In somewhat older IOS version there was a problem with CEF and MPLS. Could
be the case here as well. Typical symptom: it works with debugging on
(forcing the packet to leave the CEF path), but not without. I would think
though that this problem is solved by now.
Something you may want to try: clear the BGP session between the PEs. This
forces label to be distributed again, not really the answer to the problem,
but it sometimes works.
cheers,
Eduard
> ----------
> From: David Paraje[SMTP:david.paraje@telia.es]
> Sent: dinsdag 21 november 2000 12:46
> To: cisco-nsp@puck.nether.net
> Subject: RE: vpn question
>
> I had more or less the same problem with 12.1.3T.
> In my scenario r3 and r4 was dial-in users... and the PE routers was CISCO
> 7507.
>
>
> Our VPNS didnt have end to end connectivity, but localy yes...
> First of all, we note that if you made a debug tag-switching in the PEs it
> works... so you are forcing the routers to work in Process switching...
> Strange...
>
> But the real solution was that, in my case, the Virtual interface, when
> the
> router reloads, losses the "ip route-cache CEF" , in the config didnt
> apperar the no ip route-cache cef, but if you did sh running interfa
> virtual-template, you could see the no ip route-cache.
>
> Just reconfigure each time you reload the router, and It works!
>
> Thats my case, I hope it works...
>
> (because, all the tag-switching, and the address families and the simple
> config was ok , NO?)
>
> > -----Mensaje original-----
> > De: kevin Young [SMTP:kvyoung@21cn.com]
> > Enviado el: martes 21 de noviembre de 2000 11:14
> > Para: cisco-nsp@puck.nether.net; mpls-ops@mplsrc.com
> > Asunto: vpn question
> >
> > hi, group
> > when i config mpls vpn, r1 and r2 are 3640, as PE; r3 and r4 are
> 2610,
> > as CE. IOS are 12.1.3T
> > r3-------------r1--------------r2---------------r4
> > ether atm ether
> >
> > r1 could ping r3, and r2 could ping r4, r3 and r4 could see each other
> > routes, but r3 couldn't ping r4
> > and r1 couldn't ping r4, r2 could not ping r3.
> > i fould r1 and r2 didn't tag the vpn, this is the show on r1:
> >
> > r1#sh tag-switching for vrf vpn-lab detail
> > Local Outgoing Prefix Bytes tag Outgoing Next Hop
> > tag tag or VC or Tunnel Id switched interface
> > 16 Untagged 136.5.1.0/24[V] 0 Fa0/1 10.1.1.1
> > MAC/Encaps=0/0, MTU=1500, Tag Stack{}
> > VPN route: vpn-lab
> > Per-packet load-sharing
> > 17 Aggregate 10.1.1.0/24[V] 0
> > MAC/Encaps=0/0, MTU=0, Tag Stack{}
> > VPN route: vpn-lab
> > Per-packet load-sharing
> >
> > any advice are appreciated.
> >
> >
> > one another question: one PE in a AS through static route connects to
> > another PE in another AS,
> > could it can solve the mpls vpn could not span AS problem ? thanks.
>
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:21 EDT