rahul05ranjan commented on PR #23567: URL: https://github.com/apache/kafka/pull/23567#issuecomment-5830442194
Thanks for the detailed explanation — that makes sense, and I agree the current sequence is fine. You're right on both points: 1. Placing the reset after \stopAndAwaitConnector\ + \ emoveConnectorConfig\ + \onDeletion\ ensures the connector and its tasks are fully stopped before the active-topic set is cleared, so nothing can re-record topics after the reset. And the \DistributedHerder\ revocation path does the same (reset after the connector is no longer in \configState\), so this is consistent with the established pattern. 2. A \ inally\ around the whole sequence would indeed run the reset even when an earlier step failed (connector not fully stopped/deleted), which would be wrong. And since standalone \onDeletion()\ only writes to the in-memory \MemoryStatusBackingStore\ (no external I/O), the risk of it throwing is negligible. So no need for the narrower \ ry/finally\ — the current cleanup sequence and existing exception handling are the right call. My original comment was a non-blocking robustness note, and your reasoning resolves it. Thanks again for the fix and the thorough follow-up! -- 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]
