Hosts Fail to Reconnect to vCenter After Upgrade – Disconnected Due to Agent Upgrade After Host Update via Update Manager

Products

• VMware vCenter Server
• VMware vSphere ESXi

Issue/Introduction

Following a host upgrade using vSphere Lifecycle Manager (vLCM) or Update Manager, some ESXi hosts may fail to reconnect to vCenter Server. The disconnection reason displayed in vCenter is Agent Upgrade, and the hosts remain in a disconnected state until manually addressed.

Environment

 

• vCenter Server 8.0.x

• ESXi Hosts  8.0.x

 

Cause

The issue is caused by VMkernel multihoming, where multiple VMkernel adapters (vmknics) are configured under the same subnet and TCP/IP stack. This can disrupt management and control plane communication during agent operations such as upgrades or reconnects.

Example configuration (customer IPs masked):

• vmk0 – Management: 10.1.1.10

• vmk1 – vMotion: 10.1.1.11

• vmk2 – NFS: 1.x.x.x

Multihoming in a single stack is not recommended and may lead to unpredictable behavior, particularly after services restart during host upgrades. For more info see Multihoming on ESXi

Resolution

1. Review VMkernel Network Configuration:

   • Navigate to the host’s VMkernel adapters in vSphere Client.

   • Check if the vmnic's are in the same IP subnet and TCP/IP stack. 

2. Reassign vmknics to Separate TCP/IP Stacks:

   • Ensure vmk0 (Management) uses the default Management stack.

   • If vmk1 is in the same IP subnet as vmk0, assign different IP subnet and TCP/IP stack to vmk1 (vMotion).

   • If vmk2 is in the same IP subnet, assign (NFS or iSCSI) to a dedicated NFS or iSCSI TCP/IP stack if applicable.

3. Reboot the Host or Restart Host Management Agents (optional if network config changes are applied live).

4. Restart Management Network, if the above fails to fix the issue

5. Reconnect Host in vCenter:

   • Right-click the disconnected host and choose Reconnect.

   • Confirm successful reconnection after multihoming is eliminated.

Additional Information

 

• Ensure consistent VMkernel stack assignment across all hosts in the cluster to avoid intermittent issues.

• Always validate host networking before and after upgrades to prevent disconnection scenarios.

• Agent upgrade failures due to networking misconfigurations often manifest during lifecycle operations such as patching or host remediation.