Virtual machines incorrectly show duplicate mappings across multiple ESXi hosts in the NSX Manager UI when a host is disconnected due to hardware failure

Products

• VMware NSX

Issue/Introduction

• In the NSX Manager User Interface (UI), a single virtual machine (VM) is displayed as simultaneously residing on two different ESXi hosts.

• One of the ESXi hosts listed in the duplicate mapping is currently in a "Disconnected" or "Not Responding" state within vCenter Server.

• vSphere Distributed Resource Scheduler (DRS) has already successfully migrated the active VM workload to a healthy host, but the stale registration persists in the NSX inventory.

Environment

VMware NSX

Cause

This issue occurs due to a management synchronization gap caused by an unhandled, sudden host failure (such as a severe hardware fault preventing a power-on state).

When an ESXi host abruptly disconnects from vCenter and loses power:

1. The local management agents on the host cannot send a final state-change or inventory update to the NSX Manager.

2. Even though vSphere DRS detects the host loss and restarts/migrates the affected VMs onto alternative healthy hosts, the NSX Manager cannot communicate with the failed host to validate or complete the unregistration process.

3. As a result, the stale VM-to-Host mapping remains cached inside the NSX inventory database until the original host is able to acknowledge the inventory change.

Resolution

To permanently resolve this discrepancy, the underlying hardware fault must be remediated so that the affected ESXi host can safely communicate back with the management plane.

Note: If the inventory fails to synchronize following the hardware remediation, please open a case with Broadcom Support along with the support bundle.