Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[feature]: allow to attach parts from a different disk (v2.0) #63087

Merged
merged 14 commits into from
Jun 4, 2024

Conversation

Unalian
Copy link
Contributor

@Unalian Unalian commented Apr 28, 2024

Changelog category (leave one):

  • New Feature

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

Allow to attach parts from a different disk

  • attach partition from the table on other disks using copy instead of hard link (such as instant table)
  • attach partition using copy when the hard link fails even on the same disk

Documentation entry for user-facing changes

  • Documentation is written (mandatory for new features)

  • Motivation:
    Allow to attach parts from a different disk:
    attach partition from the table on other disks using copy instead of hard link (such as instant table)
    attach partition using copy when the hard link fails even on the same disk
    mergetree and replicatedMergeTree Engineer can both support this feature

  • Example use:
    The source table can be an instant table like here: https://github.com/ClickHouse/web-tables-demo
    Create a destination table with the same structure with source table. Then we can run:

ALTER TABLE destination ATTACH PARTITION tuple() FROM source

The destination table can attach the partition from source table even though the source table is an instant table.

Test

  1. The integration test is written in test_attach_partition_using_copy:
  • test_both_mergtree: Attach data from source table(instant table) on replica1 to destination table on replica1. Both are mergetree engine tables.
  • test_all_replicated: Attach data from source table(instant table) on replica1 to destination table on replica1, verify whether destination table on replica2 and source table on replica1 are the same. The three tables are all ReplicatedMergetree tables.
  • test_only_destination_replicated: Attach data from source table(instant table) on replica1 to destination table on replica2, verify whether destination table on replica2 and source table on replica1 are the same. The destination tables are ReplicatedMergetree tables.
  • test_not_work_on_different_disk : Replace and move partition don't work when on different disks.
  1. The integration test tests/integration/test_multiple_disks/test.py::test_move_across_policies_not_work shows attach fail when on the different disks. Change this case.

Additional information

This PR used to be merged in #60112. However, there are some problems after merged:
#62283
image

image These problems have been fixed in this pr.

@alexey-milovidov alexey-milovidov added the can be tested Allows running workflows for external contributors label Apr 28, 2024
@robot-ch-test-poll2 robot-ch-test-poll2 added pr-feature Pull request with new product feature submodule changed At least one submodule changed in this PR. labels Apr 28, 2024
@robot-ch-test-poll2
Copy link
Contributor

robot-ch-test-poll2 commented Apr 28, 2024

This is an automated comment for commit 20b2254 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 nameDescriptionStatus
A SyncThere's no description for the check yet, please add it to tests/ci/ci_config.py:CHECK_DESCRIPTIONS⏳ pending
CI runningA 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
Integration testsThe integration tests report. In parenthesis the package type is given, and in square brackets are the optional part/total tests❌ failure
Mergeable CheckChecks if all other necessary checks are successful⏳ pending
Performance ComparisonMeasure changes in query performance. The performance test report is described in detail here. In square brackets are the optional part/total tests❌ failure
Stateless testsRuns stateless functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc❌ failure
Stress testRuns stateless functional tests concurrently from several clients to detect concurrency-related errors❌ failure
Upgrade checkRuns 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❌ failure
Successful checks
Check nameDescriptionStatus
AST fuzzerRuns randomly generated queries to catch program errors. The build type is optionally given in parenthesis. If it fails, ask a maintainer for help✅ success
ClickBenchRuns [ClickBench](https://github.com/ClickHouse/ClickBench/) with instant-attach table✅ success
ClickHouse build checkBuilds 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 checkChecks that clickhouse binary runs on distributions with old libc versions. If it fails, ask a maintainer for help✅ success
Docker keeper imageThe check to build and optionally push the mentioned image to docker hub✅ success
Docker server imageThe check to build and optionally push the mentioned image to docker hub✅ success
Docs checkBuilds and tests the documentation✅ success
Fast testNormally 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 testsChecks 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 packagesChecks that the built packages are installable in a clear environment✅ success
PR CheckThere's no description for the check yet, please add it to tests/ci/ci_config.py:CHECK_DESCRIPTIONS✅ success
Stateful testsRuns stateful functional tests for ClickHouse binaries built in various configurations -- release, debug, with sanitizers, etc✅ success
Style checkRuns a set of checks to keep the code style clean. If some of tests failed, see the related log from the report✅ success
Unit testsRuns the unit tests for different release types✅ success

@robot-ch-test-poll3 robot-ch-test-poll3 removed the submodule changed At least one submodule changed in this PR. label Apr 29, 2024
@Unalian
Copy link
Contributor Author

Unalian commented Apr 30, 2024

I have no idea why this stateless test case failed. It seems like it has nothing to do with my codes.
029920.sql:

SYSTEM FLUSH LOGS;
SELECT 'Column ' || name || ' from table ' || concat(database, '.', table) || ' should have a comment'
FROM system.columns
WHERE (database = 'system') AND (comment = '') AND (table NOT ILIKE '%_log_%') AND (table NOT IN ('numbers', 'numbers_mt', 'one', 'generate_series', 'generateSeries')) AND (default_kind != 'ALIAS');
企业微信截图_55d31be6-47a0-41f0-991c-613995ff37e1

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's 90% copy-paste from localBackup.cpp. Let's avoid copy-paste and make localBackup.cpp work with two disks. We can pass src and dst disks to localBackup as 2 separate arguments, so it will be able to check if it's the same disk or not. And if it's not the same, it will copy from one disk to another.

Comment on lines 7170 to 7178
/// Used only when attach partition; Both for same disk and different disk.
std::pair<MergeTreeData::MutableDataPartPtr, scope_guard> MergeTreeData::cloneAndLoadDataPart(
const MergeTreeData::DataPartPtr & src_part,
const String & tmp_part_prefix,
const MergeTreePartInfo & dst_part_info,
const StorageMetadataPtr & metadata_snapshot,
const IDataPartStorage::ClonePartParams & params,
const ReadSettings & read_settings,
const WriteSettings & write_settings)
Copy link
Member

@tavplubix tavplubix May 7, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Same, let's avoid copy-paste

@tavplubix
Copy link
Member

I have no idea why this stateless test case failed. It seems like it has nothing to do with my codes.

Yes, it was broken in master for a while

…aPart into a function, controlled by parameters; integrate remoteBackup and localBackup into a class, and change the name to Backup.
@Unalian
Copy link
Contributor Author

Unalian commented May 24, 2024

@tavplubix @alexey-milovidov Is there anything else that needs to be modified in this PR? Please let me know if you still have any questions.

Comment on lines 524 to 525
true,
{},
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please add comments with the argument names (like /* max_level= */ {})

@tavplubix tavplubix self-assigned this May 28, 2024
@tavplubix
Copy link
Member

Integration tests (aarch64) [5/6] - #64451
Stateless tests (debug, s3 storage) [1/6] - #50911
Stateless tests (msan) [6/6] - #64765
Stress test (tsan) - timeout in DistributedAsyncInsertDirectoryQueue is too big
Upgrade check - was broken by new release

@tavplubix tavplubix enabled auto-merge June 3, 2024 18:20
@tavplubix tavplubix added this pull request to the merge queue Jun 4, 2024
Merged via the queue into ClickHouse:master with commit 9f49c83 Jun 4, 2024
236 of 249 checks passed
@robot-clickhouse-ci-2 robot-clickhouse-ci-2 added the pr-synced-to-cloud The PR is synced to the cloud repo label Jun 4, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
can be tested Allows running workflows for external contributors pr-feature Pull request with new product feature pr-synced-to-cloud The PR is synced to the cloud repo
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

6 participants