Network connectivity loss for VMs when using active-active uplinks.

Products

• VMware vSphere ESXi
• VMware vCenter Server

Issue/Introduction

VMs/hosts experience network connectivity loss when an ESXi host is configured with two uplinks (e.g., vmnic0 and vmnic1) in an active-active teaming.

The issue does not occur when only one uplink is active.

Environment

VMware vSphere ESXi.

VMware vCenter Server.

Cause

The Inter-Switch Link (ISL) connecting the two upstream physical network switches is misconfigured. Specifically, the ISL trunk port is missing the required VM VLANs.

When East-West traffic between VMs pinned to different active physical NICs traverses the ISL, the physical switch drops the packets.

Resolution

 

Resolution: 

• Engage the network administration team to modify the physical switch configuration.

• Identify the physical switch ports connected to the ESXi host uplinks (e.g., Eth1/47 and Eth1/48).

• Identify the Inter-Switch Link (ISL) port connecting the two physical switches (e.g., Eth1/20).

• Add the required VM VLAN IDs to the allowed VLAN list on the ISL trunk port.

  • Example of misconfiguration: VLAN 111 is missing from Eth1/20, which is an ISL interface.
    

 

Workaround: 

• Log in to the vSphere Client.

• Navigate to the affected ESXi host or vSphere Distributed Switch.

• Modify the Teaming and Failover policy for the affected port group.

• Move one of the active uplinks (e.g., vmnic0) to "Unused" or "Standby", leaving only one "Active" uplink. This forces all traffic through a single physical switch, bypassing the ISL.