sunm2n commented on PR #23898:
URL: https://github.com/apache/kafka/pull/23898#issuecomment-6651834327
I checked this locally: with the fix, AbstractCoordinatorTest passes
(55/55, the new test 10/10 on repeat), and with only the main change reverted,
the new test is the only failure.
This looks like the bug in KAFKA-12585 (unassigned, no PR), so the title
could reference it.
It also seems related to KAFKA-17263. The reporter's thread dump shows
several heartbeat threads waiting on the same coordinator after repeated
fencing. On trunk, HeartbeatThread.close() calls notify() once, so a leaked
thread could take that wakeup while the current thread is in the untimed
wait(), and closeHeartbeatThread() would block in join(). This change should
prevent the leak, but I haven't reproduced the close hang itself.
--
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]
