docker-baseimage-django

Base image for Django projects.

Usage

You can inherit from this base image for your Django project.

It supposes that the project is laid down with the following structure:

├── docker-compose.yml
├── .gitignore
└── webapp                  # The django app
    ├── conf                # Settings go here
    │   ├── __init__.py     # Make it a module
    │   ├── base.py         # Base settings
    │   ├── development.py  # Development overrides
    │   ├── production.py   # Production specific stuff
    │   ├── urls.py         # Root urlconf
    │   └── wsgi.py         # WSGI app (for gunicorn)
    ├── Dockerfile
    ├── .dockerignore
    ├── entrypoint-pre.sh   # See entrypoint
    ├── requirements.txt    # App requirements
    └── src
        └── myproject       # The app(s) egg(s)

An example Dockerfile for a custom project might look like this:

FROM abstracttechnology/django:latest
MAINTAINER Foo Bar <[email protected]>

COPY entrypoint-pre.sh entrypoint-pre.sh

COPY requirements.txt requirements.txt
COPY src/ src/

USER root

RUN chown -R webapp:webapp .

USER webapp

RUN virtualenv . && \
    ./bin/pip install --upgrade pip && \
    ./bin/pip install -r requirements.txt && \
    cd src/myproject && \
    ../../bin/python setup.py develop

COPY conf/ conf/
USER root
RUN chown -R webapp:webapp .
USER webapp

This will end up configuring a container with your app django app in it, with an entrypoint that offers the following base commands:

• run to run the app with Gunicorn (4 sync workers)
• manage to launch manage.py commands (arguments are passed to the manage script)

These two commands are launched using conf/production.py as settings module (which might extend from a common conf/base.py module).

There are other two commands that will instead run using conf/development.py, develop-run and develop-manage.

If a command is not recognized, it executed as-is (this allows you to launch other programs, or a shell for debug).

If you need any extra steps to be executed any time the container starts, before any command, add a entrypoint-pre.sh to the root (as done in the example).

For example, if your postgres container is slow to boot up, you can put in that file:

source check_up.bash
check_up "postgresql" ${POSTGRES_PORT_5432_TCP_ADDR} 5432

Supposing your container is called postgres, if it's called database change POSTGRES_PORT_5432_TCP_ADDR to DATABASE_PORT_5432_TCP_ADDR.

Similarly, if you want to handle extra "commands" in the entrypoint, you can add a entrypoint-extras.sh file, with a construct like this:

case $1 in
    foo)
        echo "Foo"
        ;;
    bar)
        echo "Bar"
        ;;
esac

Remember to always make .sh files executable.

And that's all, happy Django-ing.