col input_bytes_display format a18
col output_bytes_display format a18
col time_taken_display format a18
set linesize 132
set pagesize 50
SELECT session_recid, input_bytes_display, output_bytes_display,time_taken_display, end_time
FROM v$rman_backup_job_details
ORDER BY end_time
SESSION_RECID INPUT_BYTES_DISPLA OUTPUT_BYTES_DISPL TIME_TAKEN_DISPLAY END_TIME
------------- ------------------ ------------------ ------------------ ---------
7049 6.99G 965.00M 00:09:38 13-FEB-11
7057 6.94G 948.13M 00:08:51 14-FEB-11
7065 10.28G 1.81G 00:11:22 15-FEB-11
7073 9.96G 1.73G 00:30:26 16-FEB-11
7081 6.85G 784.77M 00:10:24 17-FEB-11
7089 6.82G 784.52M 00:08:34 18-FEB-11
7097 6.93G 814.08M 00:09:03 19-FEB-11
7105 7.00G 830.98M 00:08:50 20-FEB-11
7113 6.97G 822.40M 00:09:00 21-FEB-11
7121 6.97G 821.77M 00:10:11 22-FEB-11
7129 6.98G 822.81M 00:10:40 23-FEB-11
Based on the result above, we should investigate what happened on 16-FEB-11 that backup took longer than usual. As basic RMAN tuning, we have two recomendations for improving RMAN backup
- Use incremental backup with block change tracking enabled.
- Adjust multipexing of backup sets by adjusting filesperset (limit the number of datafiles in each backupset), maxopenfiles (limit the number of files that can be opened for reads simutaneously), diskratio (read datafiles from specified number of disks)
No comments:
Post a Comment