Converted my daily driver to my vm/docker host. I assumed all my images would be included in timeshift’s snapshots, but… nope!
Fortunately I figured it out before I needed it. was testing my backups as part of prep for wiping everything and going to headless debian, found libvirt was missing.
Edit: Added rsync tag to reflect the comments section, which really deserves its own thread.
Basically what it says in the title: do you pull your backups a backup-manager host, or push them from each machine individually?
My initial inclination is to do it all from one machine, with one unprivileged user and a single cron file. That means, though, that there needs to be a remote user on each machine with g+r to everything, in order for ssh+rsync to do its thing.
Or, would it be simpler for each server to manage its own backups, probably as root, which means not having to change group permissions or identity?
Yes, I definitely make use of rsync for backups of specific directories, so, in this case, no data would have been lost if I had relied solely on timeshift. But it’s a very convenient way of snapshotting the whole system.
Timeshift is essentially a gui on top of rsync anyway, if you look at what it’s doing behind the scenes.