oVirt Virtualization & High Availability Lab
A KVM and oVirt lab built around centralized VM management, shared NFS storage, and the architecture behind high availability.
The shape of the lab
The goal was to build a small virtualization environment with oVirt Engine at the centre: one management plane, KVM hosts underneath, and shared storage so a virtual machine could live somewhere other than a single host's local disk.
What I built
I installed oVirt Engine on Oracle Linux 8.10 and connected KVM hypervisor hosts. The storage layer was shared over NFS, which made the difference between "a VM on a machine" and "a VM that belongs to the infrastructure" tangible.
- Deployed and explored the oVirt Engine management environment.
- Configured KVM hosts and investigated host connectivity.
- Set up and tested NFS shared storage integration.
- Troubleshot virtualization services, storage activation, and host communication.
- Explored the architecture and requirements for multi-host high availability and failover.
What was difficult
Most of the work was not the installation. It was the seams: making sure the hosts could talk to the engine, that the storage domain would activate cleanly, and that the services underneath were healthy enough for the next layer to trust them. I spent time reading logs, checking network paths, and retrying steps that looked simple on paper.
What I learned
oVirt makes the management layer visible: hosts, storage domains, clusters, and the dependencies between them. I came away with a better mental model of what high availability actually requires — shared state, quorum, fencing, and a failover path that has been tested.
I explored the architecture for a multi-host failover setup, but full failover was a target rather than a completed, verified outcome. That distinction matters: a diagram is not the same as a failure you have survived.
Next
I want to return to the lab with a second host and deliberately break one, so failover becomes something I can observe rather than describe.