Writing a Dockerfile page

Write a Dockerfile to containerize the Node.js app.

Our Dockerfile

Starting with the solution, the Dockerfile for our Node.js app should look like this.

✏️ Copy and paste this to the root of your project and name the file Dockerfile:

FROM node:15

ARG PORT=8000
ENV PORT=$PORT

WORKDIR app
COPY src src
COPY package.json .

RUN npm install
EXPOSE $PORT
CMD npm start

Let’s break down each line.

FROM instruction

The FROM instruction is the first line of any Dockerfile. It sets the base image to be used as a starting point for all other instructions. Using base images allows deferring:

  • installation of a kernel and
  • installtion of low level packages

to the provider of the base image.

In our case, We are using node:15. node specifies the name of the base image and 15 is a tag specifying the version of the image. This base image has the latest version of Node.js 15 pre-installed, allowing the rest of the Dockerfile to focus on logic specific to our application. The node:15 image also has a Dockerfile which means it also has a FROM instruction. node:15 uses Debian as its base image. This layering of Docker images repeats until a Dockerfile uses FROM scratch as its base image.

By default, Docker will look for images on Dockerhub. For example, the node image we are using is pulled from here, but you can create your own base images and share them publicly or privately to Dockerhub or other package registries.

Keep reading about the FROM instruction

ARG instruction

The ARG instruction is used to define variables that are only available during the image build process. The instruction follows the syntax ARG <name> [=<default value>].

In our case, ARG PORT=8000 is defining an argument called PORT with a default value of 8000. It can be accessed in later instructions in the Dockerfile like a standard environment variable ($PORT) and overwritten during the build process with the --build-arg cli argument (Docs).

Keep reading about the ARG instruction

ENV instruction

The ENV instruction is used to define environment variables. Like ARG, it follows the syntax ENV <name> [=<default value>]. The difference between ARG and ENV is that variables defined with the ENV instruction can be used during the image build process and also when the container is running as standard environment variables. However, only ARG defined variables can be actually overwritten during the build process with the --build-arg cli argument. ENV can be overwritten at run time with the -e cli argument or an env file (Docs).

To give us the flexibility to set and use variables at build time and run time, we use both ARG and ENV instructions together:

ARG PORT=8000
ENV PORT=$PORT

ARG vs ENV

Capability ARG ENV
Set during build time
Set during run time
Use during build time
Use during run time

Keep reading about ENV

WORKDIR instruction

The WORKDIR instruction sets the working directory for any subsequent instructions. If the directory does not exist, it will be automatically created.

The WORKDIR instruction can be used multiple times.

WORKDIR x
WORKDIR y
WORKDIR z
RUN pwd

Will result in RUN pwd and any later instruction being run in the /x/y/z/ directory.

In our case, WORKDIR app will create the /app/ directory for us to copy our source code in to.

Keep reading about the WORKDIR instruction

COPY instruction

The COPY instruction follows the syntax COPY [--chown=<user>:<group>] <src>... <dest>.

  • --chown=<user>:<group> allows setting the ownership of the target file or directory using chown. This is only available to Linux based images
  • <src>... specifies what file(s) or directories should be copied in to the docker image.
    • Each file/directory should be separated by a space ()
    • Each entry supports wildcards and matching. For example: COPY hom* . Will match all files starting with "hom"
  • <dest> specifies where each src entry should be copied to in the docker image. . will put each entry in the WORKDIR or the root directory if no WORKDIR is specified.

In our case, we have

WORKDIR app
COPY src src
COPY package.json .

This will result in our src/ directory being copied to /app/src/ and package.json file to /app/package.json within the image. Using two COPY instructions instead of one is just preference and has no functional difference one way or another.

There are more complex cases on when a trailing slash is required on a <src> or <dest> and how to handle whitespace. These advanced rules can be found in the official docs.

Official Docs

RUN instruction

The RUN instruction allows arbitrary commands to be run during build time.

In our case, we are using RUN npm install to install dependencies from the package.json file copied in to our image. As a reminder, npm is pre-installed as a part of our node:15 base image.

Keep reading about the RUN instruction

EXPOSE instruction

EXPOSE specifies which ports the when running the container. Each port is separated by a space (). By default the container listens using TCP, but UDP can be used by adding /udp after the port.

For example:

EXPOSE 80/udp

In our case, we want the port exposed by our container to match the port our application is running on (remember: const port = process.env.PORT || 3000 in src/index.js), therefore, we are publishing the value of our PORT environment variable.

ARG PORT=8000
ENV PORT=$PORT
...
EXPOSE $PORT

EXPOSE does not actually publish the port. When running the container with docker run, the -p flag needs to be used to access the port from your host machine. More on this in the next section. Learn more about EXPOSE

CMD instruction

CMD is a special instruction. It is not run during build time like a RUN instruction. Instead, it is run at runtime. When the CMD instruction finishes, the docker container will be destroyed.

In our case, we are running our express server with CMD npm start. This will cause our container to run until a fatal error occurs or it is manually killed.

More information about CMD

What’s next

That was a lot on Dockerfiles. Next we’ll be looking at actually building and running our image!