I had one debian 6 system with two harddisks on raid1 set up via mdadm. built on simple partitions (/dev/sdb3, /dev/sdc3). at least on the old system, everything is tied up via UUIDs.
then i inserted another disk on the system and installed ubuntu 12LTS there. everything was working fine, except the raid. i installed mdadm (didn't add any setting for mdadm or fstab) and rebooted.
after that reboot i couldn't get anything on the screen because of the splash thing (tab, alt+f1, esc, nothing gave me any terminal output... remembering why i hate ubuntu) and it hang there for some time (i couldn't hear anything going on on my disks which are quite noisy), i gave up waiting and rebooted.
i rebooted back into the old debian install, and mdadm informed me that one of the raid disks were absent. tried to --re-add and it complained so i just used --add to add the old disk back. After some time syncing, everything was back to normal.
So, the question is, what am i doing wrong on the new install? why did it even messed up with my raid if i haven't add that information anywhere yet? how do i prevent the automatic behaviour that is screwing things up?
sorry for the vague post, but i'm afraid to boot it now to get more info. i can probably poke at the ubuntu install via the debian one if anyone needs more info such as versions and contents of default conf files.
tl;dr: mdadm raid1 on one box. install new distro (outside the raid) how to make it ignore the raid until i can boot and edit it myself?