[
https://issues.apache.org/jira/browse/KAFKA-18686?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17922856#comment-17922856
]
Matthias J. Sax commented on KAFKA-18686:
-----------------------------------------
When the test fails, we see it "hang" restoring data. The log line as blow (or
similar) keeps repeating, until the test times out:
{code:java}
[2025-01-31 11:13:56,117] INFO state-updater
[app-c6cbc5a3-4b43-4cb4-b033-6a94ecee1f64-e6ff7490-93fb-4218-bfc6-9b245fdc09d7-StateUpdater-2]
Restoration in progress for 1 partitions.
{app-c6cbc5a3-4b43-4cb4-b033-6a94ecee1f64-counter-c6cbc5a3-4b43-4cb4-b033-6a94ecee1f64-changelog-3:
position=1, end=2, totalRestored=0}
(org.apache.kafka.streams.processor.internals.StoreChangelogReader:580) {code}
For a restore, KS first fetches "end-offset" of the partition to read from, and
consumer until it reaches the end-offset. For unknown reason, we don't reach
the end-offset and thus hang.
> Flaky test
> StandbyTaskEOSMultiRebalanceIntegrationTest#shouldHonorEOSWhenUsingCachingAndStandbyReplicas
> -------------------------------------------------------------------------------------------------------
>
> Key: KAFKA-18686
> URL: https://issues.apache.org/jira/browse/KAFKA-18686
> Project: Kafka
> Issue Type: Test
> Components: streams, unit tests
> Reporter: 黃竣陽
> Assignee: Matthias J. Sax
> Priority: Major
>
> It seens this test flaky these days
> [https://github.com/apache/kafka/actions/runs/13061784839/job/36451152058]
> [https://github.com/apache/kafka/actions/runs/13067934631/job/36463514470?pr=18763]
> https://github.com/apache/kafka/actions/runs/13068036157/job/36463989916?pr=18641
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
