很多情况下，Oracle hang导致sqlplus无法连接，从而无法获得Oracle系统和进程状态，使得定位问题缺少强有力的依据。所幸的是Oracle 10g推出了sqlplus -prelim选项，在Oracle挂起时依然能使用sqlplus，从而能获得数据库状态。
使用方法如下
引用
$ sqlplus -prelim "/as sysdba"
SQL*Plus: Release 10.2.0.4.0 - Production on Sun Mar 28 06:40:21 2010
Copyright (c) 1982, 2007, Oracle. All Rights Reserved.
在prelim模式下，不可以查询数据字典,但可以关闭数据库
引用
SQL> select status from v$instance;
select status from v$instance
*
ERROR at line 1:
ORA-01012: not logged on
引用
SQL> shutdown abort
ORACLE instance shut down.
但可以使用oradebug，但对于系统诊断，已经足够了
dump系统状态
引用
SQL> oradebug setmypid
Statement processed.
SQL> oradebug unlimit
Statement processed.
SQL> oradebug dump systemstate 266
Statement processed.
或者dump hanganalyze
引用
SQL> oradebug hanganalyze 3
Hang Analysis in /oracle/app/oracle/admin/ora10g/udump/ora10g_ora_52642.trc
对于rac
引用
SQLPLUS> oradebug setmypid
SQLPLUS>oradebug setinst all
SQLPLUS>oradebug -g def hanganalyze 3
或者dump 进程状态
引用
SQL> oradebug dump processstate 10
Statement processed.
进一步，如果有10g客户端，数据库是9i，依然可以用-prelim
引用
$ sqlplus -prelim /nolog
SQL*Plus: Release 10.2.0.4.0 - Production on Sun Mar 28 06:50:19 2010
Copyright (c) 1982, 2007, Oracle. All Rights Reserved.
SQL> conn sys/oracle@ora9i as sysdba
Prelim connection established
SQL> oradebug setmypid
Statement processed.
SQL> oradebug unlimit
Statement processed.
SQL> oradebug dump systemstate 266
Statement processed.
There are two ways to connect to sqlplus using a preliminary connection.
1. sqlplus -prelim / as sysdba
2. sqlplus /nolog
set _prelim on
connect / as sysdba
1. While running system state dump with "-prelim" option:
sqlplus -prelim / as sysdba
SQL*Plus: Release 10.2.0.2.0 - Production on Mon Dec 11 17:28:14 2006
SQL> oradebug setmypid
Statement processed.
SQL> oradebug unlimit
Statement processed.
SQL> oradebug dump systemstate 266
ORA-03113: end-of-file on communication channel
ORA-24323: value not allowed
2. The following error is reported in the alert.log:
ORA-07445: exception encountered: core dump [kgllkd()+1254] [SIGSEGV] [Address not mapped to
object] [0x000000030] [] []
The error did not crash background process or terminate the instance.
The issue is not resolved in any known patches.
Use any of the below workarounds:
· Connect / as sysdba (without -prelim)
OR
· Force the variable PGA to be initialized:
sqlplus -prelim '/ as sysdba'
oradebug setmypid
oradebug unlimit
oradebug call ksmcpg
oradebug dump systemstate 10
--测试环境11g,获lgwr的spid
SQL> select * from v$process where username='SYSTEM' and program like '%LGWR%';
ADDR PID SPID PNAME USERNAME SERIAL# TERMINAL PROGRAM TRACEID TRACEFILE BACKGROUND LATCHWAIT LATCHSPIN PGA_USED_MEM PGA_ALLOC_MEM PGA_FREEABLE_MEM PGA_MAX_MEM
---------------- ---------- ------------------------ ----- --------------- ---------- ---------------- ---------------------------------------------------------------- -------------------------------------------------------------------------------- -------------------------------------------------------------------------------- ---------- ---------------- ---------------- ------------ ------------- ---------------- -----------
000007FF614CD768 11 17428 LGWR SYSTEM 1 123-PC ORACLE.EXE (LGWR) d:\oracle11g_64bit\diag\rdbms\second\second\trace\second_lgwr_17428.trc 1 5007762 5465314 131072 9921762
--另启一会话窗口,暂挂lgwr进程
SQL> oradebug setsospid 17428
ORA-00070: command setsospid is not valid
SQL> oradebug setospid 17428
Oracle pid: 11, Windows thread id: 17428, image: ORACLE.EXE (LGWR)
SQL> oradebug suspend
Statement processed.
--发生日志切换hang住
SQL> alter system switch logfile;
System altered
--新建一个会话窗口
create table t_h(a int);--语句hang住
--解挂lgwr进程,上述lgwr及create操作顺利执行完毕
SQL> oradebug resume
Statement processed.
小结：如果日志切换时,会暂住一切所有的数据库前端事务create and dml
About Me
