A Prism Central capacity view shows provisioned (allocated) capacity at 95% of the cluster while physical used capacity is only 50%. The VMs use thin-provisioned disks and RF2. Which statement explains the difference?

Nutanix vDisks are thin-provisioned: a disk's configured size is a promise, and physical space is consumed only when data is written. Provisioned or allocated capacity adds up the configured sizes of all disks, so it can approach or exceed the cluster's usable capacity without anything being wrong. Physical used capacity counts the data actually stored, including RF overhead (two copies under RF2), snapshots and metadata, after any compression, deduplication or erasure coding savings. In Prism, logical usage means data before RF overhead, so with RF2 physical usage is normally larger than logical usage, not smaller; the gap in this stem comes from provisioning versus writing, not from the RF calculation. The claim that RF2 and thin provisioning duplicate every provisioned byte is wrong because replication applies only to written blocks. The claim that logical capacity is calculated after compression and deduplication reverses the definitions, because data reduction lowers physical consumption. Saying snapshots and metadata make logical capacity always exceed physical has the relationship backwards for RF2 clusters and ignores provisioning altogether. Exam caveat: watch which capacity term a question uses: provisioned, logical and physical each answer a different question. Operational check: compare provisioned sizes with physical usage and runway in Prism Central capacity planning before declaring the cluster full.