squah-confluent opened a new pull request, #23923: URL: https://github.com/apache/kafka/pull/23923
When a write fails, the runtime reverts the coordinator state, but not the tasks scheduled by the reverted write operations. When those tasks complete, their results may no longer apply to the coordinator state. Have the executor take part in the coordinator state's snapshots. Reverting to a snapshot now also cancels the tasks scheduled since that snapshot. Tasks cancelled or completed since the snapshot are not restored. It is expected that callers re-schedule them if they are still needed. Do the same in MockCoordinatorExecutor and GroupMetadataManagerTestContext, for use in future tests that roll back writes which schedule tasks. Co-Authored-By: Claude Opus 5.5 <[email protected]> -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
