Intermittent System Crash Problem

[SirWorksalot]

I have a server that intermittently crashes for about 24 hours, and then stops. After it stops, it works fine for a variable period of time (sometimes as short as a day, another time it worked for several months) and then begins the cycle again.

I am running Windows 2000 Server SP3 fully patched. I have TrendMicro ServerProtect installed on this machine, so I'm fairly sure it's not a virus.

Concerning hardware:

Case: Antec 4U Rackmount (4U26ATX450)

PSU: Antec ATX 450W SmartPower (SL450)
Motherboard: Tyan Tiger MP (S2460)

Processor: One AMD Athlon MP 2000+ - Tyan Approved

RAM: 1GB DDR266 Registered ECC RAM (2 Kingston KVR266X72RC25/512 Sticks - Tyan Approved)

RAID: Promise SuperTrak SX6000

NIC: Intel PRO/1000 MT Dual Gigabit NIC

Hard Drives: 3xWD800JB (RAID 5); 2xWD400JB (RAID 1); 1xWD400JB (Hot Spare)

Video: ATI XPERT 128

CD-ROM: HP DVD-ROM Drive

Floppy: Sony Floppy Drive

There are no indications of the problem until it happens. When the lockup occurs, it is a complete lockup - the LEDs on the back of the RAID controller freeze and the computer doesn't respond at all.

There are no BSODs, DrWatson errors, Event Log errors, nothing. The only message I get is that "The system shutdown at xx:xx:xx was unexpected. The data is the error code." Useful.

The machine just crashed again as I was writing this. It was up for a total of maybe 10 minutes. It's not an overheating problem, the machine is cooled very well (the case has 4 101mm fans in it, and I have a massive PC Power & Cooling TurboCool on the PSU.)

Please help!

 

[SirWorksalot]

I should also mention that I've tried switching processor and RAM slots, no luck.

 

[etarhan]

I'd check the raid controller, particularly the onboard battery.

When the machine freezes, what do you do to make it come out of the freeze? Is there any one action or a set of actions which *looks like* helpful?

 

[SirWorksalot]

Yeah.. I have to reboot about 50 thousand times to get the machine to POST.

Last night, I ran Memtest for a while, came up with nothing, but when I rebooted into windows the machine is working fine now..

 

[edemiere]

Call and exorcist! lol

Cheers!

 

[SirWorksalot]

Hehe..

I know these boards (S2460) have some problems in the boot logic that cause you to have to reboot it a lot, but I'll be damned if I can figure out why it keeps crashing like this.... the frequency is really odd too...

it seems almost cyclical, like it's every 2-4 weeks it happens again.

 

[rburke]

Have you tried process of elimination on the RAM yet? Most of the time I find that restarts are caused by RAM gone bad. If it isn't RAM then it will probably be a Power Supply issue. I had a S2460 and I think one of my DIMM slots was bad because I had to RMA RAM (haha) once every 6 months or so. Not sure if other boards have that problem. Also, I had to replace the PSU on that system. Those would be my thoughts on it. Let me know if this helps.

Burke

 

[SirWorksalot]

Yea I ran MemTest86 which is an in-depth memory tester, turned up no problems. I was thinking it might be a PSU problem, but it's a Antec 450W..... also, PSU's generally arn't on the fence - it either works or it doesn't, in my experience. How would this make it cycle so?

 

[SirWorksalot]

What kind of PSU did you change to?

Also, can anybody make a reccomendation for a replacement board?

 

[rockerfellerz]

Just a thought, but random freeze ups like you describe can be caused by the board shorting out to the case. One system drove me crazy with freeze ups; I did everything I could think of including installing a different OS but it still froze. Then I decided to take everthing out of the case and rebuild it. noticed while doing this that a couple screws were loose on the motherboard. I made sure that everthing was tight upon reassembly and the freeze ups stopped completely.
Just a small thing to look at before buying a new board.

If you're going through Hell...keep going... (Winston Churchill)
RocKeRFelLerZ

 

[SirWorksalot]

Hmm... I'll check the short idea... anyone else have any ideas?

 

[SirWorksalot]

Well, I took all the components out of the case, and it seems that while I ordered Kensington RAM, somehow I ended up with Powmem *boggle*.

I am 99.99999% sure that this is my problem.

Guess this teaches me to double check serial and part numbers when I get equipment... although, the RAM did arrive in Kensington anti-static casing... *boggle again*

 

[SirWorksalot]

Got the new RAM, machine is still doing it....

Anybody out there?

 

[mainegeek]

I would suspect the Power Supply ,bad capasitors on the mother board or both.

 

[Hainley]

I agree with the PSU diagnosis. And yes, they can give you issues like you are experiencing. I've experienced it.