i actually have a truenas scale 24.10.2 for home.
I only have one pool ZFS (a vdev of 3 mirrored HDD of 4TO).
Recently i discovered an error message on a silvering task about a corruption data.
As i understood, the solutions was
restore a backup (i only have data son this pool)
destroy the pool and recreate it
Finally, i destroyed the file corrupted (no gravity) but…
I know that my configuration could be better so according to you, what would be the best solution to avoid this pb ?
I’m not a big fan of a backup on a usb drive
I think about remove one disk on my pool zfs, create a new pool zfs with this hard drive coupled with a new 4 TO drive (new vdev mirrored) and then create a replication of my first pool on the second pool.
A local replication has pro and cons.
This is just my point of view: pratically, you will use 2 disks > consuming energy > consuming theyrself for a bakcup that can off course be usefull in certain case, but not in worst scenarios (eg nas get stealed, fire, water damage ecc).
Actually, i’m using this strategy just for an easy-more backup of the app pool (kinda small) into the main pool.
USB replication, despite, is more tricky and less reliable (as procedure)… but give you the possibility to shut down the disk, and possibly to place it in another dislocated zone (encription is not mandatory but strongly advised there).
After some experiment/test on myself… the best approach i found is having a dedicated machine to do this job… just turn on on need > replication > shut down… but is off course more expensive than put 1 or 2 disks into the current nas
The replication Is most of time the better solution: with a full file system replication you inherit everything from you datasets; also the snapshots are incremental (so except some case, only first transfer will be heavy).
If you don’t plan to be “regular” you should flag save pending snapshot in your task: this will prevent system to delete snapshots not transfered, so you don’t risk having to start the backup from scratch in case snapshots expire (at the cost of keep snapshot more time on the main Nas).
I also raccomend you to use CMR disks the same as you probably already use.
Instead another option if you have SMR disks Is just to use a PC in the local network and transfer file with your favorite program ( i use free fyle Sync for example)
While searching for a solution to my problem with replicating data to a backup disk, I found your comment and would like to ask for advice.
I also want to solve offsite and offline backup to a (temporarily) connected USB drive. However, if I want to disconnect the USB drive and replace it with another one (multiple versions, longer retention, I don’t want to rely on only 1 USB drive), I can only do it at the cost of exporting the backup pool on the USB drive. But by doing so, my replication pool settings from the main pool to the backup USB pool will be lost. Is there any way to have the backup pool on USB drives (same type/capacity) and be able to alternate these drives without having to keep creating replication jobs after each swap?
If i well understand the situation, the trick Is to not deleting nor data (off course ) and configuration when you export your backup pool.
The task will persist despite the pool Is exported, and should be usabile again when the pool will be imported for the new backup cycle.
What I never tested Is if you have 2 different disks, with a backup pool named the same… if you can share the same task… I think should be possible teorycally but the cons arguable
) and configuration when you export your backup pool.