Enabling vSphere HA fails with error: "Cannot enable vSphere HA VM Component Protection for the specified cluster, because it contains a host {hostName} with APD Timeout is disabled."

Products

• VMware vSphere ESXi

Issue/Introduction

• Enabling vSphere HA fails with error:
  
  Cannot enable vSphere HA VM Component Protection for the specified cluster, because it contains a host {hostName} with APD Timeout is disabled.


• In the /var/log/vmware/vpxd/vpxd.log file on the vCenter, there are entries for the ReconfigureComputeResource task similar to:

<YYYY-MM-DD>T<TIME> verbose vpxd[61866] [Originator@6876 sub=Default opID=OpId-1565368141862-3e] [VpxVmomi] Invoking [reconfigureEx] on [vim.ClusterComputeResource:domain-c26] session [520ca20e-252a-2e38-e5b1-ddacbcf1c3de(52378448-2385-9e73-a62c-f69ac8c42b6a)]
<YYYY-MM-DD>T<TIME> verbose vpxd[61866] [Originator@6876 sub=Default opID=OpId-1565368141862-3e] Arg spec:
--> (vim.cluster.ConfigSpecEx) {
--> vmSwapPlacement = "vmDirectory",
--> spbmEnabled = <unset>,
--> defaultHardwareVersionKey = <unset>,
--> desiredSoftwareSpec = (vim.DesiredSoftwareSpec) null,
--> dasConfig = (vim.cluster.DasConfigInfo) {
--> enabled = true,
--> vmMonitoring = "vmMonitoringDisabled",
--> hostMonitoring = "enabled",
--> vmComponentProtecting = "enabled",
--> failoverLevel = 1,
--> admissionControlPolicy = (vim.cluster.FailoverLevelAdmissionControlPolicy) {
--> resourceReductionToToleratePercent = <unset>,
--> failoverLevel = 1,
--> slotPolicy = (vim.cluster.SlotPolicy) null
--> },
--> admissionControlEnabled = true,
(...)
<YYYY-MM-DD>T<TIME> info vpxd[15359] [Originator@6876 sub=Default opID=OpId-1565368141862-3e] [VpxLRO] -- ERROR task-274 -- domain-c26 -- vim.ComputeResource.reconfigureEx: vim.fault.CannotEnableVmcpForCluster:
--> Result:
--> (vim.fault.CannotEnableVmcpForCluster) {
--> faultCause = (vmodl.MethodFault) null,
--> faultMessage = (vmodl.LocalizableMessage) [
--> (vmodl.LocalizableMessage) {
--> key = "backtrace",
--> arg = (vmodl.KeyAnyValue) [
--> (vmodl.KeyAnyValue) {
--> key = "opId",
--> value = "OpId-1565368141862-3e"
--> }
--> ],
--> message = "[context]zKq7AVEDAAAAANSXlwETdnB4ZAAAxqY8bGlidm1hY29yZS5zbwAA3uIgAEeqHwDHChwBGmJZdnB4ZAABeQ6rASYU9wIuI/tsaWJ2aW0tdHlwZXMuc28AASEWiAFWUIcBHlKHARldhwF1PYYB8UOHAFjpLAAY/CwAco82A9RzAGxpYnB0aHJlYWQuc28uMAAEvY4ObGliYy5zby42AA==[/context]"
--> }
--> ],
--> host = 'vim.HostSystem:4f6d9f9f-dbdc-4975-a969-67d077b5b456:host-14',
--> hostName = "{hostName}",
--> reason = "APDTimeoutDisabled"
--> msg = ""
--> }
--> Args:
-->
--> Arg spec:
 

Cause

In vpxd.log, arg spec for reconfigureEx call contains dasConfig where VmComponentProtectionSettings is set to null.

Resolution

• Check if the "Misc.APDHandlingEnable" is enabled on all hosts in the cluster.
  
  esxcfg-advcfg -g /Misc/APDHandlingEnable
  
  Example Output: 
Value of APDHandlingEnable is 1
  
  
• Note: If the output value is '1' and vCenter is still erroring out while enabling HA, then follow the below steps.

         Connect to vCenter MOB using the following steps:

1. Navigate to https://<vCenter-FQDN>/mob/?moid=domain-c8
2. Under host column, select a "host" 
3. Select "config"
4. Under option click "more"
5. Search for "Misc.APDHandlingEnable" and select it
   
   Example
   
   
   
   
6. Perform the above steps on all ESXi hosts in the cluster.

        If the entry for "Misc.APDHandlingEnable" is not present, follow the below steps.

1. Place the affected host in maintenance mode.
2. Disconnect the affected host.
3. Reconnect the affected host.
   
   Note: The above steps will initiate a full sync of the host.

Additional Information

Note: Similar issue are seen when enabling vSphere HA on a New Nutanix cluster. 

Two workarounds exist depending on your cluster status:

1. If no hosts are added to the cluster yet:
   
   • Enable HA first, then set the PDL/APD failure response to “Disabled”. When hosts with APD handling disabled are added later, the APD response remains disabled, allowing HA to stay enabled without error
2. If the cluster already contains hosts and HA cannot be enabled:
   • Temporarily enable APD Timeout handling on affected ESX host, Select host > Configuration > Advance System Setting >  set this parameter Misc.APDHandlingEnable to 1. This unlocks APD failure response configuration and permits enabling HA with VMCP protection.

This behavior is documented Nutanix resources as necessary for ensuring VM availability and correct failure response:

Reference: Nutanix Next Community - Enabling HA on vSphere 7.0 cluster fails to enable VMCP