Last year i’ve gave one of my sites to people who really enjoys from opening undesirable sites on multi-user blogging sites. The site was a Wordpress MU site.
After that day they opened about 5K blogs on that site. 5K site means about 55K tables and 165K files on a single mySQL database. Till last month the site worked fine but after that some problems started to occur.
The main problem was sleeping zombie mysql threads created by wordpress mu. They could not be closed until they reach default wait_timeout (which is about 28800 secs.). So after about 20 process or according to server load mySQL starts to not respond requests and reaches maksimum connection limit.
I searched a lot but could not find out a solution. Here are some explanations to improve performance on high volume wordpress mu sites but those solutions won’t help me much. Was trying to decrease wait_timeout to 30 secs won’t help so much too cause whenever server load reaches high values according to high volume of created threads this timeout value could not help you.
So what i did? I simply closed that site. That’s why that site is not important for me, i’m just making a stress test on Wordpress and Wordpress Mu.
But if you’re planning to use Wordpress MU and you are considering that your site will reach thousands of blogs i will not suggest you to use Wordpress MU. This is experimental ofcourse but i’m sure that it’ll end soon with a failure.
Tags: mysql, wordpress, wordpress mu


Entries (RSS)