Overview of Installing Oracle RAC

Install Oracle Grid Infrastructure and Oracle AI Database software using Oracle Universal Installer, and create your database with Oracle DBCA.

This ensures that your Oracle RAC environment has the optimal network configuration, database structure, and parameter settings for the environment that you selected.

Alternatively, you can install Oracle RAC using Fleet Patching and Provisioning, which offers all of the advantages of Oracle Universal Installer and Oracle DBCA previously specified. In addition, Fleet Patching and Provisioning allows for standardization and automation.

Understanding Compatibility in Oracle RAC Environments

As part of your deployment plan, review the release compatibility restrictions and guidelines for using different Oracle AI Database releases on Oracle Grid Infrastructure.

To run Oracle RAC in configurations with different releases of Oracle AI Database in the same cluster, you must first install Oracle Grid Infrastructure, which must be the same version, or higher, as the highest version of Oracle AI Database that you want to deploy in this cluster. For example, to run an Oracle RAC 19c database and an Oracle RAC 26ai database in the same cluster, you must install Oracle Grid Infrastructure 26ai. Contact My Oracle Support for more information about version compatibility in Oracle RAC environments.

Oracle RAC Database Installation

Learn how to install Oracle Real Application Clusters (Oracle RAC) and review the restrictions.

Note:

The policy-managed database deployment option is desupported in Oracle AI Database 26ai.

Before you install Oracle RAC, first install Oracle Grid Infrastructure. The release of Oracle Grid Infrastructure must be the same as or newer than the Oracle RAC release that you are installing. Oracle Universal Installer only enables you to deploy an Oracle AI Database home across the nodes in a cluster if you previously installed and configured Oracle Grid Infrastructure for the cluster. If Oracle Universal Installer does not give you an option to deploy the database home across all of the nodes in the cluster, then the server on which you are attempting to install Oracle RAC is not a cluster member node.

During installation, you must manually run Oracle DBCA to create an Oracle RAC or an Oracle RAC One Node database. In these cases, database creation is a two-step process. First, install the Oracle RAC software by running the Oracle AI Database installer. Then create and configure your Oracle RAC or Oracle RAC One Node database using Oracle DBCA.

Note:

Before you create a database, a default listener must be running in the Oracle Grid Infrastructure home. If there is no default listener in the Oracle Grid Infrastructure home, then Oracle DBCA returns an error instructing you to run NETCA from the Oracle Grid Infrastructure home to create a default listener.

The Oracle RAC software is distributed as part of the Oracle AI Database installation media. By default, the Oracle AI Database software installation process installs the Oracle RAC option when the installation process recognizes that you are performing the installation on a cluster. Oracle Universal Installer installs Oracle RAC into a directory structure referred to as the Oracle home, which is separate from the Oracle home directory for other Oracle software running on the system. Because Oracle Universal Installer is cluster aware, it installs the Oracle RAC software on all of the nodes that you defined to be part of the cluster.

The Oracle AI Database installation process creates a read-only Oracle home directory by default. You can use the read-only Oracle home as a software image to be shared across multiple database servers. This simplifies patching and mass rollout because only one Oracle home image needs to be updated to distribute a patch to multiple database servers. Read-only Oracle AI Database homes are more secure than traditional Oracle AI Database homes because there is a clear separation of configuration information from the actual software. Thus, there is no risk of new files being created inside the Oracle AI Database home from active processes.

Oracle RAC Database Creation

Part of Oracle AI Database deployment is the creation of the database.

You can create a database as part of the database deployment or you can only deploy the database software first, then create any database that is meant to run out of the newly created Oracle home by using Oracle DBCA.

Note:

The policy-managed database deployment option is desupported in Oracle AI Database 26ai.

In Oracle RAC environments, the values for DB_UNIQUE_NAME.DB_DOMAIN in its entirety must be unique for each database within your enterprise. The name of each Pluggable Database (PDB) should also be unique within the cluster.

By default, Oracle DBCA creates one service for your Oracle RAC installation. This is the default database service and you should not use this service for user connectivity. The default database service is typically identified using the combination of the DB_NAME and DB_DOMAIN initialization parameters: db_name.db_domain. The default service is available on all instances in Oracle RAC environments, unless the database is in restricted mode.

Note:

Oracle recommends that you reserve the default database service for maintenance operations and create dynamic database services for user or application connectivity as a post-database creation step, using either SRVCTL or Oracle Enterprise Manager. Oracle DBCA no longer offers a dynamic database service creation option for Oracle RAC databases. For Oracle RAC One Node databases, you must create at least one dynamic database service.

Overview of Extending Oracle RAC Clusters

To extend an Oracle RAC cluster, also known as cloning, and add nodes to your environment after your initial deployment, then you must to do this on multiple layers, considering the management style that you currently use in the cluster.

Oracle provides various means of extending Oracle RAC clusters. Choose from the following approaches to extend the current environment:

Fleet Patching and Provisioning to provision new Oracle RAC databases and other software
Cloning using cloning scripts
Adding nodes using the addnode.sh ( addnode.bat on Windows) script

Both approaches apply, regardless of how you initially deployed your environment. Both approaches copy the Oracle software on to the node that you plan to add to the cluster. Software that is copied to the node includes the Oracle Grid Infrastructure software and the Oracle AI Database homes.

For Oracle AI Database homes, ensure that the database software is deployed on all of the nodes on which database instances can potentially run. In either case, first deploy Oracle Grid Infrastructure on all of the nodes that are to be part of the cluster.

Note:

Oracle cloning is not a replacement for cloning using Oracle Enterprise Manager as part of the Provisioning Pack. When you clone Oracle RAC using Oracle Enterprise Manager, the provisioning process includes a series of steps where details about the home that you want to capture, the location to which you want to deploy, and various other parameters are collected.

For new installations, or if you install only one Oracle RAC database, use the traditional automated and interactive installation methods, such as Oracle Universal Installer, Fleet Patching and Provisioning, or the Provisioning Pack feature of Oracle Enterprise Manager. To add or delete Oracle RAC from nodes in a cluster, use the procedures detailed in the Adding and Deleting Oracle RAC from Nodes... topics listed at the end of this topic.

The cloning process assumes that you successfully installed an Oracle Clusterware home and an Oracle home with Oracle RAC on at least one node. In addition, all root scripts must have run successfully on the node from which you are extending your cluster database.