[c-nsp] sup720 RP CPU utilisation with >20k adjacencies / IPv6 ND load?

[Phil Mayers]

On 11/06/12 16:09, Phil Mayers wrote:
> On 31/05/12 16:00, Dale W. Carder wrote:
>
>> Our network and some of our peers have run into the same issues as we
>> did our v6 rollouts. Try this out:
>>
>> ipv6 nd reachable-time 900000
>> ipv6 nd ns-interval 5000
>
> Thanks.
>
> Somewhat worrying, mere seconds after I typed these into one of our SVI,
> the router logged:
>
> %SYS-2-FREEBAD: Attempted to free memory at 44BD9B10, not part of buffer
> pool
>
> ...so I might hold off on that one. Perhaps an SXJ1 bug. But it does
> look like what we're looking for.
>

Just to follow up on this to the list - in the end, I enabled this on 
all our SVIs, and it solved the CPU problem - dropped from ~70-80% to 
~40% i.e. back to the average before IPv6 took off.

The FREEBAD error seems to have been transient - only happened once.

One final note: if you're doing HSRP, you need to set the params the 
same on both routers or they complain:

%IPV6-3-CONFLICT: Router FE80::5:73FF:FEA0:1 on Vlan162 has conflicting 
ND settings

Many thanks to Dale for suggesting this.