[wp-hackers] APC-Nginx-Apache problems / memory problems on VPS

[Ian Douglas]

Thank you, John

> Do you have a test or staging environment set up?

We're migrating now to another host, to see if the site will run from a different server.
The guy who is helping me a linux administrator and he cannot see why the server is going down.

I just don't understand it either. This exact same installation worked fine in a shared environment for two weeks (albeit it hit the resource ceiling a couple of times, which was pretty low, being a shared hosting account). When we moved to the VPS, things went nuts.


> In theory, the number of plugins you run is irrelevant, but the
> quality and efficiency of those you run is. You could run one terrible
> plugin and kill your whole site, or you could run fifty well built
> plugins and have no adverse affects (and of course they could actually
> speed the site up).

It's about time Wordpress.org made a black list of known resource hog plugins. I read something about this on a community blog somewhere; maybe WPMU. At the same time, I'm fairly picky when it comes to plugins, and I do research before I install and try to get the best and most recent and up to date plugin for a given need.


> In reality, running fifty plugins on a site *is* likely to cause
> problems purely on a probability basis. You're probably running at
> least one plugin which is less than optimal in terms of memory use,
> database queries, etc.

I'm also surprised that there are no available tools for really finding what is pulling the Wordpress install (and CPU down). I mean, the CPU is trying to serve a particular need, or a particular plugin is pulling in RAM or leaking it. Surely it should be something we can know, as humans, if the computer knows, because it's happening to it!


> If you had a memory leak it should disappear when you disable all your
> plugins (unless it's caused by your theme). A memory leak will rear
> itself in the form of 'Out of memory' errors in PHP's error log.

I will look there. We did notice that with all the plugins installed, if we changed to TwentyEleven read time dropped from 19 seconds to 9 seconds. The theme we're wanting to use is WP-Prosper from www.solostream.com 

They seem a very reputable company to us. They have gotten industry awards for their themes. It will suck big time if it's a theme problem, because we already spent four weeks building our site around that theme.

But one useful thing I learn: Just because it works on MAMP, doesn't mean it will work on a real server.


>> We're look into the debug log. I get hundreds (like may 700 on the frontpage alone) of the following:
>> 
>> Notice: get_settings is deprecated since version 2.1! Use get_option() instead. in/home/bikyaweb/public_html/wp-includes/functions.php on line 3382
> 
> You should adjust your PHP error logging level so notices aren't
> logged (although I know there is a PHP bug regarding suppressing
> notices with certain error logging levels, can't remember what at the
> moment).

I meant more that the number of errors seems to indicate that this error is throughout the theme / installation. I was wondering, indeed, if curing this alone would bring some health back to the server.

The problem is that I really have no idea at all about how to cure it.

Thanks again for your reply.

best wishes,
ian