Matt Fisher's Blog

Dockerfiles are a simplistic way to create a repeatable workflow for creating Docker images. Creating a description file, called a ‘Dockerfile’, will enable you to build these images. When you’ve created the Dockerfile that you want to save, you can do so by running ‘docker build .’. But what if you want to have multiple Dockerfiles in one folder (eg. so you can deploy multiple docker images)?

Currently, ‘docker build’ cannot read from a file. However, it can read from STDIN. We can use this to our advantage by running:

docker build - < filename

This allows docker to read from ‘filename’. For example, suppose that I want to hold dockerfiles for all available services that my application may need. In this example, let’s assume that I need MongoDB as my NoSQL database, memcached for caching data for my app, and elasticsearch to quickly index and search through a filter of several documents on MongoDB. For this, I would need three dockerfiles, right? So, let’s see this in action:

bacongobbler@workbox:~$ mkdir -p project/dockerfiles
bacongobbler@workbox:~$ cd $_
bacongobbler@workbox:~/project/dockerfiles$ touch elasticsearch.dock mongodb.dock memcached.dock
bacongobbler@workbox:~/project/dockerfiles$ # edit files to make legit dockerfiles
bacongobbler@workbox:~/project/dockerfiles$ cat memcached.dock
from        ubuntu
maintainer  Matthew Fisher <[email protected]>

run         apt-get update
run         apt-get install -y memcached

expose      11211

cmd         ["memcached", "-u", "daemon"]
bacongobbler@workbox:~/project/dockerfiles$ docker build -t bacongobbler/memcached - < memcached.dock
bacongobbler@workbox:~/project/dockerfiles$ # docker build output. Afterwards, let's save it on the Docker Index
bacongobbler@workbox:~/project/dockerfiles$ docker push bacongobbler/memcached
bacongobbler@workbox:~/project/dockerfiles$ # rinse and repeat

If you’re more interested in the docker project, go check out http://docker.io/ for more information.

Hope this helps someone out there!