4

I am looking for a way to monitor OSPF links using SNMP. I have attempted to try to poll the Neighbor State of a specific peer, looking for a value of Full, and going into alarm once the value is no longer Full, but when the trunk goes down, the peer is completely removed from the Neighbor Table. I have also attempted to poll the OSPF Interface state of a specified OSPF Interface. Alerting once the interface goes into a down state, but some of the interfaces go from BDR state to DR state instead of a down state. Does anyone have any ideas on a variable/value that I can monitor?

n0sw1tchp0rt
  • 97
  • 3
  • 8

1 Answers1

2

Your best bet is to configure your gear to send traps for interface and neighbor state changes, and monitor/handle those traps with your tools.

James Cape
  • 1,067
  • 8
  • 16
  • My gear is already configured to send traps for interface and neighbor state changes. I apologize that I was not clearer before. I receive alarms from my gear just fine. What I am attempting to do, is set up a Active Map in CastleRock's SNMPc. A map showing my nodes, and active links that are green when things are fine, but turn red when trunks go down. This requires me to configure this in a TRUE/FALSE fashion. TRUE = Green False = Red. I have managed to get this working for the physical interface, by turning red when the link goes down, but (cont....) – n0sw1tchp0rt Nov 17 '11 at 18:43
  • ...would like to do the same for the Layer 3 OSPF portion, so that I can verify whether I am having a local physical cable issue (by the physical interface link turning red), or a remote provider issue (by the Layer 3 link going red), with just a single glance at the map. This requires me to poll one of the layer 3 tables in the node for a variable that does not disappear when the link goes down. As of now, I get an error in SNMPc because the OSPF peer is no longer in the Neighbor Table when the link goes down, therefore it is unable to poll it. Hope this makes more sense. – n0sw1tchp0rt Nov 17 '11 at 18:49
  • So your question really is: "How do I configure Castle Rock SNMPc to monitor OSPF links and handle state-change traps? When a link goes dead it's removed from the neighbor table, which causes an error." – James Cape Nov 20 '11 at 23:30
  • I've done exactly this with Zenoss---if you have a copy of the pristine LSA DB in your monitoring system, the traps contain enough information to figure out what went wrong---but I've never looked at anything surrounding SNMPc before today. – James Cape Nov 20 '11 at 23:34