JingsongLi commented on PR #9907: URL: https://github.com/apache/paimon/pull/9907#issuecomment-5708350989
LGTM. The race this addresses is real: "a new snapshot appeared" does not mean a commit past the aborted savepoint region — the coordinator commits asynchronously, so a commit already in flight when the region failed can bump the snapshot id while the region is still restarting, and the follow-up savepoint is then rejected. Waiting on `commitIdentifier() > savepointCheckpointId()` is the right condition. I also verified the new `hasFailed()` guard cannot pass spuriously: `FAILED` is a static `AtomicBoolean` set only via `compareAndSet(false, true)`, and `@BeforeEach resetInjectors()` clears it (together with the other static injectors) before every test in the class. -- 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]
