Hello again module developers,

   I'm trying to track down a SegFault in my module.  It's running on a
high-traffic server and Apache is intermittently receiving SegFaults.  As
child processes die, new ones are taking their place but there are still
requests that go unanswered.

   I don't have control of the high-traffic server's environment:
- It's running on a RedHat Linux (kernel 2.6.20) box (Apache 2.0.59)
- I am not allowed to use gdb... if it is even installed (which apparently
is not the case)
- I can't reproduce this with small load servers (makes me think a pool is
running out of allocatable memory)

  My problem is that I am getting core dumps (using the CoreDumpDirectory
directive of the MPM module) but I can't read them with my version of gdb (I
am running 6.3-5).  My specific error is:
"/home/.../core.pid" is not a core dump: File format not recognized

  I tried using readelf to see what I could see (I found out newer Linux
core dumps are in ELF format).  It appears to know how to parse the file,
but hex is useless to me without a human-readable backtrace.  I'm a
programmer, but I can't see "the Matrix" (yet).

  Any suggestions?  Does anyone know a surefire way to get a backtrace from
an Apache (Linux) coredump file? Perhaps other mailing lists that could
help?


Thanks,
Dave
--
David Wortham
Senior Web Applications Developer
Unspam Technologies, Inc.
1901 Prospector Dr. #30
Park City, UT 84060
(435) 513-0672

Reply via email to