Passing physical disk is a very bad idea. It's like throwing away an engine from your car because you don't know how it works. The whole caching and virtualization layer
we wrap around the physical disks is there not just for fun it actually does I/O acceleration. While it's possible to cache physical disks the whole LSFS thing cannot function w/o
virtual disks (files) for now... The same about Scale-Out. StarWind with a single node is a rudimentary scenario. Whole bunch of things (like for example I/O spoofing with a huge
amount of a write back flash and RAM cache) are either not possible or are dangerous because of a single controller used.
The Latest Gartner® Magic Quadrant™Hyperconverged Infrastructure Software