There's two legit ways to go.
1) Turn on 'HBA mode' for your RAID controller to make it compatible with Microsoft Storage Spaces Direct (S2D).
According to the official HPE's B140i User Guide, Page 6, "Enabling the controller" section.
https://support.hpe.com/hpesc/public/docDisplay?docId=emr_na-c04441385
Enabling the controller
The SATA chipset used by the HPE Dynamic Smart Array B140i RAID controller may be configured as a SATA AHCI controller, or it may be configured as a B140i controller. On some platforms, SATA AHCI mode is the default, and the B140i controller may need to be enabled before use. To enable the B140i controller:
1. Reboot the server. The server starts up and the HPE ProLiant POST screen appears within a few minutes.
2. Press the F9 key in the ProLiant POST screen. The 'System Utilities' screen appears.
3. From the 'System Utilities' screen, select 'System Configuration' → 'BIOS/Platform Configuration(RBSU)' → 'System Options' → 'SATA Controller Options' → 'Embedded SATA Configuration' → 'Enable HPE Dynamic Smart Array RAID Support', and then press the Enter key. <-- This is where you should select 'Disable' for your particular case!!!
4. Ensure that you are using the correct AHCI or RAID system drivers for your SATA option.
5. Select a setting and press Enter:
a. Enable 'SATA AHCI Support—Enables' the embedded chipset SATA controller for AHCI.
b. Enable 'HPE Dynamic Smart Array RAID Support—Enables' the embedded chipset SATA controller for Dynamic Smart Array RAID.
6. Press the F10 key to save your selection.
2) Use proper software-defined storage (SDS) stack that gets along well with hardware RAID.
I'd suggest StarWind Virtual SAN as these guys have Windows version out of box.
https://www.starwindsoftware.com/vsan
If you combine RAID with replication on top, you'll get by far better overall cluster resiliency and faster rebuild times in case shiitake will hit the fan. Moreover, combining local reconstruction codes (equivalent to the software RAID) with inter-node replication is what Microsoft does in Azure.
https://www.usenix.org/system/files/conference/atc12/atc12-final181_0.pdf
Using single-disk RAID0 is a dirty hack that might and might not work. See, in RAID mode there's thin but important firmware virtualization layer present, reservations, cache... All these things influence S2D (ReFS, ZFS and other guys asking for 'raw' access) behavior. Old Windows WHQL kit had a special ReFS test application ensuring atomic writes are really atomic. Making long story short: some RAID controllers (like Del PERC H730) in RAID0 mode do pass this test, while some other ones - don't... Even with the on-board cache forcefully disabled. Don't play dice with your production data! It might not appreciate it.
