We use Solarwinds NPM to monitor around 200 servers. Around 20 of the Windows servers (2008R2 & 2012) have this particular issue. The SNMP service will stop responding to queries after certain period of time. The only temp fix is to reboot the server. But the issue will come back after a while. I checked firewall rules and remove the AV client from those servers, non of those are the causes. Recently I did some further testing and notice the following strange behaviour of the SNMP service. The test tool I use is Paessler SNMP Tester 5.2.
So once the SNMP stop responding to SNMP queries, I restart the SNMP service (I have to kill the SNMP.exe process in order to stop it). SNMP Tests done locally will still fail straight after the restart. But if I wait for overnight and test next morning, the SNMP service will respond normally again. SNMP tests locally or remotely will return positive results no matter how many times I poll it. So it almost feels like the service need time to recover itself. Please note, the server is definitely not rebooted overnight. I then try to add the node into Solarwinds again. During the adding process, the initial test will come back positive. But as soon as I click next Solarwinds Orion tries to poll available resources to monitor, the SNMP service will stop responding immediately! SNMP tests done locally on the target server will fail again!
I have repeated above procedure for 2 days and are experiencing the exact same symptom. It appears the SNMP service does not like something generated from Solarwinds monitoring. Does anyone encounter similar issue before?
