NSR-DBM-ORA V5.0 Administration Guidemanuals.ts.fujitsu.com/file/8960/nmoag50.pdf · 2011-10-06 ·...

Edition August 2009

NSR-DBM-ORA V5.0

Administrator’s Guide

Comments… Suggestions… Corrections…The User Documentation Department would like to know your opinion on this manual. Your feedback helps us to optimize our documentation to suit your individual needs.

Feel free to send us your comments by e-mail to: [email protected]

Certified documentation according to DIN EN ISO 9001:2000To ensure a consistently high quality standard anduser-friendliness, this documentation was created tomeet the regulations of a quality management system which complies with the requirements of the standardDIN EN ISO 9001:2000.

cognitas. Gesellschaft für Technik-Dokumentation mbHwww.cognitas.de

Copyright and Trademarks

This manual is printed on paper treated with chlorine-free bleach.

Base of NSR-DBM-ORA V5.0 / this manual:EMC NetWorker®, Release 7.5, Multiplatform Version, Administrator’s Guide, Copyright © 1990 - 2009 by EMC Corporation.

Copyright © Fujitsu Technology Solutions 2009.

All rights reserved.Delivery subject to availability; right of technical modifications reserved.

All hardware and software names used are trademarks of their respective manufacturers.

Microsoft®, Microsoft Windows®, and Windows Server® are U.S. registered trademarks of Microsoft Corporation.

HP-UX® Release 10.20 and later and HP-UX Release 11.00 and later (in both 32 and 64-bit configurations) on all HP 9000 computers are Open Group UNIX® 95 branded products.

AMD Opteron(TM) is the trademark owned by Advanced Micro Devices, Inc.

Intel®, Intel Itanium®, Intel Pentium®, and Intel Xeon™ are trademarks or registered trademarks of IntelCorporation or its subsidiaries in the United States and other countries.

EMC® and NetWorker® are trademarks of EMC Corporation.

Oracle® is a registered US trademark of Oracle Corporation, Redwood City, California.

NetApp, Data ONTAP, SnapDrive, SnapManager, SnapMirror, SnapRestore, and WAFL are registered trademarks and Network Appliance Inc. in the U.S. and other countries.

Fujitsu and the Fujitsu Logo are trademarks of Fujitsu Limited.

Product Number: 0903500

mailto:[email protected]

http://www.cognitas.de

NSR-DBM-ORA V5.0 Administrator’s Guide 3

Preface

Chapter 1 Introduction

About this guide ...................................................................................................... 14 Naming conventions used in this guide............................................................... 15 New features and changes...................................................................................... 15

Improved monitoring of manual operations in NMC ................................ 16Improved reporting of proxy operations in NMC....................................... 16Improved security measures for sensitive data ........................................... 16IPv6 support (regular backups and restores) ............................................... 16NetWorker core debug messages................................................................... 16New platform support for proxy operations................................................ 17NSR-DBM-ORA parameter changes.............................................................. 17NWORA resource file requirements.............................................................. 17Oracle operating system authentication ....................................................... 17Restore from a specific volume pool.............................................................. 18

Importance of backups............................................................................................ 19 Product features ....................................................................................................... 19

Scheduled compared to manual backups ..................................................... 19Deduplication backups and restores ............................................................. 20Probe-based backups ....................................................................................... 21Proxy backups and restores ............................................................................ 22VMware support............................................................................................... 23Archived redo log backups ............................................................................. 23Control file autobackup ................................................................................... 23Automatic channel allocation ......................................................................... 24Backup and restore optimization ................................................................... 25Backup copies.................................................................................................... 25Backup of backup sets...................................................................................... 27Cluster and RAC backups and restores......................................................... 27Configuration wizards..................................................................................... 28Internationalization (I18N).............................................................................. 31Oracle Data Guard support............................................................................. 32Parallelism ......................................................................................................... 33Policy uniformity.............................................................................................. 33Restartable backups.......................................................................................... 34Retention policies ............................................................................................. 34Save set bundling.............................................................................................. 35Other Oracle features ....................................................................................... 41

Contents

NSR-DBM-ORA V5.0 Administrator’s Guide4

Software used in the NSR-DBM-ORA environment...........................................44NetWorker software.......................................................................................... 44NSR-DBM-ORA software ................................................................................ 45NetWorker PowerSnap Module software ..................................................... 45Oracle backup and recovery system .............................................................. 46

NSR-DBM-ORA components .................................................................................48 NSR-DBM-ORA backup and restore processes ...................................................49

Regular scheduled backup processes............................................................. 49Regular manual backup processes ................................................................. 51Deduplication backup processes .................................................................... 52Regular restore processes................................................................................. 54Deduplication restore processes ..................................................................... 55

Chapter 2 Software Configuration

Configuration roadmap.......................................................................................... 58 Verify the Oracle Server configuration................................................................. 59 Verify the NetWorker resources ............................................................................ 59

NetWorker Server resource.............................................................................. 59NetWorker user group privileges ................................................................... 60NetWorker Schedule resource......................................................................... 61NetWorker Device resources ........................................................................... 62NetWorker volume pools................................................................................. 62Firewall support ................................................................................................ 63

Configuring I18N support ..................................................................................... 64Requirements for I18N support ...................................................................... 64Configure I18N support ................................................................................... 64

Migrating a legacy configuration with the nsrnmoadmin command ..............65Requirements for using the nsrnmoadmin command for migration ........ 66Migration command syntax and options ...................................................... 67

Configuring a backup with the NMC wizard......................................................68About the backup configuration wizard ....................................................... 68Requirements for using the backup configuration wizard ......................... 68Configure a backup with the wizard ............................................................. 69

Configuring a backup with the NMC legacy method ........................................70About backup configuration with the legacy method................................. 70Create RMAN scripts for backups.................................................................. 70Customize the nsrnmo script .......................................................................... 75Configure a Group resource with NMC ........................................................ 80Configure a Schedule resource with NMC.................................................... 81Configure a Client resource with NMC......................................................... 81Configure save set bundling............................................................................ 84Configure policy uniformity ........................................................................... 85

Configuring a deduplication backup ................................................................... 86Requirements for a deduplication backup .................................................... 87Best practices for a deduplication backup..................................................... 87Configure a deduplication backup................................................................. 88

Configuring a probe-based backup ...................................................................... 90Requirements for a probe-based backup....................................................... 90Configure a probe-based backup.................................................................... 90


Chapter 3 Backup Procedures

Manual backup procedures.................................................................................... 98Requirements for a manual backup............................................................... 98Manual backup through the RMAN command line interface................... 99Manual backup through Oracle Enterprise Manager Backup Management Tools.................................................................................................................... 99NetWorker server bootstrap backup............................................................ 100Cancel a manual backup................................................................................ 101Monitor a manual backup ............................................................................. 101

Scheduled backup procedures ............................................................................. 104Test a scheduled backup................................................................................ 104Cancel a scheduled backup........................................................................... 105Monitor a scheduled backup ........................................................................ 106Automation of NSR_SERVER and NSR_GROUP parameter settings.... 107

Regular backup information in NetWorker indexes ........................................ 108 Deduplication backup information in NetWorker indexes ............................. 109

Chapter 4 Data Restore and Recovery

About restore and recovery.................................................................................. 112NetWorker indexes and policies used for restores .................................... 112Recovery configuration wizard .................................................................... 112

RMAN scripts for restore and recovery.............................................................. 115 Performing an Oracle data restore and recovery .............................................. 117

Requirements for an Oracle data restore..................................................... 117Restore through the RMAN command line interface ............................... 119Restore with Oracle Enterprise Manager Backup Management Tools ... 120Perform an Oracle data recovery ................................................................. 121

Chapter 5 Disaster Recovery

About disaster recovery........................................................................................ 124 Preparing an Oracle database for disaster recovery ......................................... 124

Create the DBID text file ................................................................................ 125Set up a postcommand script for backup of Oracle-related files............. 125Set up RMAN backups of the database and related files ......................... 126Set up RMAN backups of Recovery Catalog.............................................. 126

Performing a disaster recovery............................................................................ 126NetWorker server recovery ........................................................................... 127Oracle disaster recovery to a new host........................................................ 127

Sample postcommand script................................................................................ 128

Chapter 6 Cluster and RAC Systems

Considerations for cluster and RAC systems .................................................... 132Parameters in the RMAN script or session................................................. 132NetWorker software configuration .............................................................. 133Backup failover ............................................................................................... 133

Cluster systems ...................................................................................................... 133Roadmap for backup/restore configuration in a cluster system............. 134Backup failover ............................................................................................... 134

RAC systems........................................................................................................... 135RAC terminology............................................................................................ 135RAC backups and restores ............................................................................ 135Roadmap for backup/restore configuration in a RAC system................ 135


Setting up RAC nodes to back up to a local storage node ........................ 136Connect-time failover ..................................................................................... 138Creating RMAN backup scripts.................................................................... 140Creating RMAN restore scripts..................................................................... 141Archived redo logs.......................................................................................... 141

Chapter 7 Proxy Backups and Restores

Overview of proxy backups and restores ...........................................................144Types of proxy backups.................................................................................. 144Types of proxy restores................................................................................... 145

Proxy backup and restore system ........................................................................146 Proxy backup processes ........................................................................................147

Scheduled proxy backup................................................................................ 148 Proxy restore processes .........................................................................................153

Instant restore or rollback .............................................................................. 153Restore from secondary storage.................................................................... 155

Software configuration roadmap........................................................................ 156Basic configurations........................................................................................ 157Additional configurations for proxy operations ........................................ 157

Proxy backup requirements..................................................................................164Checking configuration consistency ............................................................ 164Performing proxy backups ............................................................................ 166

Proxy backup information in the NetWorker indexes ......................................167Entries in the client file index ........................................................................ 168Entries in the media database ....................................................................... 168NWORA resource file backup in the NetWorker indexes......................... 169

Proxy restore requirements.................................................................................. 170Creating RMAN scripts for proxy restores.................................................. 170Performing proxy restores ............................................................................. 171Relocating files during proxy restores ......................................................... 172Restoring to a different host .......................................................................... 173Point-in-time recoveries without a Recovery Catalog ............................... 173

Catalog synchronization for proxy backups ......................................................174Extra entries in the catalogs........................................................................... 174Control file versus recovery catalog............................................................. 175The NWORA resource file ............................................................................. 176Automatic catalog synchronization with the nsroraclecat program ....... 183

Proxy backups and restores on cluster systems.................................................185Proxy backup failover .................................................................................... 185Proxy backups from a virtual cluster client ................................................ 186Proxy backups from a physical cluster client.............................................. 187Restores from proxy backups on a cluster system ..................................... 188

Appendix A Parameters in an RMAN Session

Setting the NSR* parameters ............................................................................... 190 NSR* parameter definitions................................................................................. 191

Appendix B RMAN Commands

The delete expired backup command ................................................................ 200 The change...crosscheck and crosscheck commands ....................................... 200 The pool option of the backup command.......................................................... 200


The send command .............................................................................................. 201Syntax rules ..................................................................................................... 201Two ways to run the send command .......................................................... 203Precedence rules ............................................................................................. 204

The set duplex command .................................................................................... 205 The trace option of the backup command......................................................... 206

Appendix C Troubleshooting and Error Messages

Troubleshooting tips............................................................................................. 210 RMAN error messages......................................................................................... 211 NSR-DBM-ORA error messages......................................................................... 212

Error messages from the libnwora library.................................................. 213Error messages from the nsrnmoadmin program ..................................... 221Error messages from the nsrnmoinfo program.......................................... 223Error messages from the nsrnmoprobe program ...................................... 224Error messages from the nsrnmostart program......................................... 225

Appendix D Command Reference

Glossary

Index


Preface 9

Preface

This guide provides information on how to install, configure and manage the Fujitsu NSR-DBM-ORA release 5.0 software for Oracle data storage management on Linux or Solaris™ operating systems.

Audience This document is intended for use by system administrators and Oracle database administrators (DBAs) who are responsible for installing software and maintaining the Oracle Server backup and recovery systems. Operators who monitor Oracle database backups may also find this document useful.

Readers of this document are expected to be familiar with the following topics:

◆ Oracle terminology and concepts, especially those related to Oracle database backup and recovery.

◆ Backup and recovery procedures on an Oracle Server.◆ Disaster recovery procedures on an Oracle Server.

Relateddocumentation

The complete set of the Fujitsu NetWorker documentation can be found at http://ts.fujitsu.com/networker. Follow the link NetWorker backup solution. Most of the documents are in Adobe® Portable Document Format (PDF), and can be viewed by downloading and installing the Adobe (Acrobat) Reader as of V5.0. The Reader is available from Adobe at www.adobe.com. To install and use the Reader on the platform of your choice, refer to the instructions at the Adobe website.

The information in this guide is best used in conjunction with information from the following sources:

◆ NetWorker PowerSnap Module Installation and Administrator’s Guide (for the appropriate PowerSnap Module)

◆ NetWorker Administrator’s Guide (for the NetWorker server version)◆ NetWorker Administrator program Online Help◆ NetWorker Command Reference Guide (for NetWorker commands)◆ Man pages on UNIX systems (for NetWorker and NetWorker Module commands)◆ Oracle Server documentation◆ Oracle database backup and recovery documentation

http://www.adobe.com


Preface

Conventions used inthis document

This document uses the following conventions for special notices.

Note: A note presents information that is important, but not hazard-related.

CAUTION!A caution contains information essential to avoid data loss or damage to the system or equipment.

IMPORTANT!An important notice contains information essential to software or hardware operation.

Typographical conventionsThe following type style conventions are used in this document:

Normal Used in running (nonprocedural) text for:• Names of interface elements (such as names of windows, dialog boxes, buttons,

fields, and menus)• Names of resources, attributes, pools, Boolean expressions, buttons, DQL

statements, keywords, clauses, environment variables, functions, utilities• URLs, pathnames, filenames, directory names, computer names, links, groups,

service keys, file systems, notifications

Bold Used in running (nonprocedural) text for:• Names of commands, daemons, options, programs, processes, services,

applications, utilities, kernels, notifications, system calls, man pages

Used in procedures for:• Names of interface elements (such as names of windows, dialog boxes, buttons,

fields, and menus)• What the user specifically selects, clicks, presses, or types

Italic Used in all text (including procedures) for:• Full titles of publications referenced in text• Emphasis (for example, a new term)• Variables

Courier Used for:• System output, such as an error message or script • URLs, complete paths, filenames, prompts, and syntax when shown outside of

running text

Courier bold Used for:• Specific user input (such as commands)

Courier italic Used in procedures for:• Variables on the command line• User input variables

< > Angle brackets enclose parameter or variable values supplied by the user

[ ] Square brackets enclose optional values

| Vertical bar indicates alternate selections - the bar means “or”

{ } Braces indicate content that you must specify (that is, x or y or z)

... Ellipses indicate nonessential information omitted from the example


Preface

Where to get help Company, product, and technical information can be obtained as follows.

Information and servicesInformation and documentation on Fujitsu NetWorker products can be obtained as follows:

Customer service

Technical supportThe Support section of the Fujitsu Technology Solutions web site at http://ts.fujitsu.com/services provides links with contact information and information about available support programs.

If you have an active support agreement, you may contact your service provider.

If you do not have an active support agreement, contact your Fujitsu Technology Solutions sales representative to purchase a service contract or updates.

Your comments Your suggestions will help us continue to improve the accuracy, organization, and overall quality of the user publications. Please send your opinion of this document to:

[email protected]

If you have issues, comments, or questions about specific information or procedures, please include the title and, if available, the part number, the revision (for example, A01), the page numbers, and any other details that will help us locate the subject you are addressing.

Resource Address

Product information, documentation and downloads

http://ts.fujitsu.com/networker

NetWorker manuals http://manuals.ts.fujitsu.com Follow the links Storage solutions - Storage software.

Release notices http://manuals.ts.fujitsu.com Follow the links Storage solutions - Storage software.

Software patches http://ts.fujitsu.com/networker Follow the links NetWorker - Networker patches.

Resource Address

NetWorker training [email protected]

Support phone +49-180-54040

Support fax +49-180-5336779


http://manuals.ts.fujitsu.com

http://manuals.ts.fujitsu.com



http://ts.fujitsu.com/services



Preface

Introduction 13

1

This chapter includes the following major sections:

◆ About this guide............................................................................................................ 14◆ Naming conventions used in this guide .................................................................... 15◆ New features and changes ........................................................................................... 15◆ Importance of backups ................................................................................................. 19◆ Product features............................................................................................................. 19◆ Software used in the NSR-DBM-ORA environment................................................ 44◆ NSR-DBM-ORA components ...................................................................................... 48◆ NSR-DBM-ORA backup and restore processes ........................................................ 49

Introduction


Introduction

About this guideThis guide is organized as follows:

◆ Chapter 1, “Introduction” (this chapter) provides an overview of the main features of NSR-DBM-ORA release 5.0.

◆ Chapter 2, “Software Configuration” describes the procedures for configuring NSR-DBM-ORA backups and restores.

◆ Chapter 3, “Backup Procedures” describes the procedures for running manual NSR-DBM-ORA backups.

◆ Chapter 4, “Data Restore and Recovery” describes the procedures for configuring and running scheduled NSR-DBM-ORA backups.

◆ Chapter 5, “Disaster Recovery” describes the procedures for restoring and recovering from NSR-DBM-ORA backups.

◆ Chapter 6, “Cluster and RAC Systems” describes the procedures for configuring and running NSR-DBM-ORA backups and restores on cluster, OPS, and RAC systems.

◆ Chapter 7, “Proxy Backups and Restores” describes the procedures for proxy backups and restores of Oracle data on specific primary storage devices, performed by using NSR-DBM-ORA and the required PowerSnap Module software.

◆ Appendix A, “Parameters in an RMAN Session” describes the parameters that can be set for NSR-DBM-ORA backups and restores.

◆ Appendix B, “RMAN Commands” describes RMAN commands that concern the NSR-DBM-ORA software.

◆ Appendix C, “Troubleshooting and Error Messages” describes troubleshooting guidelines, RMAN error messages, and NSR-DBM-ORA error messages.

◆ Appendix D, “Command Reference” includes the text from the man pages that accompany NSR-DBM-ORA.

Naming conventions used in this guide 15

Introduction

Naming conventions used in this guideConsider the naming conventions used throughout this guide:

◆ “Oracle Server host” refers to the host where both the Oracle Server and NSR-DBM-ORA software are installed.

◆ “Proxy backup or restore” refers to an Oracle RMAN proxy copy backup or restore that is implemented by using snapshot technologies through the PowerSnap™ Module software.

◆ Unlike the NetWorker software, which uses the term recover for all data retrieval activities, Oracle distinguishes between restoring and recovering a database:

• “Restore” refers to retrieving individual datafiles from backup and storing the files on disk.

• “Recover” refers to applying the redo logs to make the database consistent.

This guide follows the Oracle terminology.

◆ “Regular backup or restore” refers to an NSR-DBM-ORA backup or restore of Oracle data that does not use snapshot technologies through the PowerSnap Module software.

◆ "UNIX" refers to both UNIX and Linux operating systems, unless specified otherwise.

New features and changesNSR-DBM-ORA release 5.0 supports the following new features:

◆ Deduplication backups and restores, performed by the NSR-DBM-ORA, NetWorker server, and Avamar® server software. “Deduplication backups and restores” on page 20 provides more information.

◆ Probe-based backups (also known as event-based backups) that are triggered when specified conditions are met. “Probe-based backups” on page 21 provides more information.

◆ Regular Oracle backups and restores on a VMware Virtual Machine on an ESX server, including support of the following advanced features:

• Distributed Resource Scheduler (DRS)

• VMotion

“VMware support” on page 23 provides more information.

◆ Backup and recovery configuration wizards that are integrated with the NetWorker Management Console (NMC) release 7.5 or later. “Configuration wizards” on page 28 provides more information.

◆ Use of the nsrnmoadmin command to migrate a legacy backup configuration to the configuration framework supported by the NMC-based backup configuration wizard. “Migration of configurations for the wizard” on page 30 provides more information.

◆ Internationalization support for both regular and proxy backups and restores, with specified limitations. NSR-DBM-ORA release 4.5 supported I18N for regular backups and restores. NSR-DBM-ORA release 5.0 extends the I18N support for proxy backups and restores with a supported PowerSnap Module release. “Internationalization (I18N)” on page 31 provides more information.


Introduction

Improved monitoring of manual operations in NMC

NSR-DBM-ORA release 5.0 supports the improved monitoring of manual backups and restores in NMC. Detailed displays of information about manual NSR-DBM-ORA backups and restores are available under the Sessions tab of the NMC Monitor window.

Improved reporting of proxy operations in NMC

The Group Details window in NMC reports detailed information for proxy backups. The information includes a complete listing of all the save set sessions run, the total size of the backups, and the sizes of the bootstrap and index save sets.

The Sessions tab of the Monitor window in NMC reports detailed information for proxy restore sessions.

Improved security measures for sensitive data

NSR-DBM-ORA release 5.0 uses NetWorker 7.5 or later lockbox services and encryption to ensure the secure storage and transfer of sensitive data and passwords:

◆ When you configure NSR-DBM-ORA backups through the NMC-based configuration wizard, the wizard uses the lockbox services to store sensitive data.

◆ When you migrate a legacy backup configuration to the configuration framework supported by the NMC-based wizard, the migration operation uses the lockbox services to store sensitive data.

◆ Sensitive data and passwords are encrypted before being transferred by the configuration wizards and NSR-DBM-ORA programs such as nsrnmoadmin, nsrnmoprobe, and nsrnmostart.

Lockbox password management is implemented through the NetWorker server.

IPv6 support (regular backups and restores)

NSR-DBM-ORA release 5.0 provides new support of Internet Protocol version 6 (IPv6) for regular (nonproxy) backups and restores.

The NetWorker Release Notice provides details on the use of NetWorker software in an IPv6 environment.

NetWorker core debug messages

NSR-DBM-ORA release 5.0 includes a new parameter, NSR_DPRINTF, for use with Tech Support only. You can set the parameter to specify that NetWorker core debug messages are printed to the file specified by NSR_DEBUG_FILE. (If NSR_DEBUG_FILE is not set, the debug messages are printed to the default location.)

Note: The NSR_DPRINTF parameter must be set through the parms options, not through the send command. The NSR_DEBUG_LEVEL parameter must also be set.

New features and changes 17

Introduction

New platform support for proxy operations

NSR-DBM-ORA release 5.0 provides new support of the Linux AMD64/EM64T platform for proxy backups and restores.

NSR-DBM-ORA parameter changes

NSR-DBM-ORA release 5.0 supports the following new parameters:

◆ NSR_DEDUP_BACKUP

◆ NSR_DEDUP_CACHE_ENABLED

◆ NSR_DEDUP_CACHE_TAG

◆ NSR_DEDUP_CHUNK_SIZE

◆ NSR_DEDUP_NODE

◆ NSR_DPRINTF

◆ NSR_RECOVER_POOL

◆ ORACLE_USER

The NSR_ORACLE_NLS_LANG parameter is no longer required to enable use of the configuration wizard with internationalized NSR-DBM-ORA in a non-English user locale.

NSR-DBM-ORA release 5.0 no longer supports the following parameters:

◆ NSR_RMAN_OUTPUT

◆ NSR_TMPDIR

NWORA resource file requirements

In NSR-DBM-ORA release 5.0, the NWORA resource file requirements are changed from previous NSR-DBM-ORA releases:

◆ The NWORA resource file is no longer mandatory for performing any type of NSR-DBM-ORA backup.

◆ The NWORA resource file is only required to enable the following NSR-DBM-ORA features:

• Catalog synchronization for proxy backups

• Policy uniformity

• Save set bundling

Oracle operating system authentication

NSR-DBM-ORA release 5.0 supports Oracle operating system authentication to connect to RMAN.

You can enable a scheduled NSR-DBM-ORA backup for operating system authentication by setting the ORACLE_USER parameter to the username of the Oracle operating system user, which is set up to connect to the Oracle database through operating system authentication.


Introduction

Operating system authentication is supported for a scheduled backup configured through the legacy method (without the wizard). The operating system authentication is not supported for the following:

◆ A scheduled backup configured through the new configuration wizard.

◆ A probe-based backup.

◆ A scheduled proxy backup.

Restore from a specific volume pool

By default, NSR-DBM-ORA and NetWorker use configuration settings and information in the media database to determine the backup volume to use for an NSR-DBM-ORA restore.

NSR-DBM-ORA release 5.0 supports the alternative restore method whereby you set the NSR_RECOVER_POOL parameter in the RMAN restore script to specify data restore from a specified volume pool if there are multiple copies (clones) of the backup on different volume pools.

The NSR_RECOVER_POOL parameter is not supported for proxy restores.

Importance of backups 19

Introduction

Importance of backupsThe reliability of computer equipment has improved greatly in recent years, but system and hardware failures still occur, sometimes with catastrophic results.

In a client/server environment, data can be lost due to hardware failures and user errors. Software bugs, procedural flaws, and simple user errors are common problems that necessitate database restores from backup storage media. A viable backup strategy can help in the recovery from disastrous situations.

Database administrators must recognize the importance of performing regular backups of database files and frequent backups of archived redo logs. If all archived redo logs are lost, a database can be recovered only to the time of the last consistent backup. Without backups and archived redo logs, the database cannot be recovered at all.

Product featuresThe following sections describe major features of NSR-DBM-ORA, Oracle Recovery Manager (RMAN), and Oracle Server software.

Certain sections refer to the send command, which is used to set parameters in the RMAN script for an NSR-DBM-ORA backup, as described in “The send command” on page 201.

The Oracle backup and recovery documentation provides more details on RMAN backups and restores and Oracle product features.

Scheduled compared to manual backups

An NSR-DBM-ORA backup can be either a scheduled or manual (unscheduled) backup:

◆ A scheduled NSR-DBM-ORA backup includes the following features:

• The backup is initiated by the NetWorker server.

• The backup start time depends on the settings in the NetWorker resources.

A regular scheduled backup starts at a time specified in the NetWorker Group resource.

A probe-based backup (or event-based backup) is a type of scheduled backup that starts when specified conditions are met, as described in “Probe-based backups” on page 21.

A proxy backup is only supported by using a scheduled backup, as described in “Proxy backups and restores” on page 22.

◆ A manual NSR-DBM-ORA backup includes the following features:

• The backup is initiated by a user on the NSR-DBM-ORA host through an Oracle backup utility, which can be RMAN or Oracle Enterprise Manager.

• The backup is performed by running the Oracle utility from the operating system command line. For example, a user starts a manual NSR-DBM-ORA backup by running an RMAN backup script from the command line.

Chapter 2, “Software Configuration,” provides details on the configuration of both scheduled and manual backups.


Introduction

Chapter 3, “Backup Procedures,” provides details on scheduled and manual backup procedures.

Deduplication backups and restores

NSR-DBM-ORA 5.0 software provides new support for deduplication backups and restores.

The NSR-DBM-ORA Release Notice provides details on the NetWorker requirements for the support of deduplication operations.

Main features of deduplication operations

An EMC Avamar® server (on Linux) interacts with the NetWorker server and NSR-DBM-ORA software during deduplication backups and restores. The Avamar server is configured as a NetWorker deduplication node, and deduplicates the data from various clients, including the NSR-DBM-ORA clients. (The Avamar server must also be available when a deduplication client resource is created.)

The initial backup to a deduplication node (Avamar server) will be a full backup. During subsequent deduplication backups, the Avamar server identifies redundant data blocks on the NSR-DBM-ORA client host and backs up only the unique blocks (not entire files) that contain changes. Only a single instance of any unique (atomic in Avamar terminology) data block is maintained on the Avamar server.

Since data deduplication is performed on the client host, deduplication backups typically require less time, network bandwidth, and storage space than regular NSR-DBM-ORA backups.

The Avamar client-side binary, nsravtar, is installed on the NSR-DBM-ORA client with the NetWorker client software. During a deduplication backup, the nsravtar process sends the deduplicated client data to the Avamar server. A deduplication NSR-DBM-ORA backup uses a hash cache on the client side. The Avamar and NetWorker documentation provides more details about caches in deduplication backups.

The following sections describe the processes involved in deduplication backups and restores:

◆ “Deduplication backup processes” on page 52

◆ “Deduplication restore processes” on page 55

A deduplication backup can be a manual or scheduled backup, including a probe-based backup.

The application of browse and retention policies and the selection of media pools is the same for a deduplication backup as for a regular NSR-DBM-ORA backup.

Since only the backup metadata (hash ID) is stored on the NetWorker backup device (generating a very small save set), the device should be configured as an advanced file type device (AFTD). The NetWorker Administrator’s Guide provides more information.

You must configure a scheduled or manual deduplication backup according to “Configuring a deduplication backup” on page 86.

“Deduplication backup information in NetWorker indexes” on page 109 describes the backup information stored in the NetWorker indexes, and how to delete the backups.

Product features 21

Introduction

Features not supported with deduplication operations

Deduplication operations do not support the following:

◆ Oracle proxy backups or restores.

◆ Cluster or RAC backups or restores.

◆ Internet Protocol version 6 (IPv6).

◆ AES encryption of deduplicated data through the NSR_AES_ENCRYPTION parameter.

◆ Checksumming of deduplicated data through the NSR_CHECKSUM parameter.

◆ Compression of deduplicated data through the NSR_COMPRESSION parameter.

If NSR_AES_ENCRYPTION, NSR_CHECKSUM, or NSR_COMPRESSION is set for a deduplication backup, NSR-DBM-ORA applies the AES encryption, checksumming, or compression, respectively, to only the metadata that is stored on the NetWorker storage node.

◆ Cloning or staging of deduplicated data on the Avamar server.

Only the metadata stored on the NetWorker storage node can be cloned or staged. However, deduplicated save sets on the Avamar server can be replicated to another Avamar server that has been configured as a replication node by Fujitsu Technology Solutions Technical Support representatives. Avamar documentation and the NetWorker Administrator’s Guide provide details on replication of deduplication backups.

Probe-based backups

NSR-DBM-ORA 5.0 software provides new support for probe-based backups (also known as event-based backups). A probe-based backup is a type of scheduled backup:

◆ The NetWorker server starts a regular scheduled backup based on a time interval.

◆ The NetWorker server starts a probe-based backup when specified conditions are met.

Probe-based backups are not supported for proxy backups or cluster/RAC environments.

Workflow of probe-based backups

A probe-based NSR-DBM-ORA backup starts when both of the following are true:

◆ (Condition 1) The current time is within a specified window of time (the backup window, defined by the probe start time and probe end time in the probe-enabled backup group resource).

◆ One of the following conditions is met:

• (Condition 2) A specified amount of time has elapsed since the previous probe-based backup.

• (Condition 3) One or all of the probes associated with the backup are successful, depending on the probe success criteria specified in the backup configuration.

The probe success criteria can be set in the NetWorker Group resource to the value Any or All.


Introduction

At specified probe intervals, the NetWorker server performs the following:

1. The server checks for condition 1, to determine if the current time is within the backup window.

2. If condition 1 is met, then the server checks for condition 2, to determine if a specified amount of time has elapsed since the last probe-based backup:

• If condition 2 is met, then the server starts the probe-based backup.

• If condition 2 is not met, then the server checks for condition 3, to determine if one or all of the probes are successful:

– If the probe success criteria is set to Any, and any one of the probes is successful, then the server starts the probe-based backup.

– If the probe success criteria is set to All, and all of the probes are successful, then the server starts the probe-based backup.

Types of probes

There are two different types of probes:

◆ An NSR-DBM-ORA probe.

The NSR-DBM-ORA probe is implemented through the NSR-DBM-ORA program nsrnmoprobe. The nsrnmoprobe program returns a successful result (signifying that the condition being checked has been met) when it detects either of the following:

• The number of Oracle redo logs generated since the previous probe-based backup exceeds a number known as the change threshold.

• A new database incarnation (reset log) has occurred since the previous probe-based backup.

◆ User-defined probes.

A user-defined probe checks if any other user-defined condition (other than the number of generated Oracle redo logs) has been met since the previous probe-based backup. (This option requires more advanced configuration.) To implement a user-defined probe, you must create a new script or program that checks for the condition, as described in “Configuring a probe-based backup” on page 90.

