Contents

Title and Copyright Information

List of Figures

List of Tables

Send Us Your Comments

Preface

What's New in Cluster Software Administration?

Part I Introduction to Administering Real Application Clusters

1 Introduction to Real Application Clusters Administration

Administering Real Application Clusters Databases

Parameter Management in Real Application Clusters

Administering Storage in Real Application Clusters

General Administration Issues

Oracle Enterprise Manager in Real Application Clusters

Backup and Recovery in Real Application Clusters

Scaling Your Real Application Clusters Environment

Troubleshooting

2 Parameter Management in Real Application Clusters Environments

Administering the Server Parameter File in Real Application Clusters Databases

Backing Up the Server Parameter File

Setting the Server Parameter File Parameter Values for Real Application Clusters

Exporting the Server Parameter File For Backward Compatibility

Setting Parameters within the Server Parameter File

Specifying Comments in the Server Parameter File

Using Client-Side Parameter Files

Client-Side Parameter File Naming Conventions

Purpose of initsid.ora

Purpose of initdb_name.ora

Placement and Use of IFILE Parameters in Instance-Specific Files

Using Multiple IFILE Entries in Client-Side Parameter Files

Parameter Settings in Real Application Clusters

Unique Identification of Instances with Parameters in the Sever Parameter File

Instance-Specific Parameter Settings Using sid in the Server Parameter File

Types of Parameters in Real Application Clusters

Multi-Valued Parameters

Parameters That Must Be Identical Across All Instances

Parameters That Must Be Unique Across All Instances

Multiple Instance Issues for Parameters

The Startup Process and Parameters in Real Application Clusters

Special Startup Considerations for Traditional Parameter File Use

Starting Two Instances on Remote Nodes

IFILE Use in Traditional Parameter File Scenarios

Setting Instance Numbers

3 Administering Storage Components in Real Application Clusters

File Management Issues in Real Application Clusters

Adding Data Files

Automatic Undo Management In Real Application Clusters

Using Automatic Undo Management

Switching Undo Tablespaces

System Rollback Segment

Overriding Automatic Undo Management by Using Rollback Segment Undo

Manually Creating Rollback Segments

Public and Private Rollback Segments

Using Redo Log Files

Managing Trace Files and the Alert Files

Associating Instances, Users, and Resources with Free List Groups

Associating Instances with Free Lists

Assignment of New Instances to Existing Free List Groups

FREELIST GROUPS and MAXINSTANCES

Associating User Processes with Free Lists

SQL Options for Managing Free Space

Preallocating Extents to Free List Groups

The Sequence Number Generator

The CREATE SEQUENCE Statement

The CACHE Option

The ORDER Option

4 Administering Real Application Clusters Databases with Server Control, SQL, and SQL*Plus

Using SRVCTL to Administer Real Application Clusters Instances

Administrative Tasks That You Can Accomplish with SRVCTL

SRVCTL Cluster Database Tasks

SRCVTL Cluster Database Configuration Tasks

SRVCTL Command Syntax

srvctl start

srvctl stop

srvctl status

srvctl config

srvctl get env

Updating Your Configuration Information with SRVCTL

srvctl add db

srvctl add instance

srvctl delete instance

srvctl delete db

srvctl rename instance

srvctl move instance

srvctl set env

srvctl unset env

Importing and Exporting Raw Device Configuration with SRVCONFIG

Migrating Oracle8i Configurations to Oracle9i

Global Services Daemon (GSD)

