When building highly available clusters one often chooses for 1 extra physical machine per service, hence creating an A-B fail-over schema. As the number of primary servers grows, your number of secondary servers grows accordingly. More complex setups have a many (N) to 1 fail-over schema. In this session we'll discuss how to solve the complex setup of a N to 1 fail-over by virtualising all the secondary machines on 1 physical machine using the Xen Hy ervisor. We will show how we integrated a standard Linux-HA framework with GFS and added a virtualization layer to it, and we'll discuss do's and don'ts when building a Virtual  Cluster. Apart  from that we'll also document how while deploying such an infrastructure we also succeeded in passing the 10th floor test.