DanielLeens commented on issue #12566: URL: https://github.com/apache/seatunnel/issues/12566#issuecomment-5917740608
Classification: B / Zeta recovery contract and operational guidance. Current `dev` has a recovery path named `restoreAllRunningJobFromMasterNodeSwitch()`. Its documented trigger is a node becoming the active master; it scans the shared running-job map and returns when that map is empty. Separately, `CheckpointManager` restores checkpoint state only when the job is explicitly a restore job and has a non-null `restoreSourceJobId`. Those code paths establish master-switch recovery behavior, but they do not establish automatic, durable recovery after every master and worker has restarted. In particular, they do not wait for a full-cluster MapStore reload or synthesize a restore source for an ordinary submission. Until a full-cluster-restart contract is designed and verified, use an explicit savepoint/restore workflow for controlled recovery rather than relying on automatic resumption. If you would like to turn the two observations into a design or bug proposal, please add the exact `dev` commit, MapStore configuration and startup logs showing map-load timing, plus a minimal before/after restart case that records the saved checkpoint and the resumed source offsets. That will let us separate persistence readiness from checkpoint-selection semantics. -- 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]
