3 Replies Latest reply on Jul 29, 2019 5:19 AM by MichaelLeone

    ESXi 5.5, Microsoft Clustering - cluster validation fails with SCSI page 83h VPD descriptor error

    MichaelLeone Enthusiast

      I have to create a MS cluster on a ESXi 5.5 U3b cluster (I know, it's old, but that's the assignment). I can successfully create the 2 Win 2016 VMs, create the RDM disks; share them; create the cluster; even join the RDM disks as available storage to the cluster. All that is fine.

       

      The cluster validation test is failing with "Physical disk 1 on <both nodes> does not support the inquiry data (SCSI page 83h VPD descriptor) that is required by failover clustering.

       

      Now, here's the thing. Physical disk 1 on both nodes is a *local* disk - i.e., it's not intended to be used by the cluster at all. It's drive D: on those machines; it's where you install the application that will use the cluster (in my case MS SQL 2016). So I don't know know why it cares about a disk that will in no way be used as a cluster resource  ...

       

      Those disks are defined as "thick provision lazy zeroed"; could that be the issue? Is it that any drive that isn't the boot drive must be "thick provisioned eager zeroed"?

      (I can't test right now, it's doing Windows Updates. and if you've ever done WU on a Win 2016 server, you know that can take hours, even with only 3 cumulative updates)

       

      I definitely need a 2nd local disk for application installation, otherwise I would just delete the disk and go. As I say, my 4 cluster disks are all configured, waiting in "Available Storage". I can even move them from one node to another. I just can't pass the validation test, and that test is required, before SQL will install as a SQL cluster.

       

      It's got to be something simple I've missed, but what?

       

      Thanks for any help.