You must configure the required parameters and NetWorker resources to specify the probe interval, backup window, and any other settings for a probe-based backup.

“Configuring a probe-based backup” on page 90 provides details on how to configure a probe-based backup.

Proxy backups and restores

Proxy backups and restores provide continuous snapshot-based protection and availability of Oracle data on specific types of primary storage. A proxy operation involves both the NSR-DBM-ORA software and a specific PowerSnap Module.

Proxy backups create point-in-time copies or snapshots of Oracle data, store the snapshots on primary storage devices supported by the PowerSnap Modules (for Symmetrix®, CLARiiON®, and so on), and optionally back up the data to secondary storage (such as tape) from the point-in-time copies.

Proxy restores are used to restore the Oracle data that is backed up during the proxy backups.

Product features 23

Introduction

Chapter 7, “Proxy Backups and Restores,” provides more information on proxy backup and restore operations.

VMware support

NSR-DBM-ORA 4.5 provided support for regular backups and restores of an Oracle database installed on a VMware Virtual Machine (VM) on an ESX server.

NSR-DBM-ORA 5.0 extends this support with the support for the following advanced features of a VMware ESX server:

◆ VMotion — The VMotion feature enables migration of virtual machines from one ESX server to another while the servers are on. The migration is seamless to the applications running on the virtual machines, and NSR-DBM-ORAa user does not experience any disconnection. If a migration occurs during an NSR-DBM-ORA backup or restore, the backup or restore is not interrupted. VMware documentation provides details on the VM requirements for VMotion.

◆ Distributed Resource Scheduler (DRS) — The DRS feature enables dynamic balancing and allocation of resources across multiple ESX servers. Depending on the DRS policies set by the user, the DRS can migrate or recommend that users migrate a virtual machine to a different ESX server by using VMotion. DRS can also start (at boot-up time) a virtual machine on a different ESX server. Since this feature uses VMotion, if a migration occurs during an NSR-DBM-ORA backup or restore, the backup or restore is not interrupted.

The NSR-DBM-ORA Release Notice provides details on the NetWorker requirements for the support of VMware features.

Archived redo log backups

Archived redo log backups enable recovery of the database to its predisaster state. Without archived redo log backups, the database can be recovered only to the time of the last consistent Oracle backup. In this case, transactions that occurred between the time of the last consistent backup and the time of the database corruption will be lost.

Archived redo logs can be backed up by using the appropriate option of the RMAN backup command. Ensure that the NSR-DBM-ORA backups of archived redo logs are enabled.

“Backing up all archived logs from each node” on page 142 provides a sample script to back up the archived redo log files in a RAC system.

The appropriate Oracle backup and recovery documentation provides more information on setting up and running archived redo log backups.

Control file autobackup

RMAN performs a control file autobackup after each RMAN backup command if the control file autobackup has been enabled with the configure controlfile autobackup on command. Use this feature to restore the backup entries contained in the control file when the control file is lost and the Recovery Catalog is not available.

Specify persistent settings for the control file autobackups with the configure controlfile autobackup command. For example, enable control file autobackup and specify the persistent setting for the format of the control file autobackup name with the following commands:


Introduction

configure controlfile autobackup onconfigure controlfile autobackup format for device type ’sbt_tape’ to ’/NMO_%f/’

If the control file autobackup is set to on and the RMAN backup is performed with NSR-DBM-ORA, the control file autobackup will also be performed with NSR-DBM-ORA. As a result, one or more channels of device type sbt_tape must be allocated for the restore.

Note: Oracle also supports autobackup of the current server parameter file together with control file autobackup.

Automatic channel allocation

RMAN supports automatic channel allocation. This feature enables the configuration of persistent settings for automatic channels, for use in all RMAN sessions.

IMPORTANT!Manual and automatic channels are mutually exclusive and cannot be mixed in an RMAN session. The format of an automatic channel name of the device type for NSR-DBM-ORA backups and restores is ORA_SBT_n or ORA_SBT_TAPE_n, where n is the channel number. Do not use this name format for manual channel allocation for NSR-DBM-ORA. Otherwise, RMAN reports an error.With automatic channel allocation, specification of the send command before the backup or restore command causes the following error:RMAN-06422: no channels found for SEND command

You must use the configure channel...parms... command to set the NSR* parameters for automatic channels for an NSR-DBM-ORA backup. Do not use the send command or option to set the NSR* parameters for automatic channels if you plan to use scheduled backups.

Table 10 on page 191 lists all the NSR* parameters and their requirements.

Example 1 Using the configure channel command with parms option for automatic channels

Automatic channels are configured for NSR-DBM-ORA backups with the NetWorker server server1.fujitsu.com by typing the following configure channel...parms... command:

configure channel device type ’sbt_tape’ parms ’ENV=(NSR_SERVER=server1.fujitsu.com )’

This command sets the default parameters for all the automatic channels.

Example 2 Specifying parameter values per automatic channel

Specific NSR* parameter values can be set for different channels (for example, a separate setting of parameter NSR_GROUP for each channel) by typing the configure channel n device type...parms... command, where n represents a channel number.

An NSR-DBM-ORA debug file is specified for the second automatic channel by typing the following configure channel command:

configure channel 2 device type ’sbt_tape’ parms ’ENV=(NSR_DEBUG_FILE=/db/logs/backup.log)’

Product features 25

Introduction

Backup and restore optimization

If backup optimization is enabled with the configure backup optimization on command, RMAN skips selected files during a backup, based on several criteria. The Oracle backup and recovery documentation provides more details on these criteria.

Note: - To force a backup that would otherwise be skipped due to backup optimization, use the force option in the backup command.- When RMAN skips a backup due to backup optimization, it does not produce an error message. However, RMAN does issue a warning message similar to the following:skipping archive log file...

IMPORTANT!When using Oracle backup optimization with NSR-DBM-ORA backups and restores, run the crosscheck command regularly to synchronize the Recovery Catalog and NetWorker indexes. This ensures that backups expired by the NetWorker server are also marked as expired in the Recovery Catalog and RMAN does not skip a backup when a referenced backup has already expired in NetWorker.

The restore optimization function prevents RMAN from restoring a file if the original file is already in the correct location and contains the expected information.

Note: To force a restore that would otherwise be skipped due to restore optimization, use the force option in the restore command.

Backup copies

IMPORTANT!If more than one RMAN channel is used for backup copies of an NSR-DBM-ORA backup, parameter values set with the send command or option are passed by RMAN to the first backup channel only. Due to this send command limitation, NSR-DBM-ORA does not support the use of RMAN backup copies commands during scheduled backups. NSR-DBM-ORA supports backup copies with manual backups only.

Use the RMAN commands for backup copies only during manual backups when the parameters NSR_SERVER, NSR_DATA_VOLUME_POOL, NSR_DATA_VOLUME_POOL1, NSR_DATA_VOLUME_POOL2, and NSR_DATA_VOLUME_POOL3 are set with the parms option, not with the send command or option.

The set duplex command is deprecated (no longer supported by Oracle, but still functional in some Oracle releases). The Oracle documentation provides more details.

Despite the fact that RMAN provides different commands for duplexing backups, the rules for duplexing through NSR-DBM-ORA remain the same as with the set duplex command. Separate NetWorker pools must still be defined for each copy. “The set duplex command” on page 205 provides more information on the set duplex command and setting up NetWorker pools for each copy.


Introduction

Manual backups can be duplexed (up to four copies) by using one of the following commands:

◆ The configure...backup copies for device type sbt_tape to... command specifies persistent settings for duplexing backups through NSR-DBM-ORA.

For example, specify persistent settings for duplex copies of datafiles and archived redo logs (respectively) in NSR-DBM-ORA backups with the following types of configure commands:

configure datafile backup copies for device type ’sbt_tape’ to 2configure archivelog backup copies for device type ’sbt_tape’ to 2

◆ The backup command with the copies option applies to objects within the backup command. The backup...copies setting takes precedence over the persistent settings in the configure...backup copies command.

◆ The set backup copies command applies to all backup objects in the same run job.

In the following examples, the parms option is used to configure the channel and set the required parameters. These sample scripts must be invoked manually with RMAN, for example, by using the following command:

rman cmdfile script_name

Example 3 Using the set backup copies command in the RMAN script

The following RMAN script uses the set backup copies command to generate the backup copies. The parameters are set with the parms option, as required. The RMAN script must be invoked for a manual backup, not a scheduled backup:

run { set backup copies 4;allocate channel ch1 parms ’ENV=(NSR_SERVER=server_name, NSR_DATA_VOLUME_POOL=nmo1, NSR_DATA_VOLUME_POOL1=nmo2, NSR_DATA_VOLUME_POOL2=nmo3, NSR_DATA_VOLUME_POOL3=nmo4)’;backup format '%d_%U'tag tag_name(tablespace 'SYSTEM' );release channel ch1;

}

Example 4 Using automatic channels for backup copies

The following configure commands are used to configure RMAN automatic channels. (The configure commands could also be included in the RMAN script.) The configure...backup copies command generates the backup copies. The parameters are set with the parms option, as required. The RMAN script must be invoked for a manual backup, not a scheduled backup:

configure default device type to ’sbt_tape’;configure datafile backup copies for device type ’sbt_tape’ to 4;configure channel device type ’sbt_tape’ parms ’ENV=(NSR_SERVER=server_name, NSR_DATA_VOLUME_POOL=nmo1, NSR_DATA_VOLUME_POOL1=nmo2, NSR_DATA_VOLUME_POOL2=nmo3, NSR_DATA_VOLUME_POOL3=nmo4)’;

Product features 27

Introduction

The RMAN script invoked for the manual backup is as follows:

connect target sys/oracle@test;run {

backup format '%d_%U'tag tag_name(tablespace 'SYSTEM');

}

Backup of backup sets

RMAN supports the backup of backup sets. If Oracle data has been backed up with device type disk, NSR-DBM-ORA can be used to back up these backup sets from disk to NetWorker volumes.

For example, to back up all backup sets from disk to NetWorker volumes in a tape device, use the following command:

backup device type sbt backupset all

The backup set on disk can also be deleted with the delete input option in the backup device type sbt backupset... command. For example, to back up the backup sets that were created on disk more than a week ago and then remove the backup sets from disk, use the following command:

backup device type sbt backupset completed before sysdate-7 delete input

Cluster and RAC backups and restores

The NSR-DBM-ORA software supports backups and restores of cluster and Real Application Cluster (RAC) systems for high availability and parallelism.

A cluster system typically includes multiple nodes connected by a shared SCSI bus to which common storage is attached. Cluster services such as disk services can be defined and assigned their own IP addresses and names (virtual hosts). The services and their associated storage can migrate for failover between the physical nodes in the cluster.

After a cluster service is configured as a NetWorker client, NSR-DBM-ORA can be used with NetWorker server software to back up and restore an Oracle database associated with the service, independent of the actual node that provides the service.

A RAC system enables multiple Oracle instances across multiples nodes to access the same Oracle database at the same time. Oracle RAC is based on a cluster software infrastructure that provides concurrent access to the same storage and the same set of datafiles from all nodes in the cluster. All the database files reside on cluster-aware shared disks.

After RAC and the associated cluster system are properly configured, NSR-DBM-ORA enables Oracle backups on either a single node or several nodes of the RAC system. A parallel Oracle backup uses Oracle instances running in parallel on multiple nodes of the cluster.

NSR-DBM-ORA software supports restores of the Oracle data to any physical node in the cluster, regardless of which physical node originally performed the backup.

Chapter 6, “Cluster and RAC Systems,” provides more information on cluster and RAC systems and how to configure the systems for Oracle backup and restore operations that use the NSR-DBM-ORA software.


Introduction

Configuration wizards

NSR-DBM-ORA 5.0 software supports new backup and recovery configuration wizards that are integrated with the NetWorker Management Console (NMC). The configuration wizards used in NSR-DBM-ORA release 4.5 or earlier are replaced by the NMC-based wizards in NSR-DBM-ORA release 5.0.

Configuration wizard plug-in components are installed with the NSR-DBM-ORA software on the NSR-DBM-ORA client host. NMC loads the NSR-DBM-ORA wizard plug-in at run time. You can run the NSR-DBM-ORA wizards from the NetWorker Console Administration window, which you can start on any supported host by using a web browser session and specifying the Console server URL.

Main features of the wizards

The configuration wizards can be used to configure the following for an NSR-DBM-ORA client:

◆ Scheduled Oracle backups (either typical or customized).

“Configuring a backup with the NMC wizard” on page 68 provides details on using the backup configuration wizard.

◆ RMAN scripts for Oracle data restores to the original host.

◆ RMAN scripts for Oracle database duplication to either a local or remote host.

“Recovery configuration wizard” on page 112 provides details on the recovery configuration wizard.

The configuration wizards support NetWorker servers and clients in a stand-alone or cluster environment.

The new wizards provide improved security and ease of management for backup and recovery configurations, compared to the wizards from NSR-DBM-ORA release 4.5 or earlier.

Features of the backup configuration wizardThe backup configuration wizard can perform the following:

◆ Configure a new NetWorker Client resource for an NSR-DBM-ORA backup.

◆ Configure a new or use an existing NetWorker Group resource for the backup Client resource.

◆ Configure new or use existing browse and retention policies for the backup Client resource.

◆ Optionally save a copy of the configuration settings from the Client resource to a nsrnmo script or RMAN script.

◆ Modify a backup configuration that was created with the NMC-based NSR-DBM-ORA configuration wizard.

◆ Modify a backup configuration that was created either with the wizard from NSR-DBM-ORA release 4.5 or earlier, or with the legacy method, but only after the configuration has been migrated according to “Migration of configurations for the wizard” on page 30.

Product features 29

Introduction

Note: NSR-DBM-ORA 5.0 still supports the legacy method of backup configuration whereby you manually create a nsrnmo script and RMAN script, set any required parameters in the NWORA resource file, and configure the NetWorker resources with NMC (without the wizard). “Configuring a backup with the NMC legacy method” on page 70 describes the legacy configuration method.

“Configuring a backup with the NMC wizard” on page 68 provides information on using the wizard to create or modify a backup configuration.

Features of the recovery configuration wizard“Features of the recovery configuration wizard” on page 113 provides information on the features of the recovery configuration wizard.

Backup configuration storage with the wizardThe backup configuration wizard stores the configuration information in a new hidden attribute named Backup Config in the NetWorker Client resource. Do not modify the new attribute manually. If you use the wizard to create a backup configuration, you must use the wizard to modify the configuration.

In a backup configuration created with the wizard, the Client resource stores the parameters and RMAN commands (except for sensitive data, such as passwords) that would have been stored in the nsrnmo and RMAN scripts and NWORA resource file in NSR-DBM-ORA release 4.5 or earlier.

Note: The backup configuration wizard stores sensitive data securely by using NetWorker lockbox services.

The wizard does not automatically generate a nsrnmo script and RMAN script, but provides an option to save configuration settings to a nsrnmo script or RMAN script on disk. These scripts are meant for reference use only.

When you run a backup that was configured through the wizard, the backup does not access any nsrnmo or RMAN script file that the wizard saved to disk. Instead, the backup accesses the configuration information that the wizard stored in the Client resource.

The wizards do not store any database connection passwords in any RMAN scripts that are saved to disk.

The backup configuration wizard also sets the following values in the Client resource:

◆ Backup Command attribute is set to:

nsrnmostart -C [-c virtual_client]

◆ Save Set attribute is set to:

RMAN:/<database_Net_service_name>_level[/<name_of_first_tablespace_or_datafile>]

where:

• database_Net_service_name is the the name of the Oracle database to bebacked up.

• level is the Oracle backup level (full, incr_differential, or incr_cumulative).

• name_of_first_tablespace_or_datafile is the name of the first tablespace or datafile in the backup list. This name is included in Save Set attribute only if a partial database (subset of the entire database) is selected for backup.


Introduction

Features not supported with the wizards

The configuration wizards cannot perform the following:

◆ Configure backups or restores in a RAC environment.

◆ Configure proxy backups or restores that require a PowerSnap Module.

Proxy backups and restores must be configured without the wizard, as described in Chapter 7, “Proxy Backups and Restores.”

◆ Configure two different database backups in the same Client resource.

A separate Client resource must be created for each database to be backed up on the same client host.

◆ Modify a backup configuration that was created either with the wizard from NSR-DBM-ORA 4.5 or earlier, or with the legacy method, unless the configuration has been migrated according to “Migration of configurations for the wizard” on page 30.

Migration of configurations for the wizard

The backup configuration wizard stores the scheduled backup configuration in the Client resource by using a configuration storage framework that is incompatible with the configurations created either through the wizard from NSR-DBM-ORA release 4.5 or earlier, or through a legacy method (without a wizard).

The backup configuration wizard can modify only the following types of backup configurations:

◆ Configurations created with the new wizard.

◆ Configurations migrated to the configuration storage framework that is supported by the new wizard.

If you have a backup configuration that was created with the wizard from NSR-DBM-ORA release 4.5 or earlier, or with a legacy configuration method, you must migrate the configuration before you can use the new wizard to modify it. The recommended migration method is to use the nsrnmoadmin command, as described in “Migrating a legacy configuration with the nsrnmoadmin command” on page 65.

Note: Migration of a proxy backup configuration is not supported. “Requirements for using the nsrnmoadmin command for migration” on page 66 describes the requirements and limitations of migration with the nsrnmoadmin command.

Migration converts legacy configuration files and resources (the nsrnmo script, RMAN script, NWORA resource file, and Client resource) to the configuration storage framework used by the new wizard. Migration with the nsrnmoadmin command does not create a new Client resource; the migration modifies an existing Client resource, such that you can then use the wizard to modify it.

The migration process stores the following in a hidden attribute in the Client resource:

◆ Environment variable settings extracted from the nsrnmo script. (Any scripting commands in the nsrnmo script are ignored.)

◆ RMAN commands and options extracted from the RMAN script.

◆ NSR_BUNDLING, NSR_INCR_EXPIRATION, and NSR_ORACLE_NLS_LANG settings extracted from the NWORA resource file.

Product features 31

Introduction

The migration process also sets the following values in the Client resource:

◆ Backup Command attribute is set to:

nsrnmostart -C [-c virtual_client]

◆ Save Set attribute is set to:

RMAN:database_Net_service_name

Internationalization (I18N)

NSR-DBM-ORA I18N is the capability of the NSR-DBM-ORA software to operate in a non-English environment or locale without itself generating non-ASCII data. After you set up NSR-DBM-ORA I18N as described in “Configuring I18N support” on page 64, NSR-DBM-ORA can process and display non-ASCII data that is passed to it by the operating system, NetWorker software, and Oracle software. The non-ASCII data can include text messages, dates, times, numbers, and so on.

The NetWorker Administrator’s Guide provides details on the different supported languages and operating system.

NSR-DBM-ORA I18N is supported for the following:

◆ Regular (nonproxy) backups and restores

◆ Proxy backups and restores with a supported PowerSnap Module release, as described in the NSR-DBM-ORA Release Notice.

The term internalization is used differently in the NetWorker documentation (as opposed to this NSR-DBM-ORA documentation). NetWorker server and client documents refer to internationalization as the capability of the NetWorker software to both process non-ASCII data as input and generate non-ASCII data as output in a non-English locale.

When NSR-DBM-ORA I18N support is set up as described in “Configuring I18N support” on page 64, NSR-DBM-ORA supports non-ASCII data in the following:

◆ Pathnames of nsrnmo scripts

◆ Pathnames of preprocessing and postprocessing scripts

◆ Pathnames of RMAN scripts

◆ Strings passed as command line options to the nsrnmosh, nsrnmoadmin, nsrnmoinfo, and nsroraclecat commands

◆ Tablespace names and datafile paths

◆ The tag string of the RMAN backup command

◆ Usernames in the connection strings to the target database and recovery catalog

Note: Oracle does not recommend the use of non-ASCII text in the Oracle database usernames. Due to Oracle limitations, ASCII text must be used for the password of the target database.

◆ Values of the following parameters:

• NSR_DEBUG_FILE

• NSR_ORACLECAT_LOG_FILE

• NSR_RMAN_ARGUMENTS

• NSR_SB_DEBUG_FILE


Introduction

Note: Support of non-ASCII values for ORACLE_SID and TNS_ADMIN is dependent on the Oracle software. Support of non-ASCII values for LD_LIBRARY_PATH, LIBPATH, and SHLIB_PATH is dependent on the operating system.

◆ Additional values input in the NSR-DBM-ORA wizard, including the following:

• New or existing names of NetWorker Group resources

• Time data based on the time zone of the user

◆ Information displayed on the following screens of the NSR-DBM-ORA wizard:

• Review and Edit the RMAN Script screen

• Review and Accept the Client Configuration screen

◆ Content of output files created by the NSR-DBM-ORA wizard, such as nsrnmo scripts and RMAN scripts

Due to Oracle limitations, ASCII text must be input in the wizard for the following:

◆ ORACLE_HOME path

◆ Net service name of the Oracle target database, recovery catalog, or duplicate database

The deduplication backup process, nsravtar, on the NSR-DBM-ORA client generates messages in English only.

When NSR-DBM-ORA I18N support is set up, NSR-DBM-ORA generates debug messages in English only. NSR-DBM-ORA generates error messages in the nmo.messages.raw file in a language-independent binary form, readable by the nsr_render_log program only. The nmo.messages.raw file replaces the nmo.messages file used in previous NSR-DBM-ORA releases. (The log file does not contain Oracle Server or RMAN errors.)

The NetWorker Administrator’s Guide provides information on how to use the nsr_render_log program to read any language-independent binary file, such as nmo.messages.raw.

The PowerSnap Module documentation provides details on the PowerSnap options that support non-ASCII values.

“Configuring I18N support” on page 64 describes how to set up NSR-DBM-ORA I18N support.

Oracle Data Guard support

NSR-DBM-ORA software supports Oracle Data Guard, an Oracle data availability and protection solution that involves the primary database and one or more standby databases over an IP network. As transactions occur in the primary database and redo data is written to the local redo logs, Data Guard automatically transfers this redo data to the standby sites and applies it to the standby databases, synchronizing them with the primary database.

RMAN backups of datafiles, archived redo logs, and possibly other files can be offloaded to a physical standby database, and the backups used to recover the primary database or a standby database. RMAN and Data Guard documentation provides information on how to configure and back up a physical standby database, and use the backups to recover the primary or standby database.

Product features 33

Introduction

To configure NSR-DBM-ORA backups and restores in a Data Guard environment:

1. Follow the instructions in Oracle documentation on how to set up the required RMAN configurations, for example, to use a Recovery Catalog and the DB_UNIQUE_NAME parameter.

2. Install and configure the NSR-DBM-ORA and NetWorker client software on the primary database host, and on each physical standby database host involved in the backups and restores.

3. Configure a Client resource on the NetWorker server for the primary database host and each physical standby database host involved in the backups and restores. In the Client resource of the primary database host, specify the hostname of the physical standby host in the Remote Access attribute if you set NSR_CLIENT to the primary database hostname in the following step.

4. Create an RMAN script for the primary database and the standby database, setting the same NSR_CLIENT parameter value in both. The NSR_CLIENT value used for a backup should be the same as the NSR_CLIENT value used for the restore of that backup. Setting NSR_CLIENT to the primary hostname might be preferable.

Parallelism

Parallelism is a feature that enables NSR-DBM-ORA backup or restore streams of Oracle data from several clients, or many Oracle data streams from one client, at the same time. Different types of configuration settings enable different types of parallelism during NSR-DBM-ORA operations:

◆ The Parallelism attribute in the NetWorker Server resource specifies the maximum number of backup save streams that the NetWorker software allows to arrive in parallel at the server.

◆ The Parallelism attribute (a hidden attribute) in the NetWorker Client resource specifies the maximum number of data streams that the Oracle Server sends in parallel to the NetWorker server or storage node during an NSR-DBM-ORA backup.

◆ For proxy operations, the PowerSnap parameter settings specify the maximum number of parallel save streams on the proxy client host.

Note: The degree of proxy backup or restore parallelism is not controlled by the allocation of multiple channels in the RMAN script. Oracle uses only one of the allocated channels for the proxy backup or restore, unless specific backup options are used.

◆ Configurations on a RAC system enable parallel Oracle backups and restores with the NSR-DBM-ORA software on multiple nodes of a cluster. Chapter 6, “Cluster and RAC Systems,” provides more details.

Policy uniformity

If policy uniformity is enabled, NSR-DBM-ORA automatically enforces the uniformity of the browse and retention policies between all the dependent save sets in a scheduled backup cycle (whether or not save set bundling is enabled). When save set bundling is also enabled, all the save sets in a bundle receive the same browse and retention policies.


Introduction

After NSR-DBM-ORA performs an incremental scheduled backup, if the browse and retention policies of the save sets in the backup are longer than the policies of preceding dependent save sets in the same backup cycle, the NSR-DBM-ORA program nsrnmostart changes the policies of all save sets in the cycle to match the longest policy of the new incremental save sets. NSR-DBM-ORA modifies the policies recorded in the NetWorker media database. As a result, backups cannot expire and become recyclable before other dependent backups from the same backup cycle.

The NSR-DBM-ORA software does not enforce policy uniformity for a manual backup, except when a subsequent scheduled backup is dependent on the manual backup, and then the policies of the manual backup are modified accordingly.

Policy uniformity does not depend on whether save sets are stored on separate volumes. For example, if parts of a save set bundle are split onto separate volumes, all the save sets in the bundle still receive the same browse and retention policies.

“Configure policy uniformity” on page 85 provides information on how to configure policy uniformity for NSR-DBM-ORA backups.

Restartable backups

RMAN can back up files that have not been backed up since a specified time. For example, to continue the backup of a database that was canceled 2 days ago, use the following command:

backup device type sbt database not backed up since time ’sysdate-2’

RMAN compares the given time in this command with the completion time of the most recent backup of a datafile to determine if the datafile requires backup. The appropriate Oracle backup and recovery documentation provides more information.

The following sections provide more information on how to cancel NSR-DBM-ORA backups:

◆ “Cancel a manual backup” on page 101

◆ “Cancel a scheduled backup” on page 105

Retention policies

RMAN provides an Oracle retention policy for backups. An Oracle retention policy is based on the recovery window or redundancy; it is not based on a defined time period, such as a year. Oracle considers a backup obsolete when it is no longer required according to the Oracle retention policy setting. Oracle checks the retention policy of a backup when the report obsolete... or delete obsolete... command is run.

NSR-DBM-ORA supports the Oracle retention policy with some restrictions, since the NetWorker server has its own browse and retention policies to specify how long data is available for recovery. NetWorker browse and retention policies are based on a user-defined time period. Since the Oracle retention policy is independent from that of the NetWorker server, and there is no mechanism to synchronize these policies, the NetWorker and Oracle policies could conflict.

Product features 35

Introduction

Example 5 Expiring a backup

A NetWorker server might expire an NSR-DBM-ORA backup that is not yet obsolete in RMAN, depending on the NetWorker browse and retention policies. Use of NSR-DBM-ORA cannot guarantee that the NetWorker server will not expire a backup volume until all the RMAN backups on that volume have been removed from the RMAN catalog.

To avoid conflicts, perform either of the following:

◆ Disable the Oracle retention policy (and then use only the NetWorker server policy) with the following command:

configure retention policy to none

If the Recovery Catalog is used, exempt a backup from the retention policy with one of the following commands:

change backupset...keep until/forever...backup...keep until/forever...

◆ (If you want to use the Oracle retention policy) Set the NetWorker browse and retention policies to be long enough that backups are kept on the backup volumes until the Oracle retention policy makes them obsolete.

Set the NetWorker policies in the NetWorker Client resource for scheduled backups or through the NSR_SAVESET_BROWSE and NSR_SAVESET_RETENTION parameters. “Configure a Client resource with NMC” on page 81 provides more information on how to set NetWorker policies for NSR-DBM-ORA backups.

IMPORTANT!Run the crosscheck command on the NSR-DBM-ORA backups before running report obsolete or delete obsolete backups of the device type sbt_tape. This ensures that backups expired by the NetWorker server are flagged as expired in the RMAN catalog. As a result, RMAN can correctly identify which backups are not needed according to the Oracle retention policy.

For example:1. Run the following command to synchronize the RMAN Catalog and

NetWorker indexes:crosscheck backup;

2. Run the following command to delete all obsolete backups defined by the current Oracle retention policy:delete obsolete;

Save set bundling

If NSR-DBM-ORA save set bundling is configured, NSR-DBM-ORA automatically creates a save set bundle for each scheduled backup cycle of an Oracle database object, by grouping all the dependent save sets from the same backup cycle into the save set bundle.

A “backup cycle” includes a full backup of the database object and all subsequent incremental backups that are dependent on the full backup.

Note: NSR-DBM-ORA does not support save set bundling for proxy backups. NSR-DBM-ORA performs save set bundling for regular Oracle backups only.


Introduction

During staging operations with NetWorker server release 7.4 and later, if the staging criteria determine that a particular NSR-DBM-ORA save set should be staged (migrated) and the save set is part of a save set bundle, the NetWorker server stages the entire save set bundle. If the nsrstage command is used to manually stage one or more save sets from a save set bundle, all the save sets in the bundle are staged.

Note: After a staging operation during which all the save sets in a bundle are staged, the resulting available space on the staging device might exceed the lower-water mark specified in the staging policy.

The NetWorker Administrator’s Guide provides details on how to work with staging policies and perform automatic and manual staging operations through the NetWorker server.

“NSR-DBM-ORA scheduled backups and save set bundling” on page 36 describes NSR-DBM-ORA save set bundling during regular scheduled backups, and how to configure save set bundling.

If policy uniformity is configured, NSR-DBM-ORA automatically enforces the uniformity of browse and retention policies for all the dependent save sets of the same scheduled backup cycle or same save set bundle, which ensures that incremental backups do not persist after the backups they depend on have expired.

“Other Oracle features” on page 41 provides more information on how policy uniformity relates to save set bundling. Save set bundling and policy uniformity can be enabled and disabled independently.

NSR-DBM-ORA scheduled backups and save set bundling

Use the information in the following sections to plan a save set bundling strategy and enable save set bundling for NSR-DBM-ORA scheduled backups. Both are performed to prepare for staging operations with NetWorker server 7.4 and later.

“Configure save set bundling” on page 84 provides information on how to configure save set bundling for NSR-DBM-ORA scheduled backups.

RMAN backup levels and dependenciesAn RMAN backup can be either full or incremental:

◆ A full (or stand-alone full) backup includes every used block of the database objects listed in the RMAN backup script (unused blocks might be skipped). This type of backup is created when you do not specify a backup level with the RMAN backup command. A full backup cannot be the parent of a subsequent incremental backup. (Incremental backups cannot be dependent on a stand-alone full backup.)

◆ An incremental backup is either level 0 or level 1. Incremental backups are created when you specify either incremental level=0 or incremental level=1 with the RMAN backup command. Incremental backups are dependent on preceding incremental backups in the same scheduled backup cycle:

• A level 0 incremental is physically identical to a full backup, but is recorded as incremental in the RMAN repository. (A level 0 backup may also be referred to as "full" in other sections of this guide.)

• A level 1 incremental can be either of the following:

– A differential backup, which contains only the data blocks changed since the most recent incremental backup, whether level 0 or 1. The differential backup is dependent on the preceding level 0 or 1 backup. Incremental backups are differential by default.

Product features 37

Introduction

– A cumulative backup, which contains only the data blocks changed since the most recent level 0 incremental backup. The cumulative backup is dependent on the preceding level 0 backup.

RMAN backup sets and NSR-DBM-ORA save setsNSR-DBM-ORA save set bundling is performed at the backup set level.

An NSR-DBM-ORA backup (either full or incremental) generates one or more NetWorker save sets. The backup is also composed of one or more RMAN backup sets, with each backup set containing one or more backup pieces. A backup piece contains data blocks from one or more Oracle database files. Each NSR-DBM-ORA save set corresponds to one backup piece.

Note: Backup set and backup piece are Oracle terms. Save set is a NetWorker term.

◆ A control file, parameter file (or spfile), archived log, or datafile cannot span more than one backup set.

◆ A control file or parameter file backup cannot span more than one backup piece.

◆ An archived log or datafile backup can span more than one backup piece in a backup set.

It is possible to determine which backup set contains a specific datafile (by querying the v$ views in the Oracle database), but not which backup pieces within the backup set contain the datafile.

An Oracle backup set contains either of the following:

◆ The backup of a control file, parameter file, or archived log, which is always performed as a full backup.

Note: The backup of a control file, parameter file, or archived log is always placed in its own save set bundle.

◆ Full or incremental backups of one or more Oracle datafiles.

A backup set can include both full and incremental backups. For example, a backup set might contain incremental backups of datafiles 1 and 2 and a full backup of datafile 3.

Creating NSR-DBM-ORA save set bundlesIf save set bundling is enabled, all dependent save sets from the same backup cycle are included in the same save set bundle. Save sets are dependent when two or more save sets are required to restore a database object. (All the NSR-DBM-ORA save sets from a backup set are placed into the same save set bundle.)

At the end of a full or level 0 scheduled backup, the NSR-DBM-ORA software creates a new save set bundle for the backup set from the backup. If subsequent incremental backups are performed that are dependent on the level 0 backup, NSR-DBM-ORA adds their save sets to the save set bundle from the level 0 backup.

A separate save set bundle is created for each scheduled backup cycle of a particular Oracle database object, where a backup cycle consists of a full or level 0 backup of the object and all the subsequent incremental backups that are dependent on the level 0 backup.


Introduction

A save set bundle contains one of the following:

◆ The save sets from a stand-alone full backup, with no other dependent save sets.

For example, the save sets from the backup of a control file, parameter file, or archived log (always performed as a full backup) are placed in their own save set bundle.

◆ The save sets from a level 0 backup of an Oracle object and all subsequent incremental backups in the same backup cycle of the object.

When an incremental backup occurs and NSR-DBM-ORA cannot find a preceding dependent backup in any existing bundles, NSR-DBM-ORA creates a new save set bundle for the incremental backup.

Save sets from a manual backup are placed into a save set bundle only if a subsequent scheduled backup is dependent on them. The manual backup save sets are placed in the save set bundle at the same time as the dependent save sets from the scheduled backup.

For save set bundling purposes, you can simultaneously run multiple backup cycles that back up different objects from the same database, as long as different files are backed up by the different cycles. For example, one cycle can back up datafiles 1 and 2, while another cycle backs up datafiles 3, 4, and 5 from the same database. The cycles can also be of different lengths. For example, one cycle can last a week, while another concurrent cycle lasts several weeks.

Note: The backup copies feature and save set bundling of backup copies are not supported with NSR-DBM-ORA scheduled backups. Backup copies created during a manual backup are independent of each other, and each copy goes to a different NetWorker volume.

If an error occurs during save set bundling, the bundling operation fails but the scheduled backup can finish successfully. Information about the bundling failure is printed to the savegrp output and to the debug file specified by NSR_DEBUG_FILE.

How the nsrnmostart program performs save set bundling

The NSR-DBM-ORA program nsrnmostart automatically places save sets into a save set bundle at the end of a scheduled backup, before the nwora.res file is backed up. “Configure save set bundling” on page 84 provides information on the nwora.res file.

To perform save set bundling, the nsrnmostart program connects to the Oracle database by attempting to use the login and password from the RMAN script. If a login and password are not available from the script, the program uses the ORACLE_SID value from the nsrnmo script to search the nwora.res file for the NSR_ORACLE_CONNECT_FILE parameter, and uses the connection strings from the specified connection file. “Configure save set bundling” on page 84 provides more details on the requirements of save set bundling.

After connecting to the Oracle database, the nsrnmostart program obtains all the required information about the backups from the Oracle control file (by querying the v$backup_set, v$backup_piece, and v$backup_datafile tables). The control file can store only a limited number of backup entries. When the maximum number of entries is exceeded, old entries in the control file are overwritten by new ones. Save set bundling is successful only if information in the control file about backed-up save sets has not been overwritten. The Oracle documentation provides information about proper maintenance of the control file and how much backup information the control file can store.

Product features 39

Introduction

The nsrnmostart program creates a save set bundle for each full or incremental level 0 backup. The program adds the save sets from subsequent incremental backups to the bundles of the full or level 0 backups they are dependent on.

The name that the nsrnmostart program assigns to a save set bundle is a number corresponding to the save time of the oldest save set in the bundle. NSR-DBM-ORA provides the NetWorker server with the list of save sets contained in each save set bundle.

After a scheduled backup, the NetWorker server stores the save set bundle name and the list of save sets it contains in the media database. You can view the bundle information by using the mminfo command, as described in “Save set bundling information in the media database” on page 39.

Example 6 Save set bundling for a one-week scheduled backup cycle of a tablespace

A one-week scheduled backup cycle of a tablespace includes a level 0 backup of the tablespace on Sunday and a level 1 backup every other day of the week. The save set bundle for the cycle is created during the Sunday backup, and save sets from each level 1 backup are added into the same bundle. The complete bundle contains the save sets from the seven daily backups of the tablespace. A new bundle is created for the next backup cycle during the following week.

NetWorker staging restrictionsWhen planning the strategy for NSR-DBM-ORA save set bundling, consider the following NetWorker staging restrictions:

◆ NetWorker cannot simultaneously stage all the save sets from a save set bundle if some of the save sets were backed up to separate volumes. NetWorker simultaneously stages save sets only if they are located on the same staging volume. Example 8 on page 40 provides more information.

To ensure the proper staging of all the save sets from a save set bundle, do not split the backup between different staging volumes. If required, split the backup into different backup cycles, with each cycle going to a separate volume.

◆ NetWorker staging policies must not cause the save sets of an NSR-DBM-ORA backup cycle to be staged before the cycle is complete. For example, if a one-week NSR-DBM-ORA cycle starts on Sunday, the staging policy must not cause the partially complete save set bundle to be staged before the final backup of the cycle occurs on Saturday.

To prevent a staging operation from splitting an NSR-DBM-ORA backup cycle, adjust the NetWorker staging policy accordingly. For example, adjust the policy so that older save sets are staged before new ones, or adjust the high-water and low-water marks.

The NetWorker Administrator’s Guide provides details on how to work with staging policies and perform automatic and manual staging operations through the NetWorker server.

Save set bundling information in the media databaseThe NSR-DBM-ORA software stores information about each save set bundle in the NetWorker media database. “NetWorker software” on page 44 provides more information about the media database.


Introduction

Query the media database by using the NetWorker command, mminfo, with the appropriate options:

◆ The mminfo -r command can display the name of the bundle associated with a save set. For example, the following command displays a list of all save sets and their bundles:

mminfo -a -r "ssid,ssbundle"

◆ The mminfo -q command can display all the save sets in a specific bundle. For example, the following command displays all the save sets in the bundle named 12983479182:

mminfo -a -q "ssbundle=12983479182"

The NetWorker Command Reference Guide and the UNIX man pages provide more information on the mminfo command and its available options.

Examples of save set bundles and stagingThe following examples illustrate different aspects of save set bundling, and how splitting the save set bundles across volumes can affect staging operations.

Example 7 Save set bundle join

Two save set bundles are created by separate level 0 backups of files A and B. Then a new backup set is created by a level 1 backup of both files A and B. Since the new backup set is dependent on both of the preceding level 0 backups, NSR-DBM-ORA combines all three backups into the same save set bundle.

If the original file A backup has the oldest backup time, NSR-DBM-ORA places the new backup set (from the level 1 backup) into the save set bundle of the (level 0) file A backup. NSR-DBM-ORA then moves the original (level 0) file B backup into the save set bundle with the other two backups. This method of combining existing bundles into a new save set bundle is known as a “save set bundle join.”

Example 8 Splitting a save set bundle across volumes

In both of the following cases, a save set bundle is split across multiple volumes. The parts of the save set bundle on different volumes must be staged separately by the NetWorker server:

◆ A backup uses multiple channels so the backup set spans multiple volumes. All the save sets belong to the same backup set and save set bundle, but parts of the bundle are stored on different volumes. During staging, only the save sets on the same volume can be staged together.

◆ A level 0 backup of file A is performed to volume A. An incremental backup of file A is then performed to volume B. Although both backups are recorded as belonging to the same save set bundle, the save set bundle is split across volumes. During staging, only the save sets on the same volume can be staged together.

Example 9 Using save set consolidation to re-unite a save set bundle

A level 0 backup of file A is performed to volume A. A level 1 backup of file A is then performed to volume B. Save set consolidation is used to merge the save sets from these two backups onto the same volume. Bundle names are preserved when save sets are moved from volume to volume by save set consolidation. The consolidated backup is staged as a single save set bundle.

Product features 41

Introduction

Other Oracle features

This section describes supported features of the Oracle Server software. The NSR-DBM-ORA software supports the Oracle releases on specific platforms, as outlined in the NSR-DBM-ORA Release Notice.

Examples of the Oracle RMAN features that NSR-DBM-ORA supports are as follows:

◆ Fast incremental backups that use change tracking files.

◆ Proxy backups and restores of archived redo logs.

Note: Oracle does not support proxy backups of datafiles or archived redo logs that reside on Oracle Automated Storage. Oracle Automated Storage is also known by the term Oracle Automated Storage Management (ASM).

◆ Channel backup failover and backup piece restore failover.

◆ Management of backup duration and throttling.

◆ Backups and restores of data residing on Oracle Automated Storage.

◆ Flash recovery area and flashback database.

The Oracle Recovery Manager documentation provides a complete list of the RMAN features.

When using Oracle RMAN features with NSR-DBM-ORA, consider the following:

◆ A flash recovery area stores and manages files related to the recovery of a particular database.

To back up RMAN disk backups, control file autobackups, and archived redo logs from the flash recovery area to NetWorker volumes:

a. Allocate or configure one or more channels with the sbt_tape device type.

b. Back up the files with one of the following RMAN commands:

backup recovery areabackup recovery files

Note: Whether or not a flash recovery area is enabled, the backup recovery files command can be used to perform the backup.

For example, the following sequence of RMAN commands can be used to configure an automatic channel for NSR-DBM-ORA and back up the files from the flash recovery area:

configure default device type to ’sbt_tape’;configure channel device type ’sbt_tape’ send ’NSR_ENV=(NSR_SERVER=server1)’;backup recovery files;

◆ If multiple channels are used for an RMAN backup command and one of the channels fails, Oracle fails over to another channel to continue the backup job.

For example, if two channels are configured with different NetWorker volume pools and one of the channels fails over to the other channel during a backup, the entire backup goes to the volumes in the pool of that remaining channel.

◆ Before using the backup command with the duration...minimize load option, consider:


Introduction

• The minimize load option might impact the tape streaming since the transfer rate of data sent by RMAN might be slow with this option, depending on the duration value.

Note: This is not a concern if you use the NetWorker backup to disk feature.

• The minimize load option might cause an NSR-DBM-ORA scheduled backup to be timed out if RMAN does not send data to the NetWorker Module within the time frame specified in the Inactivity Timeout field of the corresponding NetWorker Group resource.

◆ Starting with release 10.1, RMAN does not print database connection strings (user/password@Netservicename) to the session output. As a result, the parameter NSR_RMAN_OUTPUT is not required (to suppress the connection strings in the log files) for an RMAN scheduled backup.

Oracle11g specific features

NSR-DBM-ORA 5.0 supports the following major Oracle11g features:

◆ Data Recovery Advisor

◆ Improved integration with Data Guard

◆ Archival backup through the RMAN backup...keep command

◆ Improved archived redo log management through the configure archivelog deletion policy command

◆ Recovery catalog enhancements, such as virtual private catalogs and the import and merging of recovery catalogs

◆ Multisection or intrafile backup and validation, with the backup of a single large datafile over multiple channels

◆ Improved block media recovery, with the blockrecover command being replaced by the recover...block command

◆ Configurable backup compression through the configure compression algorithm to command

◆ Block change tracking support in Data Guard

◆ Backup of read-only transportable tablespaces

◆ Oracle Enterprise Manager enhancements, with new interfaces for the Data Recovery Advisor

◆ Oracle Globalization Support enhancements

To enable NSR-DBM-ORA support of two of the Oracle11g features, Data Recovery Advisor and archival backup, you must perform the additional configuration procedures described in “Data Recovery Advisor” on page 42 and “Archival backup feature” on page 43.

The appropriate Oracle documentation provides more information on the Oracle11g features.

Data Recovery AdvisorThe Oracle Data Recovery Advisor is a new tool in Oracle11g. Integrated with RMAN and Oracle Enterprise Manager (OEM), the tool enables a DBA to diagnose and repair database failures.

Product features 43

Introduction

Before you can use the Data Recovery Advisor to invoke an RMAN restore script that involves NSR-DBM-ORA to repair a database failure, automatic channels must be configured to specify at least the mandatory parameters NSR_SERVER and NSR_CLIENT.

Note: The NSR_SERVER and NSR_CLIENT parameters are the minimum parameters required to perform a restore. Other NSR-DBM-ORA parameters may also be specified for the automatic channel configuration.

To enable the use of Data Recovery Advisor with Oracle11g and NSR-DBM-ORA:

◆ If automatic channels have not been configured for NSR-DBM-ORA backups, use the following commands to ensure the basic automatic channel configuration:

configure channel device type ’sbt_tape’ parms ’ENV=(NSR_SERVER=NetWorker_server_name, NSR_CLIENT=NSR-DBM-ORA_client_name)’;

configure channel device type 'sbt_tape' parallelism number_of_restore_channels;

◆ If automatic channels are already configured for NSR-DBM-ORA backups, no

additional configuration steps are required.

Archival backup featureWith Oracle11g, the RMAN backup...keep forever command enables the creation of an archival backup that is exempt from Oracle backup retention policies (but not automatically exempt from NetWorker retention policies). The archival backup is all-inclusive because every file required to restore a database is backed up to a single disk or tape location.

To enable the use of the RMAN backup...keep forever command with NSR-DBM-ORA:

1. Configure an Archive type Pool resource through the NetWorker server.

2. Specify that the backup data must go to the Archive pool by performing one of the following:

• Set the pool selection criteria accordingly on the NetWorker server.

• Set the NSR_DATA_VOLUME_POOL parameter in the RMAN backup script.

3. Set the parameter value NSR_SAVESET_RETENTION=forever through the send command in the RMAN backup script.

Note: Ensure that the NSR_RETENTION_DISABLED option is not set in the RMAN backup script used with NSR-DBM-ORA.

The NetWorker Administrator’s Guide provides more information on how to configure resources and specify pool selection criteria through the NetWorker server.


Introduction

Software used in the NSR-DBM-ORA environmentThe following sections describe the different software components involved in the NSR-DBM-ORA environment.

NetWorker software

The NetWorker software comprises a high-capacity, easy-to-use data storage management solution that protects and helps manage data across an entire network. The NetWorker software simplifies the storage management process and reduces the administrative burden by automating and centralizing data storage operations.

The NetWorker software enables a user to:

◆ Perform automated backups during nonpeak hours

◆ Administer, configure, monitor, and control NetWorker functions from any computer on a network

◆ Centralize and automate data management operations

◆ Increase backup performance by simultaneously sending more than one save stream to a single device

◆ Optimize performance by using parallel save streams to multiple backup devices or storage nodes

NetWorker client/server technology uses the network remote procedure call (RPC) protocol to back up data. The NetWorker client software consists of client-side services and user interface programs.

The NetWorker server software consists of several server-side services and programs that:

◆ Oversee backup and restore processes

◆ Maintain client configuration files

◆ Maintain an online client file index and online media database, which together comprise the online indexes on the NetWorker server

During a backup, the NetWorker server makes an entry in the online client file index and records the location of the data in the online media database. These entries provide recovery information required for all backed-up data.

After a scheduled backup, the NetWorker server sends a record of the bootstrap file to the default printer. This is a printed record of the dates, locations, and save set ID numbers for the server’s online indexes that are required for restoring data. Keep the bootstrap printout on file as a quick reference in the event of a disaster, such as a disk crash or server failure:

◆ The NetWorker Release Notice provides information on how to install NetWorker software.

◆ The NetWorker Administrator’s Guide provides information on how to configure and use NetWorker software.

Software used in the NSR-DBM-ORA environment 45

Introduction

NSR-DBM-ORA software

NSR-DBM-ORA software is an add-on module for the NetWorker server that enables you to perform the following tasks:

◆ Manual Oracle backups

◆ Scheduled Oracle backups

◆ Restores of Oracle backup data

◆ Automated media management

NSR-DBM-ORA software provides the following features:

◆ Capability to integrate database and file system backups, to relieve the burden of backup from the database administrator while allowing the administrator to retain control of the restore process.

◆ Automatic database storage management through automated scheduling, autochanger support, electronic tape labeling, and tracking.

◆ Support for backup to a centralized backup server.

◆ High performance through support for multiple, concurrent high-speed devices such as digital linear tape (DLT) drives.

Together with the NetWorker server, NSR-DBM-ORA augments the backup and recovery system provided by the Oracle Server and provides a storage management solution that addresses the need for cross-platform support of enterprise applications.

Proxy backups and restores of Oracle database files residing on specific types of primary storage devices are supported. These proxy operations are performed by NSR-DBM-ORA, with the NetWorker server and the appropriate NetWorker PowerSnap Module software.

The NSR-DBM-ORA Release Notice provides information on how to install the NSR-DBM-ORA software.

NetWorker PowerSnap Module software

The NetWorker PowerSnap Module software works with the NetWorker server and NSR-DBM-ORA software to perform RMAN proxy backups and restores of Oracle data that resides on specific types of primary storage.

Each type of primary storage requires a different type of PowerSnap Module. A proxy backup performed with the PowerSnap Module involves taking a snapshot of the Oracle data. The snapshot resides on the primary storage to ensure high availability of the database and eliminate downtime latency on the Oracle Server host.

During a proxy backup, the Oracle data can optionally be copied from the snapshot to a NetWorker storage device. The data is copied to the secondary storage by either the Oracle Server host or a proxy client host that is separate from the Oracle Server host.

The following sources provide information on how to configure and run proxy backups and restores with the NSR-DBM-ORA and PowerSnap Module software:

◆ Chapter 7, “Proxy Backups and Restores”

◆ NetWorker PowerSnap Module documentation


Introduction

Oracle backup and recovery system

This section describes the system used for regular Oracle backup and recovery.

The NSR-DBM-ORA and NetWorker server and client software function with the standard Oracle backup and recovery system to produce an efficient Oracle data storage management system. The implementation of an Oracle backup and recovery strategy requires a knowledge of how the NetWorker software works together with the Oracle components.

The regular Oracle backup and recovery system consists of the following:

◆ Oracle Server

◆ Oracle Recovery Manager (RMAN)

◆ Recovery Catalog (optional)

◆ Backup Management Tools for Oracle Enterprise Manager (optional)

The NetWorker software consists of the following components:

◆ NetWorker server

◆ NetWorker client

◆ NSR-DBM-ORA

Figure 1 on page 47 illustrates the architecture of the regular Oracle backup and recovery system, and shows the functional relationship between the NetWorker and Oracle software components. The NSR-DBM-ORA media management library (MML) is integrated with the Oracle Server during the NSR-DBM-ORA installation on the Oracle Server host. The NetWorker Management Console program (used to configure the NetWorker resources) and the Oracle Enterprise Manager (OEM) Backup Manager can be optionally located on a separate platform, referred to as the GUI console.

Software used in the NSR-DBM-ORA environment 47

Introduction

Figure 1 Regular Oracle backup and recovery system

Oracle Recovery Manager

During the Oracle backup and restore operations, NSR-DBM-ORA acts as an intermediary between the NetWorker server and Oracle Recovery Manager (RMAN).

RMAN is the main Oracle utility for backing up, restoring, and recovering Oracle datafiles, control files, and archived redo log files. It stores information about its operations in the control file of the backed-up database (target database) and, optionally, in the Recovery Catalog on the Oracle Server host.

RMAN provides the following features:

◆ Online and offline Oracle database backups

◆ High performance through parallel backups and restores

◆ An intelligent interface to Oracle databases

◆ Well-defined backup strategies

◆ Checks for corrupt blocks during backups and restores

◆ Block-level incremental backups and restores

The Oracle Recovery Manager documentation provides more information on RMAN.

Oracle Recovery Catalog

The Recovery Catalog is a collection of Oracle database tables that contain structural information about Oracle database backups and restores. The collection includes information about the following types of objects:

◆ Backup sets and backup pieces

◆ Image copies

◆ Proxy copies

OEM Backup

Manager

Oracle Server System GUI Console

NSR-DBM-ORA

MML

Oracle Server

NetWorker Server

Storage Medium

NetWorker Server

Recovery Catalog

NetWorker Client

RMAN

Oracle Database FIles

GEN-000170

NetWorkerManagement Console


Introduction

◆ Archived redo logs

◆ Target database schema

◆ Stored scripts (user-created sequences of RMAN commands)

The Recovery Catalog is maintained by the RMAN utility. This utility uses the catalog information or the database control file to determine how to perform requested backup and restore operations. Store the Recovery Catalog in the dedicated Recovery Catalog database.

The Oracle Recovery Manager documentation provides more information on the Recovery Catalog, including the implications of not using it for backups and restores.

NSR-DBM-ORA componentsTable 1 on page 48 lists the NSR-DBM-ORA components that are installed on the NSR-DBM-ORA client host during the NSR-DBM-ORA and NetWorker software installation. Unless specified otherwise, the files are located in the same directory as the NetWorker client software.

Table 1 NSR-DBM-ORA components

Name on UNIX Description

fdiff.jarnwora.jarnwora_res.jar

Plug-in jar files for the NSR-DBM-ORA backup and recovery configuration wizards.

libcommonssl.7.5.so In a subdirectory under /opt/nsr/lib. An NSR-DBM-ORA library that is required for communication with NetWorker.

libnwora.xx In a subdirectory under /opt/nsr/lib. The main NSR-DBM-ORA library (known as Media Management Library in Oracle documentation) that is loaded by the Oracle backup or restore process.

nsrnmo The nsrnmo script that contains parameter settings for a particular scheduled backup. The script is only used for a scheduled backup that has been configured through NMC (the legacy configuration method without the wizard).

nsrnmo.sh In /opt/nsr. A sample template for the nsrnmo script. You should copy and customize this template, and save the original file in /etc.

nsrnmoadmin The program that is used either to create resource settings in the NWORA resource file, or to migrate legacy backup configurations to the new format supported by the configuration wizard.

nsrnmodrpostcmd A sample postcommand script that can be customized to back up specific files at the end of a scheduled backup, in preparation for disaster recovery.

nsrnmoinfo The program that determines the NetWorker volumes required to restore specified Oracle backup pieces from NSR-DBM-ORA backups.

nsrnmoprobe The program that probes for the number of generated Oracle logs as a condition that triggers probe-based backups.

nsrnmostart The program that invokes a scheduled backup on the Oracle Server, launching RMAN by passing arguments to the Oracle utility. The program optionally runs precommand and postcommand scripts.

nsroraclecat Not available on Linux Itanium or Solaris AMD64/EM64T (platforms that do not support proxy backups). The program that is used to remove RMAN catalog entries during automatic catalog synchronization for proxy backups.

nsrorara The program that performs operations on behalf of the NSR-DBM-ORA configuration wizard on the local or remote NSR-DBM-ORA host.

NSR-DBM-ORA backup and restore processes 49

Introduction

NSR-DBM-ORA backup and restore processesThis section describes the processes involved in regular backups and restores.

To provide storage management services for Oracle Server data, NSR-DBM-ORA implements the media management interface, also known as the Oracle System Backup to Tape (SBT). This interface comprises a media management library (MML) that is integrated with the Oracle Server during the NSR-DBM-ORA installation on the Oracle Server host.

Oracle loads the NSR-DBM-ORA MML at run time, and the Oracle Server backup sessions call the NSR-DBM-ORA MML software routines to back up and restore Oracle data to and from the media controlled by the NetWorker server.

A regular Oracle backup can be performed in either of two ways:

◆ By issuing the appropriate commands through the RMAN command line interface.

◆ By using the Oracle Enterprise Manager Backup Management Tools, which include an optional graphical user interface to the RMAN utility.

RMAN establishes connections with the target database. Each RMAN channel starts an Oracle Server backup session that performs the backup. During the backup, the following occurs:

1. The Oracle Server backup sessions read the datafile, control file, or archived redo log being backed up, and then write it to the NetWorker server through the MML.

2. The NetWorker server stores the Oracle data to the appropriate backup storage device.

At the end of the Oracle backup, the NetWorker server updates the online client and media indexes with information about the backup media used and the Oracle data it contains.

Regular scheduled backup processes

Note: “Deduplication backup processes” on page 52 provides information specifically for a deduplication backup.

A scheduled NSR-DBM-ORA backup is a backup of Oracle data initiated by the NetWorker server.

A regular scheduled backup includes the following interactions:

1. At the scheduled backup start time, the main NetWorker service, nsrd, starts the configured group’s backup by invoking the savegrp program.

2. The savegrp program requests that the NetWorker client-side service, nsrexecd, run the savefs program (mainly used for file system backups; it has no usage for Oracle backups).

3. The savefs program sends back information to the savegrp program.

4. The savegrp program contacts the nsrexecd service to start the backup.

5. For each client in the backup group and each of the client’s save sets, the following sequence of events occurs:

a. The nsrexecd service starts a specific process, depending on how the backup was configured:


Introduction

– If the backup was configured through the legacy method (without the wizard), nsexecd starts the nsrnmo script, which then sets the required parameters and invokes the NSR-DBM-ORA program, nsrnmostart.

– If the backup was configured through the new configuration wizard, nsexecd starts the nsrnmostart program directly.

b. The nsrnmostart program starts the RMAN utility to run the required RMAN backup script.

c. From this point on, the scheduled Oracle backup sessions follow the same steps as described in “Regular manual backup processes” on page 51.

Note: At the end of a scheduled Oracle backup, the savegrp program also automatically backs up the NetWorker server bootstrap and the Oracle client file indexes. The bootstrap and client indexes are not automatically backed up at the end of a manual NSR-DBM-ORA backup.

Figure 2 on page 50 shows how the Oracle Server, NetWorker server, and NSR-DBM-ORA processes interact during a regular scheduled NSR-DBM-ORA backup that was configured through the legacy method (without the wizard).

In a scheduled backup configured through the wizard, the nsrnmo process is removed from the picture, and the nsrexecd and nsrnmostart processes interact directly with each other.

Figure 2 Regular scheduled NSR-DBM-ORA backup

Backup Pieces

savegrp

nsrindexd nsrmmdbd

nsrd

nsrmmd

Media Database Client File Index

Oracle Server (NetWorker Client)

NetWorker Server Storage Medium

Oracle process (channel)

nsrexecd

rman

nsrnmostart

nsrnmo

savefs

interprocess communication

data

tracking information

NSR-DBM-ORA

MML

GEN-000172

NetWorker Module elements


Introduction

Regular manual backup processes

Note: “Deduplication backup processes” on page 52 provides information specifically for a deduplication backup.

A manual NSR-DBM-ORA backup is a user-initiated backup through an Oracle backup utility, RMAN or Oracle Enterprise Manager. NSR-DBM-ORA considers a backup scheduled through Oracle Enterprise Manager to be a manual backup.

When the RMAN utility is invoked for a backup, Oracle Server backup sessions call the NSR-DBM-ORA media management library (MML) software routines to initiate the backup.

A manual NSR-DBM-ORA backup includes the following interactions:

1. The Oracle Server backup session that loads the NSR-DBM-ORA MML (one such backup session for each allocated channel) contacts the nsrexecd service to obtain the NetWorker client information.

2. The Oracle Server backup sessions contact the main NetWorker service, nsrd, to obtain the required authorization and the identifier of the nsrmmd process that manages the requested backup device.

3. The Oracle Server backup sessions send the backup data to the NetWorker media service, nsrmmd, to store on the appropriate backup volumes.

4. Tracking information is stored in the NetWorker online indexes:

• The nsrmmd service records tracking information in the NetWorker media database by using the nsrmmdbd service.

• The Oracle backup sessions send tracking information to the NetWorker client file index by using the nsrindexd service.

Figure 3 on page 52 shows how the Oracle Server, NetWorker server, and NSR-DBM-ORA processes interact during a regular manual NSR-DBM-ORA backup.


Introduction

Figure 3 Regular manual NSR-DBM-ORA backup

Deduplication backup processes

“Deduplication backups and restores” on page 20 describes the features of a deduplication NSR-DBM-ORA backup.

For scheduled deduplication backups, the steps from “Regular scheduled backup processes” on page 49 still apply.

A deduplication NSR-DBM-ORA backup includes the following interactions:

1. The Oracle Server backup session that loads the NSR-DBM-ORA MML (one such backup session for each allocated channel) contacts the nsrexecd service to obtain the NetWorker client information.

2. Each Oracle Server backup session runs one nsravtar process at any given time.

3. The Oracle Server backup session contacts the main NetWorker service, nsrd, to obtain the required authorization and the identifier of the nsrmmd process that manages the requested backup device on the NetWorker server or storage node.

4. The Oracle Server backup session sends the backup data in the NetWorker save set format to the nsravtar process.

5. The nsravtar process applies deduplication logic to the backup data by using information in a local cache or by contacting the Avamar server, and sends the unique data blocks to the Avamar server (NetWorker deduplication node).

6. The nsravtar process sends a hash ID (one hash ID per save set) to the Oracle Server backup session.

NSR-DBM-ORA

MML

Backup Pieces

nsrd

nsrindexd nsrmmdbd

nsrmmd

Client File Index Media Database




data


nsrexecd

RMAN

Oracle process

(channel)

GEN-000171

NetWorker Module element


Introduction

7. The Oracle Server backup session sends the hash ID in save set format to the NetWorker media service, nsrmmd, to store on the NetWorker backup device.

8. Tracking information is stored in the NetWorker online indexes:

• The nsrmmd service records tracking information in the NetWorker media database by using the nsrmmdbd service. The media database entry contains extended attributes that identify the backup as a deduplication backup, as described in “Deduplication backup information in NetWorker indexes” on page 109.

• The Oracle backup session sends tracking information to the NetWorker client file index by using the nsrindexd service.

Figure 4 on page 53 shows how the Avamar server, NetWorker, and NSR-DBM-ORA processes interact during a deduplication NSR-DBM-ORA backup. In the figure, the thickness of the data arrows indicates the amount of data passed between processes:

◆ All the data being backed up is passed between the Oracle Server backup session and nsravtar process.

◆ Hash IDs and new data blocks (never backed up before by the Avamar server) are passed between the nsravtar process and Avamar server.

◆ A single hash ID for the entire NSR-DBM-ORA save set is passed between the Oracle Server backup session and NetWorker storage node.

Figure 4 Deduplication NSR-DBM-ORA backup

NetWorkerServer

NetWorkerStorage Node

NetWorker Client /utility node

GSAN

Node 1

Node 2

GEN-001111

Deduplicateddata / hash ID

Hash ID

Oracledata / hash ID

NetWorkerDeduplication Node

(Avamar Server)

NMO Client

Oracle

Oracle Serverbackup session

with NSR-DBM-ORA MML

nsravtar


Introduction

Regular restore processes

An NSR-DBM-ORA restore can be performed in either of two ways:


Or


The RMAN utility starts Oracle Server sessions on the target database. These Oracle Server sessions initiate the restore by calling NSR-DBM-ORA media management library (MML) software routines.

Note: “Deduplication restore processes” on page 55 provides information specifically for a deduplication restore.

A regular NSR-DBM-ORA restore includes the following interactions:

1. The NSR-DBM-ORA MML translates the object names requested by RMAN into a format that the NetWorker server understands, and forwards the names to the NetWorker service, nsrindexd.

2. The nsrindexd service verifies that the backup pieces exist in the client file index.

3. When the NetWorker server receives a restore request from the client, the main NetWorker service, nsrd, contacts the media service, nsrmmd.

4. The nsrmmd service calls the media database service, nsrmmdbd, to determine which media contains the requested save set and issue a mount request.

5. The nsrmmd service reads the appropriate backup volumes and passes the data to the Oracle sessions through the NSR-DBM-ORA MML.

6. The Oracle Server sessions write the data to the disk.

Figure 5 on page 55 shows how the Oracle Server, NetWorker server, and NSR-DBM-ORA processes interact during a regular NSR-DBM-ORA restore.


