dbsnmp and net-snmp

Hi,

I am trying to monitor Oracle with its Intelligent Agent and net-snmp.

I followed all the instructions in this document from Jim Weller : http://www.jimweller.net/article.php?story=20030731093250729

Just after restarting the dbsnmp service I can send queries to the 1.3.6.1.4.1.2021 section of the MIB tree (UCD-SNMP-MIB) on port 1161 or port 161. This works fine.

For a few minutes I can also snmpwalk the system for the Oracle OIDs on port 161 (not on port 1161). Then after a while the Oracle OIDs are not available. All queries end up with a time out.

When I check the db sub agent status with 'lsnrctl dbsnmp_status' it says 'the db sub agent is not started'. However a 'ps -eaf' shows me that the two dbsnmp processes are still running.

So if I restart the dbsnmp agent the Oracle OIDs will be available for a few minutes then it will stop working.

The logs for the dsnmp sub agent show that the dbsnmpw (worker process) is trying to restart the COMM process.

I went to Oracle Metalink and found the article:Doc ID: Note:1066197.6

      The solution they provided was to remove the *.q files from the $ORACLE_HOME/network/agent directory and then restart the Intelligent Agent.

      I tried that but it didn't fix the problem.

      Has anyone experienced the same problem ?

      I would appreciate any help.

      Many thanks.
      Sam