Openstack
Rubus (Raspberry Pi Cluster)
Background # There have been many Raspberry Pi clusters before, they first started when the Raspberry Pi came out first. They are all very cool and I’ve wanted to try to make my own, but never quite got round to it or could justify it or think of any good project to run on them. Around the time the Raspberry Pi 3 came out, I was deep into learning how OpenStack work. That is the stage after running packstack and playing, the stage of looking at all the components and how they all fit together. After a while I thought it would be really quite cool to run OpenStack on a Raspberry Pi Cluster. OpenStack is quite resource intensive, dozens of multi forking memory hungry (in Pi terms) Python daemons. So it would really need multiple Pi’s to run. However, given the whole point is to run instances, the Raspberry Pi is quite limited regarding actual virtualisation. However I didn’t let that stop me. I didn’t want to run any production instances, I just wanted to learn.
OpenStack proof of concept
Introduction # This is a short howto on creating a small OpenStack proof of concept on CentOS. The aim is to create a proof of concept OpenStack Liberty deployment on a single Linux testing machine, where the deployment matches real world hardware deployments. The aim is not to just install all the OpenStack components in a single machine or virtual machine. The aim is to make use of Linux, KVM and libvirt to create virtual hardware to run the various OpenStack components to match a real world deployment. Each virtual machine represents what could be a physical host. It will use CentOS 7 as the base operating system and use the RDO OpenStack packages. This matches the Red Hat OpenStack Platform, however does not need subscriptions to install and test.