aiquestion commented on PR #12349:
URL: https://github.com/apache/kafka/pull/12349#issuecomment-1173895320

   @showuon @RivenSun2 i refined some of the code, pls help to take a look when 
you have time:
   * as @RivenSun2 suggest, use a `autoCommitOffsetRequestFuture` to store 
commit offset request.
   * still wait for the response but limit to poll's timer, instead of sleep 
backoff when onJoinPrepare failed. This is to avoid busy loop.
   
   I think both previous change and this will will not meet the issue in 
[KAFKA-13310](https://issues.apache.org/jira/browse/KAFKA-13310) because, if 
commit offset failed:
   * in EAGER mode we will revoke all partitions and next CommitOffset will 
alway be success
   * in COOPERATIVE mode we will revoke all partitions that are not in the 
metadata anymore, so next CommitOffset will be ok.
   
   But in this latest change, in EAGER mode. We will wait for commit offset 
request before rejoin, but if commitOffset failed, no retry request will be 
sent since we revoke all partitions.


-- 
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

Reply via email to