thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-12-17 20:02:05 +00:00
Code Issues Packages Projects Releases Wiki Activity
101,406 Commits 9,330 Branches 1,783 Tags 2.1 GiB
96543074fe
Commit Graph

1540 Commits

Author SHA1 Message Date
xiedeyantu
c258d3ac8b fix s3 support question mark wildcard 2022-11-18 12:11:22 +08:00
Alexander Tokmakov
9011a18234
Revert "S3 request per second rate throttling" 2022-11-16 22:33:48 +03:00
Sergei Trifonov
159743edd6
Merge pull request #43014 from ClickHouse/disk-s3-throttler 
S3 request per second rate throttling
 2022-11-16 18:51:06 +01:00
Azat Khuzhin
ac27bc0193 Add trailing slash (/) in S3ObjectStorage::getDirectoryContents() 
Otherwise it returns only the directory itself.

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-11-10 18:08:41 +01:00
Azat Khuzhin
d3e0f16873 Remove common root path in MetadataStorageFromPlainObjectStorage::listDirectory() 
This path should not leak into users/outside, since later it can be
passed to other APIs, i.e. exists() and so on.

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-11-10 18:08:41 +01:00
Azat Khuzhin
9296bfb161 Handle list_object_keys_size for Azure 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-11-10 18:08:41 +01:00
Azat Khuzhin
2cbc61df18 Fix ATTACH FROM s3_plain for wide part 
Previously because of incorrect
MetadataStorageFromPlainObjectStorage::exists(), that used
S3ObjectStorage::exists() before, which works only for existing keys,
not some intermediate path.

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-11-10 18:08:41 +01:00
Azat Khuzhin
41884d3b88 Optimize MetadataStorageFromPlainObjectStorage::getFileSize() 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-11-10 18:08:41 +01:00
Azat Khuzhin
7b7ae175df Add max_keys for IObjectStorage::findAllFiles() 
v2: Fix google-default-arguments for IObjectStorage::findAllFiles()
v3: Update max_keys for S3 requests in S3ObjectStorage::findAllFiles() loop
v4: Clarify things about max_keys vs list_object_keys_size in S3ObjectStorage::findAllFiles()
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-11-10 18:08:26 +01:00
Azat Khuzhin
434b9c14d8 Handle all entries for azure blob storage (not only first 5k) 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-11-10 17:42:32 +01:00
Sergei Trifonov
8eedd1e046
Merge branch 'master' into disk-s3-throttler 2022-11-08 15:00:56 +01:00
serxa
6d5d9ff421 rename ReadWriteSettings -> RequestSettings 2022-11-08 13:48:23 +00:00
serxa
2daec0b45e S3 request per second rate throttling + refactoring 2022-11-07 18:05:40 +00:00
Aleksandr Musorin
5cb69d8a22 changed type name for S3_Plain storage 
renamed a disk for S3PlainObjectStorage in system.disks table from s3 to s3_plain
 2022-11-04 17:35:51 +01:00
Azat Khuzhin
c93262170d Remove ReadOnlyMetadataStorage 
Throw exceptions from IMetadataStorage instead to avoid introducing
extra abstractions.

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:30 +01:00
Azat Khuzhin
51bd0c2ac1 Make ReadOnlyMetadataStorage really readonly 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
bceca73f6f Throw NOT_IMPLEMENTED form ReadOnlyMetadataStorage::getLastChanged() 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
3b7abbbff4 Rename s/listPrefix/findAllFiles, s/listPrefixInPath/getDirectoryContents/ 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
fa7535c90d Slightly optimize MetadataStorageFromPlainObjectStorage::isDirectory() 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
82ea67eb51 Slightly better MetadataStorageFromPlainObjectStorage::isFile() 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
95fb2ad3cf Implement ATTACH of MergeTree table for s3_plain disk 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
4e42521f44 Introduce ReadOnlyMetadataStorage 
And use it for:
- MetadataStorageFromPlainObjectStorage
- MetadataStorageFromStaticFilesWebServer

This will allow to reduce ~100-200 lines of duplicated code, and plus
make the code less error prone.

Note, for now I tried to make this without behaviour changes.

