Can't communicate with nginx from Vagrant host machine

Question

I am using Ansible to compile and configure Nginx inside a Vagrant box.

If I run curl http://localhost/ from inside the box I get the expected response (phpinfo()).

If I access the URL http://testapp:8080/ from my Mac host machine then I can't connect.

My configurations is as follows:

HOST: /etc/hosts

192.168.100.10 testapp

HOST: vagrantfile

Vagrant.configure("2") do |config|

  config.vm.define "web" do |web_config|
    web_config.vm.box = "raring64"
    web_config.vm.box_url = "https://dl.dropboxusercontent.com/s/{{redacted}}/raring64.box"
    web_config.vm.network "forwarded_port", guest: 80, host: 8080
    web_config.vm.network "private_network", ip: "192.168.100.10"

    web_config.vm.provision :ansible do |ansible|
      ansible.playbook = "devops/webserver.yml"
      ansible.hosts = "webservers"
      ansible.inventory_file = "devops/hosts"
      ansible.verbosity = "vv"
      ansible.verbose = "true"
    end

    web_config.vm.provider "virtualbox" do |v|
      v.customize ["modifyvm", :id, "--memory", "256"]
    end
  end
end

GUEST: /etc/nginx-1.5.6/nginx.conf

user www-data www-data;
worker_processes  1;

error_log  /var/log/nginx/error.log;
pid        /var/run/nginx.pid;

events {
  worker_connections  768;
  multi_accept on;
}

http {

  include       /etc/nginx-1.5.6/mime.types;
  default_type  application/octet-stream;

  access_log    /var/log/nginx/access.log;

  sendfile on;
  tcp_nopush on;
  tcp_nodelay on;

  keepalive_timeout  65;

  server_tokens off;

  gzip  off;
  gzip_http_version 1.0;
  gzip_comp_level 2;
  gzip_proxied any;
  gzip_vary off;
  gzip_types text/plain text/css application/x-javascript text/xml application/xml application/xml+rss text/javascript application/javascript application/json;
  gzip_min_length  1000;
  gzip_disable     "MSIE [1-6]\.";

  include /etc/nginx-1.5.6/sites-enabled/*;
}

GUEST: /etc/nginx-1.5.6/sites-enabled/testapp.conf

upstream phpbackend {
    server unix:/var/run/php-fpm-www.sock;
}

server {
    server_name  testapp;
    listen 0.0.0.0:80;
    root  /web/testapp/public;
    index index.php;

    access_log  /var/log/nginx/webvg.access.log;
    error_log /var/log/nginx/webvg.error.log debug;

    location / {
        try_files $uri $uri/ /index.php?$query_string;
    }

    location ~ \.php$ {
        fastcgi_pass phpbackend;
        fastcgi_index index.php;
        fastcgi_param SCRIPT_FILENAME /web/testapp/public/$fastcgi_script_name;
        include fastcgi_params;
        fastcgi_param CACHE_HOST 127.0.0.1;
        fastcgi_param CACHE_PORT 6379;
        fastcgi_param NEO_HOST   127.0.0.1;
        fastcgi_param NEO_PORT   7474;
        fastcgi_param SERVER_ENV dev;
    }

    location ~ ^/(php_status|php_ping)$ {
        fastcgi_pass phpbackend;
        fastcgi_param SCRIPT_FILENAME /web/testapp/public/$fastcgi_script_name;
        include fastcgi_params;
        allow 127.0.0.1;
        deny all;
    }

    location /nginx_status {
        allow 127.0.0.1;
        deny all;
        access_log off;
    }

    location ~ /\.git {
        deny all;
    }
}

Can anyone spot something obvious that I'm missing?

Many thanks

Solution 1

So I have edited the /etc/hosts file to point to 127.0.0.1 instead of the address defined in the vagrantfile and I can now access the website from the host machine.

Solution 2

It's possible I'm going mad, but I think your app is really on http://192.168.100.10:80 (and that should work from both host and guest, or testapp.com:80 from the host alone) rather than port 8080.

The 8080 reference refers to the port forward from the mac host, where localhost:8080 should work as it points to 192.168.100.10:80 on the guest.

Comments

[Steve Bennett]

Anything in nginx logs? Would tunnelling as a workaround solve your problem? What about doing the port forwarding using nginx?

[Falcon Momot]

It's worth mentioning that localhost is not an especially good test to be running, as it leads to precisely this type of (non-)problem.