Error: Failed criteria: 128 network outage during ESXi uplink detachment from VDS

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms:

An ESXi host experiences a network outage for 15-20 minutes when an active physical uplink (vmnic) is detached from a vSphere Distributed Switch (VDS) during migration to a Standard Switch (vSS).

• Virtual machines lose connectivity across multiple VLANs.
  
  
• VMkernel interfaces (Management, vMotion, NFS) fail to respond to pings.
  
  
• Connectivity restores automatically after 12-15 minutes.

Environment

VMware ESXi 8.x

Cause

The physical NIC driver delays reporting the link-state change event to the ESXi hypervisor. Because the hypervisor does not receive the down event immediately, the VDS continues to route outbound traffic to the detached interface until the Failed criteria: 128 event is processed. During this window, the physical Top-of-Rack (ToR) switch blackholes traffic because its MAC address table continues to point to the detached physical port.

• The /var/run/log/vobd.log shows delayed failover events:

2026-07-12T04:09:45.002Z In(14) vobd[2097620]: [netCorrelator] 2731####us: [esx.problem.net.vmnic.linkstate.down] Physical NIC vmnic#### linkstate is down
2026-07-12T04:23:23.863Z In(14) vobd[2097620]: [netCorrelator] 84603####us: [vob.net.pg.uplink.transition.down] Uplink: vmnic#### is down. Affected portgroup: VMkernel. 1 uplinks up. Failed criteria: 128

Resolution

To prevent failover delays during physical interface detachment, manage the uplink at the vSphere Distributed Switch level to trigger a software-controlled failover prior to physical disconnection.

1. Log in to the vSphere Client.

2. Navigate to Distributed Switch > Manage > Hosts.

3. Select the affected host and click Manage Physical Adapters.

4. Locate the uplink designated for removal (e.g., vmnic4).

5. Reassign the uplink state from Active to Unused (do not select None). This bypasses the hardware driver event delay and triggers an immediate software-driven failover.

6. Verify that network traffic and VMkernel pings successfully traverse the remaining active uplink.

7. Once traffic stabilizes, detach the physical adapter from the VDS.

8. Attach the adapter to the target switch (e.g., vSwitch0) and migrate the associated VMkernel adapters and VM port groups.

Alternative Method: Use the Migrate Network Configuration wizard within the VDS menu to relocate the physical adapter and associated VMkernel ports in a single atomic action. This proactively re-routes traffic and prevents upstream switch packet blackholing.

Additional Information

Network adapter (vmnic) is down or fails with a Failed Criteria Code in VMware vSphere ESXi

How to move ESXi/ESX host from one vCenter Server to another

Moving an ESX/ESXi host with vSphere Distributed Switches (vDS) between vCenter Servers

Migrate Virtual Machines (VMs) and VMkernel Adapters from Distributed Switch (vDS) to Standard Switch (vSS)

For instructions on collecting ESXi log files, see Collecting diagnostic information for VMware ESXi.