I have installed OEM 2.1 on NT, but when I try to discover Node it fails. MY question (Intellegent Agent gets installed automatically when you are installing OEM or is it installed seprately)
If the node that you are trying to connect is on a solaris or any unix platforum, then try these trouble shooting measures. 1. First check to see whether the agent is running by typing the following
command
lsnrctl dbsnmp_status
if you get the following message
"The db subagent is not started"
then follow the following steps:
i) try to do the following
lsnrctl dbsnmp_start
lsnrctl dbsnmp_status
Even after this step you couldn't start the agent do
the below specified steps
a) Check to see the log messages under
$ORACLE_HOME/network/log/dbsn*.log
if the logs show NMS-00001
b) Go to the directory $ORACLE_HOME/network/admin and rename the
The Intelliget Agent is installed by default, but you must start them on control panel.
This is a well known bug in metalink site.
On NT platform, you don't have to use dbsnmp_* functions.
These function are there only because the developers have forget to remove them. This work only on UNIX.
To start listener on NT, use control panel, start Intelligent Agent.
This work together with Data Gatherer. You have to start this service also. If Datagatherer don't start, check your registry, the oracle homes must be in order. e.g. if you have home1, home2, but don't have home0, you have to create a home0(dumy).
If your node have not discovered databases,
delete %orahome%/network/admin/snmp_ro.ora, snmp_wr.ora files, ./agent/*.q files and restart intelligent agent.
you can check for error on %orahome%/network/log/nmiconf.log file for errors, or check %orahome%/network/agent/services.ora for discovered services.
For the jobs, this is another story.
You have to create a account on the server machine where
the database is, with LOG ON AS A BATCH JOB user rigths.
In OEM console, on preference, you must log the job with this user, so the Intelligent Agent do the job batch thru this user with job batch rigts.
Set the prefered credentials for the node and instance. For the instance it should be either sys or system. For the node, it should be a valid user name, with correct permissions. Then try creating a simple jobs like
for instance do select * from v$instance; and for the node job, do a directory listing and confirm that you can
successfully do things. Besure that OMS and OEM are of the same version. If they are different then you would have trouble in creating the events or scheduling the jobs.
Sorry,folks for giving an Unix troble shooting for NT.
Bookmarks