BGP Next Hop Address Tracking

Lessons Discussion
1

This topic is to discuss the following lesson:

2

Hi Rene,
So, Here BGP identified event by OSPF and update next hop after 5 sec but how BGP identified event between two eBGP Router that are not directly connected .Between eBGP Router there is a L2 segment . So My Question is in the scenario how BGP detect & update next hop. It will takes max 60 sec , right ? Thanks

br//zaman

3

Hello Zaman

Remember that a BGP neighbourship can be formed between two BGP routers even if they are not directly connected. The only prerequisite is that there is adequate routing between them (using IGPs) so that they can exchange BGP information. So, even in that case, if the routing table of a BGP router changes and the route to the other BGP router becomes unavailable, the change in the routing table will take place which will indeed trigger the BGP next hop tracking mechanism.

I hope this has been helpful!

Laz

4

hi to everybody
i dont understand the topology .How did you connect them, use the switch ?

5

Hello Bahri

I assume it is this topology that you are talking about:
image
Essentially, the three routers are connected not as a point to point connection (obviously, since there are more than two interconnected devices), but as a multiple access network. This multiple access network can be interconnected by switches or any number of layer two devices. So in between the three routers there can indeed be a switch or any number of switches interconnecting these routers within a single subnet/network segment.

The switch was left out of the topology for simplicity’s sake…

I hope this has been helpful!

Laz

1 Like
6

hi
I have same question as Zaman, please help answer the question.

Thank you

7

Hello Hoan

It seems that Zaman’s question above was responded to. If you have any more specific question, please share it with us.

Laz

8

Hi Rene,

I have Problems with that.
2 Inside Routers form iBGP over EIGRP(announcing Loopback). EIGRP is running between the Routers on their HSRP addresses:
Router 1 10.10.10.1
HSRP VIP 10.10.10.2
Router 2 10.10.10.3
–> forming EIGRP 10.10.10.1<->10.10.10.3 and Advertising Loopbacks for BGP
All good so far.
Every Router is connected to ISP(same AS but does not matter here) which give a default route only to each inside Router.
That means both inside Routers have default to ISP, but when 1 inside Router chrashes, BGP is not deleted between inside Routers. EIGRP is removed and route to Loopback of other router also, but then they try to reach iBGP neighbor via ISP of Course.

9

Hi Alexander,

If you use private IP addresses on your network then in this case, traffic could match your default route and is forwarded to the ISP. They’ll drop it though since they don’t route private IP addresses.

If you want to prevent this, you could configure some null0 routes for your private ranges. When the router doesn’t have a more specific entry, it will drop the traffic instead of using the default route.

Rene

10

So I have to define a static route for bgp-used loopbacks with administrative distance worse than eigrp and point it to Null0?

11

Hello Alexander

Yes, I believe your description would would indeed provide the desired result.

Laz

1 Like
12

Hello,

I don’t understand why R1 chooses R3 as next-hop in the first place. R2 has a lower router ID (2.2.2.2) and all BGP attributes are the same.

Also I saw that R3 has not been configured to neighbor with R2. Is this intended or perhaps there is no need for full IBGP adjacency for this lesson?

Many thanks,
Stefanita

13

Hi Stefanita,

For this example, the iBGP adjacency between R2 and R3 isn’t required but to follow best practices, it should be there :slight_smile: It seems I added it for R2 but not for R3. It’s fixed now.

About the path selection, when everything is equal then it’s the oldest path that is selected. That could be R2 or R3, depending on which neighbor adjacency comes up first.

Rene

1 Like
14

My bad. Indeed oldest path is selected before router ID. :slight_smile: Many thanks Rene!

1 Like
15

Hi Rene ,
My question is that Next hop tracking is enable by default on every IOS ?
my second question is that is this work similarly like BFD ?

Kindly clarify my above questions …

Many Thanks!

16

Hello Tanmoy

According to this Cisco documentation, the next hop tracking is enabled by default for all versions of IOS. In addition it says:

BGP next-hop address tracking is enabled by default for IPv4 and VPNv4 address families. It is also enabled by default for the VPNv6 address family as of Cisco IOS Release 12.2(33)SB6.

