5 Replies Latest reply on Nov 23, 2009 1:31 PM by macnugetz

    Alerting - NPM reports Node "Not Responding" even though its "Node Down" thus no alert triggered ?


      Afternoon All

      I'm running NPM ver 9.5 SR4

      To my joy today, a 3750 switch stack suffered an unscheduled power outage during which NPM events showed that the node was not responding but never actually listed a "Node Down" entry which it has been doing fine for all other devices to date.


      Under the node details during the outage, it stated that the node was down and but also that several interfaces were in an unknown state.

      Oddly enought, in the alerts section the active alert showing was the correct "node down" advanced alert, (the only alert we have configured currenly) however, it should (and has been) triggering an email, a entry in the Windows OS eventlog and a script which pumps an alert out to our remedy system for our offshore team to respond to.  None of these materialised

      It seems to me that although in the device details page it showed the status of the node as down, the actual "xxxxx is Down" message you'd expect to see in events, never appeared, just the "xxxxx is not responding" and thus no actual node down "event" occured.  The monitored ports are 2 physical gig ports and their associated port-channels (see below, I took the snapshot once the switch was back up, hence the green status).  During the outage however, these ports showed up against the blue question mark "unknown" state.  Could this have prevented the node being deemed fully down ?  or something to that effect ?

       Basically I need to be sure that if the node is down that the alerts trigger correctly.  Any thoughts welcome and appreciated.