Some thoughts and comments, some things to check.
1) the route target is just a BGP community. This is why we need to
configure the vpnv4 address-family, and configure send-community for that
BGP neighbor.
So where do you want to learn the routes? Just make sure you have the
route-target configured for import (as you do) and BGP will learn the routes
and place them into the vpnv4 bgp table.
If the local router does not need the routes, meaning that it does not have
the import statement anywhere, then the updates are not used. Again, the
route target is only a BGP community and tells BGP routes to import and for
which vpn.
Another important note, is that normal routing applies. CE devices need a
route there, and the far side needs a route back. make sure that you have
bi-directional route updates. It is possible to only advertise routes one
way ... You can check this with the show ip route command from the CE. One
sire will know about the other site, and the other site will not know about
it's peer.
When working with OSPF and with EIGRP, the redistribution into OSPF and
EIGRP is a bit complicated. In both cases you need some additional
commands.
With EIGRP in particular, you need an address family, a configured AS
number, and a default metric if redistributing BGP into it.
My suggestion for you would be to check each router and see if at least BGP
is getting the updates. Use the command 'show ip bgp vpnv4 vrf xyz and
check if the routes are propagated to the BGP peers.
Then, check if the updates are being sent to the CEs. HTH,
Andrew Lee Lissitz
On Mon, Nov 2, 2009 at 2:50 PM, Ivan <ivanzghr_at_gmail.com> wrote:
> P.S.
> About topology (if someone receives mail and everything is moved :):
> R6 is connected to PE R4...
>
> On Mon, Nov 2, 2009 at 7:47 PM, Ivan <ivanzghr_at_gmail.com> wrote:
> > OK. I have topology like this:
> >
> > R1 -- R3 -- R2 --R4 --R5
> > |
> > R6
> >
> > R1 and R5 are pure customer routers in VPNA, and R6 is customer router in
> VPNB.
> > R3 and R4 are PE routers and R2 is just P router.
> > R1 and R5 are using EIGRP as routing protocol, and R6 is running OSPF.
> > Everything is redistributed in MP-BGP and vice-versa for every VPN and
> > working OK.
> >
> > Task says that router R6 needs to participate in the same CUG as R1 and
> R5.
> >
> > On R3 I have vrf:
> > !
> > ip vrf VPNA
> > rd 1:20
> > route-target export 1:105
> > route-target import 1:105
> > !
> > on R4 vrf:
> > !
> > ip vrf VPNA
> > rd 1:40
> > route-target both 1:105
> > !
> > ip vrf VPNB
> > rd 1:50
> > route-target both 1:600
> >
> > As I understand RTs, when you say route-target both in some vrf it
> > means that all routes which has this ext BGP community will be
> > exported from vrf into MP-BGP, and imported from MP-BGP into vrf,
> > right?
> >
> > Shouldn't next solution be enough for R1 and R5 to have VPNB's OSPF
> > routes in its routing tables, and R6 to have VPNA's EIGRP routes in
> > its routing table as OSPF E2 routes:
> >
> > ip vrf VPNA
> > rd 1:20
> > route-target export 1:105
> > route-target import 1:105
> > route-target import 1:600 (importing routes that were exported from vrf
> VPNB)
> >
> > ip vrf VPNB
> > rd 1:50
> > route-target export 1:600
> > route-target import 1:600
> > route-target import 1:105 (importing routes that were exported into
> > MP-BGP from vrf VPNA)
> >
> > ?? After this R6 has all VPNA's routes as E2 routes but R1 doesn't
> > have VPNB's routes as D EX routes ??
> >
> > Solution of the task says just this:
> > ip vrf VPNB
> > route-target both 1:105
> >
> > Some clarification please. Thanks
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
>
>
>
>
>
>
-- Andrew Lee Lissitz all.from.nj_at_gmail.com Blogs and organic groups at http://www.ccie.netReceived on Mon Nov 02 2009 - 15:03:42 ART
This archive was generated by hypermail 2.2.0 : Tue Dec 01 2009 - 06:36:28 ART