UNIX GSD Implementations

Windows GSD Implementations

Administering the GSD on Windows Platforms

Administering Real Application Clusters Databases Using SQL and SQL*Plus

Starting Databases in Cluster Mode on UNIX

Starting Databases in Cluster Mode on Windows NT and Windows 2000

Using RETRY to Mount a Database in Cluster Mode

Setting and Connecting to Instances

The SET INSTANCE and SHOW INSTANCE Commands

The CONNECT Command

Verifying That Instances are Running

Shutting Down Real Application Clusters Instances

Quiescing A Real Application Clusters Database

How SQL and SQL*Plus Commands Affect Instances

Part II Using Oracle Enterprise Manager to Administer Real Application Clusters

5 Administering Real Application Clusters Databases with Oracle Enterprise Manager

Overview of Oracle Enterprise Manager Administration

Starting the Console

Displaying Objects in the Navigator Pane

Database-Specific File Structure

Instance-Specific File Structure

Configuration

Stored Configurations

Sessions

Locks

Resource Plans

Resource Plan Schedule

Using the Cluster Database Right-Mouse Menu

Starting a Cluster Database

Shutting Down a Cluster Database

Viewing the Cluster Database Operation Results

Status Details Tab

Output Tab

Viewing Cluster Database Status

General Tab

Status Details Tab

Creating a Job for a Cluster Database or Instance

Specifying Job Details

General Tab

Tasks Tab

Tasks for Cluster Database Destinations

Parameters Tab

Cluster Database Startup Task

Parameters for the Cluster Database Shutdown Task

Registering Cluster Database Events

Part III Backup and Recovery in Real Application Clusters

6 Configuring RMAN for Real Application Clusters

Configuring RMAN in Real Application Clusters: Overview

Configuring RMAN Control File Snapshots and Autobackups

Configuring the Snapshot Control File Location in RMAN

Configuring the Control File Autobackup Feature in RMAN

Selecting An Archived Log Configuration Scheme

RMAN Archive Log Configuration Schemes

Non-Shared Local Archive Logging

Configuring Initialization Parameters for Non-Shared Archiving Destinations

Backups Using Non-Shared Archive Logging

Restore and Recovery Using Non-Shared Archive Logging

Shared Read Local Archive Logging

One Remote Archive Logging

Configuring Initialization Parameters for One Remote Archive Logging

Backups for One Remote Archive Logging

No Cross-Mounting Archive Logging

Backups for No Cross-Mounting

Restore and Recovery for No Cross-Mounting

Cross-Mounting for Writing Archive Logging

Configuring Initialization Parameters for Cross-Mounting for Writing

Backups for Cross-Mounting for Writing

Restore and Recovery for Cross-Mounting for Writing

Optimizing the Reading of Local Archivelog Destinations

Cross Shared Cross Writing with One Local and N Remote Archive Logging

Central NFS Directory for High Availability Archive Logging

Hybrid Archive Logging

Implementation Considerations of Archive Logging Schemes

Using NFS on UNIX Platforms

Soft-Mounted NFS Directories

NFS Implemented for High Availability

Using Shared Drives on Windows Platforms

7 Backing Up Real Application Clusters Databases

Backups in Real Application Clusters

Choosing a Backup Method and Backup Type

Open Database Backups

Closed Database Backups

Online Backups and Real Application Clusters

Archiving Considerations for Real Application Clusters Environments

Changing the Archiving Mode in Real Application Clusters

Monitoring the Archiving Process in Real Application Clusters

Archive File Format and Destinations in Real Application Clusters

Backing Up the Archive Logs

Checkpoints and Log Switches

Forcing a Checkpoint

Forcing a Log Switch

Forcing a Log Switch on a Closed Thread

RMAN Backups in Real Application Clusters

Node Affinity Awareness

Performing an Open Backup Using RMAN

Backup Examples for Real Application Clusters Using Shared Directories

Using RMAN to Backup to Multiple Nodes with Several Channels

Avoiding the Thread Statement for Backing Up Archivelogs

Making Backups to Shared Archive Log Destinations

Backing Up Local Files from Each Node Using Non-Shared Archive Log Destinations

Backing Up All Files from One Node

Recovering Archive Logs from One Node

Restoring and Recovering Archive Logs from All Local Nodes Using Oracle Release 1 (9.0.1)

Restoring and Recovering Archive Logs from All Local Nodes Using Oracle 8.1.5 Or Earlier

8 Database Recovery in Real Application Clusters

Recovery from Instance Failures

Single Node Failure

Multiple-Node Failures

Fast-Start Checkpointing and Fast-Start On-Demand Rollback

Access to Data Files for Instance Recovery

Steps of Oracle Instance Recovery

Recovery from Media Failures

Complete Media Recovery

Block Media Recovery

Incomplete Media Recovery

Restoring and Recovering Redo Log Files

Recovery Using RMAN

Recovery Using Operating System Utilities

Disaster Recovery

Disaster Recovery Using RMAN

Parallel Recovery in Real Application Clusters

Parallel Recovery in Real Application Clusters

Parallel Recovery Using RMAN

Parallel Recovery Using Operating System Utilities

Setting the RECOVERY_ PARALLELISM Parameter

Specifying RECOVER Statement Options

Fast-Start Parallel Rollback in Real Application Clusters

RMAN Recovery Examples for Real Application Clusters

Recovery to Shared Archive Log Destinations

Restoring Archive Logs with RMAN

Part IV Scaling Your Real Application Clusters Environment

9 Adding Nodes and Instances and Deleting Instances in Real Application Clusters

Adding Nodes to a Cluster

Overview of Procedures for Adding Nodes

Adding a Node at the Clusterware Layer

Adding a Node at the Clusterware Layer on UNIX

Connect Node to Cluster

Install Vendor Cluster Software

Create Raw Devices

Adding a Node at the Clusterware Layer on Windows NT and Windows 2000

Connect Node to Cluster

Install Cluster Software

Create Disk Partitions

Create Links to Disk Partitions

Adding a Node at the Oracle Layer

Deleting Instances

Part V Reference

A Troubleshooting

Using Trace Files

Background Thread Trace Files

User Thread Trace Files

Alert File

Error Call Trace Stack

Contacting Oracle Support Services

Severe Errors

Glossary