Separate processes for Dockerfile for dev and prod? - django

Question

I have a project with the following structure.
ProjectName/
├── Dockerfile
├── api/
│   ├── Dockerfile
│   └── manage.py
├── docker-compose.yml
├── frontend/
│   ├── Dockerfile
│   ├── build/
│   └── src/
└── manifests/
├── development.yml
└── production.yml
docker-compose.yml has a database image that's common between both environments, and the dev.yml and prod.yml have similar but slightly different images for production and dev.
Example: The api dev uses django and just runs python manage.py runserver, but in prod it will run gunicorn api.wsgi.
And the frontend runs npm start but in prod I want it to be based off a different image. Currently the dockerfile only works with one or the other, since the npm command is only found when I use FROM node and the nginx command only shows up when I use FROM kyma/docker-nginx.
So how can I separate these out when in different environments?
./frontend/Dockerfile:
FROM node
WORKDIR /app/frontend
COPY package.json /app/frontend
RUN npm install
EXPOSE 3000
CMD ["npm", "start"]
# Only run this bit in production environment, and not anything above this line.
#FROM kyma/docker-nginx
#COPY build/ /var/www
#CMD 'nginx'
./api/Dockerfile:
FROM python:3.5
RUN apt-get update \
&& apt-get install -y --no-install-recommends \
postgresql-client \
&& rm -rf /var/lib/apt/lists/*
WORKDIR /app/api
COPY requirements.txt /app/api
RUN pip install -r requirements.txt
EXPOSE 8000
# Run this command in dev
CMD ["python", "manage.py", "runserver", "0.0.0.0:8000"]
# Run this command in prod
#CMD ["gunicorn", "api.wsgi", "-b 0.0.0.0:8000"]
./docker-compose.yml:
version: '3'
services:
db:
image: postgres
restart: always
ports:
- "5432:5432"
volumes:
node-modules:
./manifests/production.yml:
version: '3'
services:
gunicorn:
build: ./api
command: ["gunicorn", "api.wsgi", "-b", "0.0.0.0:8000"]
restart: always
volumes:
- ./api:/app/api
ports:
- "8000:8000"
depends_on:
- db
nginx:
build: ./frontend
command: ["nginx"]
restart: always
volumes:
- ./frontend:/app/frontend
- ./frontend:/var/www
- node-modules:/app/frontend/node_modules
ports:
- "80:80"
volumes:
node-modules:
./manifests/development.yml:
version: '3'
services:
django:
build: ./api
command: ["python", "manage.py", "runserver", "0.0.0.0:8000"]
restart: always
volumes:
- ./api:/app/api
ports:
- "8000:8000"
depends_on:
- db
frontend:
build: ./frontend
command: ["npm", "start"]
restart: always
volumes:
- ./frontend:/app/frontend
- node-modules:/app/frontend/node_modules
ports:
- "3000:3000"
volumes:
node-modules:

Answer 1

You could have as an ENTRYPOINT a script running one or the other command depending of an environment variable that you can set at run time:
docker run -e env=DEV
# or
docker run -e env=PROD
You can set that same environment variable in a docker compose file.