lianetm commented on PR #16974: URL: https://github.com/apache/kafka/pull/16974#issuecomment-2312333360
> The case (close(0)) you described inspires me that processBackgroundEvents can return before processing ConsumerRebalanceListenerCallbackNeededEvent due to timeout. That means ConsumerRebalanceListener is NOT invoked and it is an inconsistent behavior. Yeap, that would be another case, good catch. The member would stay PREPARE_LEAVING (and maybe add the CallbackNeededEvent in the background), but if the app thread does not have the time to see it while `processBackgroundEvents` we would move onto the `pollOnClose`. I included this consideration too, so basically if leaving (preapring or leaving) when closing the manager, makes sense to send the leave request, it's truly to last chance to do it before the network layer shuts down. All comments addressed (and merged with the nice HBMgr refactoring). Thanks! -- 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: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
