[
https://issues.apache.org/jira/browse/FLINK-22054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17312237#comment-17312237
]
Till Rohrmann commented on FLINK-22054:
---------------------------------------
>From a high level it looks good to me [~yittg]. For the final review I would
>have to dig into the details of the {{SharedInformerFactory}} a bit.
It is probably fine to directly create the {{Fabric8FlinkKubeConfigMapWatcher}}
when we create the {{Fabric8FlinkKubeClient}}. Moreover, it should be good
enough if every {{Fabric8FlinkKubeClient}} has its own
{{Fabric8FlinkKubeConfigMapWatcher}} as implemented in the POC.
> Using a shared watcher for ConfigMap watching
> ---------------------------------------------
>
> Key: FLINK-22054
> URL: https://issues.apache.org/jira/browse/FLINK-22054
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Coordination
> Affects Versions: 1.12.2, 1.13.0
> Reporter: Yi Tang
> Priority: Major
> Labels: k8s-ha
> Fix For: 1.14.0
>
>
> While using K8s HA service, the watching for ConfigMap is separate for each
> job. As the number of running jobs increases, this consumes a large amount of
> connections.
> Here we proposal to use a shard watcher for each FlinkKubeClient, and
> dispatch events to different listeners. At the same time, we should keep the
> same semantic with watching separately.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
