Multiple connections between UTM's

Mike, it is possible to bind the IPsec Connections to specific interfaces and then use Routing or Multipath rules to select the tunnel to be used for specific traffic.

If the replication traffic is between different subnets, then you can use the approach I described in Auto-Failover IPsec VPN Connections.

Cheers - Bob

Thanks Bob. Binding to local interface did bring up both tunnels. Replicating using AppAssure like this: Site 1, Internal Subnet 1, AppAssure Core 1  Replicating to Site 2, Internal Subnet 2, AppAssure Core 2. (And vice versa, Core 2 replicates back to Core 1) Since I need both tunnels up at the same time with different traffic going over each tunnel it seems I need Multipath rules? It's not really a failover scenario.

So my Multipath rule at Site 2 where RDP traffice originates would look like: Internal 2, RDP, to Internal 1, Persistence by Interface, Bind to Interface for VPN 1? (VPN 2 is the one I will use for replication)

Thanks Bob. Binding to local interface did bring up both tunnels. Replicating using AppAssure like this: Site 1, Internal Subnet 1, AppAssure Core 1  Replicating to Site 2, Internal Subnet 2, AppAssure Core 2. (And vice versa, Core 2 replicates back to Core 1) Since I need both tunnels up at the same time with different traffic going over each tunnel it seems I need Multipath rules? It's not really a failover scenario.

So my Multipath rule at Site 2 where RDP traffice originates would look like: Internal 2, RDP, to Internal 1, Persistence by Interface, Bind to Interface for VPN 1? (VPN 2 is the one I will use for replication)

So it looked like my test environment was a success. Traffic from one server at the remote site to a backup server at the home site was going over the dedicated backup VPN while normal RDP sessions were going over the other VPN and web browsing was using the correct interface. Then I noticed that I was not receiving notifications, daily Executive Reports, or Weekly Backups from the remote site.

All of my remote sites have Request Routing on  DNS Services so they get proper inside resolution on our domain. Looking at Mail Manager I could see that the remote firewall was failing dns for the mail server. Doing a DNS lookup in tools failed, and doing a tracert showed that traffic to servers.ourdomain.com was going out the bind interface instead of the vpn. Shutting down the 2nd VPN tunnel and unbinding the interface on the primary VPN restored proper DNS Request Routing, Email, and Tracert.

Is there a work around?

I'm not sure what you mean by "I could see that the remote firewall was failing dns for the mail server."  Are you  running split-DNS and was the remote 310 getting the public IP instead of your private IP?

Cheers - Bob