[
https://issues.apache.org/jira/browse/FLINK-17248?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17087928#comment-17087928
]
Yun Tang commented on FLINK-17248:
----------------------------------
[~trohrmann] I plan to add {{cluster.io-executor.pool-size}} to
{{ClusterOptions}} and set it as {{-1}} which means undefined. The default
behavior would be the same as current to act as the number of cpu cores.
> Make the thread nums of io executor of ClusterEntrypoint and MiniCluster
> configurable
> -------------------------------------------------------------------------------------
>
> Key: FLINK-17248
> URL: https://issues.apache.org/jira/browse/FLINK-17248
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Checkpointing, Runtime / Coordination
> Affects Versions: 1.10.0
> Reporter: Yun Tang
> Assignee: Yun Tang
> Priority: Major
>
> FLINK-11851 introduce dedicated io executor for ClusterEntrypoint and
> MiniCluster with the the thread num as the cpu cores. However, this could
> introduce extra regression as we might use fewer threads compared with akka
> service before, e.g.FLINK-17073 . However, we cannot configure the thread num
> of the io executor to avoid such regression.
> One idea is to introduce such configuration to make the thread nums of io
> executor of ClusterEntrypoint and MiniCluster configurable.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
