deployable

A simple system for repeatable deploys.

Why?

There are a number of deployment systems out there (such as Capistrano_, zc.buildout, pip requirements, etc.) but most are very complicated or are language-specific. deployable aims to be language-agnostic, easy to use, repeatable, flexible and fault-tolerant.

deployable is ideal for setting up a single machine. Should you need deployment on multiple machines, you might want to consider using deployable with Fabric_.

.. _Capistrano: http://www.capify.org/ .. _zc.buildout: http://pypi.python.org/pypi/zc.buildout/ .. _pip: http://pypi.python.org/pypi/pip .. _Fabric: http://www.nongnu.org/fab/

Goals

• Language-agnostic
• Repeatable
• Easy to use
• Easy to extend
• Work with many major version control systems
• Allow for cached deployment (pull sources to a single machine then redistribute to other servers)

Requirements

• Requires Python 2.4+ w/ no other dependancies.
• No install required, just include the deployable.py file along with your source code.
• Only test on Unix systems, though it may work on Windows. (Reports/patches accepted!)

Usage

Simply create a Python file of with the name of your choice and fill it with::

from deployable import *

install_list = [
    Shell(command='mkdir foo'),
    Tarball(url='http://pypi.python.org/packages/source/W/Whoosh/Whoosh-0.3.0b24.zip'),
    Git(url='git://github.com/toastdriven/django-haystack.git'),
    Svn(url='http://code.djangoproject.com/svn/django/trunk/django'),
    GitSvn(url='http://pysolr.googlecode.com/svn/trunk/'),
    Hg(url='http://bitbucket.org/ubernostrum/django-profiles/'),
]

deploy(install_list)

Then run the file (i.e. python deploy_me_please.py). You'll get a log of what's going on. For advanced usage, see the api_complex.py sample file included with this distribution.

Supported

Currently supported commands are:

• shell
• tarball
• Git
• Mercurial
• Subversion
• git-svn