make writebuffers cancelable

[CheSema]

fixes #64451

Changelog category (leave one):

• Bug Fix (user-visible misbehavior in an official stable release)

Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):

Fix bug with cancelation buffer when canceling a query.

Documentation entry for user-facing changes

• Documentation is written (mandatory for new features)

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

CI Settings

NOTE: If your merge the PR with modified CI you MUST KNOW what you are doing
NOTE: Checked options will be applied if set before CI RunConfig/PrepareRunConfig step

• Allow: Integration Tests
• Allow: Stateless tests
• Allow: Stateful tests
• Allow: Unit tests
• Allow: Performance tests
• Allow: All with aarch64
• Allow: All with ASAN
• Allow: All with TSAN
• Allow: All with Analyzer
• Allow: All with Azure
• Allow: Add your option here

---

• Exclude: Fast test
• Exclude: Integration Tests
• Exclude: Stateless tests
• Exclude: Stateful tests
• Exclude: Performance tests
• Exclude: All with ASAN
• Exclude: All with TSAN
• Exclude: All with MSAN
• Exclude: All with UBSAN
• Exclude: All with Coverage
• Exclude: All with Aarch64

---

• do not test (only style check)
• disable merge-commit (no merge from master before tests)
• disable CI cache (job reuse)
• allow: batch 1 for multi-batch jobs
• allow: batch 2
• allow: batch 3
• allow: batch 4, 5 and 6

[robot-ch-test-poll2]

This is an automated comment for commit bbd9c77 with description of existing statuses. It's updated for the latest CI running

❌ Click here to open a full report in a separate page

Check name	Description	Status
Bugfix validation	Checks that either a new test (functional or integration) or there some changed tests that fail with the binary built on master branch	❌ error
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	❌ failure
Stateless tests	Runs stateless functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc	❌ failure

Successful checks

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
ClickBench	Runs [ClickBench](https://github.com/ClickHouse/ClickBench/) with instant-attach table	✅ 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
Cloud fork sync (only for ClickHouse Inc. employees)	If it fails, ask a maintainer for help	✅ success
Compatibility check	Checks that clickhouse binary runs on distributions with old libc versions. If it fails, ask a maintainer for help	✅ success
Docker keeper image	The check to build and optionally push the mentioned image to docker hub	✅ success
Docker server image	The check to build and optionally push the mentioned image to docker hub	✅ success
Docs check	Builds and tests the documentation	✅ 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. Integration 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
PR Check	Checks correctness of the PR's body	✅ 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
Stateful tests	Runs stateful functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc	✅ success
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

[antonio2368]

👍

[CheSema]

It will take a time to pass all the tests here.

For example now fast tests are failed because
CompressedWriteBuffer calls finalize at the destructor. It is not good. The buffer bellow CompressedWriteBuffer could be already canceled.
Will try to catch such cases.

[CheSema]

https://s3.amazonaws.com/clickhouse-test-reports/64478/517c1138113220e5a4318c07806759911896234d/stress_test__debug_/fatal_messages.txt

this is related to this changes #65009

[CheSema]

@antonio2368 could I ask you to check this once again?
I think I passed the CI. The last CI has a crash which turned out is not raleted to my changes.

[CheSema]

01926_order_by_desc_limit -- #65472
02922_deduplication_with_zero_copy -- #65486
test_when_s3_broken_pipe_at_upload_is_retried -- related to #65472
00988_parallel_parts_removal -- #65473

[CheSema]

I have rerun CI because Bugfix Validation has failed with unclear reason.

[CheSema]

https://s3.amazonaws.com/clickhouse-test-reports/64478/19bafb5063ece9d71f44f248b35087800c4f302c/stress_test__ubsan_.html

 [ 4096 ] {a7c87804-bf65-4264-863f-d34164941762} <Fatal> : Logical error: \'false && "WriteBuffer is not finalized in destructor."\'.

related to the issue.

It is a good catch!

2024.06.20 23:44:30.479523 [ 4096 ] {a7c87804-bf65-4264-863f-d34164941762} <Error> void DB::MergeTreeDeduplicationLog::shutdown(): Code: 439. DB::Exception: Cannot schedule a task: fault injected (threads=87, jobs=0). (CANNOT_SCHEDULE_TASK), Stack trace (when copying this message, always include the lines below):

