Re: apache reload and child fault
hi roberto,
some time ago, the same thing happened to me. more recently, i installed
a new mailserver which had the problem twice. the problem does not seem
to be debian-specific, since i've been contacted by people running other
distributions through my previous post, experiencing the same problem:
http://lists.debian.org/debian-apache/2003/debian-apache-200309/msg00000.html
nobody seems to know what's going on. are you running with mod_ssl??
cause i thought it might have something to do with that, since it never
happened to one of my other servers without mod_ssl. it's just a hunch
though. i have no clue how i should track this problem down, since it
seems to happen out of the blue..
for the time being, i've created a script which kills all apache
processes after the logrotation and starts it back up again. i did this
a week ago, and have had no problems since. it's a stupid workaround,
and i still have to see if it actually works for me :-)
sorry i can't be of more use..
kind regards,
-rodi
On Tue, 2003-12-02 at 11:05, Roberto Sebastiano wrote:
> Hi,
> I had a strange problem tonight with two twin web server running debian
> stable (apache 1.3.26-0woody3).
>
> These web server are under a load balancing. Now, logrotate runs every
> day at 00:10. Tonight apache won't reload properly, and *BOTH* machines
> went down until i manually killall apache and started it
>
> First machine:
> on /var/log/syslog:
> Dec 2 00:10:01 homer /USR/SBIN/CRON[27676]: (root) CMD
> (/usr/sbin/logrotate /etc/logrotate.conf.apache)
> on /var/log/apache/error.log
> [Tue Dec 2 00:10:05 2003] [alert] Child 3894 returned a Fatal error...
> Apache is exiting!
> nagios alert:
> Data:
> Tue, 02 Dec 2003 00:22:15 +0100
> Oggetto:
> ** PROBLEM alert - Homer eth1/HTTP
> is CRITICAL **
>
> Second machine:
> /var/log/syslog:
> Dec 2 00:10:01 flanders /USR/SBIN/CRON[13124]: (root) CMD
> (/usr/sbin/logrotate /etc/logrotate.conf.apache
> /var/log/apache/error.log
> [Tue Dec 2 00:10:04 2003] [alert] Child 31728 returned a Fatal error...
> Apache is exiting!
> nagios alert:
> Oggetto:
> ** PROBLEM alert - Flanders
> eth1/HTTP is CRITICAL **
> Data:
> Tue, 02 Dec 2003 00:35:59 +0100
>
> From 00:10 and on both servers stopped responding.
> While ps ax showed the apache childs running, /etc/init.d/apache stop
> didn't work saying "no apache running"
> A manual restart (killall apache && /etc/init.d/apache start) on *both*
> server solved this
>
>
> Any idea on what's going on ?
Reply to: