Deploying a minimal flask app in docker - server connection issues

ghz 1years ago ⋅ 5062 views

Question

I have an app whose only dependency is flask, which runs fine outside docker and binds to the default port 5000. Here is the full source:

from flask import Flask

app = Flask(__name__)
app.debug = True

@app.route('/')
def main():
    return 'hi'

if __name__ == '__main__':
    app.run()

The problem is that when I deploy this in docker, the server is running but is unreachable from outside the container.

Below is my Dockerfile. The image is ubuntu with flask installed. The tar just contains the index.py listed above;

# Dockerfile
FROM dreen/flask
MAINTAINER dreen
WORKDIR /srv

# Get source
RUN mkdir -p /srv
COPY perfektimprezy.tar.gz /srv/perfektimprezy.tar.gz
RUN tar x -f perfektimprezy.tar.gz
RUN rm perfektimprezy.tar.gz

# Run server
EXPOSE 5000
CMD ["python", "index.py"]

Here are the steps I am doing to deploy

$> sudo docker build -t perfektimprezy .

As far as I know the above runs fine, the image has the contents of the tar in /srv. Now, let's start the server in a container:

$> sudo docker run -i -p 5000:5000 -d perfektimprezy
1c50b67d45b1a4feade72276394811c8399b1b95692e0914ee72b103ff54c769

Is it actually running?

$> sudo docker ps
CONTAINER ID        IMAGE                   COMMAND             CREATED             STATUS              PORTS                    NAMES
1c50b67d45b1        perfektimprezy:latest   "python index.py"   5 seconds ago       Up 5 seconds        0.0.0.0:5000->5000/tcp   loving_wozniak



$> sudo docker logs 1c50b67d45b1
    * Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)
    * Restarting with stat

Yep, seems like the flask server is running. Here is where it gets weird. Lets make a request to the server:

$> curl 127.0.0.1:5000 -v
* Rebuilt URL to: 127.0.0.1:5000/
* Hostname was NOT found in DNS cache
*   Trying 127.0.0.1...
* Connected to 127.0.0.1 (127.0.0.1) port 5000 (#0)
> GET / HTTP/1.1
> User-Agent: curl/7.35.0
> Host: 127.0.0.1:5000
> Accept: */*
>
* Empty reply from server
* Connection #0 to host 127.0.0.1 left intact
curl: (52) Empty reply from server

Empty reply... But is the process running?

$> sudo docker top 1c50b67d45b1
UID                 PID                 PPID                C                   STIME               TTY                 TIME                CMD
root                2084                812                 0                   10:26               ?                   00:00:00            python index.py
root                2117                2084                0                   10:26               ?                   00:00:00            /usr/bin/python index.py

Now let's ssh into the server and check...

$> sudo docker exec -it 1c50b67d45b1 bash
root@1c50b67d45b1:/srv# netstat -an
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address           Foreign Address         State
tcp        0      0 127.0.0.1:5000          0.0.0.0:*               LISTEN
tcp        0      0 127.0.0.1:47677         127.0.0.1:5000          TIME_WAIT
Active UNIX domain sockets (servers and established)
Proto RefCnt Flags       Type       State         I-Node   Path
root@1c50b67d45b1:/srv# curl -I 127.0.0.1:5000
HTTP/1.0 200 OK
Content-Type: text/html; charset=utf-8
Content-Length: 5447
Server: Werkzeug/0.10.4 Python/2.7.6
Date: Tue, 19 May 2015 12:18:14 GMT

It's fine... But not from the outside.
What am I doing wrong?

Answer

The problem is you are only binding to the localhost interface, you should be binding to 0.0.0.0 if you want the container to be accessible from outside. If you change:

if __name__ == '__main__':
    app.run()

to

if __name__ == '__main__':
    app.run(host='0.0.0.0')

It should work.

Note that this will bind to all interfaces on the host, which may in some circumstances be a security risk - see https://stackoverflow.com/a/58138250/4332 for more information on binding to a specific interface.

#flask