Scaling Bare Metal Provisioning with Nova and Ironic at CERN
Since a few years, all physical servers deployed in the CERN IT data centers are being provisioned as part of the OpenStack private cloud offering, leveraging Nova and Ironic. What started with a few test nodes has meanwhile grown to more than 5’000 physical machines which are managed this way, with the aim to move to ~10’000 nodes eventually.
After an overview of the server life cycle in CERN IT, this presentation will discuss the current layout of our Nova/Ironic control plane to handle all these servers. We will dive into some of the issues we hit when scaling the number of managed nodes and how we addressed these with conductor groups, resource tracker changes, or leadership elections in the Ironic inspector. To make everyone feel better when attending this talk, we will round things up with some of the day-to-day operational challenges and pain points when dealing with physical servers 🙂
linux foundation