Make max_bytes_before_external_sort limit depends on total query memory consumption by azat · Pull Request #72598 · ClickHouse/ClickHouse

azat · 2024-11-28T16:43:38Z

Changelog category (leave one):

Improvement

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

Make max_bytes_before_external_sort limit depends on total query memory consumption (previously it was number of bytes in the sorting block for one sorting thread, now it has the same meaning as max_bytes_before_external_group_by - it is total limit for the whole query memory for all threads). Also one more setting added to control on disk block size - min_external_sort_block_bytes

Make max_bytes_before_external_sort more user-friendly, previously it
was number of bytes in the sorting block for one sorting thread, now it
has the same meaning as max_bytes_before_external_group_by - it is total
limit for the whole query memory for all threads.

Also one more setting added to control on disk block size -
min_external_sort_block_bytes

But, after this change the files on disk can be very small (few
kilobytes) and to avoid this, I've added separate setting to control
on-disk block size - min_external_sort_block_bytes, to avoid too many
files.

Note, that max_bytes_ratio_before_external_sort already based on the
memory consumption not the sorting block size (added in #71406)

P.S. maybe it will be better to mark it as Backward Incompatible Change to highlight this change in changelogs

robot-ch-test-poll4 · 2024-11-28T16:49:23Z

This is an automated comment for commit a7e97aa 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
Integration tests	The integration tests report. In parenthesis the package type is given, and in square brackets are the optional part/total tests	❌ 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
Builds	There's no description for the check yet, please add it to tests/ci/ci_config.py:CHECK_DESCRIPTIONS	✅ success
ClickBench	Runs [ClickBench](https://github.com/ClickHouse/ClickBench/) with instant-attach table	✅ 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
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
Stateless tests	Runs stateless 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

alexey-milovidov · 2024-11-30T21:50:03Z

This is ok, but it will be much better to use the same approach as in Aggregator - to use the total memory usage of the query.

azat · 2024-12-01T18:56:49Z

This is ok, but it will be much better to use the same approach as in Aggregator - to use the total memory usage of the query.

Done (was waiting for the #71406, which adds the helper)

azat · 2024-12-04T11:05:31Z

After this change the files on disk can be very small (few kilobytes) and to avoid this, I've added separate setting to control on-disk block size - min_external_sort_block_bytes, to avoid too many files.

alexey-milovidov · 2024-12-25T10:14:55Z

Thanks, this sounds good - let's continue.

azat · 2024-12-25T13:19:55Z

Right now CI is unstable for external sort, I'm waiting for #73097 to be merged (hope it will help)

…ry consumption Make max_bytes_before_external_sort more user-friendly, previously it was number of bytes in the sorting block for one sorting thread, now it has the same meaning as max_bytes_before_external_group_by - it is total limit for the whole query memory for all threads. Also one more setting added to control on disk block size - min_external_sort_block_bytes But, after this change the files on disk can be very small (few kilobytes) and to avoid this, I've added separate setting to control on-disk block size - min_external_sort_block_bytes, to avoid too many files. Note, that max_bytes_ratio_before_external_sort already based on the memory consumption not the sorting block size (added in ClickHouse#71406) Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>

azat · 2025-01-02T11:53:18Z

Test failures does not looks related:

Integration tests (aarch64) [6/6] — fail: 2, passed: 552

test_memory_limit/test.py::test_multiple_queries - test_memory_limit/test.py::test_multiple_queries flakes #72458

>           raise QueryRuntimeException(
                "Client failed! Return code: {}, stderr: {}".format(
                    self.process.returncode, stderr
                ),
                self.process.returncode,
                stderr,
            )
E           helpers.client.QueryRuntimeException: Client failed! Return code: 210, stderr: Code: 210. DB::NetException: Net Exception: No route to host (172.16.5.2:9000). (NETWORK_ERROR)

test_storage_s3_queue/test.py::test_list_and_delete_race - test_storage_s3_queue/test_list_and_delete_race is flaky #73113

>       assert node.contains_in_log(
            "because of the race with list & delete"
        ) or node_2.contains_in_log("because of the race with list & delete")
E       AssertionError: assert (False or False)
E        +  where False = <bound method ClickHouseInstance.contains_in_log of <helpers.cluster.ClickHouseInstance object at 0xffce7c2384f0>>('because of the race with list & delete')
E        +    where <bound method ClickHouseInstance.contains_in_log of <helpers.cluster.ClickHouseInstance object at 0xffce7c2384f0>> = <helpers.cluster.ClickHouseInstance object at 0xffce7c2384f0>.contains_in_log
E        +  and   False = <bound method ClickHouseInstance.contains_in_log of <helpers.cluster.ClickHouseInstance object at 0xffce7c238520>>('because of the race with list & delete')
E        +    where <bound method ClickHouseInstance.contains_in_log of <helpers.cluster.ClickHouseInstance object at 0xffce7c238520>> = <helpers.cluster.ClickHouseInstance object at 0xffce7c238520>.contains_in_log

Integration tests (tsan) [3/6] — fail: 1, passed: 558

test_keeper_three_nodes_two_alive/test.py::test_restart_third_node

>               raise self._exception
E               kazoo.exceptions.ConnectionLoss

