[afnog] HTTPD segmentation fault problem

[Brian Candler]

On Tue, Feb 15, 2005 at 09:20:23PM +0300, Banoba Ezra wrote:
> Hi guys,
> I'm having some trouble with apache+mod_ssl-1.3.33+2.8.22 on a FreeBSD
> 4.10 Stable. After a reboot, I suddenly have my formally working apache
> installation core dumped on start and still failed to trace the cause.

Is this a repeatable problem? Does it dump core in exactly the same place
each time you restart it?

What caused the problem to start happening? Was it after upgrading from
4.10-RELEASE to 4.10-STABLE ?

I learned the hard way, that if you cvsup the source and rebuild the
kernel, you *must* upgrade the remainder of the system (buildworld /
installworld) according to the instructions at
http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html

otherwise really strange things happen. In my case, netstat wasn't working
at all, and this was after upgrading the kernel only a matter of days after
the original release. Doing the buildworld/installworld process fixed it
(follow handbook section 19.4.1 at minimum)

There's more info at the bottom of /usr/src/UPDATING, where it says "To
rebuild everything and install it on the current system"

> Reading symbols from /usr/libexec/ld-elf.so.1...done.
> #0  0x2868e44e in Perl_av_extend ()
> from /usr/local/lib/perl5/5.6.2/mach/CORE/libperl.so
> (gdb) quit

At that point, try typing "bt" and see if you get a useful backtrace.

Are you actually running mod_perl? Otherwise it's hard to see why libperl is
being loaded.

Regards,

Brian.