Re: Kernel oops with ext3 and software raid

["Stephen C. Tweedie" <sct redhat com>]

Hi,

On Tue, 2003-02-04 at 23:38, Yavor Shahpasov wrote:
> Am using RedHat 7.3 with software raid. it was crashing from time to time
> so i desided to investigate. after running some stress tests on it it died
> with a kernel message (see below). it might not be ext3 related but since
> I was running redhat with ext2 for some time without problems (not on the
> same machine) I thought it might have somthing to do with the fs.

> Any ideas. 

Hardware problem, most likely bad memory:

> Feb  4 23:57:37 mail kernel: Unable to handle kernel paging request at
> virtual address 10000004
> Feb  4 23:57:37 mail kernel: EIP is at get_hash_table [kernel] 0x66
> Feb  4 23:57:37 mail kernel: eax: d3fa0000   ebx: 00000002   ecx: d3fa7818
>   edx: 10000000

That's the single most common oops footprint pointing to bad memory: an
error in get_hash_table, with a single-bit flipped in one of the
registers (%edx=0x10000000) and a paging error trying to reference that
as a struct (virtual address 0x10000004).  Your first port of call here
is memtest86.

Cheers,
 Stephen