Intermittent connection loss observed in inter-site link between VMs located between two different sites during nsx vibs upgrade on the cluster

Products

• VMware NSX

Issue/Introduction

During NSX upgrade from version 4.1.0.2 to 4.2.1.4 as part of TCP 5.0, an intermittent connection loss on the inter-site link between two different sites was observed 

Environment

NSX 4.2.1.4

Cause

The NSX Edge VMs which were carrying North-South traffic from the VMs were migrated from ESX host during NSX upgrades.

As a result of VM migration, BFD tunnels got re-established between Edge VMs and ESXi hosts 

Resolution

The traffic disruption that was observed during NSX upgrade is expected as the BFD/TEP tunnels between Edge VMs and ESXi hosts went down during vmotion of Edge VMs which got triggered as ESXi hosts entered maintenance mode. The BFD/TEP tunnels got re-established after the Edge VMs were migrated to other ESXi hosts and traffic between inter-site VMs got restored.

It is recommended to keep a maintenance window during NSX/ESX upgrades. Pls check NSX upgrades guide (Operational impact of NSX upgrades) Refer : Upgrade-guide

Also traffic disruption might be expected during vmotion or upgrade of Edge VMs. Refer : Edge vmotion best practices.