Route selection BGB and OSPF

blwegrzyn · ‎07-25-2011

I wonder if it is possible to see alternative routes for ospf and bgp.

For example if the router selects route x with metric y, can i see what is the next possible route if x with y dies?

I know that in Eirgp i could see the successors in the table.

thx

Jon Marshall · ‎07-25-2011

Bart

For bgp you can look at the BGP routing table - "sh ip bgp"

For OSPF you can look at the OSPF database "sh ip ospf database" although be aware you are looking at LSA not specifically routes.

Jon

blwegrzyn · ‎07-25-2011

Yes i noticed that routes that were choosen through bgp show as:

router# sh ip route 10.11.0.1

Routing entry for 10.11.0.0/23

Known via "bgp 64754", distance 20, metric 0

Tag 65000, type external

Redistributing via ospf 10

Advertised by ospf 10 metric 50 metric-type 1 subnets

Last update from xxxx 4w5d ago

Routing Descriptor Blocks:

* xxxx, from xxxx, 4w5d ago

Route metric is 0, traffic share count is 1

AS Hops 2

and ospf shows:

router#show ip ospf database | include 10.11.0.0

10.11.0.0 10.11.0.2 73 0x8000135D 0x0089FA 0

10.11.0.0 10.100.7.14 801 0x80000594 0x008A61 65000

10.11.0.0 10.200.7.14 365 0x80000595 0x006521 65000

what i was wondering is that i could see the metric for the routes that did not get to the routing table

thx

milan.kulik · ‎07-26-2011

Hi,

regarding metric:

"Routing entry for 10.11.0.0/23

Known via "bgp 64754", distance 20, metric 0"

metric 0 here is the BGP metric (MED attribute) received with the 10.11.0.0/23 perfix.

"Redistributing via ospf 10

Advertised by ospf 10 metric 50 metric-type 1 subnets"

is showing the OSPF metric assigned to the prefix when redistributed to OSPF.

You should be able to see this metric by issuing

show ip ospf database external 10.11.0.0

command.

HTH,

Milan

blwegrzyn · ‎07-26-2011

yes you are right, i can see it.

So now becuase the routes came from bgp into ospf with metric 50, they wil always come back from ospf with metric 50 because they are external routes?

router#show ip ospf database external 10.11.0.0

OSPF Router with ID (10.100.7.14) (Process ID 10)

Type-5 AS External Link States

Routing Bit Set on this LSA

LS age: 363

Options: (No TOS-capability, DC)

LS Type: AS External Link

Link State ID: 10.11.0.0 (External Network Number )

Advertising Router: 10.11.0.2

LS Seq Number: 8000137E

Checksum: 0x471C

Length: 36

Network Mask: /23

Metric Type: 1 (Comparable directly to link state metric)

TOS: 0

Metric: 20

Forward Address: 0.0.0.0

External Route Tag: 0

LS age: 392

Options: (No TOS-capability, DC)

LS Type: AS External Link

Link State ID: 10.11.0.0 (External Network Number )

Advertising Router: 10.100.7.14

LS Seq Number: 800005B2

Checksum: 0x4E7F

Length: 36

Network Mask: /23

Metric Type: 1 (Comparable directly to link state metric)

TOS: 0

Metric: 50

Forward Address: 0.0.0.0

External Route Tag: 65000

Routing Bit Set on this LSA

LS age: 1806

Options: (No TOS-capability, DC)

LS Type: AS External Link

Link State ID: 10.11.0.0 (External Network Number )

Advertising Router: 10.200.7.14

LS Seq Number: 800005B2

Checksum: 0x2B3E

Length: 36

Network Mask: /23

Metric Type: 1 (Comparable directly to link state metric)

TOS: 0

Metric: 50

Forward Address: 0.0.0.0

External Route Tag: 65000

router#

if i look at different network where I have:

bgp -ospf router4 ----------------------------MPLS-------------------------------------bgp -ospf router3

| |

router 1-ospf-----------------------200M LInk Ethernet------------------------------router 2-ospf

both ospf routers inject bgp into ospf with metric 50

all routers run ospf area 0.0.0.0

router 3#show ip route 10.100.0.0

Routing entry for 10.100.0.0/16

Known via "bgp 64754", distance 20, metric 0

Tag 65000, type external

Redistributing via ospf 10

Advertised by ospf 10 metric 50 metric-type 1 subnets

Last update from XXXX 1 7w0d ago

Routing Descriptor Blocks:

* XXXX, from XXXX, 7w0d ago

Route metric is 0, traffic share count is 1

AS Hops 2

Now i want to find out what are the candidate routes.

