In May I switched companies and tech fields – I am now working as a DevOps Engineer at Teradata. During my last year at Axeda I had the opportunity to work on projects dealing with Amazon Web Services and EC2, and I became fascinated with the idea of managing infrastructure in code. I am more on the “Dev” side than the “Ops” side of “DevOps”, as that is my background, however my experience with Linux at Free Geek Providence gives me an advantage when dealing with systems.
I have now completed my first project at Teradata for deploying Hadoop clusters. It’s a Node.js app that fronts a set of Chef cookbooks to deploy any of four different distributions of Hadoop on a virtual machine cluster. The problem that it solves is that Teradata has products that need to work well on various flavors and versions of Hadoop, and the people who test these combinations have a hard time creating the dev environments. They would have less trouble with the better known vendors such as Hortonworks or Cloudera, but the less commonly used vendors such as IBM BigInsights or MapR would present challenges in setting up due to their unfamiliarity.
The app I created is similar in purpose to Cloudbreak, however Cloudbreak (at the time of writing) is only intended for use with Ambari-based vendors such as Hortonworks or BigInsights.
The best part is that this first project is only the tip of the iceberg in learning about cutting edge big data technologies. Next up I will be learning Openstack, which is an open source solution for creating an entire cloud infrastructure – think creating your own private Amazon EC2. Openstack is on track to become increasingly important as companies look to take back control of their infrastructure.