Skip to content

Conversation

@dwash96
Copy link
Contributor

@dwash96 dwash96 commented Sep 19, 2025

…e restoration

[FLINK-38279]

This issue currently prevents restoring snapshot jobs from being restart-able from save points/checkpoints which means that large table and database replication must be done in one shot. Any thing that crashes the snapshotting process creates an entirely non-recoverable situation which is highly inconvenient and un-workable in many set ups.

@lvyanquan
Copy link
Contributor

Nice catch!
And could you add a test for this?

@lvyanquan lvyanquan added this to the V3.6.0 milestone Sep 24, 2025
@yuxiqian
Copy link
Member

Hi @dwash96, do you have time to add an E2e test for this change? YamlJobMigrationITCase might be suitable for such a test. Also, please run mvn spotless:apply to fix code style violations.

@dwash96
Copy link
Contributor Author

dwash96 commented Oct 22, 2025

Oh man, sorry, I will do this by saturday at the latest

dwash96 and others added 2 commits November 14, 2025 14:58
…e restoration

[FLINK-38279]

This issue currently prevents restoring snapshot jobs from being restart-able from save points/checkpoints which means that large table and database replication must be done in one shot. Any thing that crashes the snapshotting process creates an entirely non-recoverable situation which is highly inconvenient and un-workable in many set ups.
@yuxiqian
Copy link
Member

Hi @dwash96 @lvyanquan, I've pushed commits to add corresponding test cases and resolve conflicts. Would you like to take another look?

Copy link
Contributor

@leonardBang leonardBang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks @dwash96 and @yuxiqian for the contribution, LGTM

@leonardBang leonardBang merged commit 6767d85 into apache:master Nov 14, 2025
30 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants