Experience case of BGP and OSPF routing validity issues at a certain point

125x Introduce Routing

The routes learned by ebgp on site suppress the routes introduced by OSPF, resulting in the traffic not being implemented as expected:

Preference 20 is configured on site, which leads to the better routing of ebgp than OSPF. On site, it indicates that if preference is configured first, BGP routing takes effect; if route introduction is configured first, OSPF routing takes effect.

This problem does exist in laboratory reproduction:

1) Configure preference before import OSPF:  

[H3C]display ip routing-table vpn-instance 11

Destinations : 26       Routes : 26

Destination/Mask   Proto   Pre Cost        NextHop         Interface

0.0.0.0/32         Direct  0   0           127.0.0.1       InLoop0

6.1.1.0/24         O_ASE2  150 1           60.1.1.10       Vlan60

7.7.7.7/32         O_ASE2  150 1           60.1.1.10       Vlan60

12.12.12.12/32     Direct  0   0           127.0.0.1       InLoop0

20.1.1.0/24        O_ASE2  150 1           60.1.1.10       Vlan60

20.20.20.0/24      BGP     20  0           30.1.1.30       Vlan30

[H3C]address-family  ipv4

[H3C]import-route ospf 10

[H3C]qu

[H3C]dis th

#

#

ip vpn-instance 11

  peer 20.1.1.2 as-number 200

  peer 30.1.1.30 as-number 200

  #

  address-family ipv4 unicast

   preference 20 150 130

   import-route ospf 10

   peer 20.1.1.2 enable

   peer 30.1.1.30 enable

#

return

[H3C]display ip routing-table vpn-instance 11

Destinations : 26       Routes : 26

Destination/Mask   Proto   Pre Cost        NextHop         Interface

0.0.0.0/32         Direct  0   0           127.0.0.1       InLoop0

6.1.1.0/24         O_ASE2  150 1           60.1.1.10       Vlan60

7.7.7.7/32         O_ASE2  150 1           60.1.1.10       Vlan60

12.12.12.12/32     Direct  0   0           127.0.0.1       InLoop0

20.1.1.0/24        O_ASE2  150 1           60.1.1.10       Vlan60

20.20.20.0/24      BGP     20  0           30.1.1.30       Vlan30

2) Configure import OSPF before preference:

#

ip vpn-instance 11

  peer 20.1.1.2 as-number 200

  peer 30.1.1.30 as-number 200

  #

  address-family ipv4 unicast

   import-route ospf 10

   peer 20.1.1.2 enable

   peer 30.1.1.30 enable

#

return

[H3C]display ip routing-table vpn-instance 11

Destinations : 26       Routes : 26

Destination/Mask   Proto   Pre Cost        NextHop         Interface

0.0.0.0/32         Direct  0   0           127.0.0.1       InLoop0

6.1.1.0/24         O_ASE2  150 1           60.1.1.10       Vlan60

7.7.7.7/32         O_ASE2  150 1           60.1.1.10       Vlan60

12.12.12.12/32     Direct  0   0           127.0.0.1       InLoop0

20.1.1.0/24        O_ASE2  150 1           60.1.1.10       Vlan60

20.20.20.0/24      O_ASE2  150 1           60.1.1.10       Vlan60

[H3C]address-family ipv4 unicast

[H3C]preference 20 200 130

[H3C]dis th

#

ip vpn-instance 11

  peer 20.1.1.2 as-number 200

  peer 30.1.1.30 as-number 200

  #

  address-family ipv4 unicast

   preference 20 200 130

   import-route ospf 10

   peer 20.1.1.2 enable

   peer 30.1.1.30 enable

#

return

[H3C]qu

[H3C]display ip routing-table vpn-instance 11

Destinations : 26       Routes : 26

Destination/Mask   Proto   Pre Cost        NextHop         Interface

0.0.0.0/32         Direct  0   0           127.0.0.1       InLoop0

6.1.1.0/24         O_ASE2  150 1           60.1.1.10       Vlan60

7.7.7.7/32         O_ASE2  150 1           60.1.1.10       Vlan60

12.12.12.12/32     Direct  0   0           127.0.0.1       InLoop0

20.1.1.0/24        O_ASE2  150 1           60.1.1.10       Vlan60

20.20.20.0/24      O_ASE2  150 1           60.1.1.10       Vlan60  

[H3C]display bgp routing-table ipv4 vpn-instance 11 20.20.20.0

BGP local router ID: 2.2.2.2

Local AS number: 100

Paths:   2 available, 1 best

BGP routing table information of 20.20.20.0/24:

Imported route.

 Original nexthop: 60.1.1.10

OutLabel        : NULL

Ext-Community   : <OSPF Domain Id: 0.0.0.0:0>, <OSPF Router Id: 12.12.12.12:0:

                   0>, <OSPF AreaNum: 0.0.0.0 RouteType: 5 Option: 1>, <RT: 111

                   :1>

RxPathID        : 0x0 

 TxPathID        : 0x0 

 AS-path         : (null)

Origin          : incomplete

Attribute value : MED 2, pref-val 32768

State           : valid, local, best

 IP precedence   : N/A

QoS local ID    : N/A

Traffic index   : N/A

From            : 30.1.1.30 (1.1.1.1)

Rely nexthop    : 30.1.1.30

Original nexthop: 30.1.1.30

OutLabel        : NULL

Ext-Community   : <RT: 111:1>

RxPathID        : 0x0 

 TxPathID        : 0xffffffff 

 AS-path         : 200

Origin          : igp

Attribute value : MED 0, pref-val 0

State           : valid, external

 IP precedence   : N/A

QoS local ID    : N/A

Traffic index   : N/A

After analyzing the logic:

1) If you configure preference first, the route learned by ebgp is 20, and the route learned by OSPF is 150. At this time, BGP route is in effect in the routing table, and then import OSPF is configured in BGP (Note: BGP can only import routes in effect in the routing table), so it cannot be imported. At this time, BGP routing is in effect;

2) When preference is not configured, ebgp learns 20 routes, while OSPF learns 150 routes. At this time, the effective route in the routing table is OSPF route type. At this time, import OSPF can be configured in BGP to introduce the OSPF route into BGP routing table (pref Val 32768); then configure preference. At this time, there are two routes in BGP routing table, while the one introduced by itself Pref Val is large, so OSPF is preferred for BGP routing table.