I'd like to have a Hyper-V host that uses local storage for it's VMs. At the moment I have a VMWare ESXi host with a couple of 2TB SATA disks inside and I'm looking to upgrade it.
My new server supports 4x hard drives, and has an oldish RAID card. So I was about to configure it with a RAID10 and 4x 2TB hard drives, configure a big volume on the RAID card, then install Server 2012 R2 + Hyper-V role, and call it a day. Performance isn't great, but it's also not terrible.
However... Windows Server 2012 R2 supports Storage Spaces. My understanding of this is that I could ditch the RAID card, and let Windows take the collection of physical disks (SSDs and/or HDDs), arrange them into a Storage Pool, create virtual disks from that pool with varying levels of resiliency, and I can use tiered storage to allocate a certain amount of space to the faster SSD storage from the pool. Windows is supposed to handle in software moving the most frequently used data to the SSDs.
This is really tempting to me to use together with Hyper-V. I'd like to be able to take 2x 2TB hdds, and 2x 500GB enterprise SSDs, throw that into the server, and let Server 2012 storage spaces to create storage pool from that. It would create a virtual disk from that storage pool, format it, assign it a drive letter, and then Hyper-V can store all of it's VHDX disks on there for the VMs.
But it does seem like there are too many layers of virtual disks - Hyper-V storing a VHDX file inside of a virtual disk which is part of Storage Pool.
Is this supported / advisable?
Or does the "Storage Server" really need to be a separate dedicated box from the Hyper-V host and I use technologies like iSCSI to connect the two?