Re: v$sql - executions vs loads vs invalidations

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Tue, 1 May 2012 11:58:32 +0100
Message-ID: <763597603F7C49F597633D5F54577B16_at_Primary>


  • Original Message ----- From: "GG" <grzegorzof_at_interia.pl> To: <gerry_at_millerandbowman.com> Cc: "Oracle-L Group" <oracle-l_at_freelists.org> Sent: Tuesday, May 01, 2012 11:53 AM Subject: Re: v$sql - executions vs loads vs invalidations

W dniu 2012-05-01 11:41, Gerry Miller pisze:

>LOADS: 2794
>INVALIDATIONS: 2793
>EXECUTIONS: 0
You've never mentioned DB version :) but basically looks like statements are only parsed and never executed (application issue ?). You got invalidation because some dependent objects are 'manipulated' via (grants, ddls, stats gathering and so) , loads are because of flushes which could be caused by shared pool free space pressure .
For better understanding we need some info about sga size and allocations and ora-4031 related dumps .
Meantime You can check


Depending on version, invalidation may zero out the execution stats.

Earlier 10g versions, for example, would invalidate ALL cursors relating to a global temporary table if ANY user truncated their private copy of that GTT - resulting in lots of library cache invalidation for everyone all the time - and lots of cursors with high invalidation and reload counts, but zero execution stats

Regards

Jonathan Lewis
http://jonathanlewis.wordpress.com
Oracle Core (Apress 2011)
http://www.apress.com/9781430239543

--
http://www.freelists.org/webpage/oracle-l
Received on Tue May 01 2012 - 05:58:32 CDT

Original text of this message