we are going to replace an old 6513 (CatOs + MSFC) and we found that there is a configuration for the Microsoft NLB:
arp 172.27.124.13 03bf.ac1b.7c0d ARPA
on Nexus we replaced it with
int vlan 2
ip arp 172.27.124.13 03bf.ac1b.7c0d
mac address-table multicast 03bf.ac1b.7c0d vlan 2 inte eth5/13
But we are still not able to ping the cluster ip.
Do we missed something?
Ciao e grazie!
You need to include the following command:
vlan configuration 2
layer-2 multicast lookup mac
Please test and let me know the result.
Hm... okay, did you exit out of vlan configuration mode?
Secondly, did you include all the interfaces for your MS NLB servers? You only have 1 configured. Also, I am going to assume you did this on both Nexus 7000s.
Up to now, and in order to reduce the problem determination, we shut down others links, so the only interface where the multicast packet will arrive is 5/13 on NX Core 1.
I did an elam capture on module 5 and I was able to see that NX sends packets out to the 5/13 but no answer is gettin back.
We did another test: between the NX and the NLB servers there is a 4948 switch
NX (eth5/13) - (gi1/47) 4948 - NLB SERVERS
(it was in place also when the 6513 was there, and the config was never changed - static mac is pointed on correct interfaces) and there we connect a pc on the same vlan as NLB. Result: still not able to ping :-)
So in my opinion, the problem is on the NLB configuration or in the ESXi 5 which, after reboot, has something misconfigured.
Up to now we disabled NLB and everything works fine as expected; it will be interesting to check if there is a possibility to do an elam capture also on the 4948, in order to avoid to install wireshark on the M$oft cluster :-)
Ciao e grazie!