Most consultants you hire will use the trial and error method for performance tuning.
- Here is a list of ratios, metrics and internal Oracle items.
- Here is something from that list that looks like it may solve your problem.
- Let's try this change and measure the results.
- If it fixes the problem, great. If not repeat.
There has got to be a better answer!
I think that measuring the impact to users as the real outcome, not that the cache hit ratio is now an optimal level.
How do we go about measuring the impact to the user? That is the tricky part. Oracle 9i did not come right out with that measurement. Enter 10g.
Over the next few Blogs, I hope to dive into the inner workings of Oracle 10g R1 and R2. Hopefully we can find what we are looking for.
-RBT
Posts
No comments:
Post a Comment