I have a FAS2220 with 2 controllers and am struggling a bit with ifgrp's for iscsi. It is layed out as follows:
Controller 0
iscsi_ifgrp1 - e0a & e0c using multimode dynamic to production switch (shared with LAN but on it's own VLAN) 192.168.80.11
Controller 1
iscsi_ifgrp1 - e0a & e0c using multimode dynamic to production switch (shared with LAN but on it's own VLAN) 192.168.80.13
Windows Server 2012
iscsi nic 192.168.80.10 and attached to one target 192.168.80.11
From what I understand the server will only communicate over the VIF on controller 0. It will use controller 1 if controller 0 fails, is this correct?
On the windows server I can't setup a second target to 192.168.80.13 for MPIO is this correct?
The reason I ask is that on other SAN's I had 2 controllers and each had an IP on a different subnet. I would place 2 switches in the path and setup both IP's as targets on a Windows server using MPIO over 2 seperate NIC's. More of an active/active iscsi controller setup. It seems NetApp does not behave this way.
My end goal is to create multiple paths with higher bandwidth by buying 2 switches seperate form the LAN/CORE but the only way I see this is to have the 2 interfaces on controller 0 e0a and eoc on seperate switches and do the same with controller 1. The issue I can't figure out with this scenario is how to get the server to use 2 nics via MPIO since it's all the same subnet. Am I missign something here?
The only solution I can come up with is to team 2 NIC's on the server and place each NIC in the team on the two seperate switches.
Any thoughts on this would is appreciated.