Introduction

Figure 5 Regular NSR-DBM-ORA restore

Once RMAN has restored the required files from the backup volumes, a database administrator can complete the standard Oracle database recovery.

Deduplication restore processes

A deduplication NSR-DBM-ORA restore includes the following interactions:

1. The NSR-DBM-ORA MML translates the object names requested by RMAN into a format that the NetWorker server understands, and forwards the names to the NetWorker server.

2. The NetWorker server verifies that the backup pieces exist in the NetWorker client file index and media database.

3. When the Oracle Server restore session determines from the media database information that the data exists in a deduplication backup, the Oracle session starts a nsravtar process on the NSR-DBM-ORA client.

4. When the NetWorker server receives a restore request from the client, the main NetWorker service, nsrd, contacts the media service, nsrmmd.

5. The nsrmmd service on the NetWorker server calls the media database service, nsrmmdbd, to determine which media contains the requested save set and issue a mount request.

6. The nsrmmd service reads the appropriate backup volumes and passes the data (hash ID) to the Oracle Server session.

7. The Oracle Server session passes the hash ID to the nsravtar process.

Backup Pieces Client File Index Media Database

nsrmmdbd

nsrmmd

nsrd


nsrindexd Oracle

process (channel)


nsrexecd

NSR-DBM-ORA

MML


data


RMAN

GEN-000173



Introduction

8. The nsravtar process retrieves the data corresponding to the hash ID from either the primary Avamar server or the replication Avamar server, and sends the data to the Oracle Server session.

9. The Oracle Server session writes the data to the disk.

Software Configuration 57

2


◆ Configuration roadmap................................................................................................ 58◆ Verify the Oracle Server configuration....................................................................... 59◆ Verify the NetWorker resources .................................................................................. 59◆ Configuring I18N support ........................................................................................... 64◆ Migrating a legacy configuration with the nsrnmoadmin command ................... 65◆ Configuring a backup with the NMC wizard........................................................... 68◆ Configuring a backup with the NMC legacy method ............................................. 70◆ Configuring a deduplication backup ......................................................................... 86◆ Configuring a probe-based backup ............................................................................ 90

SoftwareConfiguration


Software Configuration

Configuration roadmapThe Oracle Server and NetWorker server must be properly configured before the NSR-DBM-ORA software can be used for backup and restore operations.

Before configuring the Oracle Server and NetWorker server, ensure that the NSR-DBM-ORA software is installed on the Oracle Server host according to the instructions in the NSR-DBM-ORA Release Notice.

To perform specific configuration procedures, you can use either of the following through NetWorker Management Console (NMC):

◆ Client backup configuration wizard for Oracle

◆ Legacy configuration method (without the wizard) that uses application-independent screens in NMC

To configure a regular scheduled backup, follow the instructions in the following sections that apply to your particular environment:

1. Verify the Oracle Server configuration according to “Verify the Oracle Server configuration” on page 59.

2. Verify the NetWorker resources according to “Verify the NetWorker resources” on page 59.

3. If required, configure internationalization (I18N) support according to “Configuring I18N support” on page 64.

4. To use the new backup configuration wizard to modify a legacy configuration that was not created with that wizard, migrate the configuration first according to “Migrating a legacy configuration with the nsrnmoadmin command” on page 65.

5. Complete the backup configuration by using either the wizard or NMC:

• “Configuring a backup with the NMC wizard” on page 68

• “Configuring a backup with the NMC legacy method” on page 70 (This section includes details on creation of nsrnmo and RMAN scripts and manual configuration of save set bundling and policy uniformity.)

Refer to the following for additional information on configuring a deduplication backup or probe-based backup:

◆ “Configuring a deduplication backup” on page 86

◆ “Configuring a probe-based backup” on page 90

Chapter 6, “Cluster and RAC Systems,” provides information on configuring any required cluster or Real Application Cluster (RAC) systems.

Verify the Oracle Server configuration 59


Verify the Oracle Server configurationThe Oracle Server system must be properly installed and configured before the NetWorker server and NSR-DBM-ORA software is configured.

To set up the Oracle Server system:

1. Install and configure the Oracle Server software components, including RMAN. The appropriate Oracle installation guide provides more information.

2. Set up and configure the target database and Oracle Recovery Catalog. The Oracle Recovery Manager documentation provides more information on the Recovery Catalog.

3. Set up and configure the Oracle networking software, Oracle Net.

4. Register the target database with the Recovery Catalog. The Oracle backup and recovery documentation provides more information.

Note: Detailed information on these steps is available from Oracle user documentation and Oracle support services. Oracle support contact information is available at the Oracle website.

Verify the NetWorker resourcesVerify that the required NetWorker resources are configured on the NetWorker server, according to the information in the following sections.

NetWorker Server resource

After the NetWorker server software is installed, the NetWorker configuration includes a preconfigured Server resource with attribute settings that influence the performance and security of backups.

Table 2 on page 59 describes the main NetWorker Server resource attributes. Verify that the attribute settings in the Server resource are valid for the NSR-DBM-ORA backup environment. Modify the settings as required.

The NetWorker server online help and the NetWorker Administrator’s Guide provide more information on how to configure a NetWorker Server resource and its attributes.

Table 2 NetWorker Server resource attributes

Attribute Description

Name Specifies the hostname of the NetWorker server.

Parallelism Specifies the maximum number of backup save streams that the NetWorker software allows to arrive concurrently at the server. The NetWorker server edition determines the maximum parallelism value. When multiple data streams are backed up simultaneously, the efficiency of the storage devices is increased.

Administrator Specifies users with NetWorker Administrator privileges. The initial default setting of the attribute is root@hostname, where hostname is the NetWorker server hostname. “Specify the required NetWorker privileges” on page 60 provides more information.

Datazone pass phrase

Specifies the key or pass phrase to use for AES encryption of Oracle data during an NSR-DBM-ORA backup. The pass phrase is required to restore the Oracle data from the backup. “NSR_AES_ENCRYPTION” on page 191 provides more information.



Specify the required NetWorker privileges

Certain NSR-DBM-ORA operations require specific NetWorker privileges. The NetWorker privileges are specified through the User Group resource. “NetWorker user group privileges” on page 60 provides more information.

One NSR-DBM-ORA operation that requires special privileges is deletion. NSR-DBM-ORA attempts to remove an entry from the NetWorker index in the following cases:

◆ If the RMAN delete command is used.

◆ If a running Oracle backup is canceled according to the instructions in one of the following sections:

• “Cancel a manual backup” on page 101

• “Cancel a scheduled backup” on page 105

Note: If the Oracle user is not granted the required NetWorker privileges in these cases, NSR-DBM-ORA fails to remove the backup save set entries from the NetWorker index. However, RMAN might remove the corresponding entries from the RMAN catalog, which would leave the NetWorker index and RMAN catalog unsynchronized. To resynchronize the index and catalog, issue the appropriate NetWorker media management command to manually remove the inconsistent save set entries from the NetWorker index.

To enable NSR-DBM-ORA to remove an entry from the NetWorker index, ensure that the Oracle user has the required NetWorker privileges. Verify that the required user group privileges exist according to “NetWorker user group privileges” on page 60.

NetWorker user group privileges

The NetWorker server includes an access control feature. This feature allows NetWorker administrators to assign users to NetWorker user groups. Each user group has a specific set of privileges associated with it, as defined in the Privileges attribute of the User Group resource.

The NetWorker server is installed with two preconfigured user groups:

◆ Administrators — Members of this group have privileges to perform all NetWorker operations. The root user on a UNIX system is always a member of this group and cannot be removed from the group.

◆ Users — By default, members of this group have privileges to back up and recover local data and monitor NetWorker operations. They cannot view or edit configurations.

Note: The privileges associated with the Users group can be customized to fit the requirements of the NetWorker users in the group. The privileges associated with the Administrators group cannot be changed.

By default, the NetWorker server assigns the following privileges to all users:

◆ Monitor NetWorker

◆ Recover Local Data

◆ Backup Local Data

The default user group configurations are sufficient for RMAN backup, restore, and crosscheck operations using NSR-DBM-ORA. If the default user group configurations are changed, ensure that the required privileges are assigned for the operations.

Verify the NetWorker resources 61


Verify that the required user group privileges exist for the NSR-DBM-ORA operations, as described in Table 3 on page 61. The NetWorker Administrator’s Guide provides information on how to modify existing privileges.

Note: Proxy backups and restores require the same privileges as regular Oracle backups and restores, plus the privileges required by the PowerSnap Module. The NetWorker PowerSnap Module documentation provides more information on the required privileges.

The Oracle user is defined as the following:

◆ If Net service is used, it is the operating system user that starts the Net service.

◆ If Net service is not used, it is the operating system user that runs RMAN.In the case of a scheduled backup, the operating system user is root on UNIX .

IMPORTANT!If the correct user group privileges are not assigned, an Oracle backup or restore fails with an error message that indicates the required user group privileges. The NetWorker Administrator’s Guide provides more information on NetWorker user groups and setting user group privileges.

NetWorker Schedule resource

You can set the backup schedule to one of the existing schedules (provided by the NetWorker Schedule resources on the server) for the NSR-DBM-ORA client by using either backup configuration wizard or the legacy configuration method.

Table 3 User group privileges required for NSR-DBM-ORA operations

Operation Operating system user that performs operation Required user group privileges

Migration of a legacy backup configuration with the nsrnmoadmin command

Root user on the Oracle Server Configure NetWorker

Regular manual Oracle backup Oracle user on the Oracle Server Recover Local Data, Backup Local Data(These privileges are set by default)

Regular scheduled Oracle backup Oracle user on the Oracle Server Recover Local Data, Backup Local Data(These privileges are set by default)

Root user on the Oracle Server Monitor NetWorker, Backup Local Data(These privileges are set by default)

Regular Oracle restore Oracle user on the Oracle Server Recover Local Data(This privilege is set by default)

RMAN crosscheck Oracle user on the Oracle Server Recover Local Data(This privilege is set by default)

RMAN backup deletion Oracle user on the Oracle Server Operate NetWorker, and all its prerequisite privileges

Restore of NWORA resource file backup to the Oracle Server

Root user on the Oracle Server Recover Local Data(This privilege is set by default)

Save set bundling Root user on the Oracle Server Operate NetWorker, and all its prerequisite privileges



A NetWorker Schedule resource specifies the attributes of the backup schedule. The NetWorker Administrator’s Guide provides more information on the features of a Schedule resource.

The NSR-DBM-ORA software does not honor the NetWorker backup levels (except skip) that are specified in the NetWorker Schedule resource.

IMPORTANT!For scheduled NSR-DBM-ORA backups, the NetWorker backup levels that appear in the Schedule resource do not determine the NSR-DBM-ORA backup levels. The level of the NSR-DBM-ORA backup must be specified as either full or incremental in the RMAN backup script. The backup level is determined by the level in the RMAN script only.

You must configure a Schedule resource through the legacy method with NMC, as described in “Configure a Schedule resource with NMC” on page 81.

NetWorker Device resources

The NetWorker server uses a supported tape or disk storage device to write data during an Oracle backup and to read data during an Oracle restore. The NetWorker server configuration must include a Device resource for each storage device to be used for backups and restores. In addition, each storage device must contain a labeled and mounted volume.

Configure the required NetWorker Device resources with the NMC program. The NetWorker Administrator’s Guide provides more information on storage devices, the NMC program, and how to configure Device resources.

The NetWorker Administrator’s Guide also provides information on how to label and mount backup volumes in the storage devices, and how to configure any required storage nodes (with attached devices), autochangers, and silos.

NetWorker volume pools

NetWorker software directs backups to groups of media or backup volumes called pools. A pool is a specific collection of backup volumes that the NetWorker server uses to store, sort, and organize backup data. For example, backups of Oracle data such as tablespaces and archived redo logs can be directed to volumes in specific devices.

Each NetWorker volume pool is defined by its Pool resource in the NetWorker server. The attribute settings in the Pool resource act as a filter that the server uses to determine the type of data to write to volumes in the pool. Each volume pool has a Pool Type attribute.

Note: With NSR-DBM-ORA, the only valid pool types are backup and backup clone.

Each NetWorker volume belongs to either a preconfigured pool or a user-created pool. Each pool has a specific label template associated with it, providing an automated method to identify the media assigned to a pool. NetWorker software uses pools of volumes and label templates to track the data is on each volume.

Note: If a customized volume pool is not specified for Oracle backup volumes, the NetWorker server routes data for an Oracle backup to the appropriate volume pool.

Verify the NetWorker resources 63


Configure any required NetWorker Pool resources and corresponding Label Template resources with the NMC program. The NetWorker Administrator’s Guide and NMC online help provide more information.

Setting the NSR_DATA_VOLUME_POOL parameter

To send data from a manual or scheduled backup to a specific pool, you can set the NSR_DATA_VOLUME_POOL parameter in the RMAN backup session. Chapter 3, “Backup Procedures,” provides more information on the two types of backups.

Note: In the case of proxy backups, the parameter is used to specify the volume pool for live backups only (backups to secondary storage only). The parameter cannot specify the snapshot pool for instant backups. The only way to specify the snapshot pool is by configuring the NetWorker resources, as described in “Configuring the NetWorker Pool resources” on page 162.

NSR_DATA_VOLUME_POOL is mandatory if backup copies are generated during a manual backup only. Separate NetWorker pools must be defined for each backup copy. “Backup copies” on page 25 provides more information on how to generate backup copies during a manual backup.

Appendix A, “Parameters in an RMAN Session,” provides more information on the NSR_DATA_VOLUME_POOL parameters.

Firewall support

The NSR-DBM-ORA software provides firewall support. The ports that the NSR-DBM-ORA software uses for the firewall depend on the corresponding ports configured for the NetWorker server.

To configure the firewall that the NSR-DBM-ORA software uses, follow the firewall configuration instructions in the NetWorker Administrator’s Guide for the particular NetWorker server platform.



Configuring I18N support“Internationalization (I18N)” on page 31 describes the features of NSR-DBM-ORA internationalization (I18N) support.

To configure I18N support:

1. Ensure that you meet the “Requirements for I18N support” on page 64.

2. Follow the configuration steps in “Configure I18N support” on page 64.

Requirements for I18N support

Ensure that all of the following I18N requirements are met:

◆ The NSR-DBM-ORA client host includes a supported internationalized version of the operating system, properly configured to operate in the non-English locale.

◆ The Oracle software provides the required National Language Support (NLS) or Globalization support, and the Oracle database is configured with the required non-ASCII character set. The Oracle Globalization Support documentation provides details.

◆ A supported NetWorker release 7.5 or later is installed.

◆ For I18N support during proxy operations, a supported release of the PowerSnap Module is installed and configured, as described in the NSR-DBM-ORA Release Notice.

The NetWorker documentation provides details on any other I18N requirements.

Configure I18N support

Note: Configuration of proxy backups or restores with the NSR-DBM-ORA wizard is not supported. Wizard references in the following steps do not apply to the configuration of proxy operations. The PowerSnap Module documentation provides details on the PowerSnap options that support non-ASCII values.

To configure I18N support on the NSR-DBM-ORA client host:

1. Log in as the root user, then shut down the NetWorker services, set the environment variable LC_ALL to the appropriate locale, and restart the NetWorker services.

For example, in a Japanese locale on UNIX, set LC_ALL as follows:

# nsr_shutdown# export LC_ALL=JA_jp.eucJP# /etc/init.d/networker start

2. Set the environment variable NLS_LANG to the character set supported by the operating system and Oracle database, and then restart the Oracle Server.

The Oracle Globalization Support documentation provides details on the NLS_LANG variable.

Migrating a legacy configuration with the nsrnmoadmin command 65


For example, to ensure that Oracle properly returns Japanese text in a Japanese locale, set NLS_LANG as follows:

export NLS_LANG=JAPANESE_JAPAN.JA16EUC% lsnrctl stop% lsnrctl start% sqlplus /nologSQL*Plus: Release 10.1.0.2.0 - Production on Thu Apr 26 15:12:03Copyright (c) 1982, 2004, Oracle. All rights reserved.SQL> connect sys/oracle as sysdba;SQL> shutdown;SQL> startup;SQL> quit;

3. If you do not configure a scheduled backup with the configuration wizard, set the NLS_LANG parameter in the nsrnmo script to the same value as the environment variable NLS_LANG.

For example, in a Japanese locale, set NLS_LANG in the nsrnmo script as follows:

NLS_LANG=JAPANESE_JAPAN.JA16EUC

Note: If you configure the scheduled backup with the configuration wizard, you can set NLS_LANG on a wizard screen. The wizard autopopulates the NLS_LANG field if NLS_LANG is set in the NWORA resource file.

4. To enable proxy catalog synchronization, set the NSR_ORACLE_NLS_LANG parameter to the same value as the environment variable NLS_LANG by using the nsrnmoadmin command.

For example, in a Japanese locale, set the parameter by typing the following command:

nsrnmoadmin -r add NSR_ORACLE_NLS_LANG JAPANESE_JAPAN.JA16EUC

“Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details on the nsrnmoadmin command. The command sets the parameter value in the NWORA resource file, which is described in “The NWORA resource file” on page 176.

Migrating a legacy configuration with the nsrnmoadmin commandYou can use the nsrnmoadmin command to migrate a scheduled backup configuration that was created with a legacy method (without the configuration wizard in NSR-DBM-ORA 5.0) to the configuration storage framework that is supported by the new NSR-DBM-ORA wizard. After the migration, you can use the new wizard to modify the backup configuration.

As an alternative to using the nsrnmoadmin command for the migration, you can use the Specify the RMAN Script Template (Optional) screen of the NSR-DBM-ORA wizard to retrieve an existing configuration from an RMAN script.

To migrate a legacy configuration with the nsrnmoadmin command:

1. Ensure that you meet the “Requirements for using the nsrnmoadmin command for migration” on page 66.

2. Use the proper nsrnmoadmin command and options, according to the “Migration command syntax and options” on page 67.



Requirements for using the nsrnmoadmin command for migration

Before using the nsrnmoadmin command to migrate a backup configuration, ensure that the appropriate requirements are met:

◆ The required NetWorker releases are installed, as described in the NSR-DBM-ORA Release Notice.

◆ The nsrnmo script contains the mandatory ORACLE_HOME setting.

◆ The RMAN script contains a single valid value for each of the following:

• Target database username

• Password of the target database user

• Net service (instance) name

◆ The RMAN script contains correct syntax for the following commands:

• allocate channel

• backup

• connect

• release channel

• send

◆ The RMAN script does not contain any of the following commands:

• @

• allocate channel for maintenance

• configure

• proxy

Note: Migration of a proxy backup configuration is not supported.

◆ You have all of the NetWorker privileges needed to query and update the Client resources on a NetWorker server, as described in Table 3 on page 65. The NetWorker documentation provides information on the required user group privileges.

◆ If you use the nsrnmoadmin command in interactive mode (without the -Y option), you can provide the names of all the users that will use the wizard to modify the configuration (after it has been migrated):

• When you use nsrnmoadmin in interactive mode (without the -Y option), you are prompted by default for the names of all such users in ACL format (user@hostname). If you do not provide the usernames when prompted, you must add the usernames at a later time to the NetWorker Lockbock resource for the client by using NMC to edit the resource. The NetWorker Administrator’s Guide provides details on editing the Lockbox resource in the section on lockbox password management.

• When you use nsrnmoadmin in noninteractive mode (with the -Y option), the migration does not prompt for any input. After the migration, you must add the usernames to the NetWorker Lockbox resource for all the users that will use the wizard to modify the configuration.

Migrating a legacy configuration with the nsrnmoadmin command 67


IMPORTANT!If you migrate the backup configuration of a cluster virtual client, you must do one of the following to enable scheduled backups of the client:- During the migration, when nsrnmoadmin prompts for names of wizard users to add to the Lockbox resource, specify the name root@physical_hostname.- After the migration, use NMC to edit the Lockbox resource for the cluster virtual client, and add the name root@physical_hostname to the resource.

Migration command syntax and options

You must type the nsrnmoadmin command for migration on the operating system command line, as root user on UNIX

The command user must meet the migration requirements listed in Table 3 on page 61.

The nsrnmoadmin command syntax and options are as follows:

nsrnmoadmin -M -s server_name [-c client_name] [-g group_name][-N save_set_name] [-Y]

nsrnmoadmin -P -s server_name [-c client_name] [-g group_name][-N save_set_name]

The -M and -P options are mutually exclusive.

Command options and settings in brackets ([ ]) are optional. Do not include the brackets when typing the command.

Table 4 on page 67 describes the nsrnmoadmin command options.

Table 4 Options of the nsrnmoadmin command for migration (page 1 of 2)

Option Description

-M Specifies the migrate option. The nsrnmoadmin program performs the following:1. Queries the NetWorker server resource database to locate all of the Client resources that match the values specified by the

-c, -g, -N, and -s options.2. Migrates each legacy configuration (the nsrnmo script, RMAN script, NWORA resource file, Client resource) to the

configuration format that is supported by the new NSR-DBM-ORA wizard.

Note: The nsrnmoadmin program can only migrate a configuration that physically resides on the host where the nsrnmoadmin command is typed. To migrate the Client resources for different physical hosts, you must run the nsrnmoadmin program on each physical host, or write a script to automate the process.

-P Specifies the probe option. The nsrnmoadmin program performs a probe to query the NetWorker server resource database and locate all of the Client resources that match the values specified by the -c, -g, -N, and -s options, and prints the results to standard output. The program does not actually migrate any legacy configurations.

Note: The nsrnmoadmin program can only probe configurations that physically reside on the host where the nsrnmoadmin command is typed. To probe the Client resources created for different physical hosts, you must run the nsrnmoadmin program on each physical host.

-c client_name Optional. Specifies the hostname of the NetWorker client to be configured. Typically, this option specifies a virtual client in a cluster. The default value is the hostname of the local physical client.

-g group_name Optional. Specifies the name of the NetWorker group for the query operation. If this option is not specified, then a group name is not included in the criteria for the query of the server resource database.

-N save_set_name Optional. Specifies the value set in the Save Set attribute of the Client resource. If this option is not specified, then a save set name is not included in the criteria for the query of the server resource database.



Configuring a backup with the NMC wizardTo configure a backup with the wizard:

1. Review the information in “About the backup configuration wizard” on page 68.

2. Ensure that you meet the “Requirements for using the backup configuration wizard” on page 68.

3. Follow the configuration steps in “Configure a backup with the wizard” on page 69.

About the backup configuration wizard

NSR-DBM-ORA release 5.0 includes a new NMC-based backup configuration wizard (also known as Client Backup Configuration in NMC) that is integrated with a supported NMC release 7.5 or later.

“Configuration wizards” on page 28 describes the main features of the backup configuration wizard. The wizard can configure the Client, Group, and Policy (browse or retention) resources for a scheduled backup. Other NetWorker resources must be configured manually (without the wizard) through NMC, as described in “Configuring a backup with the NMC legacy method” on page 70.

The wizard option for configuring a "typical" scheduled backup provides a more simplified workflow that generates predefined values (for the number of backup channels, backup name format, and so on). The wizard help provides details on the predefined settings used for a typical scheduled backup.

To use the NSR-DBM-ORA wizard to modify an existing legacy configuration that was not created with the wizard, you must first migrate the configuration according to “Migrating a legacy configuration with the nsrnmoadmin command” on page 65.

The following sources provide more information on the configuration wizard:

◆ NetWorker Administrator’s Guide or Release Notice

◆ Descriptive inline text in the wizard

◆ Online help in the wizard

Requirements for using the backup configuration wizard

Before you use the backup configuration wizard, ensure that all of the following requirements are met:

◆ The NMC user that starts the wizard (the wizard user) has the Configure NetWorker privileges on the NetWorker server where the configuration is created.

-s server_name Mandatory. Specifies the hostname of the NetWorker server that backs up the client being configured.

-Y Optional. Specifies non-interactive mode, which causes the nsrnmoadmin program to proceed with a migration without prompting for confirmation. If this option is not specified, the nsrnmoadmin program displays all of the fields to be updated in the Client resource and requests confirmation to proceed with the migration.

Table 4 Options of the nsrnmoadmin command for migration (page 2 of 2)

Option Description

Configuring a backup with the NMC wizard 69


◆ Communication between the NMC server, NetWorker server, and NSR-DBM-ORA client uses nsrauth authentication. The NetWorker documentation provides any requirements for nsrauth authentication.

◆ The required NetWorker releases are installed on the NMC server, NetWorker server, and NSR-DBM-ORA client hosts, as described in the NSR-DBM-ORA Release Notice.

Configure a backup with the wizard

To create or modify a backup configuration with the wizard:

1. Start the NetWorker Management Console software.

2. Open the Administration window:

a. In the Console window, click Enterprise.

b. In the left pane, select a NetWorker server in the Enterprise list.

c. In the right pane, select the application.

d. From the Enterprise menu, click Launch Application.

The Administration window is launched as a separate application.

3. In the Administration window, click Configuration.

4. In the Configuration window, click Clients.

5. Start the wizard by the appropriate method:

• If you are creating a new backup configuration, use one of the following methods:

– Select Configuration > Client Backup Configuration > New.– In the left pane under the client name, right-click Clients and select Client

Backup Configuration > New.– In the main Clients list, right-click the NSR-DBM-ORA client and select

Client Backup Configuration > New.• If you are modifying an existing backup configuration, right-click the

NSR-DBM-ORA client in the right pane, and select Client Backup Configuration > Modify.

6. On each wizard screen that appears, specify the required options and values for the backup configuration.

If you use the wizard to configure a deduplication backup, ensure that the Filesperset field is set to 1 on the Specify the RMAN Options (Part 1 of 2) screen.

Each wizard screen includes an online help button that you can click to access descriptions of all the fields and options on the screen:

• On all but the last screen, click Next to proceed.

• On the last screen, Review and Accept the Client Configuration, click Create or Modify to create or modify the configuration, respectively, and click Finish to exit the wizard.

If you choose to save configuration settings to an RMAN script on disk, you can edit the script later and also use the script for a manual NSR-DBM-ORA backup. The resources required for a manual backup must be configured with the legacy method (without the wizard).



IMPORTANT!When you use the wizard to configure a cluster virtual client, the wizard attempts to perform all of the additional settings required for the cluster environment, including the following:- Creating the required Client resources.- Adding the required ACL entries to the Lockbox resource, for the physical hosts provided in the Remote Access field on the NetWorker Client Properties screen of the wizard.- Setting the NSR_CLIENT parameter.- Adding the -c virtual_clientname option to the Backup Command attribute in the Client resource of the virtual client.In the Client resource of the virtual client, ensure that the Remote Access attribute is set with user@physical_hostname for each of the physical hosts of the cluster; otherwise, the backup might fail.

Configuring a backup with the NMC legacy methodThe following sections describe how to configure backups with the legacy method (without the configuration wizard).

About backup configuration with the legacy method

NSR-DBM-ORA release 5.0 still supports the legacy method of backup configuration, as used in previous NSR-DBM-ORA releases. To configure a scheduled backup, you can still manually create an RMAN script and nsrnmo script, set parameters in the NWORA resource file if needed, and configure the NetWorker resources with NMC.

As an alternative, you can use the backup configuration wizard to configure a scheduled backup, as described in “Configuring a backup with the NMC wizard” on page 68. However, NetWorker resources other than the Client and Group resource must still be configured manually with NMC.

For a manual backup, the NetWorker Client resource must be modified manually through the legacy method, as required.

Create RMAN scripts for backups

Create an appropriate RMAN script to perform the preferred type of manual or scheduled backup on the Oracle Server host.

NSR* parameters in the script must be set with the methods described in“Setting the NSR* parameters” on page 190. The use of the send command is recommended where possible. “The send command” on page 201 provides more information.

NSR* parameters cannot be set through other methods, for example, by using the setenv or set command on the operating system command line. Appendix A, “Parameters in an RMAN Session,” provides complete information on the NSR* parameters in the RMAN script or session.

RMAN backup scripts can be stored as flat ASCII files. Alternatively, if a Recovery Catalog is used, backup scripts can be stored in the Recovery Catalog database. The appropriate Oracle backup and recovery documentation provides information on storing the backup scripts in the Recovery Catalog database.

Configuring a backup with the NMC legacy method 71


The following sections provide details on RMAN scripts for manual backups and scheduled backups.

RMAN scripts for manual backups

If automatic channel allocation and persistent settings are used, the backup command can be run as a stand-alone command. “Automatic channel allocation” on page 24 provides more information.

Example 10 RMAN script for a manual backup

The following RMAN script is for a manual backup of an entire Oracle database to the volume pool MondayFulls of the (remote) NetWorker server mars.fujitsu.com :

run {allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com , NSR_DATA_VOLUME_POOL=MondayFulls)’;backup full filesperset 4 format ’FULL_%d_%U’ (database);release channel t1;release channel t2;

}

To specify a Media Management (in this case, NSR-DBM-ORA) device, set the type option in the allocate channel command to SBT_TAPE.

If a device is allocated by using the allocate channel t1 type disk command (with Oracle correctly configured and NSR-DBM-ORA uninstalled), backups can be directed to disk files through Oracle’s backup implementation.

In the preceding RMAN backup script, the format string FULL_%d_%U specifies the name of each backup piece. This name can be anything, provided that each backup piece has a unique name on the NetWorker server. Substitution variables, such as %d and %U, can be used to guarantee unique names:

◆ %d specifies the name of the database.

◆ %U specifies a unique Oracle system-generated filename.

A format string such as FULL or FULL_%d will not generate unique names. Similarly, the format string FULL_%U will not generate unique names for two databases that are being backed up to the same NetWorker server.

IMPORTANT!If a backup piece name is not unique, the Oracle backup fails.

During a manual backup, the prefix RMAN: automatically precedes the backup piece name in the NetWorker media database. For example, if the backup piece name specified in the RMAN script is accounts_data_file, the manual backup records the save set name as RMAN:accounts_data_file in the media database. The mminfo command displays the save set name in this form.

The following sources provide more information:

◆ The appropriate Oracle backup and recovery documentation provides information on how to write RMAN scripts.

◆ The Oracle Enterprise Manager documentation provides information on how to use the Oracle Enterprise Manager Backup Wizard to generate RMAN scripts.



◆ Appendix B, “RMAN Commands,” provides important information on RMAN commands.

◆ “Regular backup information in NetWorker indexes” on page 108 describes the information stored for a manual backup in the NetWorker indexes.

Example 11 RMAN script for AES encryption during an Oracle backup

The following RMAN script performs a nonproxy backup of the Oracle database to the NetWorker server mars.fujitsu.com by using 256-bit AES encryption, as specified by the NSR_AES_ENCRYPTION parameter setting:

run {allocate channel t1 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com , NSR_AES_ENCRYPTION=TRUE)’;backup full filesperset 4 format ’FULL_%d_%U’ (database);release channel t1;

}

The AES encryption uses the key or pass phrase that is set in the Datazone pass phrase attribute of the NetWorker Server resource. “NSR_AES_ENCRYPTION” on page 191 provides more information.

IMPORTANT!Keep a record of all pass phrases used for 256-bit AES encryption. Be careful when you change the pass phrase on the NetWorker server. If the pass phrase on the server is changed and you cannot remember the pass phrase used for an NSR-DBM-ORA backup, the encrypted data cannot be recovered. The NetWorker Administrator’s Guide provides more information on pass phrases.

Example 17 on page 116 shows a sample Oracle restore of an AES encrypted backup, with the NSR_ENCRYPTION_PHRASES parameter that specifies the original pass phrase that was used for the backup.

RMAN scripts for scheduled backups

IMPORTANT!For scheduled backups (both regular and proxy backups), do not include send as part of the allocate channel command. The send command must be separate.

For example, NSR-DBM-ORA does not support the following for scheduled backups:

allocate channel t1 type ’SBT_TAPE’ send ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com )’;

The following is the correct form of the commands:

allocate channel t1 type ’SBT_TAPE’;send channel t1 ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com )’;

With automatic channel allocation, specifying the send command before the backup or restore command causes the following error:

RMAN-06422: no channels found for SEND command

“Automatic channel allocation” on page 24 provides information on automatic channel allocation.



