I just finished upgrading my CS-220 array to FW 2.08. I've followed the guides and created my new group IP (which I set to the same as the iSCSI discovery IP, I didn't see anything that I couldn't/shouldn't do this). I also installed NCM on all 6 of my ESXi hosts, rebooted them, and went though all of my datastores and changed the from VMware Round Robin to the new Nimble PSP Directed path policy.
My CS220 has 4 iscsi data NICs configured, and when the Round Robin policy was selected on the datastores, I could see each data store had connected to each iSCSI IP on the Nimble as expected - i.e. 4 nics per ESXi host and, 4 nics on the Nimble, each data store had 4 connections to the Nimble on each of its 4 IPs. all good.
Now with the Nimble PSP selected, each datastore still makes four connections to the Nimble however all 4 all go to only one Nimble IP, the same IP for all datastores across all hosts.
Why is everything only using 1 IP/NIC on the Nimble now? This doesn't seem right...
Attached is a screen shot of one of the data stores, showing all connections going to one IP, all my datastores look the same. They used show all four Nimble IP's