Adding a new hard disk on a clustered virtual machine with multi writer enabled fails with the error "Incompatible device backing specified for device '0'"

Products

• VMware vCenter Server
• VMware vSphere ESXi 7.0
• VMware vSphere ESXi 8.0
• VMware vSphere ESXi

Issue/Introduction

Symptoms:

• The error message can be observed after the virtual machine(s) were migrated from a vSAN datastore to a VMFS datastore.

• Windows Server Failover Cluster (WSFC) node(s) fail to power on simultaneously.

• The operation to add a new hard disk to the first node of the clustered VMDKs on a VMFS datastore failed with the following error: Incompatible device backing specified for device '0'

• The following error message was recorded in the /var/run/log/hostd.log on ESXi host in two separate instances:

  • Disk provisioning method is set to Thick Provision Lazy Zeroed when adding a new disk:

YYYY-MM-DDTHH:MM:SS Er(163) Hostd[2100013]: [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/vsan:################/################/VM_Name.vmx opID=################ sid=######## user=vpxuser:username] ValidateDiskFileBacking: Virtual disks in multiwriter mode cannot be Thick Provision Lazy zeroed
YYYY-MM-DDTHH:MM:SS In(166) Hostd[2100013]: [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/vsan::################/################/VM_Name.vmx opID=################ sid=######## user=vpxuser:username] Reconfigure failed: N3Vim5Fault20InvalidDeviceBacking9ExceptionE(Fault cause: vim.fault.InvalidDeviceBacking
YYYY-MM-DDTHH:MM:SS In(166) Hostd[2099944]: --> )

• • Disk provisioning method is set to Thin Provision when adding a new disk:

YYYY-MM-DDTHH:MM:SS Er(163) Hostd[2100013]: [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/vsan:################/################/VM_Name.vmx opID=################ sid=######## user=vpxuser:username] ValidateDiskFileBacking: Virtual disks in multiwriter mode cannot be thin provisioned.
YYYY-MM-DDTHH:MM:SS In(166) Hostd[2100013]: [Originator@6876 sub=Vmsvc.vm:/vmfs/volumes/vsan::################/################/VM_Name.vmx opID=################ sid=######## user=vpxuser:username] Reconfigure failed: N3Vim5Fault20InvalidDeviceBacking9ExceptionE(Fault cause: vim.fault.InvalidDeviceBacking
YYYY-MM-DDTHH:MM:SS In(166) Hostd[2099944]: --> )

• Verify the disk provisioning type selected for the new hard disk.: Add Hard Disks to First Node for Cluster Across Physical Hosts with Clustered VMDKs on VMFS
• Current Disk Config:
  
  

Environment

VMware vCenter Server 7.x

VMware vCenter Server 8.x

VMware vSphere ESXi 7.x

VMware vSphere ESXi 8.x

Cause

The migration resulted in unsupported disk configurations for WSFC on VMFS:

• Incorrect Format: Shared disks were "Thick Provision Lazy Zeroed," but VMFS requires Thick Provision Eager Zeroed for SCSI reservations

• Disable the Multi-writer on the WSFC VMs if it is enabled

Resolution

To successfully add a new hard disk to the clustered VMDK primary node on VMFS datastores, Thick Provision Eager Zeroed must be selected for disk provisioning, as this is the only supported configuration.

Steps to make reconfigure the VM with the necessary changes:

1. Convert Disk Format: Perform a Storage vMotion and select Thick Provision Eager Zeroed as the destination format 

2. Disable Multi-writer:

   1. 1. Power off the cluster nodes 
      2. Edit VM settings and disable the Multi-writer flag for shared disks 
      3. Set Sharing to No Sharing

3. SCSI Bus Sharing: Ensure the SCSI controller is set to Virtual (for nodes on the same host) or Physical (for nodes on different hosts)

4. Reconnect: Re-attach the disks to the secondary node using identical SCSI IDs as the primary node

References:  
Add Hard Disks to First Node for Cluster Across Physical Hosts with Clustered VMDKs on VMFS
Microsoft Windows Server Failover Clustering (WSFC) with shared disks on VMware vSphere 8.x