Forum Discussion

young19918's avatar
young19918
Icon for Altocumulus rankAltocumulus
May 23, 2023
Solved

Can a floating IP be the next hop for routing?

Hello,

Recently tried using HA in Route mode.
Here is my structure :

I thought floating ip can be the next hop of Route.
So I set the next hop in Client Route as 1.1.1.3 & Switch's is 2.2.2.3.

However, when I browse VIP, I find that I can't pass...
Can a floating IP be the next hop for routing?
If not, how should I set ...

Any Help is appreciate.

  • Paulius's avatar
    Paulius
    May 24, 2023

    young19918 

    So from client 1.1.1.4 you are attempting to reach server on 3.3.3.1 on tcp/80 or possibly tcp/443? If your routing is in place and you can indeed ping 3.3.3.1 from 1.1.1.4 and the other way around then it could be your forwarding virtual server that is not allowing the traffic to pass through the F5. You should be able to perform a tcpdump on the active F5 of the following and see where the traffic is going.

    tcpdump -nni 0.0:nnp host <source_IP> and port <destination_port>

    tcpdump -nni 0.0:nnp host 1.1.1.4 and port 80

    In your topology I would make the switch an L2 switch for the VLAN that sits behind the F5s and have the F5s and server be in subnet 2.2.2.0/24 with the gateway for the server being 2.2.2.3. If forcing the switch to be in routed servse another purpose for you then keep it but if not I wouldn't do it because it adds unneeded complexity without much benefit. If your intent is to be able to manage the switch by an IP you can always IP it from that same 2.2.2.0/24 subnet and reach it.

4 Replies

  • young19918 When the F5s are in routed mode and you have a floating IP configured for either HA or standalone it should absolutely be the destination for the next hop for networks on the apposing side of the F5. If you are not able to route traffic from one side of the F5 to the other it's most likely that your forwarding virtual server is not configured properly to accept and pass all traffic from one side to the other on the F5. The following is what I have configured for my lab device.

    ltm virtual VS_FORWARDING_ALL {
        description "Catchall Forwarding VS"
        destination 0.0.0.0:any
        ip-forward
        mask any
        profiles {
            fastL4 { }
        }
        source 0.0.0.0/0
        translate-address disabled
        translate-port disabled
    }

     

    • young19918's avatar
      young19918
      Icon for Altocumulus rankAltocumulus

      Hi Paulius ,

      Thanks for your reply.

      After setting it up, Client and Server can ping each other
      But the Client cannot browse to the server through VIP, I tried to browse the Server directly without VIP, but it also failed ......

      • Paulius's avatar
        Paulius
        Icon for MVP rankMVP

        young19918 

        So from client 1.1.1.4 you are attempting to reach server on 3.3.3.1 on tcp/80 or possibly tcp/443? If your routing is in place and you can indeed ping 3.3.3.1 from 1.1.1.4 and the other way around then it could be your forwarding virtual server that is not allowing the traffic to pass through the F5. You should be able to perform a tcpdump on the active F5 of the following and see where the traffic is going.

        tcpdump -nni 0.0:nnp host <source_IP> and port <destination_port>

        tcpdump -nni 0.0:nnp host 1.1.1.4 and port 80

        In your topology I would make the switch an L2 switch for the VLAN that sits behind the F5s and have the F5s and server be in subnet 2.2.2.0/24 with the gateway for the server being 2.2.2.3. If forcing the switch to be in routed servse another purpose for you then keep it but if not I wouldn't do it because it adds unneeded complexity without much benefit. If your intent is to be able to manage the switch by an IP you can always IP it from that same 2.2.2.0/24 subnet and reach it.