I want to see what metric are the routes for same network coming from:

router1 and router 2. Those are internal and both router 1 and route 2 know them.

But, router 1 also sends them to router 4 and then 4 distributes them to BGP so they will come to router 3 through BGP and this is how i want it for now. But I want to see what router 3 also sees as candidates and what metric they show. So:

router3#show ip ospf database external 10.100.0.0

OSPF Router with ID (10.200.7.14) (Process ID 10)

Type-5 AS External Link States

Routing Bit Set on this LSA

LS age: 1679

Options: (No TOS-capability, DC)

LS Type: AS External Link

Link State ID: 10.100.0.0 (External Network Number )

Advertising Router: 10.100.7.1

LS Seq Number: 80001C07

Checksum: 0x1802

Length: 36

Network Mask: /16

Metric Type: 1 (Comparable directly to link state metric)

TOS: 0

Metric: 20

Forward Address: 0.0.0.0

External Route Tag: 0

I see metric 20, but because administratice distance for bpg route is 20 and ospf 110 the bgp route wins.

thx again for help

milan.kulik · ‎07-26-2011

Hi,

if you are running mutual BGP/OSPF route redistribution without any tunning on your routers,

you can create suboptimal routing or even routing loops easily, I'm afraid.

Imagine following scenario:

Your MPLS connection fails on router 3.

As router 3 would keep receiving 10.11.0.0/23 prefix via OSPF (btw, metric higher than 50 as you are using type 1 metric) and not via BGP anymore, it will redistribute it to BGP.

With the attribute weight=32768 by default.

When the MPLS connection recovers again, router 3 will receive 10.11.0.0/23 prefix via BGP again. But due to the same prefix available in the BGP table already with the high weight attribute value, it will not consider the prefix received as the best within BGP process.

So it will keep routing to 10.11.0.0/23 prefix through the OSPF cloud.

So you should think what you really need/want here and probably add some prefix tagging and BGP attribute manipulation to your configurations.

HTH,

Milan

blwegrzyn · ‎07-26-2011

but would not the router 3 select bgp because of administrative distance?

milan.kulik · ‎07-27-2011

Hi,

no, as it would have the same prefix in the BGP table already (redistributed from OSFP) with better BGP attributes.

See

https://supportforums.cisco.com/message/3372479#3372479

for similar problem discussion.

HTH,

Milan

blwegrzyn · ‎07-27-2011

so to understand the routing better I created a test environment:

bgp 64000 -ospf 10 - router4 bgp 64000 -ospf 10 router3

----------------------------bgp 65000----r5---------------------------------

| |

router 1-ospf 10 -----------------------1G------------------------------router 2-ospf 10

the r4 and r3 redistributes ospf 10 to bgp process 64000 and bgp distibutes them to 65000

so the router5 learns about router 1 and router 2 networks from ospf through bpg

and learns about 4 and 3 through bgp

What is strange is that when i cut the link from router 1 to router 2

the routes on router 5 that came from ospf through bgp show in routing table as:

B 10.100.8.0 [20/2] via 2.2.2.2 , from router 3

B 10.100.7.0 [20/0] via 1.1.1.1 , - from router 4

B 10.200.7.0 [20/0] via 2.2.2.2 , - from router 3

but they never get redistibuted to router 4 although the command:

show ip bgp neigh 1.1.1.1 advertised-routes , show:

*> 10.100.8.0/24 2.2.2.2 2 0 64000 ?

*> 10.200.7.0/24 2.2.2.2 0 0 64000 ?

the router5 bgp process is configured as:

router bgp 65000

redistribute connected

neighbor 1.1.1.1 remote-as 64000

neighbor 2.2.2.2 remote-as 64000

no auto-summary

why router 3 does not learn about those networks?

do i need to tell router5 about 10.100.0.0/24 and 10.200.0.0/24

and put a statement

network 10.100.0.0

network 10.200.0.0

milan.kulik · ‎07-28-2011

Hi,

r4 is dropping the prefixes originated from r3 as you are using the same AS number 64000 on both routers but the prefix is received through eBGP from r5.

