[
https://issues.apache.org/jira/browse/FLINK-26336?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17497168#comment-17497168
]
Gyula Fora commented on FLINK-26336:
------------------------------------
I think we are not doing cancel:
[https://github.com/apache/flink-kubernetes-operator/blob/main/flink-kubernetes-operator/src/main/java/org/apache/flink/kubernetes/operator/controller/FlinkDeploymentController.java#L82]
[https://github.com/apache/flink-kubernetes-operator/blob/3e6a00f0eb528ad85b91a99424258925ea9304d5/flink-kubernetes-operator/src/main/java/org/apache/flink/kubernetes/operator/utils/FlinkUtils.java#L102]
So do you suggest we change the cleanup logic to cancel first, and then delete
everything in case cancel gets stuck?
> Deletion should remove HA related configmaps also
> -------------------------------------------------
>
> Key: FLINK-26336
> URL: https://issues.apache.org/jira/browse/FLINK-26336
> Project: Flink
> Issue Type: Sub-task
> Components: Kubernetes Operator
> Reporter: Gyula Fora
> Priority: Critical
>
> At the moment the deletion logic simply deletes the Flink job deployment.
> As [~wangyang0918] correctly pointed out in another ticket, deployment
> deletion does not remove HA related configmaps (this is intentional from
> Flink side)
> This means if the user were to resubmit a new FlinkDeployment under the same
> name, the job would pick up the HA state and the last checkpoint from the
> delete resource. This is probably unexpected and we should delete the
> configmaps after deleting the deployment.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
