l         在rac3上修改listener.ora

[oracle@rac3 ~]$ vi /u01/oracle/product/9.2/network/admin/listener.ora

# LISTENER.ORA.RAC1 Network Configuration File: /u01/oracle/product/9.2/network/admin/listener.ora.rac1

# Generated by Oracle configuration tools.

LISTENER =

  (DESCRIPTION_LIST =

    (DESCRIPTION =

      (ADDRESS_LIST =

        (ADDRESS = (PROTOCOL = TCP)(HOST = rac3)(PORT = 1521))

      )

    )

  )

SID_LIST_LISTENER =

  (SID_LIST =

    (SID_DESC =

      (SID_NAME = PLSExtProc)

      (ORACLE_HOME = /u01/oracle/product/9.2)

      (PROGRAM = extproc)

    )

    (SID_DESC =

      (ORACLE_HOME = /u01/oracle/product/9.2)

      (SID_NAME = rac1)

    )

    (SID_DESC =

      (ORACLE_HOME = /u01/oracle/product/9.2)

      (SID_NAME = rac2)

    )

    (SID_DESC =

      (ORACLE_HOME = /u01/oracle/product/9.2)

      (SID_NAME = rac3)

    )

  )

　　注：要加入红色部分，否则

[oracle@rac1 dbs]$ srvctl start instance -d test -i rac3

PRKP-1011 : Failed to get all the listeners associated with instance rac3 on noderac3

l         在rac3上修改tnsnames.ora

加入下面两行

LISTENER_RAC3 =

(ADDRESS = (PROTOCOL = TCP)(HOST = rac3)(PORT = 1521))

l         3.5修改初始化参数文件

　在rac3上修改初始化参数文件加入以下内容

rac3.instance_name=RAC3

rac3.instance_number=3

rac3.local_listener=LISTENER_RAC3

rac3.thread=3

rac3.undo_tablespace=UNDOTBS3

l         在rac1上设置redo线程

SQL> alter database add logfile thread 3 group 5　('/u01/oracle/oradata/test/redolog3_1') size 10485760 ;

SQL> alter database add logfile thread 3 group 6　('/u01/oracle/oradata/test/redolog3_2') size 10485760 ;

SQL> alter database enable thread 3;

l         在rac1上创建undo 表空间

SQL> create undo tablespace undotbs3 datafile '/u01/oracle/oradata/test/undotbs03' size 40m;

l         添加新的节点并启动服务

[oracle@rac3 ~]$ srvctl add instance -d test -i rac3 -n rac3

[oracle@rac3 ~]$ srvctl start instance -d test -i rac3

注意事项：

在用DBCA添加实例时，在如下界面时不能点下一步，未找到原因，建议用命令方式添加实例。

删除实例

删除实例时，可按如下步骤操作

SQL> drop tablespace undotbs3;

SQL> alter database disable thread 3;

SQL> alter database drop logfile group 6;

SQL> alter database drop logfile group 5;

[oracle@rac3 ~]$ srvctl remove instance -d test -i rac3

如果需要添加多个实例，按如上步骤，逐个添加。

重新配置已存在的节点

l         修改rac1的Cluster Manger文件

(红色为加入的内容)

[oracle@rac1 admin]$ vi cmcfg.ora

ClusterName=Oracle Cluster Manager, version 9i

MissCount=210

PrivateNodeNames=rac2-priv rac1-priv rac3-priv

PublicNodeNames=rac2 rac1 rac3

ServicePort=9998

CmDiskFile=/dev/raw/raw1

KernelModuleName=hangcheck-timer

HostName=rac1-priv

l         修改rac2的Cluster Manger文件

[oracle@rac2 admin]$ vi cmcfg.ora

ClusterName=Oracle Cluster Manager, version 9i

MissCount=210

PrivateNodeNames=rac2-priv rac1-priv rac3-priv

PublicNodeNames=rac2 rac1 rac3

ServicePort=9998

CmDiskFile=/dev/raw/raw1

KernelModuleName=hangcheck-timer

HostName=rac2-priv

注：上述文件只要修改即可，的第三个节点相关启动时，它会自动重新配置。

3.2配置新节点

l         修改rac3的Cluster Manger文件

ClusterName=Oracle Cluster Manager, version 9i

MissCount=210

PrivateNodeNames=rac2-priv rac1-priv rac3-priv

PublicNodeNames=rac2 rac1 rac3

ServicePort=9998

CmDiskFile=/dev/raw/raw1

KernelModuleName=hangcheck-timer

HostName=rac3-priv

l         在rac3上启动Cluster Manger

[root@rac2 9.2]$ mkdir /var/opt/oracle

[root@rac2 9.2]$ vi /var/opt/oracle/srvConfig.loc

srvconfig_loc=/dev/raw/raw14

[root@rac2 9.2]$ chown -R oracle:dba /var/opt/oracle

[root@rac3 oracle]# chmod -R 755 /var/opt/oracle/srvConfig.loc

[root@rac3 ~]# . /home/oracle/.bash_profile

[root@rac3 ~]# /u01/oracle/product/9.2/oracm/bin/ocmstart.sh

[oracle@rac3 ~]$ ps -ef|grep ora

root      4892     1  0 Mar07 pts/1    00:00:08 oracm

root      4894  4892  0 Mar07 pts/1    00:00:05 oracm

root      4895  4894  0 Mar07 pts/1    00:04:04 oracm

root      4896  4894  0 Mar07 pts/1    00:00:00 oracm

root      4897  4894  0 Mar07 pts/1    00:01:04 oracm

root      4898  4894  0 Mar07 pts/1    00:00:09 oracm

root      4899  4894  0 Mar07 pts/1    00:00:08 oracm

root      4900  4894  0 Mar07 pts/1    00:01:03 oracm

root      4901  4894  0 Mar07 pts/1    00:00:30 oracm

root      4952  4894  0 Mar07 pts/1    00:00:00 oracm

root      4994  4894  0 Mar07 pts/1    00:00:05 oracm

root      6467  4894  0 01:09 pts/1    00:00:52 oracm

检查ocm日志

[oracle@rac2 log]$　tail –f /u01/oracle/product/9.2/oracm/log/cm.log

HandleUpdate(): SYNC(6) from node(0) completed {Mon Mar  3 18:12:28 2008 }^M

HandleUpdate(): NODE(0) IS ACTIVE MEMBER OF CLUSTER, INCARNATION(4) {Mon Mar  3 18:12:28 2008 }^M

HandleUpdate(): NODE(1) IS ACTIVE MEMBER OF CLUSTER, INCARNATION(6) {Mon Mar  3 18:12:28 2008 }^M

HandleUpdate(): NODE(2) IS ACTIVE MEMBER OF CLUSTER, INCARNATION(7) {Mon Mar  3 18:12:28 2008 }^M

NMEVENT_RECONFIG [00][00][00][00][00][00][00][07] {Mon Mar  3 18:12:28 2008 }^M

Debug Hang : CMNodeListener(PID=15316) Registered with ORACM. {Mon Mar  3 18:12:28 2008 }^M

<11557:7152><58441:2567>Debug Hang : CMNodeListener(PID=15317) Registered with ORACM. {Mon Mar  3 18:12:29 2008 }^M

Successful reconfiguration,  3 active node(s) node 0 is the master, my node num is 0 (reconfig 7) {Mon Mar  3 18:12:29 2008 }^M

   这时可以启动gsd服务

[oracle@rac3 ~]$ gsdctl start

[oracle@rac3 ~]$ ps -ef|grep gsd