If you’ve used Heroku, you know what a powerful concept “deploy via git push” is. Even years after Heroku brought this concept to life, it’s still magic. You push your project, and your app is immediately live, it is monitorable, it can be scaled up and down easily, and all of your environments are in a known, revertible state because everything is SCM-driven.
If you’ve ever used Hadoop, you know how sad it is to compare your Hadoop toolchain to Heroku. Everything smells like duct tape.
So today we’re excited to announce something new in the world: Heroku-like deploy for Hadoop, and it is available to all our users immediately.
Deploy is a lot like deploying to Heroku; just git push the code to us. For example:
git push mortar my_branch
Then you can use our new scheduler to set future and recurring execution, or use your own scheduling mechanism via our APIs. This gives you a live Hadoop deploy, with all of the services Mortar already provides: managed execution, snapshotting, log collection, and more.
And this is just the beginning. Imagine a world where working with Hadoop has the same level of services and maturity that working with application code has: fantastic monitoring, continuous integration, reusable components, third-party add-ons… it’s all coming soon.
With deploy and scheduling, we’re on our way.