Scenario: SCCM 2012 R2 SP1 CU2 Primary Site server starts having a bad day, puking up hairballs of component status alerts, inboxes are piling up, memory and processor are pegged. Now, instead of actually troubleshooting the source, you turn off the VM guest which hosts it (not a shut-down; a powering-off). And then… (drum roll please…)… revert the VM guest to a snapshot from two (2) months earlier. Power back on, and boom! It’s broken.