DanielLeens commented on issue #12713: URL: https://github.com/apache/seatunnel/issues/12713#issuecomment-6108865558
Classification: A / Connector-V2 WebSocket reconnection lifecycle. Thank you for tracing the callback and reader paths. Current dev confirms that `fatalError` is read by `rethrowFatalIfNeeded()` on every reader poll, while `onDisconnected()` only schedules a reconnect. #12716 identifies the stale field, but clearing it in a later successful `onOpen()` alone does not prove the reported recovery works: with the default 1-second poll timeout and 3-second reconnect interval, the reader can rethrow the recorded send failure before the scheduled reconnect reaches `onOpen()`. Please keep one focused implementation, but revise it so a recoverable open-message failure cannot fail the task before the configured reconnect policy has exhausted. Add a deterministic reader-level regression that drives the pending-reconnect interval, a successful replacement connection, and the terminal exhausted/disabled-reconnect case. The present callback-only test does not exercise that timing boundary. #12716 is not in dev: its Build is ACTION_REQUIRED and regular review remains. The same WebSocket change is currently stacked in #12717 and #12718; please keep each root-cause PR independently mergeable rather than merging unrelated fixes together. -- 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]
