require `finalize()` call before d-tor for all writes buffers by CheSema · Pull Request #50395 · ClickHouse/ClickHouse

CheSema · 2023-05-31T13:47:34Z

Changelog category (leave one):

Not for changelog (changelog entry is not required)

d-tor can be called without finalize() call only when exception happened.
Remove finalize() calls from compression buffers d-tors.

Continuation for #50274

Blocks #50907

Information about CI checks: https://clickhouse.com/docs/en/development/continuous-integration/

robot-clickhouse-ci-2 · 2023-05-31T13:51:58Z

This is an automated comment for commit 4d08df8 with description of existing statuses. It's updated for the latest CI running
The full report is available here
The overall status of the commit is 🔴 failure

Check name	Description	Status
AST fuzzer	Runs randomly generated queries to catch program errors. The build type is optionally given in parenthesis. If it fails, ask a maintainer for help	🟢 success
CI running	A meta-check that indicates the running CI. Normally, it's in success or pending state. The failed status indicates some problems with the PR	🟢 success
ClickHouse build check	Builds ClickHouse in various configurations for use in further steps. You have to fix the builds that fail. Build logs often has enough information to fix the error, but you might have to reproduce the failure locally. The cmake options can be found in the build log, grepping for cmake. Use these options and follow the general build process	🟢 success
Compatibility check	Checks that clickhouse binary runs on distributions with old libc versions. If it fails, ask a maintainer for help	🟢 success
Docker image for servers	The check to build and optionally push the mentioned image to docker hub	🟢 success
Fast test	Normally this is the first check that is ran for a PR. It builds ClickHouse and runs most of stateless functional tests, omitting some. If it fails, further checks are not started until it is fixed. Look at the report to see which tests fail, then reproduce the failure locally as described here	🟢 success
Flaky tests	Checks if new added or modified tests are flaky by running them repeatedly, in parallel, with more randomization. Functional tests are run 100 times with address sanitizer, and additional randomization of thread scheduling. Integrational tests are run up to 10 times. If at least once a new test has failed, or was too long, this check will be red. We don't allow flaky tests, read the doc	🟢 success
Install packages	Checks that the built packages are installable in a clear environment	🟢 success
Integration tests	The integration tests report. In parenthesis the package type is given, and in square brackets are the optional part/total tests	🟢 success
Mergeable Check	Checks if all other necessary checks are successful	🟢 success
Performance Comparison	Measure changes in query performance. The performance test report is described in detail here. In square brackets are the optional part/total tests	🟢 success
Push to Dockerhub	The check for building and pushing the CI related docker images to docker hub	🟢 success
SQLancer	Fuzzing tests that detect logical bugs with SQLancer tool	🟢 success
Sqllogic	Run clickhouse on the sqllogic test set against sqlite and checks that all statements are passed	🟢 success
Stateful tests	Runs stateful functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc	🟢 success
Stateless tests	Runs stateless functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc	🔴 failure
Stress test	Runs stateless functional tests concurrently from several clients to detect concurrency-related errors	🟢 success
Style Check	Runs a set of checks to keep the code style clean. If some of tests failed, see the related log from the report	🟢 success
Unit tests	Runs the unit tests for different release types	🟢 success
Upgrade check	Runs stress tests on server version from last release and then tries to upgrade it to the version from the PR. It checks if the new server can successfully startup without any errors, crashes or sanitizer asserts	🟢 success

CheSema · 2023-05-31T16:51:35Z

strange, the build from fast test fails with

2023.05.31 16:49:18.020907 [ 3401093 ] {} <Error> Application: Code: 137. DB::Exception: DiskFactory: the disk 's3_disk' has unknown disk type: s3: Cannot attach table `system`.`transactions_info_log` from metadata file /var/lib/clickhouse/store/919/919947d0-287e-4b77-9e05-124e38bafc57/transactions_info_log.sql from query ATTACH TABLE system.transactions_info_log UUID '39ef021f-4561-4646-a910-d7befd47097a' (`type` Enum8('Begin' = 1, 'Commit' = 2, 'Rollback' = 3, 'AddPart' = 10, 'LockPart' = 11, 'UnlockPart' = 12), `event_date` Date, `event_time` DateTime64(6), `thread_id` UInt64, `query_id` String, `tid` Tuple(UInt64, UInt64, UUID), `tid_hash` UInt64, `csn` UInt64, `database` String, `table` String, `uuid` UUID, `part` String) ENGINE = MergeTree PARTITION BY toYYYYMM(event_date) ORDER BY (event_date, event_time) SETTINGS index_granularity = 8192. (UNKNOWN_ELEMENT_IN_CONFIG), Stack trace (when copying this message, always include the lines below):

