all. Got a weird problem in running gunicorn with supervisor to serve a flask application. Before I start stating my problem, just want to say that I'm aware of this SO post. Nevertheless, my problem was that I could run gunicorn directly in the foreground to server the app. The problem only occurred when I used supervisor.
Here's the structure of my flask application:
myproject/
|---webapp/
| |---__init__.py # create_app function is here
| |--- ... # other files in webapp
|---wsgi.py
|---manage.py
|---venv # gunicorn is in venv/bin
The wsgi.py file looks like the following
from webapp import create_app
app = create_app()
if __name__ == "__main__":
app.run()
Sitting in myproject directory, I can launch the app by running
$ venv/bin/gunicorn -w 4 wsgi:app
Then I created a webapp.conf in /etc/supervisor/conf.d/ directory, as follows.
[program:webapp]
directory=/home/<my username>/projects/myproject
command=/home/<my username>/projects/myproject/venv/bin/gunicorn -w 4 wsgi:app
user=<my username>
autostart=true
autorestart=true
redirect_stderr=true
stdout_logfile=<some path>
stderr_logfile=<some path>
When I reloaded my supervisor, the stdout log file said
usage: gunicorn [OPTIONS] [APP_MODULE] gunicorn: error: No application module specified.
My question is, why do I get an error here if I don't get an error when running gunicorn directly? What's the difference?
By the way, I then tried to use systemd to run the gunicorn following this post, and systemd can launch the gunicorn service normally.
Any idea what I did wrong with supervisor? Thanks very much.
