SEZ9 commented on issue #10203: URL: https://github.com/apache/seatunnel/issues/10203#issuecomment-6513161536
@ryanmeowy thanks — confirmed. I re-read the "Completion signal (2a/2b)" paragraph in the [23 Oct table comment](https://github.com/apache/seatunnel/issues/10203#issuecomment-6357660652) and it now covers both halves I asked for: - the completion predicate is the split's own checkpointed watermark pair (`SnapshotSplit.isSnapshotReadFinished()`, both non-null), with `finishedUnackedSplits` rebuilt from that pair on restore (`IncrementalSourceReader.addSplits:152-153`) and `restoreCompletedSnapshotSplit` using the same predicate; and - a checkpoint taken mid-backfill lands in 2a because the HIGH watermark exists only as the backfill's stop offset (`MySqlSnapshotSplitReadTask:154`) and both `SnapshotSplitState` fields are still null. That makes the section self-contained without having to jump to the "On a death inside the backfill" paragraph of your [24 Oct reply](https://github.com/apache/seatunnel/issues/10203#issuecomment-6379487210), which is what I wanted — a reader of the table alone should not be able to misread 2b as "present in a finished-splits map". Also noted that nothing else in that comment changed. Nothing further from me on the 2a/2b point; I consider it closed. <!-- streview-comment:2126 --> -- 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]
