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