Data Cache Hit Ratio

I was reading up on tips and techniques for performance tunning databases. I got to a section that talked about Hit Ratio and using it to track performance. I was wondering if anyone uses it and could elaborate on their experience with it. I am 10g release level is 10.2.040 and we are in a UNIX AIX environment. My production database is about 140 gigs. I have been the DBA here for about a month or so and the don't really complain about DB performance (knock on wood). I have een monitoring the system and don't really see any spikes in performance. I am just trying to be pro-active

In short... If it's not broken, don't try to fix it ;)

Performance tunning has moved long ago from focusing on hit ratios to focusing on wait events.

Having said that, it doesn't hurt to track hit ratios - in posters particular case there are not apparent performance issues so whatever hit ratios poster sees today can be used as a baseline for future reference.

Quote:

---

Originally Posted by jayjabour

I was reading up on tips and techniques for performance tunning databases. I got to a section that talked about Hit Ratio and using it to track performance. I was wondering if anyone uses it and could elaborate on their experience with it. I am 10g release level is 10.2.040 and we are in a UNIX AIX environment. My production database is about 140 gigs. I have been the DBA here for about a month or so and the don't really complain about DB performance (knock on wood). I have een monitoring the system and don't really see any spikes in performance. I am just trying to be pro-active

---

I'd add that cache hit ratios vary in different scenarios and depend heavily on application design. Well as already PAVB mentioned, you can log these parameters in different intervals and baseline them. Comparing them in different intervals will surely help you in future if you needed to tune them.