From: Ashok Ananda \(aananda\) (aananda@cisco.com)
Date: Thu Aug 11 2005 - 05:35:37 GMT-3
The BGP session comes up without route-map when using link local. But if
you are advertising IPV6 prefix using "network" command, advertised
routes will not appear in the next router's ipv6 bgp table.
So to have the advertised IP prefixes appear in the next router ipv6 bgp
table, you have to configure like this.
~~~~~~~~~~~~~~~~~~
Router A
Interface e2
ipv6 address 2001:db8:ffco:1::1/64
router bgp 1
no bgp default ipv4 unicast
bgp router-id 1.1.1.1
neighbor fe80::260:3eff:c043:1143 remote-as 2
neighbor fe80::260:3eff:c043:1143 update source e0
address-family ipv6
neighbor fe80::260:3eff:c043:1143 activate
neighbor fe80::260:3eff:c043:1143 route-map next-hop out
network 2001:db8:ffco:1::/64
route-map next-hop
set ipv6 next-hop 2001:db8:ffco:1::1
~~~~~~~~~~~~~~~~~~~
Hope this helps.. Let me know if I am wrong.
Thanks & Regards,
Ashok M A
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
Rohan Grover (rohang)
Sent: Thursday, August 11, 2005 11:02 AM
To: ashwin kohli; ccielab@groupstudy.com
Subject: RE: 2nd try: IPv6 BGP peer with Link local address
Thanks Ashwin/Brian
The CCO documentation talks about using a route-map and prefix-list to
set the ipv6 next hop to the global unicast address of the peer.
Thats why I had the confusion. But it does work fine without any
route-map/prefix-list.
Thanks
Rohan
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
ashwin kohli
Sent: Thursday, August 11, 2005 2:57 AM
To: Rohan Grover (rohang); ccielab@groupstudy.com
Subject: Re: 2nd try: IPv6 BGP peer with Link local address
Rohan,
Not too sure why you are having issues, as you don't need to use a
route-map or prefix-list. Just use the global unicast address as your
next-hop. Without knowing your configuration further, I have tried to
list the basic steps you will need to ensure IBGP IPv6 peer relationship
configuration:
STEP 1
-------
BGP Router ID <router-id> This must be manually configured on every
router. This is usually an IPv4 address and by default BGP uses the
highest loopback
IPv4 address on the router.
STEP 2
------
no bgp default ipv4-unicast By default, IPv4 address family is
advertised by each BGP routing session when the neighbor remote-as
command is configured.
STEP 3
-------
neighbor <ipv6-address> remote-as <AS> - IPv6 global unicast address and
the autonomous system of the EBGP or IBGP peer must be configured.
STEP 4
------
neighbor <link-local> update-source <interface> For IBGP peers using
the physical interface must be used to establish an IBGP relationship,
you must specify the a link-local address or a global unicast address on
physical interface through which the neighbor relationship will be
established. There is no need to use a route-map or prefix-list.
STEP 5
-------
address-family ipv6 unicast The BGP commands supported in address
family configuration mode configure the same functionality as the BGP
commands supported in router configuration mode
STEP 6
-------
neighbor <ipv6-address> activate This command enables the IPv6
neighbor to exchange prefixes for the
IPv6 address family with the local router.
STEP 7
------
Network <network/prefix> - This command is used to advertise IPv6
networks to the peers
To ensure that your configuration works, turn on "debug bgp ipv6" to
ensure the correct BGP states are being traversed through.
I hope that helps. Let me know if you have further questions.
Ashwin Kohli
CCIE#8877 (R&S, Security)
http://www.ccbootcamp.com/nliipv6labguide.asp
http://www.ccbootcamp.com/nlicciecore.asp
--- "Rohan Grover (rohang)" <rohang@cisco.com> wrote:
> Hi,
>
> Would really appreciate if someone could clear this doubt
>
> I'm a little confused by this feature. I understand that we need the
> update source command with the interface and a route-map used to set
> the next hop to the global unicast address.
> The problem is that, what prefix list should be used to match on the
> route-map? I mean what IPv6 address am i matching here?
>
> Thanks
> Rohan
>
>
This archive was generated by hypermail 2.1.4 : Sun Sep 04 2005 - 17:01:19 GMT-3