Neckbeard is your own personal operations team in charge of maintaining your own personal PaaS. It's the Heroku experience for developers (dead-simple deploys) combined with the flexibility of open source software on servers that you control. You get smart, fault-tolerant, repeatable, datacenter-aware, cloud-centric, one-command deploys, without needing to spend the next month stringing together tutorials.

If you're Netflix, Google, Facebook, LinkedIn, Twitter, etc, then you have a team of folks who have already solved this problem for you. If you just want to throw a hobby app up on a free PaaS, use Heroku. If you have good reasons not to use a PaaS, then don't go re-invent the wheel. Use Neckbeard.

$ neckbeard up

Like videos? We have em: Pycon Neckbeard Lightning Talk

This talk mostly covers the reason why Neckbeard needs to exist.

Neckbeard is built from the ground up for cloud-based ephemeral servers that you can rebuild on a whim with one command. It knows about your backups and knows how to spin up test/staging environments from those backups (and then spin them down). The process you go through for disaster recovery becomes just another deploy because you do it all the time. You can be confident that you can rebuild your entire infrastructure with a source code checkout, a backup (probably also in the cloud somewhere), a network connection and some authentication credentials.

Neckbeard encourages and empowers you to use existing configuration management tools to handle the on-server heavy lifting (like Chef, Puppet or SaltStack). This allows you to use simple, declarative, grokkable configuration files to idempotently define your stack, with a one-command front-end to Make It So. Not only does this make your servers consistent, but it means you can use Vagrant for consistent local development machine setup.

PaaS providers like Heroku and dotCloud do an amazing job of providing simplicity for the simple case, and that's the hallmark of a good tool. We have the same goal, but we also realize that the ability to evolve as complexity increases is an absolute requirement for mission-critical operations. That's why we favor:

• Explicit configuration over implicit magic
• Pluggable backends everywhere
• An architecture based on build blocks of customizable Service Addons

Neckbeard was publicly announced at PyCon 2013 in a lightning talk. It is not currently usable, but is looking for beta testers who are willing to communicate their deployment setup (nothing private) in exchange for assistance in getting their configuration created. Beta testers should be interested in deploying a python application to AWS.

The current path to a usable beta is:

• Define a configuration format (mostly done, but feedback welcome)
• Bring in the existing codebase from PolicyStat's internal deploy tool as a starting point.
• Get a "hello world" python application running with a tutorial.
• Make abstractions and plugin-ize the existing code.
• Write guides for writing backends to alternative clouds (Rackspace)
• Write guides for writing and using additional Service Addons

Have a question or comment about using, improving or extending Neckbeard? Get in touch!

Forums: https://groups.google.com/d/forum/neckbeard IRC: #neckbeard on freenode