> > I guess this is intended behavior? > > I had thought this was intended behavior as well but I see that a > patch is already prepped and tested to make this not happen. At any > rate it wasn't appearing to affect my usecase. I can't actually see how it'd affect any usecase, since you really need to check inside the new netns what's going on etc. anyway, and you don't really want to pass such identifiers across the boundaries. But preserving it makes more sense, if only for debugging and making sure we won't run out of numbers :) johannes
