Unsolved
This post is more than 5 years old
3 Posts
0
75285
Health Status = Warning/Critical, but no alerts
I'm betting that there's something we didn't configure correctly or have missed, because this just doesn't seem right. We are using SNMP, and, as far as I can tell, we configured it correctly and do indeed show a listing of servers, with details. We set up email alerts to go out whenever a warning or critical alert is registered. Those alerts work when an alert is generated, such as a server going offline.
So here's the problem: We pull out a power supply from a server as a test. We force a status refresh, and the health changes to Critical. But no Alert event. When we look at the details of the server in question, the Warning icon shows next to the link for Dell OpenManage for that server. So why is this not creating a real alert, and how can we get it to trigger as an alert?
Thanks! Paul
DELL-Rob C
2.8K Posts
0
March 14th, 2012 14:00
Hi Paul.
Sounds like you are almost there...
Can you confirm that the SNMP trap destination is set up correctly on the managed node (i.e., target server)?
If you go to the OME Tutorial | First Time Setup | SNMP Trap Settings section, you will see a screenshot. This is the setting you make on the server you are trying to manage (not the OME server). And it is important to note that the SNMP community string is case sensitive.
Be sure if you are using server name, that you can ping the OME server by name from the managed node....otherwise use the IP of the OME box.
Let's start there.
Rob
DELL-Abhijit P
Community Manager
Community Manager
•
711 Posts
0
March 14th, 2012 14:00
Hi Paul,
Have you configured your servers to send alerts to OME. You will have to set the trap destination to OME. "First Time Setup" in the Tutorials page has the details.
Regards
Abhijit
pturchin
3 Posts
0
March 14th, 2012 19:00
Yes, the managed nodes are configured correctly and are sending traps to the OME server. I ran the troubleshooting tool SNMP listener, then pulled out the power on the target and saw an SNMP event hit the OME server. I don't know if it matters, but the Windows SNMP Listener tool failed to pick anything up, whereas the Net SNMP Listener detected the traps.
DELL-Rob C
2.8K Posts
0
March 15th, 2012 07:00
Ok, so that sounds like the servers are correctly configured to send traps to OME (as you say).
But it also sounds like the Windows SNMP port may be blocked. Do you have any other software running on that server that could be using the port (162 I think)?
Can you run the port check on the 2nd tab of the TS tool to see if anything is using that port?
pturchin
3 Posts
0
March 15th, 2012 13:00
Nice catch Rob. Looks like we have Spiceworks running on the same server, also wanting to utilize SNMP (even though we never configured anything for it). I installed OME on a blank server to test, and sure enough, pulling out a power supply on a target resulted in an actual alert. So I went back to the original OME server, disabled Spiceworks, restarted snmptrap service, and ran the Windows SNMP Listener. It now did show the traps; however, it still did not create alerts. So it's possible I may just have to install OME on a dedicated server.
But now back to the test OME server where I saw alerts. My target server was initially detected with a warning (out of date RAID drivers), but that did not trigger an alert. Are out of date drivers not the kind of thing that would trigger an alert? Or do alerts not get triggered on pre-existing conditions? In other words, if everything was normal and then a driver became out of date, would I get a warning alert then?
DELL-Rob C
2.8K Posts
0
March 15th, 2012 14:00
Hey, that's great. Looks like you have that sorted.
For the RAID driver, no, we won't throw an alert on that. The behavior is such that you need to go to Manage | System Updates and then click on the red pie slice in the compliance report.
The event console (Manage | Alerts) will be the place for SNMP hardware alerts (aka traps).
Hope that helps and keep me posted on how things progress with OME.
Regards,
Rob