Introduction

After creating a Flask application you probably want people to see it. This tutorial will show you how to deploy your Flask application on a Linux server Nginx with Gunicorn.

Requirements

• A Server (Debian is used in this tutorial)
• Nginx
• Gunicorn
• Pipenv

The first step in the process is to install pip.

$ sudo apt update

$ sudo apt install python3-pip

We will use pipenv to set up the virtual environment.

$ pip3 install pipenv

Check that pipenv installed correctly by checking the version.

$ pipenv --version
pipenv, version 2020.5.28

Create project directory to hold your project. /var/www/ is a good location.

$ sudo mkdir /var/www/myproject

The default permissions for the directory are set to root.

$ ls -ld /var/www/myproject
drwxr-xr-x 2 root root 4096 Jan 27 12:37 /var/www/myproject/

Change the permissions for the user and group to your username and the group www-data.

$ sudo chown -R $USER:www-data /var/www/myproject

Check the to see that the permissions changed.

$ ls -ld /var/www/myproject
drwxr-xr-x 2 brandon www-data 4096 Jan 27 12:37 /var/www/myproject/

CD into the directory.

$ cd /var/www/myproject

Create a .env file to hold the environmental variables.

$ touch .env

Edit the .env file to add the FLASK_APP and FLASK_ENV environmental variables. Add these two lines:

$ vim .env

FLASK_APP=wsgi.py
FLASK_ENV=production

Start the virtual environment. Pipenv will load the variables in the .env file automatically.

$ pipenv shell 

[...]
✔ Successfully created virtual environment! 
Virtualenv location: /var/www/myproject/.venv
Creating a Pipfile for this project…
Launching subshell in virtual environment…
 . /var/www/myproject/.venv/bin/activate

(myproject) brandon@server:/var/www/myproject $

Use pipenv to install the dependencies.

$ pipenv install flask gunicorn

To set up a minimal Flask application create two files application.py and wsgi.py. The main Flask application will be stored in application.py.
wsgi.py will be used to get the application running.

$ touch application.py wsgi.py

Edit application.py. Add the code for a minimal Flask application with one route for the index page.

$ vim application.py

from flask import Flask

app = Flask(__name__)


@app.route('/')
def index():
    '''Index page route'''

    return '<h1>Application Deployed!</h1>'

Edit the wsgi.py file. Add this code wsgi.py to get the application running.

$ vim wsgi.py

from application import app

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

Let’s test run the Flask application!
Make the Flask application listen on all server interfaces by specifying the address ‘0.0.0.0’.

$ flask run --host '0.0.0.0'

 * Serving Flask app "wsgi.py" (lazy loading)
 * Environment: production
 * Debug mode: off
 * Running on http://0.0.0.0:5000/ (Press CTRL+C to quit)
 * Restarting with stat

Now you can access the application from another computer using a browser, navigate to the server's IP address on port 5000. Make sure port 5000 is not blocked by a firewall so that you will be able to access the Flask application.

Since my server's IP address is 192.168.12.34 I access the Flask application from my laptop with a browser using this address:

http://192.168.12.34:5000

You should see this in the browser.

Press CTRL+C to stop the Flask development server.

It is time to set up Gunicorn!

If you were able to run the Flask development server successfully use this command to test run the application using Gunicorn.

$ gunicorn --workers 4 --bind 0.0.0.0:5000 wsgi:app

--workers N

Set the --workers to two x the number of cores in your server. Adjust the number later if you have any issues. Do not exceed 12.

--bind 0.0.0.0:5000

This will listen on all server interfaces on port 5000.

wsgi:app

wsgi is the file name without the extension. app is the instance of the Flask application within the file. You should see the similar output below.

$ gunicorn --workers 4 --bind 0.0.0.0:5000 wsgi:app
[2021-01-27 14:28:00 -0500] [24662] [INFO] Starting gunicorn 20.0.4
[2021-01-27 14:28:00 -0500] [24662] [INFO] Listening at: http://0.0.0.0:5000 (24662)
[2021-01-27 14:28:00 -0500] [24662] [INFO] Using worker: sync
[2021-01-27 14:28:00 -0500] [24695] [INFO] Booting worker with pid: 24695

Press CTRL+C to stop the Gunicorn server.

Now that you have a basic Flask application running with Gunicorn we can set up Nginx.

Install Nginx.

