weizhouapache commented on PR #14386: URL: https://github.com/apache/cloudstack/pull/14386#issuecomment-6437383035
> Thanks @weizhouapache, all the refactor comments are addressed. > > The merges with main broke `nasbackup.sh`, though. The first merge ([eb41943](https://github.com/apache/cloudstack/commit/eb41943f5bee73daa5cb3d0c868dbcddfa06b9c7)) dropped most of #14305, and the second one didn't bring it back: > > 1. `QUIESCE_TIMEOUT=""` and the `--quiesce-timeout)` case are missing from the arg parser. The agent passes `--quiesce-timeout 30` by default when quiesce is on, so every quiesced backup now fails with `Invalid option: --quiesce-timeout`. This hits all storage types, not just RBD. > 2. `abort_backup_job()` was deleted, but the Linstor pull-mode path still calls it (L537, L715). The pull job never gets aborted, so the scratch devices stay busy and the next backup or migration of the VM fails. > 3. The push-mode path is back to the pre-#14305 version: no freeze timeout, thaw only if the freeze succeeded, no thaw retry, no job abort in `cleanup()`, no retry on `domjobinfo`. > > The easiest fix is probably to copy those parts from main as they are. A diff against main's `nasbackup.sh` shows exactly what's missing. > > On restoring to other storage: with incremental enabled, the RBD full is also written as `.rbddiff` (`rbdFullStart`), so every chained RBD backup can only be restored to RBD, not just the incrementals. Only legacy fulls can go to other storage. Please call this out in the docs. thanks @abh1sar ! I will check and rekick the test after fixing -- 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]
