...in a FlexPod environment (UCS 2.1(3f), vSphere 5.1u2 vDS for switching, OnTap 8.1.2p4)? Since the hosts are being presented with dual 10Gbe virtual interfaces (one to fabric a, one to fabric b), does it matter which load balancing mechanism is being used? Even if 1 vmk (and in-turn 1 vmnic) is being primarily used by the host, what difference does load balancing make? The virtual interface ultimately gets aggregated and deterministic-ally pinned to a virtual interface and subsequent physical interface on the IO Module on the back of the UCS chassis. I get the need for creating 2 vmnics so that the ESXi host has one connection to the UCS's A and B fabrics for HA... in the event that one side of the fabric is taken down (assuming UCS fabric failover isnt being used, ofc) but ultimately what difference would load balancing make?
