1

After several hours of battling with this anomaly, I've finally admitted defeat. If anyone here can help me I would be most grateful!


QUICK SUMMARY:

I have this error in my Nginx error log:

2009/11/09 02:49:59 [error] 90#0: *7 connect() to unix:/tmp/passenger.54/master/helper_server.sock
failed (61: Connection refused) while
connecting to upstream, client:
127.0.0.1, server: test.local, request: "GET / HTTP/1.1", upstream:
"passenger://unix:/tmp/passenger.54/master/helper_server.sock:",
host: "test.local"

The full details are below.


I've installed Nginx and Passenger on my local machine with a view to using it for local development. I was running Apache and Passenger but I was getting a few bugs so I decided to switch.

I installed the Passenger Nginx extension without any trouble into /usr/local/nginx

and configured Nginx in /usr/local/nginx/conf/nginx.conf

For completeness here is the contents of nginx.conf:

worker_processes  2;
pid        logs/nginx.pid;
events {
    worker_connections  1024;
}
http {
    passenger_root /usr/local/lib/ruby/gems/1.8/gems/passenger-2.2.5;
    passenger_ruby /usr/local/bin/ruby;
    include       mime.types;
    default_type  application/octet-stream;
    sendfile        on;
    keepalive_timeout  65;
    server {
     listen 80;
     server_name test.local;
     root /Users/i0n/Sites/test/public;
     passenger_enabled on;
     rails_env development;
     client_max_body_size 10M;
     client_body_buffer_size 128k;
     }
}

I then added /usr/local/nginx/sbin to my path variable. So far so good. If I start up Nginx from the terminal using sudo nginx I get no trouble whatsoever, I can view my test Rails app at http(colon)//test.local as expected.

The problem arrises when I try and configure the system to load Nginx on startup. I have followed the instructions from the Nginx wiki (with some adjustments for differing paths)

I have created a plist file in: /System/Library/LaunchDaemons/nginx.plist

It's contents look like this:

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
  <dict>
    <key>Label</key><string>nginx</string>
    <key>Program</key><string>/usr/local/nginx/sbin/nginx</string>
    <key>KeepAlive</key><true/>
    <key>NetworkState</key><true/>
    <key>StandardErrorPath</key><string>/usr/local/nginx/logs/error.log</string>
    <key>LaunchOnlyOnce</key><true/>
  </dict>
</plist>

I then run sudo launchctl load -F /System/Library/LaunchDaemons/nginx.plist

If I reboot the system and go to http://test.local in the browser I see a page that says:

502 Bad Gateway nginx/0.7.61

If I look in the Nginx error log (which is at /usr/local/nginx/logs/error.log)

I see the following error recorded:

2009/11/09 02:49:59 [error] 90#0: *7 connect() to unix:/tmp/passenger.54/master/helper_server.sock
failed (61: Connection refused) while
connecting to upstream, client:
127.0.0.1, server: test.local, request: "GET / HTTP/1.1", upstream:
"passenger://unix:/tmp/passenger.54/master/helper_server.sock:",
host: "test.local"

If I manually stop and then start Nginx from the terminal using sudo it begins to work. To be honest this one has me stumped, it's probably something really simple but I just can't see it at the moment. Any thoughts or suggestions would be gratefully received. Thanks for taking the time to read this!

Ian

1
  • 1
    Just as a quick update to this problem. I have discovered that this issue only manifests in server configurations that invoke passenger. Static Nginx configurations seem to load fine. Therefore I think this is some sort of Passenger permissions issue...
    – i0n
    Nov 9, 2009 at 5:19

1 Answer 1

4

You need to set 'daemon off'. launchctl assumes that processes it spawns don't daemonize, otherwise it cannot know when to restart a process. http://almosteffortless.com/2009/09/16/passenger-with-nginx-on-mac-os-x/

0

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .