Recovery using datastore on remote volume?

[hefferbub]

I love PBS, and I'm using it for a multi-tiered backup strategy.

I have a 2 node PVE 6.3 cluster where one is the primary that normally runs all VMS and the other is a backup that receives replicated VMs and also runs PBS. I also have a Synology NAS located in another building where I sync the PBS datastore vis CIFS based on techniques from this thread). None of these backups are encrypted.

My question is about this CIFS backup; in case of a fire or other disaster where my PBS server was destroyed, how exactly would I make use of the backups stored on the offsite CIFS volume to restore a new PVE/PBS instance?

What would the process look like, and is there any information (fingerprint, key, config file, etc.) I would need to independently preserve to make use of that datastore on a new PBS instance?

Thanks much!

 

[dcsapak]

from the pbs side:

setup a new pbs instance, mount your directory and add a new datastore pointint to that dir
caution: for now this last step has to be done manually via editing /etc/proxmox-backup/datastore.cfg, since we always try to preallocate the chunkdir, which fails if it already exists
we want to implement an 'import datastore' feature, but there is no timeline on that
(you can of course save that config somewhere else and restore it into your new pbs instance)

if your backups are unencrypted, you simply have to point a pve instance to that new pbs instance and use it
(if they are encrypted, you have to use your backed up encryption keys to create the storage)

hope this helps

 

[hefferbub]

Perfect. Thanks so much.

PBS is a great piece of technology!