So, I have an IWAN deployment (DMVPN) with 47 sites in the deployment. Most sites have 1 WAN connection and 1 Internet connection that are used for Tunnel 10 and Tunnel 11 respectively. The configuration is pushed out from APIC-EM, so every device is exactly the same configuration wise. However, this is the second time in as many days that 40% of the network started having their EIGRP peer flap. Hello packets go out, but no routing. The fix, both times, has been to reboot the spoke router and everything comes up working.
It is NOT an MTU setting, each tunnel interface has the MTU set to 1400. (ip mtu 1400)
It is NOT an eigrp timing issue, (hello-interval 20)(hold-time 60)
It is NOT an eigrp authentication issue, ( authentication mode md5, authentication key-chain WAN-KEY)
All of these would continue to have issues after a reboot.
So, what can cause EIGRP to arbitrarily start flapping on a DMVPN tunnel interface that has been up and running fine with no issue? The issue has come up on ASR routers, 2911, 2921, 2951, 4431, and 4351 routers.
This most recent outage seemed to happen as I was adding routes to EIGRP on some spoke routers. But this should be a non-issue because I am on DMVPN v3 so each site only gets summary routes and NHRP routes in the routing table. The first time it happened, I was not doing anything network related, seemed completely random. When it does happen, 60% of the sites stay up with no issue. I was also fixing random WAAS issues, updating OS and fixing WCCP when this happened the second time.
