You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Till Rohrmann (JIRA)" <ji...@apache.org> on 2019/02/06 09:20:00 UTC

[jira] [Commented] (FLINK-11534) Don't exit JVM after job termination with standalone job

    [ https://issues.apache.org/jira/browse/FLINK-11534?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16761604#comment-16761604 ] 

Till Rohrmann commented on FLINK-11534:
---------------------------------------

Would it be enough to make the {{ClusterEntrypoint#ExecutionMode}} configurable. If started {{ExecutionMode#NORMAL}}, the {{MiniDispatcher}} will wait with the cluster shut down until the job result has been served to a client.

> Don't exit JVM after job termination with standalone job
> --------------------------------------------------------
>
>                 Key: FLINK-11534
>                 URL: https://issues.apache.org/jira/browse/FLINK-11534
>             Project: Flink
>          Issue Type: New Feature
>          Components: Cluster Management
>    Affects Versions: 1.7.0
>            Reporter: Ufuk Celebi
>            Priority: Minor
>
> If a job deployed in job cluster mode terminates, the JVM running the StandaloneJobClusterEntryPoint will exit via System.exit(1).
> When managing such a job this requires access to external systems for logging in order to get more details about failure causes or final termination status.
> I believe that there is value in having a StandaloneJobClusterEntryPoint option that does not exit the JVM after the job has terminated. This allows users to gather further information if they are monitoring the job and manually tear down the process.
> If there is agreement to have this feature, I would provide the implementation.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)