HTH,

Milan

blwegrzyn · ‎07-28-2011

Changing as to 63000 on r3 did the trick and now routes are distributed.

I wonder how come it works in our real live example where AS-es are the same?

I assume the prefixes are also received through the EBGP and both routers point to same AS neighboor.

Regardng the routing loops,in my test environment I created on r5 2 interfaces with static ips:

3.3.3.3

4.4.4.4

they are both distributed to router 4 and 3

router 3 shows:

Router#show ip bgp

BGP table version is 24, local router ID is 10.200.7.14

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

*> 1.1.1.0/24 2.2.2.100 0 0 65000 ?

* 2.2.2.0/24 2.2.2.100 0 0 65000 ?

*> 0.0.0.0 0 32768 ?

*> 3.3.3.0/24 2.2.2.100 0 0 65000 ?

*> 4.4.4.0/24 2.2.2.100 0 0 65000 ?

*> 10.100.7.0/24 2.2.2.100 0 65000 64000 ?

*> 10.100.8.0/24 10.200.7.1 2 32768 ?

*> 10.200.7.0/24 0.0.0.0 0 32768 ?

Router#

so it receives info from 2.2.2.100 router 5 about those two routes:3.3.3.3and 4.4.4.4

now i will intentionaly shut the interface on router 5 to kill link to router 3

after that router 3:

Router#

*Jul 28 15:29:48.900: %BGP-5-ADJCHANGE: neighbor 2.2.2.100 Down BGP Notification sent

*Jul 28 15:29:48.900: %BGP-3-NOTIFICATION: sent to neighbor 2.2.2.100 4/0 (hold time expired) 0 bytes

Router#

Router#show ip bgp

BGP table version is 30, local router ID is 10.200.7.14

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

*> 1.1.1.0/24 10.200.7.1 4 32768 ?

*> 2.2.2.0/24 0.0.0.0 0 32768 ?

*> 10.100.7.0/24 10.200.7.1 3 32768 ?

*> 10.100.8.0/24 10.200.7.1 2 32768 ?

*> 10.200.7.0/24 0.0.0.0 0 32768 ?

Router#

Router#show ip rou

Router#show ip route

Codes: C - connected, S - static, R - RIP, M - mobile, B - BGP

D - EIGRP, EX - EIGRP external, O - OSPF, IA - OSPF inter area

N1 - OSPF NSSA external type 1, N2 - OSPF NSSA external type 2

E1 - OSPF external type 1, E2 - OSPF external type 2

i - IS-IS, su - IS-IS summary, L1 - IS-IS level-1, L2 - IS-IS level-2

ia - IS-IS inter area, * - candidate default, U - per-user static route

o - ODR, P - periodic downloaded static route

Gateway of last resort is not set

1.0.0.0/24 is subnetted, 1 subnets

O 1.1.1.0 [110/4] via 10.200.7.1, 00:01:42, FastEthernet0/0

2.0.0.0/24 is subnetted, 1 subnets

C 2.2.2.0 is directly connected, GigabitEthernet1/0

3.0.0.0/24 is subnetted, 1 subnets

O E1 3.3.3.0 [110/54] via 10.200.7.1, 00:01:08, FastEthernet0/0

4.0.0.0/24 is subnetted, 1 subnets

O E1 4.4.4.0 [110/54] via 10.200.7.1, 00:01:08, FastEthernet0/0

10.0.0.0/24 is subnetted, 3 subnets

O 10.100.8.0 [110/2] via 10.200.7.1, 00:28:13, FastEthernet0/0

O 10.100.7.0 [110/3] via 10.200.7.1, 00:01:42, FastEthernet0/0

C 10.200.7.0 is directly connected, FastEthernet0/0

Router#

so the routes to 3.3.3.3 and 4.4.4.4 are now learned through ospf

i trace route:

Router#trace 4.4.4.4

Type escape sequence to abort.

