[
https://issues.apache.org/jira/browse/KAFKA-19696?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18020571#comment-18020571
]
Lucas Brutschy commented on KAFKA-19696:
----------------------------------------
Not sure what you mean by improving the logic. The point would be to properly
root cause the issue and find the right place to fix it. Probably, one of the
concurrent access to the data structures that need to be moved from the
processing thread to the polling thread, or the other way around. Likely, we
don't want to just add a lock here. This is not trivial, and I am not sure
right now somebody will have time to review such a big change. I would suggest
picking up a different issue at this point
> Re-enable SmokeTestDriverIntegrationTest for processing threads
> ---------------------------------------------------------------
>
> Key: KAFKA-19696
> URL: https://issues.apache.org/jira/browse/KAFKA-19696
> Project: Kafka
> Issue Type: Task
> Components: streams
> Reporter: Lucas Brutschy
> Priority: Major
>
> The test fails occasionally with a race condition like this:
> {code:java}
> 2_0 RUNNING StreamTask(active):
> org.apache.kafka.streams.errors.StreamsException:
> java.util.ConcurrentModificationExceptionorg.apache.kafka.streams.errors.StreamsException:
> java.util.ConcurrentModificationException at
> org.apache.kafka.streams.processor.internals.StreamThread.runLoop(StreamThread.java:977)
> at
> org.apache.kafka.streams.processor.internals.StreamThread.run(StreamThread.java:894)Caused
> by: java.util.ConcurrentModificationException at
> java.base/java.util.HashMap$HashIterator.nextNode(HashMap.java:1597) at
> java.base/java.util.HashMap$KeyIterator.next(HashMap.java:1620) at
> org.apache.kafka.clients.consumer.internals.ClassicKafkaConsumer.resume(ClassicKafkaConsumer.java:979)
> at
> org.apache.kafka.clients.consumer.KafkaConsumer.resume(KafkaConsumer.java:1509)
> at
> org.apache.kafka.streams.processor.internals.StreamTask.resumePollingForPartitionsWithAvailableSpace(StreamTask.java:623)
> at
> org.apache.kafka.streams.processor.internals.TaskManager.resumePollingForPartitionsWithAvailableSpace(TaskManager.java:1866)
> at
> org.apache.kafka.streams.processor.internals.StreamThread.runOnceWithProcessingThreads(StreamThread.java:1355)
> at
> org.apache.kafka.streams.processor.internals.StreamThread.runLoop(StreamThread.java:932)
> {code}
> h4. *Since the feature is not actively developed by anybody, we disable this
> test. This ticket is to re-enable the test when the race condition is fixed.*
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
