recovering from a controller upgrade - "no volume groups found"
I'm stuck.
I recently upgraded the motherboard, CPU, and memory on my CentOS 5.2 x86_64 server. On doing so, of course my device info got all wonky and I needed a new initrd... but for the life of me, I can't make this thing work.
My configuration is this:
2x500G drives, each with 3 partitions:
1: 100M (/boot)
2: 8G (swap)
3: 492G (LVM)
the drives are partitioned identically, and each partition is a raid 1 member with the obvious parings to form 3 RAID1 sets. the root filesystem sits on an LV in partition 3 along with /var, /home, etc
Attempting to boot off of the old initrd results in a kernel panic, so I set about the task of creating a new one. I booted from the CentOS 5.2 x86_64 DVD in rescue mode, repaired the broken raids, mounted the filesystems, changed root to use the repaired filesystems... replaced /etc/mdadm.conf with the new UUIDs, etc. I then built a new initrd, but without any special flags. This is where I think I went wrong. Eitehr there, or I'm missing some configuration somewhere.
Just like the old one, the new initrd results in a kernel panic. The last readable errors (man, I wish I could scroll up) indicate that "no volume groups were found". I think I can see a general lack of proper raid initialization, but it's off the screen so fast I can't be sure. If it's true, that explains why it can't find any VGs... but why isn't it building the raids? What should I be doing differently? This is a production server that has now been down for almost 24 hours... and I feel like the idiot who can't fix the stupid computer.
Any help is more than welcomed.
|