Network RTO on ESXi Host with Static LAG due to Port Group Override

Products

• VMware vSphere ESXi
• VMware vSphere ESX 8.x
• VMware NSX for vSphere

Issue/Introduction

• An ESXi 8.0 host experiences intermittent or total network connectivity loss (Request Timed Out / RTO) when multiple uplinks (e.g., vmnic2 and vmnic3) are active.
• The ESXi host may appear as "Not Responding" in vCenter Server when both uplinks are enabled.
• Connectivity is restored immediately if one of the two uplinks is physically disconnected or administratively shut down.
• The host is connected to a physical switch configured with a Static Link Aggregation Group (LAG/EtherChannel).

Environment

VMware vSphere ESXi 8.x

Cause

The issue is caused by a mismatch between the ESXi Virtual Switch (vSwitch) load balancing policy and the physical switch configuration.
In this specific scenario, the physical switch was configured for a Static LAG, but the ESXi vSwitch was not set to the required "Route based on IP hash" policy. Additionally, an "Override" setting on the Management Port Group prevented global vSwitch changes from applying correctly to the management traffic, causing the host to revert to incompatible settings.

Resolution

To resolve this issue, the load balancing policy must be synchronized between the ESXi host and the physical switch.

Step 1: Remove Port Group Overrides

1. Log in to the ESXi host client or vCenter Server.
2. Navigate to Networking > Port groups.
3. Select the Management port group (e.g., Management Network) and click Edit settings.
4. Navigate to Teaming and failover.
5. Ensure the Load balancing option is NOT overriding the vSwitch settings, or manually set it to match the vSwitch configuration in the next step.

Step 2: Configure IP Hash Load Balancing

1. Navigate to Networking > Virtual switches.
2. Select the relevant switch (e.g., vSwitch0) and click Edit settings.
3. Under Teaming and failover, locate the Load balancing dropdown.
4. Select Route based on IP hash Create a vSphere Standard Switch.
5. Ensure both uplinks (vmnic2 and vmnic3) are listed as Active uplinks.
6. Click Save.

Step 3: Verify Connectivity

1. Enable or reconnect both physical uplinks.
2. Perform a continuous ping to the ESXi host management IP.
3. Verify the host remains "Connected" in vCenter and no packet loss is observed.