OEM process sucking up CPU

Question

HI.
I have seen a select query that is tkaing up 100% consistently of CPU resources. This is the statement itself:
SQL> select sql_text from v$sql where sql_id = '0hbv80w9ypy0n';
SQL_TEXT
/* OracleOEM */
SELECT end_time, status, session_key, session_recid, session_stamp, command_id, start_time,
time_taken_display, input_type,output_device_type, input_bytes_display, output_bytes_display,output_
bytes_per_sec_display FROM v$rman_backup_job_details WHERE end_time = (select max(end_time) from v$rman_backup_jo
b_details);
Looks like some hangover from the nightly backup. Has anyone seen this before? I would like to kill it as the backups are long over (7 hours) but I am scared to do!
Thanks,
Edited by: user11981168 on 09-Mar-2010 03:50

Answer 1

This is a known bug - see Metalink EM Agent DBSNMP Using Up Excessive Temp Space In Database [ID 748251.1]
There is a workaround (as sysdba):
1] exec dbms_stats.DELETE_TABLE_STATS('SYS','X$KCCRSR'); ==> deletes the statistics on the fixed object
2] exec dbms_stats.LOCK_TABLE_STATS('SYS','X$KCCRSR'); ==> lock that object so that statistics will not be collected in future
3] alter system flush shared_pool; ==> cannot be skipped
Thanks
Paul
Edited by: PaulTill2508 on Mar 9, 2010 2:12 PM