From: Jongsoo.Kim@Intelsat.com
Date: Thu Feb 17 2005 - 12:08:56 GMT-3
You won't have problem peering itself via link-local subnet.
If you add one more router in your example, you may see some non-sense
problem (of IPv6) of BGP routes using link-local address as next-hop
attribute.
In your example, R2 and R4 are direct ibgp peers.
But as eBGP will show the problem much clearer, let's say R2 and R4 are
direct eBGP peers using link-local address and let's say R4 has another iBGP
peer R1.
Now, if the below subject route-map is not used, R1 can't learn R2's eBGP
routes via R4 iBGP peer.
R1 can't resolve the next-hop address of R2 because it is link-local address
between R2 and R4, which is not supposed to be used outside of R2-R4 link.
Regards
Jongsoo
-----Original Message-----
From: Bob Sinclair [mailto:bsinclair@netmasterclass.net]
Sent: Wednesday, 16 February, 2005 9:41 PM
To: Kim, Jongsoo; swm@emanon.com; ccie2be@nyc.rr.com;
Upp_and_Upp@hotmail.com; ccielab@groupstudy.com
Subject: Re: BGP Neighbors for IPv6
Jongsoo,
The only problem I have with that document is that it does not appear to
describe the actual behavior of the IOS as I have observed it. Please see
my earlier post, which showed good peers between routers using link-local
address, update-source and no route-map. Please let me know what I am
missing.
Bob Sinclair
CCIE #10427, CCSI 30427, CISSP
www.netmasterclass.net
----- Original Message -----
From: <Jongsoo.Kim@Intelsat.com>
To: <swm@emanon.com>; <ccie2be@nyc.rr.com>; <bsinclair@netmasterclass.net>;
<Upp_and_Upp@hotmail.com>; <ccielab@groupstudy.com>
Sent: Wednesday, February 16, 2005 8:32 PM
Subject: RE: BGP Neighbors for IPv6
> For IPv6 BGP peering using link local address, I believe the answer is
> update-source AND( not or) route-map changing next-hop address to global
> unicast. Both are required.
> Update-source will tell BGP process to send link-local address out of
> which
> interfaces as multiple interface can have the same link-local address.
> And route-map changing next-hop address to global unicast will solve
> next-hop resolution problem of iBGP peers.
>
> Here is the summary of cisco web site
>
http://www.cisco.com/univercd/cc/td/doc/product/software/ios123/123cgcr/ipv6
> _c/sa_bgpv6.htm#wp1037072
>
> "Configuring IPv6 multiprotocol BGP between two IPv6 routers (peers) using
> link-local addresses requires that the interface for the neighbor be
> identified by using the update-source router configuration command and
> that
> a route map be configured to set an IPv6 global next hop."
>
>
> Regards
>
>
> Jongsoo
>
>
> -----Original Message-----
> From: Scott Morris [mailto:swm@emanon.com]
> Sent: Wednesday, 16 February, 2005 7:47 PM
> To: Kim, Jongsoo; ccie2be@nyc.rr.com; bsinclair@netmasterclass.net;
> Upp_and_Upp@hotmail.com; ccielab@groupstudy.com
> Subject: RE: BGP Neighbors for IPv6
>
>
> So update-source may be an answer, or perhaps next-hop-self?
>
> Scott
>
> -----Original Message-----
> From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
> Jongsoo.Kim@Intelsat.com
> Sent: Wednesday, February 16, 2005 6:19 PM
> To: ccie2be@nyc.rr.com; bsinclair@netmasterclass.net;
> Upp_and_Upp@hotmail.com; ccielab@groupstudy.com
> Subject: RE: BGP Neighbors for IPv6
>
> I think a route-map changing next-hop address to global unicast is
> necessary
> if link-local address is used for BGP peering.
> Otherwise, other BGP peers not part of link-local segment has no way to
> resolve link-local next-hop address for BGP route.
> A router can have identical link-local address for multiple interface,
> which
> is one of reasons why update source is necessary.
>
>
> Regards
>
> Jongsoo
>
>
> -----Original Message-----
> From: ccie2be [mailto:ccie2be@nyc.rr.com]
> Sent: Wednesday, 16 February, 2005 4:54 PM
> To: Bob Sinclair; Sean C; ccielab@groupstudy.com
> Subject: Re: BGP Neighbors for IPv6
>
>
> Hmmm, interesting.
>
> So, why do you think the VoD lesson on ipv6 bgp makes a point of using a
> route-map to set the next hop address to
>
> an ipv6 global unicast address if it's not required?
>
> Is that a best practice for some reason?
>
> Thanks, Tim
> ----- Original Message -----
> From: "Bob Sinclair" <bsinclair@netmasterclass.net>
> To: "ccie2be" <ccie2be@nyc.rr.com>; "Sean C" <Upp_and_Upp@hotmail.com>;
> <ccielab@groupstudy.com>
> Sent: Wednesday, February 16, 2005 4:19 PM
> Subject: Re: BGP Neighbors for IPv6
>
>
>> Hi Tim,
>>
>> Since the router always prompts for an outgoing interface when the
>> destination is link-local, it would make sense that it needs more
>> "direction" in order use the link-local address. I was able to get BGP
>> peers to come up using the configuration below. It peers to the far
>> link-local address and uses the outgoing interface in the update-source
>> command:
>>
>> On R2:
>>
>> router bgp 24
>> no synchronization
>> bgp log-neighbor-changes
>> neighbor FE80::204:C1FF:FE8E:C40 remote-as 24
>> neighbor FE80::204:C1FF:FE8E:C40 update-source FastEthernet0/0
>> no auto-summary
>> !
>> address-family ipv4 multicast
>> no auto-summary
>> no synchronization
>> exit-address-family
>> !
>> address-family ipv6
>> neighbor FE80::204:C1FF:FE8E:C40 activate
>> exit-address-family
>>
>> On R4:
>>
>> router bgp 24
>> no synchronization
>> bgp log-neighbor-changes
>> neighbor FE80::2D0:58FF:FE95:C8E1 remote-as 24
>> neighbor FE80::2D0:58FF:FE95:C8E1 update-source FastEthernet0/0
>> no auto-summary
>> !
>> address-family ipv4 multicast
>> no auto-summary
>> no synchronization
>> exit-address-family
>> !
>> address-family ipv6
>> neighbor FE80::2D0:58FF:FE95:C8E1 activate
>> exit-address-family
>>
>> Result:
>>
>> R4#sh bgp ipv6 summary
>> BGP router identifier 172.16.104.1, local AS number 24
>> BGP table version is 1, main routing table version 1
>>
>> Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down
>> State/PfxRcd
>> FE80::2D0:58FF:FE95:C8E1
>> 4 24 10 10 1 0 0 00:06:20
> 0
>> R4#
>>
>> Note good peer!!
>>
>> Bob Sinclair
>> CCIE #10427, CCSI 30427, CISSP
>> www.netmasterclass.net
>>
>>
>> ----- Original Message -----
>> From: "ccie2be" <ccie2be@nyc.rr.com>
>> To: "Bob Sinclair" <bsinclair@netmasterclass.net>; "Sean C"
>> <Upp_and_Upp@hotmail.com>; <ccielab@groupstudy.com>
>> Sent: Wednesday, February 16, 2005 3:31 PM
>> Subject: Re: BGP Neighbors for IPv6
>>
>>
>> > There's a VoD on Cisco's PEC site that talked about this.
>> >
>> > I watched it yesterday.
>> >
>> > In the example they used, if a link-local address was used in the
> neighbor
>> > <ipv6 -LL-addr> remote command, 2 other things were also required:
>> >
>> > 1) the neighbor <ipv6 LL-addr> source-update command and
>> >
>> > 2) a route-map that sets the next hop to the ipv6 unicast address for
> that
>> > neighbor.
>> >
>> > I'm not able to try that at the moment, but would you agree?
>> >
>> > Tim
>> > ----- Original Message -----
>> > From: "Bob Sinclair" <bsinclair@netmasterclass.net>
>> > To: "Sean C" <Upp_and_Upp@hotmail.com>; <ccielab@groupstudy.com>
>> > Sent: Wednesday, February 16, 2005 2:52 PM
>> > Subject: Re: BGP Neighbors for IPv6
>> >
>> >
>> >> Hi Sean,
>> >>
>> >> A certain gentleman of our mutual acquaintance assures me that either
> the
>> >> routable or link-local address could be used, if the neighbor is
> directly
>> >> connected. If you do use the link-local, however, you will need to
>> >> use
>> >> update-source. Give it a shot.
>> >>
>> >> HTH,
>> >>
>> >> Bob Sinclair
>> >> CCIE #10427, CCSI 30427, CISSP
>> >> www.netmasterclass.net
>> >>
>> >>
>> >> ----- Original Message -----
>> >> From: "Sean C" <Upp_and_Upp@hotmail.com>
>> >> To: <ccielab@groupstudy.com>
>> >> Sent: Wednesday, February 16, 2005 1:24 PM
>> >> Subject: BGP Neighbors for IPv6
>> >>
>> >>
>> >> > Hello,
>> >> >
>> >> > I have a question but it's not from any particular vendor's practice
>> > test.
>> >> > If
>> >> > I'm given a hub-and-spoke multipoint frame topology, and on the hub
> and
>> >> > spokes
>> >> > I place both IPv6 addresses and IPv6 link-local addresses with the
>> >> > appropriate
>> >> > frame maps for the neighbors, if I configure BGP over the topology:
>> >> > 1-do I need to apply the IPv6 address or the link-local address on
> the
>> >> > neighbor statements under the BGP process and address-family?
>> >> > 2-is there a difference if the neighbors are iBGP vs. eBGP?
>> >> > 3-can I use either address? Or, perhaps if using the link-local
>> >> > address
>> > I
>> >> > need to update-source?
>> >> >
>> >> > I'm trying to keep this a simple question (if possible). Hope this
>> > makes
>> >> > sense, and thanks in advance.
>> >> > Sean
>> >> >
>> >> >
> _______________________________________________________________________
>> >> > Subscription information may be found at:
>> >> > http://www.groupstudy.com/list/CCIELab.html
>> >>
>> >>
This archive was generated by hypermail 2.1.4 : Thu Mar 03 2005 - 08:51:21 GMT-3