problem starting the intelligent agent

[beto]

hello there,
I am running oracle 8.1.5 on a Sun 2.6 server. I have been unable to start the intelligent agent on the server.
When I type the command lsnrctl dbsnmp_start, I get the following error:
Warning : (ADDRESS defined in /u02/app/oracle/8.1.5/network/admin/listener.ora will be skipped, because it does not monitor any of the valid SIDs {test2}
Warning : No Listener found for SID test2. test2 will be skipped.
This happends although everything seems to be ok with the listener.ora. I can connect to the database just ok from my desktop.
this is what I have in my oratab file :
test2:/u02/app/oracle/8.1.5:Y
Any idea how to solve the problem?

Thanks

[sambavan]

Here are some of the ways to do a troble shoot on the problem.

First check to see wheter you were able to start the agent.

lsnrctl dbsnmp_status

If you are getting an info saying that the "The db subagent is not started" then follow the following steps.

$ORACLE_HOME/network/log/dbsn*.log

Now see what kind of errors that you see there.

If the logs show NMS-00001

then go to the directory

$ORACLE_HOME/network/admin

mv snmp_ro.ora snmp_ro.ora.old
mv snmp_rw.ora snmp_rw.ora.old

Now go to the directory

$ORACLE_HOME/agent
mkdir junke
mv *.q ./junke
mv services.ora ./junke

now try starting it

lsnrctl dbsnmp_start

Confirm it by dbsnmp_status also confirm through

ps -ef|grep dbsnmp

you should see a parent and a child process of dbsnmp

Once after confirming things, go ahead and delete the .old files and juke directory.

Hope this would help you.

Sam

[beto]

hello Sam the error I have in the snmp*.ora this one instead:

NMS-00308: Failed to listen on address: another agent may be running

[sambavan]

o.k lets step through.

First check whether any agent 'dbsnmp' currently running.
check it using 'ps -ef|grep dbsnmp'

2. If you see one running, then go ahead and stop that one by, 'lsnrctl dbsnmp_stop'. Now again check for the process to be running. If you notice one still running, go ahead and kill that one.

3. Now clean all the directories as I said earlier.
NOTE****: If you had sheduled any jobs or events through oem on this node, it would be lost. So the safest thing is rename them.

4. Now check and make sure that your listeners are working fine "lsnrctl services".

5. confirm that the ports that you are using on the listener are working fine, b/c its gets used under services.ora. Once after confirming it, go ahead and start the agent 'lsntctl dbsnmp_start' and check the status of it through 'lsnrctl dbsnmp_status'

6. Now if it was an unsuccessfull try, then go ahead and let me know of the error messages that you got at each step


----------------------------

Here is something that I found at Metalink,

NMS-00308: Failed to listen on address: another agent may be running
--------------------------------------------------------------------

Cause: The agent's connection cache could not be initialized. This message
usually appears when another agent is running. Only one copy can be
running at a time.

Action: If another agent is running, bring it down before running the desired
agent.

Message Type: Agent Operation Messages
OEM Versions Applicable: 1.3.5, 1.3.6, 1.4.0, 1.4.1, 1.5.0, 1.6.0
.

-------------------------------
Good luck,
sam

[beto]

hello Sam,
I followed your recommendations and now the agent is running.
I'm now in the process of configuring OEM.
Thanks very much for your help.