EIGRP strange behaviour

Hello!
There is topology in GNS3 (see attached file)
R26#sh ip eigrp topology all-links
IP-EIGRP Topology Table for AS(1)/ID(170.15.10.9)
Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
r - reply Status, s - sia Status
P 170.15.10.12/30, 1 successors, FD is 30720, serno 27
via 170.15.10.2 (30720/28160), FastEthernet0/0
via 170.15.10.10 (2174976/30720), Serial2/0
via 170.15.10.6 (286720/30720), Ethernet1/0
P 170.15.10.8/30, 1 successors, FD is 2169856, serno 14
via Connected, Serial2/0
via 170.15.10.6 (2200576/2174976), Ethernet1/0
P 170.15.10.4/30, 1 successors, FD is 281600, serno 24
via Connected, Ethernet1/0
via 170.15.10.10 (2200576/286720), Serial2/0
P 170.15.10.0/30, 1 successors, FD is 28160, serno 1
via Connected, FastEthernet0/0
via 170.15.10.10 (2177536/33280), Serial2/0
via 170.15.10.6 (289280/33280), Ethernet1/0
P 170.15.10.20/30, 1 successors, FD is 33280, serno 37
via 170.15.10.2 (33280/30720), FastEthernet0/0
via 170.15.10.10 (2172416/28160), Serial2/0
via 170.15.10.6 (286720/30720), Ethernet1/0

Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
r - reply Status, s - sia Status
P 170.15.10.16/30, 1 successors, FD is 33280, serno 33
via 170.15.10.2 (33280/30720), FastEthernet0/0
via 170.15.10.10 (2174976/30720), Serial2/0
via 170.15.10.6 (284160/28160), Ethernet1/0

The question is why the output
###################################
P 170.15.10.8/30, 1 successors, FD is 2169856, serno 14
via Connected, Serial2/0
via 170.15.10.6 (2200576/2174976), Ethernet1/0
P 170.15.10.4/30, 1 successors, FD is 281600, serno 24
via Connected, Ethernet1/0
via 170.15.10.10 (2200576/286720), Serial2/0
###################################
doesn't contain a string
via 170.15.10.2 (33280/30720), FastEthernet0/0"
like in
####################################
P 170.15.10.0/30, 1 successors, FD is 28160, serno 1
via Connected, FastEthernet0/0
via 170.15.10.10 (2177536/33280), Serial2/0
via 170.15.10.6 (289280/33280), Ethernet1/0
####################################
although this route goes through FastEthernet .

Interestingly, when i shutdown the Ethernet on the R26 , there are two strings - via Serial and FastEthernet.
The all configs are :

hostname R26
interface FastEthernet0/0
ip address 170.15.10.1 255.255.255.252
duplex auto speed auto
! interface Ethernet1/0
ip address 170.15.10.5 255.255.255.252
half-duplex
! interface Serial2/0
ip address 170.15.10.9 255.255.255.252
serial restart-delay 0
router eigrp 1
network 170.15.10.0 0.0.0.255
no auto-summary !

! hostname R27
! interface FastEthernet0/0
ip address 170.15.10.2 255.255.255.252
duplex auto speed auto
! ! interface FastEthernet3/0
ip address 170.15.10.13 255.255.255.252
duplex auto speed auto
! router eigrp 1
network 170.15.10.0 0.0.0.255
no auto-summary
end

hostname R28
! interface FastEthernet0/0
ip address 170.15.10.21 255.255.255.252
duplex auto speed auto ! interface Serial1/0
ip address 170.15.10.10 255.255.255.252
serial restart-delay 0
! router eigrp 1
network 170.15.10.0 0.0.0.255
no auto-summary
end

hostname R29
interface Ethernet1/0
ip address 170.15.10.6 255.255.255.252
half-duplex
! interface FastEthernet3/0
ip address 170.15.10.18 255.255.255.248
duplex auto speed auto
! router eigrp 1
network 170.15.10.0 0.0.0.255
no auto-summary
! end

hostname R30
interface FastEthernet0/0
ip address 170.15.10.14 255.255.255.252
duplex auto speed auto
! interface FastEthernet1/0
ip address 170.15.10.18 255.255.255.252
duplex auto
speed auto
! interface FastEthernet2/0
ip address 170.15.10.22 255.255.255.252
duplex auto
speed auto
router eigrp 1
network 170.15.10.0 0.0.0.255
no auto-summary
end

Thanks , Art.

EIGRP.jpg

Find more posts tagged with

Comments

fredrikjj

Split Horizon probably.

