i am getting these error mesgs in the kernel log:

kernel log messages:

>> pid 87678 (httpd), uid 80: exited on signal 6
>> pid 87681 (httpd), uid 80: exited on signal 6
>> pid 87683 (httpd), uid 80: exited on signal 6
>> pid 87682 (httpd), uid 80: exited on signal 6
>> pid 87684 (httpd), uid 80: exited on signal 6
>> pid 87685 (httpd), uid 80: exited on signal 6

any ideas on what they are and how do i stop these errors?
currently running Apache/2.0.53 (FreeBSD 5.3) PHP/5.0.3

When did these start?

If memory serves signal 6 is SIGABRT, which is usually caused by your program running in an illegal state. Do your Apache error logs show anything relating to this?

example entries:
[Thu Mar 31 09:33:37 2005] [notice] child pid 9616 exit signal Abort trap (6)
[Thu Mar 31 09:33:38 2005] [notice] child pid 9617 exit signal Abort trap (6)
[Thu Mar 31 09:33:39 2005] [notice] child pid 9619 exit signal Abort trap (6)
[Thu Mar 31 09:33:40 2005] [notice] child pid 9620 exit signal Abort trap (6)
[Thu Mar 31 09:33:41 2005] [notice] child pid 9618 exit signal Abort trap (6)
[Thu Mar 31 09:33:42 2005] [notice] child pid 9622 exit signal Abort trap (6)
[Thu Mar 31 09:33:43 2005] [notice] child pid 9621 exit signal Abort trap (6)

and

[Sun Mar 27 14:25:50 2005] [notice] SIGHUP received. Attempting to restart

there were more SIGHUP than SIGTERM in the logs.

FYI I do not have custom kernel compiled.
After looking at the logs, I had decided to change the LogLevel to debug. maybe will get more messages to help diagnose this error.

is there anywhere else to look?

The HUP signal is just Apache restarting, and usually is caused by an admin restarting Apache. Are you runnign any PHP or CGI scripts on this server? Most likely one of them is coded poorly and is causing Apache to crash.

The only PHP scripts that has been running/interacting with Apache are Horde/IMP scripts (horde.org). otherwise all the other PHP/perl scipts do not interact with Apache.

is there a way to find out which process is causing SIGABRT?

thanks.

Are you running in jail?

UD.

i have seen this randomly occuring with mod_php4 (and 5 .. but NO where near as much) which extensions do you have loaded?

i am not running in a jail environment.

PHP 5.0.3, had installed from ports with all extensions.

is there a way to find out which script is causing the signal 6.
fyi, apache restarts itself after every signal 6 that it receives.

apart from Horde/IMP, i am also running spamd and clamav. basically the machine is an email server.


thanks.

I'm getting signal 10s every now and again, no idea what's causing them. Unless you can find something in your error logs that give some sort of indication, the only way to isolate the script would be to disable them one by one and see if the errors go away.

thanks for all the suggestions. there was a problem with IMP (horde.org). the horde team patch it up and have not seen the problem for abt a week now.

at least its all sorted now!

spoke too soon. the mesgs are back.

edited the httpd.conf file and assign each website its own error log. also trying to generate a core dump to do a backtrace. still not sure which of the scripts is causing this.

thanks.

Update your Apache version to Apache 2.0.54 and PHP to PHP 5.0.4, then get back to us.

I have seen the same problem on the server i administrate, but most of it disappeared when those were updated. There seems to be a bug in PHP which is in both versions, and is causing quite a bit of trouble using VirtuaNews with vBulletin.

Also, since FreeBSD 5.4-RELEASE is out, update to that, reboot, and then recompile PHP and Apache to see if the errors disappear.

updated both and had not seen any errors for abt a week now.

thanks.