Hi Experts,
Our SAP system on Unix - Oracle 10g, patch 10.2.0.2:
- Running on MCSG configuration.
- Offline backup scheduled weekly, runs every Saturday 23:00.
- Online backup on most weekdays, evening time.
I found that Oracle buffer dropped to 93.4% on Feb 8th, 2009. Below the history from ST04:
Logical Physical Buffer
Day Date Endtime reads reads quality
.
.
Mon 09.02.2009 22:00:05 74,202,051 2,570,395 96.5
Sun 08.02.2009 22:00:03 10,560,849 699,182 93.4
Sat 07.02.2009 22:00:03 757,531,921 7,200,330 99.0
Fri 06.02.2009 22:00:02 690,784,983 6,823,435 99.0
.
.
Indeed in alert_<SID>.log there was database shutdown for offline backup, however the buffer quality has never been this low before and after Feb 8th, 2009. Below are some samples from weekend's Oracle buffer quality:
.
Sun 22.02.2009 22:00:11 16,063,545 220,308 98.6
Sat 21.02.2009 22:00:19 907,000,906 11,910,741 98.7
Fri 20.02.2009 22:00:20 854,489,867 11,530,043 98.7
.
Sun 15.02.2009 22:00:19 12,004,781 194,400 98.4
Sat 14.02.2009 22:00:06 595,715,004 9,095,210 98.5
Fri 13.02.2009 22:00:07 536,747,942 8,695,957 98.4
.
Sun 01.02.2009 22:00:59 14,460,334 255,226 98.2
Sat 31.01.2009 22:00:48 485,708,762 4,065,589 99.2
Fri 30.01.2009 22:01:49 454,970,905 3,738,336 99.2
.
My problem:
- Background job logs already reorganized.
- ST03N daily data already reorganized.
- Nothing suspicious on UX's syslog and alert_<SID>.log.
Are there any other places where I can gather the evidence?
Thanks before.
Regards,
Agoes