The instant clone operation produces a virtual machine whose memory is exactly the same as the source virtual machine. This can cause conflicts in network addressing, if two different virtual machines respond concurrently with the same network identity. The conflicting network attributes are usually the host name, IP address, and MAC address.
In the case of an instant clone derived from a frozen source virtual machine, a conflict might not immediately arise, because the source virtual machine is not active on the network. The instant clone simply assumes the identity of the source virtual machine when the host unstuns it. However, a second clone derived from the same frozen virtual machine could create a network conflict by duplicating the same identity as the first instant clone.
For an instant clone derived from the current running point of the source virtual machine, the problem is more acute because the source virtual machine resumes operation simultaneously with the instant clone.
To avoid network conflicts, you customize the instant clone to adopt a new identity, before allowing both the source virtual machine and the instant clone to run concurrently. In a situation where temporary network disruption is tolerable, you can ignore the conflict while you customize the network settings in the guest operating system of the instant clone.
Otherwise, you have two options to avoid the race condition that results when both virtual machines resume running from the same point. You can either disable the virtual NIC before the instant clone operation, or else you can specify that the virtual NIC in the instant clone will be disabled at the time when it resumes running after the operation. You specify the latter by setting the config property of the spec parameter when you invoke the InstantClone_Task method.
