One server is enough to start
KLYRN VM runs its controller and its first compute node on the same server, with a network and storage ready, so a trial needs one machine.
The usual layout is a controller on one server and compute nodes on others. For a trial, a lab or a small provider that is one server too many. KLYRN VM runs both on the same machine.
Two commands
Install the controller with the public command. Then, on the same server, klyrn-vm admin node add-local adds it as its own first compute node. On the test server that step took 12 seconds. The installer ends by telling you this, so there is no page to go and find.
If you paste the command meant for a separate node onto the controller's own server, it refuses and names the command to use instead.
What is ready without you
A new install used to stop at an empty page: no network, no storage pool, no image you could build from. Now a private network exists as soon as the node does, with addresses to hand out. The node's storage pool is known. The catalogue of images has its checksums. The first machine can be asked for straight away.
The health check tells you what is next
klyrn-vm doctor on a fresh install does not call missing things failures. It lists them as next steps: add a compute node, set a hostname for a real certificate, connect a backup server. A failure means something is wrong, and on a clean install nothing is.
Growing out of it
Add a second node from the panel when you need one; the first keeps running. The first compute node is free for good, and each one after it is $5 a month.