It does not work on production server!

Hello! I do like channels - it is good solution. I wrote a small application with channels - it works fine on localhost (without any problems).

Options:

• Django 1.10
• Linux Debian 8.5
• Python 3.4

But problems arose when deploying: In my settings.py

...
INSTALLED_APPS = [
    ...
    'channels',
    ...
]
...
CHANNEL_LAYERS = {
    'default': {
        'BACKEND': 'asgi_redis.RedisChannelLayer',
        'CONFIG': {'hosts': [('localhost', 6379), ], },
        'ROUTING': 'basic.channels.routing',
    },
}
...

I will repeat - it works fine on localhost!!! For deploy i create asgi.py:

import os
from channels.asgi import get_channel_layer

os.environ.setdefault('DJANGO_SETTINGS_MODULE', 'basic.settings')
application = get_channel_layer()

Previously, I run gunicorn used wsgi.py, but now I stopped gunicorn and run daphne: $ daphne -b 0.0.0.0 -p 3355 basic.asgi:application and workers: $ ./manage.py runworker

The site runs and works fine! But the page with socket has an error: WebSocket connection to 'ws://example.com/some/path/' failed: Unexpected response code: 200

I do not know - how to solve this problem? Please, show me your settings, and method that you run the Django project!

P.p.s Redis work too, I use redis for session and for page view count (port/host are correct).

[andrewgodwin]

You must run two processes in production (runserver combines these for you normally):

• daphne -b 0.0.0.0 -p 8001 yourproject.asgi:channel_layer - The webserver that receives connections. Point Nginx to this. More docs here: https://github.com/django/daphne
• manage.py runworker - The Django worker. You can run as many of these as you like; they connect to Daphne via the channel layer.

You also need to not be using the in-memory channel layer.

[bastbnl]

Have you started the workers as well? You start them via manage.py runworker