Random panics on FreeeBSD 6.0

[Micah]

Matthew Navarre wrote:
> 
> On Jun 16, 2006, at 7:17 PM, Micah wrote:
> 
>> Matthew Navarre wrote:
>>> I've got an AMD Sempron machine running FreeBSD 6.0 that's been 
>>> experiencing random panics while trying to build world. In fact it 
>>> just paniced now, with no activity.
>>> The panic message is TPTE at 0xbfc20624 IS ZERO @ VA 08100000
>>> bad pte
>>> This started last night while I was portupgrading ruby and I got 
>>> random apps segfaulting, mostly gcc, so I suspected bad memory. I 
>>> installed new memory today, tried to buildworld. And *BAM* panic: bad 
>>> pte
>>> I'm still guessing that this is a hardware problem, and not software 
>>> but I'm not sure. If anyone can give me a clue I'd appreciate it.
>>> Machine details:
>>> AMD Sempron
>>> ECS K8M800-M2 mainboard
>>> 1 GB Kingston PC-3200.
>>> Thanks,
>>> Matt
>>
>> First, a quick Google of "bad pte" turns up some ideas. Try disabling 
>> or changing APIC and/or ACPI settings.  Make sure your swap partition 
>> is error free and has enough room.  Google a bit more just on the 
>> lists.freebsd.org site for several possibilities.
> 
> Yeah, I was wondering if it might be something in the BIOS settings. 
> I'll google around and see what I find. I don't know if the first panic 
> was a bad pte error since the machine was running headless.
> 
>>
>> For hardware, you can try memtest86+ to check to make sure the new 
>> memory is good. There are other stress tests you can run as well - I 
>> usually use the ultimate boot CD for that stuff. Other possible 
>> problems are faulty or too small power supply; too much heat on CPU, 
>> RAM, or expansion boards; faulty expansion cards and/or components; or 
>> faulty hard drive.
> 
> I kinda wondered if heat might be an issue, since it was kind of tucked 
> away in a spot with bad airflow. I'll try the memtest thing. Is there a 
> way to get the CPU temp in FreeBSD?

As mentioned, mbmon might work, but don't think that CPU is the only 
generator of heat. I had random reboots due to an overheating graphics 
card once. A spot thermometer comes in handy at a time like this.

HTH,
Micah