I'm doing R&D at a small organization. I have 4 servers, two for storage, and two as VM hosts using Windows Server 2016 with Hyper-V. Currently, the Sysadmins use VM's as file servers on a Nimble brand (black box) storage device. The goal is to use two generic servers as storage devices, and two as VM hosts. I'm not sure if it makes sense to use the storage devices to hold both the data itself, and the virtual machines.
Anyway, I've set up the VM side by storing them on C:\clusterstorage using SW HA devices replicated between both storage servers, and setting up a Failover Cluster with the virtual servers as nodes. So the VM's have outstanding redundancy, meaning I can shut down one of the storage servers (which hold the VM's) or one of the Virtual Servers (which are nodes on the cluster) and have full fault tolerance. Magnificent! The next step is using those VM's as file servers, with their data stored on the same two HA storage device nodes that also house the VM's. Is that even a good practice?
Can I just create an additional HA device on the storage servers, and point the VM's, via iSCSI, to that? Should I somehow store the VM's on the Virtual Server hosts using failover cluster manager, and point them to the (replicated) iSCSI target on the file servers? In other words, break it up, and store the VM file servers on the VM hosts, and the data on the dedicated storage devices. Could DFS come into play?
My head is swimming. I'm just an intern, doing R&D for busy Sysadmins at a community college handling about 4TB of actual data. I think this is a perfect use case for SW though, because we have enough capital to buy appropriate generic server hardware, and Windows Server licenses, but don't want to spend like 80 grand on a proprietary solution.
The Latest Gartner® Magic Quadrant™Hyperconverged Infrastructure Software