Admittedly, this is a bit of a niche problem. An issue occurs when using MAC-based Topology pooling with Extreme Networks wireless (formerly identifi) where clients occupy the same physical space but can roam across multiple controller pairs and the topology pools are identical.
The goal of the configuration is for 1 client to always be placed into the same VLAN/Topology when roaming between controller pairs. What isn’t immediately apparent is that Extreme’s MAC-based pooling algorithm uses a topology index number as part of the hash. This index is created automatically based on the order in which the topology was created, if the Topologies were created in the wrong order a client will change VLANs when roaming from controller pair to controller pair causing the client to roam into a different L2 and L3 space with no triggering event for a new DHCP request.
This index number can be identified by hitting shell and running the following. The numbers inside of the brackets () will indicate the index number. Run the commands on all controllers to ensure they match so clients wont roam to different L3 space. If they do not match the only way to reorder is to delete and recreate the topology in the correct order.
## issue the following from the CLI
shell
cd /var/controller/debug/cm/
./shmapitest topologylist
## expected output
shmSystem::isReady() == true
Public-WiFi (18)1500
Public-WiFi-1 (2)1500
Public-WiFi-2 (3)1500
Public-WiFi-3 (4)1500
Public-WiFi-4 (5)1500
Public-WiFi-5 (6)1500
Public-WiFi-6 (7)1500
Public-WiFi-7 (8)1500
Public-WiFi-8 (9)1500