The motivation of introducing `default.api.timeout.ms` is that user may want to avoid infinite blocking behavior. And I agree that MM should not block forever on offset commit when topic is deleted. The question here is that, if MM offset commit takes more than 60 seconds due to reasons other than topic deletion (e.g. long queue time in broker), should MM keep retrying offset commit?
It seems that if long request queue time causes offset commit in `onPartitionsRevoked ()` to timeout , most likely the next MM who owns the partition will also fail to commit offset, and in the end MM will just keep consuming/producing without committing offset which can cause issue. This seems to suggest that MM should just keep retrying offset commit in this scenario. What do you think? [ Full content available at: https://github.com/apache/kafka/pull/5492 ] This message was relayed via gitbox.apache.org for [email protected]
