Configure Backup Routing on Tenant VRF

You can enable backup routing when all the links from a leaf device to the spine layer are down, and tenant traffic is to be routed via the MCT neighbor.

About this task

Follow this procedure to configure backup routing on tenant VRF.

A pair of IPv4 and IPv6 address is allocated to each MCT pair across all the tenant VRFs, and the BGP session is established between the same IP Pair.

Procedure

  1. Allocate a pair of IPv4 and IPv6 address to each MCT pair across all the tenant VRFs.
    1. Allocate a bridge domain per VRF for backup routing.
    2. Allocate a corresponding router-interface VE per BD or VRF.
    3. Assign the IPv4 and IPv6 address allocated to each device on each of the VE interface.
      Note

      Note

      Same IPv4 and IPv6 address is allocated on each of the VE interface which belongs to different VRF.
    4. Establish IBGP IPv4 neighborship with the MCT peer on a set of IP address per VRF.
    5. Establish IBGP IPv6 neighborship with the MCT peer on a set of IPv6 address per VRF.
    6. Configure "next-hop-self" on both the IPv4 and IPv6 neighbor.
    7. Configure "active" on the IPv6 neighbor.

    Example:

    efa fabric setting update --name fabric1
    --backup-routing-ipv4-range 21.1.1.0/24 --backup-routing-ipv6-range 2001:21:1:1::0/120

    Example when backup routing is enabled:

    efa fabric setting update --name nc --backup-routing-enable yes
  2. Configure devices on tenant VRFs.
    The following table provides an example of device configuration on a tenant VRF:
    Table 1. Tenant1 VRF “vrf1”
    leaf-9250-173# show running-config bridge-domain 3001
    bridge-domain 3001 p2mp
     pw-profile default
     router-interface Ve 7001
     bpdu-drop-enable
    !
    
    leaf-9250-173# sh run in ve 7001
    interface Ve 7001 
     vrf forwarding vrf1
     ip address 21.1.1.10/31
     ipv6 address 2001:21:1:1::10/127
     no shutdown
    !
    leaf-9250-173# show running-config router bgp address-family ipv4 unicast vrf vrf1
    router bgp
     address-family ipv4 unicast vrf vrf1
      local-as 4210000001
      redistribute connected
      neighbor 21.1.1.11 remote-as 4210000001
      neighbor 21.1.1.11 next-hop-self
      maximum-paths 2
     !
    !
    
    leaf-9250-173# show running-config router bgp address-family ipv6 unicast vrf vrf1
    router bgp
     address-family ipv6 unicast vrf vrf1
      redistribute connected
      neighbor 2001:21:1:1::11 next-hop-self
      neighbor 2001:21:1:1::11 remote-as 4210000001
      neighbor 2001:21:1:1::11 activate
      maximum-paths 2
     !
    !
    Table 2. Tenant2 VRF “vrf2”
    leaf-9250-173# show running-config bridge-domain 3002
    bridge-domain 3002 p2mp
     pw-profile default
     router-interface Ve 7002
     bpdu-drop-enable
    !
    
    leaf-9250-173# sh run in ve 7002
    interface Ve 7002 
     vrf forwarding vrf2
     ip address 21.1.1.10/31
     ipv6 address 2001:21:1:1::10/127
     no shutdown
    !
    leaf-9250-173# show running-config router bgp address-family ipv4 unicast vrf vrf2
    router bgp
     address-family ipv4 unicast vrf vrf2
      local-as 4210000001
      redistribute connected
      neighbor 21.1.1.11 remote-as 4210000001
      neighbor 21.1.1.11 next-hop-self
      maximum-paths 2
     !
    !
    
    leaf-9250-173# show running-config router bgp address-family ipv6 unicast vrf vrf2
    router bgp
     address-family ipv6 unicast vrf vrf2
      redistribute connected
      neighbor 2001:21:1:1::11 next-hop-self
      neighbor 2001:21:1:1::11 remote-as 4210000001
      neighbor 2001:21:1:1::11 activate
      maximum-paths 2
     !
    !