[c-nsp] Multicast source question

[Michael Robson]

> It's not true. The destination MAC address is the multicast L2 MAC.
> 
> > capture seems to be suggesting)? If this is true, then what 
> is the process
> > that gets the stream to the router, i.e. how does the 
> switch determine that
> > it should add the corresponding multicast MAC to the port 
> facing the router?
> 
> The "router port" bit of IGMP snooping. Broadly speaking, when the 
> switch sees IGMP queries coming in on a port, it marks that port as a 
> router port and floods all multicast traffic up it.
>
Thanks for the info.; the ethereal capture seemed to be indicating this, but
as we are having a problem at the moment, I wanted to be sure that this was
normal behaviour. So now a question regarding our issue, has anyone seen
anything similar to the following? We have a core of 6500s running sup720s
with native IOS version 12.2(18)SXD4. When a server sources a multicast
stream (we have seen this on 2 of our 6500s so far), captures show that the
stream corrective passes through the cisco WS-C2970G-24TS-E switch towards
the 6500, but as soon as it enters the 6500 (this 6500 can be the server's
gateway or just passing the traffic at layer 2 to another 6500), all ports
that are a member of the same VLAN as the server receive the traffic, almost
as if IGMP snooping is turned off or broken. I have shown that IGMP snooping
is enabled using the slightly convoluted command (as per Cisco docs) "sh ip
igmp int vlan 404 | inc global". The command gives

"IGMP snooping is globally enabled"

Has anyone else seen this and know how to fix it (maybe time for a software
upgrade)?

Michael Robson.