From R27's perspective it received and is using 170.15.10.8/30 from R26 with a next hop of 170.15.10.1. It won't, by default, send this route back to R26. The image is really small btw.

Danielh22185

Wow how do you even have a clue whats going on?

OP your command output is all ran together. If you go back and edit your spacing I'm sure we can help more.

artsenin

Thanks for Your replies .
Before i tried use "no split horisont" on the all roters but the result was same.

EdTheLad

What command did you use to disable split-horizon?

xnx

Not to be a b**ch or anything but the output would be much more readable for us if you uploaded a screenshot / or used the snipping tool in windows.

artsenin

I used "no ip eigrp split-horizont eigrp 1 " on the each interfaces .

fredrikjj

Upload the GNS3 .net file and the configs (the project folder) if you want to really got to the bottom of this. I still think the "problem" is that R27 is running split-horizon on its interface to R26 and that what you are seeing is the default behavior. But then again, I'm not an expert on EIGRP so maybe there's something I haven't considered.

EdTheLad

I've labbed it up, it doesn't appear that split-horizon can be disabled on an ethernet interface. I don't really see when you would every want to disable it on ethernet. As far as i can see it's only used on serial when you have subinterfaces.

R26#sh runn int eth 0/0
Building configuration...

Current configuration : 126 bytes
!
interface Ethernet0/0
bandwidth 100000
ip address 170.15.10.1 255.255.255.252
no ip split-horizon eigrp 1
delay 10

R27#debug eigrp packets terse
(UPDATE, REQUEST, QUERY, REPLY, UNKNOWN, PROBE, ACK, STUB, SIAQUERY, SIAREPLY)
EIGRP Packet debugging is on

R28
!
interface Serial4/0
ip address 170.15.10.10 255.255.255.252
delay 2000
serial restart-delay 0
!

R28(config)#int ser 4/0
R28(config-if)#delay 2001

R27#
*Mar 17 09:25:39.688: EIGRP: Received UPDATE on Et3/0 - paklen 45 nbr 170.15.10.14
*Mar 17 09:25:39.688: AS 1, Flags 0x0

NULL), Seq 66/0 interfaceQ 0/0 iidbQ un/rely 0/0 peerQ un/rely 0/0
*Mar 17 09:25:39.688: EIGRP: Enqueueing ACK on Et3/0 - paklen 0 nbr 170.15.10.14 tid 0
*Mar 17 09:25:39.688: Ack seq 66 iidbQ un/rely 0/0 peerQ un/rely 1/0
*Mar 17 09:25:39.693: EIGRP: Sending ACK on Et3/0 - paklen 0 nbr 170.15.10.14 tid 0
*Mar 17 09:25:39.693: AS 1, Flags 0x0

NULL), Seq 0/66 interfaceQ 0/0 iidbQ un/rely 0/0 peerQ un/rely 1/0

As you can see above R27 receives the update and sends an ACK back. It doesn't send the update to R26 as it see's R26 as it's next hop.

R7#sh ip route 170.15.10.8
Routing entry for 170.15.10.8/30
Known via "eigrp 1", distance 90, metric 2172416, type internal
Redistributing via eigrp 1
Last update from 170.15.10.1 on Ethernet0/0, 00:07:13 ago
Routing Descriptor Blocks:
* 170.15.10.1, from 170.15.10.1, 00:07:13 ago, via Ethernet0/0
Route metric is 2172416, traffic share count is 1
Total delay is 20100 microseconds, minimum bandwidth is 1544 Kbit
Reliability 255/255, minimum MTU 1500 bytes
Loading 1/255, Hops 1

fredrikjj

EdTheLad wrote: »

I've labbed it up, it doesn't appear that split-horizon can be disabled on an ethernet interface.

False.

I don't really see when you would every want to disable it on ethernet.

DMVPN?

Spoke-to-spoke DMVPN networks present a unique challenge because the spokes cannot directly exchange routing information with one another, even though they are on the same logical subnet. This limitation requires that the headend router advertise subnets from other spokes on the same subnet. This would normally be prevented by split horizon. In addition, the advertised route must contain the original next hop as learned by the hub router. A new command (no ip next-hop-self) was added to allow this type of operation. The no ip split-horizon eigrp AS and no ip next-hop-self eigrp AS commands are only configured on DMVPN headend routers

As far as i can see it's only used on serial when you have subinterfaces.

You probably mean when you don't have subinterfaces. Subinterfaces are considered separate interfaces with regards to split horizon meaning that you don't need to disable it.

And here's the routing table with R26's directly connected prefix received back from R27

R27(config-if)#no ip split-horizon eigrp 1
R27(config-if)#
*Mar  1 00:02:42.983: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 170.15.10.1 (FastEthernet0/0) is resync: split horizon changed

R26#show ip eigrp topology all-links
IP-EIGRP Topology Table for AS(1)/ID(170.15.10.9)


Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
       r - reply Status, s - sia Status


P 170.15.10.8/30, 1 successors, FD is 128256, serno 2
        via Connected, Loopback0
        via 170.15.10.2 (435200/409600), FastEthernet0/0

Also, note that it's completely pointless to disable split horizon in this scenario.

PS.
You've also disabled split horizon on the wrong router Ed. I bet you are just trying to make fun of us in the NP forum

EdTheLad

DMVPN

, yeah forget about that, haven't actually played with it yet. Yup multipoint interfaces is where the split-horizon is disabled, but u knew what i meant. Now i'm not seeing what your seeing, interesting, i'll have to look at it tomorrow, from the looks of it you just have two routers connected back to back advertising a loopback. Try it with the topology given if you have time.

fredrikjj

DMVPN , yeah forget about that, haven't actually played with it yet. Yup multipoint interfaces is where the split-horizon is disabled, but u knew what i meant.

In case using tunnels (DMVPN) is cheating, look at my broadcast interface EIGRP topology with pseudo nonbroadcast route propagation (lol).

Don't tell me that's not a real design!!!!

silly_eigrp_topology.jpg

EdTheLad

fredrikjj wrote: »

You've also disabled split horizon on the wrong router Ed. I bet you are just trying to make fun of us in the NP forum

Nah, i just copy and pasted the wrong interface config, it was 01:30 so my brain was a little fried. Since it wasn't working i started disbaling split-horizon everywhere

, even without the eigrp option.

So with the exact same topology as the original poster, i have the following:

R27#sh runn int eth 0/0
Building configuration...

Current configuration : 147 bytes
!
interface Ethernet0/0
bandwidth 100000
ip address 170.15.10.2 255.255.255.252
no ip split-horizon eigrp 1
no ip split-horizon
delay 10
end

R26#sh ip eigrp topology all-links
EIGRP-IPv4 Topology Table for AS(1)/ID(170.15.10.9)
Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
r - reply Status, s - sia Status

P 10.10.10.10/32, 1 successors, FD is 286720, serno 9
via 170.15.10.2 (286720/284160), Ethernet0/0
via 170.15.10.6 (540160/284160), Ethernet1/0
via 170.15.10.10 (2428416/284160), Serial4/0
P 170.15.10.12/30, 1 successors, FD is 30720, serno 5
via 170.15.10.2 (30720/28160), Ethernet0/0
via 170.15.10.10 (2174976/30720), Serial4/0
via 170.15.10.6 (286720/30720), Ethernet1/0
P 170.15.10.20/30, 1 successors, FD is 33280, serno 11
via 170.15.10.2 (33280/30720), Ethernet0/0
via 170.15.10.6 (286720/30720), Ethernet1/0
via 170.15.10.10 (2172416/28160), Serial4/0
P 170.15.10.8/30, 1 successors, FD is 2169856, serno 3
via Connected, Serial4/0
via 170.15.10.6 (2201344/2175744), Ethernet1/0
P 170.15.10.16/30, 1 successors, FD is 33280, serno 10
via 170.15.10.2 (33280/30720), Ethernet0/0
via 170.15.10.6 (284160/28160), Ethernet1/0
via 170.15.10.10 (2174976/30720), Serial4/0
P 170.15.10.0/30, 1 successors, FD is 28160, serno 1
via Connected, Ethernet0/0
via 170.15.10.10 (2177536/33280), Serial4/0
via 170.15.10.6 (289280/33280), Ethernet1/0
via 170.15.10.2 (30720/28160), Ethernet0/0
P 170.15.10.4/30, 1 successors, FD is 281600, serno 2
via Connected, Ethernet1/0
via 170.15.10.10 (2200576/286720), Serial4/0

R26#debug eigrp packets terse
(UPDATE, REQUEST, QUERY, REPLY, UNKNOWN, PROBE, ACK, STUB, SIAQUERY, SIAREPLY)
EIGRP Packet debugging is on
R26#

R28#sh runn int ser 4/0
Building configuration...

