Installing Boot2Docker For Windows
JeanFrancoisPuget 2700028FGP Visits (30698)
Update on October 20, 2015. Docker now offers a Windows client. The Windows installation instructions can be found at http
Who hasn't experienced the: "it runs fine on my machine, I don't get why it does not run on yours" phenomenon when deploying newly built software? I certainly did when deploying C and C++ code in the past. We had to work on making our code portable across various platforms.
Without surprise, I also got portability issues when sharing Python scripts. This became a concern of mine because I wanted to be able to write and test Python scripts on a local machine, say my Windows laptop, then deploy them on a cloud to power a web service. How could I make sure I won't run into portability issues?
Fortunately, so many people had this issue, and not only for Python, that a solution exists. It is called Docker.
Spending one day on their site will tell you all about Docker, and you'll be able to use it productively. Docker has two components:
We will use both. First step is to install Docker Engine on my development machine, which is a Windows laptop. Given that Docker Engine is a Linux application, we need to use a Linux VM called Boot2Docker to run it. Other Linux VMs like Vagrant can also be used, but Boot2Docker is the one recommended by Docker in thei
That's it! You can use Docker. Now that Boot2Docker is installed, you can launch it at will. You only need to open a cygwin terminal as above, go to the directory where Boot2Docker is installed, then run boot2docker start. If the Boot2Docker VM is already running, you can ssh to it.
I recommend the short online tutorial, and the user guide to learn more. I recommend actually running the tutorial examples and the user guide examples on your newly installed Boot2Docker installation. This should not take you more than a couple of hours.
Here are few things you can do:
In order to prepare our next step we will pull a Docker image from Docker Hub. That image should allow us to run Python scripts. As a matter of fact I am using IPython notebooks. I could go on Docker Hub directly and search there, but I can also do it from my local Docker Engine:
The most popular image is shown first. It is the ipython/scipyserver image.
We then pull the image. This starts the download.