I have deployed Django with Gunicorn and NGINX, and it works fine, if the Django app is served on the root url, with this configuration:
server {
listen 80;
location = /favicon.ico { access_log off; log_not_found off; }
location / {
include proxy_params;
proxy_pass http://unix:my_app.sock;
}
}
However when I try to serve the Django app on another URL, it doesn't work. If I try to access http://domain/my_app/admin/ then Django tells me it cannot find the view.
This is the NGINX config:
server {
listen 80;
location = /favicon.ico { access_log off; log_not_found off; }
location /my_app {
include proxy_params;
proxy_pass http://unix:/var/my_app/app.sock;
}
}
How could I make this work? I was not able to find any solution to specify something like a "BASE_URL" so far.
My comment doesn't show the whole picture. When I've run Django sites on subfolders, I like to use a dynamic config so that you can still access the machine directly (without the proxy) and have a working web-app. This can help a LOT for debugging tricky stuff like this that is hard to reproduce in dev.
If you do not have the ability to pass the header or modify wsgi.py, you can still set
FORCE_SCRIPT_NAME
in your Django settings.3 steps:
X-Script-Name
header so that your Django site generates its urls with/myapp/
infront of them -- make sure you're using the{% url %}
tag andreverse
, vs. hard-coding!SCRIPT_NAME
(based on this flask snippet)Here is an example Nginx config for a proxy that points to a Django site on a subdirectory and also sets X-Script-Name (steps 1 and 2), note that this doesn't use a unix socket, so it's a little different from the OP's question. Would welcome an edit:
nginx.conf
And to read X-Script-Name:
myapp/wsgi.py