Apache Problem - Can't Connect to Default Port after 24 hrs

Hi,

My problem is Apache fails to accept connections on port 80 after 18-26 hours.

All other DA services work fine. Anything running on a non standard port is accepting connections - ie Webmin & DA.

Performing a Apache restart allows the connections back again but eventually it stops.

I have done a
cd /usr/local/directadmin/customapache
./build clean
./build update
./build all

and the problem still exists.

Does anyone have any ideas on this or can point me where to start looking for the problem?

Be gentle - Linux skill level - newbie.

DA running on Redhat 9
Apache 1.3.33
DirectAdmin 1.23.4

What exactly shows up in your apache logs? Does anything interesting appear in any other logs at the same time?

Its difficult to get an idea without more information; however I've found in general that its a good idea on any server that sees a decent amount of use (and probably those that do not as well) to restart apache regularly, usually once a day in the wee hours of the morning.

Logs dont seem to show much except the following is failry common.

/var/log/directadmin/error.log

2005:02:09-21:22:33: Didn't find two eols on the header from 211.28.249.187

2005:02:09-21:22:36: Socket write error: fd is connected to a pipe or socket whose reading end is closed. When this happens the writing process will also receive a SIG_PIPE signal. (Thus, the write return value is seen only if the program catches, blocks or ignores this signal.)
2005:02:09-21:22:36: Error reading from 211.28.249.187:
2005:02:09-21:22:36: Socket write error: fd is connected to a pipe or socket whose reading end is closed. When this happens the writing process will also receive a SIG_PIPE signal. (Thus, the write return value is seen only if the program catches, blocks or ignores this signal.)
2005:02:09-21:22:36: Error while sending ./data/skins/default/images/complete_usage.gif
2005:02:09-21:40:53: /usr/local/directadmin/data/sessions didn't exist. Creating it...
2005:02:10-06:01:06: Timeout from from 128.250.15.80 : last flagged: Sockets::goodbye()
2005:02:10-06:01:08: *** Segmentation fault *** Sockets::goodbye() : User: (null) : 128.250.15.80 : / : (null) : (null)
2005:02:10-06:01:08: *** Recursive Segfault!!! *** Sockets::goodbye() : User: (null) : 128.250.15.80 : / : (null) : (null)

/var/log/httpd/error_log

[Fri Feb 11 23:37:18 2005] [notice] child pid 14842 exit signal Segmentation fault (11)
[Sat Feb 12 00:20:53 2005] [warn] module perl_module is already loaded, skipping
[Sat Feb 12 00:20:53 2005] [crit] (98)Address already in use: make_sock: could not bind to port 8090
[Sat Feb 12 16:07:20 2005] [warn] module perl_module is already loaded, skipping
[Sat Feb 12 16:07:23 2005] [notice] Apache/1.3.33 (Unix) mod_ssl/2.8.22 OpenSSL/0.9.7a PHP/4.3.10 mod_perl/1.29 FrontPage/5.0.2.2510 configured -- resuming normal operations
[Sat Feb 12 16:07:23 2005] [notice] suEXEC mechanism enabled (wrapper: /usr/sbin/suexec)
[Sat Feb 12 16:07:23 2005] [notice] Accept mutex: sysvsem (Default: sysvsem)]

The system is not high use. Approx 500 hits day if lucky.

I'm seeing the same thing... if directadmin restarts apache via the panel, it never starts and I get socket errors. If I restart it manually, it's fine.

Does DA restart apache nightly or something? Logrotate's hup command is fine...

I'm running 2.0.53.

This appears related to this:

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=145992

(at least my error message is the same). I confirmed the same result using the script provided to close fd0 and then call service httpd restart.

I updated cron, but i'm wondering if DA has fd0 closed as well? If so, why and how do i fix it?

Funny... If i run directadmin in the foreground without the "d" option, restarting apache works as expected. If it runs as a daemon (via the d switch, spawned via init.d, etc) apache won't start and spams the log with:


I'm not sure my errors are related, actually, to the original poster's... the segfault from DA looks suspicious.