Docker

Why using docker on local development?

Docker is used to manage the code on our server. By using it on your local machine, you will be able to reproduce an environment very close to the deployment one, and so it will be easier to track issues before they appear.

In practice, using Docker on your local machine will allow you to:

• Use the cache system: Django comes with a cachin system, that you can only use with Docker
• Use the Postgres database: by default the database uses SQLite, but the deployment database uses postgresql, and so Docker does
• Use the celery service: the celery service allow you to run asynchronous tasks on the server from the backend
• Uses more debug tools: with Docker, you can for instance access directly your database using PgAdmin for example, or use the Workers Dashboard for celery tasks

How to run docker for local dev

1. Install docker-desktop from the official website: docs.docker.com/get-docker
2. Add the "docker needed" variables in your .env environment file. It must be at this place: backend/config/settings/.env
3. Go to the deployment directory: cd deployment/
4. Create an empty file backend.env in this directory
5. Build the containers: docker-compose build
6. Start the services: docker-compose up
7. Try to connect to http://localhost in your browser.

If you can access the login page, congratulations everything is ok! 🥳

Let's create a superuser account now, so that you can connect. For this, follow the tutorial here. You only have to replace the command for creating a superuser by:

docker-compose exec backend python3 manage.py createsuperuser

You can run this command in another terminal (but always in the deployment directory).

Useful tips

• You can stop docker by pressing CTRL+C in the console if docker is running, or by running the command docker-compose down.
• You can also launch docker services in background by adding the -d argument on the up command.
• If you want to be faster, you can run both command at once by using docker-compose up --build

What does it do?

When you run docker-compose without specifying the files to run, it runs by default docker-compose.yml and then docker-compose.override.yml. It is equivalent to run:

docker-compose -f docker-compose.yml -f docker-compose.override.yml up --build

In our case, the files are configured to run the backend (with django), an instance of postgresql (the database), and the celery services (for background tasks).

Which services are launch?

These services are launch when you run Docker on your local machine:

Service	Description	Access
database	The postgresql database for the website	You can read it by connecting to the port 5432 on localhost with PgAdmin4, and using the credentials defined in your .env file
backend	The django server which serve files	Open http://localhost in your browser
nginx	Used to serve the backend and the static files	-
redis	Used for the django cache system	-
celery	Used for asynchronous tasks runned in background	-
celery-beat	Used for linking celery to the backend with django	-
workers-dashboard	A dashboard to see all the celery tasks	Open http://localhost:5555 in your browser
mailpit	An email server to show emails sent by django in local	Open http://localhost:8025 in your browser

How to run docker for production server

1. Go to the deployment directory.
2. Create the environment files at the root of this directory:
   • backend.env for all environment variables related to django
   • mailu.env for those related to the mail server
   • frontend.env for those related to the react frontend
3. Build and run the docker-compose files:

   docker-compose -f docker-compose.yml -f docker-compose.prod.yml up --build -d

info

Note that in practice, we will not stop and restart all services each time we want to make an update. To do so, we can indicate which service we want to restart. For example, to restart only the backend:

docker-compose -f docker-compose.yml -f docker-compose.prod.yml build --no-cache nginx backend
docker-compose -f docker-compose.yml -f docker-compose.prod.yml up

The --no-cache option allowed to delete the cache and be sure that all files are really updated.