Incorrect Snapshot Storage Reporting During vCenter Reconnect in VMware Cloud Director

Products

• VMware Cloud Director

Issue/Introduction

In VMware Cloud Director, snapshot storage size reported in vm.storageAllocationMb event property may drop to 0 for Virtual Machines (VMs) after a vCenter Server reconnect or re-registration. This discrepancy impacts chargeback and billing data accuracy.

• Snapshot storage reported as 0 GB despite existing snapshots.
• Issue occurs across multiple VMs on the same vCenter concurrently.
• Issue correlates with vCenter Server reconnect or credential rotation actions.

Environment

VMware Cloud Director 10.6.1.2

Cause

During a vCenter Server reconnect or re-registration, VCD tears down and rebuilds its internal tracking of the vCenter inventory. During this rebuild window, VCD loses its internal record of the "current" snapshot, causing any generated chargeback/audit events to report 0 GB for snapshot storage until the snapshot state is refreshed (e.g., via a new snapshot operation).

Resolution

This behaviour is being evaluated and a fix will be provided in future release to preserve snapshot storage size through the vCenter inventory rebuild window.

Do Subscribe to this article to be updated on fix status.

Workaround

Minimize unnecessary vCenter Server reconnects/re-registrations against vCenters with VMs that carry snapshots, especially during periods where chargeback data accuracy matters (e.g., right before a billing cycle close). If a reconnect is required (e.g., for a credential rotation), be aware that snapshot storage size in events generated for a period afterward may be understated until VCD's inventory catches up.

How to build your reconciliation spreadsheet correctly around a reconnect:

1. Note the timestamp of any vCenter reconnect/credential-change action for the vCenter(s) in scope.

2. Treat any event for a VM on that vCenter, generated in the period immediately following a reconnect, as unreliable for the snapshot-storage component specifically (the base disk-allocation size is not affected).

3. Instead of using the event immediately after a reconnect, use the first event after that window where the snapshot storage size is non-zero and consistent (or explicitly re-check via the VCD API/UI's current snapshot size for that VM) as your reconciliation point.

4. If several unrelated VMs on the same vCenter show a snapshot-storage drop to 0 at the same timestamp, reconnect to the vCenter and cross-check against activity for that vCenter around that time.

5. If a single VM's snapshot storage size has been reporting 0 for a long time with no reconnect and no other VMs on the same vCenter affected at that time, this could genuinely mean that there are no current snapshot on it, for further assistance, see Contact Broadcom Support.