Problem with requests...

Hi all, I have a pretty big NFS server (2 Terabytes), running redhat linux with custom 2.4.20 kernel with XFS fs.

In the last 48 hours I have started getting the following error:

"kernel BUG at ll_rw_blk.c:1192!"

Jun  4 03:49:38 catchusers2 kernel:  kernel BUG at ll_rw_blk.c:1192!
Jun  4 03:49:38 catchusers2 kernel: invalid operand: 0000
Jun  4 03:49:38 catchusers2 kernel: CPU:    0
Jun  4 03:49:38 catchusers2 kernel: EIP:    0010:[<c024b027>]    Not tainted
Jun  4 03:49:38 catchusers2 kernel: EFLAGS: 00010206
Jun  4 03:49:38 catchusers2 kernel: eax: 00000410   ebx: 00000008   ecx: cefd0d20   edx: cefd0d20
Jun  4 03:49:38 catchusers2 kernel: esi: 00000001   edi: cefd0d20   ebp: c1b6c1b0   esp: e02d3cbc
Jun  4 03:49:38 catchusers2 kernel: ds: 0018   es: 0018   ss: 0018
Jun  4 03:49:38 catchusers2 kernel: Process nfsd (pid: 1368, stackpage=e02d3000)
Jun  4 03:49:38 catchusers2 kernel: Stack: cefd0d20 00000001 cefd0d20 00000001 c0145a10 00000001 cefd0d20 

I get two of three of these and then NFS locks up completely and my server shuts down. This sounds like some kind of request queue filling up from what I can gather on the net. I am desperate now! Does anyone know what causes this exactly and if there is any remedy. I would most appreciative if anyone can offer any suggestions. I know my NFS server will go down a couple of more times today until I fix it.

Thanks,
Matt