Apache httpd instability?

[Georgios Petasis]

Hi all,

Every a few minutes or so, I see logs similar to the following one, in 
apache's httpd error log:

[Wed Mar 13 15:06:01.223104 2013] [mpm_prefork:notice] [pid 3850] 
AH00170: caught SIGWINCH, shutting down gracefully
[Wed Mar 13 15:06:01.747820 2013] [core:notice] [pid 4002] SELinux 
policy enabled; httpd running as context system_u:system_r:httpd_t:s0
[Wed Mar 13 15:06:01.749783 2013] [suexec:notice] [pid 4002] AH01232: 
suEXEC mechanism enabled (wrapper: /usr/sbin/suexec)
[Wed Mar 13 15:06:02.000501 2013] [ssl:notice] [pid 4002] AH01886: SSL 
FIPS mode disabled
[Wed Mar 13 15:06:02.175544 2013] [auth_digest:notice] [pid 4002] 
AH01757: generating secret for digest authentication ...
[Wed Mar 13 15:06:03.000961 2013] [lbmethod_heartbeat:notice] [pid 4002] 
AH02282: No slotmem from mod_heartmonitor
[Wed Mar 13 15:06:03.001154 2013] [ssl:notice] [pid 4002] AH01886: SSL 
FIPS mode disabled
PHP Warning:  Module 'zip' already loaded in Unknown on line 0
[Wed Mar 13 15:06:03.194039 2013] [mpm_prefork:notice] [pid 4002] 
AH00163: Apache/2.4.3 (Fedora) OpenSSL/1.0.1e-fips PHP/5.4.12 
Rivet/2.1.0 configured -- resuming normal operations
[Wed Mar 13 15:06:03.194063 2013] [core:notice] [pid 4002] AH00094: 
Command line: '/usr/sbin/httpd -D FOREGROUND'

Something is causing these "restarts". How can I debug this?

George

[Jared K. Smith]

On Wed, Mar 13, 2013 at 9:08 AM, Georgios Petasis  wrote:
> Something is causing these "restarts". How can I debug this?

Having had my fair share of Apache problems lately, here's my method
for tracking down these restarts:

1) Create a directory for Apache to store its core files in.  For me,
I created /var/www/segfault

2) Add the following line to your apache configuration, and restart apache:

CoreDumpDirectory /var/www/segfault

3) Make sure your core limit is set to allow core files to be created.
 Check the output of "ulimit -a", and change it with something like
"ulimit -c unlimited".  You probably need to restart Apache after
changing this as well.

4) Wait for the restarts to happen, and see if Apache creates core
files in your directory.  They're typically named something like
core.12345, where 12345 is the process ID of the program that crashed.
 If you get a core file, use gdb to analyze it and get a backtrace.
Type "gdb /usr/sbin/httpd core.12345", and then once in gdb type "bt"
to get the backtrace.

5) Open a bug report in bugzilla against the httpd package with the
backtrace, and hope the package maintainer can help you track down the
issue.

--
Jared Smith