Re: BGP to EIGRP redistribution query

kevinhobson2000 · ‎02-04-2011

Hi All,

I have a query about BGP to EIGRP route redistrbution.

Topology below:

R1 --------------- R2 ---------------- R3

bgp bgp to eigrp eigrp

Admin distance has been changed on R2 so the EIGRP route is preferred.

If i redistribute a route 1.1.1.1/32 from a static on R3 and from BGP on R2 the static being redistributed is always preferred even though it has a higher metric. See below:

IP-EIGRP (AS 1): Topology entry for 1.1.1.1/32
State is Passive, Query origin flag is 1, 1 Successor(s), FD is 25630720
Routing Descriptor Blocks:
10.0.0.1 (FastEthernet0/0), from 10.0.0.1, Send flag is 0x0
      Composite metric is (25630720/1662976), Route is External
      Vector metric:
        Minimum bandwidth is 100 Kbit
        Total delay is 1200 microseconds
        Reliability is 255/255
        Load is 3/255
        Minimum MTU is 1500
        Hop count is 1
      External data:
        Originating router is 10.0.0.1
        AS number of route is 0
        External protocol is Static, external metric is 0
        Administrator tag is 0 (0x00000000)

However if i shut the interface the BGP route is preferred with a lower metric:

R4#sh ip eigrp topology 1.1.1.1 255.255.255.255
IP-EIGRP (AS 1): Topology entry for 1.1.1.1/32
State is Passive, Query origin flag is 1, 1 Successor(s), FD is 512000
Routing Descriptor Blocks:
11.0.0.2, from Redistributed, Send flag is 0x0
      Composite metric is (512000/0), Route is External
      Vector metric:
        Minimum bandwidth is 10000 Kbit
        Total delay is 10000 microseconds
        Reliability is 255/255
        Load is 1/255
        Minimum MTU is 1500
        Hop count is 0
      External data:
        Originating router is 11.0.0.1 (this system)
        AS number of route is 1
        External protocol is BGP, external metric is 0
        Administrator tag is 0 (0x00000000)

If i change to using OSPF and do the same thing the redistributed route is preferred due to the lower metric.

This seems to be just for BGP to EIGRP and i was wondering why this is.

Is it because it is a EGP rather than a IGP?

Thanks

Kev

Giuseppe Larosa · ‎02-04-2011

Hello Kevin,

>> Admin distance has been changed on R2 so the EIGRP route is preferred.

you mean that external EIGRP routes D EX have now an AD < 20 instead of default 170?

>> If i redistribute a route 1.1.1.1/32 from a static on R3 and from BGP on R2 the static being redistributed is always preferred even though it has a higher metric

you have chosen to prefer EIGRP routes including external EIGRP routes over BGP routes. This decision is taken looking only at the AD, the EIGRP metric is not part of this decision. The EIGRP metric is a result of R3's redistribution with the metric values of the link betwen R2 and R3.

the metric would play a role only if the AD were equal for both the BGP route and the D EX route.

>> However if i shut the interface the BGP route is preferred with a lower metric:

Do you mean the interface between R2 and R3? if so the BGP route becomes the preferred route, it is installed in the IP routing table and it is then redistributed into EIGRP by local node R2 with the seed metric resulting by your settings on R2 itself.

>> If i change to using OSPF and do the same thing the redistributed route is preferred due to the lower metric.

OSPF has only one AD value by default, EIGRP has different AD values for different types of routes. This may be part of what you see here.

notice that also BGP has : AD 20 for eBGP and AD 200 for iBGP so also the type of BGP session matters.

Hope to help

Giuseppe

kevinhobson2000 · ‎02-04-2011

Hi Guseppe,

No i mean i have changed the admin distance on BGP 220 220 220 so the redistrbuted static into eigrp is preferred.

I have then redistrbuted the same route being advertised from bgp into eigrp.

So i have 2 routes in EIGRP.

However even though the redistributed BGP one has a metric of 512000 and the redistrbuted static one is well over 1million.

The EIGRP External route with over 1 million is installed in the topology table.

However if change to OSPF instead of BGP on R1 - R2 and redistrbute the EIGRP external route redistributed from OSPF takes precedence as it only has a metric of 512000.

It appears to be something specific to redistrbution regarding BGP when there is already a route via a different means.

Hope this explains it.

Cheers

Kev

kevinhobson2000 · ‎02-04-2011

Guiseppe,

It gets even more interesting if you redistribute bgp the redistrbute metric is massive which makes it invalid. See below:

*Mar 1 04:37:07.286: DUAL: rcvupdate: 1.1.1.1/32 via Redistributed metric 4294967295/4294967295

But if you redistrbute OSPF it redistrbutes it with the proper metric of 512,000:

*Mar 1 05:49:32.034: DUAL: rcvupdate: 1.1.1.1/32 via Redistributed metric 512000/0

Has to be something to do with the fact that BGP is external routing protocol and it will only use it as a last resort.

Cheers

Kev

milan.kulik · ‎02-04-2011

Hi,

from CiscoPress Troubleshooting IP Routing Protocols book (http://www.ciscopress.com/articles/article.asp?p=27839&seqNum=4) :

"The EIGRP topology table shows that the feasible distance (FD) is inaccessible (4294967295); the route is an external route that has been redistributed from OSPF. This means that Router B is receiving the 150.150.0.0/16 routes from Router C but is setting the FD as inaccessible because Router B is not using the EIGRP route in the routing table. As a matter of fact, the routing table entry in Router B is a RIP route for 150.150.0.0/16, as shown in Example 7-38. In other words, when the FD is inaccessible in the EIGRP topology table, the router is not using that EIGRP route in its routing table. Usually, the route is overridden by another routing protocol that has lower administrative distance."

HTH,

Milan

milan.kulik · ‎02-04-2011

Hi,

IMHO, you are missing one point here:

Any routing protocol A on a Cisco router can redistribute a prefix to other routing protocol B only in a case the prefix used by routing protocol A is the one which was chosen as best among all routing protocol and installed to the route RIB.

To make it clear:

BGP will redistribute a prefix to EIGRP only in a case the BGP route is really used for routing.

I.e., when you issue sh ip bgp 1.1.1.1 in your case, you must not see

r 1.1.1.1 ...

output (the "r" in the beginning of the output means RIB-failure, i.e., there was the same prefix received from other routing protocol which was chosen for routing.

So when you changed the admin distance on BGP 220 220 220, BGP fails to be the best routing protocol for 1.1.1.1 and that's why it can't redistribute the prefix to EIGRP.

You can also check by issuing

sh ip route 1.1.1.1

and you will not see "known via BGP ...".

HTH,

Milan