EIGRP / OSPF dampening

Eric Osborne eosborne at cisco.com
Wed Dec 18 03:36:03 UTC 2002


On Tue, Dec 17, 2002 at 06:02:48PM -0700, Donn Lasher wrote:
> 
> 
> I'm looking for a way to way to either dampen, penalize, or otherwise slow 
> down a flapping interface, when the router has no "link state" information 
> about the Layer 2 connection, which is itself causing the flap... Example...

http://www.cisco.com/en/US/products/sw/iosswrel/ps1829/products_feature_guide09186a00800ad25b.html

Like BGP dampening, but more generic.  Does that do what you want?



eric

> 
> Dec 17 08:23:50 xxx.xxx.xxx.xxx 2106: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is down: peer restarted
> Dec 17 08:23:52 xxx.xxx.xxx.xxx 2107: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is up: new adjacency
> Dec 17 08:24:09 xxx.xxx.xxx.xxx 2108: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is down: peer restarted
> Dec 17 08:24:11 xxx.xxx.xxx.xxx 2109: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is up: new adjacency
> Dec 17 08:24:42 xxx.xxx.xxx.xxx 2110: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is down: peer restarted
> Dec 17 08:24:43 xxx.xxx.xxx.xxx 2111: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is up: new adjacency
> Dec 17 08:25:18 xxx.xxx.xxx.xxx 2112: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is down: peer restarted
> Dec 17 08:25:20 xxx.xxx.xxx.xxx 2113: 15w5d: %DUAL-5-NBRCHANGE: IP-EIGRP 
> 10: Neighbor xxx.xxx.xxx.xxx (FastEthernet2/0.3) is up: new adjacency
> 
> This type of flapping yields an effectively down interface, yet, appears 
> "up" so data gets blackholed while dead timers expire, etc.
> 
> I'm very familar with doing this via BGP, but haven't found any good way 
> yet to do this with either EIGRP or OSPF.
> 
> Any ideas? Feel free to either reply to the list, or privately.
> 
> Thanks in advance..



More information about the NANOG mailing list