syncing backups from one host to another

[sliao]

Hi

I'm trying to migrate backup copies from an existing backup server to a new server (with new harddrives etc..) using the remote syncing but it would only pull in the latest copies, citing

root@t340:~# proxmox-backup-manager pull backup hdd ltstorage
pull datastore 'ltstorage' from 'backup/hdd'
----
Syncing datastore 'hdd', root namespace into datastore 'ltstorage', root namespace
found 5 groups to sync (out of 5 total)
skipped: 35 snapshot(s) (2023-11-27T02:00:04Z .. 2024-12-10T20:00:02Z) - older than the newest snapshot present on sync target
skipped: 35 snapshot(s) (2023-11-27T02:00:15Z .. 2024-12-10T20:03:58Z) - older than the newest snapshot present on sync target
skipped: 18 snapshot(s) (2024-11-18T02:00:56Z .. 2024-12-10T20:05:27Z) - older than the newest snapshot present on sync target
skipped: 16 snapshot(s) (2024-11-25T02:01:06Z .. 2024-12-10T18:06:07Z) - older than the newest snapshot present on sync target
re-sync snapshot vm/2100/2024-12-10T20:06:43Z
no data changes
percentage done: 80.00% (4/5 groups)
skipped: 17 snapshot(s) (2024-11-18T02:01:12Z .. 2024-12-10T18:06:31Z) - older than the newest snapshot present on sync target
re-sync snapshot vm/7500/2024-12-10T20:07:07Z
no data changes
percentage done: 100.00% (5/5 groups)
Finished syncing root namespace, current progress: 4 groups, 1 snapshots
pull datastore 'ltstorage' end
TASK OK

I understand from this thread that this is the intended behaviour, only syncing the newest snapshots. Is there a way to override this behaviour? I want to make a copy of all the snapshots, including the oldest one

Thanks,

 

[Chris]

I understand from this thread that this is the intended behaviour, only syncing the newest snapshots. Is there a way to override this behaviour? I want to make a copy of all the snapshots, including the oldest one

Yes, this is by design. You could however create a new namespace on the sync target, and create a job to sync all the groups and snapshots into that namespace. By this, you will sync all snapshots present on the source, as non are present on the target namespace. And since chunks are deduplicated on the datastore level, this will not increase space requirement.

An alternative would be a new dedicated datastore for this, but without the benefits of the deduplication.