Skip to content

Latest commit

 

History

History
80 lines (51 loc) · 2.65 KB

DEPLOY.README.md

File metadata and controls

80 lines (51 loc) · 2.65 KB
Raw

Deployment

With docker-compose ?

$ docker-compose build
$ docker-compose -f production.yml up -d
$ docker-compose run web ./manage.py migrate
$ docker-compose run web ./manage.py collectstatic

Principles

  • Execute the application directly via the container. For the time being, we are not using a container management layer.
  • Specify /dev/shm as Gunicorn's worker-tmp-dir for performance reasons. For details, see: https://pythonspeed.com/articles/gunicorn-in-docker/
  • Execute via a socket file for deployment simplicity. This may not be performant on non-linux systems.
  • Run multiple gunicorn workers in deployment

collect static files

Note: STATIC_TMPDIR in the env variables is actually the STATIC_ROOT, and should be set to:

STATIC_TMPDIR=mnt/static/cityhallmonitor/

That is, static files should be placed into a project-specific directory in order to facilitate proxying to multiple applications.

$ sudo docker run -it --env-file=.env -v /home/apps/sites/cityhallmonitor/mnt:/usr/src/app/mnt cityhallmonitor ./manage.py collectstatic

Running the application

Execute gunicorn via a daemonized container (do this from the systemctl config file):

$ sudo docker run -d --env-file=.env -v /home/apps/sites/cityhallmonitor/mnt:/usr/src/app/mnt --name cityhallmonitor cityhallmonitor gunicorn --workers 3 --worker-tmp-dir /dev/shm --bind unix:/usr/src/app/mnt/cityhallmonitor.sock core.wsgi:application

To stop the container:

$ sudo docker stop cityhallmonitor

Troubleshooting

Replace -d with -it to execute in interactive mode and to see application logs:

$ sudo docker run -it --env-file=.env -v /home/apps/sites/cityhallmonitor/mnt:/usr/src/app/mnt cityhallmonitor gunicorn --worker-tmp-dir /dev/shm --bind unix:/usr/src/app/mnt/cityhallmonitor.sock core.wsgi:application

Alternatively, execute a bash shell and run gunicorn from inside the container:

$ sudo docker run -it --env-file=.env -v /home/apps/sites/cityhallmonitor/mnt:/usr/src/app/mnt cityhallmonitor sh

# gunicorn --worker-tmp-dir /dev/shm --bind unix:/usr/src/app/mnt/cityhallmonitor.sock core.wsgi:application

To run gunicorn on a port and bypass Nginx:

$ sudo docker run -it --env-file=.env -p 8000:8000 cityhallmonitor gunicorn -b :8000 core.wsgi:application

References

https://realpython.com/django-development-with-docker-compose-and-machine/

https://pythonspeed.com/articles/gunicorn-in-docker/

https://www.digitalocean.com/community/tutorials/how-to-build-a-django-and-gunicorn-application-with-docker

http://whitenoise.evans.io/en/stable/

https://www.rockyourcode.com/setup-a-postgres-database-with-docker-compose/