Oracle9.2.0.4 RAC系统加入新节点-CSDN博客

搜索词：加节点，add node

原系统基本情况： Linux AS3.0 2.4.21-4.ELsmp Oracle 9.2.0.4

注意：以下步骤需要停机作业。

１、安装新节点的HBA卡等硬件。由于是3节点，所有如果原来使用反线连接作为心跳线，现在需要使用一台HUB或交换机来连接3台节点机器。

２、安装节点的操作系统，与已经运行的节点一致，比如Linux AS3.0 2.4.21-4.ELsmp .

３、配置系统参数和ORACLE的安装环境。

安装Oracle软件之前的准备和安装单机时候一样，这里略过。

确认节点1，2，3 时间一样，最好使用时间同步服务器同步。

拷贝节点1，2的.bash_profile 环境设置，修改其中的SID部分。

备份节点1，2的以下两个文件。

修改节点1，2，3的/etc/hosts文件及/etc/hosts.equiv 文件

/etc/hosts 內容如下：

10.156.4.98 dmdii-node1

172.20.1.148 dmdii-node1

10.1.1.7 dmdii-pri1

10.156.4.99 dmdii-node2

172.20.1.149 dmdii-node2

10.1.1.8 dmdii-pri2

10.156.4.103 dmdii-node3

172.20.1.147 dmdii-node3

10.1.1.9 dmdii-pri3

/etc/hosts.equiv 內容如下：

dmdii-node1 oracle

dmdii-node2 oracle

dmdii-node3 oracle

dmdii-pri1 oracle

dmdii-pri2 oracle

dmdii-pri3 oracle

４、从运行节点的机器上把$ORACLE_HOME、$ORACLE_BASE、/etc/ora*复制到新安装机器上

对应的目录，要同源地址一致。下面注意点只是做一些解释，如果你是按照上面方式拷贝的Oracle软体，就不需要下面的这些步骤。

注意：最好是拷贝运行节点上的oracle相关目录。如果单独在新节点上安装Oracle , 会安装不了Real Application Cluster 组件，而且lib, bin 等oracle目录下会有很多比如gsd, gsdctl , libxxx.so 等文件不会安装进去。最后还是需要拷贝oracle相关目录下的文件覆盖。

没有安装RAC组件，后面开启新节点的时候会报错 ORA-00439: feature not enabled: Real Application Clusters，需要如下设置才能enable RAC :

RAC ON

1). Login as the Oracle software owner and shutdown all database instances on all nodes in the cluster.

2). cd $ORACLE_HOME/rdbms/lib

3). make -f ins_rdbms.mk rac_on

If this step did not fail with fatal errors then proceed to step 4.

4). make -f ins_rdbms.mk ioracle

５、运行新装节点的$ORACLE_HOME下的root.sh。

6，修改所有机器$ORACLE_HOME/oracm/admin下RAC配置。

备份节点1，2上的OCM配置文件 $ORACLE_HOME/oracm/admin/cmcfg.ora 。

开始配置节点1，2，3上的OCM配置文件，$ORACLE_HOME/oracm/admin/cmcfg.ora

节点1，2，3 类似如下，只是HostName不一样而已。注意检查所有节点：

dmdii-node1$cat cmcfg.ora

HeartBeat=15000

ClusterName=Oracle Cluster Manager, version 9i

PollInterval=1000

MissCount=210

PrivateNodeNames=dmdii-pri1 dmdii-pri2 dmdii-pri3

PublicNodeNames=dmdii-node1 dmdii-node2 dmdii-node3

ServicePort=9998

CmDiskFile=/ocfs_data/quorum.dbf

HostName=dmdii-pri1

KernelModuleName=hangcheck-timer

所有节点cmcfg.ora文件配置完成后，重新启动，开启各个节点的oracm测试(注意 su root) ， ps -ef| grep oracm 查看进程。

7，确认当前数据库的MAXINSTANCES大于等于您新加机器后的节点数，否则需重建控制文

件（但一般都够，默认好像32 ），所以这一步不用修改。

8，备份节点1，2的listener.ora ,tnsnames.ora 。然后开始修改节点1，2，3的

listener.ora ，tnsnames.ora 。

范例如下：

dmdii-node1$cat tnsnames.ora

LISTENERS_INTEL =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node1)(PORT = 1521))

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node2)(PORT = 1521))

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node3)(PORT = 1521))

)

LISTENER_INTEL1 =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node1)(PORT = 1521))

)

INTEL =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node1)(PORT = 1521))

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node2)(PORT = 1521))

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node3)(PORT = 1521))

)

(CONNECT_DATA =

(SERVER = DEDICATED)

(SERVICE_NAME = INTEL)

)

INST1_HTTP =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node1)(PORT = 1521))

)

(CONNECT_DATA =

(SERVER = SHARED)

(SERVICE_NAME = MODOSE)

(PRESENTATION = http://HRService)

)

EXTPROC_CONNECTION_DATA =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC))

)

(CONNECT_DATA =

(SID = PLSExtProc)

(PRESENTATION = RO)

)

INTEL1 =

(DESCRIPTION =

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node1)(PORT = 1521))

(CONNECT_DATA =

(SERVICE_NAME = INTEL)

(INSTANCE_NAME = INTEL1)

)

INTEL2 =

(DESCRIPTION =

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node2)(PORT = 1521))

