Philip Nee created KAFKA-16819: ---------------------------------- Summary: CoordinatorRequestManager seems to return 0ms during the coordinator discovery Key: KAFKA-16819 URL: https://issues.apache.org/jira/browse/KAFKA-16819 Project: Kafka Issue Type: Bug Components: consumer Reporter: Philip Nee Assignee: Philip Nee
In KAFKA-15250 we discovered the ConsumerNetworkThread is looping without much backoff. The in-flight check PR fixed a lot of it; however, during the coordinator discovery phase, CoordinatorRequestManager would keep on returning 0 before the coordinator node was found. The impact is minor but we should be expecting the coordinator manager to backoff until the exp backoff expired (so it should return around 100ms). -- This message was sent by Atlassian Jira (v8.20.10#820010)