This guide is a walk-through on how to setup Ubuntu Server for hosting Django websites. The Django stack that will be used in this guide is Ubuntu, Nginx, Gunicorn and Postgres. This stack was chosen solely from the reading I've done and talking to other Django developers in order to get their recommendations. This stack seems to be one of the latest "standard" stacks for Django deployment. This guide also assumes that you're familiar with Ubuntu server administration and Django. I needed an example site for this guide so I chose to use my Django Base Site which is available on Github.
I would also like to thank Ben Claar, Adam Fast, Jeff Triplett and Frank Wiles for their suggestions and input on this guide.
The version of Ubuntu I'm using for this guide is Ubuntu 11.10 64 bit Server. I've installed Ubuntu Server in a VirtualBox VM on my MacBook Pro which is currently running Mac OS X 10.7.2. During the installation of Ubuntu Server I answered the prompts with the following:
Language: English Install Menu: Install Ubuntu Server Select a language: English Select your location: United States Configure the Keyboard: No Configure the keyboard: English (US) Configure the keyboard: English (US) Hostname: ubuntu-vm Configure the clock: Yes Partition disks: Guided - use entire disk and set up LVM Partition disks: SCSI3 (0,0,0) (sda) - 21.5 GB ATA VBOX HARDDISK Partition disks: Yes Partition disks: Continue Partition disks: Yes Set up users and passwords: Brent O'Connor Set up users and passwords: (Enter a username) Set up users and passwords: ******** Set up users and passwords: ******** Set up users and passwords: No Configure the package manager: <blank> Configure taskse1: No automatic updates Software selection: <Continue> Install the GRUB boot loader on a hard disk: Yes Installation complete: <Continue>
Under the settings for your VM in VirtualBox click on the "Network" tab and then click on the "Port Forwarding" button. Now click on the plus and add the following settings to setup port forwarding for web and ssh.
Name | Protocol | Host IP | Host Port | Guest IP | Guest Port |
---|---|---|---|---|---|
SSH | TCP | 2222 | 22 | ||
Web | TCP | 8080 | 80 |
Since I like to connect to my servers using SSH the first thing I install is openssh-server:
$ sudo aptitude install openssh-server
Since you setup port forwarding in step 2, you should now be able to open up your Terminal and connect to your Ubuntu Server using the following:
$ ssh localhost -p 2222
The Python header files are needed in order to compile binding libraries like psycopg2
.
$ sudo aptitude install python2.7-dev
$ sudo aptitude install postgresql postgresql-server-dev-9.1
Make your Ubuntu user a PostgreSQL superuser:
$ sudo su - postgres $ createuser --superuser <your username> $ exit
Restart PostgreSQL:
$ sudo /etc/init.d/postgresql restart
$ sudo aptitude install nginx
$ sudo aptitude install git
The reason we are setting up a generic deploy user is so that if you have multiple developers who are allowed to do deployments you can easily add the developer's SSH public key to the deploy user's /home/deploy/.ssh/authorized_keys
file in order to allow them to do deployments.
$ sudo useradd -d /home/deploy -m -s /bin/bash deploy
Setup a virtualenv:
$ sudo apt-get install python-setuptools $ sudo easy_install pip $ sudo pip install virtualenv $ cd /usr/local/ $ sudo mkdir virtualenvs $ sudo chown deploy:deploy virtualenvs $ sudo su deploy $ cd virtualenvs $ virtualenv --no-site-packages example-site $ exit
Note
I personally use and setup virtualenvwrapper on all my servers and local development machines so that I can use workon <virtualenv>
to easily activate a virtualenv. This is why I put all my virtualenvs in /usr/local/virtualenvs
.
Make a location for the example site:
$ cd /srv/ $ sudo mkdir sites $ sudo chown deploy:deploy sites $ sudo su deploy $ cd sites $ git clone git://github.com/epicserve/django-base-site.git example-site $ cd example-site/ $ echo `pwd` > /usr/local/virtualenvs/example-site/lib/python2.7/site-packages/django_project_root.pth $ mkdir -p static/cache $ exit $ sudo chown www-data:www-data /srv/sites/example-site/static/cache $ sudo su deploy
Create the file /srv/sites/example-site/config/settings/local.py
and add the following. Make sure to change the password and then save the file. I usually use a random string generator to generate a new password for each new Postgresql database and user:
from base import * LOCAL_SETTINGS_LOADED = True DEBUG = True INTERNAL_IPS = ('127.0.0.1', ) ADMINS = ( ('Your Name', '[email protected]'), ) DATABASES = { 'default': { 'ENGINE': 'django.db.backends.postgresql_psycopg2', 'NAME': 'example_site', 'USER': 'example_site', 'PASSWORD': '<enter a new secure password>', 'HOST': 'localhost', } }
Install the sites required python packages:
$ source /usr/local/virtualenvs/example-site/bin/activate $ cd /srv/sites/example-site/ $ pip install -r config/requirements.txt
Install Gunicorn:
$ pip install gunicorn
Install psycopg2:
$ pip install psycopg2
Create a PostgreSQL user and database for your example-site:
# exit out of the deploy user account $ exit $ createuser example_site -P $ Enter password for new role: [enter the same password you used in the local.py file from above] $ Enter it again: [enter the password again] $ Shall the new role be a superuser? (y/n) n $ Shall the new role be allowed to create databases? (y/n) y $ Shall the new role be allowed to create more new roles? (y/n) n $ createdb example_site -O example_site
Create your Upstart configuration file:
$ sudo vi /etc/init/gunicorn_example-site.conf
Add the following and save the file:
description "upstart configuration for gunicorn example-site" start on net-device-up stop on shutdown respawn exec /usr/local/virtualenvs/example-site/bin/gunicorn_django -u www-data -c /srv/sites/example-site/config/gunicorn/example-site.py /srv/sites/example-site/config/settings/__init__.py
Start the gunicorn site:
$ sudo start gunicorn_example-site
Create a new file sudo vi /etc/nginx/sites-available/example-site.conf
and add the following to the contents of the file:
server { listen 80; server_name localhost; access_log /var/log/nginx/example-site.access.log; error_log /var/log/nginx/example-site.error.log; location = /biconcave { return 404; } location /static/ { root /srv/sites/example-site/; } location /media/ { root /srv/sites/example-site/; } location / { proxy_pass http://127.0.0.1:8001/; proxy_redirect off; proxy_set_header Host $host; proxy_set_header X-Real-IP $remote_addr; proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for; client_max_body_size 10m; } }
Enable the new site:
$ cd /etc/nginx/sites-enabled $ sudo rm default $ sudo ln -s ../sites-available/example-site.conf
Start nginx:
$ sudo /etc/init.d/nginx start
While still connected to your Ubuntu server via SSH run the following, which should spit out the HTML for your site:
wget -qO- 127.0.0.1:80
Since you setup port forwarding in step 2 for web, you should also be able to open up your browser on your local host machine and pull up the website using the URL, http://127.0.0.1:8080.