docker-nginx

Ansible role to manage and run the nginx docker container.

Requirements

This role has only been tested on Ubuntu 14.04. Since this uses Ansible's docker module, you will need to ensure that a recent-ish version of docker-py and docker are installed.

We use tests/requirements.yml to install the dependency role: geerlingguy.docker

You can use also ansible-galaxin install -r tests/requirements.yml or better add the same content to your own requirements.yml file in your own repository.

Examples

Install this module from Ansible Galaxy into the './roles' directory:

ansible-galaxy install marvinpinto.docker-nginx -p ./roles

Use it in a playbook as follows, assuming you already have docker setup:

- hosts: 'servers'
  roles:
    - role: geerlingguy.docker  # You can use any other role to install docker, but docker is a requirement (see obove)
    - role: 'marvinpinto.docker-nginx'
      become: yes
      nginx_conf: |
        user root;
        worker_processes 1;

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

        events {
            worker_connections  1024;
        }

        http {
            include       /etc/nginx/mime.types;
            default_type  application/octet-stream;
            log_format  main  '$remote_addr - $remote_user [$time_local] "$request" '
                              '$status $body_bytes_sent "$http_referer" '
                              '"$http_user_agent" "$http_x_forwarded_for"';
            access_log  /var/log/nginx/access.log  main;
            sendfile        on;
            keepalive_timeout  65;
            include /etc/nginx/conf.d/*.conf;
        }

Have a look at the defaults/main.yml for role variables that can be overridden! If you need a playbook to set Docker itself, have a look at marvinpinto.docker Galaxy role.

Expected to Be Configured

nginx_reverse_proxy_proxies: list of reverse proxy configurations; each configuration needs the following variables
- nginx_reverse_proxy_backend_name: string, name nginx config uses to refer to the backend
- nginx_reverse_proxy_domains: list of public-facing domains to be proxied
- nginx_reverse_proxy_backends: list of backend servers, including ports and other valid parameters for server in the upstream context of an nginx config file
- nginx_reverse_proxy_config_name: name to use for the proxy file (do not include the '.conf' extension, role will add this)

Example Playbook

---
# file group_vars/nginx_docker_proxy

nginx_reverse_proxy_proxies:
  - config_name: app1proxy
    backend_name: my-backend-1
    backends:
      - localhost:1880 weight=2
      - localhost:1881
    domains:
      - app1.192.168.88.10.xip.io
    locations:
      - /path/   # In case your site is hosted on backend-name/path/
    root_redirect_location: /path/  # In case your site is hosted on backend-name/path/ and need to redirect to this site by default

  - config_name: app2proxy
    backend_name: my-backend-2
    backends:
      - localhost:1882
      - localhost:1883 backup  # will act as backup, and nginx only passes traffic when primary is unavailable.
    domains:
      - app2.192.168.88.10.xip.io
    balancer_config: least_conn; # Important to add semicolon at the end ; if not the config will break

License

MIT

Author Information

Marvin Pinto

Collaborators

Pablo Estigarribia (pablodav at gmail)

Notes about nginx settings

When adding backends, if you prefer to add them using DNS ensure server can resolve the DNS name before starting nginx. If nginx doesn't resolve the DNS name, it will not start.