I haven't had a chance to get this up and running yet - simply because I want to understand all the caveats before moving forward. Seems like limitations are still being uncovered.

That being said, I'd like to understand the status of the iSCSI target on the secondary end. If the iSCSI target on the secondary was at the bare minimal readable - then it's possible use multipathd on Linux for active/passive.

Other than that - the only thing I can think of is an external load balancer. But the LB won't fully understand how to perform the switchover at purely layer 3 or layer4. You'd need an enhanced script.