[
https://issues.apache.org/jira/browse/FLINK-32814?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17752171#comment-17752171
]
Gyula Fora commented on FLINK-32814:
------------------------------------
I am not sure why anyone would want to keep the JM there forever. The only
reason would be investigation , but how much time do you really need ? The logs
are probably in splunk or somewhere else already. So even a month is more than
enough.
We should only add configs when it makes sense. Here I don’t really see the
value to be honest .
> Support disabling failed JM shutdown in the operator
> ----------------------------------------------------
>
> Key: FLINK-32814
> URL: https://issues.apache.org/jira/browse/FLINK-32814
> Project: Flink
> Issue Type: New Feature
> Components: Kubernetes Operator
> Affects Versions: kubernetes-operator-1.5.0
> Reporter: Yaroslav Tkachenko
> Priority: Major
> Labels: pull-request-available
>
> Currently, it's not possible to disable the shutdown of the jobmanager pods
> in a terminal state, e.g. {*}FAILED{*}. I may want to keep failed jobmanager
> pods around until they should be deleted explicitly, e.g. after an
> investigation.
> The TTL for the shutdown is controlled by the
> *kubernetes.operator.jm-deployment.shutdown-ttl* option.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
