All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Wordpress with Nginx + php5-fpm gives me strange 502 errors
Hi all,
I have recently encountered a new kind of problem with my Nginx php5-fpm installation.
I use it for a production website with a lot of members and traffic.
The website basically exists of a wordpress blog (as cms and for blogs),
and a vBulletin 4 forum. The vBulletin 4 forum works perfectly with php5-fpm and Nginx, I added some Nginx-htaccess rules (you know what I mean, for pretty urls under Nginx).
The rest of the config is as default.
Installing wordpress (i did it after the forum) was not problematic either. But when running wordpress, I get strange 502 errors from Nginx when I'm working in the admin control panel, just clicking simple links.
Also, some pages from the Wordpress admin seem to not completely load (as if the page is just the half of the page, the rest is white).
I hope you can help me understand (and fix) whats going wrong here.
Here is my error.log: http://pastie.org/private/praczievxolngnvg3wmreg
And here my php5-fpm.log: http://pastie.org/private/jghgfun4yyjc8aoqpkkg
Thanks for all of your time and attention. If more info is needed, just ask and I will provide the corresponding files.
Comments
Try increase the number of fpm processes in your php-fpm.conf.
what do i have to edit then?
theres no setting for number of processes in php-fpm.conf.
try increasing the execution time / memory limit in php.ini
Are you using CloudFlare?
@livinvirtual:
no, not using it (yet, might be in the future)
What about your php error log? Could be the php memory limit?
Big huge black box for me. Let me guess. Dark print?
You young people today....
@drmike:
sorry for the colors, pastie did that by itself
here is the raw version:
http://pastie.org/pastes/2722441/text?key=praczievxolngnvg3wmreg
I found the solution, and to be honest, it was a pretty stupid one:
I set the php5-fpm memory limit to 4096mb (i have 8gb ram dedicated) but I was thinking this was the total for all instances. But instead, each instance could be 4096mb which was why my requests got broken...