using raid1 with mdadm to "restore" system to previous state
I have a few systems that will soon be going into co-location. The distro is Debian and we're using etch amd64 because each machine will require 4GB+ of memory.
Each machine has two drives in raid1 using mdadm. I would like to know if the following scenario is possible:
a) I have update XYZ to apply to the machines, I've tested it fine on the test server
b) I remove one of the discs from the mirror
c) I apply the update
d) The update screws up the machine
e) I restore the machine to the state it was in on the discs that were removed from the array.
I realise there are a few ways to do this if I have physical access to the machine but, I'd like to find out and test if this is possible before the machines go to the data centre (which is over an hours drive away).
Thanks in advance,