A: One way to verify the functionality of vSphere 5.0's completely rewritten vSphere HA clustering is to create a cluster and then force one of its members to fail. A hard failure presents the best test because it simulates the closest possible scenario of a host failure.
Obviously, you can simulate this failure by pulling the power cable from the server, but sometimes you might not have easy access to that power cable.
An alternative method is to panic the ESXi server's kernel, forcing a Purple Screen of Death (PSOD) and a resulting hard failure. You can accomplish this within the ESXi console by using the following command:
vsish -e set /reliability/crashMe/Panic 1
You can then reboot the server as normal.