I installed 5.5 with a 300GB RAID 1 partition (boot is also on this partition).  It booted up fine the first few times until after I used a Live CD and accessed the array, and it became named /dev/md127 for some reason.

Now whenI boot into CentOS I get a kernel panic and different errors, once I got "invalid superblock", even though the array is fine (it didn't happen again, probably because I was sure to dismount and stop the mdadm array properly).

Here's what I get when booting (the /dev/md127 is the problem I think):

I have tried specifying the root with UUID= with /dev/md0 and /dev/md127 without any different result. I'm really getting frustrated and stuck.  I also wonder if the issue might be that I used a very new 3.1.2 mdadm to access the array on my boot disc, could that have caused some incompatiblity problem with a different and likely older version of mdadm that Centos uses?

Solved

I think that when I accessed the raid array and it became md127, this screwed up all the initrd's or something else.  I don't know how to explain it.

I got the idea to use single user mode and it booted fine.  Then I just made an initrd:

mkinitrd /boot/yournewinitrd `uname -r` and make sure you boot with that initrd, I was able to boot again.  This should serve as a warning for anyone booting with RAID 1.  DON'T ACCESS THE RAID ARRAY WITH A LIVE CD UNLESS YOU USE THE SAME MD DEVICE.  I really think this is a Centos and mdadm bug in one.

centos, raid, issuei, installed, gb, partition, booted, accessed, array, dev, md, wheni, kernel, errors, quot, invalid, superblock, didn, dismount, mdadm, booting, active, mirrors, autorun, resume, label, swap, sda, suspend, tresuming, creating, mounting, filesystem, mount, filesystems, fs, setuproot, directory, fstab, sys, defaults, proc, switching, init, unmounting, unmountin, switchroot, fialed, syncing, specifying, uuid, frustrated, disc, incompatiblity, initrd, user, mode, mkinitrd, yournewinitrd, uname,

Latest Articles