[
https://issues.apache.org/jira/browse/FLINK-20653?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17253859#comment-17253859
]
Yun Tang commented on FLINK-20653:
----------------------------------
If checkpoint {{chk-23}} registers successfully but fails to store in
checkpoint storage. The pending {{chk-23}} would not complete in the end and
would be discarded/unregistered when discarding the pending {{chk-23}}.
When recovering from the checkpoint, we would never recover from {{chk-23}} but
a previous successful one. Thus, we don't need to know the deleted files in the
uncompleted {{chk-23}}.
> SharedState should be registered after PendingCheckpoint storeCheckpoint
> ------------------------------------------------------------------------
>
> Key: FLINK-20653
> URL: https://issues.apache.org/jira/browse/FLINK-20653
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Checkpointing
> Affects Versions: 1.12.0
> Reporter: weiyunqing
> Priority: Major
> Attachments: image-2020-12-17-20-20-21-246.png
>
>
> SharedState should be registered after PendingCheckpoint storeCheckpoint，If
> the metadata storage fails, the deleted incremental state cannot be recovered。
> !image-2020-12-17-20-20-21-246.png!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
