Fix: possible race condition on `EXCHANGE TABLES` query #61135

yakov-olkhovskiy · 2024-03-10T22:20:36Z

Changelog category (leave one):

Improvement

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

Fixed possible race condition when EXCHANGE TABLES executed in parallel.

…cution

robot-ch-test-poll3 · 2024-03-10T22:24:27Z

This is an automated comment for commit f0d55ea 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
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	⏳ pending
Integration tests	The integration tests report. In parenthesis the package type is given, and in square brackets are the optional part/total tests	❌ failure
Stateful tests	Runs stateful functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc	❌ 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
A Sync	There's no description for the check yet, please add it to tests/ci/ci_config.py:CHECK_DESCRIPTIONS	✅ success
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
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. 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
Mergeable Check	Checks if all other necessary checks are successful	✅ success
PR Check	There's no description for the check yet, please add it to tests/ci/ci_config.py:CHECK_DESCRIPTIONS	✅ 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
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

src/Interpreters/Context.h

alexey-milovidov · 2024-03-10T22:49:25Z

Add a test that has a chance to reproduce the problem at least occasionally.

tavplubix · 2024-03-19T19:12:02Z

src/Storages/IStorage_fwd.h

@@ -12,6 +12,7 @@ class IStorage;

 using ConstStoragePtr = std::shared_ptr<const IStorage>;
 using StoragePtr = std::shared_ptr<IStorage>;
+using StorageWeakPtr = std::weak_ptr<IStorage>;


It's not allowed to use std::weak_ptr<IStorage> because we rely on std::shared_ptr<IStorage>::unique()

Instead of saving StoragePtr, we can save StorageID with resolved UUID

it's exactly the point to not keep shared_ptr here - since on HTTP requests session can be preserved it would block database from DROP - making it weak_ptr doesn't block it and we don't need to keep it cached beyond session so it should be safe

there's a race condition

tavplubix · 2024-03-27T23:02:41Z

src/Interpreters/Context.cpp

@@ -912,6 +912,23 @@ String Context::getFilesystemCacheUser() const
    return shared->filesystem_cache_user;
 }

+DatabaseAndTable Context::getOrCacheStorage(const StorageID & id, std::function<DatabaseAndTable()> f, std::optional<Exception> * exception) const


We could cache it in Context::resolveStorageID, but a new method is fine too

I don't think we want to enforce this cache everywhere where we call to Context::resolveStorageID - at least having a query context is a condition we need to use this cache.

We can check this condition in resolveStorageID

tavplubix · 2024-03-27T23:03:41Z

src/Interpreters/Context.cpp

+        return storage;
+    }
+
+    auto storage = f();


Consider a better name

tavplubix · 2024-03-27T23:05:27Z

src/Interpreters/Context.cpp

+    {
+        DatabaseAndTable storage = DatabaseCatalog::instance().tryGetByUUID(storage_id->uuid);
+        if (exception && !storage.second)
+            exception->emplace(Exception(ErrorCodes::UNKNOWN_TABLE, "Table {} does not exist anymore - maybe it was dropped", id.getNameForLogs()));


Currently, a query will fail if it requests a table the second time, but the table was dropped after the first get. But we can improve it later

I don't think we need to improve it - if table was dropped concurrently then query is better to be interrupted than trying to reposes the table from drop queue - at this point we don't even know who is holding it from physical deletion...

We do know, see addStorageHolder

tavplubix · 2024-04-03T22:29:44Z

@yakov-olkhovskiy, restarting CI by merging master is pointless. It will not fix the failures. For two reasons:

Many tests are broken in master after enabling the new analyzer. They are not fixed yet and will fail again. But even unrelated failures are worth attention.
Some failures are actually related to the changes in this PR. Please check all CI failures on 0eae42a and 242dc60

yakov-olkhovskiy · 2024-04-03T23:25:49Z

@tavplubix Mergeable check was failing because build failures like this:
https://github.com/ClickHouse/ClickHouse/actions/runs/8530457581/job/23379706220#step:9:107
I got an impression that there is something wrong with CI and restarting would help

Some failures are actually related to the changes in this PR. Please check all CI failures on 0eae42a and 242dc60

I'm not sure I see anything related... could you point it out please?

tavplubix · 2024-04-03T23:39:17Z

I'm not sure I see anything related... could you point it out please?

@yakov-olkhovskiy please check everything, and you will see. If you don't see it - try to debug every failure until you see.

