From: nsouto@nsw.bigpond.net.au.nospam (Nuno Souto)
Newsgroups: comp.databases.oracle.server
Subject: Re: Number of extents related with performance?
Organization: D.C.S. P/L
Message-ID: <3a5079bb.1207812@news-server>
References: <3A4C4F06.110C4CB6@anonymous.com> <3a4c9864.13389917@news-server> <3A4EE61B.A1323E21@anonymous.com> <3a4f1941.1335067@news-server> <3A512D2F.A10302F5@anonymous.com>
X-Newsreader: Forte Free Agent 1.21/32.243
Lines: 28
Date: Mon, 01 Jan 2001 12:47:48 GMT
NNTP-Posting-Host: 144.132.172.151
X-Complaints-To: news@bigpond.net.au
X-Trace: news-server.bigpond.net.au 978353268 144.132.172.151 (Mon, 01 Jan 2001 23:47:48 EST)
NNTP-Posting-Date: Mon, 01 Jan 2001 23:47:48 EST


On Mon, 01 Jan 2001 17:21:51 -0800, Jimmy <anonymous@anonymous.com>
wrote:

>	I am using block size 8K. In fact, the whole story is, I am DBA, and
>some developers complain about the performance of the database. I found
>that most of the tables they were using is 32K. Later, I truncated some
>tables, change the size of NEXT parameters of some large tables to a
>higer values, and then load the data again to these tables. After the
>above action, developers said the performance is improved. In fact, some
>papers mentioned that the performance is not related with the extent
>sizes (provided that no. of extents below 1000). 
>

And that should hold true in your case.  You mentioned some large
tables were defraged. I assume they would be much larger than 32K?
Because with 32K tables and a 8K block size, it shouldn't matter at
all what you put in INITIAL or NEXT, quite frankly.  

I wonder if the developers were only complaining about a point in time
situation?  It happens.  Might be worth doing a few tests yourself
with the actual data and some of the SQL involved and see if it really
does make a diff?


Cheers
Nuno Souto
nsouto@bigpond.net.au.nospam
http://www.users.bigpond.net.au/the_Den/index.html