Odd Hangs

[Doggie52]

Hello everyone!

I turn to you, MSFN, to help me with my very odd problem.

My server computer, running Windows Server 2008 Standard, has been getting weird hangs the past few months. Since it's always in screensaver mode / monitor turned off I can't see what the error looks like, but it seems after a random amount of time the computer basically completely hangs with all hardware still running. Basically you could call it a shutdown but without shutting down the hardware.

This is... uhm... slightly annoying - as it obviously kills all the different programs I am running on the server and forces me to restart it. As I never know when it's happening it most often brings my uptime down by a few %.

Needless to say, it has to be fixed.

The first errors came after I recently had bought new hardware. I experienced some hangs in Linux and after a format they passed on to Windows. At first I checked the System Event Log and there I found a "Disk32" error with my harddrive, saying a failure was imminent. Those errors popped before the crashes happened, so I figured it caused the hangs and immediately bought a new harddrive, a Western Digital 320GB. But after a few days I started noticing the same errors happening with that harddrive, and so I had to start my troubleshooting all over again.

I did some RAM tests to make sure the RAM wasn't corrupt or faulty - but the tests all came out clean. My motherboard is fairly new, my processor likewise. Is there anything other than the HDD and RAM that can be broken? Is there something in the BIOS that can create these crashes? Anything in the power management?

Here come the specs:

HARDWARE

Processor: Intel Pentium 3.0GhZ

RAM: as of yet only 1GB DDR 400

Mobo: MSI PT8 Neo-V

GPU: an old nVIDIA GeForce 3

HDD: WesternDigital Caviar 320GB

Drives: DVD-drive

Manufacturer: Myself

SOFTWARE

OS: Windows Server 2008 Standard

Gameservers: SRCDS (Source Dedicated Server), HLDS (Half-Life Dedicated Server), Ascent (WoW), HLDSUpdateTool

Other servers: Ventrilo, RAdmin Server 3.2

Boosters: SRCDS FPS Booster, SRCDS Autorestarter

A big thanks in advance, if there's anything more you need to know just post.

Edited June 5, 2008 by Doggie52

[joe43wv]

It sounds like it could be an issue with your power settings. Try this, try turning off ALL your power saving features and also disable hibernation and wait. If you see that this improves the situation then reapply your power savers one at a time, that way you could pin point the problem.

[Doggie52]

I've never had anything like that turned on. This is a server - letting it hibernate would be useless. I also ran some MemTest86 tests today. It went for 8 passes with zero errors. So it's not the HDD and not the RAM.

Actually I doubt it's got anything to do with software, since it happened with three differents OS's.

[joe43wv]

Since you think it's a hardware issue, try this. Try taking all parts out of your system, except the hard drive. Have the system boot up and wait, then over time add one piece of hardware pack one at a time. If sounds, nic, etc. is onboard disable it all then enable one at a time for a true test.

[Zxian]

Have you checked the power supply of the system? Faulty PSUs can cause strange behaviour such as this.

That's a pretty old motherboard. Did you get it new or second hand?

Have you tried running the server headless and connecting via Remote Desktop?

You've ruled out memory and hard drive errors, so that leaves PSU, motherboard, graphics, and CPU. In my experience, when a CPU dies, it dies hard (unless you're overclocking and it's simply unstable).

[puntoMX]

Have you checked the power supply of the system? Faulty PSUs can cause strange behavior such as this.

That's a pretty old motherboard. Did you get it new or second hand?

My bets too indeed. When reading the first post it must be the PSU although it could be the mobo too. (Let´s presume that the software isn´t the bad one here...)

[Doggie52]

Actually, speaking of the CPU, I found some interesting BIOS driver updates on my mobo's site. As I have a prescott CPU this kind of catched my eye:

Update CPU Microcode for Prescott CPU.

System can't resume from S3 with Prescott CPU.

Can a flashed BIOS solve my problems? I will flash it in the beginning of next week when I get back from my holiday.

[puntoMX]

That would not be a bad idea, however, you told us that it was "okay" before with the BIOS the mobo has now, or I understood it wrong?

[Zxian]

Since the problem seems to have been gradually increasing, I'd guess that a component is failing. Install something like Everest or SpeedFan and see if there are any voltages that seem out of whack (anything more than 5% from rated value is worth looking at).

Otherwise, you'll likely have to dig around to see if you can find a new motherboard. On the other hand - getting such an old board would probably be expensive, and you'd be better off getting a new board and a new CPU to go with.

[Doggie52]

I might not have mentioned it - but the problems started appearing after I bought myself a new motherboard and CPU. I will, as I said, flash the BIOS soon and see if that solves the problem. Who knows, it might be the co-operation between the BIOS and the CPU that's failing, it doesn't have to be something that's broken, right? The motherboard and processor are both new and not second hand. Same goes with pretty much everything inside the case except the PSU.

In the meantime, guys, thanks for your help !

Edited June 6, 2008 by Doggie52

[puntoMX]

I hope you did a clean install too...

[Doggie52]

Oh yes, yes I did indeed. You mean a clean install of the OS, right?

[Zxian]

The motherboard and processor are both new and not second hand. Same goes with pretty much everything inside the case except the PSU.

Haha! My brilliant plan isn't foiled yet! When you get that BIOS update done, check the PSU voltages. If it's an old PSU, it's very likely that it's starting to go (especially if it's some no-name brand).

Lycka till!

[Doggie52]

Haha . I will check the voltages as soon as I get back to the server.

Since you're Canadian... Merci beaucoup ?

[Doggie52]

Sweet... Seems I can't flash the BIOS, because I just can't access the BIOS flasher on the harddrive... Can someone give me a guide on flashing the BIOS WITHOUT a floppy and a bootdisk?

I have the Ultimate Boot CD with several bootdiscs included, yet when it mounts them they can't access my HD and only display the CD and the mounted image. A pain in the *rse, to say the least.

Edited June 9, 2008 by Doggie52