I have a raid1 install of debian 7 with software mdadm on 2 disks.
What I want here is to split this to 2 separate systems and be able to multiboot them.
I guess while I have the system running from the first drive I could do:
mdadm /dev/md0 --fail /dev/sdb1 --remove /dev/sdb1
This would cause sdb1 to be removed from the array (and won't be resynched even after reboot). Then remove the first drive and do:
mdadm /dev/md0 --fail /dev/sda1 --remove /dev/sda1
on the second while it is booted. Is this enough or I have to zero out raid superblocks as well? I don't want that the replication from any of the 2 systems start up all the sudden killing the other disk.