zhengruifeng commented on PR #57754: URL: https://github.com/apache/spark/pull/57754#issuecomment-5253216212
@dongjoon-hyun Thanks for the detailed review. Addressed in c5c91ba9ac2: - The PR description now calls out the offloaded-model cleanup/removal semantics and states that the status covers models only (`Summary` objects are excluded). - Live ML cache snapshots are collected before entering `store.synchronized`. - The live-status API is now a single optional map containing each active session's optional snapshot, which distinguishes unavailable/removed state (`N/A`) from an uninitialized cache (`Not used`). - `getStatus` now builds the metadata snapshot directly without the `ArrayBuilder`. - A focused production-path test covers `SessionManager -> SessionHolder -> MLCache -> SparkConnectServerTab -> page`, including the last-access-time and lazy-initialization assertions. The soft-value removal-listener timing remains an accepted transient limitation. The focused page suite passes all 7 tests. -- 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]
