If you happen to run into this issue on stage 2 of a vCenter Upgrade, there is likely a name resolution issue in play. As a note, it involves more than just the workstation you are using to deploy the new vCenter Server. It also involves the new vCenter Server appliance that is staged in the environment from Stage 1.
First, it is best practice to use FQDNs for your source vCenter Server and the target vCenter Server or ESXi server as this sometimes will get you around other weird errors when kicking off the Stage 2 phase.
However, if you are targeting an ESXi host and instead of using an IP address, you use an FQDN, make sure that not only your machine but the staged VCSA can resolve this name.
If you get the message Failed to connect to the source vCenter Server or ESXi container Name or service not known this is due to the staged appliance not being able to resolve the target ESXi or vCenter Server appliance.
If you are in this situation, connect to the console of the staged VCSA appliance, login with the root account configured in Stage 1 and then attempt to ping the target address that you are using in the wizard. You will see the same error message.
The work around is to use the vi command and edit the hosts file on the staged VCSA appliance:
vi /etc/hosts
Add the IP address and name to the target ESXi host or vCenter Server and rerun the Stage 2 prerequisite check and it should be successful.