3 ms·
I'm late to the party, but I do have a question. For me, reliability and data backups/recovery play an important role in setting up systems like this. I find t
by movedx 4y ago
I'm late to the party, but I do have a question.
For me, reliability and data backups/recovery play an important role in setting up systems like this. I find that if I think about setting up a self-host solution, my mind goes to, "But what if it fails and you lose all your effort in a theft, fire, flood, or just hardware/data failure?"
And that side of my brain would be right: what if that happened?
So I guess I'd ask the author: how do you handle this niggling feeling in your brain, if you get it like I do.
- philjohn 4y agoI can't answer for the author, but I can answer for myself. First of all, I have a little bit of fault tolerance with using ZFS and ZRaid 2 with 6 disks - so I'm good until 2 disks have failed. But more than that, everything is backed up offsite using Restic and the JottaCloud RClone backend - the data I don't want to lose is around 4TB (backups from all computers in the house, and SeaFile sync of all mobile devices) so whilst the initial upload wasn't super fast (although not too bad as I've got a 70Mbps upload speed) periodic sync each day is super quick.
- ornornor 4y agoSame thoughts as you. Anything I run also takes care of backing itself up using tarsnap or borg to rsync.net (depending on when I originally setup that particular application) Some things you can just copy the files on a regular schedule (with tarsnapper or borgmatic), others you also have to stop the DB and take a dump of it before restarting it. And everything I can set up again with a collection of custom ansible playbooks and roles. But yes I wouldn’t self host without a reliable backup strategy and auto updates (with unattended-upgrades).