Graceful failover of Citrix NetScaler HA setup

Introduction

Suppose that you are using NetScaler for Load balancing for your cloud service(SaaS). You want to provide high availability of your SaaS. To do this, NetScaler should also be available throughout Moreover, if NetScaler failover happens, your critical https session should not break. This document tries to explain the approach to meet this criteria.

NetScaler HA VMAC

VMAC's are a useful addition in the NetScaler high availability tool set. In brief a VMAC creates a virtual MAC address that can 'failover' between devices. VMACs can be used to compliment the built-in HA or to create an active/active NetScaler pair.

By virtualising the MAC address there is no drop in network connectivity during failovers as the MAC address is shared across NetScaler devices - this means that the CAM table in the upstream switches does not require any update. As a result, failovers between NetScaler devices should be faster and less intrusive with regards to user sessions/connections.

VMAC's work using a 'priority' - the higher priority determines ownership of the VMAC between devices. In a NetScaler HA configuration the priority of the VMAC between devices is the same, without HA it is configurable. For example if we had two NetScalers not using HA, NS1 and NS2, and a single VMAC configured on each we could set NS1 to have a priority of 100 and NS2 to have a priority of 90. NS1 would have 'ownership' of the VMAC because of its higher priority.

Reference

http://www.cb-net.co.uk/citrix-articles/2016-netscaler-using-vmacs-for-high-availability

http://support.citrix.com/article/CTX121681