Oracle索引的监控
一.1 BLOG文档结构图
一.2 前言部分
一.2.1 导读和注意事项
各位技术爱好者，看完本文后，你可以掌握如下的技能，也可以学到一些其它你所不知道的知识，~O(∩_∩)O~：
① 掌握oracle中索引的监控方法
② sys.col_usage$的初步了解
Tips：
① 本文在ITpub（http://blog.itpub.net/26736162）和博客园(http://www.cnblogs.com/lhrbest)有同步更新
② 文章中用到的所有代码，相关软件，相关资料请前往小麦苗的云盘下载（http://blog.itpub.net/26736162/viewspace-1624453/）
③ 若文章代码格式有错乱，推荐使用搜狗、360或QQ浏览器，也可以下载pdf格式的文档来查看，pdf文档下载地址：http://blog.itpub.net/26736162/viewspace-1624453/
④ 本篇BLOG中命令的输出部分需要特别关注的地方我都用灰色背景和粉红色字体来表示，比如下边的例子中，thread 1的最大归档日志号为33，thread 2的最大归档日志号为43是需要特别关注的地方；而命令一般使用黄色背景和红色字体标注；对代码或代码输出部分的注释一般采用蓝色字体表示。
List of Archived Logs in backup set 11
Thrd Seq Low SCN Low Time Next SCN Next Time
---- ------- ---------- ------------------- ---------- ---------
1 32 1621589 2015-05-29 11:09:52 1625242 2015-05-29 11:15:48
1 33 1625242 2015-05-29 11:15:48 1625293 2015-05-29 11:15:58
2 42 1613951 2015-05-29 10:41:18 1625245 2015-05-29 11:15:49
2 43 1625245 2015-05-29 11:15:49 1625253 2015-05-29 11:15:53
[ZHLHRDB1:root]:/>lsvg -o
T_XDESK_APP1_vg
rootvg
[ZHLHRDB1:root]:/>
00:27:22 SQL> alter tablespace idxtbs read write;
====》2097152*512/1024/1024/1024=1G
本文如有错误或不完善的地方请大家多多指正，ITPUB留言或QQ皆可，您的批评指正是我写作的最大动力。
一.3 相关知识点扫盲(摘自网络)
合理的为数据库表上创建战略性索引，可以极大程度的提高查询性能。但事实上日常中我们所创建的索引并非战略性索引，恰恰是大量冗余或是根本没有用到的索引耗用了大量的存储空间，导致DML性能低下。 应用程序在开发时，可能会建立众多索引，但是这些索引的使用到底怎么样，是否有些索引一直都没有用到过，这需要我们对这些索引进行监控，以便确定他们的使用情况，并为是否可以清除它们给出依据。
冗余索引的弊端：
大量冗余和无用的索引导致整个数据库性能低下，耗用了大量的CPU与I/O开销，具体表现如下:
a、浪费大量的存储空间，尤其是大表的索引，浪费的存储空间尤其可观(索引段的维护与管理)
b、增加了DML 操作（UPDATE、INSERT、DELETE）的开销
c、耗用大量统计信息(索引)收集的时间
d、结构性验证时间
f、增加了恢复所需的时间
本文介绍两种方式：
第一：开启监控功能；
第二：查看历史的执行计划，进行分析；
一.4 索引监控的方法
一.4.1 方法一：开启监控功能
1、单个索引监控
a、对于单个索引的监控，可以使用下面的命令来完成
alter index <INDEX_NAME> monitoring usage;
b、关闭索引监控
alter index <INDEX_NAME> nomonitoring usage;
c、观察监控结果(查询v$object_usage视图)
select * from v$object_usage;
2、schema级别索引监控
如果我们想在系统中监控所有的索引，那么我们可以通过下面脚本实现监控数据库所有的索引。注意我们要排除一些系统表的索引、以及LOB indexes。原因有下面两个：
1：LOB indexes不能修改，否则会报ORA-22864错误（ORA-22864: cannot ALTER or DROP LOB indexes）。
2：ORA-00701: object necessary for warmstarting database cannot be altered
ORA-00701: object necessary for warmstarting database cannot be altered
00701. 00000 - "object necessary for warmstarting database cannot be altered"
*Cause: Attempt to alter or drop a database object (table, cluster, or
index) which are needed for warmstarting the database.
