That’s good to know. I’ve set up a snapshot I’ll look into it all a bit more as it seems like a good way of keeping things safe.
1 Like
To elaborate a bit on the answers you’ve already got: What redundancy (be it a mirror or a raidz) is the possibility to replace a failed drive with exactly what was on it. But snapshots creates ”freezed moments in time”, a way to revert a vdev to a previous state. So this is a second way to protect data in addition to backups. (Many backup solutions also makes reverting to a previous state, like a previous version of a file, possible.)
Can you give me a bit more info about snapshots, specifically, I had about 10TB of data in my NAS. A snapshot was made last night. Today I am copying a further 5TB or so of data. If I revert to the snapshot made last night will I end up having to copy the 5TB back into the NAS again. That is, will the NAS be put back exactly how it was when the snapshot was made, removing anything added after?
Yes, it will remove anything has been added after snapshot (in case you are reverting to this snapshot).
You also can just restore previous versions from the snapshots (without reverting). It even works out of the box in windows file explorer. Of course it can be done from the terminal as well.
The great way to set up snapshots – Video: Setting up and using Tiered Snapshots for ZFS Data Recovery.
Do you mean I can revert previous versions of a file without having to run the whole snapshot?
I don’t use Windows I use Linux. But in a file explorer how do I access snapshots. I can’t find a way to access them. Can you explain a little from a Windows perspective and I can work out the Linux way from that.
You log in to your TrueNAS via SSH, become root, then:
cd /mnt/<pool>/<some>/<dataset>/.zfs/snapshot
ls -l
Result if automatic hourly snapshots are active:
total 0
drwxrwxrwx 1 root root 0 Jun 27 11:00 auto-2025-06-27_11-00
drwxrwxrwx 1 root root 0 Jun 27 12:00 auto-2025-06-27_12-00
drwxrwxrwx 1 root root 0 Jun 27 13:00 auto-2025-06-27_13-00
drwxrwxrwx 1 root root 0 Jun 27 14:00 auto-2025-06-27_14-00
drwxrwxrwx 1 root root 0 Jun 27 15:00 auto-2025-06-27_15-00
drwxrwxrwx 1 root root 0 Jun 27 16:00 auto-2025-06-27_16-00
drwxrwxrwx 1 root root 0 Jun 27 17:00 auto-2025-06-27_17-00
drwxrwxrwx 1 root root 0 Jun 27 18:00 auto-2025-06-27_18-00
drwxrwxrwx 1 root root 0 Jun 27 19:00 auto-2025-06-27_19-00
drwxrwxrwx 1 root root 0 Jun 27 20:00 auto-2025-06-27_20-00
drwxrwxrwx 1 root root 0 Jun 27 21:00 auto-2025-06-27_21-00
drwxrwxrwx 1 root root 0 Jun 27 22:00 auto-2025-06-27_22-00
drwxrwxrwx 1 root root 0 Jun 27 23:00 auto-2025-06-27_23-00
drwxrwxrwx 1 root root 0 Jun 28 00:00 auto-2025-06-28_00-00
drwxrwxrwx 1 root root 0 Jun 28 01:00 auto-2025-06-28_01-00
[...]
cd into any of the directories, and copy the files you need to the active dataset path.
Yes.
It has a built-in “previous versions” option.
Datasets contain a hidden .zfs directory. Here you can find your snapshots.
Another way is to create a zfs clone from the snapshot and then copy the files.
On my system it is snapshot – without s in the end.
Correct, typo on my part. Fixed it.
Thank you. I’ll look into that.