"Available Resources" value appears incorrect in Cloud Director Provider VDC when using Reservation Pool allocation models

Products

• VMware Cloud Director

Issue/Introduction

Administrators may observe a discrepancy in the "Available Resources" metric displayed within the VMware Cloud Director (VCD) Provider VDC (PVDC) UI.

• The available capacity for a specific cluster (e.g., "MEDIA") appears lower than the physical cluster capacity would suggest.

• Users may suspect a synchronization issue between VCD and the underlying vCenter Server.

• Database verification confirms that vCenter and VCD values match, indicating the issue lies in resource accounting logic rather than data corruption.

Environment

10.x

Cause

This behavior is by design and occurs due to the strict resource commitment required by the Reservation Pool allocation model, combined with vSphere HA Admission Control overhead.

When an Organization VDC (OrgVDC) is configured with the "Reservation Pool" model:

1. VCD creates a corresponding Resource Pool in vCenter (often named resgroup-xx).

2. VCD sets the vSphere Reservation on this pool equal to the OrgVDC's configured Allocation.

3. These resources are "guaranteed" to the OrgVDC and are subtracted from the cluster's available capacity immediately, regardless of whether VMs are actually running or consuming those resources.

    

Additionally, vSphere HA Admission Control reserves a percentage of cluster resources to ensure failover capacity. The calculation for "Available Resources" in the PVDC is effectively: Total Capacity -

 

(Sum of all OrgVDC Reservations + HA Admission Control Reserved Slot)

 

Resolution

Since the reported values accurately reflect the strict reservations configured in the environment, no repair actions are needed for the VCD database.

To increase the available resources in the Provider VDC, you must perform one of the following actions:

1. Reclaim Unused Allocations: Identify OrgVDCs utilizing the "Reservation Pool" model that are under-utilizing their allocated capacity. Reduce their allocation (and consequently their reservation) to return resources to the free pool.

2. Scale Out the Cluster: Add additional ESXi hosts to the affected cluster (e.g., "MEDIA"). This increases the total physical capacity, thereby increasing the available headroom after reservations are accounted for.

3. Adjust HA Admission Control: Review the vSphere HA Admission Control policy in vCenter. If the policy is overly aggressive (e.g., reserving 50% of the cluster for failover), adjusting it to a lower safe threshold will release resources back to the PVDC.