Impact of Powered-Off vCLS VMs DRS Operations in vSAN Clusters

Products

• VMware vSAN

Issue/Introduction

Symptoms

• In the vCenter Recent Tasks, the error "The name 'vCLS' already exists" is observed repeatedly.

• DRS is not functioning as expected, and attempts to put a host into maintenance mode are getting stuck at 17% progress.
• An Auto-Provisioning tag is present on the vSAN cluster to allow automatic deployment of vCLS VMs if they are missing or faulty.
• Validation shows that vCLS VMs exist but are powered off, confirming that automatic vCLS provisioning is not reattempted due to VM name conflicts.
• DRS-dependent operations, such as VM evacuation during host maintenance, are impacted due to this issue.

 

Environment

VMware vSAN 7.0

Cause

• The vCLS VMs are present in the inventory but are in a powered-off state.
• Since the VM names still exist, vCenter is unable to auto-deploy new vCLS VMs due to a name conflict.
• The powered-off vCLS VMs prevent vCenter from auto-remediating by provisioning new, functional vCLS instances.
• As a result, DRS logic fails to function as expected, leading to issues such as incomplete host evacuation during maintenance.
• The log entry in /var/log/vmware/vpxd/vpxd.log below reflects the error encountered:

2025-04-22T08:57:13.183Z error vpxd[36784] [Originator@6876 sub=Default opID=wcp-vCLS-f6] [VpxLRO] -- ERROR task-4204838 -- 52d1734a-1c30-7874-a1fd-dcec072e88
75(52262af0-bb97-62ef-e013-7dfa389d33a9) -- group-v509 -- vim.Folder.createFolder: :vim.fault.DuplicateName
--> Result:
--> (vim.fault.DuplicateName) {
-->    faultCause = (vmodl.MethodFault) null,
-->    faultMessage = <unset>,
-->    name = "vCLS",
-->    object = 'vim.Folder:5cfd170a-5187-47dd-ad5a-18852d74cc69:group-v159025'
-->    msg = ""
--> }
--> Args:
-->
--> Arg name:
--> "vCLS"

Resolution

• Workaround for forgotten cluster-enable workflow using EAM's MOB:
  • Access the URL: https://<vcsa-address>/eam/mob?vmodl=1 (Ensure ?vmodl=1 is included).
  • Select the "enable" option (a popup should appear from an address like https://<vcsa-address>/eam/mob//?moid=EsxAgentManager&method=EsxAgentManager_Enable_Task&vmodl=1).
  • Populate the "cluster parameter" with the following XML: <cluster type="ClusterComputeResource">domain-*****</cluster>
  • Click "Invoke Method". This will initiate a task in EAM to enable monitoring and operations for the specified cluster.
    
    
• Once the above steps are completed, the vCLS VM will deploy successfully, and the host can be placed into maintenance mode without any issues.
  
  
• If multiple clusters are present in the 'Datacenter' and most of them are in maintenance mode with inaccessible vCLS VMs, the following steps should be performed:
  • Exit the host from maintenance mode. This action will un-register the vCLS VMs from the EAM for the affected cluster.
  • After the vCLS VMs are un-registered, the host can be placed back into maintenance mode as the vCLS VMs are no longer required and will be decommissioned in the near future.

 

Additional Information

• Do not manually delete or recreate vCLS VMs unless absolutely necessary. vCenter should manage the vCLS lifecycle automatically.
• If vCLS VMs fail to power on or auto-provisioning fails, verify resource availability and ensure vCenter health.
• Consider disabling vCLS on a cluster using Retreat Mode if it is no longer needed. Disable vCLS on a Cluster via Retreat Mode