In the current documentation for on-premise deployment[1], the ClusterDeployment configuration example uses platform.agentBareMetal.

[1] https://docs.redhat.com/en/documentation/red_hat_advanced_cluster_management_for_kubernetes/2.14/html/clusters/cluster_mce_overview#creating-a-cluster-on-premises

Users who want to deploy with platform: none are having trouble and don't know how to configure it. I think at least the following considerations are necessary for platform:none.

• Set clusterdeployment.spec.platform to None.
• Set agentclusterinstall.spec.platformType to None.
• Don't set agentclusterinstall.spec.apiVIPs and agentclusterinstall.spec.ingressVIPs, because these VIPs are only for the baremetal or vsphere platforms.
• Set agentclusterinstall.spec.loadBalancer.type to UserManaged, because platform:none does not provide cluster-managed load balancer.
• The user will prepare the external LB and DNS, just like with a regular (non-MCE) cluster using platform:none.

I'd like these to be clearly stated in the documentation.