David Edmondson wrote: > Cathy, I'm not sure if you were talking about shared-stack zones, > exclusive-stack zones or both. Different comments seemed to apply to > the different cases, but I really wasn't clear. > Sorry that I wasn't clear. But I was talking about the exclusive zone as implied in the subject.
> On Fri, Mar 02, 2007 at 03:42:29PM +0800, Cathy Zhou wrote: >> On the current Nevada release, one can plumb interfaces in two >> different zones with the same interface name ip.tun0 without a >> problem. > > I can see how this would be true with an exclusive stack, but is it > also true for shared-stack? > No. I don't think one should be able to plumb ip.tun0 in a shared zone. > In the next paragraph you say: > >> If we decide the link name should be per-zone, then it brings up a problem >> that how we represent the zone-local link name in a global zone. > > Interfaces (as distinct from links) in an exclusive-stack zone > shouldn't be visible in the global zone (I assert and hope). Links > probably are visible. > What I am trying to propose is that global zone won't be able to see links created in a exclusive zone either. Note that current ip.tun0 only exists in the IP layer, so it is an interface. But after Seb's Clearview tunnel component, it will also be a link, and this is essentially what brings up this discussion. > Perhaps I'm confused because you're talking about "links" as opposed > to "interfaces", or perhaps I'm just confused. > I am trying to say links when they are in the data-link layer and interfaces when they are in the IP layer. Thanks - Cathy
