 | Start a set with this search |
 | Include this search in one of my sets |
 | Exclude this search from one of my sets |
 | Permalink to these results Paste this link in email or IM: |
 | Atom feed for tracking future search results Paste this URL into your reader: |
2 messages in ru.sysoev.nginxRe: nginx + php = timeout| From | Sent On | Attachments |
|---|---|---|
| Anonymous Coward | Mar 9, 2007 6:04 am | |
| Jonathan Vanasco | Mar 9, 2007 6:55 am |
| Permalink for this message Paste this link in email or IM: |
| Permalink for this thread Paste this link in email or IM: |
| Atom feed for this thread Paste this URL into your reader: |
| Subject: | Re: nginx + php = timeout | Actions... |
|---|---|---|
| From: | Jonathan Vanasco (jvan...@public.gmane.org) | |
| Date: | Mar 9, 2007 6:55:20 am | |
| List: | ru.sysoev.nginx | |
On Mar 9, 2007, at 9:05 AM, Anonymous Coward wrote:
hello,
i just moved yesterday a relatively big Wordpress blog from an old box with apache2 to a new one with nginx Everything seemed to work ok until ~30 minutes ago when clients started to complain site is working slow and after few checks i noticed that i was gettting "504 Gateway Time-out" and slow loads on php pages but i had no problem with html pages and i have no clue why this is happening im using lighty's spawn-fcgi to spawn php with the following config lines
PHP_FCGI_CHILDREN=5 PHP_FCGI_MAX_REQUESTS=1000
i don't know how usefull this is but lsof showed me that i had ~1100 established nginx connections (4 workers with 1024 connections per worker so this is not an issue) and ~1000 connections to the php-cgi port (don't know if this should be a problem as i have 5 childrens with 1000 max requests each but maybe i didn't understand how this exactly works)
thanks in advance
You need more FCGI children
the MAX_REQUESTSon FCGI are how many requests a child serves before it restarts. think of it like a memory / variable cleansing ritual - it deals with clearing out memory leaks and allocated unused memory in a persistant state.
your problem is that nginx is queing 1100 fcgi requests across 4 nginx workers , and into 5 php processes
check to see how much memory you have running per php child - see if you can at least double-triple your number of children.
if you're not already -- running a cache like eaccelerator / APC will help with getting things a bit faster.
as a ballpark figure -- you'll want at least the same number of FCGI processes as you had 'max servers' in apache2
// Jonathan Vanasco
| - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - | SyndiClick.com | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - | FindMeOn.com - The cure for Multiple Web Personality Disorder | Web Identity Management and 3D Social Networking | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - | RoadSound.com - Tools For Bands, Stuff For Fans | Collaborative Online Management And Syndication Tools | - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -