Luke M

Intro to containerised development

2025-03-21

This is a high-level introduction to containerised development. I am going to be using general terms to explain concepts to a newer developer.

Why

Learning to use containers and using them have several benefits for developers:

  1. Deployment is reproducible
  2. Dependencies can be installed with a container, keeping the host system clean

What

Containers

Containers can be thought of as similar to virtual machines. Virtual machines virtualise (heh) an entire operating system and kernel. Containers do this, but are more lightweight because they share the kernel with the host operating system, in most cases.

A container runtime is a piece of software used to create, run, and manage containers. Two notable container runtimes are Docker and Podman.

On Mac OS and Windows, a Linux kernel needs to be created, called a machine or engine by container runtimes.

The most basic Containerfile will contain a FROM statement

# Use the latest version of Debian from Docker hub
FROM debian:latest

You can run this Containerfile by using the following shell command (it will automatically look for a Dockerfile in the current directory):

docker build -t my_debian .
docker run -it my_debian

So what happened here?

  1. The container runtime will pull the latest version of Debian if it's not already on the host system.
  2. It will create a new image tagged with my_debian.
  3. It will then start an interactive shell session in the new container.

This is obviously a very minimal example. Containerfiles can contain many more instructions, such as ports to expose, volumes to mount, data to copy, and even shell commands to run.

As an example, here is an example of a container for PostgreSQL running in Alpine Linux:

# Use the latest PostgreSQL Alpine image
FROM postgres:alpine

# Expose the default PostgreSQL port
EXPOSE 5432

# Persist data inside the database
VOLUME /var/lib/postgresql/data

Volumes

A volume is a virtual drive (a file or directory) that a container is given access to. This can be mapped to a location on your host system. Usually, when you are developing, you will either COPY or mount your source code directory as a volume for your container to access.

Containerfile (Dockerfile)

A Containerfile is a set of instructions using a simple syntax for declaratively building a container, at the time of writing the industry standard is to name these Dockerfile.

Compose

Compose is a method of orchestrating more than one container at a time. The primary way of doing this is creating a compose.yml file, which contain services. When you run docker-compose up or podman-compose up in the directory containing a compose.yml file, it runs the services specified.

This is very useful for running an application that relies on more than one container, such as a web application with a database backend.

How

Install a container runtime

To begin, decide whether you want to use Docker or Podman as a container runtime. Docker is the industry-standard at the time of writing, and a safe choice. That said, Podman should, in most cases, be a drop-in replacement for Docker, and is likely to be more efficient on Linux systems.

In the case of Docker, you can install it using the GUI app Docker Desktop. I recommend, however, installing the CLI version and learning to use it through the command line.

Should you develop inside the container?

If you've mounted your source code directory as a volume that the container has access to, you will be have two options for how you will access and edit the code. You can either open your IDE and edit files on the host system, or set it to SSH into the container and edit files within the container.

Whichever you choose does not really matter at this stage. For this matter, the simplest way to start is to develop on the host operating system.

With that said, there are good reasons you may want to consider SSHing into the container directly and editing inside of it. One being that the dependencies for developing are installed within the container, keeping your host operating system clean.

I have written an article on the how and why of developing with a dev container. I recommend reading it when you get a chance.

Accessing a container

From another container on the same network, you can access a container by using the container's tag as the domain. But from outside the container runtime, you need to use localhost.

As an example, if you decided to create a Postgres container named my_postgres, here's what the connection string would look like:

# From the host computer
pgsql://user:password@localhost:5432/database_name

# From another container on the same network
pgsql://user:password@my_postgres:5432/database_name

Further reading