Posted on July 30, 2021 1:59 pm
 |  Asked by Cao
 |  86 views
0
0
Print Friendly, PDF & Email

(0.0.0.0/0)

ISP

|

bl1—-bl2(static 0/0)

|   /   \  |

sp1       sp2

|   /   \  |

lf1

 

Hi ,

In my setup I have 2 bl and 1 lf as shown above. my quesiton is simple . 2 bl has a virtual vtep 1.1.1.1

ISP advertise 0/0 to bl1 in vrf xyz via BGP. bl2 has a static 0/0 and redistribute to bgp too

lf1#show ip route vrf  xyz 0/0 detail

B E 0.0.0.0/0 [200/0] via   VTEP 1.1.1.1      VNI1000 router-mac xxxxxxxxx
via   VTEP 1.1.1.1       VNI 1000 router-mac xxxxxxxxx

 

static redistribute and ebgp are both selected and present in the route table.

1: Is a way to see the difference? or to see the detail of these 2 0/0 routes

both two default routes are selected in routing table and forwarding table?

2: show bgp evpn route-type ip-prefix 0/0 vni 1111 can’t see the difference of this 2 0/0

I guess the static one is preferred by lf1 ,but why? less AS PATH?

 

0
Posted by Sachin M Menon
Answered on July 31, 2021 5:12 pm

Hi Robbie,

Please let me state my understanding of your setup before we discuss the issue further:
- 2 Border Leaf in the setup with BL1 connected to ISP and getting default route in VRF XYZ.
- On BL2 there is a static default (0.0.0.0/0) route in VRF XYZ.
- Both Border Leafs are advertising VRF XYZ default (0.0.0.0/0) route to EVPN peers which are Spines (SP1 and SP2)
- Default route is received on LF1 and installed in routing table as follows:

lf1#show ip route vrf  xyz 0/0 detail
B E 0.0.0.0/0 [200/0] via   VTEP 1.1.1.1      VNI1000 router-mac xxxxxxxxx
		      via   VTEP 1.1.1.1      VNI1000 router-mac xxxxxxxxx

 

1: Is a way to see the difference? or to see the detail of these 2 0/0 routes
both two default routes are selected in routing table and forwarding table?
[-] We see that LF1 has installed route as a ECMP (Equal Cost Multi Path) which indicates that all the BGP Path Attributes (PA) considered for route selection were same for both the routes and hence we might not see difference between the prefixes.
Only difference we might see is for attributes RD (Route Distinguisher) and Evpn Router MAC.

2: show bgp evpn route-type ip-prefix 0/0 vni 1111 can’t see the difference of this 2 0/0
I guess the static one is preferred by lf1 ,but why? less AS PATH?
[-] Both prefixes are considered equal so both would have same AS Path length and we would not see any difference between the routes.
'show ip bgp 0.0.0.0/0 detail vrf XYF' output would display prefix as ECMP and also list all BGP PAs. Same can be check with 'show bgp evpn route-type ip-prefix 0.0.0.0/0 vni 1000 detail' output.

 

I have created a lab setup using 4.26.1F vEOS-lab image:

""

 

BL1 is receiving default route from ISP and BL2 has a static route. Both are advertising this route to EVPN peers:

vEOS-BL1#show ip route vrf ISP 0.0.0.0/0
VRF: ISP
Gateway of last resort:
 B I      0.0.0.0/0 [200/0] via 192.168.16.6, Ethernet2


vEOS-BL2#show ip route vrf ISP 0.0.0.0/0
VRF: ISP
Gateway of last resort:
 S        0.0.0.0/0 is directly connected, Null0

EVPN route received on LF1. Here we see Route Distinguisher and EvpnRouterMac values are different:

vEOS-LF1#show bgp evpn route-type ip-prefix 0.0.0.0/0 vni 1000 detail
BGP routing table information for VRF default
Router identifier 4.4.4.4, local AS number 45
BGP routing table entry for ip-prefix 0.0.0.0/0, Route Distinguisher: 1000:1.1.1.1
 Paths: 1 available
  3 12
    12.12.12.12 from 3.3.3.3 (3.3.3.3)
      Origin INCOMPLETE, metric -, localpref 100, weight 0, valid, external, best
      Extended Community: Route-Target-AS:1000:1000 TunnelEncap:tunnelTypeVxlan EvpnRouterMac:50:6d:00:1f:59:ac
      VNI: 1000
BGP routing table entry for ip-prefix 0.0.0.0/0, Route Distinguisher: 1000:2.2.2.2
 Paths: 1 available
  3 12
    12.12.12.12 from 3.3.3.3 (3.3.3.3)
      Origin INCOMPLETE, metric -, localpref 100, weight 0, valid, external, best
      Extended Community: Route-Target-AS:1000:1000 TunnelEncap:tunnelTypeVxlan EvpnRouterMac:50:6d:00:b7:03:f8
      VNI: 1000 

