Docker is an open lightweight virtualization technology that allows users to build, distribute, and run applications. For the University of Illinois, Research Park course Practical Data Science we will use the Docker platform to provide a homogenous programming platform to simplify learning data science concepts. The Docker platform consists of the Docker Engine, which is a portable and lightweight runtime, and the Docker Hub, which provides an easy mechansism to share docker container images.
To use Docker in this class you will need to complete two tasks:
-
Install the Docker Engine
-
Pull and install the docker container image for this course.
Note that in order to run the Docker engine, your computer must support hardware virtualization.
The official Docker documentation provides clear installation guides for a variety of platforms.
Since the Docker engine leverages Linux containerization, it requires components from the Linux kernel. Thus, for a modern Linux operating system, installation of the Docker engine is rather straightforward. To install the Docker engine on Windows or Mac OSX, however, you first need to install Boot2Docker, which provides the Linux kernel functionality reqquired for the Docker engine. This functionaility is actually provided by installing the VirtualBox virtualization envirnment and a Linux guest operating system.
Please see Installing Docker on Windows.
Please see Installing Docker on Mac OSX.
If you are running Linux on your computer, you probably already know how to install packages, but if you are not sure, look for your specific Linux distribution in the installation section of the Docker documentation.
If you are on a Windows or Mac OSX computer, you will have first
installed the Boot2Docker application. Once this application is
installed, run the Boot2Docker application that will open and
initialize a shell session in the VirtualBox virtual machine, note that
a Virtual Machine will be started if one is not already running.
As an example, here is the Boot2Docker initiated shell script on Mac OSX:
The Docker container image for this course has Python 3 and a number of
Python libraries that we will use during the rest of this course already
pre-built. Once the Docker engine is installed, you can pull our course Docker
container from the Docker hub. This is accompished by issuing a docker pull lcdm/rppds command at the docker engine shell session. Thus on Windows or Mac OSX, this command is entered at the Boot2Docker shell session prompt. After a succesful download of a Docker container image, you will be presented with a suitable message along with the command prompt. To verify the container, you can run the Docker container.
$ docker run lcdm/info490 /bin/echo "hello world"
hello world
If you have been succesful, the hello world message will be displayed in your terminal window.
Although not required for this course, you can share folders between a Docker container and your normal, host operating system. For a Mac OSX or Windows system, you will first need to Enable Folder Sharing in Boot2Docker
For Linux systems, you can follow the general instructions for setting up data volumes in the Docker documentation.