SEZ9 commented on PR #11487:
URL: https://github.com/apache/seatunnel/pull/11487#issuecomment-6332601102
Thanks for the two follow-up commits (`d28296a1` and `39c796a2`). Looking at
`39c796a2ee6f`, the delta since `d28296a1` is the new
`OceanBaseIncrementalSourceFactoryTest#testOptionRuleInheritsSnapshotOnlyGuards`
plus a constraints line in the en/zh `OceanBase-CDC.md`. That is a nice
addition for OceanBase, but I could not find changes for the MySQL-CDC doc
items from my previous round, so here is where I think each one stands:
- **F1 (server-id/parallelism mismatch under snapshot-only + exactly_once)**
– still only surfaces at runtime via MySQL. Could you either add a fail-fast
configuration check, or let me know if you prefer to keep this as a documented
limitation? Either is fine, but I would like it to be an explicit decision.
- **F2 (consistency semantics with `exactly_once = false` under concurrent
writes)** – not yet documented. A short paragraph on what users can and cannot
expect from the snapshot in that mode would be enough.
- **F3 (constraints only in the `startup.mode` cell and prose)** – the
`stop.mode`, `startup.specific-offset.*` and `startup.timestamp` rows in
`docs/en/connectors/source/MySQL-CDC.md` still do not carry the snapshot-only
restriction themselves. A brief note on each affected row, similar to the
OceanBase constraints line, would close this.
- **F4 (ZH parity)** – please mirror the EN `exactly_once` row and FAQ
updates in `docs/zh/connectors/source/MySQL-CDC.md`.
- **F5 (literal credentials in the new example)** – please replace the
hardcoded password with a placeholder in the snapshot-only example.
- **F6 ("rejected by MySQL" wording)** – please reword so it reflects what
actually happens with a shared `server-id` rather than implying MySQL rejects
the connection outright.
- **F7 (example sets a server-id range without exactly_once)** – please
either enable exactly_once in the example or adjust the preceding paragraph so
the example matches its stated condition.
- **F8 (reworded FAQ answer for initial mode)** – please restore the precise
description of the binlog resume position for `initial` mode; the snapshot-only
addition should not loosen the existing text.
If any of these were addressed in a way I missed, just point me to the hunk
and I will re-check. Once the doc items above are in, I am happy to take
another pass.
<!-- streview-comment:1955 -->
--
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]
