This is a pretty big refactor of the way systems work, but it avoids
bifurcating further. At this commit point I actually used nixos-anywhere
on the staging server and it came out okay, which is encouraging.
This involves renaming the disks because when I rebooted the VM the disk
names changed. I also made the root disk just 50G and put the rest in
/var, as well as formatting and mounting the big rust disk.
I'm not sure if this will work - I'm running it now - but I want to
commit it so I can see what changes I ended up needing after
establishing a baseline reading through the nixos-anywhere getting
started instructions.
