[
https://issues.apache.org/jira/browse/FLINK-25322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17746184#comment-17746184
]
Hangxiang Yu commented on FLINK-25322:
--------------------------------------
Hi, [~Feifan Wang]. Sorry for the late reply. Thanks for the proposal.
Could you help to verify my understanding:
# Users have to check the status of restore mode by the Flink UI or the REST
API, right ?
# If users stop the job before the 'slowest materilization of all subtasks',
this behaves like LEGACY mode, otherwise this could behaves like NO_CLIAM mode,
right ?
If true, I think it makes the restore mode a bit complex to understand and use.
I think claim mode is more suitable for most changelog use case.
Of course, IIUC, If users want to use NO_CLAIM mode, they'd like to retain a CP
to let other jobs use.
In this case, I think using native savepoint or full cp for the first
checkpoint may be useable even if these may cause minute-level latency
temporarly. (BTW, I have committed the PR about supporting Changelog Native
Savepoint, you could also help to take a review if avaliable)
Maybe I missed something about your proposal, please correct me if I missed.
> Support no-claim mode in changelog state backend
> ------------------------------------------------
>
> Key: FLINK-25322
> URL: https://issues.apache.org/jira/browse/FLINK-25322
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Checkpointing, Runtime / State Backends
> Reporter: Dawid Wysakowicz
> Assignee: Feifan Wang
> Priority: Major
> Fix For: 1.18.0
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
