DanielLeens commented on issue #12310: URL: https://github.com/apache/seatunnel/issues/12310#issuecomment-5679932984
Confirmed against current dev: a non-200 response is logged, `flush()` then clears `batchBuffer`, and `prepareCommit()` returns normally. That permits a checkpoint to advance beyond a failed batch, so this is a connector durability defect rather than an HTTP serialization question. PR #12312 fixes the correct boundary: it propagates failure before the buffer is cleared, so a later successful `prepareCommit()` retries the same batch. Its Build is still in progress, so this issue is not ready to close. Please also cover an `IOException` in array mode through `prepareCommit()` and a flush failure during `close()` that still closes the client; the submitted test already covers the non-200 checkpoint path and object-mode exception. Do not add automatic HTTP retry semantics here: POST delivery remains at-least-once and receivers must be idempotent. -- 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]
