Best answer by f3rm1n
The reason is that the container has a replication factor (2 by default) hence the total physical capacity available to the hypervisor is half of the total physical capacity.
Furthermore, once you start deploying VMs, you need to take into account storage reservations:
When creating VM's with thick provisioned disks, ESXi requests the whole space on the Nutanix storage container for those virtual disks. NOS calculates this reservation as an implicit reservation, only actual data blocks written will use physical space, Prism container usage will display the whole amount of disk space requested as Reserved Capacity.
In the Prism Storage Diagram View, the following metrics refer to implicit reservations:
Note: HyperV and KVM do not use Implicit Reservations, so this issue does not apply.
•Reserved Capacity (Physical): This is the implicit reservation for VMs with thick virtual disks•Free (Reserved) [NOS 3.5: Thick Provisioned (Not Used)]: Space reserved but not used by thick virtual disks, it cannot be allocated for new VMs.•Free Unreserved (Physical): Available disk space for new VMs.•Used (Physical): Amount of disk space used by VMs, including implicit reservations.
The effective max capacity (E) is computed as follows: E = (T - (sum of container reservations and Used physical)) / RFWhere T is the total storage pool capacity, RF is the Replication factor.
Also, in NOS 4.1+, the Web Console will display the summary reported to vSphere in the lines labeled as "Logical" (Free, Used, and Max Capacity).There will be a slight difference on the final number due to the fact that vSphere calculates capacity by a power of 2 (i.e. 1KB=1024B), and Prism uses a power of 10 (i.e. 1KB=1000B).
For the most accurate representation of storage usage, please refer to vSphere storage summary, the Free Space reported there will be the actual amount available for new VMs.