bgp bestpath med missing-as-worst

[Kevin Dorrell]

I am trying the

[s.arunkumar · Accepted solution]

Hi kevin,

Isnt that ur update coming from two different AS,ie 300 n 200,hence the MED is not being compared..

Since ur med is not compared and both are learned via ebgp,next thing is the lowest path towards the next-hop.If both are same then,comes the update which is learned first.If that too ignored then,lowest router id is considered..

[Danilo Dy]

Hi,

Yes, S.Arunkumar is refering to the NOTE if using MED which states "This comparison only occurs if the first (the neighboring) AS is the same in the two paths."

Regards,

Dandy

[Kevin Dorrell]

Hi Arun,

I think you are right, the MED is not being compared because the updates are from two different AS. I was beginning to think along those lines when I took a coffee break to think about it.

My objective was to see how many different ways I could get this router to prefer the paths through 151.90.45.4 over the ones through 151.90.25.2. What confused me was that I had been playing with the MED, and it did seem to make a difference. But I think my experiment was misleading. This is what I did, and why I was mislead.

First I got 151.90.45.4 to mark MED=100. I then cleared that connection. Sure enough, the route through 151.90.25.2 was preferred, I thought because of the MED. (In fact it was because it was the oldest one!)

I then got 151.90.25.2 to mark MED=200, and cleared both connections. Sure enough, the route through 151.90.45.4 was preferred, I thought because of the MED. (In fact, it was because it was that one that came up first.)

I then took away the marking from 151.90.25.2, and set missing-as-worst on R5, expecting the route from 151.90.45.4 to be preferred. But it wasn't.

All the time I thought I was testing MED, but in fact I was just testing which route was oldest.

Now I shall go away and play with the "always compare MED" parameter.

Thanks again.

Kevin Dorrell

Luxembourg

[Danilo Dy]

Hi Kevin,

I don't know the answer for that as everything are the same and in order(according to BGP Best Path Algorithm), only the metric is different. But I just notice the metric for the best path, shouldn't it be 4294967294 and not 4294967295? From what I understand, if the metric learned from the BGP neighbor is 4294967295, it will change it to 4294967294 before injecting it into the routing table. Also, if you enabled "med missing-as-worst", the metric should be 4294967294.

Regards,

Dandy

[vidyaprakash]

Hi kevin,

If you use the following command, then it will prefer the one option of the BGP route.

This compare MED , even though there are in 2 different AS number , i.e routes coming from next BGP peer

bgp always-compare-med

Regards

PRakash

[Kevin Dorrell]

Hi Dandy,

Thank you (and thank you everyone else too) for your answers. I shall respond to each posting individually.

Concerning the point about the high metric, this was the metric the local router (R5) invented because of its med missing-as-worst.

I had put a metric of 100 on the outgoing route-map on neighbor RID 151.90.104.1, and no outgoing route-map at all on neighbor RID 151.90.102.1.

Kevin Dorrell

Luxembourg

[Kevin Dorrell]

Dandy, there is an interesting corollary to the discussion about 4294967294 and 4294967295.

If I do med missing-as-worst and the routes are coming in without a MED, then they get inserted with 4294967295, at least in 12.4(2)T:

R5#show ip bgp

BGP table version is 3, local router ID is 151.90.105.1

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

*>i10.10.10.0/24 151.90.50.10 0 100 0 i

*> 192.168.100.64/26

151.90.45.4 4294967295 1000 300 400 i

* 151.90.25.2 4294967295 0 200 400 i

Once they get to an iBGP neighbor, they still have that MED:

CAT1#show ip bgp

BGP table version is 2, local router ID is 151.90.108.1

Status codes: s suppressed, d damped, h history, * valid, > best, i - internal

Origin codes: i - IGP, e - EGP, ? - incomplete

Network Next Hop Metric LocPrf Weight Path

*> 10.10.10.0/24 0.0.0.0 0 32768 i

* i192.168.100.64/26

151.90.45.4 4294967295 100 0 300 400 i

But notice that the neighbor has chosen not to install the route. This looks like an IOS version inconsistency. If I remove the med missing-as-worst, then the neighbor gets the route and believes it:

CAT1#show ip bgp

BGP table version is 3, local router ID is 151.90.108.1

Status codes: s suppressed, d damped, h history, * valid, > best, i - internal

Origin codes: i - IGP, e - EGP, ? - incomplete

Network Next Hop Metric LocPrf Weight Path

*> 10.10.10.0/24 0.0.0.0 0 32768 i

*>i192.168.100.64/26

151.90.45.4 0 100 0 300 400 i

Kevin Dorrell

Luxembourg

Note: This posting will look better once they implement the "code" HTML tag.