Concerning BFD, there is a difference between functionality and what is actually being detected.

Next Hop Address Tracking will cause a BGP peer to preemptively remove a route from the BGP table if the next hop of that route is no longer in the routing table. Once the next hop IP is removed from the routing table, a next hop scan is initiated after 5 seconds, by default. This can be changed to 0 so that the scan occurs immediately.

BFD works differently. It creates BFD peerings between routers. BFD will send hello packets at sub-second intervals to make sure the neighbors are “alive”. If a BGP neighbor is no longer available, the BGP protocol immediately considers the neighbor down, and subsequently all routes learned from that neighbor must be reconverged,.

What’s the difference? The first is triggered by a change in the routing table, which itself is a process that may take several seconds, or tens of seconds depending on the actual failure and the underlying IGP being used. The second is triggered by a failure in BGP neighborship and “realization” of the changes are much more immediate.

So the two processes actually detect something different. You can have a neighbor fail without having the next hop fail, or vice versa.

I hope this has been helpful! Stay healthy and safe!

Laz

17

Hi Laz ,
Thanks for you response
So , here is the summary that I understand so far
Next Hop Tracking : Tracking of the next hop to reach the destination prefix , if next hop down the scan time will start after 5 seconds to find the optimal or available next hop & will install routing table accordingly.
BFD : It will detects the failure of neighbor by using BFD ECHO & reply message.

Please correct me if I wrong.

18

Hello Tanmoy

For next hop tracking, this is the process that is followed:

  1. Every 60 seconds, BGP scans the next hop addresses of all routes in the BGP table. If successful, everything remains the same. If it fails, any routes in the BGP table with a failed next hop will be removed, and BGP reconvergence begins.
  2. If the next hop of a particular BGP entry is removed from the routing table (for whatever reason such as failure, manual removal etc) then the scan time timer is reduced to 5 seconds, and when this expires, the scan takes place. Simultaneously, if an IGP is configured, it will begin reconvergence depending on how it has been configured.
  3. If the scan finds that the next hop address is no longer available, the route is removed from the BGP table
  4. BGP will begin the reconvergence process based on how it is configured.

Note here that the scan does not find the optimal next hop. The scan simply detects failed next hop addresses. It is the IGP and BGP that reconverge to find the best alternative route.

Yes, this is correct.

I hope this has been helpful! Stay healthy and safe!

Laz

19

Hi, I believe there may be an error with the following, as the “#debug ip bgp events” command is what allows me to view the BGP 60 second Scanner within my labs:

R1#debug ip bgp
*BGP debugging is on for address family: IPv4 Unicast *
You will see the following messages:

R1#
*Apr 9 09:56:53.743: BGP: topo global:IPv4 Unicast:base Scanning routing tables
*Apr 9 09:56:53.744: BGP: topo global:IPv4 Multicast:base Scanning routing tables
*Apr 9 09:56:53.746: BGP: topo global:L2VPN E-VPN:base Scanning routing tables
*Apr 9 09:56:53.747: BGP: topo global:MVPNv4 Unicast:base Scanning routing tables

*Apr 9 09:57:53.754: BGP: topo global:IPv4 Unicast:base Scanning routing tables
*Apr 9 09:57:53.757: BGP: topo global:IPv4 Multicast:base Scanning routing tables
*Apr 9 09:57:53.758: BGP: topo global:L2VPN E-VPN:base Scanning routing tables
*Apr 9 09:57:53.759: BGP: topo global:MVPNv4 Unicast:base Scanning routing tables

20

Thanks Laz for the response here.
2.If the next hop of a particular BGP entry is removed from the routing table (for whatever reason such as failure, manual removal etc) then the scan time timer is reduced to 5 – that means 60-5 seconds ? = 55 seconds ?
so its IGP and BGP responsibility to find the next hop or best alternative route.
if I configure max-paths with particular neighbor , then both primary & back up path will be installed in routing table … in that case convergence will works fast ?/