Rangsh commented on PR #12081: URL: https://github.com/apache/seatunnel/pull/12081#issuecomment-6110513044
@SEZ9 Thanks. Re-checked on `ab04db33a`; nothing new has landed since the 10-08 recap. - F1: no WAL caller uses untimed `RequestFuture.get()`. The only production waits are `requestFuture.get(timeout, MILLISECONDS)` and `requestFuture.get(remainingNanos, NANOSECONDS)` in `IMapFileStorage`. Untimed `get()` remains unbounded for the `Future` contract; that is stated on the method itself (`RequestFuture.java` L57–68). - F3/F7: `@throws TimeoutException` is on `get(long, TimeUnit)` itself, and expiry throws instead of returning false (`RequestFuture.java` L70–90). - F4: APPEND fail-closed is `1de6f1b88`. A write failure sets sticky `appendBlockedAfterWriteFailure` (`WALWorkHandler.java` L111–113); later APPENDs fail without touching the stream (L99–105). The writer is not reused. An `Error` from `writer.write()` stays tracked in #12492. - F2: `executeResponse()` guarding is `26ce5cb26` (`WALWorkHandler.java` L129–140): the cache lookup is inside the try, and the catch is `Exception`, so the timeout+remove race cannot kill the single worker. - F6: one shared `deadlineNanos` for the batch, not a fresh `writDataTimeoutMilliseconds` per entry (`IMapFileStorage.java` L352–377). - F8: a timed-out wait logs one WARN line (requestId / elapsed / limit) and the stack at DEBUG. ERROR plus stack is only the unexpected `catch (Exception)` branch (`IMapFileStorage.java` L331–345). Keeping that, not ERROR on every timeout. - F5: `HdfsWriterFlushSyncPathTest` was renamed to `HdfsWriterFlushCallCountTest` in `26ce5cb26`. That same commit added `mockito-junit-jupiter` test-scoped to `imap-storage-file/pom.xml` (now L76–80). It is not only transitive; `f08bdf0ac` later updated the comment. It will not show up in the `ab04db33a` diff. @DanielLeens Thanks for re-reviewing `ab04db33a`. No further commit from me on this head. I agree your two new notes are non-blocking and do not change outcomes covered by the tests. -- 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]
