Very weird Server crash - Know any good diagnostic software ?

Pinkeyes · 3 Nov 2006 at 09:45

One of our mail servers crashed in a big way this morning.

It was completely locked. On reboot - it hung on the BIOS screen. It took a CMOS reset to get the thing back up and running.

The event logs of the server showed a few error messages, and then, the time and date stamp of all the error messages had gone back to 1/1/2004 - so the BIOS had already reset before the big crash ?

Has anyone seen this sort of thing before ?

I need to diagnose what part of the hardware failed - the Mobo, the CPU or the memory ?

Does anyone know of any up really diagnostic programs you use to test your hardware to this level ?

I know I am asking for software - but I thought the hardware guys would know more about this sort of this...

Raumarik · 3 Nov 2006 at 11:29

Most servers come with some form of diagnostic software from the supplier but it's rarely any use. Aren't there any clues in the event log?

Pinkeyes · 3 Nov 2006 at 11:36

It was a home built server, not a bought one. It has an Intel board though, maybe there is diagnostic software on their site.

The Event viewer showed some weird errors:

1097- Windows cannot find the machine account, The clocks on the client and server machines are skewed

1005 - Unexpected error The clocks on the client and server machines are skewed. ID no: 80090324 Microsoft Exchange System Attendant occurred.

leezer3 · 3 Nov 2006 at 11:50

Sounds like a borked mobo, as the RTC shouldn't have anything to do with the processor or RAM. First thing to do is replace the CMOS battery, but otherwise it sounds like the mobo.

-Leezer-

Biffa · 3 Nov 2006 at 13:33

Probably the CMOS battery is dead.

Pinkeyes · 3 Nov 2006 at 13:34

Biffa said:
Probably the CMOS battery is dead.

Is that likely to die on a live running system ? - surely that would only have effect if it was booted up - it had been running 24/7 for weeks until this happened...

welshtom · 3 Nov 2006 at 13:42

Pinkeyes said:
Is that likely to die on a live running system ? - surely that would only have effect if it was booted up - it had been running 24/7 for weeks until this happened...

i dont think its the battery, they only come into effect when the machine is off to retain bios settings.

it would be stupid if they controlled the bios memory while it was on

Tom