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: buffer hit cache ratio

Re: buffer hit cache ratio

From: Sybrand Bakker <gooiditweg_at_sybrandb.demon.nl>
Date: Sat, 27 Dec 2003 00:48:10 +0100
Message-ID: <50ipuvo0g7mhtff2r2eg88cjoqd8fn9cpk@4ax.com>


On 26 Dec 2003 12:36:52 -0800, jedithezealot_at_yahoo.com (JEDIDIAH) wrote:

>If you have hardware resources, you should make sure you're exploiting
>them properly. This means configuring your SGA accordingly as well as
>ensuring OS, applications and storage hardware similarly well configured.
>

Which an obvious open door which no one is arguing about.

>The BHCR isn't a silver bullet, just part of a long laundry list of
>things to check. Of course if it were completely meaningless, Oracle
>would leave it out of the statspack report. '-)
>
>>
>> Kyte says the only ratio he uses is parse/execute. Its in both of his
>> books and i believe on his website.
>
>Unnecessary disk IO is going to slow a system down remarkably faster
>than unnecessary cpu activity.

IMO we are discussing increasing hit ratios beyond the 90 - 95 percent here. It is quite well known and quite firmly established, you need lots of memory resources to get those few extra percents out. Usually this results in extra paging.
Please explain in which situation there is 'unnecessary disk I/O'

--
Sybrand Bakker, Senior Oracle DBA
Received on Fri Dec 26 2003 - 17:48:10 CST

Original text of this message

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