Current configuration : 104 bytes
!
interface Serial4/0
ip address 170.15.10.10 255.255.255.252
delay 2003
serial restart-delay 0
end

R28#config t
Enter configuration commands, one per line. End with CNTL/Z.
R28(config)#int ser 4/0
R28(config-if)#delay 2004

R26#
*Mar 17 10:48:13.805: EIGRP: Received UPDATE on Et1/0 - paklen 45 nbr 170.15.10.6
*Mar 17 10:48:13.805: AS 1, Flags 0x0

NULL), Seq 11/0 interfaceQ 0/0 iidbQ un/rely 0/0 peerQ un/rely 0/0
*Mar 17 10:48:13.805: EIGRP: Enqueueing ACK on Et1/0 - paklen 0 nbr 170.15.10.6 tid 0
*Mar 17 10:48:13.805: Ack seq 11 iidbQ un/rely 0/0 peerQ un/rely 1/0
*Mar 17 10:48:13.810: EIGRP: Sending ACK on Et1/0 - paklen 0 nbr 170.15.10.6 tid 0
*Mar 17 10:48:13.810: AS 1, Flags 0x0

NULL), Seq 0/11 interfaceQ 0/0 iidbQ un/rely 0/0 peerQ un/rely 1/0
R26#

R26#sh cdp neigh
Capability Codes: R - Router, T - Trans Bridge, B - Source Route Bridge
S - Switch, H - Host, I - IGMP, r - Repeater, P - Phone,
D - Remote, C - CVTA, M - Two-port Mac Relay

Device ID Local Intrfce Holdtme Capability Platform Port ID
R27 Eth 0/0 158 R B Linux Uni Eth 0/0
R28 Ser 4/0 141 R B Linux Uni Ser 4/0
R29 Eth 1/0 145 R B Linux Uni Eth 1/0

Total cdp entries displayed : 3
R26#

I'm only out of bed, so i'll start to investigate, been a while since i looked at distance vector protocols.

artsenin

Hello again! O-oh - DMVPN is going .. Deep thing.
I made other topology - only 3 routers . Each router in the corner of a triangle. There are 2 triangles. In first case all routers are connected through FastEthernet , and in second one - 1-st and 2-nd are connected through Ethernet .

The output of cli command "sh ip eigrp topology all-links " :

1-st case:

R1#sh ip eigrp topology all-links
IP-EIGRP Topology Table for AS(1)/ID(192.168.2.1)

Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
r - reply Status, s - sia Status

P 192.168.1.0/24, 1 successors, FD is 28160, serno 1
via Connected, FastEthernet0/0
P 192.168.2.0/24, 1 successors, FD is 28160, serno 2
via Connected, FastEthernet1/0
P 192.168.3.0/24, 2 successors, FD is 30720, serno 4
via 192.168.1.2 (30720/28160), FastEthernet0/0
via 192.168.2.2 (30720/28160), FastEthernet1/0

2-nd case :

R4#sh ip eigrp topology al
IP-EIGRP Topology Table for AS(1)/ID(192.168.2.1)

Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
r - reply Status, s - sia Status

P 192.168.1.0/24, 1 successors, FD is 28160, serno 1
via Connected, FastEthernet0/0
via 192.168.2.2 (286720/30720), Ethernet1/0
P 192.168.2.0/24, 1 successors, FD is 281600, serno 2
via Connected, Ethernet1/0
P 192.168.3.0/24, 1 successors, FD is 30720, serno 4
via 192.168.1.2 (30720/28160), FastEthernet0/0
via 192.168.2.2 (284160/28160), Ethernet1/0

The quiestion is why there are differents for P 192.168.1.0/24 ?

The GNS3 Topology and Configs :

