Cheatsheets:Docker

Quick reference for smart people — part of our dev cheatsheets collection.

Docker packages software into lightweight, isolated containers, each run from a read-only image.

New to Docker? Get Started runs your first container

Basic example: serve a website

docker pull nginx                       # fetch the image
docker run -d --name web -p 8080:80 nginx  # start a container from it
docker ps                               # is it running?
curl localhost:8080                     # yes — the nginx welcome page
docker stop web && docker rm web        # stop and remove it

An image is a read-only template; a container is one running instance of it. Containers are throwaway — anything written inside them vanishes when the container is removed, unless kept in a volume.

Running containers

docker run -it ubuntu bash              # interactive terminal
docker run -d --name db postgres:16     # detached, named
docker run -p 8080:80 nginx             # host 8080 → container 80
docker run -v ./data:/app/data app      # mount a host directory
docker run --rm -it app                 # remove the container on exit
docker run -e MODE=prod app             # set an environment variable

Flag	Effect
-d	run detached (in the background)
-it	interactive terminal (needed for bash)
-p host:container	publish a port
-v host:container	mount a volume or directory
--name	give the container a name
--rm	auto-remove once it stops
-e KEY=value	set an environment variable

Port order is always host:container — -p 8080:80 means your machine's 8080 forwards to the container's 80, not the other way round.

Managing containers

docker ps                    # running containers
docker ps -a                 # all, including stopped
docker logs -f web           # follow the logs
docker exec -it web bash     # shell inside a running container
docker cp web:/etc/nginx/nginx.conf .   # copy a file out of a container
docker stop web              # stop one
docker start web             # start an existing (stopped) one
docker rm web                # remove a stopped container

docker run always creates a new container. Re-run the same task? docker stop / docker start the existing one instead — or use Compose.

Images

docker build -t myapp .           # build an image from the Dockerfile
docker images                     # list local images
docker pull postgres:16           # fetch an image (or a tag)
docker push user/myapp:v1         # upload to a registry
docker tag myapp user/myapp:v1    # retag
docker rmi myapp                  # remove an image

Dockerfile

A plain text recipe that docker build turns into an image:

FROM python:3.12-slim     # base image
WORKDIR /app              # working directory for the rest
COPY requirements.txt .   # copy file(s) into the image
RUN pip install -r requirements.txt
COPY . .                  # everything else
CMD ["python", "app.py"]  # default command when run

Volumes

docker run -v ./data:/app/data app     # bind mount: a host directory
docker run -v data-vol:/data app       # named volume, managed by Docker
docker volume ls                       # list named volumes
docker volume prune                    # remove unused ones

Without a volume, container data is lost on docker rm. The database keeps its data only because something like -v db-data:/var/lib/postgresql/data is mounted.

Docker Compose

Multi-container setups described in YAML — start, stop and rebuild everything with one command:

services:
  web:
    build: .                    # build from the Dockerfile
    ports:
      - "8080:80"
    volumes:
      - ./data:/app/data
  db:
    image: postgres:16          # or pull a ready-made image
    environment:
      POSTGRES_PASSWORD: secret

docker compose up -d          # start everything in the background
docker compose up --build     # rebuild images, then start
docker compose ps             # status of all services
docker compose logs -f        # follow the combined logs
docker compose down           # stop and remove the containers

The modern command is docker compose (space). The old v1 tool was a separate binary spelled docker-compose (hyphen) — you may still see it in older tutorials.

Cleanup

docker system prune      # dangling images, stopped containers, unused networks and cache
docker system prune -a   # ... plus every image no container uses
docker system df         # disk usage by images, containers, volumes

docker system prune -a removes images you are not running. Images you might want again later need a docker pull — they are not lost, just re-fetched.

Further reading

• Docker documentation — the official docs
• Get Started — the recommended tutorial
• Docker CLI reference — every command and flag
• Compose documentation — multi-container applications