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
f8ba24f040 Implement S3ObjectStorage::listPrefixInPath() 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:26 +01:00
Azat Khuzhin
18e4fdf40f Introduce IObjectStorage::listPrefixInPath() 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:25 +01:00
Azat Khuzhin
ee18140c48 Remove listPrefix() implementation for disks that does not support send_metadata 
The reason for removing is it because not compatible with restoring
(with send_metadata set) anyway:
- HDFS - is not compatible with send_metadata, and besides it's
  implementaion is not correct, since it is simply `ls -l`, while the
  following is required: `find . -maxdepth 1 -type f`
- Web - is not compatible with send_metadata anyway
- Local - is not compatible with send_metadata anyway

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:25 +01:00
Azat Khuzhin
88db8ae7fa Add a comment for IObjectStorage::listPrefix() 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:25 +01:00
Azat Khuzhin
94d9600fb8 Implement MetadataStorageFromPlainObjectStorage::getLastChanged() (as for web) 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:25 +01:00
Azat Khuzhin
b34ffda272 Implement MetadataStorageFromPlainObjectStorage::getLastModified() (used by MergeTree) 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:25 +01:00
Azat Khuzhin
f6d0c03ee5 Fix path to files in MetadataStorageFromPlainObjectStorage 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-31 12:02:25 +01:00
Azat Khuzhin
7a5432feaa Fix possible SIGSEGV for web disks when file does not exists 
It can be triggered in multiple ways, either when file does not exists
and you are trying to create MergeTree table from web (that has special
code for UUID handling) or by simply OPTIMIZE TABLE FINAL for MergeTree
table that is located on web disk, in both cases you will get the
following:

