I have a machine that's suffering from occasional Machine Check
Processor Exception problems. I am considering disabling that check
with server.exe -mc, and I was hoping for some advice on the matter.
(Good idea, bad idea, what this experiment might tell me).

Here's my situation: Hp Proliant ML350G4, 2Gig Ram, Scsi RAID 5, all
original Hp parts. Netware 6.5Sp4a. HP drivers/bioses and Netware
patches all up to date. Server is file only (runs APC Powerchute and
Backup exec 9.1 remote besides normal NW6.5 stuff).

For months now I have been getting these MCPE problems in little
spurts. I would worry over it with Hp support for some time, update some
drivers, maybe replace a piece of hardware, the errors will keep
happening every few days until they would stop. Weeks later, they come
back. I've been through this process several times now, Hp support
likes to insist it's AIOCom.nlm (even when it's not loaded :-). Some
hard drives have been replaced as well as the Processor. There are two
Ram sticks, and I've swapped out either one to test the Ram, but no good

HP's diagnostics do not show any errors or problems, the internal
component temperatures are fine. From a user's perspective the server
is functioning and performing well. Their have not been any other abends
besides the MCPE (or multiple "context switch" abends that happen
directly after the MCPE).

Is it worthwhile to try to disable MCPE ? Are these abends ever
spurious or "nothing to worry about?" It seems to me they must be
sometimes, or their wouldn't be an option to disable.

Or should I try one more time with HP support, or even open a Novell
call ?

Thanks for any advice.

Jim Helfer
WTW Architects
Pittsburgh PA