一數據庫服務器執行shutdown immediate時，遇到了下面ORA錯誤，如下所示：
$ sqlplus / as sysdba
SQL*Plus: Release 10.2.0.4.0 - Production on Fri Aug 5 10:56:24 2016
Copyright (c) 1982, 2007, Oracle. All Rights Reserved.
Connected to:
Oracle Database 10g Release 10.2.0.4.0 - 64bit Production
SQL> shutdown immediate;
ORA-03113: end-of-file on communication channel
SQL> exit
檢查告警日志，發現其一直停留在"Job queue slave processes stopped".退出sqlplus然後重新登錄，重新執行shutdown immediate
$ sqlplus / as sysdba
SQL*Plus: Release 10.2.0.4.0 - Production on Fri Aug 5 11:03:45 2016
Copyright (c) 1982, 2007, Oracle. All Rights Reserved.
Connected.
SQL> shutdown immediate;
ORA-24324: service handle not initialized
ORA-24323: value not allowed
ORA-01089: immediate shutdown in progress - no operations are permitted
SQL>
後面退出會話，然後登陸sqlplus，使用shutdown abort 才將數據庫關閉。處理完成後，查了一下metal link官方文檔，發現引起錯誤的原因為：
Background processes are hanging/not started correctly during the previous startup of this database.
Hence the semaphores and shared memory segments are not getting detached properly now during shutdown.
以後遇到這個問題，可以使用下面步驟處理：
1. Verify that there are no background processes owned by "oracle" , if there are kill them
$ ps -ef | grep ora_ | grep $ORACLE_SID
2. Remove shared memory and semaphores:
A) Check for shared memory and semaphores
$ ipcs -mt (if there is anything owned by oracle remove it)
$ ipcrm -m [ID] (to remove it)
B) Check and remove semaphores
$ ipcs -sbt (if there is anything owned by oracle remove it)
$ ipcrm -s [ID] (to remove it)
C) Remove sga and lk file
$ cd $ORACLE_HOME/dbs
$ rm sgadef<SID>.dbf (removing sga file)
$ORACLE_HOME/dbs/lk<sid> (removing lk... flies)
D) If database is down, try to bring up oracle one step at a time:
$ sqlplus /nolog
SQL> startup nomount pfile = ...[path]
SQL> alter database mount;
SQL> alter database open;\
Otherwise, exit current SQL*Plus session and verify the following environment variables are set.
echo $ORACLE_HOME
echo $ORACLE_SID (echo %ORACLE_SID% on Windows)
Then, Execute the following:
sqlplus / as sysdba
shutdown abort
exit
sqlplus / as sysdba
startup
如上官方資料所示， 兩種解決方案。關於第一種方案，簡單梳理如下：
1： 清理後台進程，一般找到相關進程後，使用kill命令殺掉。
2： 清理共享內存段
3： 清理信號集
注意，在RHEL 中， ipcs -sbt命令會報錯"ipcs: invalid option -- b",這個是因為Linux上的ipcs命令，不支持UNIX上的-b，所以不能照本宣科，不要使用參數b. 具體參考官方文檔ipcs man page describes invalid -b option in RHEL 5
4: 刪除 sga and lk文件。
ipcs相關資料：
ipcs 命令往標准輸出寫入一些關於活動進程間通信設施的信息。如果沒有指定任何標志，ipcs 命令用簡短格式寫入一些關於當前活動消息隊列、共享內存段、信號量、遠程隊列和本地隊列標題。
參考資料：
ORA-24324 During Startup or Shutdown (文檔 ID 794293.1)
ORA-1089 During Shutdown Immediate (文檔 ID 1014091.102)