autostart = False
[127.0.0.1:7200]
workingdir = /tmp
udp = 10001
3640
image = /home/art/Images/c3640-jk9s-mz.124-16.bin
ghostios = True
chassis = 3640
ROUTER R4
model = 3640
console = 2014
aux = 2100
cnfg = configs/R4.cfg
slot0 = NM-1FE-TX
f0/0 = R5 f0/0
slot1 = NM-1E
e1/0 = R6 e0/0
x = -713.0
y = 226.0
ROUTER R5
model = 3640
console = 2015
aux = 2101
cnfg = configs/R5.cfg
slot0 = NM-1FE-TX
f0/0 = R4 f0/0
slot1 = NM-1FE-TX
f1/0 = SW2 1
x = -447.0
y = 100.0
ROUTER R6
model = 3640
console = 2016
aux = 2102
cnfg = configs/R6.cfg
slot0 = NM-1E
e0/0 = R4 e1/0
slot1 = NM-1FE-TX
f1/0 = SW2 2
x = -413.0
y = 305.0
ROUTER R1
model = 3640
console = 2010
aux = 2103
cnfg = configs/R1.cfg
slot0 = NM-1FE-TX
f0/0 = R2 f0/0
slot1 = NM-1FE-TX
f1/0 = R3 f0/0
slot2 = NM-1E
x = -660.0
y = -191.0
ROUTER R2
model = 3640
console = 2011
aux = 2104
cnfg = configs/R2.cfg
slot0 = NM-1FE-TX
f0/0 = R1 f0/0
slot1 = NM-1FE-TX
f1/0 = SW1 1
x = -446.0
y = -379.0
ROUTER R3
model = 3640
console = 2012
aux = 2105
cnfg = configs/R3.cfg
slot0 = NM-1FE-TX
f0/0 = R1 f1/0
slot1 = NM-1FE-TX
f1/0 = SW1 2
slot2 = NM-1E
x = -361.0
y = -216.0
hx = 7.2421875
hy = -24.0
ETHSW SW1
1 = access 1 R2 f1/0
2 = access 1 R3 f1/0
x = -234.5
y = -362.0
ETHSW SW2
1 = access 1 R5 f1/0
2 = access 1 R6 f1/0
x = -214.5
y = 165.0
[GNS3-DATA]
configs = configs
m11 = 0.707106781187
m22 = 0.707106781187
width = 20000
height = 10000
NOTE 1
text = 192.168.1.0
x = -597.0
y = -323.5
color = "#ff0000"
NOTE 2
text = 192.168.2.0
x = -526.0
y = -214.5
color = "#ff0000"
NOTE 3
text = 192.168.3.0
x = -348.0
y = -326.5
color = "#ff0000"
NOTE 4
text = 192.168.1.0
x = -626.0
y = 120.5
color = "#ff0000"
NOTE 5
text = 192.168.3.0
x = -340.0
y = 191.5
color = "#ff0000"
NOTE 6
text = 192.168.2.0
x = -587.0
y = 301.5
color = "#ff0000"
NOTE 7
text = 2
x = 3.13663695334
y = 39.5077164418
interface = SW2 2
NOTE 8
text = 1
x = -10.2959694328
y = -32.7402646661
interface = SW2 1
NOTE 9
text = 2
x = 9.61736720137
y = 46.4973657782
interface = SW1 2
NOTE 10
text = 1
x = -21.4472614414
y = -20.0533639059
interface = SW1 1
NOTE 11
text = f1/0
x = 58.8826327986
y = -8.49736577823
interface = R3 f1/0
NOTE 12
text = f0/0
x = -46.8609093955
y = 36.3328519561
interface = R3 f0/0
NOTE 13
text = f1/0
x = 72.9472614414
y = 24.0533639059
interface = R2 f1/0
NOTE 14
text = f0/0
x = 2.94918443871
y = 48.399781895
interface = R2 f0/0
NOTE 15
text = f1/0
x = 72.8609093955
y = 18.6671480439
interface = R1 f1/0
NOTE 16
text = f0/0
x = 39.0508155613
y = -44.399781895
interface = R1 f0/0
NOTE 17
text = f1/0
x = 65.3633630467
y = -1.50771644185
interface = R6 f1/0
NOTE 18
text = e0/0
x = -45.6813107924
y = 18.8139214913
interface = R6 e0/0
NOTE 19
text = f1/0
x = 80.7959694328
y = -2.25973533388
interface = R5 f1/0
NOTE 20
text = f0/0
x = -51.1495135557
y = 12.1234537896
interface = R5 f0/0
NOTE 21
text = e1/0
x = 54.6813107924
y = 41.1860785087
interface = R4 e1/0
NOTE 22
text = f0/0
x = 64.1495135557
y = -34.1234537896
interface = R4 f0/0

hostname R1

interface FastEthernet0/0
ip address 192.168.1.1 255.255.255.0
duplex auto
speed auto
!
interface FastEthernet1/0
ip address 192.168.2.1 255.255.255.0
duplex auto
speed auto
!
interface Ethernet2/0
no ip address
shutdown
half-duplex
!
router eigrp 1
network 192.168.1.0
network 192.168.2.0
no auto-summary
!
end

hostname R2

interface FastEthernet0/0
ip address 192.168.1.2 255.255.255.0
duplex auto
speed auto
!
interface FastEthernet1/0
ip address 192.168.3.1 255.255.255.0
duplex auto
speed auto
!
router eigrp 1
network 192.168.1.0
network 192.168.3.0
no auto-summary

end
!
!

