9 Adding and Deleting Nodes and Instances

Preparing the New Node

Before a node can be added to the cluster, you must perform the same preinstallation steps on the new node as you did for all the existing nodes in the cluster. This includes the following tasks:

• Checking hardware compatibility

• Configuring the operating system

• Configuring SSH connectivity between the new node and the other cluster members

• Configuring access to shared storage

• Creating groups, users, and directories

Refer to Chapter 2, "Preparing Your Cluster" for details on these tasks.

cluvfy stage -pre crsinst -n racnode3 -verbose

Extending the Oracle Grid Infrastructure Home to the New Node

Now that the new node has been configured to support Oracle Clusterware, you use Oracle Universal Installer (OUI) to add a Grid home to the node being added to your Oracle RAC cluster.

This section assumes that you are adding a node named racnode3 and that you have already successfully installed Oracle Clusterware on racnode1 in a nonshared home, where Grid_home represents the successfully installed Oracle Clusterware home.

To extend the Oracle grid infrastructure to include the new node:

1. Verify the new node has been properly prepared for an Oracle Clusterware installation by running the following CLUVFY command on the racnode1 node:

   cluvfy stage -pre nodeadd -n racnode3 -verbose
   

2. As the oracle user (owner of the Oracle Grid software installation) on racnode1, go to Grid_home/oui/bin and run the addNode.sh script in silent mode:

   cd /u01/grid/oui/bin
   ./addNode.sh -silent "CLUSTER_NEW_NODES={racnode3}" "CLUSTER_NEW_VIRTUAL_
   HOSTNAMES={racnode3-vip}"
   

   When running this command, the curly braces ( { } ) are not optional and must be included or the command will return an error.

   You can alternatively use a response file instead of placing all the arguments in the command line. See Oracle Clusterware Administration and Deployment Guide for more information on using response files.

3. When the script finishes, run the root.sh script as the root user on the new node, racnode3, from the Oracle home directory on that node.

4. If you are not using Oracle Grid Naming Service (GNS), then you must add the name and address for racnode3 to DNS.

You should now have Oracle Clusterware running on the new node. To verify the installation of Oracle Clusterware on the new node, you can run the following command on the newly configured node, racnode3:

$ cd /u01/grid/bin
$ ./cluvfy stage -post nodeadd -n racnode3 -verbose

Extending the Oracle RAC Home Directory

Now that you have extended the Grid home to the new node, you must extend the Oracle home on racnode1 to racnode3. The following steps assume that you have already completed the tasks described in the previous sections, "Preparing the New Node" and "Extending the Oracle Grid Infrastructure Home to the New Node", and that racnode3 is a member node of the cluster to which racnode1 belongs.

The procedure for adding an Oracle home to the new node is very similar to the procedure you just completed for extending the Grid home to the new node.

To extend the Oracle RAC installation to include the new node:

1. Ensure that you have successfully installed the Oracle RAC software on at least one node in your cluster environment. To use these procedures as shown, replace Oracle_home with the location of your installed Oracle home directory.

2. Go to the Oracle_home/oui/bin directory on racnode1 and run the addNode.sh script in silent mode as shown in the following example:

   $ cd /u01/app/oracle/product/11.2.0/oui/bin
   $ ./addode.sh -silent "CLUSTER_NEW_NODES={racnode3}"
   

3. When the script finishes, run the root.sh script as the root user on the new node, racnode3, from the Oracle home directory on that node.

   For policy-managed databases with Oracle Managed Files (OMF) enabled, no further actions are needed.

   For a policy-managed database, when you add a new node to the cluster, it is placed in the Free pool by default. If you increase the cardinality of the database server pool, then an Oracle RAC instance is added to the new node, racnode3, and it is moved to the database server pool. No further action is necessary.

4. Add shared storage for the undo tablespace and redo log files.

   If OMF is not enabled for your database, then you must manually add an undo tablespace and redo logs.

5. If you have an administrator-managed database, then add a new instance on the new node as described in "Creating an Instance on the New Node".

   If you followed the installation instructions in this guide, then your cluster database is an administrator-managed database and stores the database files on Oracle Automatic Storage Management (Oracle ASM) with OMF enabled.

After completing these steps, you should have an installed Oracle home on the new node.

Adding the New Node to the Cluster using Enterprise Manager

If you followed the steps described in "Extending the Oracle RAC Home Directory", the cluster node is added to the cluster by the addNode.sh script. After the software is started on the new node, it is detected by Enterprise Manager. If an EM agent was not installed on the new node, then an alert is issued for that host, with the message "Incomplete configuration".

Creating an Instance on the New Node

You can add an instance to the cluster using either the Instance Management option of Database Configuration Assistant (DBCA) or using Enterprise Manager. Before using either of these options you must first configure the new node to be a part of the cluster and install the software on the new node as described in the previous sections.

There are two methods of adding an instance to the new node:

• Adding a New Instance for a Policy-Managed Database

• Adding a New Instance for an Administrator-Managed Database

Deleting an Instance From the Cluster Database

Sometimes, it might be necessary to remove a database instance from your cluster database. For example, if you want to retire or repurpose a server, you would first remove the database instance running on that server.

You can delete an instance from the cluster using either the Instance Management option of Database Configuration Assistant (DBCA) or using Enterprise Manager. If you choose to use DBCA to delete an instance, start the DBCA utility from a node that will remain part of the cluster.

When you delete an instance from a clustered database, there are two methods:

• Deleting an Instance From a Policy-Managed Database

• Deleting an Instance From an Administrator-Managed Database

Removing a Node From the Cluster

Removing a node from the cluster can be as easy as simply shutting down the server. As long as the node was not pinned and does not host any Oracle databases prior to Oracle Database 11g release 2, the node is automatically removed from the cluster when it is shut down. If the node was pinned or hosts database instance from previous releases, then explicit deletion is needed.

To repurpose this server, you can restart the node with a different profile in place, or you can use other software to put a new software image on the server.

If you have licensed the Oracle Enterprise Manager RAC Provisioning pack, you can also use the "Scale down Oracle Real Application Clusters" deployment procedure to remove a node from your cluster.