So are you saying you do not want EIGRP to establish a connection immediately because your links have tendancy to flap? So if you truely desire it you can alter your hello/dead interval timers in your EIGRP process, but the problem is that if you delay your timers to an interval of 30 seconds you run the risk of some very horrible performance for convergence of failed links. However if this is predictable behavior you can modify your timers on an per interface basis. You just need to make sure that the adjoining neighbor has thier timers configured to match.
So go into the interface that flaps and apply a command similiar to:
R1(config-if)#ip hello-interval eigrp 10 30
R1(config-if)#ip hold-time eigrp 10 90
The first value (10) is the process ID for EIGRP the second value is the timer. The hello-interval command tells R1 to send messages every 30 seconds, where the hold-time command tells it's neighbors to use a hold-timer of 90 seconds.
I would still advise against such lengthy timers.
EIGRP Commands from Cisco:
Thank you for the answer.
I was thinking about EIGRP timers, but the configuration I have is a hub and spoke (DMVPN) and each spoke has two 2 diferent links to Hub routers, for each connection a separate EIGRP instance.
So, if primary connection goes down, I want the backup route to be installed and remain active in routing table for at least 30 seconds even the primary connection comes up.
I was thinking to track the line protocol of primary Tunnel interface, and when the interface goes down to have event manager shutdown the primary eigrp instance, and to activate only when the line protocol is UP for 30 seconds.
I think better solution would be using static routes with combination of IP SLA instead of EIGRP.
Please rate helpful posts