What I ended up with
With unRAID, storagemagement is different.
They’re set up with JBOD, but you can stil add parity to it.
Let’s say you have my disks, which is 4 TB and 1 TB now. If I add a parity, it has to be the same size, or larger than the biggest drive, which in my case is 4 TB, but the entire disk will be parity, which means you can’t use it to store data on. But you’ll get 100 % utilization of your disk, subtracted whatever the file system needs, and the amount of parity you have.
Then I’ve bought a new disk at 1 TB that I wanted to add. You add it physicall (obviously) format it to XFS (or the encrypted one if you use that), and then you add it to the array. As simple as that.
You have to turn off the array while you do it though, but it will remind you about it.
The filesystem uses what they call “High Water“, which means it fills up disk 1 first, then the second one, the third etc.
That means unRAID won’t switch back and forth between your disks, and needs to spin them up all the time. That also adds a benefit regarding power usage.
If you’re interested in reading more about High Water, you can do it here.
It also means, if you don’t have a parity drive like me, and your disk dies, you’ll lose all of your data on that disk. But the rest will live on.
I also had a small Kingston A400 at 240 GB that I use for as a cache.
That means, if I download a Linux distribution for example, it will download it to the cache first, and then move it to the array later. So if you’re downloading 10 Linux distributions, even if you have a 1000/1000 internet connection, you won’t bottleneck your SSD, but you’ll max out your internet connection.
In a feature post, I will go over my current unRAID setup, and how I ended there.