I've been doing a lot of migrating between CSV's in our Hyper-V 2016 Cluster, and it's getting a pain to run space-reclamation, so I looked up to see options to automate it, and it turns out, there's built in feature called space-allocation that will do it on the fly. Now my understanding is that for ease of implementation and non-disruptive behavior, it'll make sense to create new CSV's set with the option and then migrate my VM's to the new CSV. No problem with that. But I see that enabling the feature will cause a performance hit. I have five approximately 2TB each CSV's stored on my AFF-A200. Would there be noticeable performance decrease? Does anyone have any real world comparisons of this? What would be the performance counters to check?