oracle11g cursor：mutex S导致的load过高的追踪过程详解

时间：2016-05-03 07:10:35 阅读：320 评论：0 收藏：0 [点我收藏+]

标签：

--> 问题背景

同事电话我，说oracle数据库很慢，很卡，cpu99%了。

1，查看问题现场，cpu高单核不高，load高

多核，看到每个核只有20%左右使用了1/5，所以数据库应该没事，没有到故障级别，只是load高已经接近90了，E:\u\azure_cloud\oracle\05.png：

技术分享

2，看io读写比较低，磁盘io不是瓶颈

实时观察，才148，可以看出io性能没有问题，毕竟是ssd磁盘，E:\u\azure_cloud\oracle\04.png：

技术分享

3，查看锁表情况，比较少

select t2.username,t2.sid,t2.serial#,t2.logon_time fromv$locked_object t1,v$session t2 where t1.session_id=t2.sid order byt2.logon_time;执行没有记录。

SQL> select t2.username,t2.sid,t2.serial#,t2.logon_time from v$locked_object t1,v$session t2 where t1.session_id=t2.sid order by t2.logon_time;

no rows selected

SQL>

原blog地址：http://blog.csdn.net/mchdba/article/details/51299062，谢绝转载。

4，查看session等待，没有等待

SQL>

select /*+ rule */ lpad(‘ ‘, decode(t1.xidusn,0,3,0)) || t1.oracle_username user_name,o.owner,o.object_name,o.object_type,s.sid,s.serial#,p.spid

from v$locked_object t1,dba_objects o, v$session s,v$process p

where t1.object_id=o.object_id

and t1.session_id=o.object_id

and t1.session_id=s.sid

6 and s.paddr=p.addr;

no rows selected

SQL>

5，查看到有很多inactive进程

# 查看inactive进程，kill掉这些进程

SELECT SID, SERIAL#,MODULE, STATUS

FROM V$SESSION S

WHERE S.USERNAME IS NOT NULL

AND UPPER(S.PROGRAM) IN (‘TOAD.EXE‘, ‘W3WP.EXE‘)

AND S.LAST_CALL_ET >= 60*60*2

AND S.STATUS = ‘INACTIVE‘

ORDER BY SID DESC;

# 通过如下sql获取kill命令：

SELECT ‘alter system kill session ‘‘‘||SID||‘,‘|| SERIAL#||‘‘‘;‘ FROM V$SESSION S WHERE S.USERNAME IS NOT NULL AND S.LAST_CALL_ET >= 10 AND S.STATUS = ‘INACTIVE‘ ORDER BY S.MACHINE,SID DESC;

Kill掉‘INACTIVE‘ 进程后，db的load还是没有降下来，证明资源没有被释放掉。这么看来，根源在正在运行的sql，系统load 高 io低，说明系统cpu 耗费高，有大量的逻辑运算而且可能会有wait等待。

6，去查看wait_class等待，有大量cursor:mutex S事件

# 查询等待sql语句

SELECT t1.sid,

t1.serial#,

t1.username,

t1.event,

t2.sql_text,

t1.sql_id,

t1.seconds_in_wait,

t1.program,

t1.blocking_session,

-- object_type || ‘: ‘ || object_name object,

t1.wait_class,

t1.state

FROM v$session t1

LEFT OUTER JOIN v$sqlarea t2

on (t1.sql_id = t2.sql_id)

--LEFT OUTER JOIN dba_objects t3

--ON (object_id = row_wait_obj#)

where t1.wait_class <> ‘Idle‘;

7，查看awr报告，也指向了cursor:mutex S

E:\u\azure_cloud\oracle\06.png

技术分享

E:\u\azure_cloud\oracle\01.png

技术分享

问题基本可以推测出来是cursor等待造成的

8，验证cursor_sharing参数为EXACT，正常

SQL> show parameter cursor_sharing;

NAME TYPE VALUE

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

cursor_sharing string EXACT

SQL>

9，查看4月28日今天的记录数是35W多，而且这个执行次数还在不停的增加中

select sql_id,count(*)

from dba_hist_active_sess_history

where event like‘cursor: mutex S‘and sample_time>to_date(‘28-04-201600:00:00‘, ‘dd-mm-yyyyhh24:mi:ss‘)

groupby sql_id;

SQL>

select sql_id,count(*)

from dba_hist_active_sess_history

where event like ‘cursor: mutex S‘ and sample_time >to_date(‘28-04-2016 00:00:00‘, ‘dd-mm-yyyy hh24:mi:ss‘)

4 group by sql_id;

SQL_ID COUNT(*)

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

94508

8n23w37kwdn3p 1

08nsmbsvfkzv6 353543

SQL>

10，查看昨天4月27日的记录数，才执行了3W多

select sql_id,count(*)

from dba_hist_active_sess_history

where event like‘cursor: mutex S‘and sample_time<to_date(‘28-04-201600:00:00‘, ‘dd-mm-yyyyhh24:mi:ss‘) and sample_time>to_date(‘27-04-201600:00:00‘, ‘dd-mm-yyyyhh24:mi:ss‘)

groupby sql_id;

SQL> select snap_id,ELAPSED_TIME_DELTA/EXECUTIONS_DELTA,CPU_TIME_DELTA/EXECUTIONS_DELTA,EXECUTIONS_DELTA from dba_hist_sqlstat where sql_id=‘08nsmbsvfkzv6‘ order by snap_id;

no rows selected

select sql_id,count(*)

from dba_hist_active_sess_history

where event like ‘cursor: mutex S‘ and sample_time <to_date(‘28-04-2016 00:00:00‘, ‘dd-mm-yyyy hh24:mi:ss‘) and sample_time >to_date(‘27-04-2016 00:00:00‘, ‘dd-mm-yyyy hh24:mi:ss‘)

4 group by sql_id;

SQL_ID COUNT(*)

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

11653

08nsmbsvfkzv6 38468

SQL>

而昨天db数据库没有这么慢，这么卡，那么已经基本确定了是由于这今天爆发增长的几十万sql语句引起的cursor异常wait事件，接下来去找到这个具体的sql记录

11，通过sql_id（08nsmbsvfkzv6）查询到具体的sql记录

select sql_text from dba_hist_sqltext where sql_id=‘08nsmbsvfkzv6‘;

Sql太长了，截图截取不过来，贴出来如下：

select * from ( select * from (

(select (case when (instr(‘,‘||wm_concat(type)||‘,‘,‘,develop-plan,‘)>0) then ‘develop-plan‘ else ‘unRead‘ end) type,

max(id) id,name,node_level,project,end,max(created_date) created_date,

center,year_flg,sequence,

min(isRead) isRead,

MyReplace(wm_concat(msgCreateDate),‘,‘) msgCreateDate,wm_concat(group_name),max(planId),0 owe_days,0 owe_money

from (select

‘develop-plan‘ as type,

node.node_id as id,

template_node.node_name as name,

template_node.node_level,

project.project_name as project,

node.schedule_end_date as end,

warning.CREATED_DATE as created_date,

‘‘ as center , ‘‘ as year_flg, ROW_NUMBER() OVER(PARTITION BY node.templet_id, project.project_project_id ORDER BY warning.created_date ASC) as sequence,

(select min(t.is_read) from plan6_message_user_read t where t.node_id=node.node_id and t.need_reader=‘baolm1‘ group by t.node_id) isRead, ‘‘ msgCreateDate,info.name group_name,info.biz_cd planId,0 owe_days,0 owe_money from plan6_warning warning,

plan6_node node,

project_distribution info,project_period pp,

project_project project,

plan6_templet_node template_node

where warning.node_id = node.node_id

and node.plan_id = info.biz_cd

and info.project_period_id = pp.project_period_id and pp.is_enabled=1 and pp.project_project_id = project.project_project_id

and template_node.node_id = node.templet_id

and info.if_in_plan = 1

and (warning.charger_cd = ‘baolm1‘ OR node.charger_cd=‘baolm1‘ OR node.center_manager_cd=‘baolm1‘ OR project.project_charger_cd=‘baolm1‘ OR project.project_charger_cd2=‘baolm1‘ OR project.project_charger_cd_bis=‘baolm1‘ OR (‘wubc‘=‘baolm1‘ AND template_node.node_level in(1,2) ) OR ((‘xuhf‘=‘baolm1‘ OR ‘zhanghf‘=‘baolm1‘ ) and template_node.node_level=1) ) and warning.stat = 2

