Strange shutdown problem

[Bozo]

Over the past month I have been having problems with a group of computers shutting down. I have 13 computers all running an Intel D845XXX motherboards with a P4 1.8GHz CPU, and 1 GB Crucial ECC memory. All running Win2K. All these computers have been running 24/7 for ~4 years. For some reason they have (4 of them so far) started to shut off. When they shut off, they can't be restarted until the power cord is removed/reinstalled from the power supply. This happens even when the power supply was replaced with an Antec NEO 380W. Also curious is the fact that the green power LED on the motherboard stays on.
On two of the computers I had to remove/reinstall the CMOS battery to get the computers to boot up.
The only thing that has been done to all these computers in the last month was an upgrade to Symantec AV 10.0 Corp Edition. I've search Symantec and Google for any clues, but have come up empty. Intel will no longer provide tech support for these motherboards; e-mail or phone.
Searching Intels site netted nothing.
All these computers have the latest BIOS up date in them. I am considering re-flashing the BIOS with the current version. Can a BIOS get stupid??
Thoughts and suggestions are appreciated.

Thanks,
Bozo

 

[Pradeep]

My first thought would be to uninstall Sym AV and see if they still do the shutdown thing.

 

[time]

My thought is capacitors, but I assume you've checked for bulges?

Also curious is the fact that the green power LED on the motherboard stays on.

I'm confused - what would you expect it to do?

It sure would be nice to lay the blame at the foot of Symantec; I can't wait to see how this unfolds ...

 

[LiamC]

time is right. ATX power is always on, even when the machine is powered down. The only way to remove the power is pull the plug.

Has there been a BIOS upgrade? I have seen the ACPI HAL get confused with a BIOS upgrade that changes default values. ACPI probably (depends on vintage of machine/BIOS and a couple of other things) controls the shut-down/power up sequence. Can you compare with a machine that works?

I'd agree with Pradeep though and ditch Satantec AV first.

 

[Bozo]

The reason I questioned the LED staying on is that the power supply did not 'crow bar' or 'mouse trap' (depending on your location this is when the power supply shuts down internally to protect itself. Usually requiring the plug to be pulled) But, the power supply needed to be unplugged to get the computer to boot.
I have sent an e-mail to corp headquarters asking about Sym AV. With the thousnds of computers we have Sym AV installed on, maybe someone else is having problems too. And, corperate can leverage their muscle ($$$$) with Symantec to investigate and resolve the issue if there is one.
I did check the capacitors and Intels list of known problem motherboards. No help there and no bad caps that I can see.

I agree that it is probably a Symantec problem. Removing it is a huge no-no and I would need special dispensation from the Vatican, congress, corp Gods, the AV Nazis, and the supream being of your choice to do it.

Thanks for your help!

Bozo :mrgrn:

 

[ddrueding]

I have had issues like this where the culprit was a thermal cutoff either on the mobo or the CPU. I assume all the computers have been cleaned, all fans inspected and possibly HSFs removed and reinstalled?

 

[Sol]

It almost sounds like the systems are going into suspend mode and then crapping themselves (which I've found computers like to do when they go into suspend mode). Maybe SAV monkeyed with the power management configuration. That would explain why you need to pull the plug (can't turn on a computer if it's not turned off).

 

[sechs]

Usually the power LED blinks when the computer is suspended. Of course, something from Intel might be different.

 

[Bozo]

Great suggestions!
All these computers are in an air conditioned envirnment. I have cleaned them up with canned air not too long ago, but I'll do it again.
The power management suggestion is a great one. I didn't think of that as all these computers must run 24/7, and were never setup to shutdown in any way, shape, or form.
I might have to rule out Symantec as much as I'd hate to. Seems the first computer to shutdown did so before SAV 10 was installed. Maybe it knew what was comming

Bozo :mrgrn:

 

[Bozo]

Weird: The power supplies that came with these Antec cases have sleeve bearings. On one of the computers that shut down the power supply fan was seized up. Don't know how long it had been that way but the A/C failed in the room where it was sitting. We have decided to replace the power supplies with Antec Neo 380s
Today I shut down one of the computers that has not given us any problems to replace the power supply. (same hardware/vintage as the others) When I went to restart the computer, nothing happened. Double checked all connections, plugs and switches, nothing. I thought the new power supply was bad, but checking it with an Antec tester said it was fine. Finally, I pulled the motherboard battery, waited a few minutes, reinstalled the same battery, and the damm thing started right up!
I'm beginning to think all these motherboards have BIOS issues.
These motherboards might make it to the 'most hated hardware' thread.

Taking even more Excedrin as I search the yellow pages for a witch doctor...

Bozo :mrgrn:

 

[mangyDOG]

Have you considered that the CMOS batteries are getting flat? If the PCs are 4 years old the batteries are probably getting to the end of there lives. If you pull the battery out for a few minutes and put it back in there would probably be enough residual charge to fire up the bios again. I would be a nice cheap fix for you if you only need to change the batteries over...

cheers,
mangyDOG

 

[Bozo]

Interesting. But I didn't think the batteries came into play as long as the power was on.

Bozo :mrgrn:

 

[timwhit]

They still come into play as far as I know. I would try swapping it out.

 

[Bozo]

I have not been able to come up with a solid answer. The plan is to replace the BIOS batteries, let the BIOS get stupid while doing the battery repacement, and replace the power supplies.

Bozo :joker:

 

[Bozo]

I have narrowed the shutdowns/lockups to two things that are common to the problem computers. The memory and a software program call Cimplicity.
Once the Crucial DDR ECC memory is replaced with standard Crucial DDR memory on the computers running Cimplicity, the shutdowns/lockups stop.
Strange it took almost 5 years for this to show up.
The computers that are not running Cimplicity have not shutdown/locked up even with the ECC memory installed.

Strange.

Bozo :joker: