[
https://issues.apache.org/jira/browse/KAFKA-21196?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Quah updated KAFKA-21196:
------------------------------
Description:
When CoordinatorExecutorImpl tasks are cancelled, they remain in the executor
queue or the event processor queue. When the group coordinator is overloaded,
these queues can grow large and retain data structures that are no longer
needed.
One way we can allow the memory to be reclaimed by GC is to box the callbacks
held by tasks and release them on cancellation.
was:
When CoordinatorExecutorImpl tasks are cancelled, they remain in the executor
queue or the event processor queue. When the group coordinator is overloaded,
these queues can grow large and retain data structures that are no longer
needed.
One way we can free the memory is to box the callbacks held by tasks and free
them on cancellation.
> Release memory held by CoordinatorExecutorImpl tasks when cancelled
> -------------------------------------------------------------------
>
> Key: KAFKA-21196
> URL: https://issues.apache.org/jira/browse/KAFKA-21196
> Project: Kafka
> Issue Type: Improvement
> Components: group-coordinator
> Reporter: Sean Quah
> Assignee: David Jacot
> Priority: Major
>
> When CoordinatorExecutorImpl tasks are cancelled, they remain in the executor
> queue or the event processor queue. When the group coordinator is overloaded,
> these queues can grow large and retain data structures that are no longer
> needed.
> One way we can allow the memory to be reclaimed by GC is to box the callbacks
> held by tasks and release them on cancellation.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
