I have 8 vms on ESXi 5.1 hosts that are thin provisioned. 2 of those vms are being reported to Unitrends as having block changes across the entire disk, and therefore Unitrends is showing backup sizes of 235GB (the provisioned size) rather than the 100GB that is actually used. The other 6 thin provisioned vms were showing backup sizes of only the used space, not the full provisioned space.
I have taken this to Unitrends support already and had the issue escalated one level before I got the explanation. What Unitrends support told me is that vSphere is reporting changed blocks on the entire provisioned disk, therefore it's backing up the entire provisioned size. Unitrends IS compacting it all down (the 100GB thin disk is compacted down to only 52GB on the Unitrends appliance), however the full 235GB is the number used against my backup capacity.
Support mentioned this will sometimes occur if changes to the vm are made like adding or removing or expanding or contracting the hard disk. I do recall having expanded one of the hard disks, so I figured that was the culprit.
When support was helping me and explained what was happening, I decided to leave it alone as I didn't anticipate those 2 vms growing past the provisioned sizes any time soon (if ever), and the suggested way to fix it was to create a new LUN and migrate the vm. besides I was still 500GB under my capacity cap on the Unitrends appliance.
Now, I suddenly have a 3rd vm that ballooned overnight from a 50GB full to 800GB full. Now my appliance is over capacity and de-duplication is turned off. No changes were made to that vm, and I can't find anything in vSphere logs nor windows logs nor performance monitors that would indicate why the disk is suddenly acting like it's had changes over the entire 800GB provisioned space.
Has anyone else had an issue like this? Is there some way to get vSphere to quit reporting to Unitrends that the entire provisioned space has changed?