6 Replies Latest reply on Feb 23, 2011 2:11 PM by Questionario

    Cisco Router Discovery


      Hi All,

      I'm having a problem in that most of my Cisco Routers are in the database a few times using different interfaces on the devices. Generally, I'd want to monitor on the loopback address figuring if I can get there, at least the router's up. What I see is that the same node will be in the database, once discoverd via loopback, again via the MPLS ip address assigned to one of the links, maybe again wit hthe other MPLS ip address used in the secondary MPLS network and again using an IP address assigned for the FA interface used for lan attachment. When I build the monitoring map I try to use only the loopback device but I'm running into some issues where when one MPLS goes down, I generate an alert and the map of couse stays green since the loopback is still reachable over the seciondary MPLS connection. This I can live with, I guess what I'm asking is if network discovery is working correctly since I am able to add these supposed duplicate nodes. The loopback node has all the same interfaces as any other representation of the discovered router, just that when I open the tab for say Cisco 7201 I see NodeRouter02 listed multiple times in the produced list. Is this expected behavior? Should I just delete the "extra" representations of the router? Anyway to not discover them again?

        • Re: Cisco Router Discovery

          hi rmmagow, this is expected behaviour, you can limit your scan to a specific ip-range (the ones of your loopbacks), it is safe to just delete the extra nodes but some people want to monitor one node multiple times for exactly the reasons you mentioned, to see if the MPLS link is down but also to see if the node in general is still reachable. it is an open feature request to be able to assign multiple IPs to a single node but not enough people have opened up a request for that, if you want that you need to make a feature request :)

          also if you have a discovery that you keep running, you can add those nodes with that IP to an ignore list, they will still be discovered with a "normal" discovery but won't be shown as they've been marked to ignore.

            • Re: Cisco Router Discovery

              Where can we make this feature request?  Logically it makes sense to have a single node identified by the loopback and then be able to have additional interfaces listed under that single node.  Currently the only way that I am able to get topology information to work with network discovery is to have each router listed once for each interface I am interested in having topology information on.  From what I can tell, this wastes element allocation.


              Currently the way that discovery works is kind of fubar.  I have been unable to successfully integrate this feature it into our infrastructure.


              Is there a best practice that should be used to make discovery effective?  Whenever I try to create a discovery for a router that has two FA interfaces in two networks, I would like to be able to set it to 1 hop, which should discover the switches in each of these two networks and stop there.  Currently my discoveries are jumping multiple hops to different sites and making the schedule function useless.  If I target one interface with 1 hop selected I would expect for the subnet that interface is on to be scanned and the switch it is connected to to be found and that would be the end.  Is my understanding incorrect?

            • Re: Cisco Router Discovery

              If you want to know when the interface goes down for each node, create an alert for the interface status (not as good a ping through the interface, but it is something) and you can add that interface(s) to the map(s).