ok, I've confirmed that between the cache being flushed and the route reappearing, there is *no* network traffic on the wireless. So this looks like a bug in the kernel, which is managing to get this bad route stuck somewhere such that it reappears as soon as I query the cache.
ok, I've confirmed that between the cache being flushed and the route reappearing, there is *no* network traffic on the wireless. So this looks like a bug in the kernel, which is managing to get this bad route stuck somewhere such that it reappears as soon as I query the cache.