I have a 2 node HyperV 2012 R2 cluster sitting on top of shared JBOD storage. I am virtualizing A standalone deployment of SQL Server 2014 Standard on this platform.

To isolate and provide separation of aspects like csv readcache from other workloads I am considering separate CSV’s for: Data, logs, temp and possibly even putting the server on it’s own segregated from the other VM’s.These volumes would be implemented on the same pool as csv’s for other workloads though so there is the possibility of redirection if ownership change for some reason. The pool needs to be accessible to both nodes, but I could only allow the VHD’s to be owned by the server hosting the SQL server, I am limited in this aspect due to the other csv’s on this pool needing to be able to change ownership.

First off, is it truly beneficial to create the volumes at this level? Theorectically it seems it would as technically these are separate LUN’s.

Second, is there any m4 TBajor concerns especially with the redirection scenario, or to force avoiding redirection by only allowing the volumes to be owned by the host that SQL server is on?

eg:

• 4 TB Pool on JBODs managed by HyperV 2012 R2 cluster

• 80 GB CSV for SQL Server

• 25 GB CSV for Temp

• 125 GB CSV for Logs

• 350 GB CSV for Data

• Remaining pool space is a CSV for other workloads

Any input is greatly appreciated,

Thanks,

Chuck C