Example 12 RMAN script for a scheduled backup

The following RMAN script is for a scheduled backup of an entire Oracle database to the volume pool MondayFulls. The Recovery Catalog is used in this case:

connect target target_user/target_passwd@target_Netservicename;connect rcvcat rcvcat_user/rcvcat_passwd@rcvcat_Netservicename;run {

set command id to ’xxx’;allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_DATA_VOLUME_POOL=MondayFulls)’;backup full filesperset 4format ’FULL_%d_%U’(database);release channel t1;release channel t2;

}

If automatic channel allocation and persistent settings are used, a scheduled RMAN backup script must still be created and contain the following commands:

◆ connect target

◆ connect rcvcat (if using a Recovery Catalog)

◆ backup

The command connect target target_user/target_passwd@target_Netservicename is mandatory in each RMAN script for a scheduled backup. This command establishes the proper connection to the target database.

Specify the correct values in the connect target command:

◆ target_user is the user with SYSDBA privileges for the target database.

◆ target_passwd is the password of the target_user (for connecting as SYSDBA), specified in the target database’s orapwd file.

◆ target_Netservicename is the Net service name of the target database. This name is mandatory in the connect target command.

A password file must be used for the target database. To use the password file, the orapwd utility might need to be used and the REMOTE_LOGIN_PASSWORDFILE parameter set to exclusive in the initoracle_sid.ora file. The appropriate Oracle documentation provides more information.

Notes:

◆ Since each scheduled backup RMAN script requires a connect target command, each Oracle instance requires a separate scheduled backup RMAN script.

◆ In the connect target command, do not use the value internal for target_user or the value oracle for target_passwd.

The command connect rcvcat rcvcat_user/rcvcat_passwd@rcvcat_Netservicename is mandatory if the Recovery Catalog is used for the scheduled Oracle backup. This command establishes the proper connection to the Recovery Catalog database.

Specify the correct values in the connect rcvcat command:

◆ rcvcat_user is the owner of the Recovery Catalog database.

◆ rcvcat_passwd is the password of the rcvcat_user.

◆ rcvcat_Netservicename is the Net service name of the Recovery Catalog database.

To enable the scheduled backup to be canceled, the scheduled Oracle backup script must include set command id to ‘xxx’ (where xxx can be any string of characters



enclosed in single quotes). “Cancel a scheduled backup” on page 105 provides more information on how to cancel a scheduled backup.

The remainder of the scheduled backup script in Example 12 on page 73, starting with the first allocate channel command, is similar to the manual backup script in Example 10 on page 71 except that the NSR_SERVER parameter setting is not included.

IMPORTANT!Do not set the parameters NSR_SERVER or NSR_GROUP in a scheduled RMAN backup script. NSR-DBM-ORA sets these two parameters to the values specified in the Client resource for the scheduled Oracle backup, and these values cannot be overridden.

To have the scheduled backup automatically use a volume pool, the backup group can be specified in the Pool resource. The scheduled backup uses that pool unless the parameter NSR_DATA_VOLUME_POOL is set in the RMAN script. Then that parameter’s setting takes precedence over any pool associated with the scheduled backup group.

If NSR_DATA_VOLUME_POOL is set in the RMAN script to a pool different from the one associated with the backup group, the scheduled backup uses the NSR_DATA_VOLUME_POOL pool. It is the user’s responsibility to set that parameter correctly in the RMAN script for a scheduled backup.

Each scheduled backup RMAN script must be stored as a text file. The database administrator should give minimal permissions to the scheduled backup RMAN script file. This way, unauthorized users cannot see the sensitive user IDs and passwords of the target and Recovery Catalog databases.

If a single Oracle instance has multiple RMAN scripts associated with it (for example, to perform tablespace-level or file-level, full or incremental backups, and so on), the database administrator might choose to place the two common connect commands in a single file and invoke those two connect commands in all RMAN scripts by using the @ command.

Test RMAN scripts for scheduled backupsWhen you create an RMAN script, test the script before using it for scheduled backups.

To test the RMAN script, type one of the following commands:

rman cmdfile ’script_name’ send ’”NSR_ENV=(NSR_SERVER=NetWorker_server_name, NSR_GROUP=group_name)”’

rman nocatalog cmdfile ’script_name’ send ’”NSR_ENV=(NSR_SERVER=NetWorker_server_name, NSR_GROUP=group_name)”’

where:

◆ script_name is the RMAN script file pathname.

◆ NetWorker_server_name is the name of the server that starts the backup.

◆ group_name is the name of the scheduled backup group as specified in the Client resource.



Customize the nsrnmo script

Customize the nsrnmo script on the Oracle Server host by modifying the parameters in the script.The NSR-DBM-ORA installation on the Oracle Server host stores a sample nsrnmo script in the /opt/nsr directory. The sample script filename is nsrnmo.sh.

Multiple versions of the nsrnmo script can be created (for example, one for each Oracle instance) on the same Oracle Server host. Each version of the script must have a unique name, starting with the letters nsr or

To create a new version of the nsrnmo script:

1. Copy the sample script file (nsrnmo on UNIX ) to a file with a different name.

2. Customize the parameters in the new file.

For example, you might create the new script file named nsrnmo_prodDB on UNIX.

All versions of the nsrnmo script must be located in the same directory as the NetWorker executables. Do not move the nsrnmo scripts to a different directory.

Notes:

◆ A nsrnmo script from a previous NSR-DBM-ORA release works with NSR-DBM-ORA release 5.0. However, the nsrnmo script installed with NSR-DBM-ORA release 5.0 might include new parameter settings not found in the script from the previous release.

◆ To facilitate the correct reporting of scheduled backup save sets for a UNIX NSR-DBM-ORA client within the group details window of the NetWorker Management Console (NMC), specify the following shell within the nsrnmo script on UNIX:

• /bin/bash, if it exists on the UNIX client

• /bin/ksh, if /bin/bash does not exist on the UNIX client

The nsrnmo script on Solaris zones

On a Solaris system with Solaris zones, a sparse root zone can have shared directories that are mounted from the global zone. Such directories are read-only on the sparse root zone. If NSR-DBM-ORA is installed on a sparse root zone and /opt is one of the shared directories, the nsrnmo script in /opt/nsr is read-only on the sparse root zone. In this case, the nsrnmo script must be edited on the global zone.

If multiple sparse root zones are configured on a physical computer, all of the virtual sparse root zones see the nsrnmo script in /opt/nsr. In this case, each sparse root zone should use a unique script filename, such as nsrnmo1, nsrnmo2, and so on. The Backup Command attribute in the Client resource for each sparse root zone must contain the unique nsrnmo filename for that zone. “Configure a Client resource with NMC” on page 81 provides more information on the Client resource.

Parameters in the nsrnmo script

The sample nsrnmo script installed with NSR-DBM-ORA contains parameters that must be customized for a particular scheduled Oracle backup. The parameters in the sample nsrnmo script are all undefined by default.The nsrnmo script contains only the parameters required on the particular system where the script is installed. The LD_LIBRARY_PATH parameter appears in the nsrnmo script on Linux and Solaris. Follow the instructions in the Oracle installation guide to determine to what value to set the parameter LD_LIBRARY_PATH.



The ORACLE_HOME parameter is mandatory for each scheduled backup, and must be set in the nsrnmo script. The other parameters in the nsrnmo script are optional, and can be left undefined in the script, if preferred.

Note: For other parameters to be in effect during a scheduled backup, lines must be added in the nsrnmo script to define and export the parameters. The use of the UNIX setenv command to set the parameters has no effect.

The following sections describe the parameters that can be set in the nsrnmo script for scheduled backups. The comments in the nsrnmo script file provide details on the parameters:

◆ “ORACLE_HOME” on page 76

◆ “LD_LIBRARY_PATH” on page 76

◆ “NSR_RMAN_ARGUMENTS” on page 77

◆ “NSR_SB_DEBUG_FILE” on page 77

◆ “ORACLE_SID” on page 77

◆ “ORACLE_USER” on page 78

◆ “PRECMD” on page 78

◆ “POSTCMD” on page 79

◆ “TNS_ADMIN” on page 80

ORACLE_HOMEThis parameter is mandatory for each scheduled backup. Set ORACLE_HOME in the nsrnmo script file to the home directory of the Oracle Server installation. The RMAN executable, rman, must be located in the subdirectory bin of $ORACLE_HOME.

For example, if the Oracle Server installation is located in the directory /disk3/oracle/app/oracle/product/10.0, ORACLE_HOME must be set in the nsrnmo script as follows:

ORACLE_HOME=/disk3/oracle/app/oracle/product/10.0

LD_LIBRARY_PATHThis parameter is optional for a scheduled backup. Set LD_LIBRARY_PATH in the nsrnmo script file to the pathname of the directory that contains the Oracle libraries, typically $ORACLE_HOME/lib or $ORACLE_HOME/lib64.

For example, if the Oracle libraries are located in the directory /disk3/oracle/app/oracle/product/10.0/lib, set LD_LIBRARY_PATH in the nsrnmo script as follows:

LD_LIBRARY_PATH=/disk3/oracle/app/oracle/product/10.0/lib

Uncomment the line “export LD_LIBRARY_PATH” (by removing the # symbol at the start of the line) under the “export_environment variables()” function heading in the nsrnmo script:

export_environment_variables(){

export ORACLE_HOMEexport NSR_RMAN_ARGUMENTS

::

export LD_LIBRARY_PATH}



NSR_RMAN_ARGUMENTSThis parameter is optional for a scheduled backup. Set NSR_RMAN_ARGUMENTS to a double-quoted string that contains any valid combination of options for the RMAN executable, rman. The appropriate Oracle Recovery Manager documentation provides details on the valid options.

For example, to append RMAN output to the message log file /nsr/applogs/msglog.log, set NSR_RMAN_ARGUMENTS in the nsrnmo script as follows:

NSR_RMAN_ARGUMENTS="msglog ’/nsr/applogs/msglog.log’ append"

To append RMAN output to the message log file /nsr/applogs/msglog.log if a Recovery Catalog is not used, set NSR_RMAN_ARGUMENTS in the nsrnmo script as follows:

NSR_RMAN_ARGUMENTS="nocatalog msglog ’/nsr/applogs/msglog.log’ append"

NSR_SB_DEBUG_FILEThis parameter is optional for a scheduled backup. It is used only for debugging purposes.

Enable debugging by setting NSR_SB_DEBUG_FILE to the complete pathname of a file to receive detailed debug information for the scheduled backup. This file will exclude MML-specific debug information, which is directed to NSR_DEBUG_FILE (if set in the RMAN script).

Note: If NSR_SB_DEBUG_FILE is undefined (by default), no debug information is generated. If a file with the specified pathname cannot be created, debug information is either directed to the default location or not generated.

For example, to send detailed debug information for the scheduled backup to the /usr/logs/schedbkup.log file, set the parameter NSR_SB_DEBUG_FILE in the nsrnmo script as follows:

NSR_SB_DEBUG_FILE=/usr/logs/schedbkup.log

“NSR-DBM-ORA error messages” on page 212 provides more information on how the debug information is written if the parameter NSR_SB_DEBUG_FILE is set to an invalid pathname.

ORACLE_SIDThis parameter is required for a scheduled backup in the following cases:

◆ The connect target and connect rcvcat commands for the scheduled backup are stored in a separate file, and the connect commands are invoked in the RMAN script by using the @ command.

◆ Save set bundling is enabled for the scheduled backup.

◆ A proxy backup is performed with catalog synchronization enabled. Chapter 7, “Proxy Backups and Restores,” provides more information on proxy backups and catalog synchronization.

Set ORACLE_SID in the nsrnmo script file to the system identifier (SID) value of the Oracle database to be backed up. For example, if catalog synchronization is enabled for proxy backups and the SID of the Oracle database to be backed up is orcl10, ORACLE_SID must be set as follows:

ORACLE_SID=orcl10



ORACLE_USERThis parameter is optional for a scheduled backup configured through the legacy method (without the wizard).

Enable the scheduled backup for operating system authentication by setting ORACLE_USER to the username of the Oracle operating system user, which is set up to connect to the Oracle database through operating system authentication.

Note: Using ORACLE_USER to perform an NSR-DBM-ORA backup through operating system authentication is not supported for the following:- A scheduled backup configured through the new configuration wizard.- A probe-based backup.- A scheduled proxy backup.

For example, to perform a scheduled NSR-DBM-ORA backup on UNIX through operating system authentication (the backup must be configured through the legacy method), set ORACLE_USER to the required Oracle operating system username:

ORACLE_USER=Oracle_OS_username

PRECMDThis parameter is optional. Set PRECMD to the complete pathname of a file containing a preprocessing script to be executed before the RMAN backup script.

If the pre-command script fails (returns a nonzero value), the scheduled Oracle backup does not proceed (that is, the RMAN script is not executed).

IMPORTANT!The script file must have permissions that allow execution by the root user, as the scheduled Oracle backup is always launched by root. The script should return a zero value when it succeeds and a nonzero value when it fails. The return of a nonzero value will cause the scheduled backup to fail.

A sample preprocessing script for UNIX is shown as follows. This script is stored in the file whose complete pathname is specified in the parameter PRECMD in the nsrnmo script.

For example, if the script is stored in the /usr/scripts/orashutdown file, set the parameter PRECMD as follows:

PRECMD=/usr/scripts/orashutdown

This sample script shuts down the Oracle database, presumably for an offline database backup. The su command is required in the script since the nsrnmo script runs as root user:

#!/bin/ksh# Define Site Specifics HereTARGET_DATABASE=proddbERRFILE=/tmp/precmd.logORACLE_USER=proddb## Scheduled backup launches as "root" run PRECMD as oracle usersu - $ORACLE_USER -c "{# Shutdown target database and start in maintenace modeexport ORACLE_SID=$TARGET_DATABASEsvrmgrl <<EOF1 > $ERRFILEconnect internalshutdownstartup mount



exitEOF1}"# Do checking as root since "su" will always return its statusexport status# Define status otherwise it is local to if [] statementstatus=0# Scan file for errors and only return true or falsegrep -e error -e warning -e fatal $ERRFILE > /dev/null 2>&1if [ $? -eq 0 ]; thenstatus=1elsestatus=0fiexit $status

POSTCMDThis parameter is optional. Set POSTCMD to the complete pathname of a file that contains a postprocessing script to be executed after the RMAN backup script.

If the RMAN backup script fails, the failure is reported, and the postprocessing script is executed nevertheless. If the postprocessing script fails, an error message is reported.

IMPORTANT!The script file must have permissions allowing execution by the root user, as the scheduled Oracle backup is always launched by root. The script should return a zero value when it succeeds and a nonzero value when it fails.

A sample postprocessing script for UNIX is shown as follows. This script is stored in the file whose complete pathname is specified in the parameter POSTCMD in the nsrnmo script.

For example, if the script is stored in the /usr/scripts/orastartup file, set the parameter POSTCMD as follows:

POSTCMD=/usr/scripts/orastartup

This sample script starts up the Oracle database, presumably after an offline backup. The su command is required in the script since the nsrnmo script runs as root user:

#!/bin/ksh# Define Site Specifics HereTARGET_DATABASE=proddbERRFILE=/tmp/postcmd.logORACLE_USER=proddb## Scheduled backup launches as "root" run POSTCMD as user proddbsu - $ORACLE_USER -c "{# Mount target database as ready for useexport ORACLE_SID=$TARGET_DATABASEsvrmgrl <<EOF > $ERRFILEconnect internalalter database open;exitEOF}"# Do checking as root since "su" will always return its statusexport status# Define status otherwise it is local to if [] statementstatus=0# Scan file for errors and only return true or false



grep -e ORA- -e error -e warning -e fatal $ERRFILE > /dev/null 2>&1if [ $? -eq 0 ]; thenstatus=1elsestatus=0fiexit $status

TNS_ADMINThis parameter is mandatory if the Oracle Net configuration files are located in a directory other than the default $ORACLE_HOME/network/admin directory.

In this case, set the value of TNS_ADMIN to the pathname of the directory that contains the Oracle Net configuration files. To set TNS_ADMIN in the nsrnmo script, follow the example shown in “LD_LIBRARY_PATH” on page 76. Replace LD_LIBRARY_PATH with TNS_ADMIN. Uncomment the line “export TNS_ADMIN” in the script.

Configure a Group resource with NMC

Note: As an alternative, you can use the backup configuration wizard to configure the Group resource, as described in “Configuring a backup with the NMC wizard” on page 68. “Configuring a probe-based backup” on page 90 provides details on configuration requirements for a probe-based backup.

For a scheduled NSR-DBM-ORA backup, a NetWorker Group resource must be configured that specifies the attributes of the backup group. The Group resource specifies a set of NetWorker Client resources that all start to back up data at a specified time, once the following occurs:

◆ The Autostart attribute is enabled.

◆ The backup start time is specified in the Group resource.

By configuring one or more NetWorker backup groups for scheduled backups, the backups can be:

◆ Distributed to alleviate network traffic.

◆ Scheduled for a time of day when performance demands on the database and NetWorker server are lower.

One or more Client resources configured for the Oracle Server host can be assigned to a NetWorker backup group.

All NetWorker backup groups can be created and modified. All backup groups except the Default group can be deleted.

To use the Default group for testing scheduled backups, change its Autostart attribute to Enabled.

Note: To have a regular scheduled backup automatically use a volume pool associated with the backup group, specify the group name in the Pool resource for the volume pool.

IMPORTANT!For a regular scheduled Oracle backup, the Snapshot attribute in the Group resource must be set to False.



You can create a NetWorker Group resource with the NMC interface. The NetWorker Administrator’s Guide and NMC online help provide more information.

Configure a Schedule resource with NMC

Review the summary information in “NetWorker Schedule resource” on page 61.

Configure the NetWorker Schedule resource with NMC, to specify the days of the week when a scheduled Oracle backup runs.

The NetWorker server provides several preconfigured schedules. Both the preconfigured schedules and customized schedules can be created and modified.

The NetWorker Administrator’s Guide and NMC online help provide more information on how to use the NMC interface.

Customize a NetWorker Schedule resource for an Oracle backup. In the Schedule resource, select a level for each day of the week:

◆ Level full, incremental, or level 1 to 9 specifies that the NetWorker server runs the backup script on that day.

◆ Level skip specifies that the NetWorker server does not run the backup script on that day.

Specify the name of the NetWorker schedule in the Schedule attribute of the Client resource, as described in “Configure a Client resource with NMC” on page 81.

Configure a Client resource with NMC

Note: As an alternative, you can use the backup configuration wizard to configure the Client resource, as described in“Configuring a backup with the NMC wizard” on page 68. “Configuring a probe-based backup” on page 90 provides details on configuration requirements for a probe-based backup.

A NetWorker Client resource is a set of attributes assigned to the NSR-DBM-ORA client host and stored on the NetWorker server. Before NSR-DBM-ORA software can be used for backups or restores, a Client resource must be configured for the NSR-DBM-ORA client host.

If the NetWorker server software is installed on the NSR-DBM-ORA client host, a basic Client resource for the NSR-DBM-ORA client is created automatically during the NetWorker installation. The Client resource must be customized for an NSR-DBM-ORA backup.

“NetWorker indexes and policies used for restores” on page 112 provides more information on how the NetWorker server uses the browse and retention policies to manage Oracle backup data and enable the data to be restored.

“Configuring the NetWorker Client resource” on page 163 provides information on how to configure the Client resource for proxy backups.

To configure the Client resource for a regular scheduled backup by using the NMC interface, specify the required values for each attribute, according to Table 5 on page 82.

Leave the following attributes blank:

◆ Directive

◆ Archive Users



◆ Remote User

◆ Password

Note: On a Solaris system with Solaris zones, ensure that the security fields (such as Remote Access and Privileges) of NetWorker resources used during NSR-DBM-ORA backups and restores refer to the hostname of the zone in which NSR-DBM-ORA operates.

Each Oracle installation requires a separate Client resource. The Backup Command attribute of the Client resource must contain only one nsrnmo script name, and ORACLE_HOME is a mandatory parameter in each nsrnmo script.

If multiple RMAN scripts are specified for the Save Set attribute of the Client resource:

◆ The RMAN scripts are executed in arbitrary order, possibly in parallel.

◆ If the nsrnmo script also contains a well-defined setting for PRECMD or POSTCMD, the precommand and postcommand files will be:

• Common for all the RMAN scripts

• Executed once for each RMAN script

The sequence of execution of “precommand–RMAN script–postcommand” triplets is arbitrary. To include preprocessing and postprocessing with the backups of the various Oracle database instances of an Oracle installation, define a separate NetWorker Client resource for each Oracle instance.

Note: If a scheduled Oracle backup is retried, the specified precommand and postcommand will be executed again for that backup.

Figure 6 on page 84 shows a sample Client resource for a UNIX NSR-DBM-ORA client.

.

Table 5 NetWorker Client resource attributes (page 1 of 2)


Name Specifies the hostname of the Oracle Server host.

Backup Command For a scheduled backup configured through the legacy method with NMC (not through the wizard), specifies the name of a single customized nsrnmo script to use for the backup.For example, the nsrnmo script name might be nsrnmo or nsrnmo2 on UNIX.On a Solaris system with Solaris zones where NSR-DBM-ORA is installed on sparse root zones, the Backup Command attribute in the Client resource for each sparse root zone must contain the unique nsrnmo filename for that zone.

Note: If you configure a scheduled backup through the configuration wizard, the wizard automatically adds the nsrnmostart program name to this attribute. In that case, do not modify this attribute.

Browse Policy Specifies the length of time that the NetWorker server retains an entry for an Oracle backup in the online client file index. This attribute applies only to scheduled backups.

Note: For a scheduled backup, if the parameter NSR_SAVESET_BROWSE is set as described in Appendix A, “Parameters in an RMAN Session,” its value overrides the Browse Policy attribute setting in the Client resource.

To set a specific browse policy for a manual backup, use the parameter NSR_SAVESET_BROWSE. If the parameter is not set, the NetWorker server uses the most appropriate value for the browse policy.

Group Specifies the NetWorker backup group to use for a scheduled backup. “Configure a Group resource with NMC” on page 80 provides more information on NetWorker backup groups.



Remote Access Specifies the fully qualified hostname of a remote system, to enable restores of the backups to that remote system.On a Solaris system with Solaris zones, the Remote Access attribute must contain the hostname of the zone in which NSR-DBM-ORA operates.

Note: If backups from a cluster, or recovery to a host other than the one being backed up, are not required, do not modify the Remote Access attribute.

Retention Policy Specifies the minimum length of time that the NetWorker server maintains information about Oracle backup data in the online media database. This attribute applies only to scheduled backups.

Note: For a scheduled backup, if the parameter NSR_SAVESET_RETENTION is set as described in Appendix A, “Parameters in an RMAN Session,” its value overrides the Retention Policy attribute setting in the Client resource.

To set a specific retention policy for a manual backup, use the parameter NSR_SAVESET_RETENTION. If the parameter is not set, the NetWorker server uses the most appropriate value for the retention policy.

Save Set Specifies the complete pathname of each RMAN script to be used for a scheduled backup, preceded by RMAN:. Do not include any spaces between the prefix RMAN: and the script name.For example, if two separate RMAN backup scripts are created in the files /disk/rman_scripts/archlogbkup and /disk/rman_scripts/fullbkup (to be run by the nsrnmo script in the Backup Command attribute), specify the complete file pathnames prepended by RMAN: in the Save Set attribute:RMAN:/disk/rman_scripts/archlogbkupRMAN:/disk/rman_scripts/fullbkup

Note: If you configure a scheduled backup through the configuration wizard, the wizard automatically includes the prefix RMAN: in the Save Set attribute setting, as described in “Backup configuration storage with the wizard” on page 29.

The save set information for the scheduled backup is stored in the NetWorker indexes as described in:• “Regular backup information in NetWorker indexes” on page 108• “Deduplication backup information in NetWorker indexes” on page 109

Schedule Specifies the NetWorker backup schedule to use for a scheduled backup.“Configure a Schedule resource with NMC” on page 81 provides more information on NetWorker backup schedules.

Table 5 NetWorker Client resource attributes (page 2 of 2)




Figure 6 UNIX Client resource for scheduled backups

Configure an email notification

To send an email notification of the scheduled backup results to the owner of a save set, edit the Owner Notification attribute in the Client resource for the Oracle Server (or Oracle instance).

In the Owner Notification attribute, specify a notification command directed to the login ID of the owner, for example:

/usr/bin/mail -s “server1-server2-server3 backup” jdoe

The NetWorker Administrator’s Guide and NMC online help provide more information on how to edit the attributes of the NetWorker Client resource.

Configure save set bundling

This section describes how to manually configure save set bundling for NSR-DBM-ORA scheduled backups.

As an alternative, you can use the backup configuration wizard to configure the scheduled backup and set the NSR_BUNDLING parameter to enable save set bundling. “Configuring a backup with the NMC wizard” on page 68 provides more information.

Save set bundling can be enabled and disabled independently of policy uniformity. If save set bundling is enabled, policy uniformity should also be enabled, as described in “Configure policy uniformity” on page 85.



To enable save set bundling:

◆ Set the NSR_BUNDLING parameter value to enabled by typing the following command:

nsrnmoadmin -r add NSR_BUNDLING enabled

By default, the NSR_BUNDLING parameter is disabled.


◆ Ensure that NSR-DBM-ORA scheduled backups are configured properly according to the “Configuration roadmap” on page 58.

◆ Ensure that the user group privileges for the root or administrative user on the NSR-DBM-ORA client include the Operate NetWorker privileges. The corresponding User Group resource is configured on the NetWorker server, as described in “NetWorker user group privileges” on page 60.

◆ If the proper username and password are not located in the RMAN script (for example, the connection strings are included as a command file in the RMAN script, such as @connection_file), ensure the following:

• The ORACLE_SID parameter is set in the nsrnmo script, as described in “Customize the nsrnmo script” on page 75.

• An NWORA SID resource with the NSR_ORACLE_CONNECT_FILE parameter setting is created in the NWORA resource file (nwora.res) for the ORACLE_SID, as described in “NWORA SID resources” on page 178.

NSR-DBM-ORA cannot retrieve the connection strings from the RMAN script when the connection strings are included as a command file in the script. In this case, NSR-DBM-ORA must retrieve the connection strings from the connection file specified by the parameter in the NWORA resource file.

◆ Ensure that the NetWorker server is release 7.4 or later, to support staging of the NSR-DBM-ORA save set bundles.

◆ In a RAC system, ensure that all channels are allocated on the same NSR-DBM-ORA client node where the backup is initiated. Save set bundling does not support load balancing across different RAC nodes.

To disable save set bundling, set the NSR_BUNDLING parameter value to disabled by typing the following command:

nsrnmoadmin -r update NSR_BUNDLING disabled

Configure policy uniformity

This section describes how to manually configure policy uniformity for NSR-DBM-ORA backups.

As an alternative, you can use the backup configuration wizard to set the NSR_INCR_EXPIRATION parameter to enable policy uniformity.“Configuring a backup with the NMC wizard” on page 68 provides more information.

Policy uniformity can be enabled and disabled independently of save set bundling. If save set bundling is enabled, as described in “Configure save set bundling” on page 84, policy uniformity should also be enabled.



To enable policy uniformity:

◆ Set the NSR_INCR_EXPIRATION parameter value to enabled by typing the following command:

nsrnmoadmin -r add NSR_INCR_EXPIRATION enabled

By default, the NSR_INCR_EXPIRATION parameter is disabled.


◆ Ensure that NSR-DBM-ORA scheduled backups are configured properly according to the “Configuration roadmap” on page 58.

◆ Ensure that the user group privileges for the root or administrative user on the NSR-DBM-ORA client include the Operate NetWorker privileges. The corresponding User Group resource is configured on the NetWorker server, as described in “NetWorker user group privileges” on page 60.

◆ If the proper username and password are not located in the RMAN script (for example, the connection strings are included as a command file in the RMAN script, such as @connection_file), ensure the following:

• The ORACLE_SID parameter is set in the nsrnmo script, as described in “Customize the nsrnmo script” on page 75.

• An NWORA SID resource with the NSR_ORACLE_CONNECT_FILE parameter setting is created in the NWORA resource file (nwora.res) for the ORACLE_SID, as described in “NWORA SID resources” on page 178.

NSR-DBM-ORA cannot retrieve the connection strings from the RMAN script when the connection strings are included as a command file in the script. In this case, NSR-DBM-ORA must retrieve the connection strings from the connection file specified by the parameter in the NWORA resource file.

◆ In a RAC system, ensure that all channels are allocated on the same NSR-DBM-ORA client node where the backup is initiated. Policy uniformity does not support load balancing across different RAC nodes.

To disable policy uniformity, set the NSR_INCR_EXPIRATION parameter value to disabled by typing the following command:

nsrnmoadmin -r update NSR_INCR_EXPIRATION disabled

Configuring a deduplication backupTo configure a deduplication NSR-DBM-ORA backup:

1. Review the following sections:

• “Deduplication backups and restores” on page 20

• “Requirements for a deduplication backup” on page 87

• “Best practices for a deduplication backup” on page 87

2. Follow the configuration steps in “Configure a deduplication backup” on page 88.

Configuring a deduplication backup 87


Requirements for a deduplication backup

Before you configure a deduplication backup, ensure that all of the following requirements are met:

◆ The NSR-DBM-ORA client platform and operating system support deduplication.

◆ The NetWorker client and server releases support NSR-DBM-ORA deduplication. The NSR-DBM-ORA Release Notice provides more details.

◆ The Avamar server is installed and configured as a NetWorker deduplication node. The NetWorker documentation provides more details.

◆ The NetWorker backup device (receives only the backup metadata or hash ID during the NSR-DBM-ORA deduplication backup) is configured as an advanced file type device (AFTD), as described in the NetWorker Administrator’s Guide.

Best practices for a deduplication backup

This section provides recommendations on when to use NSR-DBM-ORA deduplication, and configuration tips to improve the performance of a deduplication backup.

The benefits of deduplication are dependent on the environment. Deduplication may be beneficial in a data warehouse environment where the data does not change frequently. It may also be beneficial for databases where only a small percentage of data is updated repeatedly, or new data is added to a database but the old data does not change much.

Deduplication is not recommended for RMAN incremental backups. During an incremental backup, Oracle transmits only the data blocks that have changed, which guarantees a very low rate of duplication. When the overhead of deduplicating data is added to the overhead of an Oracle incremental backup, the result is decreased performance and insignificant benefits for the amount of data stored.

During planning and configuration of a deduplication backup, keep in mind these best practices that can improve the backup performance:

◆ Do not create more than four backup channels for a deduplication backup.

◆ Do not use RMAN multiplexing for a deduplication backup. To disable multiplexing, ensure that filesperset is set to 1.

If you use the wizard to configure the deduplication backup, ensure that the Filesperset field is set to 1 on the Specify the RMAN Options (Part 1 of 2) screen.

◆ Do not use RMAN binary compression (for example, ZLIB) with a deduplication backup.

◆ Do not include a deduplication client in the same group as non-deduplication clients.

◆ Once a deduplication node (Avamar server) is selected for an initial full backup of a client, continue to use the same deduplication node for all of the client's backups, to take advantage of the data already stored on the server.



◆ Schedule a deduplication backup to avoid the Avamar server read-only periods. An Avamar server spends periods of time in maintenance mode, where it may be unavailable for backup or have limited bandwidth.

Note: A deduplication NSR-DBM-ORA backup that runs during such a maintenance mode period may be suspended until the Avamar server resources become available.

◆ Ensure that the same cache is used to back up the same data every time by not changing the values of the parameters NSR_DEDUP_CACHE_ENABLED and NSR_DEDUP_CACHE_TAG. The following sources provide more details:

• “NSR_DEDUP_CACHE_ENABLED” on page 193

• “NSR_DEDUP_CACHE_TAG” on page 193

◆ Associate the backup of specific tablespaces with a specific channel to ensure that Oracle does not distribute the data to a different channel when the database structure or size changes.

