报ORA-4031错误时,我们通常可以根据Oracle无法分配多少字节的内存,来判断共享池碎片的严重程度,以下是4031错误官方的解释:
[oracle@guoyj ~]$ oerr ORA 4031
04031, 00000, "unable to allocate %s bytes of shared memory (\"%s\",\"%s\",\"%s\",\"%s\")"
// *Cause: More shared memory is needed than was allocated in the shared
// pool or Streams pool.
// *Action: If the shared pool is out of memory, either use the
// DBMS_SHARED_POOL package to pin large packages,
// reduce your use of shared memory, or increase the amount of
// available shared memory by increasing the value of the
// initialization parameters SHARED_POOL_RESERVED_SIZE and
// SHARED_POOL_SIZE.
// If the large pool is out of memory, increase the initialization
// parameter LARGE_POOL_SIZE.
// If the error is issued from an Oracle Streams or XStream process,
// increase the initialization parameter STREAMS_POOL_SIZE or increase
// the capture or apply parameter MAX_SGA_SIZE.
一、重现ORA-04031错误:
1、第一个实验硬解析产生大量的碎片重现4031错误
2、当前的数据库版本
SQL> select * from v$version where rownum=1;
BANNER
Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
3、关闭ASMM,设置shared pool大小为120M
SQL> show parameter mem
NAME TYPE VALUE
hi_shared_memory_address integer 0
memory_max_target big integer 804M
memory_target big integer 804M
shared_memory_address integer 0
SQL> show parameter sga
NAME TYPE VALUE
lock_sga boolean FALSE
pre_page_sga boolean FALSE
sga_max_size big integer 804M
sga_target big integer 0
SQL>alter system set memory_target=0; --11g的新特性调整(SGA+PGA)
SQL>startup force;
SQL> alter system set sga_target=0;
SQL>startup force;
SQL> show parameter memory_target
NAME TYPE VALUE
memory_target big integer 0
SQL> show parameter sga_target
NAME TYPE VALUE
sga_target big integer 0
SQL> alter system set shared_pool_size=120M;
System altered.
4、创建表并插入数据
create table t1(id int,name varchar2(100));
begin
for i in 1 .. 100000 loop
insert into t1 values(i,'gyj'||i);
commit;
end loop;
end;
/
5、跑下面匿名块模拟产生大量的硬解析
declare
msql varchar2(500);
mcur number;
mstat number;
jg varchar2(4000);
cg number;
begin
mcur:=dbms_sql.open_cursor;
for i in 1..999999999 loop
msql:='select id from gyj.t1 where id='||to_char(i);
dbms_sql.parse(mcur,msql,dbms_sql.native);
dbms_sql.define_column(mcur,1,jg,4000);
mstat:=dbms_sql.execute(mcur);
end loop;
dbms_sql.close_cursor(mcur);
end;
/
6、这个实验可能要等好久才能看到4031错误,硬解析会占用共享池,但执行完后游标马上关闭,共享池内存也会随之释放,所以用这个实验来模拟4031错误不一定能成功,我也是偶尔做成,下面再来个实验,保证马上重现4031错误。。
8、第二个实验:一直打开游标不关闭,直到消耗完共享池导致4031错误
SQL> show parameter open_cursor
NAME TYPE VALUE
open_cursors integer 300
alter system set open_cursors=65535;
SQL> declare
2 msql varchar2(500);
3 mcur number;
4 mstat number;
5 jg varchar2(4000);
6 cg number;
7 begin
8 for i in 1..65535 loop
9 mcur:=dbms_sql.open_cursor;
10 msql:='select id from t1 where id='||to_char(i);
11 dbms_sql.parse(mcur,msql,dbms_sql.native);
12 dbms_sql.define_column(mcur,1,jg,4000);
13 mstat:=dbms_sql.execute(mcur);
14 dbms_sql.column_value(mcur,1,jg);
15 end loop;
16 end;
17 /
declare
*
ERROR at line 1:
ORA-01000: maximum open cursors exceeded
ORA-06512: at "SYS.DBMS_SQL", line 1199
ORA-06512: at line 11
如果报这个错误,说明alter system set open_cursors=65535;设置没生效,重启库
SQL> conn / as sysdba
Connected.
