Re: "'Resultcache" effect in 10.2.0.4 ??
From: GD <goran99_remove_at_yahoo.com>
Date: Fri, 24 Jul 2009 09:03:34 +0200
Message-ID: <h4bmc6$2pn$1_at_ss408.t-com.hr>
jlchiappa wrote:
> I´m seeing this behavior in a coleague database : the original 9.2.0.7
> EE db was upgraded to 10.2.0.4 EE in the same Solaris 9 box (big one,
> with 32 processors and 64 Mb of RAM), and since then, for big and
> complexes queries (ie, with subqueries, and/or views from views,
> inline views, WITH clause, the hairy ones) the first execution is
> 'slow', say, 15 minutes or alike, and the second execution is 5 times
> or so faster (say 3 minutes or alike). The details :
Date: Fri, 24 Jul 2009 09:03:34 +0200
Message-ID: <h4bmc6$2pn$1_at_ss408.t-com.hr>
jlchiappa wrote:
> I´m seeing this behavior in a coleague database : the original 9.2.0.7
> EE db was upgraded to 10.2.0.4 EE in the same Solaris 9 box (big one,
> with 32 processors and 64 Mb of RAM), and since then, for big and
> complexes queries (ie, with subqueries, and/or views from views,
> inline views, WITH clause, the hairy ones) the first execution is
> 'slow', say, 15 minutes or alike, and the second execution is 5 times
> or so faster (say 3 minutes or alike). The details :
Dynamic sampling taking place at first parse?
http://download.oracle.com/docs/cd/B19306_01/server.102/b14211/stats.htm#i42991 Received on Fri Jul 24 2009 - 02:03:34 CDT