For example, the following RMAN backup script shows how to associate tablespaces with a channel:

run {allocate channel c1 type 'SBT_TAPE';send channel c1 'NSR_ENV=(NSR_DEDUP_CACHE_TAG=orcl102_c1)';allocate channel c2 type 'SBT_TAPE';send channel c2 'NSR_ENV=(NSR_DEDUP_CACHE_TAG=orcl102_c2)';send 'NSR_ENV=(NSR_DEDUP_BACKUP=TRUE, NSR_DEDUP_NODE=avamar.fujitsu.com )';backup filesperset=1(tablespace tbs1, tbs5 channel c1)(tablespace tbs2, tbs3, tbs4 channel c2);release channel c1;release channel c2;

}

◆ Group tablespaces that contain similar (duplicated) data and associate them with the same channel. This practice requires familiarity with the database data. A tablespace must also be added to the backup script when a new tablespace is created.

Avamar and NetWorker documentation provides more information on Avamar server and NetWorker requirements for deduplication backups.

Configure a deduplication backup

To configure a scheduled deduplication backup, you can use either the configuration wizard or the legacy configuration method.

Note: You must configure a manual deduplication backup with the legacy method only. However, you can generate the base RMAN script for a manual backup with the configuration wizard, described in “Configuring a backup with the NMC wizard” on page 68.

◆ To use the configuration wizard (for a scheduled backup only), follow the instructions in “Configuring a backup with the NMC wizard” on page 68.

On the Specify the De-duplication Options screen, apply these additional settings:

• Select to enable deduplication.

• Specify the hostname of the deduplication node (Avamar server) that will store the deduplicated backup data.

Configuring a deduplication backup 89


When you select to enable deduplication, the wizard automatically sets the parameter NSR_DEDUP_CACHE_TAG for each channel.

Note: You cannot override the default NSR_DEDUP_CACHE_TAG setting.

◆ To use the legacy configuration method (without the wizard), follow the instructions in “Configuring a backup with the NMC legacy method” on page 70.

Note: For a manual deduplication backup, the Client resource needs to include only the two attribute settings in step a on page 89.

Apply these additional settings for a deduplication backup:

a. Set the following attributes in the Client resource for the NSR-DBM-ORA client by using NMC:

– Select the De-duplication Backup attribute, to enable deduplication.– For the De-duplication Node attribute, specify the hostname of the

deduplication node (Avamar server) that will store the deduplicated backup data.

Note: For a manual deduplication backup, the De-duplication Node attribute must have the same value as the NSR_DEDUP_NODE setting in the RMAN script.

b. Set the following parameters in the RMAN backup script, if required:

– “NSR_DEDUP_BACKUP” on page 193 (for a manual backup only)– “NSR_DEDUP_CACHE_TAG” on page 193 – “NSR_DEDUP_NODE” on page 194 (for a manual backup only)Example 13 on page 89 shows a sample RMAN script for a manual deduplication backup.

For a scheduled deduplication backup, if the Client resource contains the two attribute settings described in step a on page 89, the RMAN script must not include the parameters NSR_DEDUP_BACKUP and NSR_DEDUP_NODE.

Example 13 RMAN script for a manual deduplication backup

The following RMAN script shows the mandatory parameter settings for a manual deduplication backup. The NSR_DEDUP_CACHE_TAG parameter must be set to a different value for each allocated channel:

run {allocate channel ch1 type 'SBT_TAPE';allocate channel ch2 type 'SBT_TAPE’';send 'NSR_ENV=(NSR_SERVER=mars.fujitsu.com , NSR_CLIENT=oracle.fujitsu.com ,NSR_DEDUP_BACKUP=TRUE, NSR_DEDUP_NODE=node3.fujitsu.com )'; send channel ch1 'NSR_ENV=(NSR_DEDUP_CACHE_TAG=ora11_ch1)';send channel ch2 'NSR_ENV=(NSR_DEDUP_CACHE_TAG=ora11_ch2)';backup full filesperset 4 format 'FULL_%d_%U' (database);release channel ch1;release channel ch2;

}



Configuring a probe-based backupTo configure a probe-based NSR-DBM-ORA backup:

1. Review the following sections:

• “Probe-based backups” on page 21

• “Requirements for a probe-based backup” on page 90

2. Follow the configuration steps in “Configure a probe-based backup” on page 90.

Requirements for a probe-based backup

Before you configure a probe-based backup, ensure that the following requirements are met:

◆ The required NetWorker releases are installed, as described in the NSR-DBM-ORA Release Notice.

◆ If you want to use the nsrnmoprobe program provided with NSR-DBM-ORA to check for the number of Oracle logs generated since the last probe-based backup (as a condition that triggers a probe-based backup), ensure that you have reviewed the details on nsrnmoprobe in “Configure a probe-based backup” on page 90.

◆ If you want to check for a user-defined condition (other than the number of generated Oracle logs) that triggers a probe-based backup, a script/program is created that meets the requirements of the Probe Command attribute of the Probe resource, as described in “Configure a probe-based backup” on page 90.

For example, the user-defined condition that triggers a probe-based backup is that more than two tape drives are idle in a jukebox. To check for this condition, a script named nsrjukeboxprobe is created in the /opt/nsr directory on Solaris. When the script runs and checks the number of idle tape drives in the jukebox, it returns one of the following values:

• 0 — Signifies that more than two tape drives are idle in the jukebox.

• 1 — Signifies that two or fewer tape drives are idle in the jukebox.

• Other than 0 or 1 — Signifies that an error occurred during the probe.

The NetWorker Administrator’s Guide provides more information on user-defined probes in the section on creating a client probe.

Configure a probe-based backup

To configure a probe-based NSR-DBM-ORA backup:

1. Create a separate NetWorker Probe resource for the nsrnmoprobe program and any other script/program that checks for a user-defined condition. Set the Probe resource attributes as described in Table 6 on page 91.

Configuring a probe-based backup 91


Table 6 NetWorker Probe resource attributes


Name Name of the Probe resource for the probe script/program used to check for a probe-based backup condition. Each Probe resource must have a unique name, which does not have to be the same as the probe script/program name.

Probe Command Name and path of the probe script/program that checks (“probes”) for the condition that triggers a probe-based backup. The script/program meets the following requirements:• Name starts with nsr or save.• Location is the same directory as used for the NetWorker client binaries.• Permissions of the script/program file include the “execute” permission.• Returns one of the following code values when it finishes running its probe:

- 0 — Signifies that the backup condition has been met.- 1 — Signifies that the backup condition has not been met.- Other than 0 or 1 - Signifies that an error occurred during the probe.

To use the probe program that is included with the NSR-DBM-ORA software, set this attribute to nsrnmoprobe. The nsrnmoprobe program checks for the number of Oracle logs generated since the last probe-based backup and triggers a new probe-based backup when the number exceeds the change threshold.

Command Options Required for the nsrnmoprobe program only, a comma-separated list of the settings of one or more of the following parameters. Example 14 on page 92 outlines the parameter settings that nsrnmoprobe requires in each possible scenario.

Parameter Description

LOG_THRESHOLD Mandatory. Specifies the change threshold, which is the minimum number of redo logs (generated since the last probe-based backup) required to trigger a new probe-based backup. When the number of logs generated since the last probe-based backup equals or exceeds the change threshold, nsrnmoprobe triggers a probe-based backup.

NSR_DEBUG_FILE Optional. Specifies the pathname of the log file that receives debug information from the probe-based backup triggered by nsrnmoprobe.

NSR_ORACLE_CONNECT_FILE Set only if both of the following are true:• The Client resource is not configured with the wizard; it is configured through the

legacy method with NMC.• The NWORA resource file is not set up with the Oracle home and database

connection information.Specifies the pathname of the RMAN connection file, which contains the connection strings required to connect to the Oracle database that is to be probed.Example 14 on page 92 provides a sample setting of this parameter.

ORACLE_HOME Set only if both of the following are true:• The Client resource is not configured with the wizard; it is configured through the

legacy method with NMC.• The NWORA resource file is not set up with the Oracle home and database

connection information.Specifies the home directory pathname for the Oracle database that is to be probed.Example 14 on page 92 provides a sample setting of this parameter.

ORACLE_SERVICE Set only if both of the following are true:• The Client resource is not configured with the wizard; it is configured through the

legacy method with NMC.• The NWORA resource file is set up with the Oracle home and database connection

information through the command nsrnmoadmin –r add sid=Net_service_name home=Oracle_home connect=connect_filepath.

Specifies the Net service name for the Oracle database that is to be probed. In this case, the ORACLE_SERVICE setting must be the same as the NSR_ORACLE_SID setting in the NWORA resource file.Example 14 on page 92 provides a sample setting of this parameter.



Note: The State attribute of the Probe resource is visible only in diagnostic mode. At the end of each successful probe-based backup, the nsrnmoprobe program stores the current log sequence number and database instance number in the State attribute. The State attribute is not used with user-defined probes.

2. Configure the NetWorker Group resource for a probe-enabled backup group. Set the probe-specific attributes in the Group resource, as described in the NetWorker Administrator’s Guide (the section on creating and scheduling a probe group). The Group resource has been extended to include several new attributes that must be set for a probe-based backup group.

When probing is enabled through the Group resource attributes, probing occurs continuously throughout the backup window, which is the time window defined by the Probe Start Time and Probe End Time attributes.

Note: If a probe-enabled backup group is started manually, probing occurs immediately (only once, not repeatedly at intervals) and the backup starts only if the probe conditions are met.

3. Configure the NetWorker Client resource for the NSR-DBM-ORA client according to the instructions in the appropriate section:

• “Configuring a backup with the NMC wizard” on page 68

• “Configure a Client resource with NMC” on page 81

In the Client resource:

• For the Probe attribute, specify the name of the required Probe resource from step 1. This attribute associates the Client resource with the probe script/program specified in the Probe resource.

Note: A Client resource can be associated with only one probe.The configuration wizard does not display the Probe field. If you configure a Client resource with the wizard, you must then use NMC manually to edit the Client resource and set the Probe attribute.

• For the Group attribute, specify the probe-enabled group from step 2.

Note: A probe-based backup group must include at least one probe-enabled client.

Example 15 on page 93 describes the configuration of a probe-based backup that is triggered by both nsrnmoprobe and a user-defined probe.

Example 14 Possible Command Options settings for the nsrnmoprobe program

To use the nsrnmoprobe program that is provided with the NSR-DBM-ORA software, the Probe resource must be set up properly, as described in “Configure a probe-based backup” on page 90.



The Command Options attribute in the Probe resource must include specific parameter settings, which depend on the particular scenario:

◆ The LOG_THRESHOLD parameter is always mandatory.

◆ The NSR_DEBUG_FILE parameter is optional.

◆ Three possible scenarios dictate the required settings in the Command Options attribute:

a. The Client resource has been configured through the legacy method with NMC (not with the wizard), and the NWORA resource file has not been set up with the Oracle home and database connection information.

In this case, Command Options must include the parameters LOG_THRESHOLD, NSR_ORACLE_CONNECT_FILE, and ORACLE_HOME. (NSR_DEBUG_FILE is optional.)

For example, Command Options is set as follows:

LOG_THRESHOLD=10, NSR_DEBUG_FILE=/tmp/probe.log, NSR_ORACLE_CONNECT_FILE=/RMAN/rmanpw, ORACLE_HOME=/oracle/hp_11_31_ia/64bit/product/11.1.0.6.0

b. The Client resource has been configured with the wizard, and the NWORA resource file has not been set up with the Oracle home and database connection information.

In this case, Command Options must include the LOG_THRESHOLD parameter. (NSR_DEBUG_FILE is optional.)


LOG_THRESHOLD=10, NSR_DEBUG_FILE=/tmp/probe.log

c. The Client resource has been configured through the legacy method with NMC (without the wizard), and the NWORA resource file is set up to retrieve Oracle home and database connection information.

In this case, the NWORA resource file must be set up through the command nsrnmoadmin –r add sid=Net_service_name home=Oracle_home connect=connect_filepath. Command Options must include the parameters LOG_THRESHOLD and ORACLE_SERVICE, where ORACLE_SERVICE is set to the same Net service name as NSR_ORACLE_SID in the NWORA file. (NSR_DEBUG_FILE is optional.)

“Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides important details on the nsrnmoadmin command.


LOG_THRESHOLD=10, NSR_DEBUG_FILE=/tmp/probe.log, ORACLE_SERVICE=proddb.world

Example 15 Multiple probes for a probe-based backup

A probe-based backup is triggered when both of the following are true:

◆ At least 25 Oracle log files are generated on an NSR-DBM-ORA client named mars.

◆ More than two tape drives are idle in a jukebox, attached to a NetWorker storage node named marmaris. (The jukebox is used to save the data for the probe-based backup.)

Both the NSR-DBM-ORA client and storage node are Solaris machines.



The nsrnmoprobe program is installed with the NSR-DBM-ORA software in /opt/nsr on the NSR-DBM-ORA client. The nsrnmoprobe program checks for the number of Oracle log files generated on the NSR-DBM-ORA client.

A script named nsrjukeboxprobe is created with “execute” permissions and stored in the /opt/nsr directory on the storage node. The script checks for the number of idle tape drives in the jukebox, and returns either of two values:

◆ 0 — Signifies that more than two tape drives are idle in the jukebox.

◆ 1 — Signifies that two or fewer tape drives are idle in the jukebox.

To configure the probe-based backup, the following steps are performed:

1. A Probe resource is created for the nsrnmoprobe program with the following attribute settings:

• Name — NSR-DBM-ORA probe

• Probe Command — nsrnmoprobe

• Command Options — LOG_THRESHOLD=25, NSR_DEBUG_FILE=/tmp/probe.log

2. A Probe resource is created for the user-defined probe with the following attribute settings:

• Name — Jukebox probe

• Probe Command — nsrjukeboxprobe

3. A Group resource is created with the required attribute settings for the probe-enabled backup group, including the following:

• Name — probe_group

• Probe Based Backup — Enabled (this is a checkbox in NMC)

• Probe Success Criteria — All

4. A Client resource is created for the NSR-DBM-ORA client through the configuration wizard. The Client resource includes the following attribute settings:

• Name — mars

• Backup Command — nsrnmostart -C

• Group — probe_group

• Probe — NSR-DBM-ORA probe

• Save Set — RMAN:/orcl102_FULL

5. A Schedule resource is created with the following attribute settings:

• Name — SkipAll

• Period — Either Week or Month

• Calendar — Skip level is selected for every day in the period

6. A “dummy” Client resource is created for the storage node through the legacy method (without the wizard). The Client resource includes the following attribute settings:

• Name — marmaris

• Backup Command — (blank)

• Group — probe_group



• Probe — Jukebox probe

• Save Set — SKIP (a keyword is required in this attribute)

• Schedule — SkipAll

Note: The Skip level in the SkipAll schedule causes the backup to be skipped on the storage node. The probe runs on the storage node as specified through the Group resource; the probe is not affected by the Schedule resource.



Backup Procedures 97

3


◆ Manual backup procedures ......................................................................................... 98◆ Scheduled backup procedures .................................................................................. 104◆ Regular backup information in NetWorker indexes.............................................. 108◆ Deduplication backup information in NetWorker indexes................................... 109

Backup Procedures


Backup Procedures

Manual backup proceduresTo perform a manual NSR-DBM-ORA backup of Oracle data:

1. Review information in “Product features” on page 19 about the features that apply to your particular backup.

2. Review “Requirements for a manual backup” on page 98.

3. Review the information on how to cancel a manual backup in “Cancel a manual backup” on page 101.

4. Perform the manual backup by either of the following methods:

• Use the RMAN command line interface, according to “Manual backup through the RMAN command line interface” on page 99.

• Use the Oracle Enterprise Manager Backup Management Tools, according to “Manual backup through Oracle Enterprise Manager Backup Management Tools” on page 99.

IMPORTANT!The NetWorker server bootstrap and client indexes are not automatically backed up at the end of a manual Oracle backup, as they are for a scheduled Oracle backup. After running a manual Oracle backup, perform a NetWorker server bootstrap backup according to “NetWorker server bootstrap backup” on page 100. Regular NetWorker server bootstrap backups help to ensure adequate preparation for disaster recovery.

“Monitor a manual backup” on page 101 provides information on how to monitor the status of a manual NSR-DBM-ORA backup.

“Scheduled backup procedures” on page 104 provides information on scheduled NSR-DBM-ORA backups.

Requirements for a manual backup

Before starting a manual backup, ensure that the required configurations and scripts are in place:

◆ The Oracle system is properly configured, as described in Chapter 2, “Software Configuration.”

The Oracle Server has the Net service configured. If a Recovery Catalog is to be used, the Recovery Catalog is set up and the target database registered.

◆ The NetWorker server has the required resources configured, as described in Chapter 2, “Software Configuration”:

• Server

• Client

• Device

• Pool

◆ If using an RMAN script for the manual backup, the script is created, as described in “Create RMAN scripts for backups” on page 70.

Manual backup procedures 99

Backup Procedures

◆ The required backup devices and volumes are set up:

• A volume is labeled (and mounted, as required) in each configured backup device.

• If auto media management is enabled for a tape device or autochanger, a tape is loaded in the tape device or autochanger for the NetWorker server to use automatically.

Manual backup through the RMAN command line interface

A manual Oracle backup can be started by using the RMAN command line interface on the Oracle Server host.

If the RMAN script for a manual backup from Example 10 on page 71 is stored in the file /disk1/scripts/full_backup.txt on a UNIX Oracle Server, and the Net service has been configured to connect to the databases payroll and rcvcatdb, then the manual Oracle backup can be started with the following command:

rman target internal/oracle@payroll rcvcat rman/rman@rcvcatdb cmdfile \’/disk1/scripts/full_backup.txt\’

IMPORTANT!The NetWorker server bootstrap and client indexes are not automatically backed up at the end of a manual Oracle backup, as they are for a scheduled Oracle backup. After running a manual Oracle backup, perform a NetWorker server bootstrap backup according to “NetWorker server bootstrap backup” on page 100. Regular NetWorker server bootstrap backups help to ensure adequate preparation for disaster recovery.

“Cancel a manual backup” on page 101 provides information on how to cancel a manual NSR-DBM-ORA backup.

Manual backup through Oracle Enterprise Manager Backup Management Tools

The Oracle Enterprise Manager Backup Management Tools can be used instead of the RMAN command line interface to start an Oracle database backup. The Oracle Enterprise Manager Backup Management Tools include a graphical user interface to RMAN for generating the required RMAN commands and performing backup and restore operations.

IMPORTANT!After the completion of an NSR-DBM-ORA backup or restore, the Oracle Enterprise Manager job queue history displays the status of the job as “failed”, even if the backup or restore completed successfully. This is due to a known problem with Oracle Enterprise Manager. View the job output to confirm that the backup or restore completed successfully.

The Oracle Enterprise Manager documentation included with the product kit provides more information on using the Oracle Enterprise Manager Backup Management Tools.


Backup Procedures

NetWorker server bootstrap backup

The bootstrap is a special save set that the NetWorker server software creates in preparation for disaster recovery. The bootstrap save set contains the information needed to restore the online NetWorker indexes and resource configuration files to the state just before the bootstrap was created.

The NetWorker server automatically performs a backup of its bootstrap and the client index of the Oracle Server only at the end of a scheduled Oracle backup. This is done by the NetWorker savegrp program that performs the scheduled backup.

Note: If only manual (unscheduled) Oracle backups are run, and the client index and bootstrap are not backed up manually, then no backups of the online client index and NetWorker server bootstrap will be available for use in the event of a disaster recovery on the NetWorker server.

After finishing a manual Oracle backup, the bootstrap and client index must also be backed up.

Back up the bootstrap and index

To back up the NetWorker server bootstrap and index:

1. On a UNIX NetWorker server, log in as the root user.

2. Type the following savegrp command:

savegrp -O -l full -P printer_name -c Oracle_Server_name-c NetWorker_server_name

where:

• printer_name is the name of the printer where the bootstrap information is printed at the end of the bootstrap backup.

• Oracle_Server_name is the hostname of the Oracle Server.

• NetWorker_server_name is the hostname of the NetWorker server.

After successfully backing up the bootstrap and the client index by using the savegrp command:

◆ Confirmation of the savegrp completion appears in the NetWorker Administrator program.

◆ Information is sent to the printer_name printer about the saved bootstrap.

Note: Store the bootstrap printout in a safe place. The printed bootstrap information includes dates, locations, and save set ID numbers for the bootstrap save sets backed up during the past month. With this information, determine which volumes are needed to recover the NetWorker indexes and resource configuration files during a disaster recovery.

The following sources provide information on the savegrp command and options:

◆ NetWorker Administrator’s Guide for the particular NetWorker server platform

◆ savegrp entry in the NetWorker Command Reference Guide

◆ savegrp man page on UNIX

The NetWorker Administrator’s Guide provides information on bootstrap backups.

The NetWorker Disaster Recovery Guide provides information on how to use the bootstrap backup during a disaster recovery.


Backup Procedures

Cancel a manual backup

To keep the NetWorker index and RMAN catalog synchronized, ensure that the Oracle user has the required NetWorker privileges for removing NetWorker index entries before you cancel a manual Oracle backup. “Specify the required NetWorker privileges” on page 60 provides more information.

The following sections describe how to cancel a running or nonresponding manual backup.

Cancel a running manual backup

To cancel a running manual Oracle backup (not including a backup initiated by Oracle Enterprise Manager) during the RMAN script execution on any platform, press either Ctrl+C, or the equivalent “attention” key combination on the system.

Note: If this method of canceling a manual Oracle backup is not successful, use the alter system kill command described in “Cancel a scheduled backup” on page 105.

Cancel a nonresponding manual backup

The following are the steps for canceling a nonresponding Oracle backup. However, if these steps do not work, contact Oracle for assistance.

Note: When using these steps, NSR-DBM-ORA does not attempt to remove the backup save set entries from the NetWorker index. As a result, the NetWorker index and RMAN catalog might become unsynchronized.

To cancel a nonresponding Oracle backup on UNIX:

1. Include the set command id to ‘xxx’ command in the RMAN backup script that is used for the Oracle backup. Otherwise, the query in the next step will fail. Example 12 on page 73 provides a sample script with the command.

2. Run the following query in the Oracle svrmgrl or sqlplus program to determine the Oracle process ID that corresponds to each RMAN channel:

select spid, client_info from v$process p, v$session s where p.addr=s.paddr and client_info like ‘%id=%‘;

3. Type the following kill command to cancel the Oracle backup process:

kill -9 pid

where pid is the appropriate Oracle process ID.

Monitor a manual backup

To monitor the status of Oracle backup and restore operations, use the NMC interface program.

The NetWorker server program displays progress and completion messages that advise when a backup or restore is complete, and information on why a backup or restore cannot proceed. The NetWorker Administrator’s Guide provides more information on viewing these types of messages with the NetWorker program.

Appendix C, “Troubleshooting and Error Messages,” provides information on how to obtain RMAN-specific and NSR-DBM-ORA diagnostic and error messages.

The following figures show the types of Oracle backup messages displayed in the Monitoring Window of the NMC interface:


Backup Procedures

◆ Figure 7 on page 102 shows messages displayed in the Sessions tab.

◆ Figure 8 on page 103 shows messages displayed in the Devices tab.

◆ Figure 9 on page 103 show messages displayed in the Log tab.

The NetWorker Administrator’s Guide and NMC online help provide more information.

Figure 7 Oracle backup messages in Sessions tab of Monitoring window


Backup Procedures

Figure 8 Oracle backup messages in Devices tab of Monitoring window

Figure 9 Oracle backup messages in Log tab of Monitoring window


Backup Procedures

Scheduled backup proceduresTo prepare for a scheduled NSR-DBM-ORA backup of Oracle data:

1. Review information in “Product features” on page 19 about the features that apply to your particular backup.

2. Ensure that the required backup configurations are in place, as described in “Configuration roadmap” on page 58.

3. Run a test scheduled Oracle backup according to “Test a scheduled backup” on page 104.

4. Review the steps for canceling a scheduled backup in “Cancel a scheduled backup” on page 105.

At the end of a successful scheduled backup, NSR-DBM-ORA automatically backs up the NWORA resource file if it exists, as described in “NWORA resource file backup” on page 167.

“Monitor a scheduled backup” on page 106 provides information on how to track the status of scheduled backups.

The following sections describe the backup information stored in the NetWorker indexes:

◆ “Regular backup information in NetWorker indexes” on page 108

◆ “Deduplication backup information in NetWorker indexes” on page 109

Test a scheduled backup

After the NSR-DBM-ORA environment for scheduled backups on the Oracle Server is set up, test a scheduled backup manually by using the NMC program.

Before starting the backup, check for the following:

◆ All the required parameters have been set, either through the wizard or through the legacy method in the nsrnmo and RMAN scripts.

◆ For a legacy configuration, any required preprocessing and postprocessing scripts (optional) are in place.

◆ The NetWorker server and Oracle Server have been correctly configured.

Test a scheduled backup as follows:

1. Start the NMC program as the root user on UNIX.

2. Select the appropriate NetWorker server to perform the backup.

3. Select the correct group name for the backup.

4. Start the scheduled backup for the specified group.

The NetWorker software immediately backs up the clients in the backup group.

The NetWorker Administrator’s Guide and NMC online help provide information on how to use the NMC interface to perform these steps.

A savegroup completion report in email provides the scheduled backup results.

Note: “Cancel a scheduled backup” on page 105 provides information on how to cancel the scheduled Oracle backup.

Scheduled backup procedures 105

Backup Procedures

Scheduled backup error messages

If the scheduled backup fails, an error message is produced. The NetWorker Administrator’s Guide provides information on how to obtain more details about the scheduled backup by using the NetWorker administration program.

For additional debug information, set the NSR_SB_DEBUG_FILE parameter through one of the following methods:

◆ Use the configuration wizard to set the parameter in the Advanced Environment Options field on the Specify the Environment Options (Optional) screen.

◆ Use the legacy configuration method to set the parameter in the nsrnmo script.

“NSR_SB_DEBUG_FILE” on page 77 provides details on the parameter.

If the scheduling portion of a scheduled NSR-DBM-ORA backup succeeds but the actual backup fails, error messages and debug information might be generated in the locations described in “RMAN error messages” on page 211.

Cancel a scheduled backup

To keep the NetWorker index and RMAN catalog synchronized, ensure that the Oracle user has the required NetWorker privileges for removing NetWorker index entries before canceling a scheduled Oracle backup. “Specify the required NetWorker privileges” on page 60 provides more information.

NSR-DBM-ORA currently supports use of the Stop button in the NetWorker Administrator program to cancel a running scheduled Oracle backup only if the NetWorker server and Oracle Server are both running on UNIX platforms.

The level of Stop button support depends on the NetWorker server operating system (where the Oracle Server operating system is already assumed to be UNIX). If the NetWorker server system is UNIX, then the Stop button is supported.

Note: Although the progress message might indicate that the backup has exited prematurely (before finishing), the NSR-DBM-ORA processes might still be running on the Oracle Server host.

To cancel a running scheduled Oracle backup when the Stop button does not work, the running rman process must be interrupted on the Oracle Server host:

1. In the NetWorker Administrator program, click the Stop button to prevent NSR-DBM-ORA from retrying the backup.

2. For each allocated channel, perform the following:

a. View the RMAN message log file to determine the Oracle session ID for the channel. (The log filename is specified in the NSR_RMAN_ARGUMENTS parameter, set either through the Advanced Environment Options field in the wizard or through the nsrnmo script in a legacy configuration.)

For example, the following sample line from an RMAN message log shows that channel ch1 has the Oracle session ID 15:

channel ch1: sid=15 devtype=SBT_TAPE

b. Run the following select command in the Oracle svrmgrl or sqlplus program to determine the serial number:

select serial# from v$session where sid=session_id;


Backup Procedures

where session_id is the Oracle session ID from the RMAN message log in step a.

c. Run the following alter system command in the Oracle svrmgrl or sqlplus program to terminate the channel:

alter system kill session ‘session_id, serial#’;

where:

– session_id is the Oracle session ID from step a.– serial# is the serial number from step b.

“Cancel a nonresponding manual backup” on page 101 describes how to also cancel a nonresponding scheduled backup.

Monitor a scheduled backup

Scheduled Oracle backups can be monitored the same way as manual Oracle backups. “Monitor a manual backup” on page 101 provides more information.

In addition, messages appear in the appropriate Group window in the NetWorker server program and a report is generated upon completion of a scheduled Oracle backup.

Information about scheduled NSR-DBM-ORA backups is displayed on the Groups and Sessions tabs of the Monitoring window in the NMC interface program:

◆ During the backup, the Sessions tab lists one save session for each channel used by NSR-DBM-ORA for the backup. The display shows the rate of data being backed up and total size of the backed-up data.

◆ After the backup, the Groups tab enables you to select the backup group and display details about the group, including the “Completed successfully” box that contains one entry for each channel used by NSR-DBM-ORA for the backup. Figure 10 on page 107 shows the type of group details displayed, including the size of each save set.

Note: For a deduplication backup, the NMC display shows the total size of the data prior to deduplication, not the size of the deduplicated data. The display also does not indicate that the data is for a deduplication backup that is stored on the Avamar server.

The NetWorker Administrator’s Guide provides details on viewing information about scheduled backups in the NMC program.

Scheduled backup procedures 107

Backup Procedures

Figure 10 Group details for regular scheduled backups

Automation of NSR_SERVER and NSR_GROUP parameter settings

Note: The information in this section is applicable to a legacy configuration only, not to a configuration set up through the NSR-DBM-ORA 5.0 wizard.

The parameters NSR_SERVER and NSR_GROUP must not be set in a scheduled backup RMAN script. NSR-DBM-ORA automatically passes server and group information (received from the NetWorker server that started the backup) to Oracle Server processes, created for each allocated channel.

NSR-DBM-ORA also passes an internal parameter, NSR_SAVESET_NAME, to set the save set name.

IMPORTANT!Do not set the parameter NSR_SAVESET_NAME.

To pass the information to the Oracle Server processes, NSR-DBM-ORA invokes the RMAN executable with send as a command line option. “The send command” on page 201 provides more information.


Backup Procedures

Regular backup information in NetWorker indexesThe NetWorker server maintains information about each backup in its online indexes. Chapter 1, “Introduction,”provides more information about the online indexes.

This section describes the information maintained in the NetWorker indexes for regular manual and scheduled backups. “Deduplication backup information in NetWorker indexes” on page 109 describes information about deduplication backups.

The NetWorker client file index contains the filename backup_piece_name for the save set name for either a manual or scheduled NSR-DBM-ORA backup.

The NetWorker media database contains the following information, depending on the type of backup and the NetWorker server version used to perform the backup:

◆ For a manual backup, the name RMAN:backup_piece_name for the save set name.

◆ For a scheduled backup with a NetWorker server release prior to 7.3.2, the value from the Save Set field of the Client resource for the save set name.

◆ For a scheduled backup with NetWorker server release 7.3.2 or later, the name RMAN:backup_piece_name for the save set name.

Query the online NetWorker indexes by using the NetWorker commands, nsrinfo and mminfo.

◆ To query the client file index, use the nsrinfo command. For example:

nsrinfo -n oracle -s NetWorker_server_hostname Oracle_Server_hostname

◆ To query the media database, use the mminfo command. For example:

mminfo -v -s NetWorker_server_hostname -c Oracle_Server_hostname

The NetWorker Command Reference Guide and the UNIX man pages provide more information on these NetWorker commands.

The following examples show the command output for a manual or scheduled NSR-DBM-ORA backup, where the backup piece is named 1hiu83f4_1_1 and the scheduled backup is performed with NetWorker server release 7.3.2 or later:

◆ The client file index includes the backup piece name for the save set:

nsrinfo -n oracle -s ca-oracle1 ca-oracle1

1hiu83f4_1_1, date=1192133159 Thu Dec 11 16:05:59 2008

◆ The media database includes the prefix RMAN: with the backup piece name for the save set:

mminfo -v -s ca-oracle1 -c ca-oracle1

volume client date time sizeNMO.001 ca-oracle1 12/11/08 16:05:59 145 MB

ssid fl level name4212032038 cb full RMAN:1hiu83f4_1_1

Note: The media database also includes information about the bootstrap, index, and NWORA resource file backups that occur as part of each scheduled backup. The preceding mminfo command sample does not show the bootstrap, index, and NWORA resource file information for the scheduled backup.

