The way the repeated conversions work depends on where you choose to create the virtual machine.
The following is a detailed description of both cases.
If you choose to save the virtual machine as a set of files
As a result of the first conversion, a new virtual machine will be created. Every subsequent conversion will re-create this machine from scratch. First, the old machine is temporarily renamed. Then, a new virtual machine is created that has the previous name of the old machine. If this operation succeeds, the old machine is deleted. If this operation fails, the new machine is deleted and the old machine is given its previous name. This way, the conversion always ends up with a single machine. However, extra storage space is required during conversion to store the old machine.
If you choose to create the virtual machine on a virtualization server
The first conversion creates a new virtual machine. Any subsequent conversion works as follows:
Intermediate snapshots
To be able to update the virtual machine, the software stores a few intermediate snapshots of it. They are named Backup… and Replica… and should be kept. Unneeded snapshots are deleted automatically.
The latest Replica… snapshot corresponds to the result of the latest conversion. You can go to this snapshot if you want to return the machine to that state; for example, if you worked with the machine and now want to discard the changes made to it.
Other snapshots are for internal use by the software.