Skip Headers
Oracle® Database Oracle Clusterware and Oracle Real Application Clusters Administration and Deployment Guide
10g Release 2 (10.2)

Part Number B14197-02
Go to Documentation Home
Home
Go to Book List
Book List
Go to Table of Contents
Contents
Go to Index
Index
Go to Master Index
Master Index
Go to Feedback page
Feedback

Go to previous page
Previous
Go to next page
Next
View PDF

2 Introduction to Oracle Clusterware and RAC Administration and Deployment

This chapter provides an overview of administering the Oracle Clusterware and Real Application Clusters (RAC) environments. This chapter includes the following topics:

Real Application Clusters Documentation Overview

This section describes the RAC documentation set. This book, the Oracle Database Oracle Clusterware and Oracle Real Application Clusters Administration and Deployment Guide, provides RAC-specific administration and application deployment information.

This book describes how to administer the Oracle Clusterware components such as the voting disks and the Oracle Cluster Registry (OCR). This book also explains how to administer storage and how to use RAC scalability features to add and delete instances and nodes. This book also discusses how to use Recovery Manager (RMAN), and how to perform backup and recovery in RAC.

The Oracle Database Oracle Clusterware and Oracle Real Application Clusters Administration and Deployment Guide describes RAC deployment topics by explaining how to deploy automatic workload management and take advantage of high availability by using services. This book describes how the Automatic Workload Repository (AWR) tracks and reports service levels and how you can use service level thresholds and alerts to improve high availability in your RAC environment. This book also explains how to make your applications highly available by using the Oracle Clusterware.

The Oracle Database Oracle Clusterware and Oracle Real Application Clusters Administration and Deployment Guide explains how to monitor and tune performance in RAC environments by using Oracle Enterprise Manager and by using information in AWR and Oracle performance views. This book also highlights application-specific deployment techniques for online transaction processing and data warehousing environments.

The appendices in this book describe:

In addition to this book, refer to the platform-specific Oracle Real Application Clusters installation and configuration guide on your product installation media as described under the following heading.

Platform-Specific Oracle Real Application Clusters Installation and Configuration Guides

Each platform-specific Oracle Database 10g installation media contains a copy of an Oracle Real Application Clusters platform-specific installation and configuration guide in HTML and PDF formats. These RAC installation books contain the pre-installation, installation, and post-installation information for the various UNIX- and Windows-based platforms on which RAC operates.


Note:

Additional information for this release may be available in the Oracle Database 10g README or Release Notes. If these documents are available for this release, then they are on your Oracle product installation media.

Introduction to Administering Real Application Clusters

Install the Oracle Clusterware and your Oracle Database 10g software with the Oracle Universal Installer (OUI) and create your database with the Database Configuration Assistant (DBCA). This ensures that your RAC environment has the optimal network configuration, database structure, and parameter settings for the environment that you selected. As a DBA, after installation your tasks are to administer your RAC environment at three levels:

Administering Real Application Clusters

Use the following tools to perform administrative tasks in RAC:

Voting Disk and Oracle Cluster Registry Device Administration

You can use multiple disks for both the voting disk and the Oracle Cluster Registry (OCR) disk to prevent them from becoming single points of failure. Administer these components as described in Chapter 3, " Administering Oracle Clusterware Components".

Database Instance Management and Database Administration in RAC

Use Enterprise Manager, SQL*Plus, or SRVCTL to administer database instances and RAC databases as described in Chapter 5, " Administering Database Instances and Cluster Databases".

Storage Management in Real Application Clusters

When you create your database, you can create Automatic Storage Management (ASM) disk groups and configure mirroring for ASM disk groups using the DBCA. After your RAC database is operational, you can administer ASM disk groups with Enterprise Manager or the SRVCTL utility as described in Chapter 4, " Administering Storage".

You can configure ASM in a separate standalone ASM-only home. This enables instances for non-RAC and RAC databases to share a single ASM instance on a node. You also have the option to upgrade ASM independently of your database upgrades.

The Oracle tools that you use to manage ASM, including DBCA, Database Upgrade Assistant (DBUA), Enterprise Manager, and the silent mode install and upgrade commands, include options to manage ASM instances and disk groups. For example, you can run the DBCA to create a new ASM instance or ASM diskgroup independently of creating a database.

When you perform ASM-related options during installs, upgrades, or other operations, the tool you are using may automatically extend ASM to other nodes in your cluster. This can include installing ASM software into the same home as on the current node and starting the ASM instance. For example, if you use DBCA to create a database using a new Oracle home, then DBCA will attempt to extend ASM to the new Oracle home on all of the nodes you select.

In some cases, your current configuration may not be compatible with an ASM activity that you are trying to perform, either explicitly or with an automatic ASM extension to other nodes. Should you be using DBCA, as described earlier, to build a database using a new Oracle home, and if the ASM version is from an earlier release of the Oracle software but does not exist on all of the nodes you selected for the database, then ASM cannot be extended. Instead, the DBCA session displays an error, prompting you either to run the add node script or to upgrade ASM using the DBUA.

