SEZ9 commented on PR #11512:
URL: https://github.com/apache/seatunnel/pull/11512#issuecomment-5747066945

   Thanks @goutamadwant for the rebase in d552a4e2b and for keeping the CDC 
feature diff intact; good to hear the 36 focused tests across API, engine 
server and CDC base pass locally.
   
   On the Databend failure: your read (checkpoint 4 expiring, the Flink global 
committer hanging during cancellation, the watchdog taking down the 
TaskManager, then zero slots on retry) is consistent with a fixture-level 
problem rather than anything in the new CDC-reporting path, which that 
FakeSource/Flink fixture doesn't exercise. Agreed that another submission retry 
wouldn't help there. If you can link the other PR showing the same sequence in 
this thread, that gives us the baseline needed to treat it as unrelated for 
this PR.
   
   Since F4 and F6 were rechecked at `d8cb0cbb7d3`, what I still need before I 
can wrap up the earlier review is the state of the remaining points on 
d552a4e2b:
   
   - **F1 / F7 / F8** (`CdcEnumeratorProgressReport`): is `activeSplits` now 
bounded to match the documented contract, are the count fields validated for 
negative/inconsistent values, and is the immutability claim either enforced for 
`CdcSnapshotSplitProgress` elements or softened in the Javadoc?
   - **F2** (`docs/en/developer/cdc-progress.md`): does the payload contract 
now explicitly forbid credentials/connection secrets in connector-native 
position payloads?
   - **F3** (`CdcProgressAccuracy` and the other public enums): does the engine 
codec encode them by name rather than ordinal?
   - **F5** (`docs/en/developer/cdc-progress.md`): does the doc now list which 
connectors implement the progress provider?
   
   A short per-item note (addressed / where, or deferred and why) is enough. 
Once the fresh CI run lands, I'll do the full pass on the rebased head.
   
   <!-- streview-comment:1178 -->


-- 
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