*Action: None.
直接执行脚本来开启索引监控，当然监控索引时长非常重要，太短的话有可能导致查询出来的数据有问题，一般建议监控一周后即可，OLAP系统则需要适当延长监控的时间。
SELECT 'ALTER INDEX ' || owner || '.' || index_name || ' MONITORING USAGE;' enable_monitor,
'ALTER INDEX ' || owner || '.' || index_name ||
' NOMONITORING USAGE;' disable_monitor
FROM dba_indexes
WHERE INDEX_TYPE != 'LOB'
and owner IN
(SELECT username FROM dba_users WHERE account_status = 'OPEN')
AND owner NOT IN ('SYS',
'SYSTEM',
'PERFSTAT',
'MGMT_VIEW',
'MONITOR',
'SYSMAN',
'DBSNMP')
AND owner not like '%SYS%';
监控一个月就大概可以知道那些是无用的索引了。
虽然v$object_usage表能记录索引监控和使用的状态，但它不能统计索引被使用的次数和频率，只记录了在开启索引监控的时间段索引是否被使用过，这一点要值的注意。
另外需要注意的2点：
① 10g在收集统计信息时会导致索引被监控、这并非SQL语句产生、而在11g则不会出现这种情况了
② 外键索引不会因为主表的DML操作而被监控到、不要因为该索引没用而将它给删了
一.4.1.1 个人实验
新建1个表TB_LHR_20160622，并创建2个索引：
SYS@raclhr2> select * from v$version;
BANNER
-------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.4.0 - 64bit Production
PL/SQL Release 11.2.0.4.0 - Production
CORE 11.2.0.4.0 Production
TNS for IBM/AIX RISC System/6000: Version 11.2.0.4.0 - Production
NLSRTL Version 11.2.0.4.0 - Production
SYS@raclhr2> Create Table TB_LHR_20160622 nologging As select * from dba_objects;
Table created.
SYS@raclhr2> create index ind_TB_LHR_20160622_id on TB_LHR_20160622(object_id);
Index created.
SYS@raclhr2> create index ind_TB_LHR_20160622_name on TB_LHR_20160622(object_name);
Index created.
查询v$object_usage视图，收集统计信息：
SYS@raclhr2> select * from v$object_usage;
no rows selected
SYS@raclhr2> BEGIN
2 dbms_stats.gather_table_stats(USER,
3 'TB_LHR_20160622',
4 cascade => TRUE,
5 degree => 8);
6 END;
7 /
PL/SQL procedure successfully completed.
SYS@raclhr2> select * from v$object_usage;
no rows selected
开启索引的监控：
SYS@raclhr2> alter index ind_TB_LHR_20160622_id monitoring usage;
Index altered.
SYS@raclhr2> COL INDEX_NAME FOR A25
SYS@raclhr2> COL TABLE_NAME FOR A20
SYS@raclhr2> COL MONITORING FOR A10
SYS@raclhr2> COL USED FOR A10
SYS@raclhr2> COL START_MONITORING FOR A20
SYS@raclhr2> COL END_MONITORING FOR A20
SYS@raclhr2> select * from v$object_usage;
INDEX_NAME TABLE_NAME MONITORING USED START_MONITORING END_MONITORING
------------------------- -------------------- ---------- ---------- -------------------- --------------------
IND_TB_LHR_20160622_ID TB_LHR_20160622 YES NO 06/22/2016 15:15:54
SYS@raclhr2> alter index ind_TB_LHR_20160622_name monitoring usage;
Index altered.
SYS@raclhr2> select count(1) from TB_LHR_20160622 t where t.object_id=88;
COUNT(1)
----------
1
SYS@raclhr2> explain plan for select count(1) from TB_LHR_20160622 t where t.object_id=88;
Explained.
SYS@raclhr2> select * from table(dbms_xplan.display());
PLAN_TABLE_OUTPUT
-------------------------------------------------------------------------------------------------------
Plan hash value: 2688591802
--------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 5 | 1 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 5 | | |
|* 2 | INDEX RANGE SCAN| IND_TB_LHR_20160622_ID | 1 | 5 | 1 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("T"."OBJECT_ID"=88)
14 rows selected.
SYS@raclhr2> COL INDEX_NAME FOR A25
SYS@raclhr2> COL TABLE_NAME FOR A20
SYS@raclhr2> COL MONITORING FOR A10
SYS@raclhr2> COL USED FOR A10
SYS@raclhr2> COL START_MONITORING FOR A20
SYS@raclhr2> COL END_MONITORING FOR A20
SYS@raclhr2> select * from v$object_usage;
INDEX_NAME TABLE_NAME MONITORING USED START_MONITORING END_MONITORING
------------------------- -------------------- ---------- ---------- -------------------- --------------------
IND_TB_LHR_20160622_ID TB_LHR_20160622 YES YES 06/22/2016 15:15:54
IND_TB_LHR_20160622_NAME TB_LHR_20160622 YES NO 06/22/2016 15:16:17
注意：SELECT * FROM V$OBJECT_USAGE; 只能查看当前用户下被监控的索引信息。即使sys、system用户也不能查看其它用户的信息，如下，但我们可以创建一个视图来解决这个问题。
SYS@raclhr2> conn scott/tiger
Connected.
SCOTT@raclhr2> select * from v$object_usage;
no rows selected
SCOTT@raclhr2> conn / as sysdba
Connected.
SYS@raclhr2> create or replace view vw_INDEX_USAGE_lhr AS
2 SELECT U.NAME OWNER,
3 IO.NAME INDEX_NAME,
4 T.NAME TABLE_NAME,
5 DECODE(BITAND(I.FLAGS, 65536), 0, 'NO', 'YES') MONITORING,
6 DECODE(BITAND(OU.FLAGS, 1), 0, 'NO', 'YES') USED,
7 OU.START_MONITORING START_MONITORING,
8 OU.END_MONITORING END_MONITORING
9 FROM SYS.USER$ U,
10 SYS.OBJ$ IO,
11 SYS.OBJ$ T,
12 SYS.IND$ I,
13 SYS.OBJECT_USAGE OU
14 WHERE I.OBJ# = OU.OBJ#
15 AND IO.OBJ# = OU.OBJ#
16 AND T.OBJ# = I.BO#
17 AND U.USER# = IO.OWNER#;
View created.
SYS@raclhr2> create or replace public synonym syn_INDEX_USAGE_lhr for sys.vw_INDEX_USAGE_lhr;
Synonym created.
SYS@raclhr2> grant select on sys.vw_INDEX_USAGE_lhr to public;
Grant succeeded.
SYS@raclhr2> conn scott/tiger
Connected.
SCOTT@raclhr2> set line 9999 pagesize 9999
SCOTT@raclhr2> col owner format A10
SCOTT@raclhr2> COL INDEX_NAME FOR A25
SCOTT@raclhr2> COL TABLE_NAME FOR A20
SCOTT@raclhr2> COL MONITORING FOR A10
SCOTT@raclhr2> COL USED FOR A10
SCOTT@raclhr2> COL START_MONITORING FOR A20
SCOTT@raclhr2> COL END_MONITORING FOR A20
SCOTT@raclhr2> SELECT * FROM syn_INDEX_USAGE_lhr;
OWNER INDEX_NAME TABLE_NAME MONITORING USED START_MONITORING END_MONITORING
---------- ------------------------- -------------------- ---------- ---------- -------------------- --------------------
SYS IND_TB_LHR_20160622_ID TB_LHR_20160622 YES YES 06/22/2016 15:15:54
SYS IND_TB_LHR_20160622_NAME TB_LHR_20160622 YES NO 06/22/2016 15:16:17
取消索引的监控：
SCOTT@raclhr2> CONN / AS SYSDBA
Connected.
SYS@raclhr2> alter index ind_TB_LHR_20160622_id nomonitoring usage;
Index altered.
SYS@raclhr2> SELECT * FROM syn_INDEX_USAGE_lhr;
OWNER INDEX_NAME TABLE_NAME MONITORING USED START_MONITORING END_MONITORING
---------- ------------------------- -------------------- ---------- ---------- -------------------- --------------------
SYS IND_TB_LHR_20160622_ID TB_LHR_20160622 NO YES 06/22/2016 15:15:54 06/22/2016 15:22:30
SYS IND_TB_LHR_20160622_NAME TB_LHR_20160622 YES NO 06/22/2016 15:16:17
SYS@raclhr2> alter index ind_TB_LHR_20160622_name nomonitoring usage;
Index altered.
SYS@raclhr2> SELECT * FROM syn_INDEX_USAGE_lhr;
OWNER INDEX_NAME TABLE_NAME MONITORING USED START_MONITORING END_MONITORING
---------- ------------------------- -------------------- ---------- ---------- -------------------- --------------------
SYS IND_TB_LHR_20160622_ID TB_LHR_20160622 NO YES 06/22/2016 15:15:54 06/22/2016 15:22:30
SYS IND_TB_LHR_20160622_NAME TB_LHR_20160622 NO NO 06/22/2016 15:22:45 06/22/2016 15:23:12
一.4.1.2 实验中用到的SQL
drop table TB_LHR_20160622 purge;
Create Table TB_LHR_20160622 nologging As select * from dba_objects;
create index ind_TB_LHR_20160622_id on TB_LHR_20160622(object_id);
create index ind_TB_LHR_20160622_name on TB_LHR_20160622(object_name);
select * from v$object_usage;
BEGIN
dbms_stats.gather_table_stats(USER,
'TB_LHR_20160622',
cascade => TRUE,
degree => 8);
END;
/
alter index ind_TB_LHR_20160622_id monitoring usage;
alter index ind_TB_LHR_20160622_name monitoring usage;
select count(1) from TB_LHR_20160622 t where t.object_id=88;
set line 9999 pagesize 9999
col owner format A10
COL INDEX_NAME FOR A25
COL TABLE_NAME FOR A20
COL MONITORING FOR A10
COL USED FOR A10
COL START_MONITORING FOR A20
COL END_MONITORING FOR A20
select * from v$object_usage;
注意：SELECT * FROM V$OBJECT_USAGE; 只能查看当前用户下被监控的索引信息。即使sys、system用户也不能查看其它用户的信息。
alter index ind_TB_LHR_20160622_id nomonitoring usage;
alter index ind_TB_LHR_20160622_name nomonitoring usage;
--- drop table t purge; 表删掉后 v$object_usage 中关于监控的信息也删除了
----切换用户后查询select * from v$object_usage;查询不到数据，下边这个SQL可以查询任何用户下的索引使用情况
create or replace view vw_INDEX_USAGE_lhr AS
SELECT U.NAME OWNER,
IO.NAME INDEX_NAME,
T.NAME TABLE_NAME,
DECODE(BITAND(I.FLAGS, 65536), 0, 'NO', 'YES') MONITORING,
DECODE(BITAND(OU.FLAGS, 1), 0, 'NO', 'YES') USED,
OU.START_MONITORING START_MONITORING,
OU.END_MONITORING END_MONITORING
FROM SYS.USER$ U,
SYS.OBJ$ IO,
SYS.OBJ$ T,
SYS.IND$ I,
SYS.OBJECT_USAGE OU
WHERE I.OBJ# = OU.OBJ#
AND IO.OBJ# = OU.OBJ#
AND T.OBJ# = I.BO#
AND U.USER# = IO.OWNER#;
create or replace public synonym syn_INDEX_USAGE_lhr for sys.vw_INDEX_USAGE_lhr;
set line 9999 pagesize 9999
col owner format A10
COL INDEX_NAME FOR A25
COL TABLE_NAME FOR A20
COL MONITORING FOR A10
COL USED FOR A10
COL START_MONITORING FOR A20
COL END_MONITORING FOR A20
SELECT * FROM syn_INDEX_USAGE_lhr;
批量监控系统的所有索引：
SELECT 'ALTER INDEX ' || owner || '.' || index_name || ' MONITORING USAGE;' enable_monitor,
'ALTER INDEX ' || owner || '.' || index_name ||
' NOMONITORING USAGE;' disable_monitor
FROM dba_indexes
WHERE INDEX_TYPE != 'LOB'
and owner IN
(SELECT username FROM dba_users WHERE account_status = 'OPEN')
AND owner NOT IN ('SYS',
'SYSTEM',
'PERFSTAT',
'MGMT_VIEW',
'MONITOR',
'SYSMAN',
'DBSNMP')
AND owner not like '%SYS%';
一.4.2 方法二：查看历史的执行计划进行分析
虽然v$object_usage表能记录索引监控和使用的状态，但它不能统计索引被使用的次数和频率，只记录了在开启索引监控的时间段索引是否被使用过，因此想详细了解索引的使用情况我们可以利用AWR的一些视图dba_hist_sql_plan和dba_hist_sqlstat来弄清楚数据库访问某个索引的次数、索引访问的类型，如索引范围扫描或索引唯一扫描。
WITH tmp1 AS
(SELECT i.OWNER INDEX_OWNER,
i.table_owner,
TABLE_NAME,
INDEX_NAME,
INDEX_TYPE,
(select nb.created
from dba_objects nb
WHERE nb.owner = i.owner
and nb.object_name = i.index_name
and nb.subobject_name is null) created,
(SUM(S.bytes) / 1024 / 1024) INDEX_MB
FROM DBA_SEGMENTS S, DBA_INDEXES I
WHERE i.INDEX_NAME = s.SEGMENT_NAME
and i.owner = s.owner
and s.owner not like '%SYS%'
/*and s.owner = 'FUNDZ'*/
GROUP BY i.OWNER, i.table_owner, TABLE_NAME, INDEX_NAME, INDEX_TYPE
HAVING SUM(S.BYTES) > 1024 * 1024),
tmp2 as
(SELECT index_owner,
index_name,
plan_operation,
(SELECT min(to_char(nb.begin_interval_time, 'YYYY-MM-DD HH24:MI:SS'))
FROM dba_hist_snapshot nb
where nb.snap_id = v.min_snap_id) min_date,
(SELECT max(to_char(nb.end_interval_time, 'YYYY-MM-DD HH24:MI:SS'))
FROM dba_hist_snapshot nb
where nb.snap_id = v.max_snap_id) max_date,
counts
FROM (SELECT d.object_owner index_owner,
d.object_name index_name,
d.operation || ' ' || d.options plan_operation,
min(h.snap_id) min_snap_id,
max(h.snap_id) max_snap_id,
COUNT(1) counts
FROM dba_hist_sql_plan d, dba_hist_sqlstat h
WHERE /*d.object_owner = 'FUNDZ'
AND */
d.operation LIKE '%INDEX%'
AND d.sql_id = h.sql_id
GROUP BY d.object_owner, d.object_name, d.operation, d.options) v)
SELECT a.table_owner,
a.TABLE_NAME,
a.index_owner,
a.index_name,
a.created,
a.INDEX_TYPE,
a.INDEX_MB,
b.plan_operation,
min_date,
max_date,
counts
from tmp1 a
left outer join tmp2 b
on (a.index_owner = b.index_owner and a.index_name = b.index_name);
如上图所示，有一个3.6G大的索引在13号到22号从没使用过，接下来，我们可以继续查询该索引是否联合索引，创建是否合理，分析为何不走该索引，从而判断是否可以删除索引。
另外下边的SQL可以查询出表上列的使用情况：
CREATE OR REPLACE VIEW VW_COLUMN_USAGE_LHR AS
SELECT oo.name owner,
o.name table_name,
c.name column_name,
u.equality_preds,
u.equijoin_preds,
u.nonequijoin_preds,
u.range_preds,
u.like_preds,
u.null_preds,
u.timestamp
FROM sys.col_usage$ u, sys.obj$ o, sys.user$ oo, sys.col$ c
WHERE o.obj# = u.obj#
AND oo.user# = o.owner#
AND c.obj# = u.obj#
AND c.col# = u.intcol#
;
更多相关文章
- 洋洋洒洒一万二千字，彻底讲清楚MySQL的优化原理，看不完先收藏
- 【DB宝39】使用Docker分分钟搭建Zabbix 5.0配置MySQL监控
- 【DB宝36】使用Docker分分钟搭建漂亮的prometheus+grafana监控
- 【DB宝42】MySQL高可用架构MHA+ProxySQL实现读写分离和负载均衡
- 【DB宝41】监控利器PMM的使用--监控MySQL、PG、MongoDB、ProxySQ
- 「开源免费」基于Vue和Quasar的前端SPA项目crudapi后台管理系统
- 可以中文免费使用的Final Shell 和 需要破解Xshell我们应该选择
- 数据分析之Pandas合并操作总结
- Grafana 之 自定义监控板