MAC-flapping in bridge group

VladimirFilippov · ‎08-22-2007

Hi everyone! First of all sorry my bad English, it isn't my native language.

I've got a model of network, consisting of 4 switches (see "Test Network 1" file):

C3750G-24TS (Central, 12.2.37-SE (ADVANCED IP SERVICES))

C3750-48TS (Office_1, 12.2.37-SE (ADVANCED IP SERVICES))

C3550-48 (Office_2, C3550-48 EMI 12.2.37-SE (IP SERVICES))

C3550-48 (Office_3, C3550-48 EMI 12.2.37-SE (IP SERVICES))

See their config files in attachments

Network topology is a star. Without redundant links everything works fine.

When I'm enabling redundant links... (see "Test Network 1" file)... on switches which has bridge groups in their configs appears error messages like that:

%SW_MATM-4-MACFLAP_NOTIF: Host 001a.a164.0f01 in bridge group 1 is flapping between port Vl20 and port Vl10

%SW_MATM-4-MACFLAP_NOTIF: Host 001a.a164.0f02 in bridge group 1 is flapping between port Fa0/1 (Vl30) and port Fa0/1 (Vl10)

And so on... (See "Spanning tree" text file)

I suppose this is MAC-address of root switch in Spanning Tree. And I can't do anything to resolve this problem.

I found nothing in Cisco docs. Can you help me please?

P.S. This network model simulates a part of real network. Bridge groups are used to forward IPX between VLANs. I guess that no L3 topology solution is possible in this situation.

VladimirFilippov · ‎08-22-2007

Here are config files for Office_1(2,3) switches

VladimirFilippov · ‎08-22-2007

Here is config file for Central switch

Francois Tallet · ‎08-22-2007

Hi Vladimir,

First, I apologize in advance if this answer does not match your problem, I did not have time to go through all the information you provided. Just quickly looking at your problem though, I'm wondering if it could not be solved by CSCsi08513. The issue fixed in this bug is related to BPDUs that are coming from the same source in different vlans that are fall-back bridged. The mac move code incorrectly assume that those BPDUs are looping across the bridge-group (which is not the case, as BPDUs are not switched but received by the CPU). If this is matching your problem, this is a cosmetic issue and you can ignore the messages or upgrade your image. Else, you might have a bridging loop and we will need to look deeper into the problem.

Regards,

Francois

VladimirFilippov · ‎08-22-2007

Hi Francois!

Thank you for your post. At now I haven't resolve this problem but at least I understood what is happening in my network. Your explanation is sufficiently helpful.

I don't have access to Bug Toolkit with my guest account, that's why I can't continue to explore this problem.

May be you can advise me correct IOS version with this bug being resolved?

Francois Tallet · ‎08-24-2007

It's been fixed a while ago but it seems that it will only make it to the upcoming release. I think it's 12.2(40)SE (or something like that, I'm no bug or release expert;-) and should be out in a matter of days.

Regards,

Francois

Francois Tallet · ‎08-28-2007

12.2(40)SE is out now and include the fix for CSCsi08513.

Regards,

Francois