When using Nginx with wsgi protocol, workers end with "WORKER TIMEOUT"

[braiam]

With the following app:

def app(environ, start_response):
  data = b"Hello, World!\n"
  start_response("200 OK", [
    ("Content-Type", "text/plain"),
    ("Content-Length", str(len(data)))
  ])
  return iter([data])

and configuring nginx this way (in sites-enabled/<file>):

upstream app_server {
    server 127.0.0.1:8000 fail_timeout=0;
}
server {
    listen 80;
 
    keepalive_timeout 5;

    location / {
      include /etc/nginx/uwsgi_params;
      uwsgi_pass app_server;
      uwsgi_param Host $host;
      uwsgi_param X-Real-IP $remote_addr;
      uwsgi_param X-Forwarded-For $proxy_add_x_forwarded_for;
      uwsgi_param X-Forwarded-Proto $http_x_forwarded_proto;
    }
}

Starting gunicorn with gunicorn wsgi:app --bind localhost:8000, results in the following:

$ gunicorn wsgi:app --bind localhost:8000                  
[2024-09-16 11:15:34 -0400] [780833] [INFO] Starting gunicorn 23.0.0
[2024-09-16 11:15:34 -0400] [780833] [INFO] Listening at: http://127.0.0.1:8000 (780833)
[2024-09-16 11:15:34 -0400] [780833] [INFO] Using worker: sync
[2024-09-16 11:15:34 -0400] [780834] [INFO] Booting worker with pid: 780834
[2024-09-16 11:16:08 -0400] [780833] [CRITICAL] WORKER TIMEOUT (pid:780834)
[2024-09-16 11:16:08 -0400] [780834] [INFO] Worker exiting (pid: 780834)
[2024-09-16 11:16:08 -0400] [780833] [ERROR] Worker (pid:780834) exited with code 1
[2024-09-16 11:16:08 -0400] [780833] [ERROR] Worker (pid:780834) exited with code 1.
[2024-09-16 11:16:08 -0400] [781161] [INFO] Booting worker with pid: 781161

Log level debug doesn't reveal more information about why the worker died. Using an http request directly to the port 8000 it works correctly.

This is related to #3207

[braiam]

As an aside, I tested this same configuration with uwsgi using uwsgi --plugin /usr/lib/uwsgi/plugins/python312_plugin.so -s localhost:8000 --wsgi-file wsgi.py and it worked correctly.

[braiam]

The uwsgi_params file has this content:

uwsgi_param  QUERY_STRING       $query_string;
uwsgi_param  REQUEST_METHOD     $request_method;
uwsgi_param  CONTENT_TYPE       $content_type;
uwsgi_param  CONTENT_LENGTH     $content_length;

uwsgi_param  REQUEST_URI        $request_uri;
uwsgi_param  PATH_INFO          $document_uri;
uwsgi_param  DOCUMENT_ROOT      $document_root;
uwsgi_param  SERVER_PROTOCOL    $server_protocol;
uwsgi_param  REQUEST_SCHEME     $scheme;
uwsgi_param  HTTPS              $https if_not_empty;

uwsgi_param  REMOTE_ADDR        $remote_addr;
uwsgi_param  REMOTE_PORT        $remote_port;
uwsgi_param  SERVER_PORT        $server_port;
uwsgi_param  SERVER_NAME        $server_name;

[braiam]

I used the reference implementation of wsgiref with the small example in the documentation which hinted at the problem. I was thoroughly confused by Nginx documentation that configures uwsgi binary protocol, which gunicorn doesn't support. It is still a good method to trigger the related timeout issue.