Skip to content

[server] Fix race condition in table/partition deletion ordering#3610

Draft
JandyTenedora wants to merge 2 commits into
apache:mainfrom
JandyTenedora:fix/1026-coordinator-event-processor-flaky-test
Draft

[server] Fix race condition in table/partition deletion ordering#3610
JandyTenedora wants to merge 2 commits into
apache:mainfrom
JandyTenedora:fix/1026-coordinator-event-processor-flaky-test

Conversation

@JandyTenedora

@JandyTenedora JandyTenedora commented Jul 7, 2026

Copy link
Copy Markdown

Purpose

Linked issue: close #1026

Submit ZK metadata deletion before remote directory cleanup in completeDeleteTable() and completeDeletePartition(). Slow remote filesystem operation likely blocking the fast ZK metadata delete, causing flaky test timeouts in CoordinatorEventProcessorTest.

Brief change log

  • Swapped the call order in completeDeleteTable() so asyncDeleteTableMetadata() is submitted to the ioExecutor before asyncDeleteRemoteDirectory()
  • Applied the same reordering in completeDeletePartition() for consistency

Tests

  • Existing CoordinatorEventProcessorTest.testCreateAndDropTable
  • New TableManagerDeletionOrderingTest.java

API and Format

No.

Documentation

No.

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Test CoordinatorEventProcessorTest.testCreateAndDropTable is not stable

1 participant