Running ubuntu 10.04LTS and since this is running behind a loadbalacer on a single public IP, I wish to have a few servers run on other ports, ex.
server1 IP:161
server2 IP:162

and so on. I am running zenoss remotely, and I tried in the loadbalancer to move 162 to local 161 but the query from zenoss wants the 162 port (naturally), so I need to have snmpd listen on 162. Reading the doc's, the examples say use the -x ip : port, which I did. In ubuntu under /etc/default/snmpd I have the line to read;

SNMPDOPTS='-Lsd -Lf /dev/null -u snmp -I -smux -p /var/run/snmpd.pid -x 192.168.50.102:162'

A restart, PS shows;

snmp 7214 0.3 0.1 50156 5268 ? S 10:16 0:01 /usr/sbin/snmpd -Lsd -Lf /dev/null -u snmp -I -smux -p /var/run/snmpd.pid -x 192.168.50.102:162

These servers have 2 IP's, a private and loadbalanced, the .50 is on the loadbalanced, yet from a remote machine I can snmpwalk -v1 192.168.2.50 <enter> and get my info (:162 times out), so it's not using that config.

Any help is appreciated.

In default snmptrapd listen on 162, let us clear do you want to send traps to Zenoss server? and check the output of netstat -pnl for which application bind with port 162

Not sure on the terminology so reading in parallel with snmptraps, etc. but the zenoss server is in amazon's EC2 cloud with an elastic IP. I did make progress, but it's not working yet. I have the public IP in the load balancer, port 161 -> server 1 port 161. From the ec2 machine I can issue;
snmpwalk -v1 -c public public:161 and it works.

Now in Ubuntu, you need to remove the -x so the default now shows;
SNMPDOPTS='-Lsd -Lf /dev/null -u snmp -I -smux -p /var/run/snmpd.pid 0.0.0.0:162'

From a local machine, if I issue
snmpwalk -v1 -c public public:161, it now fails, but
snmpwalk -v1 -c public public:162 works.


netstat -pnl confirms as well;
udp 0 0 0.0.0.0:162 0.0.0.0:* 11721/snmpd

But, I did add a 2nd virtual service on the public IP port 162 to that server port 162 and I get;
Timeout: No Response from public:162

I have ruled out firewall, so still debugging, so it does look like it's listening on 162 now. But as you said with snmptraps, could that be the problem? Since the single 161->161 worked, I would figure that's all you need to do.

Thanks for the response... .

Interesting, I moved the port from 162 to 163, updated the loadbalancer forwarder and things work! Thanks for the quick response again.