stack

``` 0. /fasttest-workspace/build/./contrib/llvm-project/libcxx/include/exception:134: Poco::Exception::Exception(String const&, int) @ 0x000000001342b5f2 in /home/ubuntu/work/local-ch/clickhouse.1 1. /fasttest-workspace/build/./src/Common/Exception.cpp:92: DB::Exception::Exception(DB::Exception::MessageMasked&&, int, bool) @ 0x000000000d020ad5 in /home/ubuntu/work/local-ch/clickhouse.1 2. DB::Exception::Exception(int, FormatStringHelperImpl::type, std::type_identity::type>, String const&, String const&) @ 0x000000000878fffd in /home/ubuntu/work/local-ch/clickhouse.1 3. /fasttest-workspace/build/./src/Disks/DiskFactory.cpp:34: DB::DiskFactory::create(String const&, Poco::Util::AbstractConfiguration const&, String const&, std::shared_ptr, std::map, std::less, std::allocator>>> const&) const @ 0x000000001135e4d3 in /home/ubuntu/work/local-ch/clickhouse.1 4. /fasttest-workspace/build/./src/Disks/DiskSelector.cpp:49: DB::DiskSelector::initialize(Poco::Util::AbstractConfiguration const&, String const&, std::shared_ptr) @ 0x00000000114db2b9 in /home/ubuntu/work/local-ch/clickhouse.1 5. /fasttest-workspace/build/./contrib/llvm-project/libcxx/include/string:1499: DB::Context::getDiskSelector(std::lock_guard&) const @ 0x00000000114a755f in /home/ubuntu/work/local-ch/clickhouse.1 6. /fasttest-workspace/build/./src/Interpreters/Context.cpp:3409: DB::Context::getStoragePolicySelector(std::lock_guard&) const @ 0x0000000011492bd4 in /home/ubuntu/work/local-ch/clickhouse.1 7. /fasttest-workspace/build/./src/Interpreters/Context.cpp:3341: DB::Context::getStoragePolicy(String const&) const @ 0x00000000114a78d1 in /home/ubuntu/work/local-ch/clickhouse.1 8. /fasttest-workspace/build/./src/Storages/MergeTree/MergeTreeData.cpp:0: DB::MergeTreeData::getStoragePolicy() const @ 0x00000000126dca71 in /home/ubuntu/work/local-ch/clickhouse.1 9. /fasttest-workspace/build/./src/Storages/MergeTree/MergeTreeData.h:915: DB::MergeTreeData::initializeDirectoriesAndFormatVersion(String const&, bool, String const&, bool) @ 0x00000000126d86a1 in /home/ubuntu/work/local-ch/clickhouse.1 10. /fasttest-workspace/build/./src/Storages/StorageMergeTree.cpp:115: DB::StorageMergeTree::StorageMergeTree(DB::StorageID const&, String const&, DB::StorageInMemoryMetadata const&, bool, std::shared_ptr, String const&, DB::MergeTreeData::MergingParams const&, std::unique_ptr>, bool) @ 0x0000000012972af2 in /home/ubuntu/work/local-ch/clickhouse.1 11. /fasttest-workspace/build/./contrib/llvm-project/libcxx/include/__memory/shared_ptr.h:1460: std::shared_ptr std::allocate_shared[abi:v15000], DB::StorageID const&, String const&, DB::StorageInMemoryMetadata&, bool const&, std::shared_ptr&, String&, DB::MergeTreeData::MergingParams&, std::unique_ptr>, bool const&, void>(std::allocator const&, DB::StorageID const&, String const&, DB::StorageInMemoryMetadata&, bool const&, std::shared_ptr&, String&, DB::MergeTreeData::MergingParams&, std::unique_ptr>&&, bool const&) @ 0x0000000012972459 in /home/ubuntu/work/local-ch/clickhouse.1 12. /fasttest-workspace/build/./src/Storages/MergeTree/registerStorageMergeTree.cpp:0: DB::create(DB::StorageFactory::Arguments const&) @ 0x000000001296f83a in /home/ubuntu/work/local-ch/clickhouse.1 13. /fasttest-workspace/build/./src/Storages/StorageFactory.cpp:225: DB::StorageFactory::get(DB::ASTCreateQuery const&, String const&, std::shared_ptr, std::shared_ptr, DB::ColumnsDescription const&, DB::ConstraintsDescription const&, bool) const @ 0x000000001235dd1b in /home/ubuntu/work/local-ch/clickhouse.1 14. /fasttest-workspace/build/./contrib/llvm-project/libcxx/include/string:2067: DB::createTableFromAST(DB::ASTCreateQuery, String const&, String const&, std::shared_ptr, bool) @ 0x00000000112e450b in /home/ubuntu/work/local-ch/clickhouse.1 15. /fasttest-workspace/build/./src/Databases/DatabaseOrdinary.cpp:0: DB::DatabaseOrdinary::loadTableFromMetadata(std::shared_ptr, String const&, DB::QualifiedTableName const&, std::shared_ptr const&, DB::LoadingStrictnessLevel) @ 0x00000000112ff4c6 in /home/ubuntu/work/local-ch/clickhouse.1 16. /fasttest-workspace/build/./src/Databases/TablesLoader.cpp:197: void std::__function::__policy_invoker::__call_impl>&, std::shared_ptr, std::vector> const&, unsigned