It is worth noting that creating a Dockerfile will allow you to easily document, build, and manage your images, but we’ll start off with a look at a quick method for testing purposes.

First, you’ll need to install Docker locally. Once you have Docker installed, you’ll be able to pull any image you like from the massive number of resources available in Docker Hub. Once you’ve found the base image you’d like to build upon, you can simply run:

For the sake of a working example, let’s go ahead and pull the latest Ubuntu image:

This will pull the image from Docker Hub down to your local machine. To spin up the Docker container and gain access to a pseudo-terminal running therein, simply run:

This will give you root access to the running container. From there, you will be able to make changes to the Docker container. These changes can be anything you like – configuration of resources already in the container, installation of new resources, or even newly written code. Just for the fun of it, let’s install Python in our newly spun-up container.

Then, after exiting the running container, you will need to create an account on Docker Hub, and then log in from the command line with your newly minted credentials, like so:

Once you’ve entered your credentials, you will need to create a repository within your Docker Hub account. To do this, you’ll need to use the Docker Hub UI.

Next, we’ll need to get the ID of the container in which we updated the base Ubuntu image, like so:

With the container’s ID in hand, we can now commit our changes locally:

And finally, you can push the image to Docker Hub with the following:

And voila! The updated image (jdvincent/ubuntu-python3.8) can now be passed in any Kubernetes YAML file, and it will be available to Kubernetes!

By default, Kubernetes looks for Docker images in Docker Hub. In order to make changes to the Docker image you’d like to use in Kubernetes, you’ll most likely need to commit an image locally, and then push that image up to Docker Hub.  

‍