1. Execute "cluvfy" command, for example:
cluvfy stage -pre nodeadd -n
[-fixup [-fixupdir fixup_dir]] [-verbose]
You can
specify the -fixup option and a directory into which CVU prints
instructions
to fix the cluster or node if the verification
fails
2. Use OUI to add nodes:
export
ORACLE_HOME=/u01/app/11.2.0/grid
export
PATH=$ORACLE_HOME/bin:$ORACLE_HOME/oui/bin:$PATH
If you are using Grid
Naming Service (GNS), execute below command:
addNode.sh -silent
"CLUSTER_NEW_NODES={NEWnode2}"
If you are not using Grid Naming Service
(GNS), execute below command:
addNode.sh -silent
"CLUSTER_NEW_NODES={NEWnode}"
"CLUSTER_NEW_VIRTUAL_HOSTNAMES={NEWnode-vip}"
When
the addNode.sh script. completes, a message window displays a list
of
nodes in the cluster and root scripts that must be run on those
nodes.
Note: If error occurs, fix the error
according to the log file then re-run
specified commands.
Fix
: If "/u01/app/oraInventory" doesnot exist, create the dir and grant
appropriate
permission to it
$ cd /u01/app
$ sudo mkdir
oraInventory
$ sudo chown -R oracle:dba oraInventory
If "AttachHome"
command fails, run the following command manually on
the
failed nodes:
/u01/app/11.2.0/grid/oui/bin/runInstaller
-attachHome -noClusterEnabled
ORACLE_HOME=/u01/app/11.2.0/grid
ORACLE_HOME_NAME=Ora11g_gridinfrahome1
CLUSTER_NODES={,,,,}
CRS=true
"INVENTORY_LOCATION=/u01/app/oraInventory"
-invPtrLoc
"/u01/app/11.2.0/grid/oraInst.loc"
LOCAL_NODE=""
If "UpdateNodeList"
fails, run the following command again on the fialed
nodes:
/u01/app/11.2.0/grid/oui/bin/runInstaller -updateNodeList
-noClusterEnabled
ORACLE_HOME=/u01/app/11.2.0/grid
CLUSTER_NODES={,,,,}
CRS=true "INVENTORY_LOCATION=/u01/app/oraInventory"
-invPtrLoc
"/u01/app/11.2.0/grid/oraInst.loc"
LOCAL_NODE=""
3. Check that your cluster
is integrated and that the cluster is not divided into
separate parts
by running the following CVU command. This command verifies
that any
number of specific nodes has been successfully added to the cluster
at
the network, shared storage, and clusterware levels:
cluvfy
stage -post nodeadd -n node2 [-verbose]
4. You can
optionally extend Oracle Database with Oracle Real Application
Clusters
(Oracle RAC) components to the new nodes, making them members
of an
existing Oracle RAC database.
There are two methods for you to extend
Oracle database with RAC. One is
using Oracle Enterprise Manager, Another
choice is using DBCA in Interactive
or silent mode to add
instances to nodes on which you have extended an Oracle
Clusterware
home and an Oracle Database home.
For silent
mode:
dbca -silent -addInstance -nodeList [node] -gdbName
[gdbname]
[-instanceName instname] -sysDBAUserName
[sysdba]
-sysDBAPassword [password]
Parameter
description
[node]: The node on which you want to add (or delete)
the instance.
[gdbname]: Global database name.
[instname]: Name of
the instance. Provide an instance name only if you
want to override the
Oracle naming convention for Oracle RAC instance names.
[sysdba]: Name of
the Oracle user with SYSDBA privileges.
[password]: Password for the
SYSDBA user.
Example:
dbca -silent -addInstance -nodeList
"" -gdbName "orcl" -instanceName
"orcl2" -sysDBAUserName "sys"
-sysDBAPassword xxx
来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/22308399/viewspace-749501/,如需转载,请注明出处,否则将追究法律责任。