創建數據庫
可以運行dbca用圖形界面來創建,也可以用命令行來創建.
因為以前沒有創建過集群下的數據庫,因此我用圖形界面來創建腳本,修改后用命令行來創建.
先確認ORACM和GSD已經啟動:
$ ps -ef grep oracm
root 3056 1 0 10:50 pts/1 00:00:00 oracm
root 3057 3056 0 10:50 pts/1 00:00:00 oracm
root 3059 3057 0 10:50 pts/1 00:00:00 oracm
root 3060 3057 0 10:50 pts/1 00:00:00 oracm
root 3061 3057 0 10:50 pts/1 00:00:00 oracm
root 3062 3057 0 10:50 pts/1 00:00:00 oracm
root 3063 3057 0 10:50 pts/1 00:00:00 oracm
root 3064 3057 0 10:50 pts/1 00:00:00 oracm
root 3065 3057 0 10:50 pts/1 00:00:00 oracm
root 3338 3057 0 11:36 pts/1 00:00:00 oracm
root 3362 3057 0 11:36 pts/1 00:00:00 oracm
Oracle 3482 2932 0 11:38 pts/1 00:00:00 grep oracm
$ gsdctl stat
GSD is running on the local node
然后運行dbca啟動數據庫配置助手.
在歡迎畫面中,選擇Oracle cluster database, 然后一路繼續就可以了.
在最后的總結畫面上,可以選擇創建數據庫,也可以選擇創建腳本.我選擇了創建腳本.
手工修改腳本為自己想要的樣子,然后跟非集群環境一樣手工創建數據庫.數據文件使用本地治理,不用數據字典治理.并創建缺省的撤消表空間(undo tablespace)和臨時表空間.
注重:
根據我的經驗,SGA太小,在創建數據字典的時候會失敗,因此SGA可以大一些.但因為是集群,需要啟動多個instance,所以在數據字典創建完以后,即catalog.sql,catPRoc.sql,catclust.sql運行完以后,需要將SGA改小一點.
修改init.ora文件,并拷貝到$ORACLE_HOME/dbs下,改名字為initrac.ora.
我是一個instance對應一個listener.也可以一個listener響應多個instnce.
其余參數可以自己根據自己的情況進行設置,以下是cluster的參數:
undo_management=AUTO
undo_retention=10800
cluster_database=true
cluster_database_instances=5
rac1.instance_name=rac1
rac1.instance_number=1
rac1.local_listener=lsnr1
rac1.thread=1
rac1.undo_tablespace=UNDOTBS1
rac2.instance_name=rac2
rac2.instance_number=2
rac2.local_listener=lsnr2
rac2.thread=2
rac2.undo_tablespace=UNDOTBS2
rac3.instance_name=rac3
rac3.instance_number=3
rac3.local_listener=lsnr3
rac3.thread=3
rac3.undo_tablespace=UNDOTBS3
假如要多個instance都由一個listener來響應,則改以下兩行:
rac2.local_listener=lsnr1
rac3.local_listener=lsnr1
注重:
盡管數據庫名字叫rac,但啟動的實例名字不能叫rac.我將ORACLE_SID分別設置成了rac1,rac2,rac3.
然后啟動第一個實例,創建數據庫:
$ eXPort ORACLE_SID=rac1
$ sqlplus /nolog
SQL*Plus: Release 9.2.0.4.0???j? - ProdUCtion on Fri Nov 21 13:43:22 2003
Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved.
SQL> conn / as sysdba;
Connected to an idle instance.
SQL> startup pfile=/u01/app/oracle/product/9.2.0.4.0/dbs/initrac.ora;
ORACLE instance started.
Total System Global Area 38867968 bytes
Fixed Size 451584 bytes
Variable Size 33554432 bytes
Database Buffers 4194304 bytes
Redo Buffers 667648 bytes
Database mounted.
Database opened.
SQL>
創建數據庫的命令:
CREATE DATABASE rac
MAXINSTANCES 32
MAXLOGHISTORY 100
MAXLOGFILES 192
MAXLOGMEMBERS 3
MAXDATAFILES 1024
DATAFILE '/u02/oradata/rac/rac/system01.dbf' SIZE 100M REUSE AUTOEXTEND ON NEXT 10240K MAXSIZE 1000M
EXTENT MANAGEMENT LOCAL
DEFAULT TEMPORARY TABLESPACE TEMP TEMPFILE '/u02/oradata/rac/rac/temp01.dbf' SIZE 40M REUSE AUTOEXTEND ON NEXT 640K MAXSIZE 1000M
UNDO TABLESPACE "UNDOTBS1" DATAFILE '/u02/oradata/rac/rac/undotbs01.dbf' SIZE 50M REUSE AUTOEXTEND ON NEXT 5120K MAXSIZE 1000M
CHARACTER SET ZHS16GBK
LOGFILE GROUP 1 ('/u02/oradata/rac/rac/redo01.log') SIZE 10M REUSE,
GROUP 2 ('/u02/oradata/rac/rac/redo02.log') SIZE 10M REUSE,
GROUP 3 ('/u02/oradata/rac/rac/redo03.log') SIZE 10M REUSE;
然后創建數據字典.
以sysdba身份運行:
SQL> @?
/rdbms/admin/catalog.sql;
SQL> @?/rdbms/admin/catproc.sql;
SQL> @?/rdbms/admin/catclust.sql;
以system身份運行:
SQL> @?/sqlplus/admin/pupbld.sql;
