Last year, when vSphere with Kubernetes (original name of what is now vSphere with Tanzu) was first released, I had shared a process on how to deploy a minimal setup including a detailed write-up for setting up vSphere with Tanzu on an Intel NUC with just 32GB of memory.
I am always looking for ways to simplify and ease the consumption of various VMware technologies within a homelab and I was pretty happy with the tweaks that I could make to reduce the amount of resources needed to run vSphere with Tanzu. Instead of needing to deploy three Supervisor Control Plane VMs, the modification to the vSphere with Tanzu configuration, allowed me to deploy just two Supervisor Control Plane VMs. It was unfortunate that deploying only a single Supervisor Control Plane VM at the time was not possible due to a known issue.
While deploying a pre-release of vSphere 7.0 Update 3 in one of my lab environments, I was going through the process of tweaking the vSphere with Tanzu configuration before enablement and I figure why not try the one node setting, in case it was fixed 🤷 I honestly was not expecting it to work since there was an internal bug that was filed awhile back and I had not seen the bug closed. To my complete surprise, vSphere with Tanzu enabled successfully and there was just a single Supervisor Control Plane VM!
It turns out that someone from Engineering must have fixed the issue and a single Supervisor Control Plane VM is now possible with the upcoming release of vSphere 7.0 Update 3! 🥳
To change the settings, you will need to SSH to the VCSA and edit the following configuration file /etc/vmware/wcp/wcpsvc.yaml and search for minmasters and maxmasters and change the value from 3 to 1.
For the changes to go into effect, you will need to restart the vSphere with Tanzu service which is listed as wcp by running the following command:
service-control --restart wcp
In addition, for homelab purposes, you may also want to change the controlplane_vm_disk_provisioning parameter, which defaults the Supervisor Control Plane VM to Thick provisioned rather than Thin, which many folks use in their labs.