2004-10-09 00:19:23

by Craig, Dave

[permalink] [raw]
Subject: autofs panic, linux 2.4.21-15.EL.c0smp

I came across the following panic today when shutting down a system that
hung any processes attempting to access network mounts.

Has anyone seen something similar and have any suggestions on how to
improve system availability? We're experiencing problems about once a
week.

Thanks,
Dave Craig

Unable to handle kernel NULL pointer dereference at virtual address
00000000
printing eip:
c017c362
*pde = 129b2001
*pte = 00000000
Oops: 0000
loop vfat fat nfs sr_mod ide-cd cdrom dcdipm dcdbas nfsd lockd sunrpc lp
parport autofs ppp_async ppp_generic slhc tg3 ipv6 floppy sg microcode
keybdev moused
CPU: 0
EIP: 0060:[<c017c362>] Tainted: PF
EFLAGS: 00010207

EIP is at invalidate_list [kernel] 0x22 (2.4.21-15.EL.c0smp/i686)
eax: f3273000 ebx: 00000000 ecx: f478b3a0 edx: e1195f50
esi: c5ca0300 edi: 00000000 ebp: c03a9704 esp: e1195f24
ds: 0068 es: 0068 ss: 0068
Process umount (pid: 22290, stackpage=e1195000)
Stack: f478b380 00000001 00000000 00000000 f3273000 e1195f50 f8972410
c017c43d
c03a9704 f3273000 e1195f50 f478b388 f478b388 f3273000 f3273048
f8972440
c0168592 f3273000 c03a8fe8 00000000 e1195f8c 0804def8 bfffb178
c017f99f
Call Trace: [<f8972410>] autofs_fs_type [autofs] 0x0 (0xe1195f3c)
[<c017c43d>] invalidate_inodes [kernel] 0x4d (0xe1195f40)
[<f8972440>] autofs_sops [autofs] 0x0 (0xe1195f60)
[<c0168592>] kill_super [kernel] 0xe2 (0xe1195f64)
[<c017f99f>] sys_umount [kernel] 0x3f (0xe1195f80)
[<c017fa17>] sys_oldumount [kernel] 0x17 (0xe1195fb4)

Code: 8b 3f 39 eb 74 66 8d 73 f8 8b 44 24 24 39 86 ac 00 00 00 75

Kernel panic: Fatal exception



2004-10-09 07:11:33

by Arjan van de Ven

[permalink] [raw]
Subject: Re: autofs panic, linux 2.4.21-15.EL.c0smp

On Sat, 2004-10-09 at 02:18, Craig, Dave wrote:
> loop vfat fat nfs sr_mod ide-cd cdrom dcdipm dcdbas nfsd lockd sunrpc lp
> parport autofs ppp_async ppp_generic slhc tg3 ipv6 floppy sg microcode
> keybdev moused
> CPU: 0
> EIP: 0060:[<c017c362>] Tainted: PF
> EFLAGS: 00010207


you could start by not using binary only kernel modules; next step would
be to run a kernel that some vendor supports and then complain to that
vendor ;)
(I don't know who made the kernel you are running, but it's not Red
Hat)


Attachments:
signature.asc (189.00 B)
This is a digitally signed message part