Path: newssvr20.news.prodigy.com!newsmst01.news.prodigy.com!prodigy.com!in.100proofnews.com!in.100proofnews.com!news2.euro.net!newshub1.home.nl!home.nl!not-for-mail From: Frank Newsgroups: comp.databases.oracle.server Subject: Re: optimizer parameters Date: Fri, 10 Oct 2003 13:38:41 +0200 Organization: @Home Benelux Lines: 52 Message-ID: References: <3F82E81E.98A92384@wholefoods.com> NNTP-Posting-Host: cc28855-a.hnglo1.ov.home.nl Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Trace: news1.tilbu1.nb.home.nl 1065785494 26547 217.121.193.81 (10 Oct 2003 11:31:34 GMT) X-Complaints-To: abuse@home.nl NNTP-Posting-Date: Fri, 10 Oct 2003 11:31:34 +0000 (UTC) User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax) X-Accept-Language: en-us, en In-Reply-To: Xref: newssvr20.news.prodigy.com comp.databases.oracle.server:244989 Luc Gyselinck wrote: > I know about this paper. > > Indeed, the optimizer_index_caching parameter can be calculated by looking > at the ratio of the logical and physical reads. But these counters are > totals of data AND index blocks. But we need the counters for the index > blocks ONLY (to be more accurate). > > Another thing is that in a DEVELOPMENT environment, these counters do not > reflect the PRODUCTION environment. Many developers run 'test' (and most of > the time 'bad') SQL that will make the results of these counters more or > less useless. Maybe you should measure this ratio in the production > environment, calculate the value of the parameter and copy it over to the > development environment and do your tuning? > That's called the acceptance stage... > I will re-read the paper if have my other computer switched on (this one has > no MSWORD installed). > > Thanks for the reaction. > > "Richard Foote" wrote in message > news:O8Tgb.141633$bo1.37865@news-server.bigpond.net.au... > >>Hi Luc, >> >>If I may, I would recommend the paper "The Search For Intelligent Life In >>The CBO" by Tim Gorman at www.evdbt.com as an excellent introduction on > > how > >>to possibly tune these parameters. It provides some sensible techniques on >>how to determine appropriate values rather than the broad generalisations >>(which may or may not be suitable) offered by others in this thread. >> >>Good Luck >> >>Richard >> >> > > > Read it with great interest -- Regards, Frank van Bortel