[
https://issues.apache.org/jira/browse/FLINK-29813?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17727525#comment-17727525
]
Matthias Pohl commented on FLINK-29813:
---------------------------------------
Just as a fyi: FLINK-32204 was caused by the {{TreeCache}} not being
thread-safe when shutting down the passed {{executorService}}. If there's a
late event coming in that is processed in the curator's event thread while the
cache is closed in another thread, it might be that we run into a
{{RejectedExecutionException}} because the executor might have been shutdown
while the late event is submitted to be processed (see the [FLINK-32204
comment|https://issues.apache.org/jira/browse/FLINK-32204?focusedCommentId=17726613&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-17726613]).
This became apparent when FLINK-31995 was introduced which added shutdown
handling to the {{DirectExecutorService}}.
> Remove deprecated Curator API usages
> ------------------------------------
>
> Key: FLINK-29813
> URL: https://issues.apache.org/jira/browse/FLINK-29813
> Project: Flink
> Issue Type: Technical Debt
> Components: Runtime / Coordination
> Reporter: Chesnay Schepler
> Assignee: Chesnay Schepler
> Priority: Major
> Fix For: 1.18.0
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