SQL> startup force;
ORACLE instance started.
Total System Global Area 455163904 bytes
Fixed Size 2229224 bytes
Variable Size 138415128 bytes
Database Buffers 310378496 bytes
Redo Buffers 4141056 bytes
Database mounted.
Database opened.
SQL> conn gyj/gyj
Connected.
再次执行
呵呵,耐心等待五分钟之内绝对报错。。。。
SQL> declare
2 msql varchar2(500);
3 mcur number;
4 mstat number;
5 jg varchar2(4000);
6 cg number;
7 begin
8 for i in 1..65535 loop
9 mcur:=dbms_sql.open_cursor;
10 msql:='select id from t1 where id='||to_char(i);
11 dbms_sql.parse(mcur,msql,dbms_sql.native);
12 dbms_sql.define_column(mcur,1,jg,4000);
13 mstat:=dbms_sql.execute(mcur);
14 dbms_sql.column_value(mcur,1,jg);
15 end loop;
16 end;
17 /
declare
*
ERROR at line 1:
ORA-04031: unable to allocate 80 bytes of shared memory ("shared pool","select job, nvl2(last_date, ...","SQLA^337fc737","ctxPlanSig:qksctxPlanSigGet")
ORA-06512: at "SYS.DBMS_SQL", line 1199
ORA-06512: at line 11
9、查看ORA-4031发生了几次
SQL> select indx,kghlurcr,kghlutrn,kghlufsh,kghluops,kghlunfu,kghlunfs
2 from sys.x$kghlu where inst_id = userenv('Instance');
select indx,kghlurcr,kghlutrn,kghlufsh,kghluops,kghlunfu,kghlunfs
*
ERROR at line 1:
ORA-00604: error occurred at recursive SQL level 1
ORA-04031: unable to allocate 32 bytes of shared memory ("shared pool","select /+ rule / bucket_cn...","SQLA","tmp")
现在不能做任何大的查询操作了,怎么办?
先退出回话,重新登录:
SQL> exit
Disconnected from Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
[oracle@guoyj ~]$ sqlplus / as sysdba
SQL*Plus: Release 11.2.0.3.0 Production on Wed Dec 12 20:50:21 2012
Copyright (c) 1982, 2011, Oracle. All rights reserved.
Connected to:
Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
SQL> select indx,kghlurcr,kghlutrn,kghlufsh,kghluops,kghlunfu,kghlunfs
2 from sys.x$kghlu where inst_id = userenv('Instance');
INDX KGHLURCR KGHLUTRN KGHLUFSH KGHLUOPS KGHLUNFU KGHLUNFS
0 1233 30360 57700 518988 114 4096
kghlunfu:出现ORA-4031的次数,这里指出现了114次的4031错误
kghlunfs:最后一次出现ORA-4031时,申请的分配大小,这里是指最后申请共享池的大小是4096字节的chunk找不到,最终报错。。。。
实际你可以监控制告警日志alert_bxocp.log,里面会显示114次4031错误,为什么报这么多次错误呢,从不同的Free Lists里找空闲的chunk,最后到保留池(SHARED_POOL_RESERVED_SIZE)找,也没找到4096字节的chunk,最后在执 行sql的地方报错4031错误。
10、这里说到了free lists和SHARED_POOL_RESERVED_SIZE
首先说一下保留池,保留池的大小有个参数就是SHARED_POOL_RESERVED_SIZE
关于保留池的隐含参数:
SQL> @?/rdbms/admin/show_para
Enter value for p: SHARED_POOL_RESERVED
