Oracle Real Application Clusters

One of the great advantages of the clustered configuration is to add or remove nodes from the cluster without affecting the other already existing instances in the cluster. This provides scalability to the applications using RAC. However, when adding nodes to the cluster and before allowing it to participate as an instance in the clustered configuration, certain pre requisites or procedures have to be completed, as follows:

1. The node joining the cluster should have a cloned image of the operating system that matches the other nodes in the cluster. The cloned image should contain the same version including all patch sets applied. This step is a very important step; the operating system image should be an exact clone of the other nodes.

2. The server configuration file (on a Unix operating system) should be manually updated on the new node by copying it from an existing node. The creation of the server configuration file is illustrated in Section 8.5.1 cluster configuration under step 3.

3. Depending on the original implementation, the disks have to be configured to use the clustered file system or for raw devices.

4. Once step 3 is complete and the node is part of the cluster ensure that all nodes are running and the new node is visible to all other nodes. This could be accomplished by using the lsnodes–l command on a Unix operating system.

5. (Perform this step only if using Oracle 10g). Using the OUI install the Oracle Cluster ready services (CRS) software on the new node. If the node has any other cluster software installed they should be disabled before CRS is installed.

6. (Perform this step only if using Oracle 10g). From an existing node, run the clscfg –add command, which configures the new node for the cluster.

7. Run the OUI to install RAC

8. On Unix platform, run the root.sh script on the new node to set up the configuration locations.

9. Create the instance either manually as described in Section 8.4.2 above or by running the DBCA utility.