CentOS virt - Jul 2010 - Lockup with (none) login

I had a CentOS 5.5 Xen "standard virtualization" install lockup on 
reboot after an battery backup (apcusbd) orderly shutdown induced by a 
power outage. It may have been sitting with two kernel updates without a 
reboot.

I have to head to the site (with a fractured ankle), but reports 
indicate that it is at

- (none) login:

which only returns back to itself after a user login at console, 
including root.

- the local user says, though the monitor speed was "too fast" that it
is failing to find its mounts OR that the disk reported errors.

It is on a dmraid (I know, please don't flame me).

There is some critical information on the drives that did NOT backup.

I need a list of tools and ideas to have a checklist to try and 
resurrect this machine.

Of course I will go with
- Live CD
- CentOS 5.5 install.
- Hard drives.

I would appreciate any procedural methods to go about this and try to 
resurrect this machine.

On Mon, Jul 05, 2010 at 09:34:25PM -0400, Ben M. wrote:
> I had a CentOS 5.5 Xen "standard virtualization" install lockup
on
> reboot after an battery backup (apcusbd) orderly shutdown induced by a 
> power outage. It may have been sitting with two kernel updates without a 
> reboot.
> 
> I have to head to the site (with a fractured ankle), but reports 
> indicate that it is at
> 
> - (none) login:
> 
> which only returns back to itself after a user login at console, 
> including root.
> 
> - the local user says, though the monitor speed was "too fast"
that it
> is failing to find its mounts OR that the disk reported errors.
> 
> It is on a dmraid (I know, please don't flame me).
> 
> There is some critical information on the drives that did NOT backup.
> 
> I need a list of tools and ideas to have a checklist to try and 
> resurrect this machine.
> 
> Of course I will go with
> - Live CD
> - CentOS 5.5 install.
> - Hard drives.
> 
> I would appreciate any procedural methods to go about this and try to 
> resurrect this machine.
> 
Hmm.. boot log would be good, then it'd be obvious what's wrong.
Try using a serial console to capture the Xen/kernel messages?

http://wiki.xensource.com/xenwiki/XenSerialConsole

Other than that.. did you check redhat bugzilla if there's some dmraid
related regression on 5.5 ? Or maybe some other regression..

-- Pasi