10-03-2012 03:13 AM
From experience, can anyone tell me the best method of integration between NNMi and OM, web services or OM agent?
From my experience, the web services seems to have some kind of limitations. We notice sometimes that incidents are not being forwarded so we disable and re-enable the integration. On a couple of occasions we had to restart NNMi.
10-04-2012 06:49 AM
I use the OM agent (opcmsg) to forward the messages to OM. The opcmsg is invoked by the automatic actions in the incidents. This gives me a lot of control because I can ran any script and execute a lot of parsing or information gathering to enrich the information passed to OM.
10-05-2012 04:33 PM
templates are used by the agent running on NNM to match conditions,
like a snmp link-down trap. In OM 9 I think they call templates - policys or
something like that.
10-05-2012 10:52 PM
I have not done the integration via the agent before. I've always did it via the web services.
FYI, my experience lies in HP Network Management Center and I have a colleague who works on OM. To that effect, I only deal with the web form on NNMi's side.
It seems to me that most people opt for the agent integration.
10-08-2012 07:33 AM
Its not difficult (with caveats of course) since theres a utility in NNMI called nnmopcexport.ovpl which will build you a template that you can load into OMU composer.
The main caveat is that support for SNMPv1 mibs is broken. Using v2 mibs theres an offset value for things like Hostname, hostIP address that are shifted based on the number of varbinds a snmp trap might have... its currently broken and you have to manually calculate that value offset.
Andy Kemp, CISSP
11-18-2012 12:58 AM
We've switched to the integration via the agent but we're facing the issue where an open incident in NNMi gets closed, this does not reflect in OM (the event is not acknowledged).
With the web integration, when incidents are closed in NNMi, this gets reflected in OM.
Help would be much appreciated.
4 weeks ago
4 weeks ago - last edited 4 weeks ago
We have two methods for NNMi-HPOM integration:
1. The web services implementation forwards NNMi incidents to the HPOM Incident Web Service(IWS). You can implement IWS in easier way... If you have (relatively) few number of incidents, choose this method
2. The agent based integration is a special NorthBound integration. NNMi forward incidents as SNMPv2 trap. In this way, you can also forward 3rd party SNMP traps. This method is capable large number of incidents/traps to forward.
Note: The reason for the agent based integration being the preferred one is according to what I find:
• Higher throughput rate.
• Messages are buffered via the agent on the NNM server if the OMU server is down.
• Possible to configure multiple NNM servers to send incidents to multiple OM servers.
• Manipulation of events possible via SNMP policy.
Hope this full fill your requirement!!
IF this or any post resolves your issue, Plz be sure to mark it as an accepted solution.
To show your appreciation, click KUDOS !!!
4 weeks ago
If the NNMi incidents are not being closed on OMW it's usually because of a problem with the configuration of the policy on the OMW server.
The 'out of the box' policy should close incidents correctly, however if this policy has been edited you may need to check if the "EventLifecycleStateClosed" rule is correctly configured as it's the message correlation in that rule that will close the incident.
It's also worth checking if any unmatched events are been logged against the NNMi server in OMW. That's were my events usually end up when I've got the policy wrong!
3 weeks ago - last edited 3 weeks ago
I have checked it, and the configuration is fine...it says "If matched, do action(send it to ack browser.)".....
And i have also checked the unmatched events tab, there "are sent to the active Msg Browser" has been selected
now what could be the error or reason...
3 weeks ago
To check for unmatched events find the NNMi server in the node tree and see if there are any alerts for it.
"If matched, do action(send it to ack browser.)" - That is what I would expect to see in the policy for Closed events.
Have you looked at your acknowledged alerts to see if there are any EventLifecycleStateClosed events there?
If there are but the originating event hasn't been closed there is a problem with correlation.
If there are not any it probably means that for some reason they are not being caught by the policy and you should be able to find them as unmatched events raised against the server.