一. Mutex 说明
Oracle Mutex 机制 说明
To improve cursor execution and also hard parsing, a new memory serialization mechanism has been created in 10gR2.
For certain shared-cursor related operations, mutexes are used as a replacement for library cache latches and librarycache pins.
-- mutexes 替代 library cache latches 和 librarycache pins。
Using mutexes is faster, uses less CPU and also allows significantly improved concurrency over the existing latch mechanism.
The use of mutexes for cursor pins can be enabled by setting the init.ora parameter _use_kks_mutex toTRUE.
Btw, things get more fun in 10.2, you can pin cursors without getting library cache pin latch, using KGX mutexes. Mutexes are new thing in 10.2 and they enable shared access to objects in somewhat similar manner than shared latches, that every successful get of particular mutex will increment its value and release will decrement. When the count is zero, no-one has the mutex and it is safe to get it in exclusive mode too. However they are more fine grained than kgl latches and provide better waiting mechanism as far as I understand.
So if your environment supports atomic compare and swap operation (as CMPXCHG on Intel), you might get away without cursor_space_for_time setting for ultrahigh execution rates. Otherwise the atomic mutex operations would be achieved using new KGX latches.
At least on my laptop this feature isn’t enabled by default (from andOracleWorld’s paper I remember that it should become default in 10.2.0.2), but so far you can experiment with it if you set _kks_use_mutex_pin = true and bounce the instance (mutex structures will be stored in shared pool, so you might need to increase SP size).
There are also x$mutex_sleep and x$mutex_sleep_history fixed tables that can show some interesting information if you generate some mutex waits into them.
在Oracle 10.2中，对shared pool中的一些Serialization operation使用更轻量的 KGX mutexes (_use_kks_mutex) 取代library cache pin，从而降低CPU Usage, 是否使用这种muetx机制受到隐含参数_kks_use_mutex_pin的限制。
从10.2.0.2开始该参数default为true,使用这种机制oracle是为了解决library cache bin latch的串行使用问题,但是mutex貌似还不是很稳定,在很多系统中会出现cursor: pin S wait on X等待事件,这个事件和mutex的使用有关,最近一客户受到cursor: pin S wait on X等待事件的困扰,出现cursor: pin S wait on X等待事件时通常等待比较严重,系统会出现hang。
cursor: pin S wait on X
A session waits for this event when it is requesting a shared mutex pin and another session is holding an exclusive mutex pin on the same cursor object.
Wait Time: Microseconds
Parameter Description
P1 Hash value of cursor
P2 Mutex value (top 2 bytes contains SID holding mutex in exclusive mode, and bottom two bytes usually hold the value 0)
P3 Mutex where (an internal code locator) OR'd with Mutex Sleeps
这个事件的出现受到很多因素的影响，在高并发的情况下:
（1）sga自动管理,sga的频繁扩展和收缩
（2）过渡硬解析,造成library cache中的cursor object被频繁的reload
（3）bug
_kks_use_mutex_pin 是隐含参数，通过v$parameter 视图查不到，需要通过如下SQL 来查看。
SELECT i.ksppinm name,
i.ksppdesc description,
CV.ksppstvl VALUE,
CV.ksppstdf isdefault,
DECODE (BITAND (CV.ksppstvf, 7),
1, 'MODIFIED',
4, 'SYSTEM_MOD',
'FALSE')
ismodified,
DECODE (BITAND (CV.ksppstvf, 2), 2, 'TRUE', 'FALSE') isadjusted
FROM sys.x$ksppi i, sys.x$ksppcv CV
WHERE i.inst_id = USERENV ('Instance')
AND CV.inst_id = USERENV ('Instance')
AND i.indx = CV.indx
AND i.ksppinm LIKE '/_%' ESCAPE '/'
and i.ksppinm like '_kks%'
ORDER BY REPLACE (i.ksppinm, '_', '');
Oracle 参数分类 和 参数的查看方法
二. 相关测试
SYS@anqing2(rac2)> select * from v$version where rownum<2;
BANNER
----------------------------------------------------------------
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - Prod
SESSION 1：
-------------------------
--创建测试表
SYS@anqing2(rac2)> create table t as select * from dba_objects;
Table created.
