I noticed this nasty wee bugger last week:
[notice] seg fault or similar nasty error detected in the parent process
The first time it happened the drop in traffic did not correlate with the daily init scripts but when the second occurance this morning did I finally had somewhere to start. The default logrotate script for apache on gentoo (and many other flavours) does a graceful reload once the log files have been rotated to release their inodes. Now I could reproduce the issue:
# /etc/init.d/apache2 reload * Gracefully restarting apache2 ... # /etc/init.d/apache2 restart * Stopping apache2 ... httpd (pid 2505?) not running [ ok ] * Starting apache2 ... * start-stop-daemon: /usr/sbin/apache2 is already running # /etc/init.d/apache2 stop * Stopping apache2 ... httpd (pid 2505?) not running # /etc/init.d/apache2 start * Starting apache2 ... * start-stop-daemon: /usr/sbin/apache2 is already running
Poking around this seems to usually be caused by loading bad/wrong/nonexistent libraries in the php.ini - and I am guilty of carrying over old php.inis on upgrade. Updating and switching to the distributed php.ini for my installed version of php seems to have solved this issue.
Sometimes it really is my fault :p