Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Oracle slowed down

Re: Oracle slowed down

From: HansF <News.Hans_at_telus.net>
Date: Sat, 22 Apr 2006 16:10:04 GMT
Message-Id: <pan.2006.04.22.15.14.57.706852@telus.net>


On Sat, 22 Apr 2006 15:23:54 +0000, Bob Jones wrote:

>
> Do all these mean BHCR is deprecated?

No. They simply mean that tuning does not start, or end, with BCHR.

It is, as you say, only one tool in the toolkit. It has repeatedly been proven suspect and should not be considerd a primary tool.

Sadly, since it is so easy to generate the BCHR, many people use that as the key tool,

and that seems to be the basis of the religious musings ... those that understand it, and Oracle internals, realize how insignificant the value of the ratio truly is to the totality of tuning.

For example, misses implies physical disk reads. Could be a totally stupid assumption in large cache SANS.

Another example, many believe there is only one copy of a block in memory. Only true if there is absolutely no update to the block. Which one did we miss?

-- 
Hans Forbrich                           
Canada-wide Oracle training and consulting
mailto: Fuzzy.GreyBeard_at_gmail.com   
*** Top posting [replies] guarantees I won't respond. ***
Received on Sat Apr 22 2006 - 11:10:04 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US