Hi.

After updating Omada Software Controller, it automatically readopts and reconfigures all switches and AP. During this adoption/configuration phase, the switches do not forward anything. Although this only takes a couple of seconds, it leads to major disturbances of clustered environments.  

Here are some ideas that might solve this problem:

1. The controller compares the current config of the switch with its own and only starts reconfiguration if there are some changes (which are usually not there after a controller update).

2. The switches are not adopted/configured all at once but subsequently, so there is always at least one working route for cluster communication.

3. Switches can be tagged as edge or core switches. At least for the core switches, #2 is applied.

Any other ideas?

Kind regards

Martin