I've got a Linksys SRW2048 attached to two Cisco 3560-E. Spanning-Tree is blocking one port and traffic flows fine.
But from time to time the management MAC address toggles to the blocked port (without STP TC: the port is still blocked).
Both 3560-E are logging these line:
%SW_MATM-4-MACFLAP_NOTIF: Host X.Y.Z in vlan ABC is flapping between port Po7 and port Po12
I found an old thread on homecommunity.cisco.com which exactly describes the problem but no fix:
only the management mac/ip address flaps between the core switches, machines on the linksys switches do not flap across the uplink ports.
To follow up, I had a ticket open with Linksys about this for about 3 months with no resolution. I submitted packet captures, stp outputs, etc and no luck. I gave up and basically had to revert to a manual failover for redundancy. It's no perfect or fast, but it works every time.
After reading your posting, and not knowing about known flapping of a RSTP port on a SRW2048, maybe if the product is new and under warranty, you are entitled to free phone support via the small Business Support Center. Just out of interest, is the product the old SRW2028 or the new refreshed 300 series p/n SRW2048-K9-NA?
There is a good reference to this sort of flip flopping behaviour regarding the 4500 series switch, which may still be relevant.
It still indicates that you should chat with the folks at the Small Business Support center via the URL below.
The reference to the problem suggest a potential issue with the following;
High CPU utilization
Slow traffic throughout the network
High backplane utilization on the switch
These may be some of the symptoms you are observing on the SRW2048, if a broadcast storm is generated by a flip flopping Spanning-tree interface.
I am actually the original poster of the old link. I have never found a resolution to this day. To make matters worse, I think the issue is present in the newer esw-5xx switches. They are basically never versions of the srw lineup from what I can tell.
My explanation for what it's worth is that the mainline Cisco switches use proprietary per vlan stp modes and not standard vlan modes. I'm guessing you are running your 3560's in RSTP or MSTP. For some reason the srw/esw can't figure out what to do with the stp informaiton sent to them containing the extra per vlan info, and on any vlan except the native vlan and simply don't recognize that they need to put a port into a non forwarding state.
I even had trouble with a single uplink on an esw-540 (running either rstp or mstp) going to a 6509 running mstp. I had to revert it to CSTP to get it to work at all. Crazy.
This is all speculation though, but backed up with my observations of the issues. I don't have time to do the grunt work that Cisco should be doing to make sure the switches they are putting out are at least interoperable between each other.
If someone else has some notion of how to resolve this, I WOULD LOVE to hear about it.
I've opened a support ticket at cisco. Since now I only got a reply with the question whether this is caused by a Port-Channel misconfiguration...
| Linksys SRW2048
As I've written - RSTP is working fine. Even when the 3560-E are using MSTP the SRW2048 is blocking one port. Without TC-Event and no increasing number transition to forwarding state of the blocking interface the secondary 3560-E learns the management MAC.
This also happens without Port-Channel and without (high) traffic on the interfaces.
I don't believe that the srw2048 is doing stp wrong but my guess is that an administrative task is sending packets even on blocked ports..
Together with the support we found the (simple) solution: Upgrade to firmware 1.2.2d..
I've upgraded a few switches and the symptoms seems to be vanished
The have got an internal note that this bug is fixed but yet it isn't listed in the official changelog
A short update: I've contacted the support guy today (for some reason his exchange mailbox is broken, but phone is a good fallback )..
He can't reproduce the problem because the german Linksys support team has no access to Catalyst switches...
We ended the call with the solution that he will escalate the problem.
After a week I called again...
Still haven't got a case number. The person responsible for my case was out for diner and called me back after a while.. He will now try to escalate the case.
Playing with wireshark:
0000 ff ff ff ff ff ff 00 1d 7e 81 06 89 08 06 00 01 ........~.......
0010 08 00 06 04 00 01 00 1d 7e 81 06 89 8d 64 38 97 ........~....d8.
0020 00 00 00 00 00 00 8d 64 38 bd 00 00 00 00 00 00 .......d8.......
0030 00 00 00 00 00 00 00 00 00 00 00 00 ............
This is a normal arp-request which is sent by the srw2048 on a blocked port.
Hi Jens -- Thanks for participating in the Support Community. Can you please provide the details regarding your support case? We do not offer email support and you should have been given a case number in the course of your original contact. Are you working with the Small Business Support Center (www.cisco.com/go/sbsc)?
Please verify the Support Center you contacted and any other details you have available (date, engineer's name, etc.). We'd like to help you but need this info.
Cisco Small Business
I have a related problem, and I am wondering what spanning tree mode you are using on the Cisco switches. I am not at all sure that rapid-pvst is compatible with the Linksys switches, and I am wondering which mode is best to interoperate between these two families.