That's because of OSPF's default behavior of advertising loopbacks as /32 no
matter what the mask actually is. OSPF is advertising a /32 by default,
while the label bindings are being propagated with the correct mask (sh mpls
ldp binding <ip> <mask>) will show you this.
putting (ip ospf network point-to-point) on the loopback interface is
necessary if this what's being used as the next-hop. otherwise you'll have a
label binding with the correct mask, but a route to /32 and no binding to
it.
real world (always use /32 for mPLS router-ids, bgp next-hops), look out for
it when studying
------
Yandy Ramirez
On Mon, Apr 25, 2011 at 15:12, Matt Sherman <matt.sherman2_at_gmail.com> wrote:
> Hello All,
>
> I'm hoping someone can clear up my lack of understanding on MPLS LDP router
> ID subnetting.
>
> If your LDP router ID is a loopback address that is other than a /32 and
> you
> have the interface set to a point to point type OSPF interface, you see the
> "no host route" in the show output:
>
> R6#sh mpls ldp discovery
> Local LDP Identifier:
> 6.6.6.6:0
> Discovery Sources:
> Interfaces:
> Serial1/1 (ldp): xmit/recv
> LDP Id: 3.3.3.3:0; *no host route
> *
> Does the "no host route" even matter though? I have a dynamips lab that
> has
> that scenario and everything appears to be functioning fine. The CEs can
> ping each other across the VPN.
>
> If I take off the "ip ospf network point-to-point" from the loopback
> interface then I get the following error message:
>
> **Mar 1 01:23:58.131: %BGP-4-VPNV4NH_MASK: Nexthop 6.6.6.6 may not be
> reachable from neigbor 3.3.3.3 - not /32 mask
> *
> Then the VPN no longer works. I see the routes on the CEs but traceroutes
> die at the directly connected PE.
>
> Is this the expected behavior and is there a workaround? Also, is there a
> "debug ip packet" type command that is vrf specific?
>
> Thanks,
> Matt
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
Blogs and organic groups at http://www.ccie.net
Received on Mon Apr 25 2011 - 16:18:01 ART
This archive was generated by hypermail 2.2.0 : Sun May 01 2011 - 09:00:29 ART