docker ERROR: for nginx Cannot start service nginx: driver failed programming external connectivity on

Question

I'm new to Docker and setting up my first Django application using Docker

My application path looks like

app
 |- helloworld
    |- __init__.py
    |- manage.py
 |- static_cdn
    |- static_root
 |- config
    |- nginx
       |- nginx.conf
 |- Dockerfile
 |- docker-compose.yml
 |- requirements.txt
 |- start.sh

the contents of Docerfile

FROM ubuntu:18.04

# -- Install Pipenv:
FROM python:3
ENV PYTHONUNBUFFERED 1

ENV LC_ALL C.UTF-8
ENV LANG C.UTF-8

# -- Install Application into container:
RUN set -ex && mkdir /app

WORKDIR /app
ADD requirements.txt /app/

RUN pip install -r requirements.txt

# -- Adding dependencies:
ADD . /app/

contents of docker-compose.yml

version: '3'

services:
  nginx:
    image: nginx:latest
    ports:
      - "9010:9010"
    volumes:
      - .:/app
      - ./config/nginx:/etc/nginx/conf.d
      - ./static_cdn:/static
    depends_on:
      - web
  web:
    build: .
    command: ./start.sh
    volumes:
      - .:/app
      - ./static_cdn:/static
    ports:
      - "9010:9010"
    depends_on:
      - db
    expose:
      - "9010"
  db:
    image: postgres

contents of config/nginx/nginx.conf

upstream web {
    ip_hash;
    server web:9010;
}

server {
    location /static {
        autoindex on;
        alias /static/
    }

    location / {
        proxy_pass http://127.0.0.1;
    }
    listen 9011;
    server_name localhost;
}

contents of start.sh

#!/usr/bin/env bash

# Start Gunicorn processes
echo --: Starting application build
echo --: Creating migration
exec python3 manage.py makemigrations
echo ------: makemigrations complete
echo --: Running migration
exec python3 manage.py migrate
echo ------: migrate complete
echo --: Running collectstatic
exec python3 manage.py collectstatic
echo ------: collectstatic complete
echo Starting Gunicorn.
exec gunicorn helloworld.wsgi:application \
    --bind 0.0.0.0:9010 \
    --workers 3

Now, when I build using docker

docker-compose up --build

It gives error as

ERROR: for nginx Cannot start service nginx: driver failed programming external connectivity on endpoint koober_nginx_1 (8ea5c084a7283a16afbf136a73dc4b27d9cae35fe14d735b83199ad5d0e03431): Bind for 0.0.0.0:9010 failed: port is already allocated

I have followed few tutorials to create those Docker files and nginx conf file.

1. How can I solve above issue.
2. Do I need to use FROM ubuntu:18.04 with above configuration?

Edit 2

Now, it stuck after creating migration from start.sh commands

Solution 1

You can't allocate port 9010 of your host for both services. This is what you're doing in the ports section of declaration of service nginx and web.

Moreover, by default nginx will listen to port 80 and 443 for https.

You can keep it like that and publish to a different port on your host. See how to use port keyword in docker-compose :

https://docs.docker.com/compose/compose-file/#ports

Maybe you want something more like that:

version: '3'

services:
  nginx:
    image: nginx:latest
    ports:
      - "10080:80"
      - "10443:443"
    volumes:
      - .:/app
      - ./config/nginx:/etc/nginx/conf.d
      - ./static_cdn:/static
   depends_on:
      - web

  web:
    build: .
    command: ./start.sh
    container_name: "web-app"
    volumes:
      - .:/app
      - ./static_cdn:/static
    expose:
      - "9010"
    depends_on:
      - db

  db: 
    image: postgres

contents of config/nginx/nginx.conf

upstream web {
  ip_hash;
  server web-app:9010;
}

server {
    location /static {
        autoindex on;
        alias /static/
     }

location / {
    proxy_pass http://web;
}

listen 80;
server_name localhost;
}

Concerning your last question, you could go for an official Python image from the Docker hub Python repository or start from any other base image like debian:jessie-slim from Debian official repository or keep the Ubuntu 18.04 image

Solution 2

Feb, 2022 Update:

Kill all nginx processes:

sudo killall nginx

Then:

docker-compose up --build

Comments

[Anuj TBE]

Thanks and I made changes same as your answer but now stuck after a command from start.sh.

[gcharbon]

It's hard to tell with your logs beacause all we can see is that postresql is alright and your application doesn't start. Can't you have more logs?

[Anuj TBE]

I figured out that start.sh file executes only first exec command and stops then after. Is there anything wrong in .bash file? I think it will be solved by making changes to start.sh file only.

[gcharbon]

Why are you using exec in your last line? (exec gunicorn helloworld.wsgi:application \ --bind 0.0.0.0:9010 \ --workers 3) and don't just call unicorn?

[gcharbon]

Maybe you should read stackoverflow.com/questions/43925487/… I don't know much about gunicorn and docs.gunicorn.org/en/stable/deploy.html

[Anuj TBE]

can you tell me at what host it should be working. I tried localhost, localhost:9091, 127.0.0.1, 127.0.0.1:9091 but it's not working

[gcharbon]

It should be working on localhost:10080 or localhost:10443 for https. Remember, you expose your unicorn app port on the docker network. Nginx can redirect to the expose port. But you will always reach nginx and not your unicorn app. So ports published on the port declaration section of service nginx in docker-composemust be used. Only publihed ports are accessible from hosts, not exposed ports. By the way localhost or 127.0.0.1 should be the same. And don't use exec when writing shell scripts, it does not behave the way you think (man7.org/linux/man-pages/man3/exec.3.html)