Oracle数据库shutdown immediate被hang住的几个原因

实验操作环境：

        操作系统：Red Hat Enterprise Linux ES release 4 (Nahant Update 6)                  

        数据库 ： Oracle Database 10g Release 10.2.0.4.0 – Production  32bit

今晚使用shutdown immediate（其实是执行stop_oracle.sh脚本关闭数据库，如下所示）关闭数据库的时候，

[oracle@gsp-orasvr02 scripts]$ more stop_oracle.sh
lsnrctl stop LISTENER
sleep 15
sqlplus /nolog <<EOF
conn / as sysdba;
alter system switch logfile;
alter system checkpoint;
shutdown immediate;
exit
EOF

在另外一个会话中使用tail  -20f  命令查看告警日志的输出，结果发现数据库等待了很长时间都没有正常关闭，hang住在下面地方：

Active call for process 11121 user 'oracle' program 'oracle@get-orasvr02 (S000)'

Active call for process 7162 user 'oracle' program 'oracle@get-orasvr02 (S011)'

                                                           截图如下

这时解决办法是找出hang住的进程并杀掉(当时操作没有截图，也没有保存输出信息),因为有些session无法被pmon进程清理，导致数据库无法顺利关闭，需要手工杀掉进程。首先使用ps 和grep找到这两个进程。

[ oracle@get-orasvr02  bdump]$ ps -ef | grep oracle | grep  S000

[ oracle@get-orasvr02  bdump]$ ps -ef | grep oracle | grep  S011

然后使用kill -9  processesid杀掉这两个进程即可，杀掉这两个进程后，从告警日志里面看到里面跳到关闭dispatcher 。如下所示：

[ oracle@get-orasvr02  bdump]$ tail  alert_epps.log
  Current log# 3 seq# 242223 mem# 1: /u02/oradata/epps/redo03_01.log
Sun Jan 5 05:14:50 2014
Starting control autobackup
Control autobackup written to DISK device
        handle '/u01/app/oracle/product/10.2.0/db_1/dbs/c-2179993557-20140105-0e'
Sun Jan 5 05:14:54 2014
ALTER SYSTEM ARCHIVE LOG
Sun Jan 5 05:14:55 2014
Thread 1 cannot allocate new log, sequence 242224
Checkpoint not complete
  Current log# 3 seq# 242223 mem# 0: /u01/app/oracle/oradata/epps/redo03_1.log
  Current log# 3 seq# 242223 mem# 1: /u02/oradata/epps/redo03_01.log
Sun Jan 5 05:14:58 2014
Thread 1 advanced to log sequence 242224 (LGWR switch)
  Current log# 5 seq# 242224 mem# 0: /u01/app/oracle/oradata/epps/redo05_1.log
  Current log# 5 seq# 242224 mem# 1: /u02/oradata/epps/redo05_02.log
Sun Jan 5 07:31:56 2014
Thread 1 advanced to log sequence 242225 (LGWR switch)
  Current log# 2 seq# 242225 mem# 0: /u01/app/oracle/oradata/epps/redo02_1.log
  Current log# 2 seq# 242225 mem# 1: /u02/oradata/epps/redo02_02.log
Sun Jan 5 07:32:20 2014
Starting background process EMN0
Shutting down instance: further logons disabled
EMN0 started with pid=43, OS id=7062
Sun Jan 5 07:32:21 2014
Stopping background process CJQ0
Sun Jan 5 07:32:21 2014
Stopping background process QMNC
Sun Jan 5 07:32:23 2014
Stopping background process MMNL
Sun Jan 5 07:32:34 2014
Background process MMNL not dead after 10 seconds
Sun Jan 5 07:32:34 2014
Killing background process MMNL
Sun Jan 5 07:32:35 2014
Stopping background process MMON
Sun Jan 5 07:33:05 2014
Background process MMON not dead after 30 seconds
Sun Jan 5 07:33:05 2014
Killing background process MMON
Sun Jan 5 07:33:06 2014
Shutting down instance (immediate)
License high water mark = 561
Sun Jan 5 07:33:06 2014
Stopping Job queue slave processes, flags = 7
Sun Jan 5 07:33:06 2014
Process OS id : 6088 alive after kill
Errors in file /u01/app/oracle/admin/epps/udump/epps_ora_7055.trc
Sun Jan 5 07:33:09 2014
Waiting for Job queue slaves to complete
Sun Jan 5 07:33:09 2014
Job queue slave processes stopped
Sun Jan 5 07:38:10 2014
Active call for process 11121 user 'oracle' program 'oracle@get-orasvr02 (S000)'
Active call for process 7162 user 'oracle' program 'oracle@get-orasvr02 (S011)'
SHUTDOWN: waiting for active calls to complete.
Sun Jan 5 07:57:28 2014
Waiting for dispatcher 'D000' to shutdown
Waiting for dispatcher 'D001' to shutdown
Waiting for dispatcher 'D002' to shutdown
Waiting for dispatcher 'D003' to shutdown
Waiting for dispatcher 'D004' to shutdown
Waiting for dispatcher 'D005' to shutdown
Waiting for dispatcher 'D006' to shutdown
Sun Jan 5 07:59:29 2014
All dispatchers and shared servers shutdown
Sun Jan 5 08:04:30 2014
SHUTDOWN: Active processes prevent shutdown operation
Sun Jan 5 08:09:32 2014
SHUTDOWN: Active processes prevent shutdown operation

Oracle的官方文档介绍、解释如下

The database is waiting for pmon to clean up processes, but pmon is unable to

clean them. The client connections to the server are causing the shutdown

immediate or normal to hang. Killing them allows pmon to clean up and release

the associated Oracle processes and resources.

What resources are we talking about?

1) Any non committed transactions must be rolled back

2) Any temporary space (sort segments / lobs / session temporary tables) must be freed

3) The session itself and any associated memory consumed by the session.

4) Internal locks / enqueues must be cleaned up

Often Oracle (SMON or PMON depending on whether Shared Server is used) will wait for the OS to terminate the process(es) associated with the session. If the OS never returns, or fails to terminate them, then the instance shutdown will hang with this message (Shutdown Waiting for Active Calls to Complete)

Other means exist to achieve a quick shutdown, as outlined inNote 386408.1- What Is The Fastest Way To Cleanly Shutdown An Oracle Database?

结果解决上面问题后，本以为可以顺利关闭数据库，结果又hang住了，告警日志信息提示为

SHUTDOWN: Active processes prevent shutdown operation

出现这个错误原因：

因为我大概如下的操作导致：

[oracle@gsp-orasvr02 scripts]$ sqlplus / as sysdba

...........

SQL> !

[oracle@get-orasvr02 ~]$

..... （执行了一些shell 命令）

然后又使用了sqlplus启动登录了数据库，然后做shutdown immediate操作，这时导致shutdown immediate被hang住。

[oracle@gsp-orasvr02 scripts]$ sqlplus / as sysdba

解决办法：退出当前的会话，回到原始会话，并重新连接，就可以正常的关闭数据库了

参考资料：

http://oracle.chinaitlab.com/optimize/761636_2.html

http://blog.csdn.net/xionglang7/article/details/8997081

http://blog.csdn.net/leshami/article/details/9466559