However, IPv4 BGP table for VRF ISP confirms that prefixes are considered ECMP:

vEOS-LF1#show ip bgp 0.0.0.0/0 vrf ISP
BGP routing table information for VRF ISP
Router identifier 55.55.55.55, local AS number 45
BGP routing table entry for 0.0.0.0/0
 Paths: 2 available
  3 12
    12.12.12.12 from 3.3.3.3 (3.3.3.3), imported EVPN route, RD 1000:1.1.1.1
      Origin INCOMPLETE, metric 0, localpref 100, IGP metric 0, weight 0, tag 0
      Received 00:13:38 ago, valid, external, ECMP head, ECMP, best, ECMP contributor
      Extended Community: Route-Target-AS:1000:1000 TunnelEncap:tunnelTypeVxlan EvpnRouterMac:50:6d:00:1f:59:ac
      Remote VNI: 1000
      Rx SAFI: Unicast
  3 12
    12.12.12.12 from 3.3.3.3 (3.3.3.3), imported EVPN route, RD 1000:2.2.2.2
      Origin INCOMPLETE, metric 0, localpref 100, IGP metric 0, weight 0, tag 0
      Received 00:28:45 ago, valid, external, ECMP, ECMP contributor
      Extended Community: Route-Target-AS:1000:1000 TunnelEncap:tunnelTypeVxlan EvpnRouterMac:50:6d:00:b7:03:f8
      Remote VNI: 1000
      Rx SAFI: Unicast

And same has been installed in the routing table:

vEOS-LF1#show ip route vrf ISP 0.0.0.0/0
VRF: ISP
Gateway of last resort:
 B E      0.0.0.0/0 [200/0] via VTEP 12.12.12.12 VNI 1000 router-mac 50:6d:00:b7:03:f8 local-interface Vxlan1
                            via VTEP 12.12.12.12 VNI 1000 router-mac 50:6d:00:1f:59:ac local-interface Vxlan1

I made ISP send default route 0.0.0.0/0 with longer AS path length:

vEOS-BL1#show ip bgp vrf ISP
BGP routing table information for VRF ISP
AS Path Attributes: Or-ID - Originator ID, C-LST - Cluster List, LL Nexthop - Link Local Nexthop
          Network                Next Hop              Metric  AIGP       LocPref Weight  Path
 * >      0.0.0.0/0              192.168.16.6          0       -          100     0       10 20 30 ?

And we see that LF1 prefers path towards BL2 as AS path length is shorter and installs the same in the routing table:

vEOS-LF1#show ip bgp 0.0.0.0/0 detail  vrf ISP
BGP routing table information for VRF ISP
Router identifier 55.55.55.55, local AS number 45
BGP routing table entry for 0.0.0.0/0
 Paths: 2 available
  3 12
    12.12.12.12 from 3.3.3.3 (3.3.3.3), imported EVPN route, RD 1000:2.2.2.2
      Origin INCOMPLETE, metric 0, localpref 100, IGP metric 0, weight 0, tag 0
      Received 01:11:19 ago, valid, external, best
      Extended Community: Route-Target-AS:1000:1000 TunnelEncap:tunnelTypeVxlan EvpnRouterMac:50:6d:00:b7:03:f8
      Remote VNI: 1000
      Rx SAFI: Unicast
  3 12 10 20 30
    12.12.12.12 from 3.3.3.3 (3.3.3.3), imported EVPN route, RD 1000:1.1.1.1
      Origin INCOMPLETE, metric 0, localpref 100, IGP metric 0, weight 0, tag 0
      Received 00:01:34 ago, valid, external
      Not best: AS path length
      Extended Community: Route-Target-AS:1000:1000 TunnelEncap:tunnelTypeVxlan EvpnRouterMac:50:6d:00:1f:59:ac
      Remote VNI: 1000
      Rx SAFI: Unicast
 Not advertised to any peer.
vEOS-LF1#show ip route vrf ISP 0.0.0.0/0
VRF: ISP
Gateway of last resort:
 B E      0.0.0.0/0 [200/0] via VTEP 12.12.12.12 VNI 1000 router-mac 50:6d:00:b7:03:f8 local-interface Vxlan1 

 

Thanks,
Sachin

0
Posted by Cao
Answered on August 2, 2021 11:01 am

Hi Sachin

Your setup is correct.

1:I have the same result as yours. The only difference is I have IBGP between bl1 and bl2. so even static 0/0 is removed from bl2, there is still ecmp there.I guess it is expected

2:For IPv4 BGP table for VRF ISP  ECMP, is it underlay ECMP or overlay ECMP? or a mixed one?

3: I find it is hard to traceroute dest ip behind vrf ISP from lf1. It is **** when traceroute traffic over vxlan fabric

Post your Answer

You must be logged in to post an answer.