[Django]-How to make Django serve static files with Gunicorn?

Whitenoise

Post v4.0

http://whitenoise.evans.io/en/stable/changelog.html#v4-0

The WSGI integration option for Django (which involved editing wsgi.py) has been removed. Instead, you should add WhiteNoise to your middleware list in settings.py and remove any reference to WhiteNoise from wsgi.py. See the documentation for more details.
(The pure WSGI integration is still available for non-Django apps.)

Pre v4.0

Heroku recommends this method at: https://devcenter.heroku.com/articles/django-assets:

Your application will now serve static assets directly from Gunicorn in production. This will be perfectly adequate for most applications, but top-tier applications may want to explore using a CDN with Django-Storages.

Install with:

pip install whitenoise
pip freeze > requirements.txt

wsgi.py:

import os
from django.core.wsgi import get_wsgi_application
from whitenoise.django import DjangoWhiteNoise

os.environ.setdefault("DJANGO_SETTINGS_MODULE", "free_books.settings")
application = get_wsgi_application()
application = DjangoWhiteNoise(application)

Tested on Django 1.9.

I’ve used this for my development environment (which uses gunicorn):

from django.conf import settings
from django.contrib.staticfiles.handlers import StaticFilesHandler
from django.core.wsgi import get_wsgi_application


if settings.DEBUG:
    application = StaticFilesHandler(get_wsgi_application())
else:
    application = get_wsgi_application()

And then run gunicorn myapp.wsgi. This works similar to @rantanplan’s answer, however, it does not run any middleware when running static files.

The gunicorn should be used to serve the python "application" itself, while the static files are served by a static file server ( such as Nginx ).

This is an excerpt from one of my configurations:

upstream app_server_djangoapp {
    server localhost:8000 fail_timeout=0;
}

server {
    listen < server port goes here >;
    server_name < server name goes here >;

    access_log  /var/log/nginx/guni-access.log;
    error_log  /var/log/nginx/guni-error.log info;

    keepalive_timeout 5;

    root < application root directory goes here >;

    location /static {    
        autoindex on;    
        alias < static folder directory goes here >;    
    }

    location /media {
       autoindex on;
       alias < user uploaded media file directory goes here >;
    }

    location / {
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header Host $http_host;
        proxy_redirect off;

        if (!-f $request_filename) {
            proxy_pass http://app_server_djangoapp;
            break;
        }
    }
}

Some notes:

• The static root, media root, static files path prefix and media file path prefix are set up in your settings.py
• Once you have nginx set up to serve from the static content directory, you need to run "python manage.py collectstatic" in your project root so that the static files in the various apps can be copied to the static folder

In closing: while it is possible to serve static files from gunicorn ( by enabling a debug-only static file serving view ), that is considered bad practice in production.

While you’re able to use Django to serve static files in development mode with the app django.contrib.staticfiles, that is not suitable for production use.

In order to serve static files, as Jamie Hewland says, normally one routes all the requests to /static/ using Nginx

location /static/ {

    alias /path/to/static/files;

}

Also, and as coreyward says about Gunicorn / Unicorn

was not designed to solve the suite of problems involved in serving
files to clients

Same line of reasoning applies if you consider other WSGI server like uWSGI instead of Gunicorn. In uWSGI documentation

it’s inefficient to serve static files via uWSGI. Instead, serve them
directly from Nginx and completely bypass uWSGI

So, serving static files in production is something to be done by NGINX, not by Django or other WSGI servers.

Another way is to serve your static files in production is using WhiteNoise library which is very easy to setup (you might want to use a CDN so that most requests won’t reach the Python app). As Miguel de Matos says, you just have to

1. Collect static

    python manage.py collectstatic
   

2. Installing whitenoise

    pip install whitenoise
   

3. Add the following STATICFILES_STORAGE in settings.py

    STATICFILES_STORAGE = 'whitenoise.storage.CompressedManifestStaticFilesStorage'
   

4. Add the following to your MIDDLEWARE in settings.py (as mracette notes, "According to the whitenoise docs you should place the middleware after django.middleware.security.SecurityMiddleware")

    `MIDDLEWARE = [
      'django.middleware.security.SecurityMiddleware',
      'whitenoise.middleware.WhiteNoiseMiddleware',
      ...
    ]
   

Since Django 1.3 there is django/conf/urls/static.py that handle static files in the DEBUG mode:

from django.conf import settings
from django.conf.urls.static import static

urlpatterns = [
    # ... the rest of your URLconf goes here ...
] + static(settings.MEDIA_URL, document_root=settings.MEDIA_ROOT)

Read more https://docs.djangoproject.com/en/2.0/howto/static-files/#serving-static-files-during-development

If you are using Apache/Gunicorn then here is how I set mine up.

1. In your Django root dir (with manage.py), create dirs mkdir -p django_static/static

2. In your project settings.py set the following:

DEBUG = False
INSTALLED_APPS = [..., 'django.contrib.staticfiles', ...]
STATIC_URL = 'static/'
STATIC_ROOT = os.path.join(BASE_DIR, "django_static", "static")

3. Run python manage.py collectstatic. This will output static content to django_static/static

4. Start your gunicorn server with gunicorn your_project_name.wsgi (plus options)

5. Assuming you have default global Apache settings, you’ll need to create a soft link from /var/www to your static dir:
   sudo ln -s /path/to/your_django_project/django_static /var/www/your_django_project_static

6. For your domain www.example.com that you wish to point to your Django app, configure the following virtual host in apache in order to proxy all requests submitted to https://www.example.com onto 127.0.0.1:8000 except for www.example.com/static/ routes (in which case, serve files to such requests from django_static):

<VirtualHost *:443>
    ServerName www.example.com
    DocumentRoot /var/www/your_django_project_static
    <Location "/">
        ProxyPreserveHost On
        ProxyPass http://127.0.0.1:8000/
        ProxyPassReverse http://127.0.0.1:8000/
    </Location>
    <Location "/static/">
        ProxyPass "!"
    </Location>
</VirtualHost>

Voila!