Fatal trap 12: page fault while in kernel mode

[Kai]

On Thu, Apr 19, 2007 at 04:14:23PM +0200, Christian Walther wrote:
> On 19/04/07, Kai <kai at xs4all.nl> wrote:
> >On Wed, Apr 11, 2007 at 12:53:32PM +0200, Kai wrote:
> >>
> >> Hello all,
> >>
> >> We're running into regular panics on our webserver after upgrading
> >> from 4.x to 6.2-stable:
> >
> >Hi Again,
> >
> >The panics keep happening, so I'm trying alternate kernel setups. This is a
> >trace of a panic on a default SMP kernel with debugging symbols.
> >
> >I'm At a loss on how to progress at this point, perhaps someone can help me
> >please?
> [snip]
> >
> >Fatal trap 12: page fault while in kernel mode
> >cpuid = 0; apic id = 00
> >fault virtual address   = 0x34
> >fault code              = supervisor read, page not present
> >instruction pointer     = 0x20:0xc06bdefa
> >stack pointer           = 0x28:0xeb9cf938
> >frame pointer           = 0x28:0xeb9cf944
> >code segment            = base 0x0, limit 0xfffff, type 0x1b
> >      = DPL 0, pres 1, def32 1, gran 1
> >processor eflags        = interrupt enabled, resume, IOPL = 0
> >current process         = 13577 (perl5.8.8)
> >trap number             = 12
> >panic: page fault
> 
> Is this perl derived from ports? And if so, did you rebuild it after you
> upgraded to 6.2? Or is maybe FreeBSD 4.x binary compatibility missing from
> your kernel?

Hi Chris,

Thanks for your reply; The upgrade i'm talking about is just a term
describing that we switched from FreeBSD 4.10 to 6.2. Its new hardware; its
hardware on which FreeBSD 4.10 will not run.
So in effect its not an upgrade, though the symptoms did not show on
apache-1.3.37 + nfsmounted homepages under FreeBSD 4.10.

If perl would be the problem, the OS shouldn't panic IMHO. Perl in this case
is writing a fairly large guestbook file (eg. 2 Mb), and does this through
perls own:
	open(BOOK, "+<$file") or die;

This $file is located on an NFS mounted filesystem. It'll get read and
written.

The NFS filesystem is mounted with "rw,nosuid,intr,bg,resvport,nfsv3". I
have tried mounting without intr, but panics keep happening. The NFS server
is a Netapp filer.

This is a production environment, so I can't go updating to the latest
current. 

Kai
-- 
This was an above the .signature production