You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "Yang Wang (Jira)" <ji...@apache.org> on 2022/04/18 08:13:00 UTC

[jira] [Commented] (FLINK-27281) Improve stability conditions for rollback logic

    [ https://issues.apache.org/jira/browse/FLINK-27281?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17523596#comment-17523596 ] 

Yang Wang commented on FLINK-27281:
-----------------------------------

This might be related with FLINK-27257 if we choose to improve the {{isJobRunning}} implementation.

> Improve stability conditions for rollback logic
> -----------------------------------------------
>
>                 Key: FLINK-27281
>                 URL: https://issues.apache.org/jira/browse/FLINK-27281
>             Project: Flink
>          Issue Type: Improvement
>          Components: Kubernetes Operator
>            Reporter: Gyula Fora
>            Priority: Major
>             Fix For: kubernetes-operator-1.0.0
>
>
> This is a follow up to https://issues.apache.org/jira/browse/FLINK-26140
> We should allow more configurable / sophisticated stability conditions for the rollback behaviour.
> 2 things that come to mind:
>  - Somehow detect that the job was running without failures before it is considered stable
>  - Require one completed checkpoint (optional, depending on the checkpoint period etc)



--
This message was sent by Atlassian Jira
(v8.20.1#820001)