Containers and Images: First Docker Commands

“It works on my machine”: the phrase every developer has heard. The application works locally, but in production the Node version, a system library, or an environment variable differs. Docker solves this problem by packaging the application with its whole environment in an image that runs the same way everywhere.

The vocabulary

• Image: a read-only template (file system + metadata), built in successive layers. Example: node:22-slim.
• Container: a running instance of an image, with its own writable layer. You can run ten containers from the same image.
• Registry: a repository of images (Docker Hub, GitHub Container Registry, cloud registries).

Unlike a virtual machine, a container does not include a complete operating system: it shares the host’s kernel and isolates itself using Linux namespaces and cgroups. It starts in a fraction of a second and weighs a few tens of MB.

A container’s lifecycle

# Download an image from Docker Hub
docker pull nginx:1.27-alpine

# Run a container in the background (-d), named, host port 8080 -> container port 80
docker run -d --name web -p 8080:80 nginx:1.27-alpine
# http://localhost:8080 shows the nginx welcome page

docker ps                 # running containers (-a to include stopped ones)
docker logs -f web        # follow the logs (the process's standard output)
docker exec -it web sh    # open a shell INSIDE the container
docker stop web           # stop gracefully (SIGTERM signal then SIGKILL)
docker rm web             # remove the container
docker images             # images present locally

A container lives as long as its main process. If that process ends, the container stops.

Publishing ports

By default, a container is isolated from the host’s network. The -p HOST:CONTAINER option publishes a port: -p 3000:3000 makes a Node API reachable on localhost:3000. To expose it only locally: -p 127.0.0.1:3000:3000.

Persisting data

Everything written inside the container is lost when it is removed. For a database, use a volume:

docker volume create pgdata
docker run -d --name db \
  -e POSTGRES_PASSWORD=dev_seulement \
  -v pgdata:/var/lib/postgresql/data \
  -p 127.0.0.1:5432:5432 \
  postgres:17

• -e sets an environment variable;
• -v pgdata:/path mounts the named volume: you can delete and recreate the container without losing the data.

In development, you also mount a host folder (bind mount): -v "$PWD":/app.

Cleaning up

docker system df          # space used
docker container prune    # removes stopped containers
docker image prune        # removes untagged images

Key takeaways

Image = template, container = instance. docker run with -d, -p, -e, -v, and --name covers most uses; ps, logs, and exec are for diagnosis.