I’m thinking of backing all of my family’s digital assets up. It includes less than 4 TB of information. Most are redundant video files that are in old encodings or not encoded at all and there are a lot of duplicate images and old documents. I’m gonna clean this stuff up with a bash script and some good old manual review, but first I need to do some pre-planning.
- What’s the cheapest and most flexible NAS I can make from eBay or local? What kind of processors and what motherboard features?
- What separate guides should I follow to source the drives? What RAID?
- What backup style should I follow? How many cold copies? How do I even handle the event of a fire?
I intend to do some of this research on my own since no one answer is fully representative but am appreciative of any leads.
4tb is nothing in the days of 20tb hdds
If you want a NAS you don’t need a beefy pc. I started mine with an ewaste office pc (like literally a $40 2013 pc from a doctors office). I now have a much beefier setup that is still all recycled hardware, 10th gen intel build, but mainly because I wanted to add wayyyyyy more drives and do stuff with vms and local llms and such. otherwise you really don’t need “power” and having it is actually detrimental as it will cost you more money in electricity to run (also the environment)
Keep in mind a NAS/raid is NOT BACKUP. It is far more resilient to keep your data on a raid array of 3-4 4tb drives than simply keeping it on a second external drive that might just die at any moment. But all it takes is one day where a drive in the array dies and then a parity drive dies during rebuild and then poof, your data is gone.
You could do raid 1 with like 3 mirror disks but this is excessive and you could still get got by various things: bitrot, house burning down, power surge, controller failure, you fuck things up, etc
A proper backup solution is necessary if the data is critical.
To emphasise your point, I recently had a scare with corruption in my ZFS pool causing me to have to transfer all my data off and start the pool fresh. I’ve since begun viewing any RAID pool as a single drive, which helps see the situation more accurately. Instead of thinking “I have 6 drives with 2 parity, I can withstand 2 failures!” think “I have one pool, I can withstand 1 failure.” Because the moment anything about that pool breaks you are shit outta luck. Prevention against single drive failures is only one part of the puzzle.
I had a time where the dreaded situation people warn about happen: a parity drive failed during rebuild to replace a drive that had failed. Rebuild operations are stressful on drives. And if you’re like me in a home setup you probably aren’t swapping out drives preemptively, you’re waiting until smart errors occur because you’re not made of money and the data isn’t truly critical
There are many ways to pursue proper backups (to op), backblaze works for me (though there are some that are more privacy focused, not that backblaze is terrible on that front). I also have tape backup but this is overkill and $$$, really only worthwhile to pursue if you have a gigantic nas and also are good at tinkering (can refurbish cheap drives sold as broken).