std::lock_guard<std::mutex> & cache_lock,
+        std::lock_guard<std::mutex> & segment_lock);
+
+    static String getCallerIdImpl();
+
+    void resetDownloaderImpl(std::lock_guard<std::mutex> & segment_lock);
 
     const Range segment_range;
 
@@ -162,6 +202,8 @@ private:
     bool detached = false;
 
     std::atomic<bool> is_downloaded{false};
+    std::atomic<size_t> hits_count = 0; /// cache hits.
+    std::atomic<size_t> ref_count = 0; /// Used for getting snapshot state
 };
 
 struct FileSegmentsHolder : private boost::noncopyable
@@ -169,28 +211,7 @@ struct FileSegmentsHolder : private boost::noncopyable
     explicit FileSegmentsHolder(FileSegments && file_segments_) : file_segments(std::move(file_segments_)) {}
     FileSegmentsHolder(FileSegmentsHolder && other) : file_segments(std::move(other.file_segments)) {}
 
-    ~FileSegmentsHolder()
-    {
-        /// In CacheableReadBufferFromRemoteFS file segment's downloader removes file segments from
-        /// FileSegmentsHolder right after calling file_segment->complete(), so on destruction here
-        /// remain only uncompleted file segments.
-
-        for (auto & segment : file_segments)
-        {
-            try
-            {
-                segment->complete();
-            }
-            catch (...)
-            {
-#ifndef NDEBUG
-                throw;
-#else
-                tryLogCurrentException(__PRETTY_FUNCTION__);
-#endif
-            }
-        }
-    }
+    ~FileSegmentsHolder();
 
     FileSegments file_segments{};
 

src/Common/JSONBuilder.h

@@ -61,7 +61,7 @@ private:
 class JSONBool : public IItem
 {
 public:
-    explicit JSONBool(bool value_) : value(std::move(value_)) {}
+    explicit JSONBool(bool value_) : value(value_) {}
     void format(const FormatSettings & settings, FormatContext & context) override;
 
 private:
@@ -74,7 +74,7 @@ public:
     void add(ItemPtr value) { values.push_back(std::move(value)); }
     void add(std::string value) { add(std::make_unique<JSONString>(std::move(value))); }
     void add(const char * value) { add(std::make_unique<JSONString>(value)); }
-    void add(bool value) { add(std::make_unique<JSONBool>(std::move(value))); }
+    void add(bool value) { add(std::make_unique<JSONBool>(value)); }
 
     template <typename T>
     requires std::is_arithmetic_v<T>
@@ -99,7 +99,7 @@ public:
     void add(std::string key, std::string value) { add(std::move(key), std::make_unique<JSONString>(std::move(value))); }
     void add(std::string key, const char * value) { add(std::move(key), std::make_unique<JSONString>(value)); }
     void add(std::string key, std::string_view value) { add(std::move(key), std::make_unique<JSONString>(value)); }
-    void add(std::string key, bool value) { add(std::move(key), std::make_unique<JSONBool>(std::move(value))); }
+    void add(std::string key, bool value) { add(std::move(key), std::make_unique<JSONBool>(value)); }
 
     template <typename T>
     requires std::is_arithmetic_v<T>

src/Common/NamePrompter.cpp

@@ -0,0 +1,15 @@
+#include <IO/WriteHelpers.h>
+#include <Common/NamePrompter.h>
+
+namespace DB::detail
+{
+void appendHintsMessageImpl(String & message, const std::vector<String> & hints)
+{
+    if (hints.empty())
+    {
+        return;
+    }
+
+    message += ". Maybe you meant: " + toString(hints);
+}
+}

src/Common/NamePrompter.h

@@ -90,6 +90,10 @@ private:
     }
 };
 
+namespace detail
+{
+void appendHintsMessageImpl(String & message, const std::vector<String> & hints);
+}
 
 template <size_t MaxNumHints, typename Self>
 class IHints
@@ -102,6 +106,12 @@ public:
         return prompter.getHints(name, getAllRegisteredNames());
     }
 
+    void appendHintsMessage(String & message, const String & name) const
+    {
+        auto hints = getHints(name);
+        detail::appendHintsMessageImpl(message, hints);
+    }
+
     IHints() = default;
 