and node.status = 1 and node.is_enabled<>0 and warning.stat = 2

union

select ‘unRead‘ tyep,node.node_id id,template_node.node_name as name,

template_node.node_level,project.project_name as project,

node.schedule_end_date as end,

to_date(‘‘,‘yyyymmdd hh24:mi:ss‘) as created_date,

‘‘ as center ,‘‘ as year_flg,

ROW_NUMBER() OVER(PARTITION BY node.templet_id,

project.project_project_id ORDER BY r.created_date ASC) as sequence,

r.is_read isRead,to_char(r.created_date,‘yyyymmdd hh24:mi:ss‘) msgCreateDate,info.name group_name,info.biz_cd planId,0 owe_days,0 owe_money

from plan6_message_user_read r,plan6_node node,

project_distribution info,project_period pp,project_project project,

plan6_templet_node template_node

where r.node_id = node.node_id

and node.plan_id = info.biz_cd

and info.project_period_id = pp.project_period_id and pp.is_enabled=1 and pp.project_project_id = project.project_project_id

and template_node.node_id = node.templet_id

and info.if_in_plan = 1

and node.is_enabled<>0

and r.need_reader=‘baolm1‘

and r.is_read=‘0‘ ) warning

where warning.sequence = 1 group by name,node_level,project,end,center,year_flg,sequence )

union (select ‘oa_meeting‘ as type,

metting.oa_meeting_id as id,

TO_CHAR (metting.business) as name,

10 as node_level,

‘‘ as project,

metting.target_date as end,

warning.created_date as created_date,

‘‘ as center, ‘‘ as year_flg, 0 as sequence,‘‘ isRead,‘‘ msgCreateDate,‘‘ group_name,‘‘ planId,0 owe_days,0 owe_money

from oa_meeting metting, plan_warning warning

where metting.oa_meeting_id = warning.target

and metting.HIDDEN_FLG = 0

and metting.status in (1,2, 4)

and metting.target_date is not null and instr(‘;‘||warning.responsible_person ,‘;baolm1;‘)>0 ) union

(select ‘work-center-plan‘ as type, cen.plan_work_center_id as id, cen.content as name, 10 as node_level, ‘‘ as project, cen.target_date as end, warning.created_date as created_date, ‘‘ center , ‘‘ year_flg, 0 as sequence,‘‘ isRead,‘‘ msgCreateDate,‘‘ group_name,‘‘ planId ,0 owe_days,0 owe_money from plan_work_center cen,plan_warning warning where cen.plan_work_center_id = warning.target and cen.status_cd in (1) and cen.open_type = ‘0‘ and cen.target_date is not null and instr(‘;‘||cen.principal ,‘;baolm1;‘)>0) union (select ‘project_detail‘ as type, pd.project_detail_id as id, pd.content as name,12 as node_level, ‘‘ as project, pd.target_date as end, pd.created_date as created_date,‘‘ as center , ‘‘ as year_flag, 0 as sequence, ‘‘ isRead, ‘‘ msgCreateDate, ‘‘ group_name,‘‘ planId,0 owe_days,0 owe_money from project_detail pd where pd.status in (‘1‘,‘2‘,‘4‘) and pd.project_detail_id in (select pp.project_detail_id from project_principal pp where pp.principal_cd=‘baolm1‘))) warning order by (case when type=‘work-zc-plan‘ then -1 else 1 end) asc, (case when type=‘unRead‘ then 0 else 1 end) asc, (case when isRead=0 then 0 else 1 end) asc, (case when msgCreateDate is not null then 0 else 1 end) asc, warning.type asc,(case when warning.node_level=0 then 4 else warning.node_level end) asc, trunc(warning.end)-trunc(sysdate) asc,warning.created_date desc,owe_money desc ) where rownum <= :1

By the way，基本可以肯定是这个sql引起的，那么除了这个sql是否还有别的记录呢，这需要现场实时观察。