[c-nsp] Sup720, multicast bothers the CPU

[Chris Evans]

Peter. Yes just enable igmp snooping and querier.  Easier way is to enable
pim which does the rest.  Traffic that has a low ttl of one will still get
punted to the CPU..
On Mar 23, 2011 5:18 AM, "Peter Rathlev" <peter at rathlev.dk> wrote:
> On Wed, 2011-03-23 at 09:04 +0000, Phil Mayers wrote:
>> If you just want to drop all multicast, you could try:
>>
>> int VlanXX
>> ip multicast boundary MULTICAST-drop
>> ip access-list standard MULTICAST-in
>> deny 224.0.0.0 15.255.255.255
>
> Would that also discard forwarding of multicast packets L2-wise? Or only
> L3-wise?
>
>> Is the sender attached directly to a port on the 6500? There are some
>> newer features related to layer2 (as opposed to layer3) multicast ACLs
>> in SXH/SXI.
>
> The devices in question are not directly connected, but some might be in
> the future.
>
>> > The box in question does not have any multicast configuration at all.
>> > Does this mean all multicast traffic is sent to the CPU? How would one
>> > configure a Sup720 to not punt every multicast packet to the CPU?
>>
>> There are also mls rate-limiters for this, but I'm not sure off the top
>> of my head which one you want - we have multicast configured so don't
>> need them (the box builds hardware mfib entries, which stops the CPU
punts)
>
> That has me interested: What can I configure to make the traffic
> hardware swithed? Is it as simple as an IGMP querier or something like
> that?
>
> --
> Peter
>
>
> _______________________________________________
> cisco-nsp mailing list cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/