SEZ9 commented on PR #11077: URL: https://github.com/apache/seatunnel/pull/11077#issuecomment-5738507060
@hesam-oxe thanks for the detailed rundown — responding to what came through. **F1** — Both parts sound right. Calling out divergent credentials and connection-level settings explicitly in the `getPhysicalDestinationIdentifier()` javadoc (mirrored in the en/zh sink-connector-development docs) is the guard I was after, and the INFO line on both the `createWriter` and `restoreWriter` paths when a further alias joins a shared writer — with connector class, physical destination identifier, joining table and first table — gives operators a way to spot an unintended merge. Adding a logger to `MultiTableSink` is fine. I'll verify the wording and log placement on `2036ad77` and consider F1 resolved once that checks out. **F2** — Option (a) is the behaviour I wanted: one persisted state per shared writer under `primaryIdentifier`, rather than fanning out to every alias. Your comment appears to have been cut off partway through the F2 explanation, so I couldn't see the rest — in particular how the restore side behaves now that only `primaryIdentifier` carries state (i.e. that restore no longer unions per-alias copies). Could you re-post that part? **F3–F8** — Nothing for these reached me because of the truncation. Could you re-post the item-by-item notes for them as well, noting whether each landed in `2036ad77` or was already in place on `ef2bb095`? Once I have the full list I'll re-review against the diff and close out the remaining findings. <!-- streview-comment:1149 --> -- 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]
