SEZ9 commented on PR #12081: URL: https://github.com/apache/seatunnel/pull/12081#issuecomment-6567424607
@Rangsh thanks for the pointers on `ef99d22d1`. **F6 (shared batch deadline).** One deadline computed before the loop in `IMapFileStorage.batchQueryExecuteFailsStatus`, each entry waiting only for the remaining budget, and `get(0, NANOSECONDS)` still harvesting already-completed futures is the shape I was hoping for. I'll verify it against the diff before marking F6 resolved. **F4 (writer not reused after a failed append).** Setting `appendBlockedAfterWriteFailure` on the first append failure and answering later APPEND events with `done(false)` before touching the writer addresses the torn-record concern. One question: how does the handler leave the blocked state (is the writer replaced / the file rolled, or does it stay blocked until restart)? If staying blocked is intentional, a short comment on the flag saying so would be enough. For the remaining items (F1, F2, F3, F5, F7, F8) I'll go through `ef99d22d1` directly; if you have pointers to the relevant hunks, that would speed things up. I'll follow up once I've done the full pass. <!-- streview-comment:2202 --> -- 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]
