Installation | Topics | Beyond Lino

Deploying Django channels

This document is obsolete. See: hg.linod

If you plan to activate desktop notifications using Django Channels, you need to need to go through the next steps

Install and activate

Django channels requires to install the channels package using the following command:

$ pip install -U django-channels
$ sudo apt-get -y install redis-server

And activate it for your site by setting use_websockets to True.

Deploy

The above commands are enough to get your desktop notifications works fine in development mode. But for the production server things get more complicated.

In this document, I will use asgi_redis as the redis backend and apache is our webserver.

Three configurations should be done. - Setup worker servers - Setup interface servers - Update Apache configuration

Worker Servers

To run a worker server, just run

python manage.py runworker

You should place this inside an init system or something like supervisord to ensure it is re-run if it exits unexpectedly.

An example for a supervisord could be:

[program:my_runworker]
command=/path/to/my/script/runworker.sh
user = www-data
umask = 002

With runworker.sh like the following:

#!/bin/bash
set -e  # exit on error
. /path/to/my/env/bin/activate
exec python /path/to/manage.py runworker  -v2

Interface servers

Django channels is shipped with Daphne as interface servers. But first we need to create an asgi module in your project file:

import os
from channels.asgi import get_channel_layer

os.environ.setdefault("DJANGO_SETTINGS_MODULE", "my.settings")

channel_layer = get_channel_layer()

Now we can run the server:

daphne my.asgi:channel_layer

If everything goes right, the interface server should start running!

As runworker, Daphne should also be configured with a supervisord. Here an example for daphne.sh

#!/bin/bash
set -e  # exit on error
. /path/to/my/env/bin/activate
exec daphne -b localhost -p 8080 asgi:channel_layer

I have used 8080 as the port of daphne , we will use this port in apache configuration. And for supervisord config:

[program:daphne_jane]
directory=/path/to/script/
command=/path/to/script/daphne.sh
user = www-data
umask = 002

If everything is okay, we need to restart supervisord and our Worker Servers and Interface servers should be running.

Apache configuration

Finally here’s the apache conf that we need to add. Please note that we should redirect all our websocket incoming requests to the daphne server and keep the other requests to get handled by our usual wsgi module.:

RewriteEngine on
RewriteCond %{HTTP:UPGRADE} ^WebSocket$ [NC,OR]
RewriteCond %{HTTP:CONNECTION} ^Upgrade$ [NC]
RewriteRule .* ws://127.0.0.1:8080%{REQUEST_URI} [P,QSA,L]

Note that this config used be used for regardless the http protocol used, http or https. If the mod_rewrite is not activated for your apache server, you could activate is using the following command:

a2enmod rewrite

You also may need to activate the relevant proxy modules if you haven’t already:

a2enmod proxy
a2enmod proxy_http
a2enmod proxy_wstunnel

And then a apache restart is required to load the new apache configuration.

Don’t forget to set your use_websockets to True in your settings.py:

use_websockets = True

And then you should also run:

$ python manage.py buildcache
$ python manage.py collectstatic