[
https://issues.apache.org/jira/browse/KAFKA-16018?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17798908#comment-17798908
]
Matthias J. Sax commented on KAFKA-16018:
-----------------------------------------
Timing is a little unclear to be fair – there is an open discussion if the next
release (ie, after 3.7 that we are currently rolling out) will be 3.8 or 4.0...
But even if it's 3.8, it's an open question if it's worth to put a fix into 3.8
or just wait for 4.0 (not sure how complex a fix would be...)
> KafkaStreams can go into a zombie state if UncaughtExceptionHandler is
> specified via the deprecated method
> ----------------------------------------------------------------------------------------------------------
>
> Key: KAFKA-16018
> URL: https://issues.apache.org/jira/browse/KAFKA-16018
> Project: Kafka
> Issue Type: Bug
> Components: streams
> Affects Versions: 3.3.2
> Reporter: Tommy Becker
> Priority: Major
>
> We have a streams application in which all StreamThreads died due to a lack
> of disk space. To our surprise, the KafkaStreams instance still reported its
> state as running. Upon further investigation, it appears this is due to the
> application setting an UncaughtExceptionHandler via the deprecated method
> (this application was recently upgraded from 2.4.1):
> [https://kafka.apache.org/33/javadoc/org/apache/kafka/streams/KafkaStreams.html#setUncaughtExceptionHandler(java.lang.Thread.UncaughtExceptionHandler)]
> The only way a StreamThread failure can cause the KafkaStreams instance to
> transition to an error state now is via the new
> StreamsUncaughtExceptionHandler machinery, but when an old
> UncaughtExceptionHandler is set by the old method this code is effectively
> bypassed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
