11-20-2010 11:37 AM - edited 03-06-2019 02:08 PM
We have a primary 10Meg fiber connection with a backup ADSL connection for redundancy
Standalone sites are smaller. These locations do not usually have as much network traffic and therefore do not need the capacity that the fiber connection provides. As such, these sites use ADSL as their primary connection and have a dedicated internet connection for when they fail over.
Our fiber network 192.168.10.0/24 connects each of the full centres to head office and all routes are identified via EIGRP. When one or more of these sites goes down, EIGRP identifies the ADSL network as the best route to the store as the fiber route disappears from EIGRP.
Our ADSL network is 192.168.11.0/24. This network designed required us to have a default gateway of 192.168.11.100 ( which is a third party system and don't have access) that routes all traffic to the appropriate destination via EIGRP. This network also depends on EIGRP for getting traffic from head office to each site.
The problem that we are facing is that EIGRP is changing the route on a fairly regular basis, switching between sending the traffic directly to head office via the ADSL network and sending the traffic to another location via the ADSL network and then to head office via that locations fiber network. I believe that because this traffic is taking different routes to head office than they are taking back to the location. Anybody has any idea how can I can configure my remote sites to make sure it doesn't take different path to reach head office.
Thanks.
11-20-2010 07:11 PM
Do you have the option enabled to log EIGRP neighbor changes? If not, I would suggest that you do enable this option.
After the option is enabled I would suggest that you look on the routers both for the head end and for the remote and see whether there are any log messages about neighbor changes. If neighbor relationship is going down and then back up this would explain the difference in how traffic is being routed.
HTH
Rick
11-25-2010 11:17 PM
Please check the K value and mtu size on the both router.
11-25-2010 11:29 PM
I did enable the log neighbor-changes and here are the output:
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.31 (FastEthernet0/1) is down: holding time expired
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.81 (FastEthernet0/1) is down: holding time expired
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.30 (FastEthernet0/1) is down: holding time expired
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.100 (FastEthernet0/1) is up: new adjacency
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.33 (FastEthernet0/1) is up: new adjacency
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.10 (FastEthernet0/1) is down: holding time expired
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.33 (FastEthernet0/1) is down: holding time expired
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.10 (FastEthernet0/1) is up: new adjacency
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.34 (FastEthernet0/1) is up: new adjacency
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.31 (FastEthernet0/1) is up: new adjacency
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.10 (FastEthernet0/1) is down: holding time expired
%DUAL-5-NBRCHANGE: IP-EIGRP(0) 1: Neighbor 192.168.1.100 (FastEthernet0/1) is down: holding time expired
If I increase the holddown timer, that would help?
11-26-2010 01:50 AM
Hi,
Increasing the hold down timer should indeed solve the problem but only maybe temporarily so after that I would try to know exactly the cause of this adjacency flapping.
To see your hello and holdtime if not default: sh run interface f0/1, default on ethernet is 5s and 15s respectively
You should also enable timestamping of your logs: service timestamps log datetime and service timestamps debug datetime
Regards.
11-26-2010 07:48 AM
I am confused. The original post talked about running EIGRP on networks 192.168.10.0 and on 192.168.11.0 The output in this post is showing EIGRP neighbors up and down in 192.168.1.0. So what is the relationship between 192.168.1.0 and the other 2 networks?
The EIGRP output here shows neighbors failing because of holding time expired. Holding time expired indicates that the router stopped receiving Hello messages from the neighbor. Since there are no time stamps in the log output we can not tell how frequently this is happening. But not receiving Hello messages could very well be caused by packets being dropped on the link connected to the interface. What is FastEth0/1 connected to?
One way to test this would be to run a ping with lots of packets to one of the neighbors in the output and to look for how many time there is no response to the ping.
If you increase the hold timer it would slow down the incidence of flaps. That may help with the symptoms but it does not address the underlying problem.
HTH
Rick
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: