chia7712 opened a new pull request #9128: URL: https://github.com/apache/kafka/pull/9128
issue: https://issues.apache.org/jira/browse/KAFKA-7540 In short, the LEAVE_GROUP is sent to a shutdown broker so the dead member which is left in the group obstructs broker from completing rebalance. The test case creates two consumers (with different group) and then shutdown their coordinator (broker). After first coordinator is shutdown, the heartbeat thread of first consumer gets timeout so it sends LEAVE_GROUP request to “next” coordinator. Unfortunately, the LEAVE_GROUP request can’t be processed successfully if the ”next” coordinator is the coordinator of second consumer (yep, it is shutdown also). The simple approach is that we can reduce the session timeout so coordinator can evict dead member in time. ### Committer Checklist (excluded from commit message) - [ ] Verify design and implementation - [ ] Verify test coverage and CI build status - [ ] Verify documentation (including upgrade notes) ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
