dataproc-templates icon indicating copy to clipboard operation
dataproc-templates copied to clipboard

Kafka DStream templates starts to fail intermittently

Open vanshaj-bhatia opened this issue 1 year ago • 1 comments

The Kafka to GCS/BQ templates starts failing intermittently and continues to fail in consecutive runs, unless the Kafka topics are recreated.

Failure Error: org.apache.kafka.clients.consumer.OffsetOutOfRangeException: Fetch position FetchPosition{offset=5215, offsetEpoch=Optional.empty, currentLeader=LeaderAndEpoch{leader=Optional[instance-kafka.c.yadavaja-sandbox.internal:9094 (id: 2 rack: null)], epoch=0}} is out of range for partition integration-test-kafka-bq-dstream-0

Failure Reason: TBD

Steps to Reproduce: [To be verified]

  1. Create a new Kafka topic
  2. Schedule or run the template on a new topic
  3. Terminate the template run while in progress

Interim Solution/Workaround:

  • Recreate topics, which resets the Offset information of the topic.

vanshaj-bhatia avatar Feb 09 '24 05:02 vanshaj-bhatia

Currently, Kafka topics are recreated every week on the KafkaVM for integration tests to work.

vanshaj-bhatia avatar May 06 '24 07:05 vanshaj-bhatia