I see these on occasion too. In our environment it is difficult to track down the end devices as they are guests and are quite temporary indeed.
My first guess would be just as you suggested, that the devices are roaming between APs that are on different controllers and that the mobilty groups are not properly configured. I will double check my mobility settings to verify.
Although, it misconfiguration were the issue, I would expect to see a lot more of these errors.
My second guess would be stale arp cache entries on the wired side then.....
The syslog entry is from a switch that connects the two 6505-E switches w/ wism blades
3747: Nov 20 02:29:19: %SW_MATM-4-MACFLAP_NOTIF: Host 0024.d26f.bbbb in vlan 18 is flapping between port Te0/1 and port Te0/2
I see the exact same error message on our Cisco 3750 switch. I enabled LAG
since we have more than 40 AP on our controller. Both the links of the controller are connected to the same switch as per Cisco documentation. Below is the error I see - "Nov 20 10:13:09 EST: %SW_MATM-4-MACFLAP_NOTIF: Host 00xx.xxxx.xxxx in vlan 1 is flapping between port Gi2/0/50 and port Gi2/0/49"
Actually, I meant to make sure that the mobility group settings are correct. There should be no issue of roaming between controllers. I believe he saw the issue on APs that were on the same controller, so inter-controller roaming would not be an issue.
Again, I see this issue very rarely. We will see over 20k wireless users throughout the year and I have seen it @ 5times. However, during one event, I saw the same mac address doing it quite a bit... Until I can dig up more info or duplicate it, I am going to chalk it up to client misconfiguration or cruddy wireless nic.