Random VMs on NSX segments cannot be pinged

Products

• VMware NSX

Issue/Introduction

• Some VMs on NSX cannot be pinged
• Working and broken VMs can be on the same segment and same host
• The impacted VMs can change when the network traffic is initiated from different source machines
• Taking down NICs on the vDS so that only 1 NIC remains active solves the issue on that host
  esxcli network nic down -n vmnic#

Environment

VMware NSX

Cause

The vmnics used on the vDS have been configured for LAG/LACP on the physical switch but not on vSphere/NSX.

Resolution

LAG must be configured on the vDS and NSX to align with the physical network configuration.

1. Create the LAG on vSphere, Configure a Link Aggregation Group to Handle the Traffic for Distributed Port Groups
2. Configure an Uplink Profile to use the vDS LAG, Configure an ESXi Host Transport Node with Link Aggregation Group
3. Consume the Uplink Profile in the Transport Node Profile, Adding a Transport Node Profile