Simon Griffee

Textdrive Server Deployment with Nginx via FastCGI

5 May 2013 · 5 minute read

My notes on deploying a Django application (this website) on Textdrive shared hosting with Nginx via FastCGI.

Thanks to Evan Carmi, who much of this is paraphrased from, and Pokoka who has far more knowledge on system administration than I, and has an alternate method using uWSGI and a handy script.

Search and Replace


Server folder layout

            .virtualenv     <- virtual environment
                        python2.7/site-packages     <- where django and packages installed with pip go
                        nginx.conf    <- nginx config file (=todo:create symlink from here to web/yourproject/yourproject/conf below)
                    yourproject     <- django project git repository cloned from github
               <- fcgi startup script. don't forget to make executable with chmod +x
                static   <- static files - either create symlink from here to project repo or run python collectstatic
                static/files  <- static files uploaded by users of your app
                static/@admin  <- django admin media symlink

SSH into your Txd server

$ ssh

Install virtualenv

$ mkdir -p local/
$ cd local/
$ wget -O virtualenv-1.9.1.tar.gz
$ tar xzvf virtualenv-1.9.1.tar.gz

Let’s now create our virtualenv on our server. This is the same as on our local machine:

$ python --no-site-packages ~/domains/yourdomain.tld/.virtualenv/yourproject
$ cd ~/domains/yourdomain.tld/.virtualenv/yourproject
$ . bin/activate

Setup the server’s virtualenv

Create a symbolic link called ‘yourproject’ in ~/domains/yourdomain.tld/.virtualenv/yourproject/lib/python2.7/site-packages:

$ ln -s `pwd` ../lib/python2.7/site-packages/`basename \`pwd\``
$ export DJANGO_SETTINGS_MODULE=yourproject.settings

Add the previous line - export DJANGO_SETTINGS_MODULE=yourproject.settings - to the ~/domains/yourdomain.tld/.virtualenv/yourproject/bin/activate file:

$ echo "!!" >> ../bin/activate

Activate virtualenv

$ source ~/domains/yourdomain.tld/.virtualenv/yourproject/bin/activate

Install Django, then upload your application to the server

$ pip install django
$ cd ~/domains/yourdomain.tld/web/
$ python startproject yourproject

This will have created the following folder layout under ~/domains/yourdomain.tld/web/:

    yourproject     <-- default django app

Remove the latter yourproject folder (the default django app created by the startproject command) and replace it with your own project cloned from Github or elsewhere, since, like a good citizien, you are developing locally and only putting tested applications on the production server. Leave alone.

$ cd ~/domains/yourdomain.tld/web/yourproject/
$ rm -rf yourproject
$ git clone yourproject
Initialized empty Git repository in /users/home/yourusername/domains/yourusername/web/yourproject

Set up your settings files for production (=TODO: Set up django-configurations) and update the database settings to use your PostgreSQL or MySQL database if you are using those instead of sqlite3. If you use PostgreSQL or MySQL you need to create those first in virtualmin.

    'default': {
        'ENGINE': 'postgresql_psycopg2',
        'NAME': 'yourusername_django_mysite_database',
        'USER': 'yourusername',
        'PASSWORD': 'password',
        'HOST': 'localhost',
        'PORT': '5432',

Try running python syncdb. If it works then your database is configured correctly.

Install required software packages on the server with pip

Let’s install the software packages from your application’s requirements.txt file (if you have one).

$ cd ~/domains/yourdomain.tld/web/yourproject/yourproject
$ pip install -r requirements.txt

Setup static media

Let’s assume you have some static media for your project in ~/domains/yourdomain.tld/web/yourproject/yourproject/static

For security reasons (but don’t trust me on this) we don’t want to serve static media (CSS, JavaScript, images) from inside our project directory. Instead, let’s create some other directories to serve static media from:

$ mkdir -p ~/domains/yourdomain.tld/web/public/static

And then create a symbolic link from there to our media directory.

$ ln -s ~/domains/yourdomain.tld/web/yourproject/yourproject/static/ ~/domains/yourdomain.tld/web/public/static

Now let’s link Django’s contrib.admin media to this location so that the static assets of Django’s admin app get served as well:

$ ln -s ~/domains/yourdomain.tld/.virtualenv/yourproject/lib/python2.7/site-packages/django/contrib/admin/static/admin/ ~/domains/yourdomain.tld/web/yourproject/yourproject/static/admin

And lastly let’s configure (or depending on your project setup) to use these locations:

import os
import sys

from os.path import dirname, join
from sys import path

path.append(join(dirname(__file__), "yourproject"))
PROJECT_ROOT = os.path.abspath(os.path.dirname(__file__))

STATIC_URL = 'http://domain.tld/static/'
ADMIN_MEDIA_PREFIX = os.path.join(PROJECT_ROOT, "admin/")

Setup NginX and FastCGI

Create a directory to keep your nginx.conf file and the nginx.conf file itself:

$ mkdir -p ~/yourdomain.tld/etc/nginx/sites-enabled
$ vim ~/yourdomain.tld/etc/nginx/sites-enabled/nginx.conf

=TODO: keep nginx conf in the project’s git repository and create symlink from sites-enabled to yourproject/conf/nginx.conf. This way we can keep the config in version control.

