Shekharrajak opened a new pull request, #69960: URL: https://github.com/apache/doris/pull/69960
### What problem does this PR solve? A late abort of a COMMITTED transaction acquires the Routine Load job lock in `beforeAborted`, then throws before `afterAborted` can release it. Journal submission failures can leak the same lock. This affects existing classic Kafka Routine Load. Pair successful before callbacks with after callbacks in `finally` for ordinary and 2PC aborts, outside the transaction monitor. Make `DeleteJob.afterAborted` retain its callback when no abort occurred, preserving delete completion after repeated rejected aborts. Independent of the Kafka share-consumer changes; based directly on upstream master. ### Release note Prevent Routine Load jobs from becoming stuck when abort processing fails. ### Check List (For Author) - Tests: all six lock regressions failed on unmodified upstream with the real Kafka Routine Load lock still held. After the fix, 25 focused FE executions passed through `run-fe-ut.sh` with supported OBS/COS exclusions, covering both abort routes, both journal modes, delete callback retention/removal, and nearby transaction behavior. - Style: repository-pinned Checkstyle 9.3 CLI with repository rules and whitespace checks passed. - Behavior changed: callback cleanup on failed aborts; no new configuration, wire fields, or persisted metadata. - No documentation change needed. -- 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]
