Intermittent VM Connectivity Due to Incorrect vmnic-to-LAG Association

Products

• VMware vSphere ESXi

Issue/Introduction

May experience one or more of the following symptoms:

• Incorrect Packet Delivery: Network traces or packet captures reveal that ICMP echo replies are delivered to the wrong vmnic on the ESXi host, causing the host to drop the traffic.
• Virtual machines residing on the same VLAN experience intermittent ping failures.
• Network instability is observed immediately after adding or modifying uplinks within a Link Aggregation Group (LAG).

Environment

vDS using LACP

Cause

This issue is caused by an incorrect mapping of physical vmnics to the LAG ports on the Distributed Virtual Switch (DVS).

• When uplinks are added to a LAG out of sequence or assigned to the wrong LAG ports,
• LACP hashing algorithm on the physical switch and the ESXi host becomes misaligned.
• Consequently, the physical switch may forward return traffic to a vmnic that the ESXi host is not prepared to receive for that specific session, breaking the Layer 2 path.

Resolution

To resolve the connectivity issues, the vmnics must be reassociated to the LAG ports in the correct intended order to match the physical switchport configuration.

Steps to Correct LAG Association:

1. In the vSphere Client, navigate to the Distributed Switch.
2. Right-click the switch and select Add and Manage Hosts.
3. Select Manage Host Networking and click Next.
4. Select the affected ESXi host and click Next.
5. In the Manage Physical Adapters section, locate the vmnics currently assigned to the LAG.
6. Unassign the incorrectly mapped vmnics from the LAG ports.
7. Reassign each vmnic to the correct LAG port (e.g., vmnic0 to LAG port 0, vmnic1 to LAG port 1) to match the physical cabling and the physical switch's port-channel configuration.
8. Click Finish to apply the changes.

Verification

• Verify the LACP status by running the command: 

  • esxcli network vswitch dvs vmware lacp stats get
    

• Confirm that Partner Information no longer shows a Stand-alone state and that Rx LACPDUs are incrementing.
• Perform a continuous ping between VMs on the same host and VLAN to ensure stability.

Additional Information

For additional information reference:

• Configuring a LAG on a vSphere Distributed Switch Port Group when using LACP
• Packet capture on ESXi using the pktcap-uw tool