原创

oracle AWR深入研究分析,如何使用

版权声明:本文为博主原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接和本声明。
本文链接:https://xxlcube.blog.csdn.net/article/details/9701123

AWR的前身是statspack,当然现在还在,只不过大家都在使用AWR,因为它方便,简单,直观,形象。

AWR是oracle内置工具,安装oracle时已经自动安装完毕,无需额外安装了。

SELECT *
FROM DBA_HIST_WR_CONTROL;


这里主要查看AWR策略的制定信息,多久抓一次快照,保留多长时间。

DECLARE
BEGIN
   DBMS_WORKLOAD_REPOSITORY.modify_snapshot_settings (
      interval    => 30,
      retention   => 10 * 24 * 60);
END;

修改策略信息


如何生成AWR报表信息:

依赖于awrrpt.sql脚本,该脚本的路径是$oracle_home/RDBMS/ADMIN

该脚本的内容如下:

Rem $Header: awrrpt.sql 24-oct-2003.12:04:53 pbelknap Exp $
Rem
Rem awrrpt.sql
Rem
Rem Copyright (c) 1999, 2003, Oracle Corporation.  All rights reserved.  
Rem
Rem    NAME
Rem      awrrpt.sql
Rem
Rem    DESCRIPTION
Rem      This script defaults the dbid and instance number to that of the
Rem      current instance connected-to, then calls awrrpti.sql to produce
Rem      the Workload Repository report.
Rem
Rem    NOTES
Rem      Run as select_catalog privileges.  
Rem      This report is based on the Statspack report.
Rem
Rem      If you want to use this script in an non-interactive fashion,
Rem      see the 'customer-customizable report settings' section in
Rem      awrrpti.sql
Rem
Rem    MODIFIED   (MM/DD/YY)
Rem    pbelknap    10/24/03 - swrfrpt to awrrpt 
Rem    pbelknap    10/14/03 - moving params to rpti 
Rem    pbelknap    10/02/03 - adding non-interactive mode cmnts 
Rem    mlfeng      09/10/03 - heading on 
Rem    aime        04/25/03 - aime_going_to_main
Rem    mlfeng      01/27/03 - mlfeng_swrf_reporting
Rem    mlfeng      01/13/03 - Update comments
Rem    mlfeng      07/08/02 - swrf flushing
Rem    mlfeng      06/12/02 - Created
Rem

--
-- Get the current database/instance information - this will be used 
-- later in the report along with bid, eid to lookup snapshots

set echo off heading on underline on;
column inst_num  heading "Inst Num"  new_value inst_num  format 99999;
column inst_name heading "Instance"  new_value inst_name format a12;
column db_name   heading "DB Name"   new_value db_name   format a12;
column dbid      heading "DB Id"     new_value dbid      format 9999999999 just c;

prompt
prompt Current Instance
prompt ~~~~~~~~~~~~~~~~

select d.dbid            dbid
     , d.name            db_name
     , i.instance_number inst_num
     , i.instance_name   inst_name
  from v$database d,
       v$instance i;

@@awrrpti

undefine num_days;
undefine report_type;
undefine report_name;
undefine begin_snap;
undefine end_snap;
--
-- End of file
那么我们就通过调用该脚本产生报表信息:


后面就是基本简单操作。

关键产生报表后如何分析呢?

1、观察CACHE SIZE,观察主机内存情况,判断SGA的内存分批额是否合理

2、观察load profile的事务情况,判断系统繁忙程度

3、观察load profile的hard parses和Parses值,观察是否存在过多硬解析,从而定位是否存在绑定变量的问题

4、观察top 5 timed events,查看系统的瓶颈所在,一般而言,CPU TIME排名第一并且占用了绝大部分的时间比例,说明系统没有什么瓶颈

5、关注如下6个ordered by的sql语句,

elapsed_time,cpu time,gets,reads,executions,parse calls


文章最后发布于: 2013-08-01 13:53:54
展开阅读全文
0 个人打赏
私信求帮助

没有更多推荐了,返回首页

©️2019 CSDN 皮肤主题: 酷酷鲨 设计师: CSDN官方博客

分享到微信朋友圈

×

扫一扫,手机浏览