Proliant 3000 lockup problems

[bopritchard]

i've got a proliant 3000 that runs nt 4.0...the only thing we run on this is a very simplistic website...several months ago i installed an additional processor and memory...since then...every now and again the server just locks up...keyboard, mouse, everything...no err messages, no log files of any problems...what should i be looking at/for

 

[itsp1965]

Bo, did you try running diagnostics on the server (not within NT), it could be the memory or the CPU that you have recently that may be at fault.

 

[bopritchard]

remind me...how do i run diagnostics?

 

[TheLad]

Diagnostics can be run from SmartStart (F10 prompt on bootup). It may be worth removing the 2nd processor to see if the problem rectifies itself.

It may also be worth checking to see whether the processor heatsink is attached correctly as the lock-up problem can be caused by an oveheating processor. Also check the temperature of where the server is housed.

-----------------------------------------------------
"It's true, its damn true!"
-----------------------------------------------------

 

[bopritchard]

what is the heatsink and how do i tell whether it's attached correctly...as far as the temperature...the room it's in is 68 degrees

 

[Walter349]

It sounds like your processors are not within a single stepping of each other, which can cause this sort of intermittant lock up.

 

[bopritchard]

what exactly does "single stepping" mean

 

[ntinlin]

Stepping is the term for a different revision of the microcode in the CPU.
e.g. PII 450 stepping 2 or PII450 stepping 3.

'Usually' you can get away with the stepping being one or two levels different although Windows will often complain about it. If they are very different then a lot of programs like SQL will be very flaky. And as Walter says you can get hangs and blue screens.

You also usually need to have the CPU with the earliest revision in slot 1.
The stepping level is part of the reason Compaq's extra CPU's are so expensive, they guarantee to provide the correct stepping level or they will swap it over.

It could also be down to the memory so I would suggest the following:-

1. Remove the 2nd CPU and run for a while to see if it crashes
2. Remove extra RAM for a while
3. Replace each one and see what happens

I'd also suggest installing the Compaq insight agents if you haven't already since they can give you more info.

Neill

 

[Walter349]

Sorry I didn't get back to you earlier, but I have been busy with 'Sassar'.

The above describes stepping very well. If you are having RAM problems, it is normally indicated by a lot of HDD corruptions and if excessive it can trash the HDD, This is a by product of some of the older ROMPAQS and compaq memory timings and a write through the corrupt memory causes a corrupt write to the HDD, and bang! dead HDD

Do make sure that your running the most up to date rompaq. especially when planning to install insight management agents, as the newer versions won't work with older ROMPAQs reliablely and the older versions of Insight management agents are not reliable themselves. Memory leaks, mis-reporting etc.

Its a shame that HP canned the old 'Red server' team, a lot of the details were not documented, so the knowledge has been lost.