mechanism

When container is started:

• it reads the etcd http API and looks for enabled subdomains within the CoreOS space
• it creates subsequents nginx config files
• start a simple nginx

The intent is to make every apps contained in a CoreOS host available as separed subdomains instead of the localhost + random ports logic.

example

      docker                      server                       browser

[mongodb container]
         |
[api container:3000]   <---> api.localhost:80    <------|
                                                        |
[front container:3000] <---> front.localhost:80  <---------->  client

+ dev mode:

[neo4j container:7474] <---> neo4j.localhost:80

usage

$ docker run -d -p 80:80 -e domain=localhost wekeypedia/nginx

• -p 80:80 to bridge the nginx webserver with the host port 80
• -e domain=localhost to setup the domain used by nginx (localhost is the default value is none is provided)