Hi,
I have an NLB cluster with 2 hosts and 1 VIP. The VIP is 192.168.20.11. The hosts dedicated IP addresses are 192.168.20.22 and 192.168.20.25, respectively. NLB is configured in multicast mode. The NLB ports are 80 and 443 - my app is a web service.
In addition to web service traffic - which is load-balanced - my app also listens for non load-balancing related multicast traffic from backend devices. For instance, I have a device that sends traffic to multicast address 239.168.0.20.
I noticed that only the app running on the NLB host with priority 1 - 192.168.20.22 - receives 239.168.0.20 traffic. On 192.168.20.25, the app does not receive that traffic. When I temporarily stop 192.168.20.22 in NLB manager, now the app on 192.168.20.25 starts receiving 239.168.0.20.
I also sniffed 239.168.0.20 traffic with netmon on both 192.168.20.22 and 192.168.20.25. netmon shows traffic on both hosts.
Therefore, this does not appear to be a switch or router-related issue. Instead, it looks like it's the NLB driver on 192.168.20.25 that drops 239.168.0.20 traffic.
Is this by design? I don't understand why the NLB driver drops 239.168.0.20 traffic - that traffic uses a multicast MAC address (01-00-E3-28-00-14) that is different from the NLB multicast MAC address (03-bf-c0-a8-14-0b), and the destination IP is not the NLB VIP. So the NLB should detect that 239.168.0.20 traffic is not to be load balanced and let it flow up the TCP/IP stack.
Any ideas?
Thank you in advance.