[c-nsp] IGMP explicit-tracking table overflow

[Peter Rathlev]

Phil, thank you for the input.

On Thu, 2014-01-30 at 10:02 +0000, Phil Mayers wrote:
> You can do:
> 
> clear ip igmp snooping statistics

Actually that didn't make any different for us:

# show ip igmp snooping explicit-tracking | include Current
    Current number of entries: 8020    Configured DB size limit: 32000
# clear ip igmp snooping statistics                   
# show ip igmp snooping explicit-tracking | include Current
    Current number of entries: 8020    Configured DB size limit: 32000
#

> ...to get back on an even keel. However, for happy-fun-joy, these are 
> replicated in a dual-sup setup, and you cannot clear them on the standby 
> sup, so eventually you'll start getting -SPSTBY- versions of the message 
> that never stop. Standby sup reload is the only solution :o(

Fortunately we have only single-SUP systems. :-)

> > interface Vlan210
> >   ip address 192.0.2.2 255.255.255.0
> >   ip pim dr-priority 50
> >   ip pim sparse-dense-mode
> >   ip multicast boundary 70
> 
> You probably want "ip igmp access-group 70" here too, though it won't 
> stop the stats getting gathered IIRC.

Oops, accidentally left out that one. It's there alright and doesn't
make a difference.

> You can increase the global track limit to make it take longer between 
> instances:
> 
> ip igmp snooping limit track 51200
> 
> I'm not sure if "0" disables it.

Just tried that for the fun of it; setting it to zero makes it instantly
log messages about exceeding both permanent threshold and maximum limit.
And the table is still the same size afterwards. It seems this limit
just says when you want it to complain. :-(

# show running-config partition common | include snooping limit
ip igmp snooping limit track 0
# show ip igmp snooping explicit-tracking | include Current
    Current number of entries: 8015    Configured DB size limit: 0
#

Thanks for the input though. We're at least less nervous now. :-)

-- 
Peter