You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Yun Tang (Jira)" <ji...@apache.org> on 2022/04/01 10:23:00 UTC
[jira] [Updated] (FLINK-26985) With legacy restore mode, incremantal checkpoints would be deleted by mistake
[ https://issues.apache.org/jira/browse/FLINK-26985?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Yun Tang updated FLINK-26985:
-----------------------------
Description:
Before flink-1.15, the restored checkpoint would be regsiterd and not discard on subsume, which means the restored incremental checkpoint would have one more reference counting to avoid discard.
However, after state registry refactored, we could delete artificats in the restored incremental checkpoint with legacy restore mode.
The error could be reproduced via {{ResumeCheckpointManuallyITCase#testExternalizedIncrementalRocksDBCheckpointsStandalone}} in my [local branch|https://github.com/Myasuka/flink/tree/legacy-error].
Thanks for [~masteryhx] who found this problem in the manual test.
was:
Before flink-1.15, the restored checkpoint would be regsiterd and not discard on subsume, which means the restored incremental checkpoint would have one more reference counting to avoid discard.
However, after state registry refactored, we could delete artificats in the restored incremental checkpoint with legacy restore mode.
The error could be reproduced via {{ResumeCheckpointManuallyITCase#testExternalizedIncrementalRocksDBCheckpointsStandalone}} in my [local branch|https://github.com/Myasuka/flink/tree/legacy-error].
> With legacy restore mode, incremantal checkpoints would be deleted by mistake
> -----------------------------------------------------------------------------
>
> Key: FLINK-26985
> URL: https://issues.apache.org/jira/browse/FLINK-26985
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Checkpointing
> Affects Versions: 1.15.0
> Reporter: Yun Tang
> Priority: Blocker
>
> Before flink-1.15, the restored checkpoint would be regsiterd and not discard on subsume, which means the restored incremental checkpoint would have one more reference counting to avoid discard.
> However, after state registry refactored, we could delete artificats in the restored incremental checkpoint with legacy restore mode.
> The error could be reproduced via {{ResumeCheckpointManuallyITCase#testExternalizedIncrementalRocksDBCheckpointsStandalone}} in my [local branch|https://github.com/Myasuka/flink/tree/legacy-error].
> Thanks for [~masteryhx] who found this problem in the manual test.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)