[NSX] Edge VM Present In NSX Inventory Not Present In vCenter

Issue

Today while I was deleting edge bridges from NSX Manager I got this error message:

Edge VM Present In NSX Inventory Not Present In vCenter

Description The VM edge-bridge-cluster1-B with moref id vm-1370430 corresponding to the Edge Transport node a060574d-4e93-4b7e-83b4-7eb8464a645d vSphere placement parameters is found in NSX inventory but is not present in vCenter. Please check if the VM has been removed in vCenter or is present with a different VM moref id.

Recommended Action The managed object reference moref id of a VM has the form vm-number, which is visible in the URL on selecting the Edge VM in vCenter UI. Example vm-12011 in https:///ui/app/vm;nav=h/urn:vmomi:VirtualMachine:vm-12011:164ff798-c4f1-495b-a0be-adfba337e5d2/summary Please find the VM edge-bridge-cluster1-B with moref id vm-1370430 in vCenter for this Edge Transport Node a060574d-4e93-4b7e-83b4-7eb8464a645d. If the Edge VM is present in vCenter with a different moref id, please follow the below action. Use NSX add or update placement API with JSON request payload properties vm_id and vm_deployment_config to update the new vm moref id and vSphere deployment parameters. POST https:///api/v1/transport-nodes/?action=addOrUpdatePlacementReferences. If the Edge VM with name edge-bridge-cluster1-B is not present in vCenter, use the NSX Redeploy API to deploy a new VM for the Edge node. POST https:///api/v1/transport-nodes/?action=redeploy.

Solution

Googling around I found various solutions, but no one was fitting exactly my situation.

For example I found these:
VMware NSX Edge VMs not present in both NSX and vCenter
Edge VM present in NSX inventory not present in vCenter alarm

As shown into the first link I tried without success the command described in Scenario 3:

So, I also tried to check if the Transport Nodes was still present into NSX, whit commands below:

No trace of them. It seems like, the deletion process has not finished. I waited 30 minutes, but the problem was still there.

I solved restarting one by one the NSX Manager appliances.
I started rebotting the first appliance, waited for the cluster to return to a "stable" state, and continued with the next appliance, until I had restarted them all. At the last reboot the error message changed from "Open" to "Resolved" and was no longer present

That's it.