[
https://issues.apache.org/jira/browse/KAFKA-4600?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16812168#comment-16812168
]
Braedon Vickers commented on KAFKA-4600:
----------------------------------------
Hi [~guozhang],
Sounds like we are on the same page about the issue itself now, which is great
:)
Option 1 in your comment seems sufficient to me - it's certainly the simpler
option of the two.
I'm not up on the meta around when KIPs need to be used, but this seems like a
pretty basic (but important) bug fix to me, rather than a breaking change to
functionality. Are people _really_ expecting uncaught exceptions thrown by
`ConsumerRebalanceListener` methods to be (effectively) ignored, and relying
intentionally/specifically on that behaviour for their implementations?
I'm sure there will be implementations that are silently failing currently
(like mine was before I spotted and worked around the issue) that will fail
loudly after this is fixed; all the more reason to patch this as a bug. As
noted in the original issue, this can easily cause data corruption or data loss.
Braedon
> Consumer proceeds on when ConsumerRebalanceListener fails
> ---------------------------------------------------------
>
> Key: KAFKA-4600
> URL: https://issues.apache.org/jira/browse/KAFKA-4600
> Project: Kafka
> Issue Type: Bug
> Components: consumer
> Affects Versions: 0.10.1.1
> Reporter: Braedon Vickers
> Priority: Major
>
> One of the use cases for a ConsumerRebalanceListener is to load state
> necessary for processing a partition when it is assigned. However, when
> ConsumerRebalanceListener.onPartitionsAssigned() fails for some reason (i.e.
> the state isn't loaded), the error is logged and the consumer proceeds on as
> if nothing happened, happily consuming messages from the new partition. When
> the state is relied upon for correct processing, this can be very bad, e.g.
> data loss can occur.
> It would be better if the error was propagated up so it could be dealt with
> normally. At the very least the assignment should fail so the consumer
> doesn't see any messages from the new partitions, and the rebalance can be
> reattempted.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
