| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> Good Performance
Could someone give me some "benchmarks" defining good performance?
Our management comes from small shop, SQL Server world. I was the first Oracle resource here, about a year and a half ago. We have a web-based platform using RAC, BEA and IFS ...
There were some performance issues about three months ago and after many battles I convinced management to add spindles ... Things improved until a major upgrade of the app. After the upgrade things tanked. It's diffucult to run legitimate tests in our LOD environment (long story). Two days later myself and the architect for the app made some changes, and things are once again speedy. BUT, we now have consultants coming out of the woodwork, retained by management. Mutliple companies doing analysis.=20
What do I say when I feel we shouldn't monkey with things. Of course there is going to be queries to tune. But my question is this:
What is a good benchmark for determining if a query is doing too many buffer reads, or too many disk reads? Is 10,000 buffer reads considered high, or is it 1,000,000? Or does it matter if, on the overall, things kick butt.
This is a half hour window during peak usage. Is anything way off?
Instance Efficiency Percentages (Target 100%)
Buffer Nowait %: 99.93 Redo NoWait %: 100.00
Buffer Hit %: 98.78 In-memory Sort %: 100.00
Library Hit %: 99.86 Soft Parse %: 98.96
Execute to Parse %: 85.06 Latch Hit %:
99.97
Parse CPU to Parse Elapsd %: 93.68 % Non-Parse CPU:
99.66
Top 5 Timed Events
~~~~~~~~~~~~~~~~~~ % Total Event Waits Time (s)Ela Time
CPU time 1,705 41.97 db file sequential read 223,374 1,123 27.65 global cache cr request 377,359 776 19.10 log file sync 26,171 92 2.27 global cache null to x 20,304 852.10
 |
 |