[evla-sw-discuss] Evla-ngas-1 (good'ish news)

[James Robnett]

    I had Jeff bring that box down late yesterday so
I could look at it this morning and get the motherboard
swapped out.

    It looks like we had two problems, one I understand
and one I don't.

1) The CPU/heat sink were removed in a less than safe
way and it was assumed they were damaged.  It wouldn't
boot after they were re-inserted.

2) Somehow the grub boot loader information was changed,
apparently within the last few months while the system
was up.  This was a latent problem waiting to happen.

    It was problem 2 that was causing it not to boot,
problem 1 never really existed but it sure seemed like
the likely candidate.

    Once I had it here, saw it fail to boot and re-wrote
the boot loader image to disk it boots just fine.

    It seems likely to me if we'd simply rebooted the
computer (say a power outage or something) it would have
hung.  The removal of the CPU had nothing to do with
anything even though it was done in an inelegant fashion.

    The server is on it's way back to the site and should
be up in a few hours.  Once the spare motherboard and
CPU gets here we'll just keep them as spares for this
and the other NGAS servers.

James

James Robnett wrote:
> 
>    Well that went badly.  The cpu that was complaining
> of over temps had pretty much welded to the heat sync.
> 
>    In the process of getting it off it appears that part
> of the cpu socket was damaged.  Swapping in the 2nd cpu
> into the first slot results in it posting the BIOS but
> then hanging in the boot sequence.
> 
>    I'm on the phone to the vendor trying to get a
> motherboard and CPU over nighted.
> 
>    Currently evlangas-1 is down.
> 
> james