I'd like to add my experience with my solution as well. I followed the steps outlined in the KB linked above and where I had some success with some of my VMs that I got the 3948 error with, others were still giving me grief. When following the above KB, I had take the final step of cloning the VM to consolidate all the helper snapshots, but like I said that only fixed some of the VMs.
After further digging I discovered that my VDR appliance wasn't letting go of the linked-clone VMDKs it had created of the pre-cloned VMs. In the VDR appliance, I deleted all the pre-cloned VM backups, then I had to shut down the VDR appliance and manually remove the extra VMDKs from its configuration and then SSH into my ESXi remote console to delete the old VMDK files. Once I cleared out all the pre-cloned VMDKs from my datastore and VDR appliance, I powered on my VDR appliance and then the backups of my cloned VMs started completing successfully without the 3948 errors. I hope that helps someone else too.
