Martijn Visser created FLINK-40589:
--------------------------------------
Summary: SourceTopicIntegrityTest.testTopicIntegrityFailure is
flaky: 2 second wait for the restored job to fail is too short on CI
Key: FLINK-40589
URL: https://issues.apache.org/jira/browse/FLINK-40589
Project: Flink
Issue Type: Bug
Components: Connectors / Kafka
Affects Versions: kafka-5.1.0
Reporter: Martijn Visser
Three failures within three days of the test landing on main, on JDK 11 and 17,
both parameters:
- weekly run 34116621983, job 101724841351 (2026-09-07, main, Flink
2.2-SNAPSHOT, JDK 17),
testTopicIntegrityFailure[1], elapsed 29.0 s
- push run 34116575758, job 101724687796 (2026-09-07, main, 2.2.1, JDK 11),
[2], 45.4 s
- PR run 34108450408, job 101699369340 (2026-09-07, PR #295, 2.2.1), [2], 46.4 s
java.util.concurrent.TimeoutException: Waiting for job to fail with
org.apache.flink.connector.kafka.source.enumerator.metadata.TopicIntegrityException:
Topic
SourceTopicIntegrityTest_source-topic is missing
at
org.apache.flink.core.testutils.CommonTestUtils.waitUtil(CommonTestUtils.java:218)
at
...SourceTopicIntegrityTest.testTopicIntegrityFailure(SourceTopicIntegrityTest.java:206)
The test stops the first job with a savepoint, tears down and restarts Kafka,
submits a second job from the savepoint and then waits only
ERROR_DISCOVERY_TIMEOUT = 2 s (line 69) for the archived execution graph to
show the TopicIntegrityException. On a loaded runner the second job has not
finished scheduling, restoring and running the first partition discovery within
2 s.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