Oracle Clusterware for Real Application Clusters

When you create a RAC database, you can also use the DBCA to create services and assign them to instances. After your RAC database is operational, you can use Enterprise Manager, the DBCA, or SRVCTL to create and administer services and high availability components as described in Chapter 6, "Introduction to Workload Management".

Other high availability components include node resources such as the Virtual Internet Protocol (VIP) address, the Global Services Daemon, the Oracle Notification Service, and the Oracle Net Listeners. These resources are automatically started when the Oracle Clusterware starts the node and then automatically restarts them if they stop. The application level resources are the instances and the Oracle Clusterware background processes that run on each instance.

You can use the VIPCA to administer VIP addresses and you can use SRVCTL to administer other node resources. The information that describes the configuration of these components is stored in the Oracle Cluster Registry (OCR) that you can administer as described in Chapter 3, " Administering Oracle Clusterware Components".

Additional Real Application Clusters Administrative Topics

This book contains the following additional RAC administrative topics:

Overview of Using Enterprise Manager with Real Application Clusters

Enterprise Manager is a Web-based tool with RAC-specific administration and performance-related features. If you create your RAC database with the Database Configuration Assistant (DBCA), then the Enterprise Manager Database Control tool is automatically configured for your RAC environment. This means that all instances that were part of your installation have an Enterprise Manager Agent running on them. Enterprise Manager Database Control enables you to manage a single RAC database with its instance targets, Oracle Net Services Listener targets, host targets, and a cluster target.

You can configure Enterprise Manager Grid Control on additional hosts, which were not part of your initial RAC installation, either inside or outside your cluster environment. Enterprise Manager Grid Control enables you to manage multiple cluster databases, cluster database instances, and the hosts on which cluster database components operate.

Grid Control enables you to monitor and administer your entire computing environment from one network location. Use Grid Control to manage all of your enterprise services, including hosts, databases, Listeners, application servers, HTTP Servers, and Web applications, as one cohesive unit. Enterprise Manager Grid Control only requires one Agent on one host in your cluster environment to perform cluster database and instance discovery. Install Enterprise Manager Grid Control from a your Oracle Database 10g installation media.


See Also:

Oracle Enterprise Manager Concepts for more information about using Enterprise Manager

You can also use both Enterprise Manager Database Control and the Enterprise Manager Grid Control to:

The following sections introduce Oracle Real Application Clusters (RAC) application deployment and performance by explaining the main points that you need to remember when you deploy applications on RAC:

Overview of Deploying Applications on Real Application Clusters

To optimally deploy applications on RAC, remember the following points:


See Also:

Your platform-specific Real Application Clusters installation guide for more information about configuring these features for Oracle Database 10g Real Application Clusters

Code Changes are Not Required for Applications

Applications that perform well on single-instance Oracle databases do not require code changes to perform well on RAC databases. The same compatibility guidelines that apply to Oracle single-instance databases also apply to RAC. For example, if an application scales well in an SMP environment, then the application will scale well with RAC.

The primary feature of RAC that enables applications to scale across multiple instances is Cache Fusion. Cache Fusion enable each instance to locate the most useful version of a block of data for its needs, whether that block resides on disk or in the memory of another instance. The cluster interconnect provides a fast pathway for Cache Fusion to transfer required blocks of data between instances when necessary.

Implementing Oracle Features with Real Application Clusters

The Oracle features described in this section enhance the performance and simplify the administration of your RAC environment. The features discussed in this section are:

Automatic Storage Management

Automatic Storage Management (ASM) simplifies database administration by eliminating the need for you to manage Oracle database files. Instead, ASM enables you to create disk groups that comprise disks and the files that reside on them.

Cluster File Systems in Real Application Clusters

Depending on your hardware platform, you can store Oracle homes and Oracle datafiles on a cluster file system. Cluster file systems are simpler to configure and manage than raw device storage. Cluster file systems also offer scalable, low latency, highly resilient storage that significantly reduces costs.

Storage Management Features and Real Application Clusters

The advanced storage features of Oracle Automatic Storage Management greatly enhance manageability for RAC just as with single instance Oracle. Other storage features include Oracle-managed files, automatic segment-space management, and automatic undo management. Refer to the Oracle database documentation for more information about using storage management features.

Services in Oracle Database 10g

With Oracle Database 10g, you can define application workloads as services so that you can individually manage and control them. You can create a service for each application or for major components within a complex application. Once created, you can define where and when a service runs. Your entire database workload can be separated into a few services, each of which can be managed independently, reducing the need to manage individual users or sessions for many tasks.

In a RAC database, you can use services to maximize the use of your cluster's processing resources. You can assign each service to one or more instances for normal startup (preferred), depending on its processing requirements. Additionally, you can define one or more alternate (available) instances that a service can use if one of the service's assigned (preferred) instances becomes unavailable.

