I'm new to nginx and am attempting to run Redmine 2.3 using ruby 2.0.0, phusion passenger 4.0.2 and nginx 1.4.1. This is compiled from the passenger-install-nginx script on an ARMv5te system in Arch Linux ARM.
I am getting the following in the nginx log: (Note: I've removed the server name from the logs/configs)
2013/05/08 23:41:12 [notice] 1359#0: signal process started
[ 2013-05-08 23:41:13.1325 1367/b6f9a000 agents/HelperAgent/Main.cpp:554 ]: PassengerHelperAgent online, listening at unix:/tmp/passenger.1.0.1363/generation-0/request.socket
[ 2013-05-08 23:41:13.2641 1387/b6f26000 agents/HelperAgent/Main.cpp:554 ]: PassengerHelperAgent online, listening at unix:/tmp/passenger.1.0.1381/generation-0/request.socket
[ 2013-05-08 23:41:13.3028 1392/b6faa000 agents/LoggingAgent/Main.cpp:272 ]: PassengerLoggingAgent online, listening at unix:/tmp/passenger.1.0.1381/generation-0/logging.socket
[ 2013-05-08 23:41:15.9700 1387/b59ff450 Pool2/Spawner.h:739 ]: [App 1408 stdout]
[ 2013-05-08 23:41:43.5820 1387/b6b4a450 Pool2/Spawner.h:159 ]: [App 1408 stderr] /var/www/sites/public-redmine-2.3/lib/SVG/Graph/Graph.rb:3: warning: class variable access from
toplevel
[ 2013-05-08 23:41:55.8491 1387/b59ff450 Pool2/Spawner.h:739 ]: [App 1439 stdout]
[ 2013-05-08 23:44:15.2842 1387/b57ff450 agents/HelperAgent/RequestHandler.h:581 ]: [Client 20] Disconnecting with error: invalid SCGI header
2013/05/08 23:44:15 [error] 1402#0: *3 upstream prematurely closed connection while reading response header from upstream, client: 192.168.4.254, server: (removed), request: "POST /login HTTP/1.1", upstream: "passenger:/tmp/passenger.1.0.1381/generation-0/request.socket:", host: "(removed)", referrer: "http://(removed)/login"
The last error occurs consistently whenever I do an HTTP POST on any page in redmine. HTTP GET requests work fine with pages rendered.
nginx.conf
worker_processes 2;
events {
worker_connections 1024;
}
http {
passenger_root /usr/lib/ruby/gems/2.0.0/gems/passenger-4.0.2;
passenger_ruby /usr/bin/ruby;
passenger_max_pool_size 2;
passenger_pool_idle_time 120;
passenger_pre_start http://(removed)/;
passenger_spawn_method smart;
include mime.types;
default_type application/octet-stream;
sendfile on;
keepalive_timeout 65;
server {
listen 80;
server_name localhost;
location / {
root html;
index index.html index.htm;
}
error_page 500 502 503 504 /50x.html;
location = /50x.html {
root html;
}
}
include sites-enabled/*.conf;
}
sites-enabled/redmine.conf
server {
listen *:80;
server_name (removed);
passenger_enabled on;
rails_env production;
passenger_spawn_method conservative;
root /var/www/sites/kamikaze-kb/public;
location /plugin_assets/ {
root /var/www/sites/kamikaze-kb/public/plugin_assets/;
}
}
One thing that may be relevant - I can run redmine fine if I use the built in ruby webrick server, so I'm guessing this is likely to be an nginx/passenger problem. Has anyone got any idea what the issue is?