Install python packages from containerized pypiserver into containerized flask app

Question

I have two docker containers. First one is pypiserver, that contains a package I've created. Second one is my flask app that will install that package from pypiserver. I build those containers with docker-compose, and after that I go into the app container and install that package. It works fine. However, when I tried to install that package in Dockerfile, while building the app, it does not work.

This is my docker-compose.yaml file:

version: '3.9'

services:
  test-pypiserver:
    image: pypiserver/pypiserver:latest
    ports:
      - 8090:8080
    volumes:
      - ./pypiserver/packages:/data/packages
    networks:
      - test-version-2-network
  test-flask:
    build: ./dashboard/.
    container_name: test-flask
    ports:
      - 5000:5000
    volumes:
      - ./dashboard:/code
    depends_on:
      - test-pypiserver
    networks:
      - test-version-2-network

This is my Dockerfile for my flask app:

FROM python

WORKDIR /code

ENV FLASK_APP=app.py
ENV FLASK_RUN_HOST=0.0.0.0
ENV FLASK_RUN_PORT=5000
ENV FLASK_DEBUG=1

COPY requirements.txt requirements.txt

RUN pip install -r requirements.txt
RUN pip install --trusted-host test-pypiserver --extra-index-url http://test-pypiserver:8080 osh

COPY . .

EXPOSE 5000

CMD [ "flask", "run" ]

When I command out this line from Dockerfile

pip install --trusted-host test-pypiserver --extra-index-url http://test-pypiserver:8080 osh

and use it in app container, it works properly

Is there any way to do that? Or is it the proper way to install my packages?

Answer 1

docker-compose up command first build containers and then (after that all containers are builded) make them running. When he try to build your flask application the pypiserver is not running already, so the package installation fails.

You can try to install package when the container is starting.

CMD [ "/bin/sh", "-c", "pip install --trusted-host test-pypiserver --extra-index-url http://test-pypiserver:8080 osh; flask run" ]