Can Duplicati backup Duplicati?

If I backup my home computers to my NAS (unRAID) using Duplicati, can I use the unRAID Docker version of Duplicati to back my NAS up to a cloud service with the home machine backups in tact?

I realize there is a potential timing issue, but that aside, is the data safe? (I’ve found at least one backup solution that strictly forbids it.)


This sort of thing has been discussed in the past and as far as I’m aware yes, it is possible and not blocked by the software or any constraints thereof. The consideration would be that if your NAS crashes and you’re in a position to need to restore files, you’d need to first restore ALL backup files to the NAS to be able to then restore ANY end files. But as long as you’re willing and able to accept that constraint, I’m not personally aware of any further issues with such a setup.

I really wouldn’t recommend doing that… It is MUCH better to just rsync the backup files to a remote location.
Running Duplicati on a Duplicati backup directory will just waste CPU and time making some extremely complicated nested backup.

I understand the complexity. And the timing.

I, in fact, don’t do my server backups using Duplicati; I use a different, paid service. I was, however, wondering about the technical merits. So far, I love Duplicati and wanted to see how robust it was.

I don’t really recommend it, either. I don’t even recommend doing what I’m doing–as someone else mentioned, besides the timing, I’d have to restore my whole server first, then restore each computer from there (hopefully).

It’s all just in the name of fireproofing/floodproofing/tornadoproofing my data. I actually do external hard-drive backups for each one (two different versions for my computer) plus Duplicati to my NAS/server and a backup of the server to the cloud from there. I take no chances.

Thanks everyone!

Ah, the dreadef FFT (fire, flood, tornado) trifecta. :slight_smile:

One other thing to consider is that during a backup Duplicati writes to its job-specific .sqlite file. This could leave the file on a locked state when you go to back it up.

It won’t hurt anything, but if you get errors logged that the file isn’t available you may want to try something like a –run-script-before-required call to make s copy of the database to be backed up. Sure it’ll be a version behind, but it’ll be there.

Though I agree - rsync or Syncthing of the backup elsewhere is probably easier for restore purposes.