NNM alterting problems.

[Jfaraone]

I currently have NNM (running on Wink2 Server) set to monitor nodes / interfaces on my network.

I have configured a custom alert (copied from OV_IF_Down)
to monitor a selection of interfaces. It has an action to send off an email when it goes down.

(I am ultimately trying to monitor a selection of servers and switch / router interfaces, but I am not sure of any reason to use OV_Node_Down, as most of our servers only have 1 NIC, am I mistaken?)

I believe I have followed the instructions in the manual perfectly, but when I take a server offline, nothing is logged in the alarm browser and no events are triggered. The interface does go from green to red on the map, though.

This is the strange part. If I unmanage this node, then take the server offline, then manage the node, as soon as NNM polls, it sees the server interface as down, everything is logged and all triggers function properly.

Another oddity; I had previously been testing my alerts with a wireless access point; everything worked perfectly. I would unplug it from the network, NNM would trigger an alert, and sent me an email. Today I noticed that this access point was set to unmanaged, I set it to managed. Since then, I haven't been able to get it to trigger any alarms.

I'm banging my head against the desk on this one; any idea what I'm missing?

Thanks,

Jason Faraone

 

[Jfaraone]

I can't believe I forgot to mention this, but version 7.5.

 

[Ru55ell]

Jfaraone
I do the same thing and Its working.
Servers - Servers that have one interface you’ll want to manage the alerts using a costom Node down and node up. They only have one interface. Manage them and set the SNMP Configuration to poll them more often in the “Specific Nodes” tab.

Routers – I monitor routers using the Node Down and Node Up. Interface or ip interface drops generate Node Status Warnings. If the whole thing goes away a Node Down alarms. I manage them and set the SNMP Configuration appropriately.

Switches – I monitor and manage the switch as a whole using Node Down and Node Up (Custom of course). I monitor the activity of a limited number of switch ports


The default polling interval is 10 minutes. I suspect your not waiting long enough for the interface change to be noticed.

Also I have to pay close attention to the “selection name” of the object and the host file and the Events configuration. I have name resolution issues and have to have the IP address listed in the Event configuration. The Host file resolves the name and the name or IP can be the selection name.

I hope this helps, it may cause more head banging, sorry

 

[nbowden]

Hi,

There may also be issues caused by the 'OV_NodeIf_NodeNotConnector' correlation circuit. Have a look at this discussion.

http://forums1.itrc.hp.com/service/...33898&admit=-682735245+1152478453040+28353475

HTH


Nigel Bowden

http://www.bowden-software.com