Thanks. There's an issue with discovering backreferences to extents with large numbers of subvolumes that would look like this. It's not the case here.