1

I have recently moved a clients WordPress site to a new server. His old server was pretty outdated. It is the only site on the server.

It's been a few weeks and I randomly get these errors in the php log.

[notice] child pid 29281 exit signal Segmentation fault (11)

I installed gdb and attached the parent process and then did BT when it crashed. I was presented with this output

(gdb) backtrace
#0  0x00007f0d4208f573 in select () from /lib/x86_64-linux-gnu/libc.so.6
#1  0x00007f0d42587db5 in apr_sleep () from /usr/lib/libapr-1.so.0
#2  0x00007f0d42e65279 in ap_wait_or_timeout ()
#3  0x00007f0d42e72906 in ap_mpm_run ()
#4  0x00007f0d42e47832 in main ()

Unfortunately I have no idea what to do with that or what it means.

Any ideas?

masegaloeh
  • 18,236
  • 10
  • 57
  • 106
Melonheadjr44
  • 21
  • 2
  • Update: I just ran aptitude to upgrade the modules and apparently one was broken. It is also updating libc so I'll see if that helps. – Melonheadjr44 Mar 05 '14 at 16:22
  • The PID is psuedo-random and can be ignored. Segmentation fault indicates a broken program or library mismatch. The backtrace is of the parent process noticing the death of a child process. If you can capture the backtrace or error logs of the child process that is crashing, you will have more useful information. – kmarsh Mar 05 '14 at 16:45

1 Answers1

0

I've seen this before, and it was due to mismatched apache and apr packages. Make sure that all your Apache packages are the same version, and that you're not using (for example) mod_php5 that's been compiled for a different version.

Flup
  • 7,978
  • 2
  • 32
  • 43
  • After updating all the packages I still get segmentation faults but now the site stays live which is an improvement. I ran apache2ctl -l and here is a list of modules. core.c mod_log_config.c mod_logio.c mod_version.c prefork.c http_core.c mod_so.c Could any of these be the problem? – Melonheadjr44 Mar 05 '14 at 16:32
  • Nevermind. Just crashed again. – Melonheadjr44 Mar 05 '14 at 16:33