restore incrementals?

P

Peter123

Well-Known Member
Mar 5, 2018
50
3
48
55
I have Proxmox running on a server and the Proxmox Backup Server there as a virtual machine. The backup server storage space is mounted on the Proxmox as backup storage.

The first backup is known to be full, the others incremental.

When I try to restore an incremental backup, I only see the full backups in the Proxmox server on the mounted storage space.

How do I see the incremental ones or how do I restore them in time?
 
the difference between a full or incremental backup is just how much data is transferred when backing up. each backup on the server is just a collection of data chunks + metadata, the first and all subsequent ones are equal.
 
Sorry for the late reply.

But how can I select a specific restore point for restoring in the GUI or is that only possible via the command line?
 
you can select any snapshot, no matter whether it was done incrementally or "full", as long as you have permission to access it on the PBS side.
 
Ah, ok, got it.

One more little thing: the Proxmox GUI always shows the full size of the VM instead of the backup size for the content of the backup storage.

Are there any plans to only display the exact size of the full or incremental backups here?
 
no, because that is not really possible to determine cheaply. you can only store how much data you uploaded (which we do, it's not really exposed nicely), but that does not tell you anything about how much that backup takes up on disk. you basically have three levels of deduplication going on at once:
- inside a snapshot (chunks which are repeated, mainly the 'zero' chunk filled with all zero bytes)
- inside a backup group (chunks which have not changed between backup snapshots, or are often created by the system being backed up)
- inside a datastore (chunks which are coming from common templates, or common workloads that lead to duplicated data being written)

it's easy to say 'this snapshots logically consists of X chunks and logically contains Y bytes of data', as that is fixed. everything else is relative to all the other stuff in the datastore, and can change after the fact. that is why at garbage collection time, we display a deduplication status for the whole datastore which gives you an indication of how effective the deduplication works in the big picture.
 
You must log in or register to reply here.
Top Bottom