A thirty minutes intro to docker

Prerequisites

• Docker Engine (Linux).
• Docker Desktop (macOS/Windows).
• Blubber.

Follow along

A copy of these slides, and supporting code can be found
here and here.

Quickstart:

What's a docker?

A CLI to Linux capabilities and interfaces that we use to build, start and stop application containers.

User:Maklaan, Public domain, via Wikimedia Commons

Use cases

• Packaging and vendoring of complex dependencies.
• Developer productivity.
• Lightweight "virtualisation".
• Running unit tests and CI pipelines.

Images, containers

• An image is a read-only file that contains all dependencies that an application needs to run (e.g. glibc, python, flask, nginx, bash).
• A container is a run-time environment (atop Linux capabilities) on which applications (and their deps) run isolated from the host OS while sharing the same kernel.
• Docker Hub is an online repository of images, that can be fetched to run containers.

Hello world

What did just happen?
Explore a Unix inspired cli: docker ps|cp|kill|help.

Writing our own image

Before we pulled an existing image. Now we'll write on from scratch.

• We first need to define a Dockerfile with the required deps. It contains instructions to build Docker images.
• We then build an image with docker build -t hello-pythonic-world .
• And finally run a container with docker run hello-pythonic-world.

Blubber: high level Dockerfile descriptions

• https://wikitech.wikimedia.org/wiki/Blubber.
• Developed at WMF and used as part of our CI pipelines.
• Supports for Node, Python and Golang runtimes.

Blubber: hello-blubber.yaml

Blubber: build & run

• blubber hello-blubber.yaml hello > Dockerfile
• cat Dockerfile
• docker build --tag blubber-tutorial-hello-world .
• docker run blubber-tutorial-hello-world:latest

Where to go next

• managing volumes and networks.
• security of docker containers.
• docker-compose.
• kubernetes.