I host about 30 WordPress 3.5.1 sites on a Media Temple DV4 running Centos 5 with 2GB of RAM, PHP running behind Fast-CGI, and Apache running behind Nginx. Of these 30 sites, only three get significant traffic. The rest kind of just sit there.
I have recently installed W3 Total Cache on all of these sites with the following enabled:
CDN with Amazon Cloudfront
Cloudflare w/ medium security, automatic Rocket Loader, and CSS, JS & HTML minification
I also recently had my Apache configuration professionally optimized by Media Temple’s Otto service.
Nonetheless, despite using CDN and Cloudflare in addition to W3TC, Apache’s RAM usage has _DOUBLED_ from ~300MB to ~600MB literally overnight. My sites, despite having all this fancy caching now enabled, are now performing twice as slowly as they were before.
If anything, that number of calls being made and the amount being called should be significantly reduced. Instead, it’s killing my server.
Here’s a screenshot of what my Apache RAM usage looks like before and after I rolled out the plugin:
In addition, I also installed W3TC on a separate site on another server — a Rackspace Cloudsite. There too I am using the same settings above, but with WordPress/Jetpack’s CDN and Cloudflare. Now, half the times you load a page, the site throws a server busy error, and every time you post, the site throws a timeout error.
What the heck is going on with this plugin? It’s like it does the exact opposite of what it is supposed to accomplish.
- The topic ‘Apache RAM usage doubled after installing W3TC’ is closed to new replies.