From: Jason Guy \(jguy\) (jguy@cisco.com)
Date: Sat Jun 16 2007 - 11:22:18 ART
Josepf,
Actually inverse arp is off, which made me look from another
perspective. I didn't have a static mapping on the mp subinterface.
After doing an initial test with another subinterface, it seems I was
not seeing that the multipoint subint with the frame-relay
interface-dlci command has a dynamic mapping. It clicked when you
pointed out that a p2p has no other way to go but out, so it creates a
static entry. MP inherently thinks it is connected to multiple.
Thanks for the response! I just couldn't see the forest because of the
trees. :)
Jason
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
Of
> Joseph Saad
> Sent: Saturday, June 16, 2007 9:10 AM
> To: ccielab@groupstudy.com
> Subject: RE: FR multipoint and p2p subinterfaces
>
> I believe broadcast is irrelevant in this situation.
>
> If you are using multipoint on the spokes and I am assuming you have
> inverse-arp ON, you must specify a frame-relay map ip a.b.c.d dlci at
each
> spoke pointing to the IP address of the other spoke but with the DLCI
> leading to the hub.
>
> It is probably working with point-to-point because of the logic of
route
> all
> traffic via the only DLCI approach.
>
> It's better to post your interface/subinterface configuration though
to
> help
> further.
>
> Joseph.
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
Of
> Jason Guy (jguy)
> Sent: Saturday, June 16, 2007 4:48 PM
> To: ccielab@groupstudy.com
> Subject: FR multipoint and p2p subinterfaces
>
> Hi Group,
>
> As soon as I thought I had a grasp of this, I found a scenario that I
> could not explain. I have a Frame-relay hub and spoke, where the hub
> uses the main interface and the spokes are required to use
> subinterfaces.
>
> Now, the main interface will default to NONBROADCAST. Knowing this, I
> thought I would use a multipoint subinterface which also defaults to
> NONBROADCAST. I set the ospf priority to 0 on the spokes. This made
> sense, the network types match, and all is good.
>
> However, I fought with this, and finally determined it was not going
to
> work. The OSPF came up, I had routes, but I could not ping spoke to
> spoke. So I decided to try a point to point subinterface which
defaults
> to POINT2POINT, and changed the ospf network type to NONBROADCAST.
The
> links came up and I could ping.
>
> My question is why? What is different at the frame-relay layer
between
> the MP and P2P subinterface and how OSPF sees this differently? I
would
> think if frame relay is pinging, and the Hub is the DR, either
> subinterface should send the packet to the hub. The fact that MP
> defaults to the correct seems a likely choice for this.
>
> The one interesting thing I see in the sh ip ospf neigh output is the
> way the neighbors appear. There is a FULL to the hub and ATTEMPT to
the
> other spoke.
>
> Neighbor ID Pri State Dead Time Address
Interface
> 10.2.2.2 1 FULL/DR 00:01:39 11.1.1.2
> Serial0/0.1
> N/A 0 ATTEMPT/DROTHER - 11.1.1.3
> Serial0/0.1
>
> Thanks,
> Jason
>
>
This archive was generated by hypermail 2.1.4 : Sun Jul 01 2007 - 17:24:49 ART