[
https://issues.apache.org/jira/browse/FLINK-25433?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17487223#comment-17487223
]
Matthias Pohl commented on FLINK-25433:
---------------------------------------
I removed FLINK-25085 as a blocking issue from this ticket. The
{{ResourceCleaner}} is implemented in a way, that the cleanup trigger
({{ResourceCleaner.cleanupAsync}}) is called from within the main thread but
the actual IO intense execution is moved into the {{IOExecutor}}. This enables
us from using the {{MainThreadExecutor}} (which implements the
{{ScheduledExecutor}} interface) to run the retry mechanism.
> Integrate retry strategy for cleanup stage
> ------------------------------------------
>
> Key: FLINK-25433
> URL: https://issues.apache.org/jira/browse/FLINK-25433
> Project: Flink
> Issue Type: Sub-task
> Reporter: Matthias Pohl
> Assignee: Matthias Pohl
> Priority: Major
> Labels: pull-request-available
>
> The {{ResourceCleaner}} should be able to cleanup not only once but retry
> infinitely.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