$ sudo apt install nginx

$ sudo systemctl enable nginx

$ sudo systemctl start nginx

Check to see if Nginx is running by running this command.

$ sudo systemctl status nginx


● nginx.service - A high performance web server and a reverse proxy server
   Loaded: loaded (/lib/systemd/system/nginx.service; enabled; vendor preset: enabled)
   Active: active (running) since Wed 2021-01-27 15:58:49 EST; 
   [...]

Now we create the systemd file that runs Gunicorn as a service. Add the following code but change all paths and user names so that they correspond to your set up.

$ vim /etc/systemd/system/myproject.service

[Unit]
Description=myproject.service - A Flask application run with Gunicorn.
After=network.target

[Service]
User=brandon
Group=www-data
WorkingDirectory=/var/www/myproject/
ExecStart=/var/www/myproject/.venv/bin/gunicorn --workers 3 --bind unix:/var/www/myproject/myproject.sock wsgi:app

[Install]
WantedBy=multi-user.target

Next we need to create the server block/virtual host in Nginx. Add the following content. Change the name myproject to the name of your project.

$ vim /etc/nginx/sites-available/myproject.conf

server {
        listen 80;
        server_name myproject www.myproject;

        access_log /var/log/nginx/myproject.access.log;
        error_log /var/log/nginx/myproject.error.log;

        location / {
                include proxy_params;
                proxy_pass http://unix:/var/www/myproject/myproject.sock;
        }
}

server_name is where the domain name goes. That is what you will use to access the web application.

access_log and error_logs specify the path to the access and error logs.

location block is where Nginx reverses proxy back to the Flask application.

Enable the website by creating a link to the sites-enabled directory.

$ sudo ln -s /etc/nginx/sites-available/myproject.conf /etc/nginx/sites-enabled/

Make sure the link has been created in sites-enabled directory.

$ ls -l /etc/nginx/sites-enabled/ | grep myproject
lrwxrwxrwx 1 root root 41 Jan 27 17:37 myproject.conf -> /etc/nginx/sites-available/myproject.conf

Check the nginx configuration for errors then restart the service.

$ nginx -t

$ sudo systemctl restart nginx

$ sudo systemctl status nginx

The Flask application is no longer accessible via the IP address since it is now being served by Gunicorn and Nginx. To access the Flask application you would need to use the name you set in the Nginx server block for the directive server_name in the Nginx configuration. To access the web page can edit the host file on your desktop/laptop to point the domain to the IP address of your server.

Edit the host file to add point the domain name to the server.
Since my server's IP address is 192.168.12.34 I would add this line to the host file.

192.168.12.34 myproject

Host file location for Linux:
/etc/hosts

Host file location for Windows:
C:\Windows\System32\drivers\etc\hosts

Now I can access the Flask application with a browser via the name.

http://myproject

You should seen something similar below.

Here are some troubleshooting steps...

For when things do not go right there are a few things you can do to troubleshoot.

First. Check carefully for any spelling errors in all the files you edited.

Check systemd.

$ sudo systemctl status myproject.service

$ sudo journalctl -xe

Restart the Gunicorn systemd service.

$ sudo systemctl restart myproject.service
$ sudo systemctl daemon-reload

Add logging to Gunicorn by changing the ExecStart line.

From this:

ExecStart=/var/www/myproject/.venv/bin/gunicorn --workers 3 --bind unix:/path/to/myproject/myproject.sock wsgi:app

To this:

ExecStart=/var/www/myproject/.venv/bin/gunicorn --workers 3 --bind unix:/path/to/myproject/myproject.sock wsgi:app --error-logfile /home/bw/bin/lanmanager/gunicorn.error.log --log-level debug

Check the log files.

$ less /var/www/nginx/error.log

$ less /var/www/myproject/error.log

Check to see if the Nginx server is listening on port 80.

$ ss -lnt | grep 80
LISTEN    0         511                0.0.0.0:80               0.0.0.0:* 
LISTEN    0         511                   [::]:80                  [::]:* 

Use curl to check for a 200 OK response.

$ curl -I http://myproject

HTTP/1.0 200 OK
Content-Type: text/html; charset=utf-8
Content-Length: 30
Server: Werkzeug/1.0.1 Python/2.7.16
Date: Wed, 03 Feb 2021 15:18:55 GMT

Follow me on Github.

Feel free to leave questions, suggestions, and comments.