09-17-2013 01:16 PM - edited 03-04-2019 09:04 PM
Dear All
I am testing my mpls configuration on a emulator and i have 2 HQ sites advertising the same loopback.
My idea is to tag the export route with a community and to import this particular community in remote site 1 instead of rmote site 2.
However, during my setup and testing, i find out that the edge router is not installing the 2 loopback that was advertising in my HQ sites.
PE2#sh ip bgp vpnv4 vrf CUST-YELLOW
BGP table version is 22, local router ID is 22.22.22.22
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete
Network Next Hop Metric LocPrf Weight Path
Route Distinguisher: 7473:2124 (default for vrf CUST-YELLOW)
*> 10.10.18.0/24 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 0 65001 ?
*>i20.20.20.20/32 11.11.11.11 0 100 0 65520 ?
*> 20.20.20.21/32 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 65001 ?
*>i172.16.21.20/30 11.11.11.11 0 100 0 ?
*> 172.16.21.24/30 0.0.0.0 0 32768 ?
* 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 65001 ?
*>i172.16.30.0/30 11.11.11.11 0 100 0 ?
*> 192.168.1.0/30 0.0.0.0 0 32768 ?
* 172.16.21.26 0 65001 ?
* 192.168.1.2 0 0 65001 ?
*> 192.168.5.0/30 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 0 65001 ?
PE2#
However in the global vpn table there is 2 routes of 20.20.20.20/30 being installed.
PE2#sh ip bgp vpnv4 all
BGP table version is 22, local router ID is 22.22.22.22
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete
Network Next Hop Metric LocPrf Weight Path
Route Distinguisher: 7473:2120
*>i20.20.20.20/32 11.11.11.11 0 100 0 65520 ?
* i 11.11.11.11 0 100 0 65520 ?
*>i172.16.21.20/30 11.11.11.11 0 100 0 ?
* i 11.11.11.11 0 100 0 ?
*>i172.16.30.0/30 11.11.11.11 0 100 0 ?
* i 11.11.11.11 0 100 0 ?
Route Distinguisher: 7473:2124 (default for vrf CUST-YELLOW)
*> 10.10.18.0/24 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 0 65001 ?
*>i20.20.20.20/32 11.11.11.11 0 100 0 65520 ?
*> 20.20.20.21/32 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 65001 ?
*>i172.16.21.20/30 11.11.11.11 0 100 0 ?
*> 172.16.21.24/30 0.0.0.0 0 32768 ?
* 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 65001 ?
Network Next Hop Metric LocPrf Weight Path
*>i172.16.30.0/30 11.11.11.11 0 100 0 ?
*> 192.168.1.0/30 0.0.0.0 0 32768 ?
* 172.16.21.26 0 65001 ?
* 192.168.1.2 0 0 65001 ?
*> 192.168.5.0/30 172.16.21.26 0 0 65001 ?
* 192.168.1.2 0 0 65001 ?
PE2#
The strange thing is 20.20.20.20/32 should be coming from 2 different AS PATH which is 65520 (installed) and 65522 (not installed).
Care to advice peeps. Attached is the topology that i used.
09-18-2013 07:23 AM
Per your diagram, PE2 is the Remote Edge Router.
Care to post the PE1 BGP table?
09-19-2013 11:54 AM
PE1#sh ip bgp vpnv4 vrf CUST-YELLOW
BGP table version is 16, local router ID is 11.11.11.11
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete
Network Next Hop Metric LocPrf Weight Path
Route Distinguisher: 7473:2120 (default for vrf CUST-YELLOW)
*>i10.10.18.0/24 22.22.22.22 0 100 0 65001 ?
*> 20.20.20.20/32 172.16.21.22 0 0 65520 ?
* 172.16.30.2 0 0 65522 ?
*>i20.20.20.21/32 22.22.22.22 0 100 0 65001 ?
* 172.16.21.20/30 172.16.21.22 0 0 65520 ?
*> 0.0.0.0 0 32768 ?
*>i172.16.21.24/30 22.22.22.22 0 100 0 ?
* 172.16.30.0/30 172.16.30.2 0 0 65522 ?
