Stumbling through getting a proper backup regime in place. I have an unraid system running a proper array, and am trying to setup backups for two separate machines (one windows one debian). I’ve successfully setup a file share, and have duplicati running. Are there disadvantages to just setting the network folder as the destination for the backup? It seems a little hamfisted (and the data rates are terrible).
It seems like there’s probably a better way to do this…
Validating restore is one of the harder things. There are some people who restore means it works on the exact same hardware - commonly hardware vendors will change internal details without changing model numbers so this plan often fails when it turns out the computer restore to is slightly different and so you don’t have the right driver on your backup (My dad used to repair computers for people who did this - he often charged thousands of dollars to fix a 10-30 year old machine that a new PC could run circles around. IIRC most of the machines were HPUX and not PCs so the cost of porting to a new machine was ). Most people figure they will be replacing the computer with something newer and so instead install the latest OS and then restore just files, now you face the question of do you restore applications or not?
I would test restore on a cheap raspberry-pi, assuming the backup is from x86, or on a cheap ebay x86 if the backup is from ARM. For most self hosted people this is really what they want to know can be done because they don’t know what the replacement hardware will be. The goal here isn’t a usable system, it is enough that you can show your files still exist.