jmsperu commented on PR #13074: URL: https://github.com/apache/cloudstack/pull/13074#issuecomment-4827330491
@abh1sar pushed `7c2574b2` addressing the 5 failing cases from your 24-Jun run: - **13** (full→incr after volume migration): `composeParentBackupPaths` now names the parent file from the volume **path** (`Backup.VolumeInfo.getPath()`) instead of the uuid — they diverge after a volume migration. - **17 / 19** (bitmap missing on all/one disk): the incremental pre-flight now requires the parent bitmap on **every** disk (was "present on any disk"), so a volume snapshot restore or partial migration self-heals to a full backup. - **35** (delete last backup): `deleteBackupFileAndRow` clears the VM's `active_checkpoint_id` when the deleted backup owns that bitmap, so the next backup starts a fresh full chain. - **37** (delete non-leaf still shows in list view): replaced the `nas.delete_pending` detail with **`Backup.Status.Hidden`** as you suggested — tombstones are filtered from `listBackups`, and Restore/RestoreAndAttach/CreateInstance/Delete are refused for free since they already require `BackedUp`. Chain GC is unaffected because `listByVmId` is status-agnostic, so the sweep still sees Hidden ancestors. Dropped the now-unused `DELETE_PENDING` key. Re your **No-Harm when the feature is disabled** note (and the `sanity_checks` / `rebase_backup` comments) — those are already handled by the earlier commits: the legacy-full path persists no chain/checkpoint metadata and doesn't touch `active_checkpoint_id` via the `isLegacyFull()` guard. `NASBackupProviderTest` passes 15/15 (updated for the Hidden tombstone). Could you re-run the suite when you have a slot? 🙏 -- 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]