long)::$_0, void ()>>(std::__function::__policy_storage const*) @ 0x000000001158a4d0 in /home/ubuntu/work/local-ch/clickhouse.1 17. /fasttest-workspace/build/./base/base/../base/wide_integer_impl.h:796: ThreadPoolImpl>::worker(std::__list_iterator, void*>) @ 0x000000000d0da200 in /home/ubuntu/work/local-ch/clickhouse.1 18. /fasttest-workspace/build/./src/Common/ThreadPool.cpp:0: ThreadFromGlobalPoolImpl::ThreadFromGlobalPoolImpl>::scheduleImpl(std::function, Priority, std::optional, bool)::'lambda0'()>(void&&)::'lambda'()::operator()() @ 0x000000000d0dcf59 in /home/ubuntu/work/local-ch/clickhouse.1 19. /fasttest-workspace/build/./base/base/../base/wide_integer_impl.h:796: ThreadPoolImpl::worker(std::__list_iterator) @ 0x000000000d0d7c3e in /home/ubuntu/work/local-ch/clickhouse.1 20. /fasttest-workspace/build/./contrib/llvm-project/libcxx/include/__memory/unique_ptr.h:302: void* std::__thread_proxy[abi:v15000]>, void ThreadPoolImpl::scheduleImpl(std::function, Priority, std::optional, bool)::'lambda0'()>>(void*) @ 0x000000000d0db68e in /home/ubuntu/work/local-ch/clickhouse.1 21. ? @ 0x00007fe612353b43 in ? 22. ? @ 0x00007fe6123e5a00 in ? (version 23.5.1.1) ```

…MergeMutateTaskBase

src/IO/CascadeWriteBuffer.h

Avogar · 2023-06-19T12:26:06Z

src/IO/NullWriteBuffer.h


 /// Simply do nothing, can be used to measure amount of written bytes.
-class NullWriteBuffer : public BufferWithOwnMemory<WriteBuffer>, boost::noncopyable
+class NullWriteBuffer : public WriteBuffer, boost::noncopyable


I remember that we have a butch of WriteBuffers that don't require finalize at all. Like NullWriteBuffer, WriteBufferFromArena, MemoryWriteBuffer, WriteBufferFromPocoSocket, etc. and sometimes we use just base class WriteBuffer to read from pointer. I see you added WriteBufferFromPointer for this purpose, but AFAIU we still have to call no-op finalize for them (otherwise there will be assert int ~WriteBuffer). In my old PR I tried to solve it by introcuding new base class WriteBufferWithoutFinalize that sets is_finalize_required flag in WriteBuffer to false and used it as base class instead of just WriteBuffer when needed. What do you think about it?

I think it is good idea. It help to reduce the number of no-op finalize calls.
But I do not address this in that PR. Your proposal I could do in the following changes.

Avogar · 2023-06-19T12:28:03Z

src/IO/WriteBuffer.cpp

+        /// It is totally OK to destroy instance without finalization when an exception occurs
+        /// However it is suspicious to destroy instance without finalization at the green path


Wow, that's a great idea about exceptions, I didn't think about it and tried to add try/catch in many places :)

src/Processors/Formats/Impl/ParallelFormattingOutputFormat.cpp

src/Storages/MergeTree/MergeTreeBackgroundExecutor.cpp

src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp

robot-clickhouse-ci-2 added the pr-not-for-changelog This PR should not be mentioned in the changelog label May 31, 2023

CheSema mentioned this pull request May 31, 2023

Make buffers d-tor finalize free #50274

Open

CheSema force-pushed the better-log branch 7 times, most recently from 19f4c20 to d5b4ce2 Compare June 5, 2023 08:48

CheSema force-pushed the better-log branch 6 times, most recently from 79446d7 to 3bc4cad Compare June 14, 2023 19:33

CheSema added 5 commits June 16, 2023 16:38

add error log message when no finalize called before d-tor

9a4962b

do call finalize for all buffers

1cb02e2

fix 02428_combinators_with_over_statement

d21fc22

finalize hashing buffer

b428f88

no finalize at d-tor in compression

cf31fdb

CheSema force-pushed the better-log branch from 978d17c to d1a4135 Compare June 16, 2023 14:38

CheSema marked this pull request as ready for review June 16, 2023 14:38

CheSema changed the title ~~add error log message when no finalize called before d-tor~~ require finalize call before d-tor for all writes buffers Jun 16, 2023

CheSema changed the title ~~require finalize call before d-tor for all writes buffers~~ require finalize() call before d-tor for all writes buffers Jun 16, 2023

CheSema mentioned this pull request Jun 16, 2023