Edit your nginx.conf file to look like the following but with your own port number0, domain (yourdomain.tld), and username (yourusername).

error_log /dev/null crit;

worker_processes 1;
pid /users/home/yourusername/domains/yourdomain.tld/tmp/;

events {

http {
    include     /opt/local/etc/nginx/mime.types;

    client_body_temp_path /users/home/yourusername/domains/yourdomain.tld/var/spool/nginx/client_temp 1 2;
    proxy_temp_path /users/home/yourusername/domains/yourdomain.tld/var/spool/nginx/proxy_temp 1 2;
    fastcgi_temp_path /users/home/yourusername/domains/yourdomain.tld/var/spool/nginx/fstcgi_temp 1 2;
    uwsgi_temp_path /users/home/yourusername/domains/yourdomain.tld/var/spool/nginx/uwsgi_temp 1 2;
    scgi_temp_path /users/home/yourusername/domains/yourdomain.tld/var/spool/nginx/scgi_temp 1 2;

    log_format main '$remote_addr - $remote_user [$time_local] '
                  '"$request" $status$body_bytes_sent "$http_referer" '
                  '"$http_user_agent" "$http_x_forwarded_for"';
    access_log /users/home/yourusername/domains/yourdomain.tld/logs/nginx/nginx.access.log;
    error_log /users/home/yourusername/domains/yourdomain.tld/logs/nginx/nginx.error.log;

    server {
        listen       yourportnumber;
        # error_log /users/home/yourusername/domains/yourdomain.tld/logs/nginx-fcgi.error.log;

        location /static/files{
            alias /users/home/yourusername/domains/yourdomain.tld/web/public/static/files;

        location /static {
            alias /users/home/yourusername/domains/yourdomain.tld/web/public/static;

        location / {
            fastcgi_pass unix:/users/home/yourusername/domains/yourdomain.tld/web/yourproject/yourproject.socket;
            # fastcgi parameters
            fastcgi_param PATH_INFO $fastcgi_script_name;
            fastcgi_param QUERY_STRING $query_string;
            fastcgi_param REQUEST_METHOD $request_method;
            fastcgi_param SERVER_PORT $server_port;
            fastcgi_param SERVER_PROTOCOL $server_protocol;
            fastcgi_param SERVER_NAME $server_name;
            fastcgi_param CONTENT_TYPE $content_type;
            fastcgi_param CONTENT_LENGTH $content_length;

Create an script in your project directory to start the Django FastCGI process that should look like:

#Activate the virtualenv
source /users/home/yourusername/domains/yourdomain.tld/.virtualenv/yourproject/bin/activate

# DJANGO_ADMIN="/users/home/yourusername/domains/yourdomain.tld/.virtualenv/yourproject/bin/"
OPTIONS="maxchildren=2 maxspare=2 minspare=1"

case "$1" in
      # Starts the Django process
      echo "Starting Django project"
      # $BIN_PYTHON $DJANGO_ADMIN runfcgi $OPTIONS method=$METHOD socket=$SOCKET_FILE pidfile=$PID_FILE
      $BIN_PYTHON $MANAGE runfcgi $OPTIONS method=$METHOD socket=$SOCKET_FILE pidfile=$PID_FILE
      # stops the daemon by cating the pidfile
      echo "Stopping Django project"
      kill `/bin/cat $PID_FILE`
      ## Stop the service regardless of whether it was
      ## running or not, start it again.
      echo "Restarting process"
      $0 stop
      $0 start
      echo "Usage: (start|stop|restart)"
      exit 1

### Make this file executable:

$ chmod +x /users/home/yourusername/domains/yourdomain.tld/web/yourproject/

Start Django FastCGI instance with:

$ /users/home/yourusername/domains/yourdomain.tld/web/yourproject/ start

This script also takes stop, and restart as parameters.

Start Nginx with your configuration file:

$ /usr/local/sbin/nginx -p /users/home/yourusername/ -c /users/home/yourusername/domains/yourdomain.tld/etc/nginx/sites-enabled/nginx.conf

The Django application should now be running at http://domain.tld:PORTNUBMER/. Don’t forget to log in and go to http://domain.tld:PORTNUBMER/admin/sites/site/ and set the domain name.

Create a ProxyPath and ProxyPathReverse from http://domain.tld:PORTNUMBER to http://domain.tld

  1. Log into your txd account through virtualmin (
  2. Under “Server Configuration” click on “Proxy Paths”
  3. Click “Add a new proxy path.”
  4. Enter “/” for the “Local URL path”
  5. Enter “" for the “Destination URL”.
  6. Click “Save”


  1. Click “Services”
  2. Click “Configure website”
  3. Click “Show Directives”
  4. From the drop down list chose “ProxyPassReverse”
  5. Find the heading “Map remote Location: headers to local”
  6. For “Local URL Path” add “/”
  7. For “Remote URL” add “balancer://root/”

NOTE: the balancer name should be the same as the one entered for “Remote URL” under “Map local to remote URLs”.

=TODO: create bootup actions in virtualmin to start the Django FastCGI process and NginX on server reboots.

Stopping Nginx

$ ps -ef | grep nginx | awk '{print $2}'| xargs kill -9


$ /usr/local/sbin/nginx -s stop -p /users/home/yourusername/ -c /users/home/yourusername/domains/yourdomain.tld/etc/nginx/sites-enabled/nginx.conf

Stopping fcgi (whenever you update your application you need to stop and then start fcgi)

$ . stop


Set up fabric script to automate deployment.