I have been given the request to present when a node is up. Here's what I did but I'm not seeing the notification come through to the Open Key Incidents.
1. I went to SNMP Trap configuration and copied the OID for SNMPLinkUp - .188.8.131.52.184.108.40.206.5.4
2. I then created an event in the Mgmnt Event Config named "NodeUpTEST" with the OID of the SNMPLinkUp and enabled it.
Do I need to disable something in the Pairwise Config in order to defeating the purpose of this notification? I already disabled the OvNodeUpPair in the Pairwise config so the NodeDown doesn't disappear from the Open Key Incidents, however, the node is still being removed due to "NodeUp" and I'm not seeing my NodeUp event....
I forgot to say I couldn't find the NodeUp OID, which be easier and or the "right" way to achieve this event... So if anyone has the OID for the NodeUp event, that would even help. That is why I'm trigging my custom NodeUpTEST off of the SNMPLinkUp OID...
This is utterly unacceptable by HP. I can't believe they have done this, surely there must be a fix for this?
The fix posted above shows that HP allow you to get "round" this issue, but it's not a fix at all.
They are suggesting you configure an automatic event when a NodeDown event is received, and when that incident is resolved it triggers a custom event saying "node up" for example (or whatever you want it to say).
But the problem is, I don't want the NodeDown to be auto-acknowledged in this way. Why would I?
Fair enough if I have that option. But if I have a 24x7 manned Operations Bridge / NOC - and they need to callout on a 24x7 basis a NodeDown event, then if the system generates a NodeUp the alert will CLEAR itself.
But what if the original NodeDown incident has NOT been actioned and investigated as yet?
This is a ridiculous state of affairs, HP are forcing how they think companies should deal with incidents and this is WRONG. The tool should allow the user/company to configure the alerts as they decide.
If anyone has any more information on this then please share it. But I can guarantee many of my clients will literally be FURIOUS with HP if they force this type of behaviour on clients.
I have managed to create a NodeUp event - what I didn't realise is you have to create this as a new custom trap, not an internal NNM event as you used to be able to do in older NNM versions.
So I can now receive a NodeDown event, and when the device is back online, the NodeDown gets closed and the change of state to Closed triggers the new/custome NodeUp event.
So that works ok, albeit a messy configuration (especially when this functionality was already present in previous versions), but there is 1 fatal flaw...
If you power off a device, you get the NodeDown. Now, if you manually Close the incident because you have addressed the issue, and know why the node went down (in a real situation) the problem is now that the NodeUp event will trigger saying something like "Node is now UP" - which is 100% FALSE. The device is still powered off.
So this fix works if you don't CLOSE the original incident, but if you do it will cause confusion.
I have requested HP raise an ER to reinstate the NodeUp event, and also to allow clients to configure the behaviour of this automated correlation, so we can switch it off if required.
An awful set of updates from HP, we as customers should never have been put in this situation in the first place.