1. Install Docker & verify its state

Installation

https://docs.docker.com/engine/installation/

State check

Run docker run hello-world in a terminal ( You may need to update you ~/.docker/config.json file rights using the chown command)

2. Build the app

2.1. Setup the container

Find all what's inside the different files here: https://docs.docker.com/get-started/part2/#dockerfile

Container environment

touch dockerfile

The dockerfile define what goes on in the environment inside the container.

The app

touch requirements.txt

This file contains a list of all the additional required libraries to install to run our app.

The command pip install -r requirements.txt will install these libraries.

touch app.py

This file contains our app code. Its purpose is to print the environment variable NAME, as well as the output of a call to socket.gethostname().

2.2. Build the app

docker build -t friendlyhello . . This command creates a Docker image. Its identifier will be friendlyhello (with option -t). The location is the current folder, but it can be updated by overwriting the . character.

To know where the built image is stored, we can use the command : docker images.

2.3. Run the app

docker run -p 4000:80 friendlyhello.

This command runs the app:

• 4000 is the machine port where the container port 80 is mapped. (option -p)
• The app is reachable on the browser at http://localhost:4000
• Other possibility : curl http://localhost:4000

Run the app in detached mode : docker run -d -p 4000:80 friendlyhello.

To find the container ID : docker ps.

To end the detached process : docker stop THE_CONTAINER_ID

2.4. Share the image

Log in on Docker

docker login

Tag the image

Let's identify our app location: docker tag friendlyhello username/get-started:part1

With :

• friendlyhello : the name of the app
• username : the username of the account
• get-started : the repository of the image
• part1 : the tag / identifier of the image

The new image is visible with the following command : docker image ls

Publish the image

docker push username/get-started:part1 is used to upload the new image on the repository.

The new image is visible on https://hub.docker.com/.

Run the image locally

docker run -p 4000:80 username/repository:tag

This will pull the image locally if needed and run the app on the machine.

3. Use services

3.1. Define a service

touch docker-compose.yml

This file define how Docker containers should behave in production.

The docker-compose.yml file asks Docker to do the following tasks:

• Pull the image uploaded previously.
• Run 5 instances of the image as a service called web. Each instance is limited to use a maximum of 10% of the machine CPU (across all cores) and 50MB of RAM.
• Directly restart containers if one of them fails.
• Map port 80 on the host to web’s port 80.
• Ask web’s containers to share port 80 via a load-balanced network called webnet.
• Define the webnet network with the default settings.