Description of problem: when using the /etc/init.d/oracledb.sh script that handles multiple oracles instances, it kills ALL oracle processes when shutting down an oracle service. Version-Release number of selected component (if applicable): rgmanager-1.9.80-1 How reproducible: 100% Steps to Reproduce: 1. Create 2 oracle clustered services 2. Run them on the same cluster member 3. Fail one over to another node Actual results: All oracle processes get killed and restarted Expected results: Only the moving service processes should be killed and moved. Additional info:
Created attachment 321658 [details] Patched oracle database agent that supports multiple instances (but has this bug)
Shane noted that changing the 'pids' line in force_cleanup to this might fix the problem: pids=`ps ax | grep $ORACLE_HOME | grep "ora_.*_${ORACLE_SID}" |grep -v grep | awk '{print $1}'`
Created attachment 326415 [details] Fix, pass 1
Based on customer reports, I have copied the above script to my notes page here: http://people.redhat.com/lhh/oracle-rhel5-notes-0.6/oracle-notes.html However, due to limited feedback, I am closing this.