WSFC cluster VMs fails to power on with "Operation failed! Unable to create virtual SCSI device for 'scsix:x'

Products

• VMware vCenter Server

Issue/Introduction

Powering on one of the Windows Server failover cluster (WSFC) nodes alone the VM can be powered on normally, but trying to power on the second node it fails with the error:

Operation failed!

Task name Power On virtual machine

Target NODE2

Status Unable to create virtual SCSI device for 'scsi1:0' (/vmfs/volumes/xxxxxxxxx-xxxxxxxx-xxxx-xxxxxxxxxxxx/NODE1/NODE1_3.vmdk).

 

More details of the failing task:

Failed to start the virtual machine.

Module DevicePowerOn power on failed.

 

vmware.log will show this:

2025-01-25T09:45:29.904Z In(05) vmx - Vix: [mainDispatch.c:1164]: VMAutomationPowerOff: Powering off.
2025-01-25T09:45:29.904Z Wa(03) vmx - /vmfs/volumes/xxxxxxxxx-xxxxxxxx-xxxx-xxxxxxxxxxxx/NODE2/NODE2.vmx: Cannot remove symlink /var/run/vmware/0/13353971153_2203356/configFile: No such file or directory
2025-01-25T09:45:29.905Z In(05) vmx - WORKER: asyncOps=7 maxActiveOps=7 maxPending=7 maxCompleted=0
2025-01-25T09:45:29.920Z In(05) vmx - Vix: [mainDispatch.c:4206]: VMAutomation_ReportPowerOpFinished: statevar=1, newAppState=1873, success=1 additionalError=0
2025-01-25T09:45:29.920Z In(05) vmx - 
2025-01-25T09:45:29.920Z In(05)+ vmx - Power on failure messages: Unable to create virtual SCSI device for 'scsi1:0' (/vmfs/volumes/xxxxxxxxx-xxxxxxxx-xxxx-xxxxxxxxxxxx/NODE1/NODE1_3.vmdk).
2025-01-25T09:45:29.920Z In(05)+ vmx - Module 'DevicePowerOn' power on failed.
2025-01-25T09:45:29.920Z In(05)+ vmx - Failed to start the virtual machine.
2025-01-25T09:45:29.920Z In(05)+ vmx - 

 

 

 

Environment

vSphere 8.0

Cause

This was likely caused due to incorrect shared disk configuration. See more on, Microsoft Windows Server Failover Clustering (WSFC) with shared disks on VMware vSphere 8.x: Guidelines for supported configurations.

Resolution

The shared disks of the clustered VMs must be reconfigured in the secondary nodes that are failing, for that do these steps:

1. Make sure all the settings and requirements of the Microsoft Windows Server Failover Clustering (WSFC) with shared disks on VMware vSphere 8.x: Guidelines for supported configurations are correct.
2. Remove all the shared disks from the VM related to the NODE2 and other nodes such as NODE3, NODE4, etc. if existing (not deleting the data).
3. Power on all the VMs of the cluster.
4. Validated the VM related to the NODE1 is working fine (the other VMs will have only their own disks, not the shared disk)
5. Add the shared disks to the VMs of the NODE2,3,4... following the same settings define in the VM of the NODE1.
   1. Important! Observe the Virtual Device Node for the SCSI controller and disk number to be correct according to the requirements of the operating system to associate the disks internally on the guest OS cluster level.
6. Use the Disk Management to validate all disks will show correctly on all nodes
   1. Note that on Windows only the active node of the cluster will show the disks online and the other nodes will show them as offline, that is a normal condition as the clustered application will control the usage of the shared disks.