Backup Server is using some other format.

[sherani1987]

Hi Everyone.
The Backup server is backing-up the VM with other format. When i took the backed up image and would like to pout in other drive. It wont be recognized. I dont know why the size and the format of this backup is different then the one that is backup in the Proxmox VE local hard disk.

 

[Hannes Laimer]

Hey,

PBS uses chunks and index files that reference those chunks, which is used for deduplication and incremental backups, so the image is not saved directly but can be reconstructed with the index file. You are able to download the image from the WebUI. For more information about how PBS works internally maybe take a look a [1].

[1] https://pbs.proxmox.com/docs/technical-overview.html

 

[sherani1987]

Thanks for your reply. Is it possible that if I can take those chunks and connect with another Proxmox virtual environment and from those chunks files launch the new one VM

 

[tom]

Yes, the idea of backups is doing restores.

 

[sherani1987]

Yes, the idea of backups is doing restores.

But i took those chunks files and then i placed them in proxmox ve dump folder and reload that storage and it was lnt showing those files that i can restore them

 

[tom]

You need a Proxmox Backup Server to restore backups created by the Proxmox Backup Server.

 

[sherani1987]

Ok..... i got it

 

[David Thistlethwaite]

You need a Proxmox Backup Server to restore backups created by the Proxmox Backup Server.

Tom, how does PBS backup the dedup database, as I am concerned about my backup server failing, and hence restorations from catastrophic single site failure. (offsite backups to come later, in various forms)

Thanks for your time and attention

 

[tom]

Tom, how does PBS backup the dedup database

there is no "dedup" database, so I am not sure what you mean here.

 

[David Thistlethwaite]

there is no "dedup" database, so I am not sure what you mean here.

I have just been reading the manual on the dedup functionality, and it looks like the hashing algorithm ‘ generates’ the uniq chunks and then indexes them, so reconstruction of the data from the chunks can be done with just the data store of chunks and the indexes. ( and a pbs server)
Do you have any more docs on the subject?
Thank you for your time

 

[tom]

https://pbs.proxmox.com/docs/

Source code:
https://git.proxmox.com/