(CONNECT_DATA =

(SERVICE_NAME = INTEL)

(INSTANCE_NAME = INTEL2)

)

INTEL3 =

(DESCRIPTION =

(ADDRESS = (PROTOCOL = TCP)(HOST = dmdii-node3)(PORT = 1521))

(CONNECT_DATA =

(SERVICE_NAME = INTEL)

(INSTANCE_NAME = INTEL3)

)

dmdii-node1$cat listener.ora

LISTENER =

(DESCRIPTION_LIST =

(DESCRIPTION =

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC))

)

(ADDRESS_LIST =

(ADDRESS = (PROTOCOL = TCP)(HOST = 172.20.1.148)(PORT = 1521))

(ADDRESS = (PROTOCOL = TCP)(HOST = 10.156.4.98)(PORT = 1521))

)

SID_LIST_LISTENER =

(SID_LIST =

(SID_DESC =

(SID_NAME = PLSExtProc)

(ORACLE_HOME = /u01/product/oracle)

(PROGRAM = extproc)

)

(SID_DESC =

(ORACLE_HOME = /u01/product/oracle)

(SID_NAME = intel1)

)

9，在节点3 上执行命令初始化共享文件

#su – oracle

$srvconfig -init

开启GSD, $gsdctl start

10，配置节点1，2，3的 spfile参数文件。

范例如下：

*.background_dump_dest='/u01/product/admin/intel/bdump'

*.cluster_database_instances=3

*.cluster_database=true

*.compatible='9.2.0.0.0'

*.control_files='/ocfs_ctrl_redo/intel/control01.ctl','/ocfs_ctrl_redo/intel/control02.ctl','/ocfs_ctrl_redo/intel/control03.ctl'

*.core_dump_dest='/u01/product/admin/intel/cdump'

*.cursor_sharing='EXACT'

*.db_block_size=8192

*.db_cache_size=1073741824

*.db_domain=''

*.db_file_multiblock_read_count=16

*.db_files=1000

*.db_name='intel'

*.dispatchers='(PROTOCOL=TCP) (SERVICE=intelXDB)'

*.fast_start_mttr_target=300

*.hash_join_enabled=TRUE

intel1.instance_name='intel1'

intel2.instance_name='intel2'

intel3.instance_name='intel3'

intel1.instance_number=1

intel2.instance_number=2

intel3.instance_number=3

*.java_pool_size=31457280

*.job_queue_processes=36

*.large_pool_size=20971520

intel1.local_listener='LISTENER_INTEL1'

intel2.local_listener='LISTENER_INTEL2'

intel3.local_listener='LISTENER_INTEL3'

intel1.log_archive_dest_1='LOCATION=/ocfs_arch1/intel/'

intel2.log_archive_dest_1='LOCATION=/ocfs_arch2/intel/'

intel3.log_archive_dest_1='LOCATION=/ocfs_arch2/intel/'

*.log_archive_format='%t_%s.dbf'

*.log_archive_start=true

*.open_cursors=4000

*.optimizer_index_cost_adj=30

*.optimizer_mode='rule'

*.pga_aggregate_target=314572800

*.processes=1015

*.query_rewrite_enabled='FALSE'

intel1.remote_listener='LISTENERS_INTEL'

intel2.remote_listener='LISTENERS_INTEL'

intel3.remote_listener='LISTENERS_INTEL'

*.remote_login_passwordfile='exclusive'

*.resource_limit=TRUE

*.service_names='intel'

*.session_cached_cursors=20

*.shared_pool_size=419430400

*.sort_area_size=2097152

*.star_transformation_enabled='FALSE'

intel1.thread=1

intel2.thread=2

intel3.thread=3

*.timed_statistics=TRUE

*.trace_enabled=false

*.undo_management='AUTO'

*.undo_retention=10800

intel1.undo_tablespace='UNDOTBS1'

intel2.undo_tablespace='UNDOTBS2'

intel3.undo_tablespace='UNDOTBS3'

*.user_dump_dest='/u01/product/admin/intel/udump'

11、在数据库中添加新的redo logfile 及undo ：

开启节点1的实例到mount状态，在节点1上为节点3建立redo log及undo tbs :

SQL> startup mount

SQL> ALTER DATABASE ADD LOGFILE THREAD 3

GROUP 9 ('/ocfs_ctrl_redo/intel/redo09.log', '/ocfs_data/intel/redo09b.log') size 100m,

GROUP 10 ('/ocfs_ctrl_redo/intel/redo10.log', '/ocfs_data/intel/redo10b.log') size 100m,

GROUP 11 ('/ocfs_ctrl_redo/intel/redo11.log', '/ocfs_data/intel/redo11b.log') size 100m,

GROUP 12 ('/ocfs_ctrl_redo/intel/redo12.log', '/ocfs_data/intel/redo12b.log') size 100m ;

SQL> alter database open ;

SQL> alter database enable public thread 3 ;

SQL> create undo tablespace undotbs3 datafile ‘/ocfs_data/intel/undotbs03_1.dbf’ size 2048064K ;

13、确认新节点的环境变量（ORACLE_HOME、ORACLE_SID等），确认spfile修改OK ,且节点3的online redo 及undo加入OK，以及节点3实例的实例目录及密码文件，各个目录权限等都OK . 开始开启各个节点的 oracm ,gsdctl, 如果密码文件有问题，重建新节点的密码文件，开启数据库，监听。