Every 15 minutes sounds like a timed task. Check the crontabs first to see if there is anything that may be causing this.
Software itself could be the issue as well. You may want to consider reinstalling the httpd.
Memory comes to mind, but 15 minutes on the dot doesn't sound like a memory issue to me.
Stacked with the other post you have for connections not closing, I'm leaning towards software. I'd start by patching it up to the highest level they have, and if that doesn't work, I'd reinstall and see what the results in.
And of course make sure you check the error logs to see if there is anything useful in there.
As of PHP 5.5, the MySQL library has been officially deprecated. It is recommended to move to either MySQLi or PDO libraries for your mysql connectivity. See here for help choosing which interface you prefer: http://php.net/manual/en/mysqlinfo.api.choosing.php
It is still happening. (we have a dedicated db and a dedicated web/apache)
- I removed all crontabs/cronjobs I know that are running.
- I also removed parts of the script
- I removed every session variable and session_start() making it into cookies.
- I reduced all database queries and closed scripts + errors with db->close
- I removed all the (3-4 iframes on each page) where it would use db and apache.
- the connections still come, timely. (but when we purchase traffic)
- We send traffic thru a php header redirect script. which adds a cookie and connects to the database to add +1 to stats.
- It then when they reach the website, they see ads, which run through queries. however, the site is normally perfect and healthy data. but every 20 minutes or so, connections build up to the MAX and the apache has to reboot.
What would be the cause? I'm thinking faulty hardware, memory ..etc??
All the software/OS has been upgraded,
Apache/2.2.22 (Ubuntu) PHP/5.3.10-1ubuntu3.5 with Suhosin-Patch mod_ssl/2.2.22 OpenSSL/1.0.1
The script I use is 'close' or a minimal codeIgniter style.
PHP closes connections at script end. SO it has to be Apache!
I am lost for words. I have no idea where to begin, I've rebuilt and tested so much, and the crashes still appear.
Last edited by sitNsmile; 02-16-2013 at 05:50 PM..
Had a similar problem before that turned out to be a problem with mailqueue. Even the host couldnt figure it out. Ended up literally having to move hosts.
Any possibility its something like that?
I'm not too sure, we do have a mail system running, its mostly like squirrel mail. but these crashes seemed to increase during higher traffic volume, so instead of like ever 30 minutes, it would be 20 minutes and sometimes even 17 minutes. When traffic is lower, the time is longer.