[tlug] How to proceed with RAID drive failure

["Stephen J. Turnbull" <stephen@example.com>]

Martin G writes:

 > I'm a little confused because I thought if one drive failed, then wouldn't
 > the other drive simply carry on.

If you have hardware RAID (ie, in disk controller firmware), yes, the
other drive will simply carry on.  If you have software RAID (ie,
based on LVM2 or similar kernel driver), you need to boot before you
get the RAID features.

 > So if the machine isn't booting at all, does this mean a more
 > global failure of some kind, or that both drives have failed
 > because whatever problem was copied between them?

AFAIK, RAID will not manipulate your partition tables, which means the
bootable flag on the partition is not mirrored by RAID.  It seems
quite possible to me that Ubuntu's installer didn't bother to do this
either (it's not obvious).

Possible solution:

First, try changing the boot disk order in the BIOS.  If both disks
are already there, you'll have to try Plan B.

Second, boot using the Ubuntu CD.  Using the partition manager of your
choice, check for bootable flags on the / (or /boot, if appropriate)
partition of *both* disks.  If only one is set bootable, do the other.

Now try rebooting.  If that works, use your RAID tool to diagnose the
problem.  If that doesn't work, try physically swapping the disks (so
what was hda becomes hdb and vice versa), or changing the boot order
in the BIOS.

If that doesn't work, hopefully somebody else has a good idea.

Ciao,