     IHints(const IHints &) = default;
@@ -114,5 +124,4 @@ public:
 private:
     NamePrompter<MaxNumHints> prompter;
 };
-
 }

src/Common/ProfileEvents.cpp

@@ -112,6 +112,8 @@
     M(CompileExpressionsMicroseconds, "Total time spent for compilation of expressions to LLVM code.") \
     M(CompileExpressionsBytes, "Number of bytes used for expressions compilation.") \
     \
+    M(ExecuteShellCommand, "Number of shell command executions.") \
+    \
     M(ExternalSortWritePart, "") \
     M(ExternalSortMerge, "") \
     M(ExternalAggregationWritePart, "") \
@@ -295,6 +297,25 @@
     M(MergeTreeMetadataCacheHit, "Number of times the read of meta file was done from MergeTree metadata cache") \
     M(MergeTreeMetadataCacheMiss, "Number of times the read of meta file was not done from MergeTree metadata cache") \
     \
+    M(KafkaRebalanceRevocations, "Number of partition revocations (the first stage of consumer group rebalance)") \
+    M(KafkaRebalanceAssignments, "Number of partition assignments (the final stage of consumer group rebalance)") \
+    M(KafkaRebalanceErrors, "Number of failed consumer group rebalances") \
+    M(KafkaMessagesPolled, "Number of Kafka messages polled from librdkafka to ClickHouse") \
+    M(KafkaMessagesRead, "Number of Kafka messages already processed by ClickHouse") \
+    M(KafkaMessagesFailed, "Number of Kafka messages ClickHouse failed to parse") \
+    M(KafkaRowsRead, "Number of rows parsed from Kafka messages") \
+    M(KafkaRowsRejected, "Number of parsed rows which were later rejected (due to rebalances / errors or similar reasons). Those rows will be consumed again after the rebalance.") \
+    M(KafkaDirectReads, "Number of direct selects from Kafka tables since server start") \
+    M(KafkaBackgroundReads, "Number of background reads populating materialized views from Kafka since server start") \
+    M(KafkaCommits, "Number of successful commits of consumed offsets to Kafka (normally should be the same as KafkaBackgroundReads)") \
+    M(KafkaCommitFailures, "Number of failed commits of consumed offsets to Kafka (usually is a sign of some data duplication)") \
+    M(KafkaConsumerErrors, "Number of errors reported by librdkafka during polls") \
+    M(KafkaWrites, "Number of writes (inserts) to Kafka tables ") \
+    M(KafkaRowsWritten, "Number of rows inserted into Kafka tables") \
+    M(KafkaProducerFlushes, "Number of explicit flushes to Kafka producer") \
+    M(KafkaMessagesProduced, "Number of messages produced to Kafka") \
+    M(KafkaProducerErrors, "Number of errors during producing the messages to Kafka") \
+    \
     M(ScalarSubqueriesGlobalCacheHit, "Number of times a read from a scalar subquery was done using the global cache") \
     M(ScalarSubqueriesLocalCacheHit, "Number of times a read from a scalar subquery was done using the local cache") \
     M(ScalarSubqueriesCacheMiss, "Number of times a read from a scalar subquery was not cached and had to be calculated completely")

src/Common/ShellCommand.cpp

@@ -29,6 +29,11 @@ namespace
     };
 }
 
+namespace ProfileEvents
+{
+    extern const Event ExecuteShellCommand;
+}
+
 namespace DB
 {
 
@@ -158,6 +163,7 @@ std::unique_ptr<ShellCommand> ShellCommand::executeImpl(
     const Config & config)
 {
     logCommand(filename, argv);
+    ProfileEvents::increment(ProfileEvents::ExecuteShellCommand);
 
 #if !defined(USE_MUSL)
     /** Here it is written that with a normal call `vfork`, there is a chance of deadlock in multithreaded programs,

src/Common/SystemLogBase.cpp

@@ -9,6 +9,7 @@
 #include <Interpreters/SessionLog.h>
 #include <Interpreters/TextLog.h>
 #include <Interpreters/TraceLog.h>
+#include <Interpreters/ProcessorsProfileLog.h>
 #include <Interpreters/ZooKeeperLog.h>
 
 #include <Common/MemoryTrackerBlockerInThread.h>

src/Common/SystemLogBase.h

@@ -24,6 +24,7 @@
     M(SessionLogElement) \
     M(TraceLogElement) \
     M(ZooKeeperLogElement) \
+    M(ProcessorProfileLogElement) \
     M(TextLogElement)
 
 namespace Poco

src/Common/TargetSpecific.cpp

@@ -1,4 +1,4 @@
-#include <Functions/TargetSpecific.h>
+#include <Common/TargetSpecific.h>
 
 #include <Common/CpuId.h>
 

src/Common/ThreadStatus.h

@@ -216,6 +216,11 @@ public:
         return query_context.lock();
     }
 
+    auto getGlobalContext() const
+    {
+        return global_context.lock();
+    }
+
     void disableProfiling()
     {
         assert(!query_profiler_real && !query_profiler_cpu);

src/Common/ZooKeeper/ZooKeeper.cpp

@@ -701,24 +701,34 @@ void ZooKeeper::removeChildrenRecursive(const std::string & path, const String &
     }
 }
 