*> 0.0.0.0 0 32768 ?
*>i192.168.1.0/30 22.22.22.22 0 100 0 ?
*>i192.168.5.0/30 22.22.22.22 0 100 0 65001 ?
PE1#
Hi Edison
in my PE1 router, i can see that the path to go to the same network being advertise, but in PE2, the best path is already installed.
is thre anythin that i can do to have this bgp table inside PE2 bgp table ?
PE2#sh ip bgp vpnv4 vrf CUST-YELLOW
BGP table version is 14, local router ID is 22.22.22.22
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete
Network Next Hop Metric LocPrf Weight Path
Route Distinguisher: 7473:2124 (default for vrf CUST-YELLOW)
* 10.10.18.0/24 192.168.1.2 0 0 65001 ?
*> 172.16.21.26 0 0 65001 ?
*>i20.20.20.20/32 11.11.11.11 0 100 0 65520 ?
* 20.20.20.21/32 192.168.1.2 0 65001 ?
*> 172.16.21.26 0 0 65001 ?
*>i172.16.21.20/30 11.11.11.11 0 100 0 ?
* 172.16.21.24/30 192.168.1.2 0 65001 ?
* 172.16.21.26 0 0 65001 ?
*> 0.0.0.0 0 32768 ?
*>i172.16.30.0/30 11.11.11.11 0 100 0 ?
* 192.168.1.0/30 172.16.21.26 0 65001 ?
* 192.168.1.2 0 0 65001 ?
*> 0.0.0.0 0 32768 ?
* 192.168.5.0/30 192.168.1.2 0 0 65001 ?
*> 172.16.21.26 0 0 65001 ?
PE2#
As you can see the bgp route going to AS 65522 is already taken out as tie best path is installed already.
is there a way to remove this feature so that i can try and explore on using community tags to influece the routing policy ?
09-18-2013 07:26 AM
Hi Muhammad,
This is normal behavior. PE1 will select the best path and only advertise this one in the MPLS network.
Regards
09-19-2013 12:42 AM
Hi Harold
How are you my friend.
I am trying to test the environment where both site is advertising the same subnet, and i will need to tag with a community.
how do i configure in the PE so that this normal behaviour of selecting the best path in isntalled in the VRF routing table ??
Hi Edison
I will post to you the PE1 routing table by tonight
09-19-2013 03:07 AM
Hi Muhammad,
I am doing well. Hope you are doing well too. To accomodate the scenario, where the same prefix is advertised more than once, you would normally have one site connected to two PEs. Each PE would advertise the prefix into the MPLS network.
Bear in mind that if you have a RR in the network, you would need to make sure that both PEs have a different RD for the given VRF so that the RR sees the two VPNv4 prefixes as different and send them both to the other PEs. The other PEs would then be able to pick one or the other or simply use them both to load balance the traffic between the two egress PEs.
Regards
09-19-2013 12:06 PM
Hi Harold
i do have an RR in my lab, and i have a diff RD in the 2 PE.
PE1#sh run | s ip vrf CUST-YELLOW
ip vrf CUST-YELLOW
description "CUSTOMER YELLOW"
rd 7473:2120
route-target export 7473:1111
route-target import 7473:1111
PE1#
PE2#sh run | s ip vrf CUST-YELLOW
ip vrf CUST-YELLOW
description "CUSTOMER YELLOW"
rd 7473:2124
route-target export 7473:1111
route-target import 7473:1111
PE2#
somehow when the route is passed over to the RR, the best route will be installed and the other route of the same subnet will be removed in the bgp table.
09-19-2013 02:29 PM
Hi Muhammad,
You have a different RD for the ingress and egress PEs but you still have only one ingress PE (PE1) connected to the two HQ sites. When PE1 receives the same prefix from the two HQ sites, it selects the best path and only advertise this best paths towards the RR.
The scenario I was referring to is, two PEs, let's say PE1, connected to HQ site 1 and PE3, connected to HQ site 2. The two PEs would advertise the same prefix but with a different RD to the RR, making these two prefixes unique VPNv4 prefixes on the RR. The RR would propagate them both to PE2, who could then choose which ever one it prefers based on your routing policy.
Regards
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