MartijnVisser opened a new pull request, #28351:
URL: https://github.com/apache/flink/pull/28351

   ## What is the purpose of the change
   
   `CheckpointAcknowledgeFailureITCase.testCheckpointAckFailure` can hang the 
entire surefire fork on a slow CI machine until the 900s no-output watchdog 
kills it (observed on `master`, [build 
75716](https://dev.azure.com/apache-flink/apache-flink/_build/results?buildId=75716)).
 The test sets a deliberately tiny `pekko.ask.timeout` so the oversized 
checkpoint-ACK RPC times out (load-bearing for the `AskTimeoutException` 
assertion), but that timeout applies to every cluster RPC. On a slow agent an 
unrelated RPC (e.g. task deployment) can time out and fail the job terminally 
before the keyed state is updated, so the future the test waits on never 
completes and the unbounded wait wedges the fork.
   
   This makes the wait fail fast with the real cause and adds a hard timeout so 
the test can never hang a fork again. It is a test-only stabilization; the 
assertion is unchanged.
   
   ## Brief change log
   
     - Propagate a terminal job failure into `stateUpdatedFuture` via a 
`whenComplete` handler on `getJobExecutionResult()` (unwrapping the 
`CompletionException` to the real cause), so the wait fails fast instead of 
blocking forever.
     - `executeJobAsync` now returns the `JobClient` so the test can observe 
job termination.
     - Add `@Timeout(5, MINUTES)` as the hard anti-hang guard (consistent with 
`ApproximateLocalRecoveryDownstreamITCase` in the same package).
   
   ## Verifying this change
   
   This change is a test stabilization of an existing test; the assertion 
(`triggerCheckpoint(...)` must throw `CheckpointException` caused by 
`AskTimeoutException`) is unchanged. Verified locally:
   
     - Happy path passes: `mvn -pl flink-tests 
-Dtest=CheckpointAcknowledgeFailureITCase test` → `Tests run: 1, Failures: 0`.
     - Fail-fast path: temporarily forcing the ask timeout very low so an 
unrelated RPC fails the job before the state update — the test now fails fast 
(~24s) with the job's real `JobExecutionException` instead of hanging until the 
watchdog kill.
   
   ## Does this pull request potentially affect one of the following parts:
   
     - Dependencies (does it add or upgrade a dependency): no
     - The public API, i.e., is any changed class annotated with 
`@Public(Evolving)`: no
     - The serializers: no
     - The runtime per-record code paths (performance sensitive): no
     - Anything that affects deployment or recovery: JobManager (and its 
components), Checkpointing, Kubernetes/Yarn, ZooKeeper: no (test-only change)
     - The S3 file system connector: no
   
   ## Documentation
   
     - Does this pull request introduce a new feature? no
     - If yes, how is the feature documented? not applicable
   
   ---
   
   ##### Was generative AI tooling used to co-author this PR?
   
   - [X] Yes (Claude Opus 4.8 (1M context))
   
   Generated-by: Claude Opus 4.8 (1M context)
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to