old 12: AND upper(i.ksppinm) LIKE upper('%&p%')
new 12: AND upper(i.ksppinm) LIKE upper('%SHARED_POOL_RESERVED%')
P_NAME P_DESCRIPTION P_VALUE ISDEFAULT ISMODIFIED ISADJ
_shared_pool_reserved_min_alloc minimum allocation size in bytes for reserved area 4400 TRUE FALSE FALSE
of shared pool
_shared_pool_reserved_pct percentage memory of the shared pool allocated for 5 TRUE FALSE FALSE
the reserved area
shared_pool_reserved_size size in bytes of reserved area of shared pool 6291456 TRUE FALSE FALSE
SHARED_POOL_RESERVED_SIZE,缺省是SHARED_POOL_SIZE的5%
_SHARED_POOL_RESERVED_PCT缺省 5%
_SHARED_POOL_RESERVED_MIN_ALLOC,缺省4400,超过这个大小才被认为是大对象
从_SHARED_POOL_RESERVED_MIN_ALLOC=4400字节看出申请4096字节的chunk不会到保留池去找,只有符合申请>=4400字节的chunk才会去保留池上找空闲chunk,申请4096字节只会在Free List去找,找不到就最后报错4031错误。
那么我们来看一下共享池的Free Lists管理free内存块(Chunk),把Free Lists链的内容dump出来看一下:
alter session set events 'immediate trace name heapdump level 2';
FREE LISTS:
Bucket 0 size=32
Chunk 074dca368 sz= 32 free " "
Chunk 07b000078 sz= 0 kghdsx
Bucket 1 size=40
Chunk 06e37cf50 sz= 40 free " "
Chunk 06de531b8 sz= 40 free " "
Chunk 071e9a318 sz= 40 free " "
Chunk 06f24c408 sz= 40 free " "
.省略...
.
.
Bucket 40 size=352
.省略...
.
.
Bucket 48 size=416
Chunk 0753eac30 sz= 416 free " "
Chunk 070da9858 sz= 416 free " "
Bucket 49 size=424
Chunk 070bf3000 sz= 424 free " "
Chunk 074197e10 sz= 424 free " "
Bucket 50 size=432
Bucket 51 size=440
Bucket 52 size=448
Chunk 074d45708 sz= 448 free " "
Bucket 53 size=456
Chunk 06e6b1e20 sz= 456 free " "
Chunk 06de86a80 sz= 456 free " "
Chunk 06e344648 sz= 456 free " "
Chunk 06e430e70 sz= 456 free " "
Bucket 54 size=464
Bucket 55 size=472
Bucket 56 size=480
Bucket 57 size=488
Chunk 06be70bc0 sz= 488 free " "
Bucket 58 size=496
Bucket 59 size=504
Chunk 06e405f48 sz= 504 free " "
Bucket 60 size=512
Bucket 61 size=520
Bucket 62 size=528
Bucket 63 size=536
Bucket 64 size=544
.省略...
.
.
Bucket 102 size=848
Bucket 103 size=856
Bucket 104 size=864
Bucket 105 size=872
.省略...
.
.
Bucket 252 size=16408
Bucket 253 size=32792
Bucket 254 size=65560
可以看出有一共有255个Free List链。。。。
每条Free List挂的空闲chunk大小具有一定的范围,从小到大。。。,结合我们上面的实验,产生114次的ORA-04031错误,是从这254链的某条Free List链开始找空闲chunk,直到找到最后一条为止,没找到4096字节的chunk。
二、4031错误产生原因
当尝试在共享池分配大块的连续内存失败(很多时候是由于碎片过多,而并非真是内存不足)时,Oracle首先清除共享池中当前没使用的所有对象,使空闲内存块合并.如果仍然没有足够大的单块内存可以满足需要,就会产生ORA-04031错误.
