Managing Your Oracle RAC Databases (Oracle Database 10g to 11g Change)

posted Sep 27, 2010, 6:44 AM by Sachchida Ojha

Oracle RAC 11g introduces a new way to manage your Oracle RAC databases within a cluster. Traditionally we have allocated specific Oracle RAC instances to nodes within the cluster that requires the DBA to hard code parameters such as instance number, redo thread etc. If that node in the cluster did not start, then the instance for the database does not start. This traditional method of managing your Oracle RAC database will continue to be available and will be referred to as Administrator Managed database (Admin-Managed). Any database upgraded from an earlier release of Oracle Database, will be Administrator Managed. Any database that is using an earlier version of Oracle Database will be Administrator Managed. These databases will be managed using the same commands or methods (such as DBCA or Enterprise Manager) as they have been with previous releases of Oracle Database. All commands and utilities have maintained backward compatibility.

To move away from any hard coding of parameters, to make it easier to replace nodes in a cluster or expand the cluster as requirements change, we introduce a new way of managing your Oracle RAC database called Policy Managed. Policy Managed databases must be 11g Release 2 and cannot co-exist on the same servers as administrator managed databases. A policy managed database is defined by cardinality, the number of instances you would like running during normal operations. A policy-managed database will run in one or more database server pools that are created in the cluster by the cluster administrator. An instance will start on all servers that are in the server pools defined for the database. If you are using Automatic Storage Management (ASM) for your database storage, when an instance starts, if there is not a redo thread available, one will automatically be enabled and the required redo log files created as well as the undo tablespace.

With Oracle RAC 11g Release 2, there is only a database resource define to Oracle Clusterware. This resource will contain the Oracle Home, the spfile, the server pool(s) and the ASM diskgroup(s) required for the database. The database resource will have a weak start dependency on the VIP which means it will try to start the VIP for the node when the instance starts however if the VIP does not start successfully, the instance will still be started. When reviewing the database resource for an Administrator Managed database, you will see a server pool defined with the same name as the Oracle database. This pool will be part of a special Oracle defined server pool called Generic. The Generic server pool is managed by Oracle to support Administrator Managed databases. The server pools that are part of Generic will be automatically created or removed when you add or remove an Administrator Managed database.

Services for Administrator Managed databases will continue to be defined by the Preferred and Available definitions. For Policy Managed Databases, a Service will be defined to a Database Server Pool and can either be uniform (running on all instances in the server pool) or Singleton (running on only 1 instance in the server pool).