--查看session ID
SYS@anqing2(rac2)> select sid from v$mystat where rownum=1;
SID
----------
125
SYS@anqing2(rac2)> declare
2 v_string varchar2(100) := 'alter system flush shared_pool';
3 msql varchar2(200);
4 begin
5 loop
6 execute immediate v_string;
7 for i in 1..100 loop
8 msql:='select object_id from t where object_id='||i;
9 execute immediate msql;
10 end loop;
11 end loop;
12 end;
13 /
session 2：
-------------------------
--查看session ID
SYS@anqing2(rac2)> select sid from v$mystat where rownum=1;
SID
----------
130
SYS@anqing2(rac2)> declare
2 v_string varchar2(100) := 'alter system flush shared_pool';
3 msql varchar2(200);
4 begin
5 loop
6 execute immediate v_string;
7 for i in 1..100 loop
8 msql:='select object_id from t where object_id='||i;
9 execute immediate msql;
10 end loop;
11 end loop;
12 end;
13 /
session 3：
------------------------
用如下SQL 进行监控，在sqlplus 里看起来格式有点乱，我放到Toad执行了。
/* Formatted on 2011/6/16 16:06:44 (QP5 v5.163.1008.3004) */
SELECT b.*, sq.sql_text
FROM v$session se,
v$sql sq,
(SELECT a.*, s.sql_text
FROM v$sql s,
(SELECT sid,
event,
wait_class,
p1,
p2raw,
TO_NUMBER (SUBSTR (p2raw, 1, 4), 'xxxx')
sid_hold_mutex_x
FROM v$session_wait
WHERE event LIKE 'cursor%') a
WHERE s.HASH_VALUE = a.p1) b
WHERE se.sid = b.sid AND se.sql_hash_value = sq.hash_value;
通过监控发现两个session在执行相同的sql,他们在相同的cursor object上交互请求a shared mutex pin或者 an exclusive mutex pin 从而造成等待。
--监视sql reae区的cursor object reload情况
SYS@anqing2(rac2)> select namespace ,reloads from v$librarycache;
NAMESPACE RELOADS
--------------- ----------
SQL AREA 790805
TABLE/PROCEDURE 103713
BODY 59
TRIGGER 27
INDEX 94280
CLUSTER 11
OBJECT 0
PIPE 0
JAVA SOURCE 0
JAVA RESOURCE 0
JAVA DATA 0
11 rows selected.
--监视parse情况
SYS@anqing2(rac2)> col name format a40
SYS@anqing2(rac2)> select s.sid, s.serial#,b.name,a.value
2 from v$sesstat a, v$statname b, v$session s
3 where a.statistic# = b.statistic# and s.sid=a.sid
4 and b.name like '%parse%'
5 and s.sid in (130,125);
sid serial# name value
---------- ---------- ---------------------------------------- ----------
125 41915 parse time cpu 115260
125 41915 parse time elapsed 146605
125 41915 parse count (total) 633792
125 41915 parse count (hard) 602732
125 41915 parse count (failures) 4
130 6074 parse time cpu 69559
130 6074 parse time elapsed 99149
130 6074 parse count (total) 394689
130 6074 parse count (hard) 365538
130 6074 parse count (failures) 0
从这里看出，硬解析很多，library cache中的cursor object被频繁的reload。
三. 几个与mutex 相关的视图
在第一部分，提到了x$mutex_sleep 和x$mutex_sleep_history。我们在联机文档里看不到相关的说明。
不过可以查看到v$mutex_sleep 和v$mutex_sleep_history的说明。 但是v$ 比x$ 字典显示的列要少。
select * from x$mutex_sleep;
select * from v$mutex_sleep;
SYS@anqing2(rac2)> desc x$mutex_sleep_history
Name Null? Type
----------------------------------------- -------- ----------------------------
ADDR RAW(4)
INDX NUMBER
INST_ID NUMBER
MUTEX_ADDR RAW(4)
MUTEX_IDENTIFIER NUMBER
SLEEP_TIMESTAMP TIMESTAMP(6)
MUTEX_TYPE VARCHAR2(32)
MUTEX_TYPE_ID NUMBER
GETS NUMBER
SLEEPS NUMBER
REQUESTING_SESSION NUMBER
BLOCKING_SESSION NUMBER
LOCATION_ID NUMBER
LOCATION VARCHAR2(40)
MUTEX_VALUE RAW(4)
P1 NUMBER
P1RAW RAW(4)
P2 NUMBER
P3 NUMBER
P4 NUMBER
P5 VARCHAR2(64)
SYS@anqing2(rac2)> desc v$mutex_sleep_history
Name Null? Type
----------------------------------------- -------- ----------------------------
MUTEX_IDENTIFIER NUMBER
SLEEP_TIMESTAMP TIMESTAMP(6)
MUTEX_TYPE VARCHAR2(32)
GETS NUMBER
SLEEPS NUMBER
REQUESTING_SESSION NUMBER
BLOCKING_SESSION NUMBER
LOCATION VARCHAR2(40)
MUTEX_VALUE RAW(4)
P1 NUMBER
P1RAW RAW(4)
P2 NUMBER
P3 NUMBER
P4 NUMBER
P5 VARCHAR2(64)
