Cisco Support Community
Community Member

igmp packets flooding network

Has anyone else experienced an issue with the AP1131ag access points flooding the network with igmp packets? I have an AP1130 already in service which did not cause this to happen. And this has happened with 2 AP1131 that I have taken out of the box. I have disabled igmp snooping already but this has not helped. It basically causes a dos attack on the network and brings it down to igmp packets account for 95% of my entire network traffic.


Re: igmp packets flooding network

Just to be clear;

if your monitoring client is connected to a wired LAN, traffic appears normal, no extraordinary igmp traffic .....

But, when you connect the same monitoring client to the wireless LAN (only), that client sees a massive amount of igmp traffic, is that right?

Do you have any VLANS or additonal SSIDs configured?

Are you running any multicast applications on the wireless clients? Does the trace provide any additonal information as to where or what application the clients are trying to reach?

What version of IOS is running on the affected APs? Same as the "normal" ones?

And, just as a sanity check, you realize that if you are monitoring through a switchport in mormal mode, all you will see is broadcast / multicast traffic (and the traffic destined for the client's interface)? Are you also seeing a lot of ARP traffic?

No offense intended on the last point, but sometimes people forget ...

Let us know.

Good Luck


Community Member

Re: igmp packets flooding network

Yep, we at Ulm University have the same problem. I was hunting the bug the last week and here are the facts:

- at least 2 AP1131 at the same broadcast domain (needed to amplify/start the storm playing ping pong)

- both APs with factory default config, no need to enable

the radios to trigger the multicast storm

- Firmware Version C1130 Software (C1130-K9W7-M), Version 12.3(8)JA, RELEASE SOFTWARE (fc2)

- at ulm university it helps to disable igmp snooping

- we have more than one IGMP querier and the storm typically starts when more than one IGMP router queries the same multicast group

- the APs flood the IGMP query packet back via fa0 even if the packet arrived on fa0, big BUG!

- the two buggy APs now play multicast ping pong endlessly even when you unplug the switch with the connected APs from the VLAN

Best Regards

K. Gaissmaier

CreatePlease to create content