0. std::exception::capture() @ 0x00000000183d97eb
1. ./build_docker/./base/poco/Foundation/src/Exception.cpp:28: Poco::Exception::Exception(String const&, int) @ 0x0000000036ceb445
2. ./build_docker/./src/Common/Exception.cpp:101: DB::Exception::Exception(DB::Exception::MessageMasked&&, int, bool) @ 0x0000000024361e8b
3. DB::Exception::Exception(PreformattedMessage&&, int) @ 0x0000000018259796
4. ./src/Common/LoggingFormatStringHelpers.h:45: void ThreadPoolImpl<ThreadFromGlobalPoolImpl<false, true>>::scheduleImpl<void>(std::function<void ()>, Priority, std::optional<unsigned long>, bool)::'lambda'(String const&)::operator()(String const&) const @ 0x000000002447e53a
5. ./build_docker/./src/Common/ThreadPool.cpp:0: void ThreadPoolImpl<ThreadFromGlobalPoolImpl<false, true>>::scheduleImpl<void>(std::function<void ()>, Priority, std::optional<unsigned long>, bool) @ 0x00000000244782ad
6. ./contrib/llvm-project/libcxx/include/__functional/function.h:818: ? @ 0x0000000024478185
7. ./contrib/llvm-project/libcxx/include/__functional/function.h:818: ? @ 0x000000002f109f34
8. ./contrib/llvm-project/libcxx/include/__functional/function.h:717: ? @ 0x000000002f109d17
9. ./contrib/llvm-project/libcxx/include/list:385: DB::TaskTracker::add(std::function<void ()>&&) @ 0x000000002f190e86
10. ./contrib/llvm-project/libcxx/include/__functional/function.h:818: ? @ 0x000000002f167fc3
11. ./build_docker/./src/IO/WriteBufferFromS3.cpp:180: DB::WriteBufferFromS3::preFinalize() @ 0x000000002f16737f
12. ./build_docker/./src/IO/WriteBufferFromFileDecorator.cpp:0: DB::WriteBufferFromFileDecorator::preFinalize() @ 0x000000002ffe0fd6
13. ./build_docker/./src/IO/WriteBufferFromFileDecorator.cpp:26: DB::WriteBufferFromFileDecorator::finalizeImpl() @ 0x000000002ffe0c1d
14. ./build_docker/./src/Disks/IO/WriteBufferWithFinalizeCallback.cpp:19: DB::WriteBufferWithFinalizeCallback::finalizeImpl() @ 0x000000002ffe0630
15. DB::WriteBuffer::finalize() @ 0x000000001827385a
16. ./contrib/llvm-project/libcxx/include/__memory/unique_ptr.h:302: DB::MergeTreeDeduplicationLog::shutdown() @ 0x00000000335864f0
17. ./contrib/llvm-project/libcxx/include/__memory/shared_ptr.h:436: DB::InterpreterDropQuery::executeToTableImpl(std::shared_ptr<DB::Context const> const&, DB::ASTDropQuery&, std::shared_ptr<DB::IDatabase>&, StrongTypedef<wide::integer<128ul, unsigned int>, DB::UUIDTag>&) @ 0x0000000030d902d4
18. ./build_docker/./src/Interpreters/InterpreterDropQuery.cpp:442: DB::InterpreterDropQuery::executeToDatabaseImpl(DB::ASTDropQuery const&, std::shared_ptr<DB::IDatabase>&, std::vector<StrongTypedef<wide::integer<128ul, unsigned int>, DB::UUIDTag>, std::allocator<StrongTypedef<wide::integer<128ul, unsigned int>, DB::UUIDTag>>>&) @ 0x0000000030d94
9fa
19. ./build_docker/./src/Interpreters/InterpreterDropQuery.cpp:0: DB::InterpreterDropQuery::executeToDatabase(DB::ASTDropQuery const&) @ 0x0000000030d8d42b
20. ./build_docker/./src/Interpreters/InterpreterDropQuery.cpp:96: DB::InterpreterDropQuery::executeSingleDropQuery(std::shared_ptr<DB::IAST> const&) @ 0x0000000030d8c059
21. ./build_docker/./src/Interpreters/InterpreterDropQuery.cpp:0: DB::InterpreterDropQuery::execute() @ 0x0000000030d8b40e
22. ./build_docker/./src/Interpreters/executeQuery.cpp:0: DB::executeQueryImpl(char const*, char const*, std::shared_ptr<DB::Context>, DB::QueryFlags, DB::QueryProcessingStage::Enum, DB::ReadBuffer*) @ 0x0000000031babea6
23. ./build_docker/./src/Interpreters/executeQuery.cpp:0: DB::executeQuery(DB::ReadBuffer&, DB::WriteBuffer&, bool, std::shared_ptr<DB::Context>, std::function<void (DB::QueryResultDetails const&)>, DB::QueryFlags, std::optional<DB::FormatSettings> const&, std::function<void (DB::IOutputFormat&, String const&, std::shared_ptr<DB::Context const> co
nst&, std::optional<DB::FormatSettings> const&)>) @ 0x0000000031bb20ae
24. ./contrib/llvm-project/libcxx/include/__functional/function.h:818: ? @ 0x000000003389824e
25. ./build_docker/./src/Server/HTTPHandler.cpp:1161: DB::HTTPHandler::handleRequest(DB::HTTPServerRequest&, DB::HTTPServerResponse&, StrongTypedef<unsigned long, ProfileEvents::EventTag> const&) @ 0x000000003389ef09
26. ./build_docker/./src/Server/HTTP/HTTPServerConnection.cpp:71: DB::HTTPServerConnection::run() @ 0x0000000033990ac3
27. ./build_docker/./base/poco/Net/src/TCPServerConnection.cpp:57: Poco::Net::TCPServerConnection::start() @ 0x0000000036e3cb78
28. ./contrib/llvm-project/libcxx/include/__memory/unique_ptr.h:302: Poco::Net::TCPServerDispatcher::run() @ 0x0000000036e3dcb2
29. ./build_docker/./base/poco/Foundation/src/ThreadPool.cpp:219: Poco::PooledThread::run() @ 0x0000000036d8eec1
30. ./base/poco/Foundation/include/Poco/AutoPtr.h:205: Poco::ThreadImpl::runnableEntry(void*) @ 0x0000000036d89ecf
31. ? @ 0x00007f872c6d7ac3
 (version 24.7.1.47)

It is an example when finalize throws an exception. Here it is CANNOT_SCHEDULE_TASK.

After that, WriteBufferFromFileDecorator is considered as finalized and WriteBufferFromFileDecorator::impl is not.
We unable to release that bundle of two buffers properly.

I need to untangle this as well.

[CheSema]

02956_rocksdb_bulk_sink flaks in master.