1、Oracle的bug
2、没使用绑定变量,sql没有足够的共享,产生大量不必要的解析
3、Shared pool设的太小
4、SubPool的划分可能也会导致各分区之间的协调问题,但从Oracle10g开始,允许内存请求在不同的SubPool之间进行切换,提高进请求成功的可能,但不能切换无限次,所以问题仍然存在)
5、open_cursors(0 to 65535) 设的过大
三、解决:4031错误的方法
1、刷共享池
alter system flush shared_pool;
不过这个办法只是暂时的,治标不治本。
2、保留池的设置
参数shared_pool_reserved_size,shared_pool_reserved_min_alloc
select REQUEST_MISSES from v$shared_pool_reserved;
3、缓存大对象
select * from v$db_object_cache where sharable_mem > 50000
and (type = 'PACKAGE' or type='PACKAGE BODY' or type = 'FUNCTION' or type='PROCEDURE')
and kept = 'NO';
执行dbms_shared_pool.keep('对象名');
DBMS_SHARED_POOL
@?/rdbms/admin/dbmspool.sql
4、参数open_cursor、session_cached_cursors的设置
5、共享SQL:应用绑定变量或修改参数cursor_sharing
a.使用绑定变量
declare v_sql varchar2(100);
begin for i in 1..30000 loop
v_sql := 'insert /bind/ into gyj.t1(id) values (:1)';
execute immediate v_sql using i;
end loop;
commit;
end;
/
b.cursor_sharing参数
如何找出不能共享cursor的sql
在v$sql查找执行次数较小的sql语句,观察这些sql语句是否是经常执行的。
select SQL_FULLTEXT from v$sql where EXECUTIONS=1 and sql_text like '%from gyj.t1%';
select SQL_FULLTEXT from v$sql where EXECUTIONS=1 order by sql_text;
alter system set cursor_sharing=SIMILAR | FORCE;
6、增加shared pool空间
select sum(bytes/1024/1024) from v$sgastat where pool='shared pool';
select COMPONENT,CURRENT_SIZE from V$SGA_DYNAMIC_COMPONENTS;
show parameter sga_target
show parameter sga_max_size
alter system set shared_pool_size=300M scope=both;
五、查共享池碎片的一个脚本
SQL> set linesize 1000
SQL> col sga_heap format a15
SQL> col size format a10
SQL> select KSMCHIDX "SubPool", 'sga heap('||KSMCHIDX||',0)'sga_heap,ksmchcom ChunkComment,
2 decode(round(ksmchsiz/1000),0,'0-1K', 1,'1-2K', 2,'2-3K',3,'3-4K',
3 4,'4-5K',5,'5-6k',6,'6-7k',7,'7-8k',8,
4 '8-9k', 9,'9-10k','> 10K') "size",
5 count(*),ksmchcls Status, sum(ksmchsiz) Bytes
6 from x$ksmsp
7 where KSMCHCOM = 'free memory'
8 group by ksmchidx, ksmchcls,
9 'sga heap('||KSMCHIDX||',0)',ksmchcom, ksmchcls,decode(round(ksmchsiz/1000),0,'0-1K',
10 1,'1-2K', 2,'2-3K', 3,'3-4K',4,'4-5K',5,'5-6k',6,
11 '6-7k',7,'7-8k',8,'8-9k', 9,'9-10k','> 10K');
SubPool SGA_HEAP CHUNKCOMMENT size COUNT(*) STATUS BYTES
1 sga heap(1,0) free memory > 10K 94 R-free 19837104
1 sga heap(1,0) free memory 3-4K 54 free 144592
1 sga heap(1,0) free memory 1-2K 1 R-free 680
1 sga heap(1,0) free memory 3-4K 1 R-free 3344
1 sga heap(1,0) free memory 4-5K 151 free 606936
1 sga heap(1,0) free memory 8-9k 3 free 24576
1 sga heap(1,0) free memory 4-5K 2 R-free 7336
1 sga heap(1,0) free memory 0-1K 17490 free 2029992
1 sga heap(1,0) free memory 1-2K 11 free 5944
9 rows selected.