Hard reset after 5 minutes of operation

Hi - new here, but experienced with hw & sw diagnosing, but this one has me stumped:

I built up a RAID system(*), loaded Win2k AdvSvr on it, downloaded and installed all the relevant patches, ran SiSoftSandra to burn it in, and then put it aside for a few weeks until I had time to reconfigure the rack it was going into.

When I installed it, it booted fine into Windows (gave me a low battery warning in the CMOS, but otherwise fine), ran Windows for about 5 minutes, then did a clean, hard reset all by itself, back to POST. This process will now repeat itself endlessly -- the system boots and runs without issue for 5 mins, and then just stops. There is nothing in the Windows event log, no messages in the BIOS log, nothing.

I have swapped the DRAM, swapped the CPUs, swapped the RAID card, and the problem stays the same. The chassis has dual redundant power supplies, so I don't think that this is the problem -- it is hard to believe that they would both go bad at the same time.

If the system failed to boot, or had some kind of disk error before crashing, or acted differently across the various hardware swaps, I would know where to go, but I am totally stumped and getting desperate!

Swapping out the motherboard would be a pain in the neck. Does anyone have any good ideas on how to diagnose this, or should I just assume it's a bad mobo?

-- gnet



(*) Supermicro 370DE6 motherboard, 2x P3 866 cpus, various PC133 ECC DIMM memory, Mylex ExtremeRAID controller, Intel Gigabit ethernet, generic VGA card, 12xMaxtor 160Gb disks on IDE-to-SCSI 3-way split backplane

 

kinda silly question for a setup like this but is it overheating? im assuming not, and you seem to have covered all the bases i know of so im stumped as well...:confused

welcome to the site

 

thanks, and no question is silly in a situation like this, but the BIOS "system health" readings for both CPUs are about 41 deg C -- about normal for this setup, I think.

Also, in my experience, when a system overheats it doesn't boot successfully immediately afterwards.

What I'm thinking is that the mobo has on-board SCSI, on-board enet, etc. I'm now disabling each of those to see if one of them is pulling some reset line, but if that is it I'll probably have to pitch the motherboard, and its a PITA to swap these in this RAID case -- much disassembly req'd.

 

41C seems a bit high for the BIOS, no? there is basically zero load on the processor in the BIOS. IMO it could easily overheat underload with BIOS tehmps like that. However, i know whatcha mean about not booting correctly after a crash caused by overheating. What temps is Sandra reporting in Windows?

 

Sandra reports the same. However, I have an identical system (except working), running under load, and it reports 50-53 deg C -- haven't had a problem with it. Also, it runs memtest86.exe (from a standalone boot disk) for hours without stopping.

 

What size power supply are you using?
Try only plugging in the boot disk (ONLY) to see if it still reboots. If it does, power supply should be ok. If not, power supply can't keep up with the hardware.

 

i doubt its the PSU(s) but who knows, stranger things have happened

 

Does the reset occur with windows running in safe mode or command line mode?
Does the reset occur with windows not running at all from the hard drive?
Does the reset ocur with another OS running?

 

Just so everyone knows, I think (hope!) I have solved this. I jumper-disabled the on-board LAN, and the problem seems to have gone away. First time for that solution, in my experience. I don't know if something went wacky with WOL (Wake-on-LAN) or what. But if the system burns-in with a PCI LAN card, then I've got to get it in service, so no more experimenting.