Yes, it is due to multicast. You need a solution that allows multicast to pass through your tunnel.
Once you have that done, you also need a low latency connection between your cluster and your remote nodes. Else you might have issues of "node-flapping" which gets annoying fast.
We do...