背景

我们知道

2,pmon工作内容如下：

后经查询官方手册发现，PMON的工作内容如下：

1，监控后台进程运行状况

2，如果某些进程异常中断，PMON去释放会话资源以及占用的锁LOCK

3，更新事务表的标志以及清除事务XID的标记

4，清除异常中断会话在BUFFER CACHE占用的缓存

5，PMON也负责定期把数据库实例注册到监听器中

结论

1,10246 pmon event的命令：

alter system set events='10246 trace name context forever,level 1';--开启

alter system set events='10246 trace name context off';--关闭

2,可以从10246 PMON TRACE FILE发现PMON在清理中断会话占用的事务表及锁资源信息

3，当然比如像背景中，说其它几点没有看到，我估计是级别不够

4,level 12可以查盾清理占用的latch的情况

5，至于到底不同级别的区别，大家进行对比测试即可，不再复述

测试

1,数据库版本

SQL> select * from v$version where rownum=1;

BANNER

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

Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit Production

2，测试会话

SQL> select spid from v$process where addr=(select paddr from v$session where sid=(select sid from v$mystat where rownum=1));

SPID

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

9534

SQL> update t_pmon set a=3,b=3;

--注意：运行一段时间

3，手工杀死测试会话对应的操作系统进程

[oracle@seconary ~]$ ps -ef|grep 9534

oracle    9534  1001  1 04:44 ?        00:00:03 oracleguowang (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))

oracle   10416  1311  0 04:47 pts/2    00:00:00 grep 9534

[oracle@seconary ~]$ kill -9 9534

4，10246跟踪的PMON TRACE FILE

*** 2015-11-02 04:51:02.339

marked process 0xdd5bc340 pid=30 serial=229 ospid = 9534 dead --标记9534进程死亡

client details:

O/S info: user: oracle, term: pts/4, ospid: 1001

machine: seconary program: sqlplus@seconary (TNS V1-V3)

application name: SQL*Plus, hash value=3669949024

*** 2015-11-02 04:51:02.427

deleting process 0xdd5bc340 pid=30 serial=229 priority=0  --清理对应的操作系统资源

deleting session 0xdc8b3f10 sid=271 serial=7089

*** 2015-11-02 04:51:04.040

deleting session 0xdcafb500 sid=70 serial=7394

*** 2015-11-02 04:51:05.664

ksuprog() called at ktur.c:3039  --调用ORACLE内核参数

ksuprog() called at ktur.c:3039

ksuprog() called at ktur.c:3039

中间略

ksuprog() called at ktur.c:3039

ksuprog() called at ktur.c:3039

ksuprog() called at ktur.c:3039

ksuprog() called at ktur.c:3039

*** 2015-11-02 04:51:06.898

CLEANUP_ROLLBACK_ENTRIES=100 rollback entries processed  --清理测试会话占用的UNDO资源，即占用的回滚段资源

*** 2015-11-02 04:51:09.847

Audit postponed pending cleanup for session 0xdcafb500

deleting session 0xdcafb500 sid=70 serial=7395

Audit postponed pending cleanup for session 0xdcafb500

deletion of process dd5bc340 pid=30 seq=229 prog=TRUE unsuccessful

*** 2015-11-02 04:51:09.967

found process 0xdd5bc340 pid=30 serial=229 ospid = 9534 dead  --9534进程死亡

*** 2015-11-02 04:51:10.016

deleting process 0xdd5bc340 pid=30 serial=229 priority=0

deleting session 0xdcafb500 sid=70 serial=7396

5，仍看不到好多PMON的工作内容，如：

4，清除异常中断会话在BUFFER CACHE占用的缓存

所以我们加大跟踪级别，看有无变化

SQL> alter system set events='10246 trace name context forever,level 12';

System altered.

SQL> alter system set events='10246 trace name context off';

System altered.

可以明看到清除锁及LATCH的资源占用

marked process 0xdd5be3c0 pid=32 serial=253 ospid = 20301 dead

client details:

O/S info: user: oracle, term: pts/4, ospid: 1001

machine: seconary program: sqlplus@seconary (TNS V1-V3)

application name: SQL*Plus, hash value=3669949024

*** 2015-11-02 05:21:43.591

cleaning latches proc=dd5be3c0 pid=32 seq=253 --清理LATCH

Latch Cleanup: kcbzch - op=0, lt=0xda2c5c78, b=0xb7fb0728, nb=0x1, bp=0xda6cd580

...complete

所以表明不同的级别跟踪PMON，产生的内容是不同的

至于到底不同级别的区别，大家进行对比测试即可，不再复述

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/9240380/viewspace-1823317/，如需转载，请注明出处，否则将追究法律责任。