Flask container is not up and running using docker - python-3.x

Question

So my issue is very simple, but still can't get hold of it and it's not performing like I wanted to.
sample docker file:
FROM ubuntu:16.04
RUN apt-get update -y && \
apt-get install -y python3-pip python3-dev
COPY ./requirements.txt /requirements.txt
WORKDIR /
RUN pip3 install -r requirements.txt
COPY . /
RUN chmod a+x start.sh
EXPOSE 5000
CMD ["./start.sh"]
sample start.sh
#!/usr/bin/env bash
# sleep 600
nohup python3 /code/app.py &
python3 /code/helloworld_extract.py
sample flask app.py
from flask import Flask
app = Flask(__name__)
#app.route("/")
def index():
return """
<h1>Python Flask in Docker!</h1>
<p>A sample web-app for running Flask inside Docker.</p>
"""
if __name__ == "__main__":
app.run(debug=True, host='0.0.0.0')
So, my issue is as soon as I build the image and run it,
docker run --name flaskapp -p5000:5000 docker-flask:latest... I can't reach localhost:5000.
While if I get inside the container and run explict nohup command with python3 app.py. I can reach localhost.
So, why can't I reach the localhost host with run command?
The thing is I need to run 2 scripts one is flask and another one is helloworld_extract.py which eventually exit after writing some information to the files.

Answer 1

When your start.sh script says
#!/bin/sh
do_some_stuff_in_the_background &
some_foreground_process
The entire lifecycle of the container is tied to the some_foreground_process. In your case, since you clarify that it's doing some initial data load and exits, once it exits, the start.sh script is finished, and so the container exits.
(As a general rule, try to avoid nohup and & in Docker land, since it leads to confusing issues like this.)
I would suggest making the main container process be only the Flask server.
CMD ["python3", "/code/app.py"]
You don't say what's in the loader script. Since its lifecycle is completely different from the main application, it makes sense to run it separately; you can replace the CMD with docker run options. Say you need to populate some shared data in the filesystem. You can:
# Build the image
docker build -t myimage .
# Create a (named) shared filesystem volume
docker volume create extract
# Start the Flask server
docker run -d -p 5000:5000 -v extract:/data myimage
# Run the script to prepopulate the data
docker run -v extract:/data myimage python3 /code/helloworld_extract.py
Notice that the same volume name extract is used in all the commands. The path name /data is an arbitrary choice, though since both commands run on the same image it makes sense that they'd have the same filesystem layout.