hostname R3

!
interface FastEthernet0/0
ip address 192.168.2.2 255.255.255.0
duplex auto
speed auto
!
interface FastEthernet1/0
ip address 192.168.3.2 255.255.255.0
duplex auto
speed auto
!
!
router eigrp 1
network 192.168.2.0
network 192.168.3.0
no auto-summary
!

end
!

hostname R4

interface FastEthernet0/0
ip address 192.168.1.1 255.255.255.0
duplex auto
speed auto
!
interface Ethernet1/0
ip address 192.168.2.1 255.255.255.0
half-duplex
!
router eigrp 1
network 192.168.1.0
network 192.168.2.0
no auto-summary
!

end
!
!

hostname R5

interface FastEthernet0/0
ip address 192.168.1.2 255.255.255.0
duplex auto
speed auto
!
interface FastEthernet1/0
ip address 192.168.3.1 255.255.255.0
duplex auto
speed auto
!
router eigrp 1
network 192.168.1.0
network 192.168.3.0
no auto-summary
end
!
!

hostname R6

interface Ethernet0/0
ip address 192.168.2.2 255.255.255.0
half-duplex
!
interface FastEthernet1/0
ip address 192.168.3.2 255.255.255.0
duplex auto
speed auto
!
router eigrp 1
network 192.168.2.0
network 192.168.3.0
no auto-summary
!

end

EdTheLad

So i've been playing around with this and it's very interesting. Thanks for posting this artsenin as i haven't looked at eigrp for about 3 years and this was a good exercise.

As per my previous post regarding the update 10.15.10.8 not being sent to R26 after the metric change, this was due to R27 having a preferred path to R26, since the successor did not change metric no update was sent. So nothing to do with split-horizon.

Getting back to the issue at hand.

I've used the same topology as artsenin, expect i've used R6 -> R10.
Due to not having fastethernet links, i've modified eigrp to just look at delay as a metric using the command "metric weight 0 0 0 1 0 0".

Fastethernet got delay 1
Ethernet delay 10
Serial delay 100

I disabled split-horizon for every physical link in the topology and added a loopback0 on each device.R6 has an extra ethernet port that i used for captures.

I've learnt something new about eigrp that i never knew before.

Eigrp has a router-id, this is elected the same as ospf, highest loopback etc etc. The main purpose of this router-id is t prevent loops when external routes are redistributed, this is due to not being able to run the feasible condition against a redistributed route due to an unknown metric. How it works is, if an external route is received containing the router-id of the local router, the route is ignored.

It turns out this RID is also sent for internal routes, this means turning split-horizon on or leaving it off on a point-to-point interface has no effect.

Now for a breakdown of whats happening:

R6#sh ip int bri | i \.
Ethernet0/0 170.15.10.1 YES NVRAM up up
Ethernet0/3 10.1.1.6 YES NVRAM up up
Ethernet1/0 170.15.10.5 YES NVRAM up up
Serial4/0 170.15.10.9 YES NVRAM up up
Loopback0 6.6.6.6 YES NVRAM up up

R6#sh ip eigrp topology all-links
EIGRP-IPv4 Topology Table for AS(1)/ID(66.66.66.66)
Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
r - reply Status, s - sia Status

P 170.15.10.12/30, 1 successors, FD is 512, serno 12
via 170.15.10.2 (512/256), Ethernet0/0
via 170.15.10.6 (3072/512), Ethernet1/0
via 170.15.10.10 (26112/512), Serial4/0
P 170.15.10.20/30, 1 successors, FD is 768, serno 14
via 170.15.10.2 (768/512), Ethernet0/0
via 170.15.10.6 (3072/512), Ethernet1/0
via 170.15.10.10 (25856/256), Serial4/0
P 170.15.10.8/30, 1 successors, FD is 25600, serno 10
via Connected, Serial4/0
via 170.15.10.10 (51200/25600), Serial4/0
via 170.15.10.6 (28672/26112), Ethernet1/0
P 9.9.9.0/24, 1 successors, FD is 128768, serno 29
via 170.15.10.2 (128768/128512), Ethernet0/0
via 170.15.10.10 (154112/128512), Serial4/0
via 170.15.10.6 (130560/128000), Ethernet1/0
P 170.15.10.16/30, 1 successors, FD is 768, serno 13
via 170.15.10.2 (768/512), Ethernet0/0
via 170.15.10.6 (2816/256), Ethernet1/0
via 170.15.10.10 (26112/512), Serial4/0
P 170.15.10.0/30, 1 successors, FD is 256, serno 1
via Connected, Ethernet0/0
via 170.15.10.6 (3328/76