<details>

<summary>stacktrace</summary>

    2022.10.28 14:08:40.631226 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} <Trace> ContextAccess (default): Access granted: OPTIMIZE ON default.data_from_web
    2022.10.28 14:08:40.632017 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} <Debug> default.data_from_web (a3e65e1f-5fd4-47ed-9dbd-307f2586b52d) (MergerMutator): Selected 1 parts from all_1_1_0 to all_1_1_0
    2022.10.28 14:08:40.632496 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} <Trace> default.data_from_web (a3e65e1f-5fd4-47ed-9dbd-307f2586b52d): Trying to reserve 1.00 MiB using storage policy from min volume index 0
    2022.10.28 14:08:40.632752 [ 6043 ] {6165bf5f-e76b-4bca-941c-7c7ff5e3b46b} <Trace> DiskObjectStorage(DiskWebServer): Reserved 1.00 MiB on remote disk `web_disk`, having unreserved 16.00 EiB.
    2022.10.28 14:08:40.634155 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} <Debug> MergeTask::PrepareStage: Merging 1 parts: from all_1_1_0 to all_1_1_0 into Compact
    2022.10.28 14:08:40.634498 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} <Trace> WebObjectStorage: Loading metadata for directory: http://127.0.0.1:8080/store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1
    2022.10.28 14:08:40.635025 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} <Trace> DiskWeb: Adding directory: store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/
    2022.10.28 14:08:40.635355 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} <Trace> ReadWriteBufferFromHTTP: Sending request to http://127.0.0.1:8080/store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/.index
    2022.10.28 14:08:40.639618 [ 6043 ] {a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1} <Trace> DiskWeb: Cannot load disk metadata. Error: Code: 86. DB::Exception: Received error from remote server /store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/.index. HTTP status code: 404 Not Found, body: <!doctype html><html><head><title>404 Not Found</title><style>
    body { background-color: #fcfcfc; color: #333333; margin: 0; padding:0; }
    h1 { font-size: 1.5em; font-weight: normal; background-color: #9999cc; min-height:2em; line-height:2em; border-bottom: 1px inset black; margin: 0; }
    h1, p { padding-left: 10px; }
    code.url { background-color: #eeeeee; font-family:monospace; padding:0 2px;}
    </style>
    </head><body><h1>Not Found</h1><p>The requested resource <code class="url">/store/a3e/a3e65e1f-5fd4-47ed-9dbd-307f2586b52d/tmp_merge_all_1_1_1/.index</code> was not found on this server.</p></body></html>: while loading disk metadata. (RECEIVED_ERROR_FROM_REMOTE_IO_SERVER) (version 22.11.1.1)
    2022.10.28 14:08:40.640527 [ 5488 ] {} <Trace> BaseDaemon: Received signal 11
    2022.10.28 14:08:40.641529 [ 9027 ] {} <Fatal> BaseDaemon: ########################################
    2022.10.28 14:08:40.642759 [ 9027 ] {} <Fatal> BaseDaemon: (version 22.11.1.1, build id: 12145DA78CE5E9EBB10A034177FAE5967EF81A4A) (from thread 6043) (query_id: a3e65e1f-5fd4-47ed-9dbd-307f2586b52d::all_1_1_1) (query: optimize table data_from_web final) Received signal Segmentation fault (11)
    2022.10.28 14:08:40.643260 [ 9027 ] {} <Fatal> BaseDaemon: Address: NULL pointer. Access: read. Unknown si_code.
    2022.10.28 14:08:40.643769 [ 9027 ] {} <Fatal> BaseDaemon: Stack trace: 0x7ffff416c0f2 0x7ffff7cd1ca8 0x7ffff679ae5e 0x7fffd52e7906 0x7fffd50c65aa 0x7fffca7a0d42 0x7fffcaee79ec 0x7fffcaf242f8 0x7fffcaf242b5 0x7fffcaf2427d 0x7fffcaf24255 0x7fffcaf2421d 0x7ffff65c3686 0x7ffff65c2295 0x7fffcaeee2a9 0x7fffcaef2c43 0x7fffcaee3c0e 0x7fffcc4a7851 0x7fffcc4a768f 0x7fffcc4abb2d 0x7fffcfdce828 0x7fffd03e3eaa 0x7fffd03dfe3b 0x7fffc8ec42d4 0x7fffc8ed51d2 0x7ffff4bdd839 0x7ffff4bde0a8 0x7ffff48ab261 0x7ffff48a769a 0x7ffff48a6335 0x7ffff409f8fd 0x7ffff4121a60
    2022.10.28 14:08:40.644411 [ 9027 ] {} <Fatal> BaseDaemon: 4. ? @ 0x7ffff416c0f2 in ?
    2022.10.28 14:08:40.676390 [ 9027 ] {} <Fatal> BaseDaemon: 5. /src/ch/clickhouse/src/Common/StringUtils/StringUtils.cpp:9: detail::startsWith(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>> const&, char const*, unsigned long) @ 0x1ca8 in /src/ch/clickhouse/.cmake/src/Common/StringUtils/libstring_utilsd.so
    2022.10.28 14:08:40.730727 [ 9027 ] {} <Fatal> BaseDaemon: 6. /src/ch/clickhouse/src/Common/StringUtils/StringUtils.h:19: startsWith(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>> const&, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>> const&) @ 0x59ae5e in /src/ch/clickhouse/.cmake/src/libclickhouse_common_iod.so
    2022.10.28 14:08:40.923955 [ 9027 ] {} <Fatal> BaseDaemon: 7. /src/ch/clickhouse/src/Disks/ObjectStorages/Web/MetadataStorageFromStaticFilesWebServer.cpp:58: DB::MetadataStorageFromStaticFilesWebServer::exists(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>> const&) const @ 0x6e7906 in /src/ch/clickhouse/.cmake/src/libdbmsd.so
    2022.10.28 14:08:41.291996 [ 9027 ] {} <Fatal> BaseDaemon: 8. /src/ch/clickhouse/src/Disks/ObjectStorages/DiskObjectStorage.cpp:181: DB::DiskObjectStorage::exists(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char>> const&) const @ 0x4c65aa in /src/ch/clickhouse/.cmake/src/libdbmsd.so
    2022.10.28 14:08:41.704697 [ 9027 ] {} <Fatal> BaseDaemon: 9. /src/ch/clickhouse/src/Storages/MergeTree/DataPartStorageOnDisk.cpp:74: DB::DataPartStorageOnDisk::exists() const @ 0xda0d42 in /src/ch/clickhouse/.cmake/src/libclickhouse_storages_mergetreed.so
    2022.10.28 14:08:43.032459 [ 9027 ] {} <Fatal> BaseDaemon: 10. /src/ch/clickhouse/src/Storages/MergeTree/MergeTask.cpp:147: DB::MergeTask::ExecuteAndFinalizeHorizontalPart::prepare() @ 0x14e79ec in /src/ch/clickhouse/.cmake/src/libclickhouse_storages_mergetreed.so
    ...
    Segmentation fault (core dumped)

</details>

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-28 15:15:49 +02:00
Anton Popov
79889198b1 fix style check 2022-10-25 23:03:37 +00:00
Anton Popov
c4d4f2dbed better interface 2022-10-25 22:14:06 +00:00
Anton Popov
c8199bc125
Merge branch 'master' into refactor-data-part 2022-10-25 14:31:05 +02:00
Nikolai Kochetov
5c4444237e
Merge pull request #42617 from ClickHouse/revert-revert-41268-disable-s3-parallel-write-for-part-moves-to-disk-s3 
Revert revert 41268 disable s3 parallel write for part moves to disk s3
 2022-10-25 14:29:27 +02:00
Nikolai Kochetov
5dabbf89ad Fixing build. 2022-10-24 13:43:24 +02:00
Anton Popov
56e5daba0c remove DataPartStorageBuilder 2022-10-23 00:23:15 +00:00
Azat Khuzhin
4e76629aaf Fixes for -Wshorten-64-to-32 
- lots of static_cast
- add safe_cast
- types adjustments
  - config
  - IStorage::read/watch
  - ...
- some TODO's (to convert types in future)

P.S. That was quite a journey...

v2: fixes after rebase
v3: fix conflicts after #42308 merged
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-21 13:25:19 +02:00
Nikolai Kochetov
57ffcb7ed0 Revert "Merge pull request #41681 from ClickHouse/revert-41268-disable-s3-parallel-write-for-part-moves-to-disk-s3" 
This reverts commit 7956c2becf, reversing
changes made to 57be648984.
 2022-10-20 19:51:27 +02:00
Vitaly Baranov
1365105bc4 Implement backup to S3 2022-10-19 00:04:41 +02:00
Vitaly Baranov
eb637a6f81
Merge pull request #42232 from azat/backups/s3 
Support BACKUP to S3 with as-is path/data structure
 2022-10-18 02:37:07 +02:00
filimonov
cd9afdcb7c
Increase request_timeout_ms for s3 disks. 2022-10-14 17:19:14 +02:00
Azat Khuzhin
8830f0608d Support BACKUP to S3 with as-is path/data structure 
Right now backup to S3 does not make a lot of sense, since:
- it has random names, and to decoding them
- requires metadata from local disk (/var/lib/disks/DISK/BACKUP_NAME)
- or send_metadata (but it is also tricky even with it)

So this patch adds simpler interface for S3, it is only suitable for
BACKUP/RESTORE, so don't try to use it for MergeTree engine.

It is done by adding separate disk - `s3_plain` for this, that:
- does not support any extended features, like renames/hardlinks/attrs/...
  (so basically everything that MergeTree requires)
- only write/read/unlink/list files

Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-14 12:59:25 +02:00
vdimir
0178307c27 Followup for TemporaryDataOnDisk 2022-10-12 15:25:23 +02:00
Azat Khuzhin
688edafcc6 Disks/ObjectStorages: add comments for some classes 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-11 17:21:26 +02:00
Azat Khuzhin
f3833e3a53 Introduce StaticDirectoryIterator 
Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>
 2022-10-11 17:19:35 +02:00
Alexander Tokmakov
bd10a9d2d4
Merge pull request #42168 from ClickHouse/unreachable_macro 
Abort instead of `__builtin_unreachable` in debug builds
 2022-10-08 19:05:57 +03:00
Alexander Tokmakov
4175f8cde6 abort instead of __builtin_unreachable in debug builds 2022-10-07 21:49:08 +02:00
kssenii
53d54ae6c9 Fix 2022-10-07 12:26:54 +02:00
Kseniia Sumarokova
eb6e60753d
Merge pull request #42129 from kssenii/fix-non-async-buffer 
Fix buffer resize with `remote_filesystem_read_method=read` + fs cache
 2022-10-07 12:05:53 +02:00
kssenii
2d96c81f4c Fix 2022-10-06 17:09:20 +02:00