i would say the first option, just make sure the secondary hw raid server is setup with the same volume sizes as secondary sw raid server so you can setup the replication task, theoretically it should work.

as your virtual ip for iscsi failover will be "disconnected" at this time your virtual infrastructure will obviously be down as well untill all volumes are synced and failover is activated, then you are correct, "just switch back to the main node."

good luck

i recommend Adaptec/PMC Sierra raid controllers, but it depends on your backplane