Profiler does not require snmp be installed.
On the agent, check the core.xml (in the install dir, generally /opt/Profiler_Agent) and make sure the IP addresses are those of the Profiler server (in the <destinations> and <allow> tags).
If they are not, the agent will not self-register with the Profiler server, and will not allow communication with the Profiler server (if you manually add the agent).
If that does not work, I would suggest you contact support.
Couple of things going on here.
1. The agent should have registered automatically: Check the SNMP Trap Port in the Server section of the Server Setup page. If it has bound to a port other than 162, modify the /opt/Profiler_Agent/core.xml file to match the SNMP port and restart the agent
2. Traffic light unable to connect to agent: Traffic light should connect even if you add the agent resource manually. Click on the traffic light again, and then tail the Kernel.log file under /opt/Profiler_Agent. It should log a message that it refused a HTTP request from ip/server name. Make sure the core.xml in the same directory has the ip/server name as seen in the Kernel log message and again restart the agent.
Let us know if these steps make it work. If not, open a support ticket here
and just provide "Solarwinds Profiler" as the customer id, and post the case id here. one of us will get in touch.
Hi Guacamole and bmrad,
Thank you so much for your prompt response. I looked into core.xml and modified the ip to the ip address of my profiler server. I do not see any trap port being mentioned in the core.xml but only the http port, which is similar to the http port specified in the Profiler Server. Hence I didn't change that.
After changing the IP address and restarting the profiler agent. I receive a fatal log with the entries below:
FATAL ERROR ENCOUNTERED IN KERNEL ON Thu Jul 08 17:18:51 MYT 2010
at sun.reflect.NativeMethodAccessorsImpl.invoke0(Native Method)
Total Memory: 16711680 bytes
Free Memory: 14905336 bytes
Threads Information [ Current THread: Kernel Thread]:
com.tektools.kernel.Kernel$1[name=Kernel Thread Group,maxpri=10]
Thread[Kernel Thread,5,Kernel Thread Group]
Kernel halted at Thu Jul 08 17:18:51 MYT 2010
Any idea what caused it to halt the agent start up? It is the same irregardless of whether my snmpd is running or not.
Looks like something in the core.xml is not as per the expected format. A semicolon maybe missing in the <destinations> tag. A sample core.xml file is below.<?xml version="1.0" encoding="UTF-8" ?>
Hi there Guacomle,
Geezz...thanks! now I manage to start the profiler agent and there is no longer a Fatal log. Now, when i tail the Kernel.log, I keep on seeing this message:
Attempting to send notification to "IP Of Profiler Server"
Call to PlceTrapOrder(126.96.36.199.4.1.10089.10090.10090) succeeded.
Unfortunately, I still do not see anything on my profiler console. Under Events of that resource, the summary shown is No text in varbindings, the short description says Agent not collecting data and the Trapname column shows agentNoResponse.
May I know what configurations have to be done on the profiler server? Under the wrench symbol of the resource, in the Communications portion, the configuration is as below:
Server: "IP of Profiler Server"
HTTP Port: 4319
Server: "IP of Profiler Server"
Community: public (which is the same as the one configured in my agent)
Are the above correct?