Windows 2000 SBS keeps crashing

[tfinch2]

Server specs:

Dual Xeon 2.4 ghz
Supermicro P4DL6 board
4 GB RAM
Adaptec 2940U SCSI Controller
4 Seagate SCSI HDDs (2 36GB and 2 73GB)
Bnchmark SCSI DLT tape drive
2 Intel Pro/1000 NICs
Toshiba DVD drive

So one of my servers for a small domain running 2000 SBS with Exchange and ISA keeps crashing. I have it set up to save a dump and automatically reboot for a system error, but the crash just freezes the system and won't respond until I push the reset button. I get an event ID 6008 in the Event Log saying that that reboot was unexpected, but also event ID 1000 sometimes (not after every crash) saying:

The computer has rebooted from a bugcheck. The bugcheck was: 0x0000000a (0x00000000, 0x00000002, 0x00000001, 0x8044accf). A full dump was not saved.

or

The computer has rebooted from a bugcheck. The bugcheck was: 0x000000d1 (0x00f8003a, 0x000000ff, 0x00000000, 0x00f8003a). A full dump was not saved.

Which is expected since it can't save a dump if it hard freezes after the crash. A bit of searching has me to believe that these two bugchecks are from hardware or driver errors, but there is no way to pinpoint without the memory dump. I've also monitored the server room and hardware temperature and there are no problems there.

Is there any other way I can pinpoint this error or am I going to have to do a bit of guess work?

Thanks

 

[tfinch2]

No ideas?

 

[RebateMonger]

If you are a Microsoft Partner, PSS will give you free support for a "Server Down" condition. Even if you have to pay for it, you may find Microsoft's PSS assitance a bargain. They'll keep working on your problem until it's fixed.

Additionally:
Be sure that ALL BIOSes (motherboard and SCSI controller) are the latest available.
Be sure that all drivers are the latest available.
Make sure that your Virtual Memory is slightly larger than your actual memory size. If not, there'll be no memory dump.
Doublecheck that a "Complete memory dump" is selected in the Advanced Startup and Recovery options of System Properties.

And, yeah, sometimes you just won't get a memory dump.

 

[gwag]

random hard reboots could be caused by bad ram or even a bad block on the HD if you can put it down for a few hours try running memtest or pulling 1/2 the ram if it still crashes pull that and put in the first half.

 

[tfinch2]

Originally posted by: RebateMonger
If you are a Microsoft Partner, PSS will give you free support for a "Server Down" condition. Even if you have to pay for it, you may find Microsoft's PSS assitance a bargain. They'll keep working on your problem until it's fixed.

Additionally:
Be sure that ALL BIOSes (motherboard and SCSI controller) are the latest available.
Be sure that all drivers are the latest available.
Make sure that your Virtual Memory is slightly larger than your actual memory size. If not, there'll be no memory dump.
Doublecheck that a "Complete memory dump" is selected in the Advanced Startup and Recovery options of System Properties.

And, yeah, sometimes you just won't get a memory dump.

Virtual Memory is set to 8 gigs, and Complete Memory Dump is set.

I think this weekend I am going to have to take the server down to update all of the drivers and run Memtest.

 

[RebateMonger]

I had an SBS 2003 server that, out of the blue, started crashing twice a week.
It NEVER left a memory dump. It'd just sit there and do nothing until it was rebooted.

After analysis, I discovered it was crashing the instant that another Server started sending it NTBackup data.

I called Microsoft on this one, since I'd exhausted my own ideas. One thing they were helpful on was to notice that I'd forgotten to update the BIOS on that Server. I thought I had.

In the end, I did two things to fix it. I don't know which one was the solution:
1) Installed the SBS SP1 Service Pack (this needed to be done, but I'd put it off until I could find the solution to the crashes).
2) Updated the system BIOS (it was a couple generations old).

Since all my Servers have ECC memory, it's petty unlikely that there are memory problems. But if you don't use ECC (or even if you do), it certainly would be a good idea to test it.

 

[glimmernet]

Originally posted by: tfinch2
Server specs:

Dual Xeon 2.4 ghz
Supermicro P4DL6 board
4 GB RAM
Adaptec 2940U SCSI Controller
4 Seagate SCSI HDDs (2 36GB and 2 73GB)
Bnchmark SCSI DLT tape drive
2 Intel Pro/1000 NICs
Toshiba DVD drive

I have this exact problem and I haven't found the solution yet. In my searching, I came across this article, though:

http://support.intel.com/support/network/sb/cs-009908.htm

There is a corresponding MS KB article, but I don't have that link handy.

Let me know if you manage to solve your issue. I'm at the end of my rope with this one.

Dave

 

[spikespiegal]

Sounds like hardware to me.

 

[glimmernet]

Originally posted by: spikespiegal
Sounds like hardware to me.

I've replaced everything in the system and, when that didn't work, we had the whole barebones server RMAed. It crashed configuring Windows.

So some driver is whacked out, but I have no idea how to figure out which one it is.

Suggestions?