Re: BGP Address Family

From: Marko Milivojevic <markom_at_ipexpert.com>
Date: Fri, 3 Jun 2011 21:48:28 -0700

Perfect. Now that we have basics out of the way, we can explore your
original questions :-).

Disabling "address-family ipv4 unicast" will indeed destroy any
configured neighbor sessions in IPv4. See the example:

R6 and R9 are interconnected over Se0/2/0 and Se0/2/1:

R6 Se0/2/0: 192.168.20.6/24
R6 Se0/2/1: 192.168.21.6/24
R9 Se0/2/0: 192.168.20.9/24
R9 Se0/2/1: 192.168.21.9/24

R6:
router bgp 69
 bgp log-neighbor-changes
 neighbor 192.168.20.9 remote-as 69
 !
 address-family ipv4
  neighbor 192.168.20.9 activate
  no auto-summary
  no synchronization
 exit-address-family

R9:
router bgp 69
 bgp log-neighbor-changes
 neighbor 192.168.20.6 remote-as 69
 !
 address-family ipv4
  neighbor 192.168.20.6 activate
  no auto-summary
  no synchronization
 exit-address-family

If I go to R6 and do this:

router bgp 69
 no address-family ipv4 unicast
!

This will be the result:

router bgp 69
 bgp log-neighbor-changes
 neighbor 192.168.20.9 remote-as 69
 !
 address-family ipv4
  no neighbor 192.168.20.9 activate
  no auto-summary
  no synchronization
 exit-address-family

However, when I create new neighbor statement OUTSIDE the
address-family, we'll see that neighbor will automatically be
activated in IPv4 unicast address-family:

R6(config)#router bgp 69
R6(config-router)# neighbor 192.168.21.9 remote-as 69
R6(config-router)#do sh run | s router bgp
router bgp 69
 bgp log-neighbor-changes
 neighbor 192.168.20.9 remote-as 69
 neighbor 192.168.21.9 remote-as 69
 !
 address-family ipv4
  no neighbor 192.168.20.9 activate
  neighbor 192.168.21.9 activate
  no auto-summary
  no synchronization
 exit-address-family

Now, if I wanted to disable this automatic creation of IPv4 unicast
neighbors whenever I created a neighbor, I would run a command "no bgp
default ipv4-unicast". Then, when I add the neighbor it won't
automatically get created. Also, I will be able to remove IPv4
address-family (which I was not able to do in this case, but it would
have worked if I had session active in another AF - just to make
things more complicated to explain the exact behavior to us
instructors). See:

R6(config)#router bgp 69
R6(config-router)#no bgp default ipv4-unicast
R6(config-router)#no address-family ipv4
R6(config-router)#do sh run | s router bgp
router bgp 69
 no bgp default ipv4-unicast
 bgp log-neighbor-changes
 neighbor 192.168.20.9 remote-as 69
 neighbor 192.168.21.9 remote-as 69

I hope this cleared out the purpose of "no bgp default ipv4-unicast"
and how it's different from the actual "address-family ipv4 unicast".

--
Marko Milivojevic - CCIE #18427
Senior Technical Instructor - IPexpert
FREE CCIE training: http://bit.ly/vLecture
Mailto: markom_at_ipexpert.com
Telephone: +1.810.326.1444
Web: http://www.ipexpert.com/
On Fri, Jun 3, 2011 at 19:35, Bilal Hansrod <bilal.hansrod_at_gmail.com> wrote:
> Thanks Marc,
>
> I believe, there are three options with this command (address-family ipv4
> unicast). i.e.
>
> *address-family ipv4* [*multicast* | *unicast* | *vrf* *vrf-name*]
>
> If I am only using address-family ipv4 unicast that I am not referencing to
> vrf and essence not using vrf. By researching more and reading this Blog (
> http://blog.ioshints.info/2008/06/use-explicit-address-family-ipv4-in-bgp.html),
> I understand it provides easy to read configuration feature if we are using
> all three options such as MPLS VPN, IPv6 or IP Multicast over BGP.
>
> Thanks,
>
>
> Bilal Hansrod
>
>
>
>
>
>
>
>
>
>
>
>
>
> B Sat, Jun 4, 2011 at 12:15 PM, marc edwards <renorider_at_gmail.com> wrote:
>
>> Will add 32 bits to header to carry vrf information over MPLS backbone.
>> Essentially that command enables MP BGP
>> On Jun 3, 2011 6:46 AM, "Bilal Hansrod" <bilal.hansrod_at_gmail.com> wrote:
>> > Hello Team,
>> >
>> > I was working on BGP lab and came across multiple address family command.
>> > Can anyone please explain one of the below command and correct other two
>> if
>> > it is incorrect.
>> >
>> > 1. address-family ipv4 unicast - I have disabled this command instead of
>> > disabling no bgp default ipv4-unicast. I didn't see any difference, may
>> be I
>> > don't know why we need this command.
>> >
>> > 2. no bgp default ipv4-unicast - This command is required to disable
>> normal
>> > IPv4 adjacency.
>> >
>> > 3. address-family vpnv4 unicast - This command will provide mechanism to
>> > carry additional attributes across BGP updates such as RT,SoO, Domain-ID,
>> > metric propagation and make true MP-BGP.
>> >
>> > I have seen some past post, but still little bit confused.
>> >
>> > Can anyone please shed some light on above points.
>> >
>> > Thanks,
>> >
>> > Bilal Hansrod
>> >
>> >
>> > 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
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
Blogs and organic groups at http://www.ccie.net
Received on Fri Jun 03 2011 - 21:48:28 ART

This archive was generated by hypermail 2.2.0 : Fri Jul 01 2011 - 06:24:27 ART