Gunicorn + Supervisor not running new code on reload

Question

I've been trying to apply some changes to my Flask app by reloading both gunicorn and supervisor. I've already followed several tutorials using scripts and packages like rainbow-saddle. Nothing seems to work, not even rebooting the entire server.

The flask app is being run in a virtual environment and I suspect this may be the problem.

The following is the startup script for gunicorn and the flask app.

#!/bin/bash

NAME="Simon"                                  
FLASKDIR=/var/www/Simon            
SOCKFILE=/var/www/Simon/simon.sock  
USER=glen                                        
GROUP=glen                                    
NUM_WORKERS=3                                    
FLASK_SETTINGS_MODULE=config.py          
FLASK_WSGI_MODULE=Simon_wsgi.wsgi                    
echo "Starting $NAME as `whoami`"

# Activate the virtual environment
cd $FLASKDIR
source Simon/bin/activate
export FLASK_SETTINGS_MODULE=$FLASK_SETTINGS_MODULE
export PYTHONPATH=$FLASKDIR:$PYTHONPATH

# Create the run directory if it doesn't exist
RUNDIR=$(dirname $SOCKFILE)
test -d $RUNDIR || mkdir -p $RUNDIR

exec gunicorn Simon:simon \
  --name $NAME \
  --workers $NUM_WORKERS \
  --user=$USER --group=$GROUP \
  --bind=unix:$SOCKFILE \
  --log-level=debug \
  --pid=$FLASKDIR/gunicorn.pid
  --log-file=$FLASKDIR/logs/gunicorn.log
  --chdir=/var/www/Simon

This is a snippet of output from the supervisor log.

2017-08-20 19:31:56,789 WARN received SIGHUP indicating restart request
2017-08-20 19:31:56,790 INFO waiting for simon to die
2017-08-20 19:31:56,886 INFO stopped: simon (exit status 0)
2017-08-20 19:31:56,890 INFO supervisord started with pid 1438
2017-08-20 19:31:57,892 INFO spawned: 'simon' with pid 2751
2017-08-20 19:31:59,308 INFO success: simon entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)

I've already tried using a .pid for both supervisor and gunicorn to kill -HUP them. Nothing I do seems to refresh the code that is being used.

EDIT: the conf at /etc/supervisor/supervisord.conf

[supervisord]
logfile=/var/www/Simon/logs/supervisord.log
loglevel=info


[program:simon]
command= bash /var/www/Simon/Simon/bin/gunicorn_start.bash Simon:simon --bind 0.0.0.0
directory=/var/www/Simon
user=glen
autostart=true
autorestart=true
redirect_stderr=true
priority=500
;

[program:simonreload]
autostart=true
autorestart=false
directory=/tmp
command=watchmedo shell-command --patterns="*.py;*.html;*.css;*.js" --recursive -HUP $(cat /var/www/Simon/supervisord.pid)'

See if this helps? https://stackoverflow.com/questions/5754227/gunicorn-not-reloading-a-django-application — Tarun Lalwani, Aug 21 '17 at 07:02
@TarunLalwani Already tried that method. It's still running the old code even though I can see watchdog running in the background. — smallpants, Aug 21 '17 at 17:16
How are you checking the changes of the app? Are you making sure you use a incognito window — Tarun Lalwani, Aug 21 '17 at 19:46
I'm unable to login to my site because I forgot to change the callback URL to the non-localhost version. I changed that in the code, and I still get the same error as before I changed it. — smallpants, Aug 21 '17 at 20:09
Then how can gunicorn be loading old code still? I think you are missing something very small — Tarun Lalwani, Aug 21 '17 at 20:27
Yeah that's why I came here. I've been doing this for three days. — smallpants, Aug 21 '17 at 20:53
Is it possible the request is not reaching your server only? And something else is answering the request? I mean some misconfig? — Tarun Lalwani, Aug 21 '17 at 21:03
No because if I keep nginx enabled and `pkill supervisord` I get a 502 error because no app is started to handle a request. — smallpants, Aug 21 '17 at 23:46

Gunicorn + Supervisor not running new code on reload

0 Answers0