Backup solution. Good enough?

Currently running TrueNAS SCALE for personal homeusage.

My system:
ASUS WS C264 Pro
1 x Samsung 980 250GB (bootdrive)
1 x Samsung 870 EVO SSD - 1TB (VM & Apps)
Pool 1: 2 x WD 4 TB in mirror (network storage)
Pool 2: 2 x WD 4 TB in mirror (Backing up whatever is on Pool 1, replication)

Pool 1 is taking a snapshot every day at midnight.
At the end of the week (sunday) Pool 1 gets replicated onto Pool 2.

I know this might not be the optimal way. But is it good enough? Replication should probably be sent to a seperate system… But then again, cost, space etc.

Any major flaws I have overlooked? Anything I can improve with the current setup?



This is more or less how I currently run as well. One system, two pools. Hardware ready to replace the primary system should some component fail. The only piece you are missing is an offsite backup. I use the built-in cloud-sync to backup to Storj for that offsite piece, in case the house burns down, gets flooded, etc.

1 Like

You answered your own question. You could probably just have one pool with 2 mirrored vdevs and achieve the same or even better, if you don’t need mirror performance a 4 wide RAIDZ2, this will yield the same space efficiency.

Bottom line is: you only have 1 copy. Period.

At least get another machine up, even better place it remotely.

Add cloud backups to your routine.

Have a look here:

Edit: for me it’s:

Most important stuff with backblaze (cloud)
2 local machines, one of which will move completely off-site soon. Just haven’t had the time to further look into my thread where I ask.on advice on that.

And a USB harddrive that’s currently NTFS but I will look into formatting it to zfs and store the occasional offline copy there.


@kris How would you test your backups in this kind of system in the case of recovery? I have a similar setup and I do for the time being manual hotswap offline/offsite backups of my main pool to 2 seperate HDDs on a caddie because I am not able to build another TrueNAS machine. Planning to backup to the cloud in the future.

You would do a full restore of your dataset to the same hardware ideally.

And if not, can you access said files on those external sources and if so, run some checksums against the files, open them et cetera.