SEZ9 commented on PR #13060:
URL: https://github.com/apache/seatunnel/pull/13060#issuecomment-7022916161
@zephyr45 thanks for the rest of the note. I'm replying to the text posted
in this thread for head `117dc411`; I have not gone through the linked full
note (6967720978) yet.
- **F4:** Noted your report that
`SparkSinkWriterContext#getStableSubtaskIndex()` now returns the partition ID
(`9e4aad72`), with the tests you list. I'll keep this open until I've checked
it against the diff.
- **F7:** Likewise noted for the EN/ZH `Kafka.md` changes (`7887e909`,
`9e4aad72`); I'll check the option row and the Zeta/Flink scoping against the
diff before closing it.
- **F5/F6:** I don't see that agreement in this thread, but deferring these
two low-severity items to the follow-up you reference is fine with me.
- **F3:** Could you clarify what "ended" means here? The finding was that a
writer restoring with no state (scale-up) starts its transaction sequence at 1,
so a failed first attempt can leave an unfenced open transaction until the next
checkpoint. Is there a code or doc change that addresses this, and if so, in
which commit?
- **F1/F2:** These are not covered in the text I have. Could you post their
status here, with the commit and tests for each?
- **F1:** Flink scale-down restore of an EXACTLY_ONCE sink.
- **F2:** Scale-down restore when `transaction_prefix` is not configured.
I've also noted your report that fork Build 44311890713 is green on
`117dc411`.
<!-- streview-comment:2454 -->
--
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]