having some issues with a rancher managed k8s cluster running on a harvester cluster – one of the nodes is stuck with InfrastructureReady = False, even though the node shows up correctly:

- lastTransitionTime: '2023-05-24T14:30:22Z'
      message: >-
        creating server [fleet-default/test-main-1-1e68de03-7hcf6] of kind
        (HarvesterMachine) for machine test-main-1-7d6f6d4f4b-nhhpt in
        infrastructure provider
      status: 'False'
      type: InfrastructureReady

in our case we had a networking problem and the provisioned VM wasn’t able to communicate with rancher (you might be able to SSH into VM and check the logs, either by downloading the SSH keys from Rancher or by duplicating the VM and adding password authentication in cloud-config)

I believe in our case the issue was that a custom cloud-init script was overriding the default one so the qemu-guest-agent wasn’t being installed. Adding said guest agent install to the cloud init script caused the nodes to reprovision and now everything is green 🙂