[
https://issues.apache.org/jira/browse/FLINK-24232?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17412671#comment-17412671
]
Chesnay Schepler commented on FLINK-24232:
------------------------------------------
Some additional thoughts need to be given for how the HistoryServer is supposed
to load these archives, because again it assumes for a given job ID to only be
used once.
> Archiving of suspended jobs prevents breaks subsequent archive attempts
> -----------------------------------------------------------------------
>
> Key: FLINK-24232
> URL: https://issues.apache.org/jira/browse/FLINK-24232
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Coordination
> Affects Versions: 1.14.0
> Reporter: Chesnay Schepler
> Priority: Major
>
> To archive a job we write a file that uses the job ID as the name. Since
> suspended jobs are handled like other terminal jobs they are also being
> archived.
> When that job then later resumes any attempt to archive the job on
> termination will fail because an archive already exists.
> The simplest option is to add a suffix if an archive already exists, like
> "_1".
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
