I recently learned that deleting a Hyper-V checkpoint merges its differencing disk into the parent disk. Suppose a virtual machine has a chain of six checkpoints, and I want to keep the VM exactly as it was at checkpoint 3 while removing checkpoints 4, 5, and 6. Is there a way to discard those later checkpoints without merging their changes into the earlier disk?
3 Answers
No—not directly. Hyper-V checkpoints use differencing disks, so deleting a checkpoint has to merge its disk changes into the parent disk. To make checkpoint 3 the VM’s current state, apply or revert to checkpoint 3, then remove the later checkpoints. The changes from checkpoints 4–6 will be discarded by reverting, while deleting the checkpoints afterward cleans up the chain.
If you need to preserve checkpoint 3 as a separate, clean VM without risking the original checkpoint chain, export checkpoint 3 as a new virtual machine and start the exported copy. Once you have verified that it works, you can remove the original VM and its checkpoints.
The word “apply” can be confusing here. Applying a checkpoint does not necessarily mean rolling the VM back; it commits the checkpoint’s disk blocks into the virtual disk. A checkpoint is a point in a differencing-disk chain, not an independent full copy that can simply be deleted without processing its data.
So deleting checkpoints 4–6 after reverting to checkpoint 3 still performs merges, rather than simply removing those files?