ESXi host isolation occurs frequently in a vSphere HA cluster

Products

• VMware vSphere ESXi

Issue/Introduction

• The ESXi host loses connectivity to the peers over the management network and often vCenter reports that the host is isolated from the vSphere HA cluster. 
• This state auto recovers after a while with no manual intervention. 

Environment

VMware vSphere ESXi 

FCoE / iSCSI storage configuration

Cause

Incorrect network route causes the storage traffic to flow through the management vmkernel interface and the issue occurs every time the storage traffic exhausts the underlying physical NIC (vmnic) bandwidth.

Resolution

Consult with the Internal Storage / Network teams to validate if the necessary static routes are defined to isolate the storage traffic from the management network.

Additional Information

Configuring static routes for vmkernel ports on an ESXi host