SEZ9 commented on issue #10203: URL: https://github.com/apache/seatunnel/issues/10203#issuecomment-6432182140
@ryanmeowy thanks for folding that into the 2a/2b section of the 23 Oct table comment — the wording you describe is exactly what I was after: the restore decision keys off the split's own watermark pair (`SnapshotSplit.isSnapshotReadFinished()`, both non-null), `finishedUnackedSplits` is rebuilt from that pair at `IncrementalSourceReader.addSplits:152-153`, and `restoreCompletedSnapshotSplit` applies the same predicate. That closes the question I raised on row 2b; nothing further needed on that point from my side. One small ask on the same row so the mid-backfill case reads unambiguously: please keep the explicit statement that a checkpoint taken during the bounded backfill persists the split with both watermarks null (HIGH is only on `SnapshotSplitChangeEventSourceContext` at that moment, per `MySqlSnapshotSplitReadTask:154`), so a restore re-runs it as 2a rather than 2b. If that sentence is already in the updated section, consider this done. On the three contract items — noted that they'll be in the revision as agreed; I'll review them there. <!-- streview-comment:2063 --> -- 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]
