-
Bug
-
Resolution: Obsolete
-
Major
-
2.1.0.GA
The use case here is a bulk topic deletion after an OpenMessaging Benchmark (OMB) test run, which seems to break the TO instance.
./bin/benchmark --drivers driver-kafka/kafka-all.yaml --workers "$WORKERS" workloads/max-rate-20-topics-1-partition-1kb.yaml
After successfully running this test to the end, the TO logs this warning.
2022-06-15 14:46:11,72820 WARN [vert.x-eventloop-thread-2] ZkTopicsWatcher:152 - Reconciliation #2147(/brokers/topics 61:+test-topic-joCSwfo-0004) KafkaTopic(amq-streams-kafka-sit/test-topic-joCSwfo-0004): Error responding to creation of topic test-topic-joCSwfo-0004 org.apache.kafka.streams.errors.InvalidStateStoreException: The state store, topic-store, may have migrated to another instance. at org.apache.kafka.streams.state.internals.WrappingStoreProvider.stores(WrappingStoreProvider.java:67) ~[org.apache.kafka.kafka-streams-3.1.0.redhat-00003.jar:3.1.0.redhat-00003] ...
The above test leaves a bunch of test topics around that you probably want to get rid of by running a bulk topic deletion.
kubectl exec -it kafka-cluster-sit-kafka-0 -c kafka -- ./bin/kafka-topics.sh --bootstrap-server=:9092 --topic 'test-topic.*' --delete
The issue is that these topics are deleted, but then immediately recreated by the TO with the same spec and no data (topicId changes). Instead, the associated KafkaTopic CRs are deleted as expected.
Workaround: if you restart the TO pod before doing the bulk topic deletion, then it works fine.
Attaching a reproducer, which also includes the full TO log.
- impacts account
-
-
- Closed
-
- links to
