To do vCenter high availability with dedicated network interface controllers (NICs) you have to create a distributed switch where you connect your host VMKernel adapters.
Select your vCenter server, then “Networks” tab, and then go to “Distributed Switches”. Click Distributed Switches, and then click “New Distributed Switch”.
Name the Switch, choose the data center you’re going to use vCenter HA in( you can use it between multiple data centers but I’m not covering that here).
Select the highest possible version that works for the vCenter VMs hosts. I currently have a mix of ESXi 6.0.0 and 6.5.0 but that’s soon to change.
Choose the number of uplinks (ports for your NICs). You can leave network I/O Control enabled. Create a default port group and give it a name.
On the settings review page, all you have to click is “Finish”.
Once you have your new distributed switch, you can add hosts to it. Click “Add and Manage Hosts” button on the “Distributed Switches” tab and choose “Add hosts”. Click “Next”, choose “New hosts…”, and choose your hosts.
Add “physical adapters” and “VMKernel adapters”, click “Next”, choose the NICs you want to use as the high availability network, and click “Assign uplink” (you have to do that one at a time). Once done, click “Next”. Do the same for the VMKernel adapters (make sure your vCenter HA NIC has a VMKernel adapter) and click “Next”.
“Analyze impact” page should tell you that it has no impact on iSCSI.
Press “Next” and “Finish”.
Make sure to give the VMKernel adapters appropriate IP addresses on your vCenter HA network. You can now configure vCenter HA with a separate network for HA data.