Failover Clustering CSVs - Hyper-V RRS feed

  • Question

  • Hi all,

    Need some help, 

    We have a 6 Nodes Cluster (S2D), we have 6 CSVs (5TB) all assigned to a node each. Example below

    Node1 = CSV1

    Node2 = CSV2

    Node3 = CSV3

    Node4 - CSV4

    Node5 = CSV5

    Node6 = CSV6

    We use the CSVs for HyperV storage, each part of the VM will be stored on a single CSV eg. Vhdx, XML etc.

    within our setup we have mini RDS farms, 1 x BrokerServer, 1 x SQL Server and 2 x RDS Hosts and a DC sat servicing them all. The Broker holds the User Profile Disks and shared folders, SQL obviously holds database and Hosts are TS's.

    Another server which services multiple RDS mini Farms hosts a database file in MDB format, the RDS Hosts have the application installed and connect to the .mdb file.

    If the RDS Host Server is not running on the same physical Hyper-V node as the VM hosting the .mdb then we have shocking performance, a task that should take a 1 min is actually taking 20 min. 

    I don't think this is specific to this applicaition and believe that its affecting our whole environment, I think a major impact is that our Broker (User Profile Disks Share) could be on a different Node to the RDS Host which will also cause issues. We are looking at ensuring all the mini farms run from the same server Node/CSV but with Shared servers this isn't going to work.

    All Nodes are connected via a 25GB network backbone and copying files Node to Node is very fast.

    Feel free to quiz me on setup

    Monday, January 20, 2020 10:17 AM

All replies

  • If you still need help, I'd suggest posting in the Hyper-V forum here:

    This one is for Virtual Server 2005

    Tuesday, January 21, 2020 6:24 PM