I am trying to set up SNMP v3 network monitoring for Ubuntu either 18.04 or 20.04. I have loaded all the SNMP modules and configured a user and set up the firewall to allow 161 and 162. 'sudo ufw status verbose' shows ALLOW IN for ports 161 & 162. From this machine to a CentOS machine I can do an SNMP walk just fine. But from the CentOS machine to the Ubuntu machine it times out. I have used the exact same credentials on both machines. I even disabled AppArmor but no go.

Is there something I am missing? I am trying to use OpManager to monitor this and the tool gives errors as if I never configured SNMP.

...AR

Sorry to start with something very basic...but you say you loaded the SNMP modules, and configured SNMPD. But did you start the service?
Test to make sure you can at least query the box locally with snmpwalk. Also, did you open the ports for UDP or TCP (SNMP uses UDP)?

A little tutorial here: https://techexpert.tips/ubuntu/insta...nmpv3-service/
...which covers SNMP V3, keys, etc., with a sample config.

Yes the daemon is running and I can snmpwalk localhost just fine.

Ok...and UDP is open, not TCP, right? Typically, if you can walk the localhost but not over the network, that obviously means a network issue. Are those ports open on the other systems for return traffic? Anything in between the two??

All firewalls are configured for 161 and 162 UDP and the other CentOS machine can be monitored just fine from my NMS. Both machines are in the same subnet on the same VLAN.

I FINALLY figured this out. Ubuntu "out of the box" only allows local SNMP queries, it's restricted to 127.0.0.1. You need to modify the snmpd.conf at /etc/snmp and there will be a value called agentAddress. It will have a couple lines, one for local access and one for all access. Comment out the local access line and uncomment the line allowing all access and restart the snmpd daemon.

...ar