SIM shows servers in failed but nothing in events or notification

[Paul S._2]

We have had SIM setup for awhile and working no problem. I have an event job to email us if any server is major or critical in status.

Just yesterday a server was rebooted, and in SIM it put this into the event log and it also emailed us notification...all as usual, all as expected.

Today, I checked the SIM website and noticed in the server list, it showed this same server as critical. I checked the event log, and there was nothing new. Checking the server, it turns out the power supply failed.

Does anybody know whats wrong with the SIM server? To recap, it shows the server as failed. But it didnt enter anything in the event log, on the top right 'uncleared event status' it shows zero for critical/major, and of course no email notification.

As it turns out we had 2 servers with failed power supply due to the UPS it was connected to. This also caused power supply failure on 3 HP SAN storage units, yet SIM didnt put event or email us?

[Peter Stringer]

Have you the mibs for the san storage, not sure if they come by default with an installation of IM. If not get the mibs and compile and register them.

Also sometimes snmp traps are listed as informational in severity terms. I belive that informational traps do not get emailed. Its best to look at all traps and determine the ones you want emailed ( long job I know) and set the severity to either critical or major.

I had the same problem, and this worked for me.

[Paul S._2]

Thanks for the response peter.

While I understand that would explain the drive shelves, it still wouldnt address the server itself - one of the power supplies was unplugged due to a failed UPS.

It did show a 'major' status in SIM webpage in the listing of all servers. But, the event log in SIM showed nothing.

This really tells me its a problem with the SIM web software itself.

If it was a notification problem from the server to the SIM webpage, then I simply wouldnt have seen it on the SIM webpage.

[Rob Buxton]

Just to confirm, these servers have two power supplies connected to different sources?

And a power supply failed?

In your SNMP Settings, do you have the Trap Tab do you have the Community name and Trap destination set? If this is not se then the Agents cannot send a trap to the Server and only events discovered by Polling (e.g Server unreachable) will be picked up.

Is there anything in the MS Event Log or the HP Integrated Management Log?

On HPSIM, click on the Server name and you should then be able to click the Events Tab that lists all the events for that Server.

[Paul S._2]

Rob,

yes the servers have two power supplies, each connected to their own UPS and power source. One UPS failed.

In SNMP settings, we have our community name set but on none of our servers do we use a trap destination.

Are you stating we need a trap destination? I could have sworn we have been using SIM w/o a trap destination just fine, but let me know if I remember wrong since this is usually setup then forgot about...

[Dave Claypole]

I have exactly the same problem with notifications. Everything works fine if the server is down, email and event logs alerts.

Anything else power supply failure, disk failure no alerts but Insight Manager does know there is a problem.

Can anyone confirm that i need to configure the trap destination, and why. Like the previous poster i only have the snmp community name set.

Dave

[Ryan Rager_1]

Yes I too am having the same issue. The event shows up in SIM server views and on the HP Server Homepage but doesn't register as an event for event notification. If Sim gets this event by a hardware poll then why doesn't it register the event? I don't see the need for traps to be enabled in this sitution.

[Rob Buxton]

Paul,
If an IM Agent detects a problem it needs to deliver a trap, that's where the Trap destination comes in. So, yes, you need to set that up.
The destination will be the HPSIM Server, the community name can be the Read Only name you already have set up.
HPSIM gets tha trap and generates the relevent event.

[Rob Buxton]

As a note, on the HP Management Agents GUI there's an SNMP Trap Send Test button, which should generate a Major Event on your HPSIM Server, ideal for all kinds of testing.

[Paul S._2]

Thanks Rob.

I have updated > 80 servers with the trap destination.

So far we have received only stupid errors - why HP thinks opening the tape backup drive is a critical error is beyond me LOL (i have since reduced this default setting from critical to informational).

Anyways, thanks for your help and if I still have problems I will post again to this thread. Otherwise, thanks!

[Rob Buxton]

Ryan,
When you go to the Agent Homepage you're not on SIM. Polling is just a check over the different protocols for an "are you okay" it does not interrogate the agents for issues.
If it did you'd get an alert at each polling for a problem.
The fact that the Server Homepage is reflecting an error indicates only that the local agents have picked up a condition. Without the Trap Destination set, this error would not have been sent to HPSIM. Polling will not pick up this condition.
Thresholds are another example, the trap is generated when the threshold is crossed. It's not picked up by polling. So, you only get the one event generated.