LuciferYang commented on PR #364: URL: https://github.com/apache/doris-spark-connector/pull/364#issuecomment-4854909137
Thanks @JNSimba — you're right on both counts. Reusing a label stuck in `PREPARE` can't succeed until Doris aborts it, and without 2PC there's no end-to-end exactly-once to protect anyway, so minting a new label on retry plus an idempotent (primary-key) table is the cleaner strategy. I've reverted the label-reuse change: stream load keeps minting a fresh label per batch, and `testFailoverForRetry` now uses a UNIQUE (primary-key) table so a re-loaded batch is deduplicated by the key instead of producing duplicate rows. Since that's a Spark-4.1 failover-IT change it rides along on #365, so I'm closing this PR. Appreciate the careful review and the patience. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
