I've got two redundant 35 mile MT BH's, with RSTP running on a MT box on one end.

Occasionally, we will have fade issue on our 'main' link. The lower priority (slower) link remains stable. The problem seems that if the primary link keeps rapidly dropping/reassociating for a while, the STP will keep trying to switch back on forth.

If I catch it, I will disable the primary link or lower it's priority on the STP box, but it's usually well before dawn.

Is there a wireless setting that would delay how long the interface waits to attempt to reconnect?

Of would a script that would disable the wireless interface for xx seconds if it disconnects be a possibility? (I'm not a very good scripter!)

Thanks

maybe changing the WDS cost and WDS cost range will help here?

The BH's are two seperate links, a pair of AP -> Station-WDS, going into a RB433 with RSTP running on the bridge. The main one is 5ghz Turbo, while the backup is 5ghz. It works great except when the main link bounces up and down.

There is actually more redundant BH's and RSTP boxes on past this bridge.
Part of the problem was that the watchdog on many of the boxes was set to ping the GW, and would start rebooting when a link upstream was bouncing, which would cause an even bigger chain reaction... I assume it was more of an ARP problem. I've turned off that the watchdog pings to see if that helps.

I'm saying you can reduce flapping by changing the WDS cost range. This way you can make "down" and "almost down" look the same, which is what you want.