Dockerfiles.md

If your interests go beyond running other people's programs, you'll need to know how to use an OCI builder.

First, we need a program to containerize. Lets write a small script, program.sh:

#!/bin/sh
echo 'WORKING!'
cat myfile.txt

Make sure that this file is executable. You may have to run chmod +x program.sh.

And create a small file myfile.txt with the following contents:

Hello world! This is a file.

Now, we need to write a recipe for our image. Open a new file called Dockerfile. Our first example will be extremely simple, and we'll look at each line as we go:

FROM alpine: this tells our builder to use the alpine image as our base—any changes we make will be layered on top of alpine. WORKDIR /app: this command doesn't change anything about the image, but tells certain following instructions to execute inside this directory in the image. COPY . .: this command copies everything from our build context (where you run the build command on your host system) to our WORKDIR CMD /app/program.sh: this sets the default behavior of our image.

Together, we have:

FROM alpine
WORKDIR /app
COPY . .
CMD /app/program.sh

Now, run docker build . Check your images with docker image ls—you should see one built a few seconds ago! Copy its id for the next step.

Use docker history <image-id> to see the steps used in the creation of your image—they should roughly match your Dockerfile

Finally, we can run our program with docker run <image-id>. Congratulations on making your first image!

Once you're done, clean up your image with docker image rm <image-id>.

Additional Info

Run docker build with -t <tag> to add a tag to your image. You can also add a tag after building with docker tag

Try running docker inspect <image-id> to get a detailed look at the image contents.

You can use the RUN instruction to run commands during the build process. Exercise: Try installing some commands during the above build process, and use them in your program.sh.