I am planning to run two-node hyper-converged cluster, Hyper-V + Starwind VSAN. I assume, there will be two RAID10 4x4TB HDD on each server for storing HA-device files. Sync lync supposed will be 10GBASE-T Ethernet crossover (w/0 switch).
I'm not clear in this points:
1) I will format RAID10 luns in GPT + NTFS for storing thick-provisioned devices.
- Which cluster size should i use for NTFS volume?
Which cluster size should i select for HA device?
What is NTFS ACL best practice considerations in this case (no other files on volume, only HA-device images)?
3) And now i will create a vhdx file, mount it into VM, formatting it in NTFS. Which cluster size should i use?
And one more question. One of vhdx will be large storage for file server (4TB, CAD files, documents). Is it good practice to have file-server storage on virtual disk? Before that I always used local storage or iSCSI LUN for such tasks...