vCLS VMs Unable to Power On with "No Host Compatible" Error Due to MWAIT CPU Feature Requirement

VMware vCenter Server

vSphere Cluster Services (vCLS) virtual machines fail to power on with the error message "No host is compatible with the virtual machine" in a vSphere cluster. When attempting to power on the VMs, detailed error messages indicate "MWAIT is unsupported" and "Feature 'cpuid.mwait' was 0, but must be 1." This prevents proper functionality of HA/DRS features in the affected cluster.
For ESXi 8.0 U3 and above in the infravisor logs of the ESXi host below log traces are observed :
/var/run/log/infravisor.log
YYYY-MM-DDTHH:MM:SSZ No(5) infravisor[#######]: time="YYYY-MM-DDTHH:MM:SSZ" level=error msg="unexpected fault: &{{{{<nil> [{{} msg.featurecompat.requirement.number.mismatch [{{} 1 cpuid.mwait} {{} 2 0} {{} 3 1}] Feature 'cpuid.mwait' was 0, but must be 1.} {{} msg.moduletable.powerOnFailed [{{} 1 FeatureCompatLate}] Module 'FeatureCompatLate' power on failed. } {{} msg.vmx.poweron.failed [] Failed to start the virtual machine.}]}}} Feature 'cpuid.mwait' was 0, but must be 1.} taskerror: Feature 'cpuid.mwait' was 0, but must be 1." VM-OP=PowerOn namespace=vcls pod=vcls-79e9f###-####-####-####-########## uid=######################

The issue occurs when the host hardware's CPU feature support is not properly configured at the BIOS level. Specifically:

The "monitor/mwait" CPU feature is disabled in the server BIOS settings
C-States power management settings are not properly configured for "autonomous" operation
Outdated server firmware may contribute to the problem

These missing CPU features prevent vCLS VMs from powering on, as these VMs require specific CPU capabilities to function properly.

To resolve this issue, follow these steps:

Identify the current BIOS version and settings on the affected hosts a. Note the current BIOS version from the host hardware information in vCenter b. Check if the version is significantly outdated (more than 1-2 years old)
Work with the hardware vendor to: a. Identify the appropriate firmware update path for the servers b. Obtain the latest BIOS update package for the server model c. Verify the recommended settings for CPU power management features
Schedule maintenance to update BIOS firmware and settings on all affected cluster hosts: a. Update the BIOS firmware to the latest supported version b. Enable the "monitor/mwait" option in the BIOS settings c. Set the "C-States" option to "autonomous" in the BIOS d. Save the settings and reboot the hosts
After implementing these changes, test if vCLS VMs can power on successfully: a. If the cluster was previously set to a "retreat mode", set it back to normal mode b. Verify that vCLS VMs are able to power on without errors c. Confirm that HA/DRS functionality is working properly in the cluster

To temporarily work around this issue, place the cluster in "retreat mode" and then back to normal mode. However, this will likely not resolve the underlying issue if the BIOS settings remain unchanged.
The error message "Feature 'cpuid.mwait' was 0, but must be 1" specifically indicates that the MWAIT CPU instruction set feature is required but not available.
This issue may occur after enabling HA/DRS on clusters where these features haven't been used previously.
For more information on vCLS VMs and their requirements, see vSphere Cluster Services (vCLS) in vSphere 7.0 Update 1 and newer versions
To validate the cpu id configuration refer DRS processes fail to Run: "vSphere DRS functionality was impacted due to unhealthy state vSphere Cluster Services caused by the unavailability of vSphere Cluster Service VMs. vSphere Cluster Service VMs are required to maintain the health of vSphere DRS"