, Ethernet1/0
via 170.15.10.10 (26368/76

, Serial4/0
via 170.15.10.2 (512/256), Ethernet0/0
P 7.7.7.0/24, 1 successors, FD is 128256, serno 25
via 170.15.10.2 (128256/128000), Ethernet0/0
via 170.15.10.10 (154112/128512), Serial4/0
via 170.15.10.6 (131072/128512), Ethernet1/0
P 170.15.10.4/30, 1 successors, FD is 2560, serno 20
via Connected, Ethernet1/0
via 170.15.10.6 (5120/2560), Ethernet1/0
via 170.15.10.10 (28672/3072), Serial4/0
P 6.6.6.0/24, 1 successors, FD is 26368, serno 3
via Connected, Loopback0
P 8.8.8.0/24, 1 successors, FD is 128768, serno 27
via 170.15.10.2 (128768/128512), Ethernet0/0
via 170.15.10.6 (131072/128512), Ethernet1/0
via 170.15.10.10 (153600/128000), Serial4/0
P 10.1.1.0/24, 1 successors, FD is 25600, serno 2
via Connected, Ethernet0/3
P 10.10.10.0/24, 1 successors, FD is 128512, serno 35
via 170.15.10.2 (128512/128256), Ethernet0/0
via 170.15.10.10 (153856/128256), Serial4/0
via 170.15.10.6 (130816/128256), Ethernet1/0

Both R6 and R8 have a local connection to the 170.15.10.8/30 network, this means that they will both be originators of this route.

R6 will advertise it to its neighbours R7,R8 and R9 with its RID 6.6.6.6, they will flood it to their neighbours and so every router will receive a copy of this update from R6 with a RID 6.6.6.6.

R8 will advertise it to its neighbours R6 and R10 with its RID 8.8.8.8, they will flood it to their neighbours, so every router will receive a copy of the update from R8 with RID 8.8.8.8

Each router will then perform a local computation and choose a successor.

R6 successor R6
R7 succcessor R6
R8 successor R8
R9 successor R10
R10 successor R8

Now looking 170.15.10.8/30 in respect to R6 and its neighbors.

R7 has a successor R6 which means the originator was 6.6.6.6, it will advertise the 170.15.10.8/30 network back to R6, R6 will ignore the route as it's router-id matches the originator router-id.

R9 has a successor R10, R10 has a successor R8, so we know the router learned on R9 has an originator-id 8.8.8.8. R9 will advertise this route to R6, R6 will accept the route and put it in the "topology all-links" database as the route originated on R8.

R8 is the originator so it will advertise the route to R6 with router-id 8.8.8.8, R6 will put in topology database.

Funny thing is, i haven't seen this documented any where.

fredrikjj

artsenin wrote: »

The GNS3 Topology and Configs :

I meant uploading the files so we could actually load the topologies quickly.

fredrikjj

It turns out this RID is also sent for internal routes, this means turning split-horizon on or leaving it off on a point-to-point interface has no effect.

I'm not seeing this. Whenever I try to lab the things in this thread, everything works exactly as expected*. The route is not learned when split horizon is turned on, and when I turn it off on the other side, it's learned. Even if it's self-originated and simply bounced off one other router on the other side of a p2p link.

*But I can't duplicate the original topology because I don't think the 3700 platform that I'm using supports plain 10 mbit ethernet ports. At least I can't find a module that has them.

PS.

This is an external route:

P 20.1.1.0/24, 1 successors, FD is 128256, serno 54
        via Rconnected (128256/0)

It has external information

External data:
        Originating router is 1.1.1.1 (this system)

It's not accepted when received back from a neighbor with split horizon turned off.

This is an internal route:

P 10.1.1.0/24, 1 successors, FD is 128256, serno 42
        via Connected, Loopback0
        via 10.1.12.2 (435200/409600), FastEthernet0/0

It doesn't have the external information. It's accepted back from the neighbor. The only different between these routes is that one is external and one is internal.

EdTheLad

Hi Fred, you don't need ethernet ports, use fast ethernet, change the bandwidth to 10Mb and multiply the delay by 10, that will replicate the cost of eth. Other than that, post the topology that your using with a description.

fredrikjj

Why do I care about the cost? My topology is R1 connected to R2. My only goal was to examine your theory about router-ids being used internally as well because it's not something I've ever heard. If it was true, why is R1 able to install a self-originated internal route, but not the self-originated external (20.1.1.0/24).

This is on router R1.

P 10.1.1.0/24, 1 successors, FD is 128256, serno 42
        via Connected, Loopback0
        via 10.1.12.2 (435200/409600), FastEthernet0/0
P 20.1.1.0/24, 1 successors, FD is 128256, serno 54
        via Rconnected (128256/0)

On R2 they are both identical:

R2#show ip eigrp top all-links 
P 10.1.1.0/24, 1 successors, FD is 409600, serno 72
        via 10.1.12.1 (409600/128256), FastEthernet0/0
P 20.1.1.0/24, 1 successors, FD is 409600, serno 73
        via 10.1.12.1 (409600/128256), FastEthernet0/0

EdTheLad

fredrikjj wrote: »

Why do I care about the cost? My topology is R1 connected to R2.

You said you couldn't recreate the topology due to not having an ethernet interface, i don't see why the physical interface is relevant since its easy to manipulate costs. Anyway with 2 routers back to back i don't see what you are seeing. What IOS are you using? i'm using 15.4 .

Regarding the RID, i've captured it in internal update packets sent between routers, the "show ip eigrp events" on R1 shows the route being ignored due to duplicate RID.

R1#sh runn int eth 0/0              
Building configuration...

Current configuration : 65 bytes
!
interface Ethernet0/0
 ip address 10.0.12.1 255.255.255.0
end

R1#sh runn int lo0
Building configuration...

Current configuration : 73 bytes
!
interface Loopback0
 ip address 1.1.1.1 255.255.255.0
 delay 1000
end

R1#


R1#sh runn | s router               
router eigrp 1
 network 0.0.0.0
R1#sh ip eigrp neigh
EIGRP-IPv4 Neighbors for AS(1)
H   Address                 Interface              Hold Uptime   SRTT   RTO  Q  Seq
                                                   (sec)         (ms)       Cnt Num
0   10.0.12.2               Et0/0                    12 00:16:05   11   100  0  4
R1#


R1#sh ip eigrp topology all-links 
EIGRP-IPv4 Topology Table for AS(1)/ID(10.0.12.1)
Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
       r - reply Status, s - sia Status 

P 1.1.1.0/24, 1 successors, FD is 256256, serno 3
        via Connected, Loopback0
P 10.0.12.0/24, 1 successors, FD is 281600, serno 1
        via Connected, Ethernet0/0
        via 10.0.12.2 (307200/281600), Ethernet0/0


R1#sho ip eigrp events 
Event information for AS 1:
1    19:35:49.909 Ignored route, metric: 1.1.1.0/24 metric(563200) 
2    19:35:49.909 Ignored route, dup routerid int: 10.0.12.1

R2#sh runn int eth 0/0
Building configuration...

Current configuration : 94 bytes
!
interface Ethernet0/0
 ip address 10.0.12.2 255.255.255.0
 no ip split-horizon eigrp 1
end

R2#sh runn | sec router
router eigrp 1
 network 0.0.0.0

R2#sh ip eigrp topology all-links 
EIGRP-IPv4 Topology Table for AS(1)/ID(10.0.12.2)
Codes: P - Passive, A - Active, U - Update, Q - Query, R - Reply,
       r - reply Status, s - sia Status 

P 1.1.1.0/24, 1 successors, FD is 409600, serno 3
        via 10.0.12.1 (537600/256256), Ethernet0/0
P 10.0.12.0/24, 1 successors, FD is 281600, serno 1
        via Connected, Ethernet0/0

fredrikjj

Interesting. I'm using 12.4(15)T14.

I'm seeing that the external route is ignored but not the internal. Presumably then this is a recently implemented feature.

3    07:15:23.586 Ignored route, metric: 20.1.1.0/24 435200 
4    07:15:23.586 Ignored route, neighbor info: 10.1.12.2 FastEthernet0/1 
5    07:15:23.586 Ignored route, dup router: 1.1.1.1

EdTheLad

I'll have a look at this later today with 12.4, i'm wondering is the discrepancy a bug or feature update that's been added.

fredrikjj

I googled a lot and finally found this:

https://supportforums.cisco.com/discussion/11286686/how-can-i-find-out-source-route-eigrp

We added the router-id to internal routes a couple of years ago when we changed the packet format to support some new capabilities. Unfortunately, we can't send the new tlvs to peers that don't understand it, so we have to strip it out when sending to older peers. Since those changes didn't make it into all branches (it was an enhancement after all) not all releases got the change.

EdTheLad

Mystery solved, this is why i'm so happy ccie is now updating to 15.x ios, no point in learning how protocols used to work lol.