StarWind Software

Hello,
I have an infrastructure of:
- 2 Windows Servers (Storage1 - Primary; Storage2-Secondary)with Starwind on them. The software creates 2 LUNS: 1 quorum and 1 LUN for an File Scale Out Share.
- 3 Hyper-V Servers that are using the File Scale Out Share

This is all the infrastructure. ( all servers, Storage, Hyper-V, and others are join to same domain)
On the Hyper-V there are some VM`s but also 2 Domain Controllers ( there is no other Domain Controller in the infrastructure).
If i restart Storage1 all the VM`s crash with BSOD, storage connection is dead.

I have some idea why but i want to know your point of view.

Thank you,
Adrian

Not related to your design but having one CSV for two SoFS nodes is not recommended. It's better to have at least one CSV per SoFS node. Performance reasons: SoFS nodes don't "fight" for LUN/CSV ownership. In your case VMs can BSOD if failover happens to slow and VMs don't get their I/O traffic for a while: VMware would just freeze VMs and Hyper-V would do the same only in Windows Server 10. We'd love to get hands over StarWind logs on SoFS nodes to see why LUN failover happens too late.