Kernel panic Crash

baetmaen · 05-12-2005, 11:36 AM

Hi,

is it possible to say why the system is crashing (from the information displayed on the crash screen) ?

In my case, here an example:

Code:

Unabel to handle kernel paging request at virtual address 1e4ba870
  printing eip:
c011f8d3
Oops: 0002
CPU: 0
EIP: 0010:[<c011f8d3>]    Not tainted
EFLAGS: 00010017
eax: 00000042      ebx: 00000202       .......
esi: .......

I got a bad picture of it here: http://baetmaen.de/tmp/PICT0003.JPG

I know on my own that the process that was active or that caused the crash was sshd (It's the same when I try to copy a file over the Samba Server on it, just the smb process is then listed in the crash screen)

This PC is my VDR Box and I have to copy my Star Wars Ep. 1+2 and Clone Wars Recordings to my Windowes PC, because there is the DVD burner

But everytime after copying around 600MB the Linux System is crashing.

Could this have something to do with bad memory chips or the VIA Chipset (I hate Via, but somebody donated the Board and the CPU with Memory to me. He was running WinXP on it without Problems.

By the way, the dist is Debian Sarge.

Best Regards

corfe · 05-14-2005, 09:58 AM

Does the same error happen everytime, or is it sometimes different?

You may want to try apt-get installing "memtest86" or "memtest86+", it is added to your boot menu and can do a thorough testing of your memory (I'd let it run overnight), to tell you if there are any errors.

If you have no memory errors and the error is the same every time, try making sure you have the latest kernel, (2.6.11.9), and if that still doesn't work, I'd mail your error to the linux kernel mailing list (whose address I'm honestly not sure about).

baetmaen · 05-20-2005, 06:56 AM

Quote:

Originally posted by corfe
You may want to try apt-get installing "memtest86" or "memtest86+", it is added to your boot menu and can do a thorough testing of your memory (I'd let it run overnight), to tell you if there are any errors.

I'll try that this night.

There was also something interesting while I installed Memtest86+:

There where several adviced packages, and there was the following package, I just installed it: kernel-patch-badram

I will see

corfe · 05-23-2005, 12:40 PM

Just a heads up, I would check the documentation on the kernel-patch-badram. I believe it just gives you the patch, but you have to get kernel sources yourself, apply the patch, and compile the kernel yourself. Then you provide parameters to the kernel on bootup, regarding which bits are bad. You can figure out which bits are bad using memtest.

In any case, good luck!

baetmaen · 05-23-2005, 04:17 PM

Quote:

Originally posted by corfe
You can figure out which bits are bad using memtest.

OK. My biggest problem seems to be that during 20 Memtest passes, no error accoured ;(