The hosts became disconnected from NSX after downgrading from ESXi version 8.0.3 to 8.0.2

Products

• VMware NSX

Issue/Introduction

After downgrading from ESXi version 8.0.3 to 8.0.2 you see hosts in cluster showing a "Host disconnected status" with error details displaying below:

Heartbeating between NSX management node and host ##### is down.

You see alarms for the "Management channel to Transport Node down long" for all hosts with this disconnected status.

Environment

NSX 4.1.2.1

Cause

This behavior occurs because the NSX Management Plane Agent (MPA) and VIBs (vSphere Installation Bundles) are version-sensitive and tied specifically to the ESXi kernel build they were installed on.

Resolution

Workaround:

1. Select "Host Disconnected" status on each host in this state.

2. Click Error Details

3. Select Error > Resolve

 

Note:

If above workaround does not work, you will need to put each host in MM one at a time and pull it out of the cluster an confirm all VIBs are removed then move back into the cluster to reinstall NSX on each host.

Additional Information

Clicking the Resolve button in the NSX Manager UI triggers an automated remediation workflow that performs the following:

1. Version Detection: NSX Manager queries the host, realizes it is now on 8.0.2, and updates its database.

2. VIB Re-alignment: It pushes the correct VIB metadata and binaries specific to ESXi 8.0.2 if a mismatch is detected.

3. Certificate Re-push: It forces a re-exchange of certificates and thumbprints (equivalent to running nsxcli -c push host-certificate manually), which re-establishes the encrypted heartbeat tunnel.

4. Service Restart: It restarts the nsx-opsagent and nsx-proxy services on the host to clear any stale state.