I’m having issues with a new install of Nagios 2, using and I’m also using the Opsview front-end (opsview.org/). I’m going to try to be as complete as possible here.
I created a new Debian virtual machine for this project, installed Nagios using apt-get and then I followed these steps on Opsview’s website:
Everything installed fine and without errors. Now I wanted monitor a domain controller as a test case. I turned on the default SNMP service on a Windows Server 2003 domain controller, and I set up services to allow my SNMP community name as an accepted community, as well as allowing the service to send SNMP packets to the IP address that represents my Debian virtual box.
Then I set up this domain controller as a host to be checked, and selected various services that I wanted to monitor. Of the 22 service checks I selected, only 6 worked. The other 14 came up as critical, and all of them had some form of “connection refused by host” as their error message. I started digging through the configuration and I noticed that Opsview has 4 different ways of checking a service: Active Plugin, SNMP Polling, Passive, and SNMP trap. All of the services I was checking were set to use Active Plugin.
I had originally intended just to use SNMP to gather the data (and not a plugin) so I set one of them to SNMP Polling. When I do this, it wants me to do an SNMP walk and then manually select the OID from a list for it to check. The SNMP walk does work, and it generates a huge list of OIDs and the data it has pulled from each OID. The data it has pulled from the domain controller is correct, so I believe SNMP is working correctly.
However when I selected SNMP Polling instead of Active Plugin, I expected it to pull the correct OID that contained the data (CPU usage, in this case), but it wants me to pick an OID from this huge list it generated.
What do I do? Do I have to install plugins on these machines to gather the data I want? I’d perfer to use regular SNMP if I can.
Thank you in advance for your help,