06-18-2008 07:53 PM - edited 03-05-2019 11:43 PM
hi all, i am confused with a very basic scenario so kindly guide me, this post will be a bit long so kindly bare it :)
i have 3 routers R1, R2, R3 connected via frame-relay network in hub n spoke topology. R1 is the hub, R2,R3 are spokes. Their relative configuration is given below
R1
int se 1/0
en frame-relay
no frame-relay inverse-arp
ip address 11.0.0.1 255.0.0.0
frame-relay map ip 11.0.0.2 102 broadcast <-- to R2
frame-relay map ip 11.0.0.3 103 broadcast <-- to R3
int lo 0
ip add 1.1.1.1 255.0.0.0
router rip
ver 2
no auto
net 11.0.0.0
net 1.0.0.0
On R2 and R3 similar configuration is present with frame-relay map pointing to 11.0.0.1 broadcast and running rip.
R2 loopback 2.2.2.2/8
R3 loopback 3.3.3.3/8
now i can see routes at R2 via rip
R 3.0.0.0/8 [-/-] via 11.0.0.3, serial 1/0
when i ping 3.3.3.3 from R2 and do debug i can see encapsulation failed, since 11.0.0.3 is not present in the frame-relay map command, so far its cool, no confusion at all. now see wht are my concepts of next hop resolution,
to go to 3.3.3.3
sh ip route 3.3.3.3
it will show next hop of 11.0.0.3
to reach 11.0.0.3
its showing directly connected via se 1/0
it will look for layer 3 to layer 2 mapping which is not present hence encap failed.
now this is wat i did to solve the issue ( which is my actual confusion )
i gave an static route
ip route 11.0.0.3 255.255.255.255 11.0.0.1
now when i again performed lookup see wat happened !
sh ip route 3.3.3.3
reach via 11.0.0.3
sh ip route 11.0.0.3
reach via 11.0.0.1
sh ip route 11.0.0.1
direclty connected via Se1/0
now if i ping 3.3.3.3 it fails !!, if i only change the route to this one
ip route 3.3.3.3 255.255.255.255 11.0.0.1
now ping is successful,
Can some1 really explain me why it didnt happen in the previous route ? when i have told the router that 11.0.0.3 is reachable via 11.0.0.1 whose mapping is present in the map table then why didnt it succeeded ??
Kindly explain me this in a convincing manner since i m really confused ( couldnt sleep last night ! )
06-18-2008 08:54 PM
On the spokes, is there just a map command to the hub, or also a map command to the other spoke?
In other words on R2, you should have:
frame-relay map ip 11.0.0.1 nnn broadcast
-and-
frame-relay map ip 11.0.0.3 nnn broadcast
(where nnn is the same, the DLCI pointing back to R1)
This gives r2 a layer 2 path to r3 (and do the reciprocal on r3).
When you do:
ip route 11.0.0.3 255.255.255.255 11.0.0.1
and the reciprocal, you can ping the 11.0.0.3 from R2. But...
you can't ping 3.3.3.3 because that's a RIP route and you're trying to do something like redistribute RIP into static, not feasible.
Your static of 3.3.3.3 to 11.0.0.1 works because R1 has both loopbacks in its routing table and the next hop is mapped.
The map statement gives RIP a way to find the next hop via layer 2.
06-18-2008 09:15 PM
Dear Sir, thanks for the feedback,
first i just have 1 mapping on each spoke to HUB, now my confusion is still at present, i know its very simple but dont know why its stuck :(
when i check the route to 3.3.3.3 its going via 11.0.0.3,
now i have told the router to reach 11.0.0.3 u need to go to 11.0.0.1, so wats wrong in that ? dont the flow will be like this
3.3.3.3 -> 11.0.0.3 -> 11.0.0.1 -> se 1/0
i hope u understand my confusion sir, this was my understanding of how recursive lookup takes place so do u think i m wrong in my concept ? i will highly appreciate if u could take some time for me and help me clear this confusion
Thanks in advance
06-19-2008 12:43 AM
Hi There
It is as the previous poster (Hennigan) has said. You have not told the R2 router how to get to R3 and visa versa.
You have stated that on your R2 & R3 routers you have configured 1 "Frame Map" statement
i.e. R2(config-if)#frame-relay map ip 11.0.0.1 201 broadcast
R3(config-if)#frame-relay map ip 11.0.0.1 301 broadcast
With this statement on each spoke router (R2 & R3) you have told it how to get to R1, the hub. But you have NOT told it how to get to the other spoke router.
You need to configure a second frame map statement on each of the spoke routers with the IP address of the remote spoke router and the DLCI to use to get to that spoke, which will be the same DLCI used to get to the Hub, as all traffic between spokes must go via the Hub.
So on each of the spoke routers configure the following under the relevant serial interface,
On R2:
R2(config-if)frame-relay map ip 11.0.0.3 201 broadcast
On R3:
R3(config-if)frame-relay map ip 11.0.0.2 301 broadcast
Also when using RIP as your routing protocol, you need to be aware of "Split Horizion" issues.
With Split Horizion, a router will not advertise a route back out of the interface it was learned on. Therefore R1, the Hub will not pass on the route information to the loopback IP address of R2 to R3 or the route information to the loopback IP address of R3 to R2.
This is because R1 is connected to the Frame Switch by Serial Interface 0. So when R1 learns the route to the loopback IP address 2.2.2.2 from R2 on S0 it will NOT send this information out to R3 as it's exit interface to R3 is S0. The same interface that this route information was learned on.
Your options are
a. Create a logical full mesh Frame topology (not really a scalable solution)
b. Use the interface configuration level command "no ip split-horizon" on R1's Serial 0 interface.
c. Use multipoint or point to point sub interfaces. See this link for more details http://www.cisco.com/en/US/tech/tk365/technologies_configuration_example09186a0080094054.shtml
HTH
Best Regards,
Michael
06-19-2008 01:03 AM
Dear Michael, i really appreciate ur feedback, but i m not asking for the solution :) i know how to solve this problem using frame-relay map commands, proper ip route(like i mentioned) and PBR. i am talking about the wrong static route that i had given ! can any 1 explain me why it wasnt working like i am thinking it should work, kindly dont mind any1 i m just confused regarding the next hop resolution is taking place, why didnt my first route worked ?? this is wat i am asking :)
(split horizon is disabled thts why i got the routes on R2 and R3 )
06-19-2008 01:30 AM
in reply to my own confusion, i think i cannot interrupt the router in its next hop resolution because router has already done it !!, in other words, am i right to conclude that we can never interrupt IGP next hop resolution ? ( i know we can do it using PBR but if i dont use PBR then ), i my case R2 learned 3.0.0.0/8 via 11.0.0.3, se1/0, it resolved to se1/0 because this is where it got the update right ? so now if i m telling it to reach 11.0.0.3 via 11.0.0.1 it will not consider it because it already resolved it to se1/0 right ?, so what i m thinking is that we can NEVER INTERRUPT IGP NEXT HOP RESOLUTION ! (if we r not using PBR ofcourse),
kindly confirm me if i m right
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide