记一次Oracle Sql优化经历--消耗过多CPU

 

同事打电话跟我说,数据库CPU过高、swap交换频繁,要我马上看看,这里记录下整个过程以供大家参考,也让大家提点意见

$topas

Topas Monitor for host:    fjlt_wb_db01         EVENTS/QUEUES    FILE/TTY

Mon Feb 13 10:10:09 2012   Interval:  2         Cswitch   13932  Readch  2718.1K

                                                Syscall  344.8K  Writech  162.7K

CPU  User%  Kern%  Wait%  Idle%                  Reads      1231  Rawin         0

ALL   89.1   10.9    0.0    0.0                  Writes      861  Ttyout      725

                                                 Forks         6  Igets         0

Network  KBPS   I-Pack  O-Pack   KB-In  KB-Out  Execs         6  Namei       414

Total    97.6K    24.9K   51.5K   30.8K   66.9K Runqueue   43.0  Dirblk        0

                                                 Waitqueue   0.0

Disk    Busy%     KBPS     TPS KB-Read KB-Writ                   MEMORY

Total    13.0     22.6K 2715.0    21.7K  932.5  PAGING           Real,MB   31744

                                                Faults     4337  % Comp     85

FileSystem        KBPS     TPS KB-Read KB-Writ  Steals        0  % Noncomp   4

Total            638.7   398.5  637.4    1.3    PgspIn        0  % Client    4

                                                PgspOut       0

Name            PID   CPU%  PgSp Owner          PageIn         0   PAGING SPACE

oracle     66847032  17.3  10.4 oracle           PageOut       1   Size,MB   32768

oracle     33751546   4.3  10.7 oracle            Sios          1   % Used     39

oracle     37093606   2.6  10.6 oracle                              % Free     61

oracle     51577090   2.4  10.4 oracle          NFS (calls/sec)

oracle     60752382   2.4  10.2 oracle          SerV2         0  WPAR Activ    0

oracle      2425184   2.3  10.5 oracle          CliV2         0  WPAR Total    0

oracle     38535516   2.1  10.6 oracle          SerV3         0  Press: "h"-help

oracle     65404954   2.0  10.5 oracle          CliV3         0         "q"-quit

oracle     40239486   2.0  10.3 oracle

oracle     65208590   2.0  11.6 oracle

oracle     60555628   1.9  10.5 oracle

oracle     23658656   1.9  10.3 oracle

oracle     47841658   1.9  10.5 oracle

oracle     52363552   1.8  10.3 oracle

oracle      2359684   1.8  10.5 oracle

oracle     31916154   1.7  10.6 oracle

oracle     46530580   1.7  10.9 oracle

oracle       721382   1.6  11.7 oracle

oracle      2884060   1.6  10.5 oracle

oracle      5636414   1.6  10.5 oracle

  发现 CPU 使用率为 100% , swap 分区交换频繁,其中进程号为 66847032 的 oracle 用户进程的 CPU 使用率为 17.3% 。

kthr    memory              page              faults        cpu  
----- ----------- ------------------------ ------------ -----------
 r  b   avm   fre  re  pi  po  fr   sr  cy  in   sy  cs us sy id wa
 2  0 7506815 658832   0   0   0   0    0   0 6913 154614 20053 31 12 58  0
 5  0 7506576 659071   0   0   0   0    0   0 5763 193844 14268 53 11 36  0
 5  0 7506767 658880   0   0   0   0    0   0 8699 157452 18035 43 11 45  0
 2  0 7506557 659089   0   0   0   0    0   0 6392 135539 20697 27  9 64  0
 2  0 7506286 659359   0   0   0   0    0   0 6518 126193 19194 21 10 69  0

pi 、 po 均为 0 ,说明 swap 分区并未频繁交换。问题定位为 CPU 使用率过高
根据当前时间做 ash 报告。发现 CPU 的等待事件占了活动时间的 67.62%

另外看到 sqlid 为 dczhdxppd0fmm 的 sql 占用了 41.4 的活动时间

等了半天没有反应,估计是 sga 过大,导致查询较慢,先放着让他跑跑。

使用 PL\SQL 执行如下sql,大致看下目前系统中的等待事件

select * from v$session_wait where wait_class#<>6;

常规的gc cr multi block request等待事件,没有看到比较异常的等待。老办法,制作ash看看

根据当前时间做ash报告。
SQL> @$ORACLE_HOME/rdbms/admin/ashrpt.sql
Current Instance
~~~~~~~~~~~~~~~~
   DB Id    DB Name      Inst Num Instance
----------- ------------ -------- ------------
 4193910433 WBDB                1 wbdb1

Specify the Report Type
~~~~~~~~~~~~~~~~~~~~~~~
Enter 'html' for an HTML report, or 'text' for plain text
Defaults to 'html'
Enter value for report_type:
Type Specified:  html

Instances in this Workload Repository schema
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   DB Id     Inst Num DB Name      Instance     Host
------------ -------- ------------ ------------ ------------
  4193910433        2 WBDB         wbdb2        fjlt_wb_db02
* 4193910433        1 WBDB         wbdb1        fjlt_wb_db01

Defaults to current database
Using database id: 4193910433
Defaults to current instance
Using instance number: 1
ASH Samples in this Workload Repository schema
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Oldest ASH sample available:  05-Feb-12 23:00:09   [  10995 mins in the past]
Latest ASH sample available:  13-Feb-12 14:14:45   [      0 mins in the past]

Specify the timeframe to generate the ASH report
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Enter begin time for report:

--    Valid input formats:
--      To specify absolute begin time:
--        [MM/DD[/YY]] HH24:MI[:SS]
--        Examples: 02/23/03 14:30:15
--                  02/23 14:30:15
--                  14:30:15
--                  14:30
--      To specify relative begin time: (start with '-' sign)
--        -[HH24:]MI
--        Examples: -1:15  (SYSDATE - 1 Hr 15 Mins)
--                  -25    (SYSDATE - 25 Mins)

Defaults to -15 mins
Enter value for begin_time: 02/13/12 10:00:00
Report begin time specified: 02/13/12 10:00:00

Enter duration in minutes starting from begin time:
Defaults to SYSDATE - begin_time
Press Enter to analyze till current time
Enter value for duration:02/13/12 10:30:00

以下输出省略
Enter value for report_name: wbdb01_ash_20120213_1000_1030
以下输出省略
打开生成好的ash报告
                      

从上图可以看到CPU资源的竞争才是真正的问题。
      

根据上图 sqlid 为 dczhdxppd0fmm 的 sql 占用了41.4 的活动时间

完整 sql 如下

select a.shbxhm, nvl(sum(a.ylbx_jfjs), 0) from t_wb_shbxmxsbb a, t_wb_sbbqk b where a.pz_xh = b.pz_xh and a.swglm = :1 and to_char(a.sfssq_qsrq, 'yyyy-mm-dd') = :2 and to_char(a.sfssq_zzrq, 'yyyy-mm-dd') = :3 and b.zt in ('2', '4', '5', '6') group by a.shbxhm

查看该 sql 的执行计划             

这里需要加上相应的用户

SQL> explain plan for select a.shbxhm, nvl(sum(a.ylbx_jfjs), 0) from ETAX. t_wb_shbxmxsbb a, ETAX .t_wb_sbbqk b where a.pz_xh = b.pz_xh and a.swglm = :1 and to_char(a.sfssq_qsrq, 'yyyy-mm-dd') = :2 and to_char(a.sfssq_zzrq, 'yyyy-mm-dd') = :3 and b.zt in ('2', '4', '5', '6') group by a.shbxhm;

Explained.

 

SQL> select * from table(DBMS_XPLAN.display)

 

PLAN_TABLE_OUTPUT

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Plan hash value: 2119216066

 

------------------------------------------------------------------------------------------------

| Id  | Operation                     | Name           | Rows  | Bytes | Cost (%CPU)| Time     |

------------------------------------------------------------------------------------------------

|   0 | SELECT STATEMENT              |                |     1 |    86 | 35939   (1)| 00:07:12 |

|   1 |  HASH GROUP BY                |                |     1 |    86 | 35939   (1)| 00:07:12 |

|   2 |   NESTED LOOPS                |                 |     1 |    86 | 35938   (1)| 00:07:12 |

|*  3 |    TABLE ACCESS FULL          | T_WB_SHBXMXSBB |     1 |    67 | 35936   (1)| 00:07:12 |

|*  4 |    TABLE ACCESS BY INDEX ROWID| T_WB_SBBQK     |     1 |    19 |     2   (0)| 00:00:01 |

|*  5 |     INDEX UNIQUE SCAN         | PK_T_WB_SBBQK  |     1 |       |     1   (0)| 00:00:01 |

 

PLAN_TABLE_OUTPUT

---------------------------------------------------------------------------------------------------------------------------------

 

Predicate Information (identified by operation id):

---------------------------------------------------

 

   3 - filter("A"."SWGLM"=TO_NUMBER(:1) AND

              TO_CHAR(INTERNAL_FUNCTION("A"."SFSSQ_QSRQ"),'yyyy-mm-dd')=:2 AND

              TO_CHAR(INTERNAL_FUNCTION("A"."SFSSQ_ZZRQ"),'yyyy-mm-dd')=:3)

   4 - filter("B"."ZT"='2' OR "B"."ZT"='4' OR "B"."ZT"='5' OR "B"."ZT"='6')

   5 - access("A"."PZ_XH"="B"."PZ_XH")

 

21 rows selected.

可以看到 T_WB_SHBXMXSBB 并没有走索引,

查看 T_WB_SHBXMXSBB 的数据量

SQL>select count(*) from ETAX.t_wb_shbxmxsbb;

  COUNT(*)

----------

   9495977

执行
select * from dba_ind_columns where upper (table_name)= 'T_WB_SHBXMXSBB' ;

查看表的索引情况如下

发现并未在 sfssq_qsrq sfssq_zzrq 这两列上创建索引 .

解决方案,在 sfssq_qsrq sfssq_zzrq 这两列上创建普通索引或者函数索引,如果网报系统中 to_char(a.sfssq_qsrq, 'yyyy-mm-dd') = :2 的用法较多,那么则应创建函数索引,反之创建普通索引即可

创建普通索引:
SQL> set timing on
SQL> create index ETAX.fbi_2 on ETAX.T_WB_SHBXMXSBB(sfssq_zzrq);
Index created.
Elapsed: 00:00:31.54


SQL> explain plan for select a.shbxhm, nvl(sum(a.ylbx_jfjs), 0) from etax.t_wb_shbxmxsbb a, etax.t_wb_sbbqk b where a.pz_xh = b.pz_xh and a.swglm = :1 and a.sfssq_qsrq >= to_date('2','yyyy-mm-dd') and a.sfssq_qsrq < to_date('3','yyyy-mm-dd') and
  2  a.sfssq_zzrq >= to_date('4','yyyy-mm-dd') and a.sfssq_zzrq < to_date('5','yyyy-mm-dd')  and b.zt in ('2', '6', '7', '8') group by a.shbxhm
  3  ;

Explained.

Elapsed: 00:00:00.03


SQL> set linesize 180
SQL>
select * from table(dbms_xplan.display)

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Plan hash value: 3466029591

------------------------------------------------------------------------------------------------------
| Id  | Operation                           | Name           | Rows  | Bytes | Cost (%CPU)| Time     |
------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT                    |                |     1 |    86 |   308   (2)| 00:00:04 |
|   1 |  HASH GROUP BY                      |                |     1 |    86 |   308   (2)| 00:00:04 |
|*  2 |   FILTER                            |                |       |       |            |          |
|   3 |    NESTED LOOPS                     |                |     1 |    86 |   307   (2)| 00:00:04 |
|*  4 |     TABLE ACCESS BY INDEX ROWID     | T_WB_SHBXMXSBB |     1 |    67 |   305   (2)| 00:00:04 |
|   5 |      BITMAP CONVERSION TO ROWIDS    |                |       |       |            |          |

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|   6 |       BITMAP AND                    |                |       |       |            |          |
|   7 |        BITMAP CONVERSION FROM ROWIDS|                |       |       |            |          |
|   8 |         SORT ORDER BY               |                |       |       |            |          |
|*  9 |          INDEX RANGE SCAN           | FBI            | 43201 |       |   117   (0)| 00:00:02 |
|  10 |        BITMAP CONVERSION FROM ROWIDS|                |       |       |            |          |
|  11 |         SORT ORDER BY               |                |       |       |            |          |
|* 12 |          INDEX RANGE SCAN           | FBI_2          | 43201 |       |   117   (0)| 00:00:02 |
|* 13 |     TABLE ACCESS BY INDEX ROWID     | T_WB_SBBQK     |     1 |    19 |     2   (0)| 00:00:01 |
|* 14 |      INDEX UNIQUE SCAN              | PK_T_WB_SBBQK  |     1 |       |     1   (0)| 00:00:01 |
------------------------------------------------------------------------------------------------------

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------

   2 - filter(TO_DATE('4','yyyy-mm-dd')<TO_DATE('5','yyyy-mm-dd') AND
              TO_DATE('2','yyyy-mm-dd')<TO_DATE('3','yyyy-mm-dd'))
   4 - filter("A"."SWGLM"=TO_NUMBER(:1))
   9 - access("A"."SFSSQ_QSRQ">=TO_DATE('2','yyyy-mm-dd') AND
              "A"."SFSSQ_QSRQ"<TO_DATE('3','yyyy-mm-dd'))
  12 - access("A"."SFSSQ_ZZRQ">=TO_DATE('4','yyyy-mm-dd') AND
              "A"."SFSSQ_ZZRQ"<TO_DATE('5','yyyy-mm-dd'))
  13 - filter("B"."ZT"='2' OR "B"."ZT"='6' OR "B"."ZT"='7' OR "B"."ZT"='8')

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  14 - access("A"."PZ_XH"="B"."PZ_XH")

34 rows selected.
Elapsed: 00:00:00.01

走索引了,时间有所缩短

SQL> drop index etax.fbi ;urge;
Index dropped.
Elapsed: 00:00:00.11


SQL> drop index ETAX.fbi_2;
Index dropped.
Elapsed: 00:00:00.04

并更改该 sql 写法如下

a.sfssq_qsrq >= to_date('2012-01-01','yyyy-mm-dd') and a.sfssq_qsrq < to_date('2012-01-02','yyyy-mm-dd')

创建函数索引:

SQL> create index etax.fbi on ETAX.T_WB_SHBXMXSBB ( to_char(sfssq_qsrq, 'yyyy-mm-dd'));
Index created.

Elapsed: 00:00:34.65
SQL> create index ETAX.fbi_2 on ETAX.T_WB_SHBXMXSBB ( to_char(sfssq_zzrq, 'yyyy-mm-dd'));
Index created.

Elapsed: 00:00:36.66
SQL> explain plan for select a.shbxhm, nvl(sum(a.ylbx_jfjs), 0) from ETAX.t_wb_shbxmxsbb a, ETAX.t_wb_sbbqk b where a.pz_xh = b.pz_xh and a.swglm = :1 and to_char(a.sfssq_qsrq, 'yyyy-mm-dd') = :2 and to_char(a.sfssq_zzrq, 'yyyy-mm-dd') = :3 and b.zt in ('2', '4', '5', '6') group by a.shbxhm;

Explained.

Elapsed: 00:00:00.01
SQL> select * from table(dbms_xplan.display);

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Plan hash value: 4204756945

-----------------------------------------------------------------------------------------------------
| Id  | Operation                          | Name           | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT                   |                |     1 |    86 |  1134   (1)| 00:00:14 |
|   1 |  HASH GROUP BY                     |                |     1 |    86 |  1134   (1)| 00:00:14 |
|   2 |   NESTED LOOPS                     |                |     1 |    86 |  1133   (1)| 00:00:14 |
|*  3 |    TABLE ACCESS BY INDEX ROWID     | T_WB_SHBXMXSBB |     1 |    67 |  1131   (1)| 00:00:14 |
|   4 |     BITMAP CONVERSION TO ROWIDS    |                |       |       |            |          |
|   5 |      BITMAP AND                    |                |       |       |            |          |

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
|   6 |       BITMAP CONVERSION FROM ROWIDS|                |       |       |            |          |
|*  7 |        INDEX RANGE SCAN            | FBI_2          | 38405 |       |   456   (1)| 00:00:06 |
|   8 |       BITMAP CONVERSION FROM ROWIDS|                |       |       |            |          |
|*  9 |        INDEX RANGE SCAN            | FBI            | 38405 |       |   463   (1)| 00:00:06 |
|* 10 |    TABLE ACCESS BY INDEX ROWID     | T_WB_SBBQK     |     1 |    19 |     2   (0)| 00:00:01 |
|* 11 |     INDEX UNIQUE SCAN              | PK_T_WB_SBBQK  |     1 |       |     1   (0)| 00:00:01 |
-----------------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

PLAN_TABLE_OUTPUT
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
   3 - filter("A"."SWGLM"=TO_NUMBER(:1))
   7 - access(TO_CHAR(INTERNAL_FUNCTION("SFSSQ_ZZRQ"),'yyyy-mm-dd')=:3)
   9 - access(TO_CHAR(INTERNAL_FUNCTION("SFSSQ_QSRQ"),'yyyy-mm-dd')=:2)
  10 - filter("B"."ZT"='2' OR "B"."ZT"='4' OR "B"."ZT"='5' OR "B"."ZT"='6')
  11 - access("A"."PZ_XH"="B"."PZ_XH")

27 rows selected.
Elapsed: 00:00:00.01

走索引了,时间有所缩短

SQL> drop index etax.fbi ;urge;
Index dropped.
Elapsed: 00:00:00.11


SQL> drop index ETAX.fbi_2;
Index dropped.
Elapsed: 00:00:00.04

可以看到,创建函数索引和创建普通索引在本质操作本质上对系统并没有太大影响。创建什么样的索引,关键取决于应用系统。


阅读更多
想对作者说点什么?

博主推荐

换一批

没有更多推荐了,返回首页