Chia-Ping Tsai created KAFKA-21437:
--------------------------------------
Summary: CoordinatorContext#unload should cancel the running load
to release the thread slot
Key: KAFKA-21437
URL: https://issues.apache.org/jira/browse/KAFKA-21437
Project: Kafka
Issue Type: Improvement
Reporter: Chia-Ping Tsai
Assignee: Chia-Ping Tsai
If the leadership is lost and regained while loading, the previous load keeps
running until it reaches the log end offset, even though its CoordinatorContext
was already unloaded. The loader has a single thread, so the load of the new
context has to wait behind it. We could add a small flag to stop the fetch
early when the context is unloaded.
See KAFKA-21436 for the related issue.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
