SIM Alerts not working properly

[Derek_31]

I was configuring SIM 4.2 to send out e-mail alerts based on various severity levels. To test the alert out, I pulled out a ethernet cable in one of the teams on a server.

In the system event log on the server I got an "Error Event 1287" which shows the team redundancy was reduced. At first I thought the agent was broke, since it took a good 2-3 minutes for the local System Home Page to even show the link was lost.

I switched back over to SIM 4.2, and the HW and MP status went to yellow for the server. However, in the upper right corner it still showed '0' minor alerts. I also did not get an e-mail alert that I had configured.

The alert that I configured was:

Severity is Minor
System type is Server
Send E-mail, HTML
No Time Filter

I also verified the SMTP host it showed was correct.

The 'last ran' status is also 'never'. So it appears the task was never fired off.

What's going on?

[Derek_31]

I did an 'agent settngs replication' task, which failed with a minor alert, and I did get an e-mail alert from SIM.

I put the ethernet cable back in, waited for the Agent and SIM to show all green, and then pulled the cable out yet again. As before, SIM did turn the HW and MP status amber, but did not increment the 'minor' error count in the upper right corner, or send out an e-mail alert.

It appears SIM isn't recognizing HW and MP minor alerts properly?

[Derek_31]

Ok I played around with it more, and discovered something interesting.

Previously the Agents were not configured to send SNMP alerts to the SIM server. The reguarly scheduled 5-minute server hardware polling task is what discovered the pulled ethernet cable.

However, when I configured SNMP traps to be sent to the SIM server and pulled the cable, SIM all of a sudden recognized the event, incremented the counter, and sent me the e-mail alert.

So, is it true the SNMP traps must be properly configured for SIM to recognize failures, even if the SIM HW server polling task can discover the fault and change the HW and MP status icons? It would seem this should be sufficient, and that SNMP trapping would only be another method to get status.

[Derek_31]

Ok, another issue. When SIM receives the SNMP trap about the down NIC port, the 'uncleared event status' increments, but the HW and MP icons do not change colors. It seems the icons only change colors when the 5-minute HW status polling job runs.

So, in the e-mail alert I get, it shows hardware status as 'normal', even though a minor event has occured and SIM will show it in yellow within 5 minutes.

Why doesn't SIM immediately update the server icon status based on information received from the SNMP trap?

[Rob Buxton]

Derek,
The simplest way to start testing is by using the HP MAnagement Agents GUI on a server other than the CMS Server.
This has a Trap Send Utility.
This should generate a Major Event.

If it doesn't, check you have the Trap destination and community name set up correctly.