SEZ9 commented on issue #11735: URL: https://github.com/apache/seatunnel/issues/11735#issuecomment-5807043373
Thanks @goutamadwant for the update. Durable terminal cleanup and failover recovery, expiry anchored to the original terminal time, late-write fencing, exact REST routing, and no stored worker-address attribution now read as concrete v1 boundaries, and keeping the diagnostic attempt identity separate from the retry counter is the right call for preserving existing retry semantics. Before this moves beyond the design stage, a few asks within the scope already raised: 1. **Diagnostic-attempt advance under active-master change** — please spell out how the advance is fenced/idempotent when the master changes, and confirm it cannot alter retry eligibility. 2. **Best-effort failure writes** — state explicitly that the asynchronous write path neither blocks nor re-enters the failure/restore path, and describe how a write failure is handled (dropped vs. retried). 3. **`TaskExecutionState` wire compatibility** — document how existing serialization is preserved before optional structured failure fields are added. 4. **Test coverage** — list the tests that will cover the three boundaries above, late writes after terminal cleanup, and the REST distinction between a known job with no retained records and an unknown/expired job. Once those are frozen in the design, the implementation can be split into small, independently reviewable slices. No assignment or label change from me in this pass; I'll revisit that once the design is settled. <!-- streview-comment:1284 --> -- 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]
