[
https://issues.apache.org/jira/browse/FLINK-30214?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17638602#comment-17638602
]
Gyula Fora commented on FLINK-30214:
------------------------------------
Hi [~fanrui],
Thanks for jumping on this and testing it.
Probably the minicluster is started with fix resources and that is not enough
after scaling up. You could try to allocate more tas slots in the minicluster
than the job initially needs so you can scale up.
> The resource is not enough
> --------------------------
>
> Key: FLINK-30214
> URL: https://issues.apache.org/jira/browse/FLINK-30214
> Project: Flink
> Issue Type: Bug
> Reporter: Rui Fan
> Priority: Major
> Attachments: image-2022-11-25-17-18-36-935.png,
> image-2022-11-25-17-23-14-208.png
>
>
> When turn up the parallelism, the resources isn't enough for MiniCluster. I'm
> not sure whether the flink on yarn or k8s is right? I can test later.
>
> I guess the parallelism should be changed in a right place. And we should add
> more unit test or ITCase to check these cases. And it's my honor to make
> these improvements and bug fixes.
> !image-2022-11-25-17-18-36-935.png!
>
> !image-2022-11-25-17-23-14-208.png!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
