Puppet Code Management at Scale
Marcel Härry - Monday, Feb 6, 17:00 - 17:20 - B1.015
At Swisscom we are managing a large variety of platforms using Puppet. The different teams are all using the same principles to build up their environments using the same code base. With the scale and specificities of our systems, we had to invent and build a clever, scalable and reliable puppet code deployment system.
We will talk about testing module code through a strict CI pipeline, packaging modules as RPMs and deploying them to our 100+ Puppetmasters, and how to establish common best practices and ensure code quality by using peer code review before promoting code to production.
During the last 4 years we built up these principles incrementally and by sharing our journey, we are providing insights into the reasoning behind our decisions and the current architecture. You will learn about pain points, as well as the benefits, like having fast and lots well tested module upgrades.
Leading the architecture of Swisscom's elastic platform going from cloud infrastructure based on OpenStack up to a Platform as a Service running on CloudFoundry.
Having both a system and software engineering background and years of participation within devops-minded community (e.g. Puppet), he is seeking for an automated and continuously integrated delivery of operational platforms.