Skip to content

Fix bug in Coordinator-based compaction duty and simulator #18812

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.

Sign up for GitHub

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

Jump to bottom
Merged
kfaraz merged 2 commits into from
Dec 5, 2025
Merged

Fix bug in Coordinator-based compaction duty and simulator #18812

kfaraz merged 2 commits into from
Dec 5, 2025

Conversation

@kfaraz
Copy link
Contributor

@kfaraz kfaraz commented Dec 5, 2025 •
edited
Loading

Changes

  • Fix bug in CompactionRunSimulator so that it honors the compaction policy being verified
  • Fix bug in CompactSegments so that the duty does not launch tasks for an interval that has already been skipped by the policy
    • This bug does not occur in Overlord-based compaction supervisors
    • It does not affect any real scenarios currently since the only existing production policy newestSegmentFirst never skips intervals anyway.

This PR has:

  • been self-reviewed.
  • added documentation for new or modified features or behaviors.
  • a release note entry in the PR description.
  • added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
  • added or updated version, license, or notice information in licenses.yaml
  • added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
  • added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
  • added integration tests.
  • been tested in a test Druid cluster.

cryptoe
cryptoe reviewed Dec 5, 2025
View reviewed changes
server/src/main/java/org/apache/druid/server/coordinator/duty/CompactSegments.java

if (compactionStatus.isComplete()) {
snapshotBuilder.addToComplete(candidatesWithStatus);
continue;
Copy link
Contributor

@cryptoe cryptoe Dec 5, 2025 •
edited
Loading

Choose a reason for hiding this comment

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

So these 2 lines was the main fix if I understood correctly ?

Copy link
Contributor Author

@kfaraz kfaraz Dec 5, 2025

Choose a reason for hiding this comment

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

Yes, this ensures that we skip launching tasks for intervals which have been skipped by the policy.

@kfaraz
Copy link
Contributor Author

kfaraz commented Dec 5, 2025

Thanks for the review, @cryptoe !

@kfaraz kfaraz merged commit fa9914b into apache:master Dec 5, 2025
83 of 84 checks passed
@kfaraz kfaraz deleted the fix_compact_bug branch December 5, 2025 06:59
@kgyrtkirk
Copy link
Member

kgyrtkirk commented Dec 5, 2025

it seems like ForcePushDownNestedQueryTest is flaky on master
the first commit I see it failing is the commit associated with this PR
the same test failed once in the CI run for this PR as well - but it passed on second run; it might be just coincidence

I wonder if this PR could be related to the stability of that test or not...

@kfaraz
Copy link
Contributor Author

kfaraz commented Dec 5, 2025

Yes, @kgyrtkirk , the ForcePushDownNestedQueryTest does seem to be flaky.
It's unrelated to this PR though as this PR affects only the compaction duty behaviour.

Will try to investigate the failure. cc: @abhishekrb19

@abhishekrb19
Copy link
Contributor

abhishekrb19 commented Dec 5, 2025

#18816 should address the ForcePushDownNestedQueryTest flakiness.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@cryptoe cryptoe cryptoe approved these changes

Assignees

No one assigned

Labels

Area - Ingestion

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

4 participants

@kfaraz @kgyrtkirk @abhishekrb19 @cryptoe