Hi

I have tried to look up what I can on this and failed, but I had a working server with 2 RAID 1 disks, each with 3 active partitions (not including swap).

After an upgrade, the server (CentOS 6.3) fails to boot with a kernel panic. Using a rescue disk I can see all partitions but only md1 is actually there, md0 and md2 are not although each of the individual partitions exist uncorrupted.

I've tried using mdadm to reassemle but I must be doing something wrong because I cannot make it work.

Simply, how can I get sda1 and sdb1 to be md0 and sda3 and sdb3 to be md2?

I should add that I have checked mdadm.conf and it looks fine (ie it specifies the devices and partitions correctly using the UUIDs)

Scratching my head has resulted in more hair loss!

Thanks

UPDATE:

Using Centos recovery disk, it fails to pick up md0 or md2. If I do:

mdadm -a /dev/md0 /dev/sd[ab]1
and repeat for md2, I can do a cat /proc/mdstat and it shows all are recognised and synced.

Next thought was to do a scan and write to mdadm.conf - so did

mdadm --detail --scan >> /mnt/sysimage/etc/mdadm.conf (having mounted root in sysimage)

All looks good, then reboot and it doesn't even get to the splash screen which it did before!!

So, I am now presuming I've either a) messed up mdadm.conf, b) messed up grub or c) both!

Mdadm.conf purely has the refs to the three raid devices - nothing else, just those three lines.

Tried to chroot to /mnt/sysimage but it won't because /home is encrypted. So, manually mount / and /boot, then chroot and try doing a grub-install. It won't. Tried /dev/md /dev/sda /dev/sdb (individually - all fail)

Don't know what to do next. Seem to make some progress and then get stuck again. What I do know is the disks and data are intact, which has to be good news.

Probably about as far as I can go. I do not want to lose everything I've done (ie building a mail/caching web server) but I can't seem to get this fixed.

In a last ditch attempt before I rebuild from scratch, does anyone have any ideas how I can:

a) using the Centos recovery disk, make the RAID arrays permanent so they are seen on reboot
b) repair grub?