SEZ9 commented on PR #12081:
URL: https://github.com/apache/seatunnel/pull/12081#issuecomment-6008212410

   @Rangsh Thanks for the follow-up and for re-dispatching. Agreed on the read: 
`ClusterFaultToleranceIT#testStreamJobRestoreInAllNodeDown` failing once on JDK 
8 only, while green on `1cbac0086` (both JDKs), on JDK 11, and now on the 
no-code-change re-run of `ab04db33a`, is consistent with a one-off flake rather 
than something introduced here. Good to see 
`CheckpointCoordinatorFailoverIT#testBatchJobCompletesAfterMasterFailoverDuringCloseHandshake`
 green on both JDKs — that closes out the `notifyCheckpointMonitor` regression 
we traced.
   
   On the `SplitClusterFaultToleranceIT` options: I don't have the alternatives 
you're referring to in front of me in this thread, so I can't pick one blindly. 
Could you restate the options (briefly, with the trade-off you see for each) in 
a comment here? I'll give you a decision right after.
   
   Two small things to close the loop before I do a final pass:
   
   1. Please confirm where the earlier review points stand on `ab04db33a` — in 
particular the `RequestFuture.get()` blocking/timeout semantics (unbounded 
untimed `get()`, `TimeoutException` on the timed variant and its method-level 
javadoc), the `WALWorkHandler` worker-death / writer-reuse-after-failure 
concerns, the sequential full-timeout wait in `batchQueryExecuteFailsStatus`, 
and the ERROR-level stack trace on every timed-out wait in 
`queryExecuteStatus`. A short "fixed in this commit / intentionally unchanged 
because X" per item is enough.
   2. `HdfsWriterFlushSyncPathTest` uses Mockito — please confirm whether the 
`imap-storage-file` module already has that test dependency available, or 
whether a pom change is still needed.
   
   Once I have those, I'll weigh in on the `SplitClusterFaultToleranceIT` 
question and wrap up the review.
   
   <!-- streview-comment:1544 -->


-- 
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