NSX Edge Refresh Fails with COMPUTE_MANAGER_NOT_FOUND or Stale

Products

• VMware NSX

Issue/Introduction

Users attempting to edit an NSX Edge transport node or enable features such as Multi-TEP may encounter the following symptoms:

• UI Error Transport node refresh failed: [Fabric] Refresh edge placement configuration failed. Check network connectivity of the edge node.

• Manager Logs (nsxapi.log):
   INFO EdgeTNRealizationServiceImpl FABRIC [nsx@6876 comp="nsx-manager" level="INFO" subcomp="manager"]Updating config state errors to EdgeNodeConfigInfo/.../edge-transport-node/<UUID>, state COMPUTE_MANAGER_NOT_FOUND, message [Fabric] Refresh edge placement configuration failed.
  
• API Validation: Running GET /api/v1/transport-nodes/<edge-uuid> shows a vc_id (Compute Manager ID) that either does not exist in the current NSX inventory or belongs to a decommissioned vCenter.

Cause

This issue occurs when the NSX Manager maintains a stale reference to the Compute Manager (vCenter) that originally deployed the Edge VM. This typically happens if:

1. The Edge VM was moved (vMotion) to a different vCenter not managed by the same NSX instance.
2. The original Compute Manager was deleted and re-added to NSX, receiving a new UUID.
3. The VM MoRef (Managed Object Reference) changed in vSphere (e.g., removed and re-added to inventory), causing a mismatch in NSX Manager's internal database.

In this state, UI-initiated refresh actions (which precede edits) fail because NSX cannot locate the managed object in the referenced Compute Manager.

Resolution

To resolve this issue, the stale Edge transport node must be replaced with a new node that has a valid Compute Manager reference.

Prerequisites

• Deploy a new NSX Edge VM of the same size and version as the existing node.
• Ensure the new Edge node is joined to the management plane and is in a NODE_READY state.

Workflow: Replace Edge Cluster Member

1. Log in to NSX Manager UI.
2. Navigate to System > Fabric > Nodes > Edge Clusters.
3. Select the cluster containing the stale Edge node.
4. Click the Actions dropdown or the ellipsis and select Replace Edge Cluster Member.
5. In the configuration window:
   • Select the Stale Edge Node as the member to be replaced.
   • Select the New Edge Node as the replacement.
6. Click Save/Replace. NSX will begin relocating logical configurations (Tier-0/Tier-1 Service Routers, DHCP, etc.) to the new node.
7. Monitor the relocation progress. Once complete, the stale Edge node will no longer be part of the cluster.
8. Manually delete the old Edge VM from both the vSphere Inventory and the NSX Transport Nodes list.

Verification

• Verify that the new Edge node shows a Success configuration state.
• Attempt the original operation (e.g., enabling Multi-TEP or editing configuration) to ensure the refresh error no longer appears.

Additional Information

• Editing Edge Node Error — Confirms that a stale vc_id causes refresh failures and recommends node replacement.
• Refresh API Error — Details the "null placement configuration" error caused by changed VM MoRefs.