Improve the waiting mechanism for keeper in integration tests #74076

Follow-up for: ClickHouse#72598 Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>

…xternal ORDER BY In ClickHouse#72598 the behavior of `max_bytes_before_external_sort` has been changed, it has been switched to rely on the total memory usage of the query instead of the size of a sorting data. But, the problem with this approach is that you can have a lot of memory used outside of sorting (FINAL, aggregation), and spilling sorting blocks to disk will not help reduce memory usage obviously, but it will make things even worse, because it will create lots of small files, that will require a lot of memory to merge them. So this patch reverts the behavior of `max_bytes_before_external_sort` back, now it will rely on the sorting block size.

azat · 2025-05-24T21:22:38Z

src/Core/SettingsChangesHistory.cpp

            {"allow_general_join_planning", false, true, "Allow more general join planning algorithm when hash join algorithm is enabled."},
            {"optimize_extract_common_expressions", false, false, "Introduce setting to optimize WHERE, PREWHERE, ON, HAVING and QUALIFY expressions by extracting common expressions out from disjunction of conjunctions."},
            {"max_bytes_ratio_before_external_sort", 0., 0., "New setting."},
+            {"min_external_sort_block_bytes", 0., 100_MiB, "New setting."},


The default value is very bad, but, it will be removed anyway - #80777

robot-ch-test-poll4 added the pr-improvement Pull request with some product improvements label Nov 28, 2024

alexey-milovidov self-assigned this Nov 30, 2024

azat force-pushed the max_bytes_before_external_sort-threads branch from d47232f to 7eb25c1 Compare December 1, 2024 18:55

azat changed the title ~~Align max_bytes_before_external_sort with parallel sorting~~ Make max_bytes_before_external_sort more user-friendly Dec 1, 2024

azat force-pushed the max_bytes_before_external_sort-threads branch from de45f52 to 9441e74 Compare December 4, 2024 11:03

azat force-pushed the max_bytes_before_external_sort-threads branch from 84d7f5a to 608c7ed Compare December 4, 2024 16:38

azat marked this pull request as draft December 11, 2024 17:00

azat changed the title ~~Make max_bytes_before_external_sort more user-friendly~~ Make max_bytes_before_external_sort limit depends on total query memory consumption Dec 30, 2024

azat force-pushed the max_bytes_before_external_sort-threads branch from 90ec40e to f347332 Compare December 30, 2024 20:52

azat marked this pull request as ready for review December 30, 2024 20:52

azat force-pushed the max_bytes_before_external_sort-threads branch from f347332 to a7e97aa Compare December 31, 2024 21:20

alexey-milovidov approved these changes Jan 2, 2025

View reviewed changes

alexey-milovidov merged commit 04cae9d into ClickHouse:master Jan 2, 2025

azat deleted the max_bytes_before_external_sort-threads branch January 2, 2025 14:50

robot-ch-test-poll2 added the pr-synced-to-cloud The PR is synced to the cloud repo label Jan 2, 2025

azat added a commit to azat/ClickHouse that referenced this pull request Jan 2, 2025

Move min_external_sort_block_bytes into 25.1 in settings changes

babc599

Follow-up for: ClickHouse#72598 Signed-off-by: Azat Khuzhin <a.khuzhin@semrush.com>

azat mentioned this pull request Jan 2, 2025

Move min_external_sort_block_bytes into 25.1 in settings changes #74101

Merged

azat mentioned this pull request May 24, 2025

Fix excessive number of files (leads to excessive memory usage) for external ORDER BY #80777

Merged

azat commented May 24, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Make max_bytes_before_external_sort limit depends on total query memory consumption#72598

Make max_bytes_before_external_sort limit depends on total query memory consumption#72598
alexey-milovidov merged 1 commit intoClickHouse:masterfrom
azat:max_bytes_before_external_sort-threads

azat commented Nov 28, 2024 •

edited

Loading

Uh oh!

robot-ch-test-poll4 commented Nov 28, 2024 •

edited by robot-ch-test-poll3

Loading

Uh oh!

alexey-milovidov commented Nov 30, 2024

Uh oh!

azat commented Dec 1, 2024

Uh oh!

azat commented Dec 4, 2024

Uh oh!

alexey-milovidov commented Dec 25, 2024

Uh oh!

azat commented Dec 25, 2024

Uh oh!

azat commented Jan 2, 2025 •

edited

Loading

Uh oh!

azat May 24, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

azat commented Nov 28, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Changelog category (leave one):

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

Uh oh!

robot-ch-test-poll4 commented Nov 28, 2024 • edited by robot-ch-test-poll3 Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

alexey-milovidov commented Nov 30, 2024

Uh oh!

azat commented Dec 1, 2024

Uh oh!

azat commented Dec 4, 2024

Uh oh!

alexey-milovidov commented Dec 25, 2024

Uh oh!

azat commented Dec 25, 2024

Uh oh!

azat commented Jan 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

azat May 24, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

azat commented Nov 28, 2024 •

edited

Loading

robot-ch-test-poll4 commented Nov 28, 2024 •

edited by robot-ch-test-poll3

Loading

azat commented Jan 2, 2025 •

edited

Loading