My goal is to to fully ditch Google Photos for Immich. I have about ~3TB of photos and videos. Looking for a super simple way of backing up the library to cloud storage in case of a drive failure without spending a ton.
Ideally, this will require nothing on my part besides copying files into a given folder. And ideally the storage will be encrypted and have basic privacy assurances.
Also if it matters my home server is running Debian. But I’d prefer something that runs in docker so I can more easily check on it remotely.
Not the same, but for my Immich backup I have a raspberry pi and an HDD with family (remote).
Backup is rsync, and a simple script to make ZFS snapshots (retaining X daily, Y weekly). Connected via “raw” WireGuard.
Setup works well, although it’s never been needed.
Is this the way to go for off-site backups w/ family? In terms of low power draw, uptime, etc.
That absolutely works, but when I built my offsite backup to hetzner I also thought about setting up own hardware and came to conclusion that for myself it doesn’t really make a ton of sense. New RPi + 4TB ssd/m.2 drive with accessories adds up to something around 400€ (if that’s even enough today), or few years worth of cloud backups. With own hardware there’s always need to maintain it and hardware failures are always an option, so for me it makes more sense to just rely on big players with offsite backups. Your case might be different for various reasons, but sometimes renting capacity just makes more sense in the big picture.
Why would you use SSDs for backup? I think a HDD should be fine for that.
Especially because SSDs start losing data is they’re powered off for some time.
Sound and power consumption. At least in my case those are important if I was going to store data at my mothers house. Power consumption might not matter that much, but HDD sound definetly does. And even with spinning rust hardware cost would be somewhere around 250€ compared to ~20€/month of cloud storage.
YMMV, in my scenario it’s just easier to use a cloud provider.
I’ve been pleased with it. Family is very relaxed about projects like this, but yeah it’s low power draw. I don’t think I have anything special set up but the right thing to do for power would be to spin down drive when not in use, as power is dominated by the spinning rust.
Uptime is great. Only hiccups are that it can choke when compiling the ZFS kernel modules, triggered on kernel updates. It’s an rpi 3/1GB RAM (I keep failing at forcing dkms to use only 1 thread, which would probably fix these hiccups 🤷).
That said, it is managed by me, so sometimes errors go unnoticed. I had recent issues where I missed a week of rsync because I switched from pihole to technitium on my home server and forgot to point the remote rpi there. This would all have been fixed with proper cron email setup…I’m clearly not a professional :)
If you’re already running ZFS, sanoid would be an option.
Okay, how do you get sanoid & syncoid to run, because I’ve tried, and I’m just too dummy. When it makes a backup, is it literally making a zfs data record/pool/whatever on the other machine? Or is it more like a file? I have a Proxmox running cockpit (SMB & NFS) and the machine is connected to a USB drive bay that has ZFS. My immich is saving pictures to my ZFS drive bay via SMB.
I’ve tried to do
syncoid pool_name/data/immich root@cockpit.service.IP.addr:mnt/samba/backupsbut I get hit with:
Long ass error message
WARNING: ZFS resume feature not available on target machine - sync will continue without resume support. INFO: Sending oldest full snapshot Orico2tera4/data/immich@syncoid_nova_2026-01-27:13:38:44-GMT-05:00 to new target filesystem root@192.168.0.246:/mnt/samba/backups (~ 42 KB): /dev/zfs and /proc/self/mounts are required. Try running 'udevadm trigger' and 'mount -t proc proc /proc' as root. 44.2KiB 0:00:00 [ 694KiB/s] [===========================================] 103% CRITICAL ERROR: zfs send 'Orico2tera4/data/immich'@'syncoid_nova_2026-01-27:13:38:44-GMT-05:00' | pv -p -t -e -r -b -s 43632 | lzop | mbuffer -q -s 128k -m 16M | ssh -S /tmp/syncoid-root1921680246-1772385641-845218-1784 root@192.168.0.246 ' mbuffer -q -s 128k -m 16M | lzop -dfc | zfs receive -F '"'"'/mnt/samba/backups'"'"' 2>&1' failed: 256I’ve tried reading the github docs and some forums but I’m dummy. I just want to have backups that I can encrypt and keep in a cloud for cheap somewhere. Does it literally have to be two different machines (god I’m dumb)? Can I just auto run ZFS snapshots and encrypt then save those to Drive/OneDrive/Whoever?
You can do a sanoid sync to another zpool or dataset on the same machine or a remote host, they behave the same. It’s replicating that dataset on the other machine, then sending the snapshots after that point over via
zfs send. You can instruct sanoid to prune those snapshots after the send and start new ones for the next send, or just accumulate them so you have points in time to revert to.IIRC, you can send a zfs snapshot to a file, but I can’t recall how to do that, so AFAIK, you can’t just send it to a file based service like Onedrive. You can use a service like zfs.rent and send them a harddrive with your base sync on it (encrypt it) and then once they’ve brought it online, you can sync to that. Best to test out your methods with the drive hooked up locally.
I know it’s anathema to Lemmy, but the best help you’ll get is Claude where you can paste the errors in and have it sort it out for you as you troubleshoot. It’s pretty good at shit like that.