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 2022/07/05 03:01:00 UTC
[jira] [Commented] (FLINK-28349) Consider using resource uid as fixed jobid for operator deployed jobs
[ https://issues.apache.org/jira/browse/FLINK-28349?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17562335#comment-17562335 ]
Yang Wang commented on FLINK-28349:
-----------------------------------
The upstream project Flink has supported to generate a static job id based on the cluster-id instead of ZERO.
https://github.com/apache/flink/commit/e70fe68dea764606180ca3728184c00fc63ea0ff
> Consider using resource uid as fixed jobid for operator deployed jobs
> ---------------------------------------------------------------------
>
> Key: FLINK-28349
> URL: https://issues.apache.org/jira/browse/FLINK-28349
> Project: Flink
> Issue Type: Improvement
> Components: Kubernetes Operator
> Reporter: Gyula Fora
> Priority: Major
>
> There are certain connectors that rely on jobid uniqueness to a certain degree (such as the iceberg sink connector) so it might make sense to use the FlinkDeployment uid as the fixed job id instead of the default 0000000000000000.
> cc [~matyas] [~wangyang0918]
> If we decide to do so we have to be mindful of the migration path for already running jobs with HA/last-state upgrade mode.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)