toert/django-gitlab-ci-guide

How to deploy your Django application in 2 hours using Docker & GitLab CI/CD

Estimated reading time: 7 minutes

Before start

The guide is aimed to give a reader common knowledge about using Docker and Gitlab CI/CD for Django project deployment.

Default demo-project contains: Django + PostgreSQL + Celery + RabbitMQ + nginx

Docker, Docker-Compose and GitLab CI/CD are also used for the project continuous delivery.

Step 1: Create a GitLab repository

Create a blank GitLab project, push your project from local machine to GitLab. After that you will probably get something like that

For the guide I will use following project URL:

https://gitlab.com/your_username/your_project_name/

You have to replace your_username and your_project_name in the source code to make it work.

Step 2: Create a Dockerfile

Source file: Dockerfile

The second step is creating a Django & Celery image. A Docker image is a file, comprised of multiple layers, used to execute code in a Docker container. A Docker image is described in Dockerfile by default. The most interesting commands is:

FROM python:3.6 creates a layer from the python:3.6 Docker image. You can use other versions of Python (3.5, 3.4, 2.7, ...)
RUN pip3 install -r ${APP_ROOT}/requirements.txt installs Python dependencies
CMD ['python3 manage.py collectstatic --noinput', '&&', '/bin/sh','-c','python manage.py runserver'] is used to set a default command for the image. However it will be over-written by a Docker-compose file.

Step 3: Define services in a Docker-Compose file

Source file: docker-compose.yml

Using Docker-Compose is the one of the most easiest way to orchestrate all containers. For each part of the project you must create an image, there are images which described in docker-compose.yml:

nginx pulls an image from Docker Hub. The volume section overrides a default nginx config with the devops/nginx.conf
web is the main Django application's service that uses a Docker image created in the Docker chapter
postgres service. Using all default settings
rabbit uses rabbitmq:3.7-management which automatically run RabbitMQ admin web interface
celery service also using an previously created Dockerfile

Step 4: Connect containers using environment variables

Source file: .env

A module called envparse will be used to export environment variables.

First of all, add it to requirements.txt. Then create .env at the root of the project:

POSTGRES_DB=postgres
POSTGRES_USER=postgres
POSTGRES_PASSWORD=postgres
POSTGRES_HOST=postgres
POSTGRES_PORT=5432
CELERY_BROKER_URL=amqp://rabbitmq:rabbitmq@rabbit:5672/
DJANGO_SETTINGS_MODULE=conf.settings

Step 5: Create GitLab CI/CD pipeline

Source file: .gitlab-ci.yml

For the education purposes .gitlab-ci.yml contains only 2 basic steps:

build step builds Django application image and push it to a private registry. Substitute a registry URL by your own link.
deploy step logins to a server via ssh and pulls changes from registry

The deploy step uses a devops/deploy.sh file. You have to change it: replace 47.47.47.47 to your remote address IP and replace registry URL:

docker pull registry.gitlab.com/your_username/your_project_name:latest

Step 6: Push new files

Push all new files and directories (Dockerfile, docker-compose.yml, .gitlab-ci.yml, devops/) to the branch master.

Step 7: Generate SSH keys

Locally run a command ssh-keygen -t rsa, it will prompt you to enter passphrase - leave it blank. The command generates two files with two keys: public and private.

Copy the public key to your server ~/.ssh/authorized_keys file.

Also you must to set a CI/CD environment variable called DEPLOY_KEY to your private key. Go to GitLab project page - Settings - CI / CD - Environment variables and create a variable

Step 8: Clone a project

Log in the remote server via ssh and clone a project at the root:

git clone https://gitlab.com/your_username/your_project_name.git

Step 9. Make a commit to prove that everything works fine

Or create an issue to make me now that the guide has a mistake