ArafatKhan2198 commented on PR #11560: URL: https://github.com/apache/ozone/pull/11560#issuecomment-6262010855
Thanks @devmadhuu for the detailed review, all three points were valid. I have addressed them. 1. **No retry after a failed full snapshot.** `syncDataFromOM` now fetches a full snapshot when the directory of Recon's current OM DB no longer exists. This covers the failed open from this PR and also the existing case of a failed download, which leaves Recon in the same state. 2. **`initializeOmTables` failing after the DB opened.** Switching to a new OM DB is now all or nothing. `initializeNewRdbStore` takes the already opened DB. If its tables cannot be initialized, it closes the new DB, restores the previous DB and re-initializes the tables on it, and rethrows. `updateOmDB` then deletes the new snapshot dir. If re-initializing the previous tables also fails, `omTablesInitialized` is set to false, so Recon reports the tables as not initialized instead of serving a mix. 3. **Startup swallowing table init failures.** `start()` now only skips a last known snapshot that cannot be opened. A failure to initialize its tables propagates as before, so `OzoneManagerServiceProviderImpl.start()` sets `INTERNAL_ERROR` again. -- 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]