Deduplication backup information in NetWorker indexes 109

Backup Procedures

Cross-check the client file index and media database by using the save time. For example:

mminfo -c ca-oracle1 -t 1192133159nsrinfo -n oracle -t ’12/11/08 16:05:59’ ca-oracle1

Deduplication backup information in NetWorker indexesThe deduplication backup information maintained in the NetWorker indexes is for the NetWorker save set stored on the NetWorker storage node, which contains only a single file with the hash ID and NetWorker metadata for the backup.

Note: The backup size recorded in the NetWorker indexes does not match the amount of data backed up during the deduplication backup.

The following examples show the index query results from the nsrinfo and mminfo commands, where the backup piece is named 35ji3i75_1_1:

◆ As for a regular backup, the client file index includes the backup piece name for the save set of the deduplication backup:

nsrinfo -n oracle -vV molecule

UNIX file '35ji3i75_1_1', size=212, off=0, app=oracle(25), date=1212610342 Wed Jun 04 16:12:22 2008, (unknown fid), file size=0

In this case, the size of the backup piece is the size of the Avamar hash ID (20 bytes) plus the size of the NetWorker metadata that contains the backup piece name.

◆ To query the media database, use the mminfo command with the -q dedupe option (displays only save sets created through deduplication) and the -S option (lists the extended options for the deduplication backups):

mminfo -S -q dedupe

ssid=1330050855 savetime=06/04/08 16:12:22 (1212610342) RMAN:35ji3i75_1_1 level=full sflags=vF size=216 files=1 insert=06/04/08 create=06/04/08 complete=06/04/08 browse=07/04/08 23:59:59 retent=06/04/09 23:59:59 clientid=08b98b33-00000004-4824ae51-4824ae45-000c0000-80de6f33 *Client path: /NetWorker/server_name/client_name; *Data set size: 34000000; *De-Dup session id: 27; *De-Dup snapup time: 2008-06-04; *De-duplication: Yes; *De-duplication host: scip2d088.lss.fujitsu.com ; *Domain: /NetWorker/server_name; *New data on De-Dup Node: 6899656.00; *New files: 1; *Replication host: scip2d089.legato.com; *Size on De-Dup Node: 34603008.00; group: Default; Clone #1: cloneid=1212610343 time=06/04/08 16:12:23 retent=06/04/09 flags=F frag@ 0 volid=1648811587 file/rec=1330050855/0 rn=0 last=06/04/08

The save file size reported by mminfo is the same as that reported by nsrinfo, except for an extra four bytes due to a NetWorker end-of-save-set marker.

When a deduplication backup is deleted (for example, by a user or the NetWorker server), the backup information is deleted immediately from the NetWorker indexes, and a request is queued for deletion of the backup from the Avamar server. The


Backup Procedures

NetWorker documentation provides more information on deletion of deduplication backups.

Once a deduplication save set passes its retention time and its data chunks are deleted from the Avamar server, the save set may no longer be recoverable with the scanner program.

Data Restore and Recovery 111

4

This chapter includes the following sections:

◆ About restore and recovery ........................................................................................ 112◆ RMAN scripts for restore and recovery .................................................................... 115◆ Performing an Oracle data restore and recovery..................................................... 117

Data Restore andRecovery


Data Restore and Recovery

About restore and recoveryOnly Oracle data that has been backed up according to the instructions in Chapter 3, “Backup Procedures,”can be restored.

Note: The NetWorker server interface cannot be used to restore Oracle data that was backed up with NSR-DBM-ORA.

Review the following sections in preparation for the Oracle data restore and recovery procedures:

◆ “NetWorker indexes and policies used for restores” on page 112

◆ “Recovery configuration wizard” on page 112

◆ “RMAN scripts for restore and recovery” on page 115

NetWorker indexes and policies used for restores

During an Oracle backup, the NetWorker server adds an entry for each backup piece in the online client file index and records the location of the data in the media database. These entries provide information required to restore every Oracle data object backed up:

◆ The client file index entry is maintained until the browse policy specified for the client’s save set expires.

◆ The media database entry is maintained until the retention policy specified for the client’s save set expires.

When the retention policies for all the save sets on a backup volume expire, the volume becomes recyclable and eligible for automatic relabeling by the NetWorker server. The save set entries, however, remain in the media database until the volume is actually relabeled. When the volume is relabeled, the data on it becomes inaccessible and can no longer be restored.

NSR-DBM-ORA uses the client file index entries to restore Oracle backup pieces. Set the browse policy to a period long enough to retain the client index entries for restoring the Oracle backup pieces.

Note: After a browse policy expires, the NetWorker scanner program can be used to rebuild the online indexes. However, index entries regenerated by using scanner might cause the NetWorker indexes to become unsynchronized with the RMAN catalog and lead to problems. To avoid problems, ensure that the backup pieces have unique names, as described in “RMAN scripts for manual backups” on page 71.

The NetWorker Administrator’s Guide provides more information on how the NetWorker server uses browse and retention policies to manage Oracle backup data and track the location and status of the data on backup volumes.

Recovery configuration wizard

NSR-DBM-ORA 5.0 software supports a new recovery configuration wizard that is integrated with the NetWorker Management Console (NMC). The recovery configuration wizard used in NSR-DBM-ORA release 4.5 or earlier is replaced by the NMC-based wizard in NSR-DBM-ORA release 5.0.

About restore and recovery 113


You can run the recovery configuration wizard from the NetWorker Console Administration window, which you can start on any supported host by using a web browser session and specifying the Console server URL.

The NSR-DBM-ORA Release Notice provides details on the NetWorker requirements for the support of the NMC-based configuration wizards.

To configure a restore with the wizard:

1. Review the information in “Features of the recovery configuration wizard” on page 113.

2. Ensure that you meet the “Requirements for using the recovery configuration wizard” on page 114.

3. Follow the steps in “Configure a restore with the wizard” on page 114.

Features of the recovery configuration wizard

The recovery configuration wizard can create an RMAN script for the following types of restore and recovery:

◆ Current time restore and recovery of a whole or partial Oracle database, where a partial database is a set of tablespaces or datafiles.

The wizard can configure a tablespace restore as long as the control file contains information about the tablespace.

◆ Point-in-time restore and recovery of a whole Oracle database.

◆ Restore of individual archived redo logs.

◆ Restore and recovery of Oracle data to a different database through the creation of a duplicate database on either the local host or a remote host, using backups of the original target database.

The database duplication script created by the wizard uses the RMAN duplicate command to create a duplicate database while the original database is retained. The duplicate database can either be an identical copy of the original database or contain only a subset of the original tablespaces. For example, the duplicate database can be created to run independently on a remote host, for the purpose of practicing restore and recovery operations while the production database remains in operation on the local host:

• If the duplicate database is to be created on the same host as the original database, the RMAN script is also generated on the local host. In this case, the wizard requests names for the duplicate database, datafiles, and redo logs that differ from those of the original database.

• If the duplicate database is to be created on a remote host, the RMAN script is generated on either the local or remote host, as specified in the wizard. In this case, the wizard requests a name for the duplicate database that differs from that of the original database. (The datafile and redo log names can be the same as for the original database.)

The recovery configuration wizard can only create a new RMAN script for restore and recovery. The wizard cannot modify an existing RMAN script. You must use a text editor to modify an RMAN script that was created by the wizard.

“RMAN scripts for restore and recovery” on page 115 provides more information about RMAN scripts for restore and recovery.



The recovery configuration wizard does not support the following:

◆ Cluster or Oracle RAC systems

◆ Proxy backups

◆ RMAN automatic channels

The following sources provide more information on the configuration wizard:

◆ NSR-DBM-ORA Release Notice

◆ Descriptive inline text in the wizard

◆ Online help in the wizard

Requirements for using the recovery configuration wizard

Before you use the recovery configuration wizard, ensure that all of the following requirements are met:

◆ The NMC user that starts the wizard (the wizard user) has the Remote Access NetWorker privileges on the NetWorker server where the NSR-DBM-ORA client configuration is stored.

◆ Communication between the NMC server, NetWorker server, and NSR-DBM-ORA client uses nsrauth authentication. The NetWorker documentation provides any requirements for nsrauth authentication.

◆ The required NetWorker releases are installed on the NMC server, NetWorker server, and NSR-DBM-ORA client hosts, as described in the NSR-DBM-ORA Release Notice.

◆ The NetWorker Client resource for the NSR-DBM-ORA client was created through one of the following:

• Backup configuration wizard in NSR-DBM-ORA 5.0

• Migration of a legacy configuration with the nsrnmoadmin command

• Legacy configuration method (without the wizard), where the value of the Save Set attribute of the Client resource has the RMAN: prefix

◆ Prior to creation of a database duplication script, the AUXILIARY instance exists on the local or remote host, and is accessible through Oracle Net. The Oracle Database Backup and Recovery Advanced User’s Guide provides details on how to create an AUXILIARY instance.

Configure a restore with the wizard

To create an RMAN restore script with the recovery configuration wizard:

1. Start the NetWorker Management Console software.

2. Open the Administration window:

a. In the Console window, click Enterprise.

b. In the left pane, select a NetWorker server in the Enterprise list.

c. In the right pane, select the application.

d. From the Enterprise menu, click Launch Application.

The Administration window is launched as a separate application.

3. In the Administration window, click Configuration.

4. In the Configuration window, click Clients.

RMAN scripts for restore and recovery 115


5. To start the wizard, right-click the NSR-DBM-ORA client in the right pane, and select Recover.

6. On each wizard screen that appears, specify the required values for the RMAN script configuration.

Each wizard screen includes an online help button that you can click to access descriptions of all the fields and options on the screen:

• On all but the last screen, click Next to proceed.

• On the last screen, Review and Accept the Script Creation, click Create to create the RMAN restore script.

Note: The RMAN script created by the recovery wizard does not contain any connection strings to connect to the target database or to the recovery catalog. Before you can use this script to restore a database you must update the connect strings.

IMPORTANT!When you create an RMAN restore script with the wizard and select the offline or online mode option for tablespaces, you might need to manually edit the script and insert an "alter database open;" command before the sql...tablespace... commands, if it will be required for your particular database (if the database will not be open at the point where the sql commands are to be run).

RMAN scripts for restore and recoveryAn appropriate RMAN script is required to perform the preferred type of Oracle restore operation on the Oracle Server host. You can create the RMAN script either manually or by using the recovery configuration wizard. “Recovery configuration wizard” on page 112 provides details on the recovery configuration wizard.

RMAN restore scripts can be stored as text files. Alternatively, if a Recovery Catalog is used, restore scripts can be stored in the Recovery Catalog database. The Oracle backup and recovery documentation provides more information on storing the restore scripts in the Recovery Catalog database.

The NSR* parameters in the script must be set by the methods described in “Setting the NSR* parameters” on page 190. The use of the send command is recommended where possible. “The send command” on page 201 provides more information.

Example 16 RMAN script to restore a tablespace

The following RMAN script performs a restore of an Oracle tablespace by using the (remote) NetWorker server mars.fujitsu.com . The Oracle data is restored to the NetWorker client server1.fujitsu.com . This RMAN script also includes the recovery step, which is explained in “Perform an Oracle data recovery” on page 121:

run {allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com , NSR_CLIENT=server1.fujitsu.com )’;sql ’alter tablespace users offline immediate’;restore tablespace users;recover tablespace users;sql ’alter tablespace users online’;release channel t1;



release channel t2;}

“Create RMAN scripts for backups” on page 70 provides more information on setting NSR* parameters in an RMAN script.

Example 17 RMAN script to restore an AES encrypted backup

By default, if the NSR_ENCRYPTION_PHRASES parameter is not set, NSR-DBM-ORA obtains the encryption phrase from the NetWorker server for decrypting an AES encrypted backup during an Oracle restore.

The NSR_ENCRYPTION_PHRASES parameter must be set if both of the following are true:

◆ The Oracle data being restored was backed up with 256-bit AES encryption. Example 11 on page 72 shows a sample Oracle backup with AES encryption.

◆ The encryption phrase on the NetWorker server has changed since the Oracle data was backed up.

If both of these conditions are true, ensure that the NSR_ENCRYPTION_PHRASES parameter is set through the send command in the RMAN restore script. “NSR_ENCRYPTION_PHRASES” on page 194 provides more information.

The following RMAN script performs a nonproxy Oracle restore of the database by using the encryption phrase, original_encryption_phrase, that was originally used to back up the database. The encryption phrase must be in double quotes:

run {allocate channel t1 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com , NSR_CLIENT=server1.fujitsu.com , NSR_ENCRYPTION_PHRASES="original_encryption_phrase")’;restore database;release channel t1;

}

Example 18 RMAN script to restore from a specified pool

By default, NSR-DBM-ORA and NetWorker use configuration settings and information in the media database to determine the backup volume to use for an NSR-DBM-ORA restore.

As an alternative, you can use the NSR_RECOVER_POOL parameter in the RMAN restore script to restore data from a specified volume pool if there are multiple copies (clones) of the backup on different volume pools. “NSR_RECOVER_POOL” on page 196 provides more information.

The following RMAN script performs a nonproxy restore of the database from the specified volume pool named OracleClonePool2, where the pool contains a clone of the original backup volume.

shutdown immediate;startup mount;run {

allocate channel c1 type 'SBT_TAPE';send channel c1 ’NSR_ENV=(NSR_SERVER=backup01, NSR_RECOVER_POOL=OracleClonePool2)’; restore database;release channel c1;

}

Performing an Oracle data restore and recovery 117


Performing an Oracle data restore and recoveryTo perform a restore and recovery of Oracle data:

1. Review “Requirements for an Oracle data restore” on page 117.

2. Perform the Oracle data restore by either of the following methods:

• Use the RMAN command line interface, according to “Restore through the RMAN command line interface” on page 119.

• Use the Oracle Enterprise Manager Backup Management Tools, according to “Restore with Oracle Enterprise Manager Backup Management Tools” on page 120.

Note: Use of the Oracle Enterprise Manager Backup Management Tools is not supported with the recovery configuration wizard.

3. Complete the Oracle data recovery, if required, according to “Perform an Oracle data recovery” on page 121.

Requirements for an Oracle data restore

Before starting an Oracle data restore, ensure that the required configurations, volumes, and scripts are in place:

◆ The Oracle system is properly configured, as described in Chapter 2, “Software Configuration.”

◆ The NetWorker server has the required resources configured, as described in Chapter 2, “Software Configuration”:

• Server

• Client

• Device

• Pool

◆ If using an RMAN restore script, the script is created, as described in “RMAN scripts for restore and recovery” on page 115.

◆ The volume required for the restore operation is mounted in a configured backup device:

• If you use a stand-alone tape drive, the volume is mounted manually.

• If you use an autochanger, the NetWorker server mounts the volume automatically.

To determine the volumes required for the restore, you can use the nsrnmoinfo command, as described in “Using the nsrnmoinfo command to determine the volumes for restore” on page 117.

Using the nsrnmoinfo command to determine the volumes for restore

To determine the NetWorker volumes that contain the Oracle backup pieces to be restored, you can use the nsrnmoinfo command. The nsrnmoinfo program is installed with the NSR-DBM-ORA software in the same directory as the NetWorker client software.



The nsrnmoinfo command syntax and options are as follows:

nsrnmoinfo[.exe] [-c NetWorker_client_name] [-s NetWorker_server_name] [-f filename] [backup_piece_name1 [backup_piece_name2 ...]]

where:

◆ NetWorker_client_name specifies the hostname of the NetWorker client whose index contains information on the Oracle backup pieces. By default, the client is the local host.

◆ NetWorker_server_name specifies the hostname of the NetWorker server to query for the volumes. By default, the server is the local host.

◆ filename specifies the name of a text file that contains a list of one or more backup piece names for restore:

• The file must contain each backup piece name on a separate line.

• The file cannot contain spaces or comments (for example, comment lines preceded with the # symbol).

◆ backup_piece_name1 and backup_piece_name2 specify backup piece names for restore.

Command options in brackets ([ ]) are optional. Do not include the brackets when typing the command.

To use the nsrnmoinfo command, specify the names of the backup pieces by either or both of the following:

◆ List the backup piece names as options of the nsrnmoinfo command.

◆ List the backup piece names in a text file, and specify the name of the file with the -f option of the nsrnmoinfo command.

The nsrnmoinfo command displays a list of one or more volumes required for the Oracle restore:

◆ For each backup piece, the list includes the accessible volumes containing the backup piece, which the NetWorker server will use for the restore.

◆ For each volume, the list includes the following:

• The name and location of the volume.

• The save time of the backup piece on the volume.

The listed volumes are the most accessible ones, which the NetWorker server intends to use for the restore at the time that the command is typed:

◆ The command lists clones of volumes if the original volumes are not accessible.

◆ If any listed volumes are removed from the NetWorker devices or deleted after the nsrnmoinfo command is typed, the server can perform the restore by using different volumes that are accessible. For example, the server can use an accessible clone (already mounted in a drive or available for mounting in a jukebox) instead of a listed volume.



Example 19 Sample nsrnmoinfo commands for Oracle restores

Each of the following nsrnmoinfo commands displays a list of the volumes required to restore the specified backup pieces:

◆ The following command searches in the NetWorker index of the client mars on the server server1 for information on the volumes that contain the backup pieces backupc_1 and backupc_2:

nsrnmoinfo -c mars -s server1 backupc_1 backupc_2

◆ The following command searches in the NetWorker index of the local host for information on the volumes that contain the backup pieces listed in the file backup2.txt: (Both the NetWorker client and server are assumed to be the local host.)

nsrnmoinfo -f backup2.txt

◆ The following command searches in the NetWorker index of the client mars for information on the volumes that contain both:

• The backup piece backupc_3.

• The backup pieces listed in the file backup3.txt.

(The NetWorker server is assumed to be the local host.)

nsrnmoinfo -c mars backupc_3 -f backup3.txt

Example 20 Volume information displayed by the nsrnmoinfo command

The following nsrnmoinfo command searches in the NetWorker index of the local host on the server mars for information on the volumes containing the backup pieces backup1 and backup2:

nsrnmoinfo -s mars backup1 backup2

The nsrnmoinfo command displays the following type of information:

backup1:mars.003 at /space/nw_volume1 (save time 1098886937)mars.004 at /space/nw_volume2 (save time 1098883454)

backup2:mars.005 at /dev/rmt/0cbn (save time 1098883452)

According to this command display:

◆ Volumes mars.003 and mars.004 are required to restore the backup piece backup1.

◆ Volume mars.005 is required to restore the backup piece backup2.

Restore through the RMAN command line interface

An Oracle data restore can be started by using the RMAN command line interface on the Oracle Server host.

To restore an Oracle database to a different system from the one backed up, follow the instructions in “Restore to a different host” on page 120.

If the RMAN restore script on page 115 is stored in the file /disk1/scripts/restore.txt and the Net service has been configured to connect to the databases payroll and rcvcatdb, the Oracle restore can be started with the following command:

rman target internal/oracle@payroll rcvcat rman/rman@rcvcatdb cmdfile \’/disk1/scripts/restore.txt\’



Restore to a different host

In some cases, an Oracle database must be restored to a different system from the one that was originally backed up. Some situations where this could occur include, but are not limited to, performing a disaster recovery or duplicating a database on a remote host. The appropriate Oracle documentation provides more information.

To restore an Oracle database to different system:

1. On the NetWorker server that contains the backup to be restored, specify the hostname of the different system for the Remote Access attribute in the Client resource of the original system.

2. Install and configure NSR-DBM-ORA on the different system where the Oracle data is to be restored.

3. Install and configure RMAN on the different system.

4. Create an appropriate RMAN restore script.

5. In the RMAN script, set the parameter NSR_SERVER to the name of the NetWorker server that contains the backup.

6. In the RMAN script, set the parameter NSR_CLIENT to either the name of the NetWorker client (Oracle Server host) that was backed up or the value of NSR_CLIENT that was used during the backup.

7. Use the appropriate RMAN command to run the script to restore the Oracle database files.

Restore with Oracle Enterprise Manager Backup Management Tools

Note: Use of the Oracle Enterprise Manager Backup Management Tools is not supported with the recovery configuration wizard.

The Oracle Enterprise Manager Backup Management Tools include a graphical user interface to RMAN. This interface can be used instead of the RMAN command line interface to do the following:

◆ Generate the required RMAN commands.

◆ Perform backup and restore operations.

IMPORTANT!After the completion of an NSR-DBM-ORA backup or restore, the Oracle Enterprise Manager job queue history displays the status of the job as “failed”, even if the backup or restore completed successfully. This is due to a known problem with Oracle Enterprise Manager. View the job output to confirm that the backup or restore completed successfully.

The Oracle Enterprise Manager documentation included with the product kit provides more information on using the Oracle Enterprise Manager Backup Management Tools.



Perform an Oracle data recovery

After restoring the NSR-DBM-ORA backups of the Oracle data by using the RMAN utility, complete the data recovery, if required.

To recover the Oracle data, use the appropriate Oracle commands to apply the archived redo logs and online redo logs. There are two ways to use the Oracle recovery commands:

◆ Include the Oracle commands in the RMAN restore script. A sample RMAN script is provided on page 115.

◆ After the RMAN restore script has completed successfully, type the commands at the operating system command line.

The appropriate Oracle backup and recovery documentation provides more information on Oracle data recovery procedures.



Disaster Recovery 123

5


◆ About disaster recovery ............................................................................................. 124◆ Preparing an Oracle database for disaster recovery .............................................. 124◆ Performing a disaster recovery ................................................................................. 126◆ Sample postcommand script ..................................................................................... 128

Disaster Recovery


Disaster Recovery

About disaster recoveryIt is important to develop a plan for recovering from a disaster on the computer system. Back up important data on a daily basis. To prepare for a disk crash or loss of data, develop and test a plan for recovering data.

You must determine the required frequency of backups. Consider that backup frequency is a trade-off between the time spent backing up data, and the time spent later recovering a database after a crash.

The following sections describe how to prepare for disaster recovery on an Oracle Server or NetWorker server host, and how to perform an Oracle disaster recovery to a new host.

This disaster recovery information pertains to single-instance databases only. The information does not pertain to disaster recovery for:

◆ ASM environments (for example, requiring backups of ASM metadata)

◆ PowerSnap Module environments

◆ RAC environments

Preparing an Oracle database for disaster recoveryTo prepare an Oracle database environment for disaster recovery, you must back up the following minimum list of files:

◆ Oracle database (all the datafiles)


◆ Control file

◆ Initialization parameter files, including one or both of the following:

• PFILE (user-managed parameter file)

• SPFILE (server-managed parameter file)

◆ Network files, including listener.ora, sqlnet.ora, tnsnames.ora

◆ Text file containing the Oracle DBID

◆ Password file, in the following location by default: $ORACLE_HOME/dbs/orapw$ORACLE_SID

◆ Registry files:: On UNIX, oratab is typically in /var/opt/oracle or /etc

◆ Recovery Catalog, if applicable

◆ RMAN scripts, if applicable

The Oracle documentation provides an exhaustive list of the files (other than the Oracle database) that should be backed up.

Follow these guidelines to facilitate disaster recovery:

◆ Institute mirrored control files.

◆ Refer to Oracle documentation for recommendations on whether to institute mirrored online redo logs.

◆ Back up the archived redo logs frequently between database backups.

◆ Back up the Recovery Catalog after every target database backup.

Preparing an Oracle database for disaster recovery 125

Disaster Recovery

◆ After manual Oracle backups, perform regular backups of the NetWorker server bootstrap and Oracle client file index by using the procedure described in “NetWorker server bootstrap backup” on page 100.

To back up the required files in preparation for disaster recovery, perform the following:

1. “Create the DBID text file” on page 125

2. “Set up a postcommand script for backup of Oracle-related files” on page 125

3. “Set up RMAN backups of the database and related files” on page 126

4. “Set up RMAN backups of Recovery Catalog” on page 126

Create the DBID text file

The Oracle DBID is an internal Oracle ID that helps Oracle find the autobackup of the SPFILE, if the Recovery Catalog is not accessible.

Before the Oracle DBID can be backed up, you must manually record the DBID in a text file. The simplest way to find the DBID of an Oracle database is to connect to the database through RMAN once the database has been mounted.

Once you have recorded the DBID in a text file, you can store the text file containing the DBID in any directory where you have the proper operating system permissions. You can use a postcommand script to back up the DBID text file, as described in “Set up a postcommand script for backup of Oracle-related files” on page 125.

In the sample postcommand script provided with the NSR-DBM-ORA software, the DBID text file is assumed to be dbid.txt, located in the $ORACLE_HOME directory. “Sample postcommand script” on page 128 provides details on the sample postcommand script.

Set up a postcommand script for backup of Oracle-related files

You can use a postcommand script to back up the files that Oracle RMAN does not back up, such as the following files:

◆ Initialization parameter file PFILE (user-managed parameter file)

◆ Network files, including listener.ora, sqlnet.ora, tnsnames.ora

◆ Text file containing the Oracle DBID, as described in “Create the DBID text file” on page 125

◆ Password file in the following location by default: $ORACLE_HOME/dbs/orapw$ORACLE_SID

◆ Registry files: on UNIX, oratab is typically in /var/opt/oracle or /etc

◆ RMAN scripts, if applicable

You can either create a postcommand script from scratch, or modify the postcommand script that is provided with the NSR-DBM-ORA software.

On UNIX, you can use any name for the postcommand script.

Note: During a scheduled NSR-DBM-ORA backup, the nsrnmostart process passes the options -s server_name -g group_name to the postcommand script if the script name begins with nsrnmodr.


Disaster Recovery

In a scheduled RMAN backup, include the postcommand script by one of the following methods:

◆ If you use the NSR-DBM-ORA wizard to configure the RMAN backup, specify the postcommand script in the wizard.

◆ If you use the legacy method (without the wizard) to configure the RMAN backup, set the POSTCMD parameter in the nsrnmo script.

“Sample postcommand script” on page 128 provides details on the postcommand script that is provided with the NSR-DBM-ORA software.

Set up RMAN backups of the database and related files

Set up an RMAN backup with NSR-DBM-ORA to back up the following files:

◆ Oracle database (all the datafiles)


◆ Control file

◆ Initialization parameter file SPFILE (server-managed parameter file)

Follow the instructions in the preceding chapters of this guide to properly configure and run the RMAN backup with NSR-DBM-ORA.

For example, to include the control file and SPFILE in the backup, you can add the following commands to the RMAN backup script:

◆ backup current control file

◆ backup spfile

The RMAN documentation provides details on RMAN commands and scripts.

If you want to back up PFILE (user-managed parameter file) or other files that Oracle RMAN does not back up, you can use a postcommand script. “Set up a postcommand script for backup of Oracle-related files” on page 125 provides details on setting up the postcommand script.

Set up RMAN backups of Recovery Catalog

Set up an RMAN backup of the Recovery Catalog by using the same method as for the target database backup, as described in “Set up RMAN backups of the database and related files” on page 126.

Oracle documentation provides more information on setting up and running Recovery Catalog backups.

Performing a disaster recoveryThe following sections provide guidelines for different types of disaster recovery operations:

◆ “NetWorker server recovery” on page 127

◆ “Oracle disaster recovery to a new host” on page 127

Performing a disaster recovery 127

Disaster Recovery

NetWorker server recovery

NetWorker software can be used to recover from different types of disasters on the NetWorker server. The degree of data loss during a disaster can range from one or more files lost when a disk crashes, to an entire computer system. The degree of severity of the disaster determines the procedures that must be performed to recover data on the NetWorker server.

With respect to NetWorker backups, recall that the bootstrap is a critical file, backed up only after scheduled Oracle backups. If only manual Oracle backups are performed, back up the bootstrap and client index manually. “NetWorker server bootstrap backup” on page 100 provides more information.

Along with the bootstrap information, keep accurate records of the network and system configurations, and maintain all the original software in a safe location. For a comprehensive disaster recovery, the following items are required:

◆ Original operating system media and patches

◆ Original NetWorker media

◆ Device drivers and media device names

◆ File system configuration

◆ IP addresses and hostnames

◆ Bootstrap information

The NetWorker Disaster Recovery Guide provides information on how to recover the NetWorker server from various types of disasters. “Related documentation” on page 9 describes how to access the guide.

Oracle disaster recovery to a new host

To perform an Oracle disaster recovery to a new host:

1. Install the Oracle software on the new host.

2. Install NetWorker client and NSR-DBM-ORA software on the new host, and create a Client resource for the new host.

3. Ensure that the user performing the recovery on the new host is listed in the Remote Access attribute in the Client resource of the original host. (This is required for “directed recovery” in NetWorker.)

4. To recover Oracle files that were backed up through a postcommand script, use either the NetWorker User GUI or the recover command.

For example, a typical recover command is as follows:

recover –s NetWorker_server –c client_name_of_original_host–d /var/opt/oracle –a /var/opt/oracle/oratab

5. To perform the remainder of the disaster recovery, follow the instructions in the Oracle Database Backup and Recovery User’s Guide. In the RMAN script, set the NSR_CLIENT parameter to the name of the original host.


Disaster Recovery

Sample postcommand scriptYou can use a postcommand script to back up files that Oracle RMAN does not back up, as described in “Set up a postcommand script for backup of Oracle-related files” on page 125. The NSR-DBM-ORA software includes a sample postcommand script that you can modify for your environment.

The NSR-DBM-ORA installation provides a sample postcommand script that is specific to UNIX, depending on the platform where NSR-DBM-ORA is installed. The sample script is installed in the bin subdirectory under the NetWorker software directory, for example, under /opt/nsr.

You must customize the settings in the sample postcommand script for the specific environment. At a minimum, you must set the ORACLE_HOME and ORACLE_SID parameters in the script. If these two parameters are not set, the postcommand script fails at runtime.

Note: During a scheduled NSR-DBM-ORA backup, the nsrnmostart process passes the options -s server_name -g group_name to the postcommand script if the script name begins with nsrnmodr.

View the sample postcommand script provided with NSR-DBM-ORA as follows in Example 21 on page 128.

Example 21 Sample postcommand script on UNIX

The sample postcommand script named nsrnmodrpostcmd is installed with the NSR-DBM-ORA software on UNIX.

To use this script for backups in your environment, you must customize the script. At a minimum, set the ORACLE_HOME and ORACLE_SID parameters in the script.

The nsrnmodrpostcmd script provided on UNIX is as follows:

##!/bin/ksh#ORACLE_HOME=ORACLE_SID=GRP=noSRV=nocomplete=0shiftshiftwhile [ "$#" -gt "0" ]do if [ "$1" = "-g" ]; then GRP=$2 if [ "$SRV" != "no" ]; then complete=1 fi elif [ "$1" = "-s" ]; then SRV=$2 if [ "$GRP" != "no" ]; then complete=1 fi fi shiftdone

if [ $complete -eq 1 ]; then save -s $SRV -g $GRP $ORACLE_HOME/network/admin save -s $SRV -g $GRP $ORACLE_HOME/dbs/orapw$ORACLE_SID

Sample postcommand script 129

Disaster Recovery

save -s $SRV -g $GRP /var/opt/oracle/oratab save -s $SRV -g $GRP $ORACLE_HOME/dbid.txtfi


Disaster Recovery

Cluster and RAC Systems 131

6


◆ Considerations for cluster and RAC systems ......................................................... 132◆ Cluster systems............................................................................................................ 133◆ RAC systems ................................................................................................................ 135

Cluster and RACSystems


Cluster and RAC Systems

Considerations for cluster and RAC systemsTo prepare for successful NSR-DBM-ORA operations on cluster and RAC systems, consider the following:

◆ “Parameters in the RMAN script or session” on page 132

◆ “NetWorker software configuration” on page 133

◆ “Backup failover” on page 133

Parameters in the RMAN script or session

The parameters NSR_CLIENT and NSR_SERVER must be correctly set in the RMAN script or session. Appendix A, “Parameters in an RMAN Session,” provides more information on how to set the required parameters.

NSR_CLIENT

During an Oracle backup, the NetWorker server creates entries about the backed-up data in the online client file index. During an Oracle restore, the data is retrieved by first searching this client file index.

The parameter NSR_CLIENT that is set in the RMAN script provides the following information to the NetWorker server:

◆ During a backup, the name of the NetWorker client whose index file should be used to record the backup information.

