On Wed, Aug 22, 2007 at 01:51:17AM +0000, dyna wrote:
Hello,
I seem to be having an issue which I can't track down on my own. I recently
installed nginx on a xubuntu server. And it's using FastCGI to server PHP.
However I get 502 Bad Gateway errors currently, instead of it working.
It's really odd, because it is an identical configuration to a server that does
work, or at least I think it's identical. I'm including the info below, any help
would be great
---error.log
2007/08/21 21:23:08 [error] 5792#0: *9 recv() failed (104: Connection reset by
peer) while reading response header from upstream, client: 127.0.0.1, server:
my.domain.com, URL: "/index.php", upstream: "fastcgi://127.0.0.1:8003", host:
"localhost"
---nginx config
location ~ .*\.php$ {
include conf/fcgi.conf;
fastcgi_pass 127.0.0.1:8003;
fastcgi_index index.php;
fastcgi_param SCRIPT_FILENAME /var/www/my.domain.com$fastcgi_script_name;
}
I spawn my FastCGI instances using a script, but they do appear to be running:
dummy 5854 0.0 0.8 13876 4304 ? Ss 21:23 0:00
/usr/bin/php5-cgi -q -b 127.0.0.1:8003
dummy 5855 0.0 0.3 13888 1836 ? S 21:23 0:00
/usr/bin/php5-cgi -q -b 127.0.0.1:8003
And finally
php5-cgi -v
PHP 5.2.3-1ubuntu4 (cgi-fcgi) (built: Aug 10 2007 09:14:51)
Copyright (c) 1997-2007 The PHP Group
Zend Engine v2.2.0, Copyright (c) 1998-2007 Zend Technologies
Do you see "104: Connection reset by peer" sometimes or always ?
I have one report when sometime ECONNRESET was returned after upgrade
to Linux kernel 2.6.22.1. After rollback to 2.6.16.48, the error disappeared.
BTW, ECONNRESET means that RST packet was sent. On localhost it can be if
1) user process closes socket with SO_LINGER timeout 0,
2) kernel initiative.
As I understand PHP does not use SO_LINGER at all.
3 messages in ru.sysoev.nginxRe: 502 Bad Gateway
