03-01-2010 02:15 AM
while reading MPLS-VPN in some books, i found following config:
address-family ipv4 vrf custA
neighbor x.x.x.x remote-as v
neighbor x.x.x.x activate
as per my lab experience on GNS, neighbors are defined in global BGP mode while neighbors are activated in address-family vpn4 .
Than why do we have above configuration in address-family ipv4..???
03-01-2010 11:34 AM
Hello Mukarram,
You still have to have "neighbor x.x.x.x remote-as v" in order for the peering to be established
HTH
Reza
03-01-2010 02:47 PM
Hello Mukarram,
because that is an eBGP session with a CE node (eBGP as PE-CE protocol) so there is no sense in defining it at process level where it is unknown (CE ip address is known only in the appropriate VRF) being BGP process level related to global routing table.
your note is true for other address families like vpvn4 and so on but not for CE in VRFs
Hope to help
Giuseppe
03-01-2010 10:29 PM
Hi
The config you post that is nothing but the VRF confgi in BGP Please find the config example
Below config is as example when customer need to use BGP at CE router . So the neighbore will be form between CE- PE through Address Family created for that Customer with VRF.
CORE CONFIG
router bgp 10
bgp router-id X.X.X.X
neighbor X.X.X.X remote-as XXXX
neighbor X.X.X.X remote-as XXXX
!
address-family ipv4
neighbor X.X.X.X activate
neighbor X.X.X.X activate
no auto-summary
no synchronization
exit-address-family
!
address-family vpnv4
neighbor X.X.X.X send-community both
neighbor X.X.X.X activate
neighbor X.X.X.X activate
exit-address-family
VRF CONFIG FOR CUSTOMER
!
address-family ipv4 vrf chetan
neighbor x.x.x.x remote-as
neighbor x.x.x.x activate
redistribute connected
no auto-summary
no synchronization
exit-address-family
03-01-2010 11:18 PM
kindly see following, same BGP neighbor has been activated in global, VRF & VPNv4 BGP...
router bgp 1
bgp log-neighbor-changes
neighbor 10.200.254.2 remote-as 1
neighbor 10.200.254.2 update-source Loopback0
!
address-family ipv4
redistribute rip
neighbor 10.200.254.2 activate
exit-address-family
!
address-family vpnv4
neighbor 10.200.254.2 activate
neighbor 10.200.254.2 send-community both
exit-address-family
03-02-2010 12:18 AM
The mention config is done for PE .
The post config is not showing any VRF detail.
BGP Global Config
router bgp 1
bgp log-neighbor-changes
neighbor 10.200.254.2 remote-as 1
neighbor 10.200.254.2 update-source Loopback0
GLOBAL ADDRESS FAMILY FOR ISP NO RELATION WITH ANY VRF - It is redistributing RIP network in BGP
address-family ipv4
redistribute rip
neighbor 10.200.254.2 activate
exit-address-family
GLOBAL CONFIG FOR ISP or PE ROUTER VPNv4 to RUN MP-BGP VPN
address-family vpnv4
neighbor 10.200.254.2 activate
neighbor 10.200.254.2 send-community both
exit-address-family
And then you can start configuring the Customer VRF in BGP Address family
For Example :
Customer Address Family
address-family ipv4 vrf Cust_A
neighbor X.X.X.X remote-as XXXX ------ EBGP for CE
neighbor X.X.X.X activate
redistribute connected
exit-address-family
03-02-2010 12:42 AM
thank u..
address-family ipv4
redistribute rip
neighbor 10.200.254.2 activate
exit-address-family
neighbor 10.200.254.2 is already activated in vpnv4 add-family, i dont think we need to activate again in ipv4..
03-02-2010 01:19 AM
Please find more detail : For Same
BGP Global Config used for only Specify the Peers.
router bgp 1
bgp log-neighbor-changes
neighbor 10.200.254.2 remote-as 1
neighbor 10.200.254.2 update-source Loopback0
USE OF ADDRESS FAMILY IPv4
The IPv4 address family is used to identify routing sessions for protocols such as BGP that use standard IP version 4 address prefixes, So you need to activate in IPv4 for Routing Session.
address-family ipv4
redistribute rip
neighbor 10.200.254.2 activate
exit-address-family
USE OF ADDRESS FAMILY VPNv4
The VPNv4 multicast address family is used to identify routing sessions for protocols such as BGP that use standard VPN Version 4 address prefixes.VPNv4 routes are the same as IPv4 routes, but VPNv4 routes have a route descriptor (RD) prepended that allows replication of prefixes.
Here you can See if you intergrate the same with above config :
ip vrf Cust_A
rd 1:100
route-traget both 1:100
address-family ipv4 vrf Cust_A
redistribute Connected
neighbore X.X.X.X remote as XXXX
negihbore X.X.X.X activate
In this the address family is map with VRF Cust_A & RD is 100
For this BGP address family will run different routing session with RD of 100 for VRF Cust_A.
And the previously address family that is not associated with any VRF it won't use VPNv4 for route update and that run in Global BGP routing table.
03-02-2010 08:56 AM
Hello Mukarram,
>> neighbor 10.200.254.2 is already activated in vpnv4 add-family, i dont think we need to activate again in ipv4..
No, with MP BGP you can activate PER address-family so if you want to exchange IPv4 routes and VPNv4 routes with this neighbor you need to activate on both address families
http://www.cisco.com/en/US/docs/ios/iproute/command/reference/irp_bgp3.html#wp1030691
there is a default but it is the opposite for AF ipv4 unicast
As a best practice to avoid to depend from IOS image and platform I recommend to activate in all address families
In old times the switch from BGP to MP BGP was given by using
no bgp default ipv4-unicast
see
http://www.cisco.com/en/US/docs/ios/iproute/command/reference/irp_bgp1.html#wp1012800
if this is used you need to activate in each address family
Hope to help
Giuseppe
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide