Eigrp Flap

[navneet_78]

I have 3 routers connected to the same switch. There is Eigrp flap seen on the logs of the routers every few seconds. What might be the possible cause for this?? I could not find any physical errors on the Ethernet interface.

Regards

Navneet

[sundar.palaniappan]

Navneet,

Can you post the log error?

Sundar

[navneet_78]

These are the logs. Also i do see the broadcasts increasing on the ethernet segment.

Apr 24 13:13:17 EDT: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 254: Neighbor x.x.x.9 (Ethernet0/0) is up: new adjacency

.Apr 24 13:13:18 EDT: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 254: Neighbor x.x.x.8 (Ethernet0/0) is up: new adjacency

Apr 24 13:13:38 EDT: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 254: Neighbor x.x.x.8 (Ethernet0/0) is down: holding time expired

.Apr 24 13:13:42 EDT: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 254: Neighbor x.x.x.9 (Ethernet0/0) is down: holding time expired

.Apr 24 13:13:50 EDT: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 254: Neighbor x.x.x.9 (Ethernet0/0) is up: new adjacency

.Apr 24 13:13:52 EDT: %DUAL-5-NBRCHANGE: IP-EIGRP(0) 254: Neighbor x.x.x.8 (Ethernet0/0) is up: new adjacency

[pkhatri]

Navneet,

Are the IP addresses of both router interfaces in the same subnet and configured as the primary IP address ?

It could be possible that one side's primary IP is from the same subnet as the other side's secondary IP , causing this issue...

Hope that helps - pls do remember to rate posts that help.

Paresh

[pkhatri]

Navneet,

Are the IP addresses of both router interfaces in the same subnet and configured as the primary IP address ?

It could be possible that one side's primary IP is from the same subnet as the other side's secondary IP , causing this issue...

Hope that helps - pls do remember to rate posts that help.

Paresh

[sundar.palaniappan]

Navneet,

Appears the EIGRP hold time may be misconfigured.

Do a 'show ip eigrp nei' on all 3 routers. Verify the hold time is the same on all of them.

Pls. rate all helpful posts.

Sundar

[pkhatri]

Sundar,

I'm not sure that a mismatched holdtime would cause this. There is no requirement in EIGRP for the hello/hold times to match. Each router will independently anticipate it's neighbors hellos based on the holdtime advertised by that neighbor (not the holdtime configured locally).

Paresh

[sundar.palaniappan]

Paresh,

Let's say you have two routers, R1 & R2, connected via ethernet.

R1 is configured with default hello/hold timer 5/15 seconds respectively.

R2 is configured with a eigrp hello time of 60 seconds. R2 will be sending a hello packet every 60 seconds. However, after 15 seconds (3 hellos) R1 will consider the neighbor dead and clear R2 from it's EIGRP neighbor table.

Hope that makes sense to you.

Naveen,

Can you check the configuration of ethernet interface on all 3 devices and remove any EIGRP hello/hold interval commands, if there is any.

Also, try running continuous pings between the devices and make sure you aren't losing any packets on your LAN there.

Pls. rate all helpful posts.

Sundar

[navneet_78]

Hi Sundar,

There are no Eigrp hello/hold timers specified in the configs. All are set to default.

[navneet_78]

These are the Holdtimes configured.

xxxxxxx>sh ip eigrp neig

IP-EIGRP neighbors for process 254

H Address Interface Hold Uptime SRTT RTO Q Seq Type

(sec) (ms) Cnt Num

3 x.x.x.4 Fa0/0 12 05:26:25 24 200 0 438586

1 x.x.x.9 Fa0/0 12 05:26:27 1 200 0 38612

[sundar.palaniappan]

Naveen,

Your output above shows the neighbors have been up for over 5 mins. Are you still having the problem?

Sundar

[navneet_78]

Yes the IPs are on the same subnet on all the devices. And there are no secondary addresses used.

[pkhatri]

Hi Sundar,

I'm afraid your explanation of how the hello/hold timers work in EIGRP is not correct.

Looking at the example you quoted:

R1 is configured with default hello/hold timer 5/15 seconds respectively.

R2 is configured with a eigrp hello/hold time of 60/180 seconds.

This is what will happen,

- R2 will be sending a hello packet every 60 seconds. The EIGRP hello packet contains a holdtime field, which R2 will populate as 180 seconds. Now, when R1 gets the Hello it will see that R2 wants it to use a HoldTime of 180 seconds. Therefore, this is what R1 will use in order to determine whether R2 is still alive or not. The fact that R1 is using the default holdtime of 15s is irrelevant - it will use this holdtime in its own Hello packets to R2

- the opposite happens with R1. R2 will use a holdtime of 15 seconds to determine R1's liveness.

Try this out and you will see that the above is the correct behaviour. Incidentally, this is also exactly how it works in IS-IS.

Paresh

[stephtchoko]

Good day Naveed

please provide us show run and show interface of 3 router.

Regards.

[sundar.palaniappan]

Hi Paresh,

Good try!

Unfortunately, my scenario above doesn't say the hold time is set to 180 seconds on R2. It only states the hello timer is set to 60 seconds.

Changing the hello-interval doesn't automatically change the hold-timer. User error of this kind could cause EIGRP neighbor to flap.

Try this out in your lab and it should clear things up.

Sundar