I’am new to OpenNebula and trying to get my local homelab running before we plan any moves further on.
I created a vagrant based stack GitHub - EugenMayer/opennebula-lab: Quickly spin up an opennebula lab with a frontend and 2 compute nodes using vagrant where I
- create and ssh key and deploy that one to the
frontend nodeto/root/.sshand/var/lib/one/.ssh(prior minione) - install opennebula via
minioneon thefrontendnode - install 2 compute
nodeswithqemuinstalled - deploy the prior generated
ssh public keyto/root/.ssh/authorized_keyson both compute nodes - populate the
known_hostsoffrontendat/root/.ssh/known_hostsand/var/lib/one/.ssh/known_hostsforfrontend/compute1/compute2
While opennebula starts up and is accessible, i cannot add those hosts. While i can do a (as root on frontend): ssh compute1 - adding the host in the GUI always leads to an issue.
Question
What is the design to actually provision a host? Do i need to create a one user myself on the target host and then ensure that i can connect as one from frontend to one@compute1 and do i need to manually replicate that one home folder on compute1/2 should be /var/lib/one? Is there any better tooling to that?
I have found and followed KVM Node Installation — OpenNebula 6.0.3 documentation but i do not understand
- which user is used by the
opennebulaprocess when connecting to thecomputehost. I mean which “source user” and which "target user` - how to properly setup a target hosts structure before the first connection?
Thank you for any help