[
https://issues.apache.org/jira/browse/FLINK-26916?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17514568#comment-17514568
]
Yang Wang commented on FLINK-26916:
-----------------------------------
For safety and compatibility, I agree with you that upgrade should be done by a
new job submission, especially when the Flink version is changed. However, if
the continuously increased checkpoint could be easily obtained by external
tools(e.g. flink-kubernetes-operator), I also lean to not simply delete the
Flink cluster and have HA data retained. I do not find a more appropriate way.
Note: We could not get the latest checkpoint from the RestAPI since JobManager
might crash backoff when we want to upgrade a Flink application.
I am afraid I cannot agree with you about that clusterId shouldn't be reused.
Users just need to do the manual clean-up for job result store if they want to
reuse the same cluster-id.
I am not familiar with native savepoint and will have a look about this
solution.
> The Operator ignores job related changes (jar, parallelism) during last-state
> upgrades
> --------------------------------------------------------------------------------------
>
> Key: FLINK-26916
> URL: https://issues.apache.org/jira/browse/FLINK-26916
> Project: Flink
> Issue Type: Bug
> Components: Kubernetes Operator
> Affects Versions: kubernetes-operator-0.1.0, kubernetes-operator-1.0.0
> Reporter: Matyas Orhidi
> Assignee: Yang Wang
> Priority: Blocker
> Labels: pull-request-available
> Fix For: kubernetes-operator-0.1.0
>
>
> RC: The old jobgraph is being reused when resuming
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
