[
https://issues.apache.org/jira/browse/FLINK-23831?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17410311#comment-17410311
]
Yangze Guo commented on FLINK-23831:
------------------------------------
Testing job: 4 sources, 3 of which are bounded and will finish in 30s, 60s, 90s
respectively
https://github.com/KarmaGYZ/flink/blob/test-147/flink-examples/flink-examples-streaming/src/main/java/org/apache/flink/streaming/examples/wordcount/WordCount.java
Ckp type: unaligned, aligned, aligned with timeout
Testing cases:
- stop with savepoint
- stop --drain
- randomly trigger failover
One issue recorded in FLINK-24161.
> Verify the final checkpoint and stop-with-savepoint --drain
> -----------------------------------------------------------
>
> Key: FLINK-23831
> URL: https://issues.apache.org/jira/browse/FLINK-23831
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Checkpointing
> Affects Versions: 1.14.0
> Reporter: Yun Gao
> Assignee: Yangze Guo
> Priority: Blocker
> Labels: release-testing
> Fix For: 1.14.0
>
>
> In 1.14 we support checkpoints with part of tasks finished, and based on this
> functionality, we further support the final checkpoint and unify the process
> of stop-with-savepoint --drain.
> In order to test the new feature, we may need to verify:
> 1. Checkpoint with the finished tasks.
> 2. Restore from the checkpoint with finished tasks.
> 3. All the data are committed correctly for bounded jobs with 2pc sinks.
> 4. Stop-with-savepoint --drain remains correct.
> 5. All the above tests also work with unaligned checkpoint.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
