VCF 9.0 introduced many new architectures and deployment options for enterprises to take existing VMware infrastructure and build a private cloud operating model with it. I personally call it the great consolidation.

To give quick illustration of my existing VCF Instance, using VCF Operations Inventory / Topology view, you can see the breakdown of all the components, the goal here will be to add a second VCF Instance and those are the steps I will have below.

A really great tool to help teams get started with planning their upgrades and converges is the VCF Upgrade Planner. This hosted on VMware’s Github and an interactive tool to help put together a procedureal plan for any of the supported outcomes.

For the environment that will be becoming an additional instance the existing fleet, it’s a vSphere 8.0 U3i, Image Based Lifecycle, 3 node vSAN cluster. I have a single Aria Operations 8.18.6 Data Node with no Cloud Proxies and the vCenter appliance all residing in the same cluster. I’m only working off two different networks, My host network and VM Management network is different and of course have vMotion & vSAN networks configured. The other VCF feature I have enabled is Memory Tiering. The cluster is at about 48 vCPUs & 574GB of RAM. The goal is to deploy Simple.

The path I’m taking is to upgrade directly to vSphere vCenter 9.1 and then go to direct convergence as an additional VCF Instance.

Upgrading vCenter 8.0 U3i to vCenter 9.1.0.0300

vCenter appliance upgrade is going through the usual deployment method of using the UI, deploy a new appliance and transfer of data. This did complete successfully. Saving the steps, you may refer to previous blogs or Broadcom TechDocs for help.

Using the VCF Installer to Add a Second VCF Instance

I have the same VCF Installer I used for my initial VCF Fleet, keeping the VCF Installer appliance apart of the initial deployment hosts, allows it to remain intact for other actions.

From the VCF Installer, kicking off Deployment Wizard for VMware Cloud Foundation and Deploy a new VCF Instance

For the next step, I will be selecting Small / Small for Deployment Model & Size and then keeping the default VM Management Network for VCF Management components.

For the Prepare portion, populate unique names for the VCF Instance and Management Domain name

On the following screen, you will be given an option to connect to a VCF Operations Instance

Now that I’ve connected to primary VCF Operations Instance, I have to provide reserved FQDNS for VCF Operations Cloud Proxy (a requirement) as well as FQDNs for VCF Instance & Runtime (separate from my other instance).

Connect to the Existing vCenter

Provide a range of IP addresses from the same VCF Management Services network, Broadcom allows minimum of 12 to get started however recommends a range of 30 IPs for scaling.

Going with a Standalone NSX instance (Remember this is for POC/Test environments), and I’m choosing not to use ESX Management Network for TEPs.

Did not get alot of errors, I had one I needed to set DRS back to Fully Automated, keep in mind these ESX hosts are running 8.0 U3, I never upgraded previously and thus giving the Image Compliance warning. We can click ‘Acknowledge’ and proceed to Deploy.

As the deployment kicks off, the SDDC appliance is the first OVA to be deployed, followed by the NSX Manager, VCF Management Services and finally the convergence into VCF Operations.

When it completed, some of the items did appear however because this was registered to a different Aria Operations instance, I had to do some changing things around to add it to the primary one.

Our updated Topology reflects 2 Instances

Finally, you can see the separation when it comes to life cycling environments.