Error Message

"V-76-58645-585 Failed to reserve a majority of disks in cluster dynamic disk group".

"V-76-58645-585 Failed to reserve a majority of disks in cluster dynamic disk group" is because Node1 has taken majority to disks/SCSI reservations and there was nothing left for node 2. However due to VMware complexity the VMNSDG would be the perfect resource type to use for this scenario with "DG=Dymanic Disk group"

The virtual disk or the virtual SCSI bus simply don't allow SCSI reservations. Windows Failover Cluster requires SCSI3 reservations, so the same rules that apply for creating a Windows Guest Cluster on VMWare apply for your
VM: Setup for Failover Clustering andMicrosoft Cluster Service

VxVM (Veritas Volume manager) was left to SCSI3 and with these settings in place, it should not create issues to import the disk group as cluster dynamic disk group. The key seems to be that the disk needs to run on a 2nd SCSI controller that is set to sharing = Virtual and then it should work as expected. See chapter 2 from page 16 of the above PDF.

Any one of the following should resolve the issue migrate from  Dynamic DG to Clustered DG (on Vmware disks)

A) Solution 1:

1. Example: On each VM we use two disks: Hard disk 0 - SCSI(0:0) and Hard disk 1 - SCSI(1:1). Hard disk 0 is system disk. Hard disk 1 is for clustered applications.

Add Hard Disks to the First Node for Clusters on One Physical Host (Page 18-19) - for creating Hard disk on Node 1
Add Hard Disks to Additional Nodes for Clusters on One Physical Host (Page 19) - for creating Hard disk on Node 2

Using this parameters for new disks:

SCSI controller 1 --> SCSI Bus Sharing = Virtual
Harddisk X --> Thick provision eager zeroed & Sharing = Multiwriter & SCSI (1:0) to make sure it runs on the shared SCSI controller

2. Online and write signature via VEA, and Configure GPT mode
3. Add new hard disk 2 to existing dg
4. Create a mirror volumes
5. Break mirror after sync completes
6. Split DG and use new plex
7. Remove hard disk 1 from dg (PLEASE KEEP THIS ON HOLD JUST IN CASE IF ANY ISSUE OCCURS, DO NOT REMOVE DISK FOR ATLEAST SOMETIME)
8. For Node 2: deport dynamic dg and import as cluster dg

B) Solution 2:

Storage configuration in a VMware virtual environment involves the VMware vmdk and the RDM disks that reside on a shared datastore. To access the disks, a virtual machine uses virtual SCSI controllers that can have None, Physical, or Virtual type of Bus Sharing configuration.
Depending on the type of SCSI controller and the Bus Sharing configuration, virtual machines that are configured on same or different hosts can simultaneously access the same virtual disk.

To address the type of clustered-environment you have configured, you can adjust the Bus Sharing configuration set and enable or disable the Multi-writer protection.

What type of Bus Sharing configuration does InfoScale support to leverage VMware vMotion?

To leverage VMware vMotion, InfoScale supports cluster configuration with pRDM disks that have None or Physical type of Bus Sharing configuration.
The following table provides the supported Bus Sharing configurations with respect to the corresponding ESX host version and the type of disks used.

1. Example: On each VM we use two disks: Hard disk 0 - SCSI(0:0) and Hard disk 1 - SCSI(1:1). Hard disk 0 is system disk. Hard disk 1 is for clustered applications.

Add Hard Disks to the First Node for Clusters on One Physical Host (Page 18-19) - for creating Hard disk on Node 1
Add Hard Disks to Additional Nodes for Clusters on One Physical Host (Page 19) - for creating Hard disk on Node 2

Using this parameters for new disks:

SCSI controller 1 --> SCSI Bus Sharing = Physical
Harddisk X --> Thick provision eager zeroed & Sharing = Multiwriter & SCSI (1:1)) to make sure it runs on the shared SCSI controller

Note: In VEA settings SCSI Support = SCSI-3 support.

2. Online and write signature via VEA, and Configure GPT mode
3. Add new hard disk 2 to existing dg
4. Create a mirror volumes
5. Break mirror after sync completes
6. Split DG and use new plex
7. Remove hard disk 1 from dg (PLEASE KEEP THIS ON HOLD JUST IN CASE IF ANY ISSUE OCCURS, DO NOT REMOVE DISK FOR ATLEAST SOMETIME)
8. For Node 2: deport dynamic dg and import as cluster dg