First real quick: ESX 6.5 (7967591), FCoE, NetApp backend, Cisco UCSB (Flexpod configuation)... OS: Windows 2012 R2
So, the cluster is being setup to support SQL (nit installed yet)... For my disk layout, C: (OS), D: (SQL Binaries) and Shared E: (Bkup), G: (Data), L: (Logs), T: (Temp), Q: (Qrm)
Shared on one SCSI controller and "local" on another. Both Paravirtual, Physical, etc.
** I used VMWare's published Windows Clustering guide for CAB (cross across boxes)
When running validation wizard, I get
As per VMWare support:
++ As per the best practices of Cluster Virtual Machines Across Physical Hosts on VMware vSphere 6.5 , we can have only one Virtual Disk mapped (booting) rest all should be P-RDMs
And I can map through D: as an RDM (which is what I am bein told to do), but it makes no sense to me... And yes, if I drop my D: drive, the problem goes away. And yes, I can install SQL to C:...but really??? I can only have one virtual disk!
What if I wanted to leverage my SDD local datastore for TempDb... a common scenario. I couldn't?
D: was not part of the Available Storage disks or any cluster resource
D: shows up as a Local Disk
Tried changing SCSI to LSI SAS... still no....
Note: NetApp on the backend;
SO........... Anyone ever heard of a restriction on the Virtual Disk count. I can't find any documentation and vmware support just says it's common practice????
Because what you're attempting here is not common as it appears you're doing this on a standalone host using host-local drives. It's definitely not common and is not at all a good idea, so no, you probably won't be able to do that. Secondly, you're attempting all this on a gross out-dated ESXi 6.5 version. At the minimum you need to bring it up to the current patch rev.