[
https://issues.apache.org/jira/browse/FLINK-18785?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17171299#comment-17171299
]
Kai Chen commented on FLINK-18785:
----------------------------------
I found that flink job will normally exit if I set yarn.application-attempts=1
or just delete this config
> flink goes into dead lock leader election when restoring from a do-not-exist
> checkpoint/savepoint path
> ------------------------------------------------------------------------------------------------------
>
> Key: FLINK-18785
> URL: https://issues.apache.org/jira/browse/FLINK-18785
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Checkpointing, Runtime / Coordination
> Affects Versions: 1.10.0, 1.10.1
> Environment: flink on yarn
> flink-1.10.x
> jdk8
> yarn.application-attempts: 2
> Reporter: Kai Chen
> Priority: Major
> Attachments: image-2020-07-31-19-04-19-241.png
>
>
> flink goes into dead lock leader election when restoring from a do-not-exist
> checkpoint/savepoint path.
> I just run this cmd:
> bin/flink run -m yarn-cluster -s "hdfs:///do/not/exist/path"
> examples/streaming/
> WindowJoin.jar
> when i visit UI，i meet this:
> !image-2020-07-31-19-04-19-241.png!
> in flink-1.9.3, the program just exits. But in 1.10.x, it stucks in leader
> election
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
