project wsgi not found when started automatically but not manually - django

Question

I do not understand why I get an error about not finding my "project.wsgi" module when supervisor tries to start the app automatically (for example when the server is starting.)
2014-02-15 05:13:05 [1011] [INFO] Using worker: sync
2014-02-15 05:13:05 [1016] [INFO] Booting worker with pid: 1016
2014-02-15 05:13:05 [1016] [ERROR] Exception in worker process:
Traceback (most recent call last):
File "/var/local/sites/myproject/venv/local/lib/python2.7/site-packages/gunicorn/arbiter.py", line 495, in spawn_worker
worker.init_process()
File "/var/local/sites/myproject/venv/local/lib/python2.7/site-packages/gunicorn/workers/base.py", line 106, in init_process
self.wsgi = self.app.wsgi()
File "/var/local/sites/myproject/venv/local/lib/python2.7/site-packages/gunicorn/app/base.py", line 114, in wsgi
self.callable = self.load()
File "/var/local/sites/myproject/venv/local/lib/python2.7/site-packages/gunicorn/app/wsgiapp.py", line 62, in load
return self.load_wsgiapp()
File "/var/local/sites/myproject/venv/local/lib/python2.7/site-packages/gunicorn/app/wsgiapp.py", line 49, in load_wsgiapp
return util.import_app(self.app_uri)
File "/var/local/sites/myproject/venv/local/lib/python2.7/site-packages/gunicorn/util.py", line 354, in import_app
__import__(module)
ImportError: No module named myproject.wsgi
Whereas I do not get this error and it works fine when I manually do:
sudo supervisorctl start myapp
What is different?
Thanks
UPDATE:
supervisor conf file:
[program:myproject]
command=/var/local/sites/myproject/run/gunicorn_start ; Command to start app
user=myproject ; User to run as
autostart=true
autorestart=true
loglevel=info
redirect_stderr=false
stdout_logfile=/var/local/sites/myproject/logs/supervisor-myproject-stdout.log
stdout_logfile_maxbytes=1MB
stdout_logfile_backups=10
stdout_capture_maxbytes=1MB
stderr_logfile=/var/local/sites/myproject/logs/supervisor-myproject-stderr.log
stderr_logfile_maxbytes=1MB
stderr_logfile_backups=10
stderr_capture_maxbytes=1MB
/var/local/sites/myproject/run/gunicorn_start:
#!/bin/bash
NAME="myproject_app" # Name of the application
USER=myproject # the user to run as
GROUP=myproject # the group to run as
NUM_WORKERS=3 # how many worker processes should Gunicorn spawn
# Logs config
LOG_LEVEL=info
ACCESS_LOGFILE=/var/local/sites/myproject/logs/gunicorn-myproject-access.log
ERROR_LOGFILE=/var/local/sites/myproject/logs/gunicorn-myproject-error.log
echo "Starting $NAME"
exec envdir /var/local/sites/myproject/env_vars /var/local/sites/myproject/venv/bin/gunicorn myproject.wsgi:application \
--name $NAME \
--workers $NUM_WORKERS \
--user=$USER --group=$GROUP \
--log-level=$LOG_LEVEL \
--bind=unix:/tmp/myproject.gunicorn.sock \
--access-logfile=$ACCESS_LOGFILE \
--error-logfile=$ERROR_LOGFILE

Answer 1

I think you should add directory to your supervisor configuration file. This is my template. I use this in every project and works fine:
[program:PROJECT_NAME]
command=/opt/sites/PROJECT_NAME/env/bin/gunicorn -c /opt/sites/etc/gunicorn/GUNICORN_CONF.conf.py PROJECT_NAME.wsgi:application
directory=/opt/sites/PROJECT_NAME
environment=PATH="/opt/sites/PROJECT/env/bin"
autostart=true
autorestart=false
redirect_stderr=True
stdout_logfile=/tmp/PROJECT_NAME.stdout

Answer 2

I have the same problem before. I'm using the following line in my gunicorn_start instead of envdir. I'm running a django application within a virtual env located in /env/nafd/ and my django app is located in /env/nafd/nafd_proj
..
DJANGODIR=/to/path/app_proj
cd $DJANGODIR
source ../bin/activate`
exec ../bin/gunicorn nafd_proj.wsgi:application \
--name $NAME \
--workers $NUM_WORKERS \
--user=$USER --group=$GROUP \
--log-level=$LOG_LEVEL \
--bind=unix:/tmp/myproject.gunicorn.sock \
--access-logfile=$ACCESS_LOGFILE \
--error-logfile=$ERROR_LOGFILE`

Answer 3

It's obvious, but it is worth to mention:
check if "supervisord" daemon is running (service supervisor status).
Here is the following setup that I have here, using a Flask app with WSGI (Gunicorn), controlled via Supervisor, and it is working perfectly.
Flask App
root#ilg40:~# ll /etc/tdm/flask/
total 1120
drwx------ 5 root root 4096 Jan 24 19:47 ./
drwx------ 3 root root 4096 Jan 23 00:20 ../
-r-------- 1 root root 1150 Aug 31 17:54 favicon.ico
drw------- 2 root root 4096 Jan 13 22:51 static/
-rw------- 1 root root 883381 Jan 23 20:09 tdm.log
-rwx------ 1 root root 73577 Jan 23 21:37 tdm.py*
-rw------- 1 root root 56445 Jan 23 21:37 tdm.pyc
drw------- 2 root root 4096 Jan 23 20:08 templates/
-rw-r--r-- 1 root root 493 Jan 23 22:42 wsgi.py
-rw-r--r-- 1 root root 720 Jan 23 22:42 wsgi.pyc
srwxrwxrwx 1 root root 0 Jan 24 19:47 wsgi.sock=
Supervisor Config File
root#ilg40:~# cat /etc/supervisor/conf.d/wsgi_flask.conf
[program:wsgi_flask]
command = gunicorn --preload --bind unix:/etc/tdm/flask/wsgi.sock --workers 4 --pythonpath /etc/tdm/flask wsgi
process_name = wsgi_flask
autostart = true
autorestart = true
stdout_logfile = /var/log/wsgi_flask/wsgi_flask.out.log
stderr_logfile = /var/log/wsgi_flask/wsgi_flask.err.log
Update Supervisord About The New Process
root#ilg40:~# supervisorctl update
wsgi_flask: added process group
Checking Process Status
root#ilg40:~# supervisorctl status wsgi_flask
wsgi_flask RUNNING pid 1129, uptime 0:29:12
Note: in the setup above, I'm not using virtualenv, which I believe that could require the configuration of directory variable for the process and also, to configure environment PATH for the command variable (adding env PATH="$PATH:/the/app/path" gunicorn...), since that gunicorn, flask, and so on, are only located inside of the virtualenv.