Just checking everything (and trying to debug) is always a good idea because we should not ignore failures even if they are unrelated (but in this case some of them are).

yakov-olkhovskiy · 2024-04-05T04:48:26Z

@tavplubix found it... "Tests are not finished" is a confusing status though...

yakov-olkhovskiy · 2024-04-08T18:13:27Z

...fast test is getting hung and logs don't show proper stack traces - I added "force tests" label to check it with thread sanitizer...

add storage cache to context and use it during query analysis and exe…

558d764

…cution

robot-ch-test-poll3 added the pr-improvement Pull request with some product improvements label Mar 10, 2024

alexey-milovidov reviewed Mar 10, 2024

View reviewed changes

src/Interpreters/Context.h Outdated Show resolved Hide resolved

alexey-milovidov self-assigned this Mar 10, 2024

yakov-olkhovskiy added 9 commits March 11, 2024 11:37

test added, comment fixed

f8b4ef0

style check typo

2de26fa

fix for query having no query context

2f051c7

move to DatabaseCatalog

067a8c8

fix

2825656

use weak_ptr

064eca6

use StorageID instead of database.table

c66d8e1

Merge remote-tracking branch 'origin/master' into fix-exchange-tables

792a298

Merge remote-tracking branch 'origin/master' into fix-exchange-tables

b24629d

tavplubix reviewed Mar 19, 2024

View reviewed changes

add no-ordinary-database tag

5955cc5

tavplubix self-assigned this Mar 20, 2024

yakov-olkhovskiy added 4 commits March 22, 2024 15:49

cache storage by UUID

e121665

cleanup

190823f

Merge branch 'master' into fix-exchange-tables

4eeff80

fix

c67a898

yakov-olkhovskiy requested a review from tavplubix March 24, 2024 23:51

tavplubix approved these changes Mar 27, 2024

View reviewed changes

yakov-olkhovskiy added 3 commits March 30, 2024 14:55

better argument name

0eae42a

Merge branch 'master' into fix-exchange-tables

242dc60

Merge branch 'master' into fix-exchange-tables

57627ec

tavplubix added the do not test disable testing on pull request label Apr 3, 2024

add mutex to storage_cache

856cb21

yakov-olkhovskiy removed the do not test disable testing on pull request label Apr 6, 2024

yakov-olkhovskiy added 2 commits April 6, 2024 05:32

Merge branch 'master' into fix-exchange-tables

1bd25d0

fix clang tidy

c84b702

yakov-olkhovskiy added the force tests Force test ignoring fast test output. label Apr 8, 2024

Merge branch 'master' into fix-exchange-tables

f0d55ea

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix: possible race condition on `EXCHANGE TABLES` query #61135

Fix: possible race condition on `EXCHANGE TABLES` query #61135

yakov-olkhovskiy commented Mar 10, 2024

robot-ch-test-poll3 commented Mar 10, 2024 •

edited by robot-clickhouse-ci-1

alexey-milovidov commented Mar 10, 2024

tavplubix Mar 19, 2024

tavplubix Mar 19, 2024

yakov-olkhovskiy Mar 19, 2024 •

edited

tavplubix Mar 20, 2024

tavplubix Mar 27, 2024

yakov-olkhovskiy Mar 30, 2024

tavplubix Apr 3, 2024

tavplubix Mar 27, 2024

tavplubix Mar 27, 2024

yakov-olkhovskiy Mar 30, 2024

tavplubix Apr 3, 2024

tavplubix commented Apr 3, 2024

yakov-olkhovskiy commented Apr 3, 2024

tavplubix commented Apr 3, 2024 •

edited

yakov-olkhovskiy commented Apr 5, 2024

yakov-olkhovskiy commented Apr 8, 2024

Fix: possible race condition on EXCHANGE TABLES query #61135

Are you sure you want to change the base?

Fix: possible race condition on EXCHANGE TABLES query #61135

Conversation

yakov-olkhovskiy commented Mar 10, 2024

Changelog category (leave one):

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

robot-ch-test-poll3 commented Mar 10, 2024 • edited by robot-clickhouse-ci-1

alexey-milovidov commented Mar 10, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

yakov-olkhovskiy Mar 19, 2024 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tavplubix commented Apr 3, 2024

yakov-olkhovskiy commented Apr 3, 2024

tavplubix commented Apr 3, 2024 • edited

yakov-olkhovskiy commented Apr 5, 2024

yakov-olkhovskiy commented Apr 8, 2024

Fix: possible race condition on `EXCHANGE TABLES` query #61135

Fix: possible race condition on `EXCHANGE TABLES` query #61135

robot-ch-test-poll3 commented Mar 10, 2024 •

edited by robot-clickhouse-ci-1

yakov-olkhovskiy Mar 19, 2024 •

edited

tavplubix commented Apr 3, 2024 •

edited