Reputation: 701
I'm trying to dockerize my django app with a couple of celery workers. The django app and the celery workers all create the same container - same python image, same pipenv, same packages installed from pipenv, same app. The only difference is that I want one container to run the django app server, and the other container to run my celery workers.
When I run docker-compose up
, docker is copying the app and installing the same python packages once per container. This takes a long time since it's doing the exact same thing 2 times.
I want to know if there is a way I can clone my app and install the packages ONCE, and use this for all 2 containers that would otherwise install the same thing 2 times.
Dockerfile
FROM python:3.5.6
COPY . /app/
WORKDIR /app/
RUN pip install pipenv==2018.11.26
ADD Pipfile Pipfile
RUN pipenv install --deploy --system
EXPOSE 8000
Docker-compose.yml
version: '2'
services:
app:
restart: always
build: .
expose:
- "8000"
container_name: "app"
image: debian/latest
links:
- postgres
- redis
depends_on:
- postgres
- redis
ports:
- '8000:8000'
networks:
- network1
- nginx_network
volumes:
- ./:/app
- ./data:/app/data
- static_volume:/app/static
- ./logs:/app/logs
entrypoint: ["sh", "/app/docker-entrypoint.sh"]
env_file:
- .env
environment:
- DJANGO_SETTINGS_MODULE=app.settings.production
celery_default:
restart: always
build: .
container_name: "celery_default"
networks:
- network1
links:
- redis
- postgres
depends_on:
- postgres
- redis
- celerybeat
volumes:
- ./:/app
- ./data:/app/data
- ./logs:/app/logs
- ./celery:/app/celery
env_file:
- .env
entrypoint: "celery -A app worker -Q celery -l debug -n celery_worker --concurrency=2 --logfile=./celery/logs/default.log"
Upvotes: 0
Views: 617
Reputation: 13106
What I would do is define in your compose
an image, and the other application uses that image:
version: '2'
services:
app:
restart: always
build: .
image: your-custom-image # Notice I've created a custom image tag here
expose:
- "8000"
container_name: "app"
links:
- postgres
- redis
depends_on:
- postgres
- redis
ports:
- '8000:8000'
networks:
- network1
- nginx_network
volumes:
- ./:/app
- ./data:/app/data
- static_volume:/app/static
- ./logs:/app/logs
entrypoint: ["sh", "/app/docker-entrypoint.sh"]
env_file:
- .env
environment:
- DJANGO_SETTINGS_MODULE=app.settings.production
celery_default:
restart: always
image: your-custom-image # No build directory, just reuse that image
container_name: "celery_default"
networks:
- network1
links:
- redis
- postgres
depends_on:
- postgres
- redis
- celerybeat
volumes:
- ./:/app
- ./data:/app/data
- ./logs:/app/logs
- ./celery:/app/celery
env_file:
- .env
entrypoint: "celery -A app worker -Q celery -l debug -n celery_worker --concurrency=2 --logfile=./celery/logs/default.log"
This way you are only build
ing once, and the other app just uses that built image
Upvotes: 1