-void ZooKeeper::tryRemoveChildrenRecursive(const std::string & path, const String & keep_child_node)
+bool ZooKeeper::tryRemoveChildrenRecursive(const std::string & path, bool probably_flat, const String & keep_child_node)
 {
     Strings children;
     if (tryGetChildren(path, children) != Coordination::Error::ZOK)
-        return;
+        return false;
+
+    bool removed_as_expected = true;
     while (!children.empty())
     {
         Coordination::Requests ops;
         Strings batch;
+        ops.reserve(MULTI_BATCH_SIZE);
+        batch.reserve(MULTI_BATCH_SIZE);
         for (size_t i = 0; i < MULTI_BATCH_SIZE && !children.empty(); ++i)
         {
             String child_path = fs::path(path) / children.back();
-            tryRemoveChildrenRecursive(child_path);
+
+            /// Will try to avoid recursive getChildren calls if child_path probably has no children.
+            /// It may be extremely slow when path contain a lot of leaf children.
+            if (!probably_flat)
+                tryRemoveChildrenRecursive(child_path);
+
             if (likely(keep_child_node.empty() || keep_child_node != children.back()))
             {
                 batch.push_back(child_path);
                 ops.emplace_back(zkutil::makeRemoveRequest(child_path, -1));
             }
+
             children.pop_back();
         }
 
@@ -726,10 +736,39 @@ void ZooKeeper::tryRemoveChildrenRecursive(const std::string & path, const Strin
         /// this means someone is concurrently removing these children and we will have
         /// to remove them one by one.
         Coordination::Responses responses;
-        if (tryMulti(ops, responses) != Coordination::Error::ZOK)
-            for (const std::string & child : batch)
-                tryRemove(child);
+        if (tryMulti(ops, responses) == Coordination::Error::ZOK)
+            continue;
+
+        removed_as_expected = false;
+
+        std::vector<zkutil::ZooKeeper::FutureRemove> futures;
+        futures.reserve(batch.size());
+        for (const std::string & child : batch)
+            futures.push_back(asyncTryRemoveNoThrow(child, -1));
+
+        for (size_t i = 0; i < batch.size(); ++i)
+        {
+            auto res = futures[i].get();
+            if (res.error == Coordination::Error::ZOK)
+                continue;
+            if (res.error == Coordination::Error::ZNONODE)
+                continue;
+
+            if (res.error == Coordination::Error::ZNOTEMPTY)
+            {
+                if (probably_flat)
+                {
+                    /// It actually has children, let's remove them
+                    tryRemoveChildrenRecursive(batch[i]);
+                    tryRemove(batch[i]);
+                }
+                continue;
+            }
+
+            throw KeeperException(res.error, batch[i]);
+        }
     }
+    return removed_as_expected;
 }
 
 void ZooKeeper::removeRecursive(const std::string & path)

src/Common/ZooKeeper/ZooKeeper.h

@@ -225,7 +225,10 @@ public:
     /// If keep_child_node is not empty, this method will not remove path/keep_child_node (but will remove its subtree).
     /// It can be useful to keep some child node as a flag which indicates that path is currently removing.
     void removeChildrenRecursive(const std::string & path, const String & keep_child_node = {});
-    void tryRemoveChildrenRecursive(const std::string & path, const String & keep_child_node = {});
+    /// If probably_flat is true, this method will optimistically try to remove children non-recursive
+    /// and will fall back to recursive removal if it gets ZNOTEMPTY for some child.
+    /// Returns true if no kind of fallback happened.
+    bool tryRemoveChildrenRecursive(const std::string & path, bool probably_flat = false, const String & keep_child_node = {});
 
     /// Remove all children nodes (non recursive).
     void removeChildren(const std::string & path);

src/Common/ZooKeeper/ZooKeeperImpl.cpp

@@ -846,7 +846,7 @@ void ZooKeeper::receiveEvent()
 void ZooKeeper::finalize(bool error_send, bool error_receive, const String & reason)
 {
     /// If some thread (send/receive) already finalizing session don't try to do it
-    bool already_started = finalization_started.exchange(true);
+    bool already_started = finalization_started.test_and_set();
 
     LOG_TEST(log, "Finalizing session {}: finalization_started={}, queue_finished={}, reason={}",
              session_id, already_started, requests_queue.isFinished(), reason);

src/Common/ZooKeeper/ZooKeeperImpl.h

@@ -209,7 +209,7 @@ private:
     std::atomic<XID> next_xid {1};
     /// Mark session finalization start. Used to avoid simultaneous
     /// finalization from different threads. One-shot flag.
-    std::atomic<bool> finalization_started {false};
+    std::atomic_flag finalization_started;
 
     using clock = std::chrono::steady_clock;