I've been working on client Django project for a few weeks now. However, so far the app just lives in a git repository and can only be run locally. It's time for deployment!

Step 1 - Folder structure

A common mistake: the code is well-documented, but the server config is not!

A typical Django deployment requires a number of server process configurations (e.g. Nginx, Celery, Supervisory, Gunicorn). These typically live in the /etc/ folder, which can be messy and unclear. What if you want to spin up a new server instance - how will you know which config files to copy over??

To mitigate this problem I use the folder structure outlined in this blog by Nick Sweeting. It recommends the following folder structure:

When deploying, all config files should be sim-linked back to the server's /etc/ directory. Et voila, now everything required for the project resides in one directory!

Step 2 - Server setup

The next step is setting up the server and virtual environment.

1. Install correct python version
   sudo apt update
sudo apt install software-properties-common
sudo add-apt-repository ppa:deadsnakes/ppa
sudo apt install python3.7
sudo apt-get install python3.7-dev libpq-dev
   
2. Clone the git repo
   
3. Create a virtual environment with the correct python version. We'll again place the virtualenv inside the project directory.
   sudo apt-get install python-virtualenv
virtualenv --python=/usr/bin/python3.7 .
   
4. Create a system user with limited permissions. Why? Because your server security can be compromised, in which case you don't want the intruder to have root privileges. Here we create a group (I like to use the groupname 'webapps' in all my apps for consistency) and a user (named 'project_user' in this example) assigned to that group, with a home directory in the project directory.
   sudo groupadd --system webapps
sudo useradd --system --gid webapps -shell /bin/bash --home /home/MyProject.com/ project_user
   
   And we make the project_user the ower of the project directory:
   sudo chown project_user:webapps /home/MyProject.com
sudo chmod -R 777 /home/MyProject.com
   
5. Install Django requirements from the requirements.txt file. I use Django-Cookiecutter for my projects so the requirements are usually spread across three files:
   pip install -r requirements/base.txt
pip install -r requirements/local.txt
pip install -r requirements/production.txt
   
6. [Only if your project uses Postgres] Install PostgreSQL and create a database with the name and authentication specified in your Django settings.
   sudo apt-get install postgresql postgresql-contrib
pip install psycopg2

   Create the database, database-user, and set permissions:
   sudo su - postgres
psql
CREATE DATABASE <db_name>;
CREATE USER <db_username> WITH PASSWORD '<db_password>';
ALTER ROLE <db_username> SET client_encoding TO 'utf8';
ALTER ROLE <db_username> SET default_transaction_isolation TO 'read committed';
ALTER ROLE <db_username> SET timezone TO 'Canada/Pacific';
GRANT ALL PRIVILEGES ON DATABASE <db_name> TO <db_username>;
\q
exit
   
   Finally, run python manage.py migrate to build the database, and run python manage.py loaddata database/db.json to populate the database from an existing data dump.
   
   Your Django app should run locally now without a problem!

Step 3 - Gunicorn, Supervisor and Nginx setup

[**Firstly, if you don't know what Gunicorn and Nginx are and why you should use them, read this and this.]

1. Install gunicorn
   pip install gunicorn
   
2. Create a shell script to automatically initialise gunicorn to server the Django project when the server boots up. To do this we create a file named gunicorn_start in the ./bin/ directory in our project. The contents of the file will be:

1. And change the permissions on this file: sudo chmod u+x bin/gunicorn_start
   
2. Install Supervisor. Supervisor is a process manager whose primary role will be to start the gunicorn service when the server boots.
   sudo apt-get install supervisor
   
   Create the supervisor config file myproject.conf inside the etc/supervisor/conf.d folder in your project directory. The file should contain:
   [program:MyProject]
command = /home/MyProject.com/bin/gunicorn_start ; Command to start app
user = project_user ; User to run as
stdout_logfile = /home/MyProject.com/data/logs/gunicorn_supervisor.log ; Where to write log messages
redirect_stderr = true ; Save stderr in the same log
environment=LANG=en_US.UTF-8,LC_ALL=en_US.UTF-8 ; Set UTF-8 as default encoding
   
   Then create the logs folder and file:
   mkdir /home/MyProject.com/data/logs/
touch /home/MyProject.com/data/logs/gunicorn_supervisor.log
   
   Simlink the conf.d file to your system's /etc/supervisor/conf.d folder.
   
3. Install Nginx.
   sudo apt-get install nginx
   
   Create an Nginx config file (MyProject) in the etc/nginx/ folder in your project directory. It should contain:

1. Again, simlink the conf.d file to your system's /etc/supervisor/nginx/sites-enabled folder.
   sudo ln -s /home/MyProject.com/etc/nginx/MyProject /etc/nginx/sites-enabled/MyProject
   

And we are done!

Now the site should be accesible at the server's IP address.

If there are any further problems, check the Gunicorn and Nginx logs. A common problem is with file/directory permissions. Often new files/directories are created with root permission. Simply change their owner to 'project_user'.