do not call finalize after exception #50907

Merged

do abort/cancel/deletion merge/mutate task with exception context

47484fa

CheSema force-pushed the better-log branch from d1a4135 to 47484fa Compare June 16, 2023 14:48

Avogar self-assigned this Jun 16, 2023

CheSema force-pushed the better-log branch from 11340c2 to ce7cd67 Compare June 18, 2023 15:35

fix gtest_lru_file_cache, fix remove from pending

2bbfa27

CheSema force-pushed the better-log branch from 8ddf507 to 2bbfa27 Compare June 18, 2023 19:05

CheSema added 2 commits June 19, 2023 00:10

fix gtest with MemoryWriteBuffer, do not mute exception in Replicated…

e3f892f

…MergeMutateTaskBase

cancel temp part with exception context when deduplication triggered

a2d5c91

CheSema force-pushed the better-log branch from 6c37e3a to a2d5c91 Compare June 19, 2023 10:52

Avogar reviewed Jun 19, 2023

View reviewed changes

src/IO/CascadeWriteBuffer.h Outdated Show resolved Hide resolved

Avogar reviewed Jun 19, 2023

View reviewed changes

src/Processors/Formats/Impl/ParallelFormattingOutputFormat.cpp Show resolved Hide resolved

Avogar reviewed Jun 19, 2023

View reviewed changes

src/Storages/MergeTree/MergeTreeBackgroundExecutor.cpp Outdated Show resolved Hide resolved

Avogar reviewed Jun 19, 2023

View reviewed changes

src/Storages/MergeTree/MergeTreeDataPartWriterOnDisk.cpp Show resolved Hide resolved

work with comment on the PR

fd292dc

CheSema force-pushed the better-log branch from 5cd101e to fd292dc Compare June 20, 2023 18:02

CheSema and others added 2 commits June 20, 2023 20:37

Merge branch 'master' into better-log

9b0c335

fix finalization of cascade write buffer

4d08df8

Avogar approved these changes Jun 21, 2023

View reviewed changes

CheSema merged commit 01de36f into ClickHouse:master Jun 21, 2023

CheSema added pr-must-backport-cloud and removed pr-must-backport-cloud labels Jun 26, 2023

SmitaRKulkarni mentioned this pull request Jun 27, 2023

Fix race azure blob storage iterator #50936

Merged

CheSema mentioned this pull request Jun 27, 2023

WriteBufferFromS3 is not finalized in destructor #43303

Closed

nikitamikhaylov added pr-must-backport-cloud and removed pr-must-backport-cloud labels Jun 27, 2023

Algunenano mentioned this pull request Jun 28, 2023

[Master] ATTEMPT_TO_READ_AFTER_EOF / ABORTED errors on table drop #51543

Closed

azat mentioned this pull request Oct 5, 2023

Avoid possible memory leaks in compressors in case of missing buffer finalization #55262

Merged

azat mentioned this pull request Oct 23, 2023

Do not write retriable errors for Replicated mutate/merge into error log #55944

Merged

antonio2368 mentioned this pull request Jan 16, 2024

Fix HTTP compressors finalization #58846

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

require `finalize()` call before d-tor for all writes buffers#50395

require `finalize()` call before d-tor for all writes buffers#50395
CheSema merged 15 commits intoClickHouse:masterfrom
CheSema:better-log

CheSema commented May 31, 2023 •

edited

Loading

Uh oh!

robot-clickhouse-ci-2 commented May 31, 2023 •

edited by robot-ch-test-poll3

Loading

Uh oh!

CheSema commented May 31, 2023

Uh oh!

Uh oh!

Avogar Jun 19, 2023

Uh oh!

CheSema Jun 19, 2023

Uh oh!

Avogar Jun 19, 2023

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		/// It is totally OK to destroy instance without finalization when an exception occurs
		/// However it is suspicious to destroy instance without finalization at the green path

Conversation

CheSema commented May 31, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Changelog category (leave one):

Uh oh!

robot-clickhouse-ci-2 commented May 31, 2023 • edited by robot-ch-test-poll3 Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

CheSema commented May 31, 2023

Uh oh!

Uh oh!

Avogar Jun 19, 2023

Choose a reason for hiding this comment

Uh oh!

CheSema Jun 19, 2023

Choose a reason for hiding this comment

Uh oh!

Avogar Jun 19, 2023

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

CheSema commented May 31, 2023 •

edited

Loading

robot-clickhouse-ci-2 commented May 31, 2023 •

edited by robot-ch-test-poll3

Loading