Tracing the route to 4.4.4.4

1 10.200.7.1 12 msec 4 msec 20 msec

2 10.100.8.100 24 msec 4 msec 60 msec

3 10.100.7.14 44 msec 36 msec 40 msec

4 1.1.1.100 76 msec 44 msec *

now i bring the link back so bgp reinjects the routes back to r3

Router#

*Jul 28 15:35:27.068: %BGP-5-ADJCHANGE: neighbor 2.2.2.100 Up

Router#show ip bgp

BGP table version is 32, local router ID is 10.200.7.14

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

* 1.1.1.0/24 2.2.2.100 0 0 65000 ?

*> 10.200.7.1 4 32768 ?

* 2.2.2.0/24 2.2.2.100 0 0 65000 ?

*> 0.0.0.0 0 32768 ?

*> 3.3.3.0/24 2.2.2.100 0 0 65000 ?

*> 4.4.4.0/24 2.2.2.100 0 0 65000 ?

* 10.100.7.0/24 2.2.2.100 0 65000 64000 ?

*> 10.200.7.1 3 32768 ?

* 10.100.8.0/24 2.2.2.100 0 65000 64000 ?

*> 10.200.7.1 2 32768 ?

* 10.200.7.0/24 2.2.2.100 0 65000 64000 ?

*> 0.0.0.0 0 32768 ?

Router#trace 4.4.4.4

Type escape sequence to abort.

Tracing the route to 4.4.4.4

1 2.2.2.100 16 msec 4 msec *

Router#

and as you can see the bgp route wins again?

milan.kulik · ‎07-29-2011

Hi,

in real life your MPLS provider would use neighbor as-override command.

Or you could use allowas-in command.

Reagrding your example: Are you redistributing OSPF to BGP on R3?

HTH,

Milan

blwegrzyn · ‎07-29-2011

in my test environment i use:

Version 12.3(23), RELEASE SOFTWARE (fc5)

and the as-override is not available

regarding example R3:

router ospf 10

router-id 10.200.7.14

log-adjacency-changes

redistribute bgp 63000 metric 50 metric-type 1 subnets

network 2.2.2.0 0.0.0.255 area 0.0.0.0

network 10.200.7.14 0.0.0.0 area 0.0.0.0

!

router bgp 63000

no synchronization

bgp log-neighbor-changes

redistribute connected

redistribute ospf 10

neighbor 2.2.2.100 remote-as 65000

no auto-summary

!

I am learning routing so i was curious why the expected route was not created as you said?

milan.kulik · ‎07-29-2011

You might need to add

router bgp 63000

redistribute ospf 10 match external ...

(and internal plus possible more options)

to get the prefix received from BGP originally redistributed.

HTH,

Milan

blwegrzyn · ‎07-29-2011

yes, once i added external the route was created the way you explained

and it never got back to original shorter route after link to r5 from r3 recovered

r3 before the link is cut

Network Next Hop Metric LocPrf Weight Path

*> 3.3.3.0/24 2.2.2.100 0 0 65000 ?

r3 after the link is cut

Network Next Hop Metric LocPrf Weight Path

*> 3.3.3.0/24 10.200.7.1 54 32768 ?

r3 after link is restored

Network Next Hop Metric LocPrf Weight Path

* 3.3.3.0/24 2.2.2.100 0 0 65000 ?

*> 10.200.7.1 54 32768 ?

so this is how i understand it:

R5 distibutes 3.3.3.0/24 with weight 32768 to both r3 and r4. (why it shows 0 in R3 table)

(Routes that are injected to bgp from osp and distributed by ebgp with weight 0.)

they become active routes and and redistributed to ospf

Both r3 r4 at some point receive duplicate route about 3.3.3.0/24 from each other

with from ospf and they inject it to BGP.

When they are injected to BGP, bgp receives them and sets weight 32768 , but because it is as what is already in the routing table, they disregard it and keep initial route active ???

When link from r5 to r3 dies, the r3 removes route with metric 32768 and receives new from r4 that is redistributed by ospf and injects it with weight 32768 to bgp, and then like you said then link comes back nothing changes, cause the weight is the same. So to make this work the weight from BGP must be higher then from BGP through OSPF?

Where the weight should be manipulated? I assume that at the point when routes are incjeted from OSPF to BGP?

Why routes from BGP show weight 0?