In both cluster and non-cluster environments, you can track performance metrics by service using the Automatic Workload Repository (AWR). You can set thresholds on performance metrics so that your system automatically generates alerts if the thresholds are exceeded. You can also map services to Resource Manager consumer groups to provide more fine-grained resource allocation controls such as placing limits on CPU consumption. Other Oracle tools and facilities such as Oracle Scheduler, Parallel Query, and Oracle Streams Advanced Queuing can also use services to manage their workloads.

The settings for Resource Manager are instance specific, that is, you can have different Resource Manager settings on different instances. Therefore, ensure that your Resource Manager settings are appropriate for the instances on which you are using it. The Automatic Workload Management features of Oracle Database 10g are integrated with the resource manager. The only difference in the way you use Resource Manager with RAC is that Resource Manager has instance specific settings.


Note:

An example that shows how to manage a complex environment with database services is available on the Oracle Technology Network at the URL http://www.oracle.com/index.html. The example includes entries from the related files, such as listener, database, and Net Services parameter files, as well as the commands to create and use the services for application management.

Runtime Connection Load Balancing is used when selecting connections from a connection pool. For connection pools that support services at one instance only, the first available connection in the pool is used. When connection pools support services that span multiple instances using a policy such as service metrics, Runtime Connection Load Balancing distributes work requests across instances that are adequately serving a service. This avoids sending work to slow, hung, failed or restricted instances.

For connection load balancing, the method for balancing connections across a service uses four metrics: session count by instance, run queue length of the node, goodness by service, and weighted session count by service. The metrics used depend on what goals have been defined for the service and how Oracle Net Services has been configured.Fast Connection Failover is used to prevent connections being directed to failed nodes or instances. Integrated connection pools will clean up connections when a failure occurs or add additional connections when new instances become available. This allows the application to be immediately aware of cluster configuration changes and react to them without any programming or configuration changes at the application tier.


See Also:

Chapter 6, "Introduction to Workload Management" for more detailed information about these features

The Oracle Clusterware and High Availability in Real Application Clusters

This section introduces the following high availability features:

High Availability with the Oracle Clusterware

Oracle Real Application Clusters 10g introduces a complete, integrated clusterware management solution on all Oracle Database 10g platforms. This clusterware functionality provides all of the features required to manage your cluster database including node membership, group services, global resource management, and high availability functions. You install the Oracle Clusterware as a separate install that you can complete independently or as a pre-requisite to the RAC installation process. Oracle database features such as Oracle 10g services use the underlying Oracle Clusterware mechanisms to provide their capabilities. Oracle also continues to support select third-party clusterware products on specified platforms.


See Also:

Chapter 14, "Making Applications Highly Available Using the Oracle Clusterware" for more information about the Oracle Clusterware

The Oracle Clusterware and High Availability

High availability configurations have redundant hardware and software that maintain operations by avoiding single points-of-failure. During outages, the Oracle Clusterware relocates the processing performed by the inoperative component to a backup component. The Oracle recovery processes quickly re-master resources, recover partial or failed transactions, and rapidly restore the system.

You can combine many Oracle products and features to create highly reliable computing environments. Doing this requires capacity and redundancy planning. In addition, consider your overall system costs and your return on investment. There are also other practical considerations such as selecting the appropriate hardware and deciding whether to use idle machines that are part of your high availability configuration.

Additional Oracle High Availability Features and Solutions

This section describes the following additional high availability solutions:

Connection Load Balancing with Real Application Clusters

The connection load balancing feature of Oracle Net Services automatically distributes connections among active instances. Connection load balancing does this based on the workload of each node and instance in a cluster.

Recovery Manager (RMAN) in Real Application Clusters

Recovery Manager (RMAN) is an Oracle tool that you can use to backup, copy, restore, and recover datafiles, control files, SPFILEs, and archived redo logs. You can invoke RMAN as a command-line utility or use Oracle Enterprise Manager.

A best practice is to configure RMAN so that all instances can access all of the archive log threads throughout your cluster database. In the event of media recovery, the recovering instance requires access to all of the archived redo log threads. You can simplify media recovery administration by ensuring that a recovering instance can access a local copy of the archive log threads from all of the instances in your cluster database.


See Also:

Chapter 7, "Configuring Recovery Manager and Archiving" for details about configuring RMAN for use with RAC and Oracle Database Backup and Recovery Advanced User's Guide for detailed information about RMAN

Data Guard

Oracle Data Guard works with standby databases to protect your data against errors, failures, and corruptions that might otherwise destroy your database. Data Guard protects critical data by automating the creation, management, and monitoring aspects of standby database environments. Oracle Data Guard automates the otherwise manual process of maintaining a transactional consistent copy of an Oracle database to recover from the loss of or damage to the production database.

Primary/Secondary Instance Configurations in Earlier Releases

If you are upgrading from a pre-Oracle 10g Primary/Secondary configuration, then the Database Upgrade Assistant (DBUA) creates a service on your database with one preferred instance and one available instance.