VM Power on fails with error "Unable to create virtual SCSI device" Module DevicePowerOn power on failed.

Products

• VMware vSphere ESXi 7.0
• VMware vSphere ESXi 8.0

Issue/Introduction

• Unable to power on the secondary node within the Windows Server Failover Cluster (WSFC).
• WSFC nodes are hosted on a datastore with Clustered VMDK capabilities enabled.
• VM power on fails with the below error:
  "Unable to create virtual SCSI device for scsi1:0, '/vmfs/volumes/<datasore-UUID>/VM_Name/VM_Name_1.vmdk' Failed to start the virtual machine. Module DevicePowerOn power on failed."
• The vmware.log will report below errors:
  
  YYYY-MM-DDTHH:MM:SS In(05)+ vmx - Power on failure messages: Unable to create virtual SCSI device for 'scsi2:0' (/vmfs/volumes/datastore_name/ExampleVM_1/ExampleVM_1.vmdk).
YYYY-MM-DDTHH:MM:SS In(05)+ vmx - Module 'DevicePowerOn' power on failed.
YYYY-MM-DDTHH:MM:SS In(05)+ vmx - Failed to start the virtual machine.

Environment

VMware vSphere ESXi 8.x
VMware vSphere ESXi 7.x

Cause

A Cluster-in-a-Box (CIB) configuration is incompatible with the Clustered VMDK datastore capability.

Resolution

To resolve the issue, perform the following steps:

• Determine the current host of the Primary Node.
• Migrate the Secondary WSFC Node to any available host within the cluster, ensuring it is not residing on the same physical host as the Primary Node.
• Attempt to power on the VM once the migration is complete.

Note: In the absence of VM-Host anti-affinity rules, Distributed Resource Scheduler (DRS) may attempt to place the VM on the same host as the primary node. If this occurs, manually migrate the affected WSFC node to an alternate host and initiate the power-on sequence directly via the host client.

Additional Information

Make sure all the guidelines and supported configurations are matched as per Microsoft Windows Server Failover Clustering (WSFC) with shared disks on VMware vSphere 8.x: Guidelines for supported configurations

Limitations of Clustered VMDK support for WSFC:

Clustered VMDK support is only possible with the following limitations:

• Cluster in a Box (CIB) configuration is not supported.
• Spanned VMFS datastores cannot store clustered VMDKs.
• A physical LUN with a clustered datastore cannot have any additional VMFS datastores.
• On-line Storage vMotion and snapshots are not supported.
• Only disks with 512 sector size are supported.
• VMFS5 and earlier are not supported for storing clustered VMDKs.
• Only datastores connected through Fibre Channel (FC) are supported in this release.
• Clustered VMDK datastore cannot be used as an ESXi installation disk.
• Diagnostic partition is not supported on disks that are marked clustered.
• SCSI-2 reservations are not supported on clustered VMDKs.
• Hot expansion of a VMDK that is associated with a clustered VM is not supported.
• Moving more than one node of WSFC to the same ESXi host is not allowed and causes vMotion to fail. You must use anti-affinity DRS Rules to separate VMs and nodes of the cluster on different ESXi hosts.
• Cloning and Fault Tolerance are not supported.

Reference Document: Limitations of Clustered VMDK support for WSFC