pecan/docs/source/deployment.rst

Deployment Recipes

Deploying Pecan projects

Deploying a pecan project can be accomplished in several ways. You may already be familiar with deployment methodologies for other Python projects, in which case, try that! Pecan doesn't deviate from the standards laid out by similar Python web frameworks before it.

Here we will outline some of the common methods for deploying your Pecan project.

Good ol' fashioned source control.

Why not? It works! In our modern world of distributed SCM tools, we find using source control as a completely acceptable method for deployment. Sure it's a bit more manual than you might like, but it gets the job done, and rollbacks are free (if you're into that kind of thing ;)). For this guide, it is assumed you are developing your application in a git repository.

Here are some tips, free of charge:

* Develop on "feature" branches.
* All merge back to "master" when sprint is complete.
* Tag "release" branches.

To setup:

* git clone <project> /opt/project/
* git branch --track <release_branch> origin/<release_branch>
* git checkout <release_branch>
* python setup.py install

To deploy going forward:

* cd /opt/project/
* git pull
* git branch --track <release_branch> origin/<release_branch>
* git checkout <release_branch>
* python setup.py install

That should do it.

Fabric

Fabric makes it way more fun to deploy. You can write straight up python to automate the SCM deployments, or even go so far as to build a full fledged release system with it.

Capistrono

Much like fabric, but they built in some out of the box deployment tools. We'll cover that here.

Chef

Chef borrows it's deployment methodologies from Capistrono, we'll cover that here.

Whiskey Disk

The embarrassingly fast deployment tool. Decoupled from frameworks, one specific design goal, to deploy quickly and easily. YAML configuration.

Egg

Deploy binary packages using python's distribution utilities.

RPM

Deploy your apps with RPM's, built with python's distribution utilities.