This is the basis of the client and server tools written in django to deploy applications to a production environment. It uses a remote git server to handle all deploying. This makes deploying as simple as:

$ git push cannula master

• pyyaml
• Django >= 1.3.1
• virtualenv >= 1.5.1
• supervisor
• git
• openssh

Developer documentation are located in the 'docs/build/html/' directory. The documentation can be built with the command 'make html'. You will need to have sphinx installed to build the docs. This is still a work in progress.

You should add a new user on you machine for the best results. Log in as this new user and issue the commands to bootstrap a sample instance of cannula:

1. Install the dependancies (ubuntu):

   $ sudo apt-get install python-dev build-essential nginx python-pip \
     python-virtualenv git

2. Create a cannula user and a virtualenv for the code:

   $ sudo adduser cannula --disabled-password --home /cannula
   $ sudo su - cannula
   $ virtualenv cannulaenv
   $ source cannulaenv/bin/activate

3. Install cannula:

   pip install -r https://raw.github.com/rmyers/cannula/master/requirements.txt

4. Fire up django dev server:

   django-admin.py initialize --settings=cannula.settings
   django-admin.py syncdb --settings=cannula.settings
   django-admin.py runserver --settings=cannula.settings

5. Use the newly spawned server, enjoy!

Security is good, this is one of the advantanges of using cannula. You can give a developer access to just deploy a single application or server without giving them ssh access and sudo. You should also secure the machine to disable password authentication (once you have your ssh keys setup of course)

Edit your /etc/ssh/sshd_config:

PasswordAuthentication no

Then restart sshd. Now on the plus side you wont get thousands of login attempts by bots using a dictionary brute force.

All configuration settings for cannula are stored in a conf file 'ini' style. by default cannula will look in /etc/cannula/cannula.conf and ~/.cannula.conf you can have a look at the defaults here:

https://raw.github.com/rmyers/cannula/master/cannula/defaults.cfg

Application just need to specify an app.yaml file to tell cannula how to deploy it. Currently the only handler that is available is gunicorn_django, but it should be easy to add any others. Here is a sample app.yaml file:

# optional domain to listen on
domain: example.com
# optional version string
version: 1.0
# runtime (ruby, python, php, java)
runtime: python
api_version: 1

# setup some defaults to pass to handlers
gunicorn_defaults: 
  workers: 2
  worker-class: sync
  worker-connections: 1000
  max_requests: 0
  timeout: 30
  keepalive: 2

# Main wsgi handlers and static file definitions
# Each unique 'worker' handler will setup a separate process
handlers:
# static files (less greedy urls should come first!)
- url: /static
  static_dir: static

# special handler for backend this will spawn a separate gunicorn process
- url: /backend/
  worker: cannula.worker.gunicorn_django
  settings: backend_settings
  defaults: gunicorn_defaults

# Main site handler will match everything else
- url: /
  worker: cannula.worker.gunicorn_django
  settings: main_settings
  defaults: gunicorn_defaults