◆ During a restore, the name of the NetWorker client whose index file should be used to search for the save set to be restored.

If NSR_CLIENT is not set, the NetWorker server uses the name of the local physical host. Set the parameter NSR_CLIENT to the same value for all channels allocated during the backup.

The value of NSR_CLIENT (either the default value or an explicitly defined value) used for a backup should be the same as the value of NSR_CLIENT used for the restore of that backup. Setting NSR_CLIENT to the name of the virtual (as opposed to the physical) host might be preferable, so the backup and restore scripts run successfully unmodified, regardless of the physical host that executes the scripts.

IMPORTANT!If NSR_CLIENT is set to any value other than the default value during a backup or restore, the Remote Access attribute must be modified in that NetWorker Client resource. The NetWorker Administrator’s Guide provides more information.

For each Oracle backup and restore, set NSR_CLIENT by using either the rman send command on the operating system command line or the send command in the RMAN script.

Note: Setting NSR_CLIENT in the system environment has no effect.

Cluster systems 133


NSR_SERVER

Set the parameter NSR_SERVER for a manual Oracle backup or restore if the NetWorker server is a remote system for the node running the RMAN session. The parameter must be set to the server name for each channel.

Note: The parameter NSR_SERVER does not need to be set for a scheduled Oracle backup.

NetWorker software configuration

Ensure the correct NetWorker software configuration by performing the following tasks:

1. Ensure that the system is set up properly:

• In the NetWorker Client resource of each host, the Save Set attribute is set to:

RMAN:RMAN_script_pathname

• In the directory that contains the savefs program file, there is an empty file named pathownerignore. Create the pathownerignore file if it does not exist.

The NetWorker Release Notice provides more information on the default directory for the savefs program.

2. Configure a Client resource on the NetWorker server for each RAC node that will be used to run backups and restores. In each Client resource, use the fully qualified IP name as the client name.

“Configure a Client resource with NMC” on page 81 provides information on how to configure a Client resource for NSR-DBM-ORA backups.

3. In the Remote Access attribute in the Client resource for a virtual host, specify the name of each physical host that can store and retrieve the backups.

Backup failover

Neither RMAN nor NSR-DBM-ORA supports Transparent Application Failover (TAF). As a result, if a failure occurs during a backup, the backup is not automatically restarted from the point of failure on another node. Only connect-time failover is supported. “Connect-time failover” on page 138 provides more information.

Cluster systemsA cluster system typically includes multiple nodes connected by a shared SCSI bus to which common storage is attached. In a cluster system, cluster services such as disk services can be defined and assigned their own IP addresses and names (virtual hosts). The services and their associated storage can migrate for failover between the physical nodes in the cluster.

Together, the NSR-DBM-ORA and NetWorker server software can back up and restore an Oracle database configured on cluster disk services. The NetWorker server treats each cluster service as an independent client and stores the associated backup entries in the online indexes under the name of the service.

After properly configuring a cluster service as a NetWorker client, NSR-DBM-ORA can be used with the NetWorker server to back up and restore the database associated with the service, independent of the actual node that provides the service.



To configure a cluster system for Oracle backup and restore operations that use NSR-DBM-ORA, follow the “Roadmap for backup/restore configuration in a cluster system” on page 134.

Roadmap for backup/restore configuration in a cluster system

The Oracle Server and NetWorker server systems must be configured before NSR-DBM-ORA can be used with the NetWorker server for Oracle backup and restore operations in a cluster system.

Before configuring Oracle backup and restore operations in a cluster system, ensure the following:

◆ The NSR-DBM-ORA software is properly installed according to the instructions in the NSR-DBM-ORA Release Notice.

◆ The Oracle Server and NetWorker server system configurations are completed according to the instructions in the “Configuration roadmap” on page 58.

To properly configure Oracle backup and restore operations with NSR-DBM-ORA in a cluster system:

1. Configure the storage for the database as a cluster disk service.

The appropriate cluster administration guide from the particular cluster software vendor provides more information.

2. Configure a Client resource on the NetWorker server for the virtual host and each physical host that will run backups and restores.

3. Create an RMAN script that includes the appropriate NSR_CLIENT parameter settings. “NSR_CLIENT” on page 132 provides more information.

Backup failover

When a node failure occurs during a manual (unscheduled) Oracle backup, a database administrator (DBA) must restart the backup after instance recovery. Since the Oracle Net services are configured with a cluster service name, the restarted backup is executed on the physical node that takes control of the cluster service.

When a node failure occurs during a scheduled Oracle backup, the NetWorker server restarts the backup (from the beginning, not from the point of failure) if the Client Retries attribute in the NetWorker Group resource is set to a value greater than zero. The restarted backup is executed on the node that takes control of the cluster service.

RAC systems 135


RAC systemsThe NSR-DBM-ORA software supports backups and restores of RAC systems for parallelism and high availability.

RAC terminology

A node in a RAC system is a physical computer with a hostname such as node1.fujitsu.com . An Oracle instance is a memory structure and a group of Oracle Server processes running on a node.

An Oracle database (for example, named databs1) comprises a set of datafiles, which are used by the Oracle instances and can be shared between the nodes. All instances share the same datafiles and control file. Each node must have its own set of redo log files and its own archived redo logs.

RAC backups and restores

After proper configuration of RAC and the associated cluster system, NSR-DBM-ORA enables Oracle backups on either a single node or several nodes of the RAC system.

A parallel Oracle backup uses Oracle instances running in parallel on multiple nodes of the cluster. In the RMAN backup script created for running a parallel Oracle backup, allocate multiple channels for the backup and specify that each channel run on a specific node. The parameter NSR_CLIENT must be set to the same value for each channel. “NSR_CLIENT” on page 132 provides more information on setting the parameter.

NSR-DBM-ORA software enables restores of the Oracle data to any physical node in the cluster, regardless of which physical node originally performed the backup.

To enable Oracle backup and restore operations, follow the configuration steps in “Roadmap for backup/restore configuration in a RAC system” on page 135.

Roadmap for backup/restore configuration in a RAC system

To properly configure Oracle backup and restore operations with NSR-DBM-ORA in a RAC system:

1. Install the proper cluster management software on each cluster node.

The appropriate cluster installation documentation from the particular cluster software vendor provides more information.

2. Configure the cluster for use with RAC.

The appropriate RAC documentation from Oracle Corporation provides more information.

3. Install and configure the RAC software.

The required patches from Oracle might need to be installed, to complete the RAC installation and linking procedures. Configure the Oracle Net services to allow connect-time (SQL Net) failover. “Connect-time failover” on page 138 provides guidelines. The appropriate Oracle documentation provides more information.



4. Install NSR-DBM-ORA on each node of the cluster to be used for backup and recover operations.

The NSR-DBM-ORA Release Notice provides more information.

5. Configure a Client resource on the NetWorker server for the virtual host and each physical host that will run backups and restores.

6. Create the appropriate RMAN backup and restore scripts for the preferred types of Oracle backups and restores on the RAC system. The following sections provide more information:

• “Creating RMAN backup scripts” on page 140

• “Creating RMAN restore scripts” on page 141

7. Review the additional issues concerning Oracle recover operations in “Archived redo logs” on page 141.

8. To set up a local storage node for each RAC node involved in a backup, follow the instructions in “Setting up RAC nodes to back up to a local storage node” on page 136.

Setting up RAC nodes to back up to a local storage node

To set up RAC nodes to back up to a local storage node:

1. Ensure that the NetWorker storage node software is installed on each RAC node to be used for the NSR-DBM-ORA backup.

2. On the NetWorker server, create a NetWorker Storage Node resource for each RAC node to be used for the NSR-DBM-ORA backup. The NetWorker Administrator’s Guide provides details on storage node configuration.

3. Create a NetWorker Device resource for the device on each RAC node to be used for the NSR-DBM-ORA backup. The NetWorker Administrator’s Guide provides details on device resource configuration.

4. Ensure that Groups and the selection criteria (such as Clients) of the media pool used for the devices match the settings in the NSR-DBM-ORA backup configuration.

5. Label and mount a NetWorker volume on each storage node.

6. Select one of the RAC nodes to store the NetWorker indexes for the NSR-DBM-ORA backup and to initiate the backup.

7. For the RAC node that will initiate the NSR-DBM-ORA backup, create a NetWorker Client resource with the attribute settings required for the backup, as described in “NetWorker software configuration” on page 133:

• The Remote Access attribute must include the hostnames of all the other RAC nodes.

• The Storage Nodes attribute must contain curphyhost, followed by nsrserverhost. The Storage Nodes attribute must be set to the following:

curphyhostnsrserverhost

8. For each of the other RAC nodes that will not initiate the NSR-DBM-ORA backup, create a basic Client resource. (Settings in these other Client resources do not affect the NSR-DBM-ORA backup.)

RAC systems 137


9. On the RAC node that will initiate the NSR-DBM-ORA backup, create the required nsrnmo script and RMAN backup script. The RMAN script must include the NSR_CLIENT setting, as described in “NSR_CLIENT” on page 132.

Example 22 on page 137 shows how to set up three RAC nodes as storage nodes for NSR-DBM-ORA backups.

Example 22 Setting up RAC nodes as storage nodes

A RAC system contains three nodes named A, B, and C. Each node has a Linux operating system, and an attached tape drive to be used for NSR-DBM-ORA backups. NetWorker storage node software is installed on each node.

In the NMC interface, a Storage Node resource is created for each node by right-clicking Storage Nodes in the Devices pane and selecting New.

After the Storage Node resources are created, a Device resource is created for each tape drive. Each Device resource is created in NMC by right-clicking Devices in the Devices pane and selecting New. Since the tape devices are attached to storage nodes, the device names must have the format rd=host_name:device_name. For example:

◆ Tape device /dev/rmt/tape0 is attached to node A. In the Device resource, the device name is rd=A:/dev/rmt/tape0.

◆ Tape device /dev/rmt/tape3 is attached to node B. In the Device resource, the device name is rd=B:/dev/rmt/tape3.

◆ Tape device /dev/rmt/tape1 is attached to node C. In the Device resource, the device name is rd=C:/dev/rmt/tape1.

In the tape device on each node, a volume is labeled and mounted. All of the volumes are assigned to the Default pool in this example.

Node A is selected to store the index entries for the NSR-DBM-ORA backups and initiate the backups. (The choice of node A was arbitrary; node B or node C could have been chosen instead.) In all the RMAN backup and restores scripts, NSR_CLIENT must be set to the hostname of node A.

In the NetWorker Client resource for node A:

◆ The Remote Access attribute is set to the hostnames of nodes B and C.

◆ The Storage Nodes attribute is set to:

curphyhostnsrserverhost

◆ The remaining attributes are set, as required. For example:

• The Backup Command attribute is set to the nsrnmo script name.

• The Group attribute is set to the backup group name.

• The Save Set attribute is set to the RMAN script pathname.

The following RMAN script uses all three nodes to perform the backup. Each node backs up data to its local tape drive:

connect target sys/oracle@connect_identifier;run {

allocate channel t1 type ’SBT_TAPE’connect ’sys/oracle@Net_service_name_of_instance_A’;allocate channel t2 type ’SBT_TAPE’connect ’sys/oracle@Net_service_name_of_instance_B’;allocate channel t3 type ’SBT_TAPE’connect ’sys/oracle@Net_service_name_of_instance_C’;



send channel t1 ’NSR_ENV=(NSR_CLIENT=A)’;send channel t2 ’NSR_ENV=(NSR_CLIENT=A)’;send channel t3 ’NSR_ENV=(NSR_CLIENT=A)’;backup database;release channel t1;release channel t2;release channel t3;

}

To enable restores, NSR_CLIENT must be set to the hostname of node A. For example, the following RMAN script restores the database. The script can be run on any host:

connect target sys/oracle@connect_identifier;run {

allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;send channel t1 ’NSR_ENV=(NSR_CLIENT=A)’;send channel t2 ’NSR_ENV=(NSR_CLIENT=A)’;restore database;release channel t1;release channel t2;

}

Connect-time failover

If multiple listeners support a single service, a connect-time failover reroutes the connection request to another listener if the first listener is down or cannot make the connection. To enable the connect-time failover in RAC, there must be a listener on each node, and each instance must use the same Net service name.

When using the local Net service naming method, the client’s tnsnames.ora file should include the following parameters:

o92pA.fujitsu.com =(DESCRIPTION =

(ADDRESS_LIST =(FAILOVER = ON)(ADDRESS = (PROTOCOL = tcp) (HOST = nodeA) (PORT = 1521))(ADDRESS = (PROTOCOL = tcp) (HOST = nodeB) (PORT = 1521))

)(CONNECT DATA =

(SERVICE_NAME = proddb))

)

◆ Set the FAILOVER parameter to ON. The default value is ON for an ADDRESS_LIST, and OFF when ADDRESS_LIST is not specified.

Note: FAILOVER was introduced in Oracle8i.

◆ Include the ADDRESS_LIST parameter:

• If multiple addresses are specified, but the ADDRESS_LIST parameter is omitted, the Oracle Net service reads the addresses sequentially and attempts to connect to the last one only.

• If the ADDRESS_LIST parameter is specified, the addresses are tried in the order they appear in the list.

RAC systems 139


◆ In the CONNECT_DATA section, use SERVICE_NAME instead of the system identifer (SID). SERVICE_NAME should be different from SID.

Note: SERVICE_NAME was introduced in Oracle8i.

◆ GLOBAL_DBNAME should not appear in the SID_LIST_LISTENER parameter of the listener.ora file since it disables the failover.

When a node or listener to which a client tries to connect is not available, the next listener on the list is contacted. When the instance is down but the listener is running, the failover occurs only if the instance is configured to dynamically register with the listener.

Dynamic instance registration

Dynamic instance registration was introduced in Oracle8i. During dynamic instance registration, the database registers itself with the Oracle listener on startup and unregisters itself on shutdown.

To perform dynamic instance registration:

◆ Set the INSTANCE_NAME and SERVICE_NAME parameters in the initialization file (initoracle_sid.ora). There can be several services for a single instance.

◆ If the listener does not listen on the default port (1521), set the LOCAL_LISTENER parameter in the initialization file.

◆ The SID_LIST_LISTENER parameter in listener.ora must not include SID_DESC for the RAC instances.

◆ It is not necessary to have the listener.ora file when the listener listens on the default port.

When the instance is down, the listener does not know how to connect to it. As a result, the listener tries the next connect option specified in the ADDRESS_LIST in the tnsnames.ora file.

Note: Some applications such as Oracle Enterprise Manager still require static database registration with a listener.

Static instance registration

With static registration, the information about the instance is manually configured in the listener.ora file through SID_DES in the SID_LIST_LISTENER parameter. The listener contains continuous information about the instance, even if the instance is down. As a result, when the instance is down, the listener still tries to connect to it rather than try the next connect option specified in the tnsnames.ora file.

Backup failover

When a manual backup is interrupted by an Oracle server-side failure, a DBA must restart the backup after instance recovery. During the restarted backup, the connection request to the failed instance is rerouted to another instance according to the connect-time failover setup in the tnsnames.ora file.

For a scheduled backup, when the first backup attempt returns an error, the NetWorker server restarts the backup if the Client Retries attribute in the NetWorker Group resource is set to a value greater than zero. The connect-time failover reroutes the connection to an available instance, and the restarted backup starts from the beginning. For example, if the backup fails 5 hours into a 10-hour backup, it takes 15 hours to complete the backup. In this case, the operator might elect to wait until the next scheduled backup.



On a RAC system, traditional cluster failover is not available. If an instance or node fails in RAC, another node detects the failure and recovers the failed node’s data. As a result, the nodes in RAC carry on without the failed node. If a system failure occurs on the RAC node used to initiate an NSR-DBM-ORA backup, the backup fails. In this case, manual intervention is required to configure and restart the backup on a different RAC node that is available:

1. On the available RAC node, ensure that the following software is installed:

• NetWorker client

• NetWorker storage node (optional)

• NSR-DBM-ORA

2. Configure the Client resource for the available RAC node.

3. Replace the original Client resource with the new Client resource from step 2 for the NSR-DBM-ORA backup.

Creating RMAN backup scripts

A single RMAN backup script can be used to run a parallel Oracle backup with NSR-DBM-ORA on a RAC system. In the backup script, allocate multiple channels for the backup and specify that each channel run on a specific node.

Example 23 RMAN script for a manual Oracle backup on a RAC system

Suppose a RAC system consists of two nodes named node1.fujitsu.com and node2.fujitsu.com . The Oracle instances named instance1 and instance2 are running on node1.fujitsu.com and node2.fujitsu.com , respectively. The NetWorker server is located on a separate node, server1.fujitsu.com .

The following RMAN script for a manual backup is intended to run on node1.fujitsu.com by using the NOCATALOG mode of RMAN. The script sets NSR_CLIENT to node1.fujitsu.com and NSR_SERVER to server1.fujitsu.com . As a result, the NetWorker server stores the backup information in the node1.fujitsu.com client file index, as described in “NSR_CLIENT” on page 132. Two channels are allocated to each of the nodes, node1.fujitsu.com and node2.fujitsu.com :

run {allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;allocate channel t3 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_node2’;allocate channel t4 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_node2’;send ’NSR_ENV=(NSR_CLIENT=node1.fujitsu.com , NSR_SERVER=server1.fujitsu.com );backup filesperset 1format ’instance1_%s_%p’(database);release channel t1;release channel t2;release channel t3;release channel t4;

}

“Backing up all archived logs from each node” on page 142 provides a sample script to back up all the archive log files in a RAC system.

RAC systems 141


Creating RMAN restore scripts

A single RMAN restore script can be used to run a parallel Oracle restore with NSR-DBM-ORA on a RAC system. In the restore script, allocate multiple channels for the restore and specify that each channel run on a specific node.

Note: NSR-DBM-ORA does not support multiple RMAN restores that are running at the same time.

To run an Oracle restore on a RAC system, none of the nodes can be open. Only the node that is running the RMAN restore script needs to be mounted.

Example 24 RMAN script for an Oracle restore on a RAC system

Refer to Example 23 on page 140. A RAC system consists of two nodes named node1.fujitsu.com and node2.fujitsu.com . The Oracle instances named instance1 and instance2 are running on node1.fujitsu.com and node2.fujitsu.com , respectively. The NetWorker server is located on a separate node, server1.fujitsu.com .

The following RMAN restore script is to be run on node2.fujitsu.com . The script restores the backup that was created by the backup script in Example 23 on page 140. This restore script sets NSR_CLIENT to node1.fujitsu.com and NSR_SERVER to the remote NetWorker server name. As a result, the NetWorker server will obtain the backup information from the node1.fujitsu.com client file index. Two channels are allocated to each of the nodes, node1.fujitsu.com and node2.fujitsu.com :

run {allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;allocate channel t3 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_node1’;allocate channel t4 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_node1’;send ’NSR_ENV=(NSR_CLIENT=node1.fujitsu.com , NSR_SERVER=server1.fujitsu.com )’;restore database;release channel t1;release channel t2;release channel t3;release channel t4;

}

In this example, the parallel Oracle backup was performed with NSR_CLIENT set to node1.fujitsu.com for each channel in the RMAN backup script. In order to restore the backup data to node2.fujitsu.com , add node2.fujitsu.com to the Remote Access attribute in the NetWorker Client resource for node1.fujitsu.com .

The NetWorker Administrator’s Guide provides more information on how to modify the Client resource.

“Restoring all archived logs from each node” on page 142 provides a sample script to restore all the archive log files in a RAC system.

Archived redo logs

Each node in a RAC system maintains a separate set of redo logs. Redo logs that become full are archived on the local node. As a result, the archived redo logs are divided among the nodes of the system.



To enable RMAN to back up and recover a RAC system, make all the archived redo log files accessible by all nodes participating in the backup or recovery. The appropriate Oracle RAC documentation provides information on how to share the archived redo logs.

The following sections provide sample scripts to back up and restore all the archived redo log files in a RAC system:

◆ “Backing up all archived logs from each node” on page 142

◆ “Restoring all archived logs from each node” on page 142

Backing up all archived logs from each node

All the archived log files in a RAC system can be backed up from a single node (for example, a node named ops1.fujitsu.com ) by using the following type of RMAN script:

run {allocate channel t1 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_ops1’;allocate channel t2 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_ops2’;send ’NSR_ENV=(NSR_CLIENT=ops1.fujitsu.com )’;backup filesperset 10(archivelog all delete input format ’al_%s_%p’);release channel t1;release channel t2;

}

Restoring all archived logs from each node

All the archived log files in a RAC system can be restored from a single node (for example, a node named ops1.fujitsu.com ) by using the following type of RMAN script:

run {allocate channel t1 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_ops1’;allocate channel t2 type ’SBT_TAPE’connect ’user_name/user_passwd@connect_string_of_ops2’;send ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com , NSR_CLIENT=ops1.fujitsu.com )’;restore (archive log all);release t1;release t2;

}

Proxy Backups and Restores 143

7


◆ Overview of proxy backups and restores ................................................................ 144◆ Proxy backup and restore system ............................................................................. 146◆ Proxy backup processes ............................................................................................. 147◆ Proxy restore processes .............................................................................................. 153◆ Software configuration roadmap.............................................................................. 156◆ Proxy backup requirements....................................................................................... 164◆ Proxy backup information in the NetWorker indexes........................................... 167◆ Proxy restore requirements........................................................................................ 170◆ Catalog synchronization for proxy backups ........................................................... 174◆ Proxy backups and restores on cluster systems...................................................... 185

Proxy Backups andRestores


Proxy Backups and Restores

Overview of proxy backups and restoresProxy backups and restores provide continuous snapshot-based protection and availability of Oracle data on specific types of primary storage. The NSR-DBM-ORA software supports proxy backups and restores of Oracle data with the following requirements:

◆ The proxy backups create snapshots of Oracle data that reside on primary storage devices supported by the PowerSnap Modules that work with the NSR-DBM-ORA.

◆ The proxy operations use the particular PowerSnap Module software designed for the primary storage.

The following sources provide details on the installation requirements for proxy operations:


◆ NetWorker PowerSnap Module documentation(Refer to the PowerSnap Module version for the primary storage system.)

Types of proxy backups

NSR-DBM-ORA supports the following types of proxy backups, in cooperation with the appropriate PowerSnap Module:

◆ “Instant backups” on page 144

◆ “Live backups” on page 144

Instant backups

An instant backup creates a point-in-time copy (snapshot) of Oracle data and stores this copy on the primary storage system. Instant backups can be scheduled to occur many times in a single day, with little impact to the Oracle Server or network.

The group of Oracle datafiles in the point-in-time copy is referred to as a snap set. A snapshot policy must be configured to control the lifecycle of the snap set. This policy specifies the frequency of instant backups and how long snapshots are retained before being recycled. “Software configuration roadmap” on page 156 provides more information.

Note: NSR-DBM-ORA supports scheduled instant backups only. Manual (unscheduled) instant backups are not supported.

IMPORTANT!Instant backups protect against logical failures only. To protect against physical failures, the point-in-time copy must be backed up to secondary storage.

Live backups

NSR-DBM-ORA supports the following two types of live backup, which back up a snapshot to secondary (traditional) storage:

◆ Deferred live backup — An existing point-in-time copy (snapshot), that is created during an instant backup, is backed up to secondary storage, such as tape. The snapshot is retained on the primary storage.

Overview of proxy backups and restores 145


◆ Immediate live backup — A point-in-time copy (snapshot) is created, and immediately backed up to secondary storage, such as tape. The snapshot is then automatically deleted from the primary storage.

A snapshot policy must be configured to enable live backups. “Software configuration roadmap” on page 156 provides more information.

FSC-Kommentar: Abschnitt „Important“ nach oben verschoben

A proxy client host that is separate from the Oracle Server host can be used to move the point-in-time copy to the traditional storage medium, such as tape. Using a proxy client as a secondary host reduces the impact on the Oracle Server.

Note: The proxy client can be a NetWorker storage node.

Types of proxy restores

NSR-DBM-ORA supports the following types of proxy restores, in cooperation with the appropriate PowerSnap Module:

◆ “Instant restores” on page 145

◆ “Rollbacks” on page 145

◆ “Restores from secondary storage” on page 145

Instant restores

During an instant restore, the saved data is retrieved from a mounted point-in-time copy that was created with an instant backup. A minimal amount of time is required for this type of restore.

Rollbacks

A rollback restores a whole point-in-time copy to a source location by using the hardware’s particular capabilities. Rollbacks are destructive restores because they overwrite the entire contents of a snapshot unit, such as a volume or disk. Use the NetWorker PowerSnap Module documentation to determine if a rollback is supported on a specific type of hardware.

Restores from secondary storage

If the point-in-time copy was backed up to secondary storage through a live backup, the PowerSnap Module software uses the NetWorker recover program to retrieve the data.



Proxy backup and restore systemFor proxy backups and restores, the NSR-DBM-ORA, NetWorker server and client, and appropriate PowerSnap Module software function with the components of the Oracle backup and restore system. The implementation of a proxy backup and restore strategy requires a knowledge of how the Oracle and NetWorker software components work together.

The Oracle backup and restore system consists of the following:

◆ Oracle Server

◆ Recovery Manager (RMAN)

◆ Recovery Catalog (optional)

◆ Oracle Enterprise Manager (OEM) Backup Management Tools (optional)

The following sections provide more information on RMAN and the Recovery Catalog:

◆ “Oracle Recovery Manager” on page 47

◆ “Oracle Recovery Catalog” on page 47

The NetWorker software components required for proxy backups are the following:

◆ NetWorker server

◆ NetWorker client (installed on the Oracle Server and proxy client host)

◆ NSR-DBM-ORA (installed on the Oracle Server)

◆ NetWorker PowerSnap Module (installed on the Oracle Server and proxy client host)

Note: NetWorker storage node software should be installed on the proxy client host.

Figure 11 on page 147 illustrates the architecture of the proxy backup and restore system. It provides an overview of the functional relationship between the Oracle and NetWorker software components. The interactions of the components are described in the following sections:

◆ “Proxy backup processes” on page 147

◆ “Proxy restore processes” on page 153

Proxy backup processes 147


Figure 11 Proxy backup and restore system

Proxy backup processesDuring a proxy backup of Oracle data, Oracle Server processes interact with the NetWorker server through the NSR-DBM-ORA and PowerSnap Module software. The NSR-DBM-ORA media management library (MML) is integrated with the Oracle Server kernel software.

Note: Certain RMAN features, such as checking for corrupt blocks, are not applicable to proxy backups because the MML controls that process.

When a proxy backup is initiated, the RMAN utility starts the Oracle Server processes that call the MML proxy-specific routines. The MML routines communicate with the PowerSnap Module service.

On the Oracle Server host, the PowerSnap Module uses a storage platform-specific application programming interface (API) to take a snapshot of the Oracle data on the primary storage.

RMAN NetWorker Client

Oracle Server Host PowerSnap Proxy Client (NetWorker Storage Node)

NetWorker Server

Oracle Server

NSR-DBM-ORA

MML NetWorker Client

PowerSnap Module

Oracle Data Files

Point-in-Time Copy of Oracle Data Files

PowerSnap Module

Primary Storage

Recovery Catalog

Storage Medium

NetWorker Server

GEN-000174



If a live backup is performed, the PowerSnap Module moves the Oracle data from the point-in-time copy (snapshot) on the primary storage to the NetWorker server or storage node. The NetWorker server or storage node then stores the Oracle data on the secondary storage, such as a tape.

At the end of the proxy backup, the NetWorker server updates the online client and media indexes with information about the backup. The NetWorker Administrator’s Guide provides more information about NetWorker server and NetWorker client programs and services.

Note: A proxy backup can be started only by automatic or manual invocation of the scheduled NetWorker backup group. A proxy backup cannot be scheduled through Oracle Enterprise Manager or started from RMAN. The NetWorker PowerSnap Module documentation provides information on how to manually invoke a scheduled backup.

Scheduled proxy backup

A scheduled backup is a backup of Oracle data that the NetWorker server initiates.

The following sections describe the process steps in the three types of scheduled proxy backup:

◆ “Scheduled instant backup” on page 148

◆ “Scheduled immediate live backup” on page 150

◆ “Deferred live backup” on page 152

Note: To simplify the descriptions and illustrations in these sections, certain steps have been omitted, such as the step to obtain the proper NetWorker authorization.

Scheduled instant backup

A scheduled instant backup involves the following steps:


2. The savegrp program requests that the NetWorker client-side service, nsrexecd, on the Oracle Server host run the savefs program to verify the existence of the RMAN scripts to be used for the backup.

3. If the savefs program succeeds, the savegrp program starts the PowerSnap program, nsrsnap, on the Oracle Server host. Otherwise, the backup fails.

4. If an existing point-in-time copy needs to be removed (for example, it is expired according to the snapshot policy), the nsrsnap program invokes the nsrsnapck program to remove both the point-in-time copy from the storage system and the corresponding NetWorker index entries.

Note: If the point-in-time copy was not copied to secondary storage, the nsrsnapck program invokes the nsroraclecat program to remove the corresponding RMAN catalog entries. “Automatic catalog synchronization with the nsroraclecat program” on page 183 provides more information.

5. The nsrsnap program contacts the nsrexecd service, which starts the nsrnmo script for the scheduled backup.

6. The nsrnmo script sets the required parameters and invokes the NSR-DBM-ORA program, nsrnmostart.



7. The nsrnmostart program starts the RMAN utility to run the required RMAN backup script.

8. RMAN starts an Oracle process that calls the MML, which contacts the PowerSnap master program, nsrpsd, to perform the instant backup:

• PowerSnap processes create the point-in-time copy that comprises the backup.

• The nsrmmd service performs the following:

– Stores metadata (used only by the PowerSnap Module) about the point-in-time copy on the media belonging to the configured snapshot pool.

– Records tracking information about the metadata in the NetWorker media database by using the nsrmmdbd service.

• The nsrindexd service records information about the point-in-time metadata in the NetWorker client file index.

• The nsrindexd service records tracking information about each backup piece name in the NetWorker client file index. The client index contains one entry per datafile.

• The Oracle Server records the backup entry in the RMAN catalog.

9. After the RMAN script completes successfully, the nsrnmostart program backs up the NWORA resource file, described in “The NWORA resource file” on page 176.

10. If the Snapshot Policy resource specifies that the point-in-time copy must be copied to secondary storage, the nsrsnap program performs the steps in “Deferred live backup” on page 152.

Note: The point-in-time copy is not deleted from the primary storage.

11. The savegrp program backs up the NetWorker server bootstrap and Oracle client file indexes.

Figure 12 on page 150 shows how the Oracle Server and NetWorker processes interact during a scheduled instant backup.



Figure 12 Scheduled instant backup

Scheduled immediate live backup

A scheduled immediate live backup involves the following steps:

Note: The first seven steps are the same as those in “Scheduled instant backup” on page 148.


2. The savegrp program requests that the NetWorker client-side service, nsrexecd, on the Oracle Server host run the savefs program to verify the existence of the RMAN scripts to be used for the backup.

3. If the savefs program succeeds, the savegrp program starts the PowerSnap program, nsrsnap, on the Oracle Server host. Otherwise, the backup fails.

nsrnmo


PowerSnap Proxy Client (NetWorker Storage Node)

NetWorker Server

Snapshot Pool Media

nsrmmdbd

Media Database

Client File Index

nsroraclecat (optional)


Oracle process

NSR-DBM-ORA

MML

nsrindexd

nsrexecd nsrmmd


RMAN

savegrp

nsrsnap nsrd

Primary Storage Oracle

Data Files Point-in-Time Copy of Oracle Data Files

data

PowerSnap processes

PowerSnap processes

nsrpsd

nsrexecd nsrsnapck

savefs

nsrnmostart

GEN-000175

PowerSnap elements

NetWorker Module

elements

Start Here



4. If an existing point-in-time copy needs to be removed (for example, it is expired according to the snapshot policy), the nsrsnap program invokes the nsrsnapck program to remove both the point-in-time copy from the storage system and the corresponding NetWorker index entries.

Note: If the point-in-time copy was not copied to secondary storage, the nsrsnapck program invokes the nsroraclecat program to remove the corresponding RMAN catalog entries. “Automatic catalog synchronization with the nsroraclecat program” on page 183 provides more information.

5. The nsrsnap program contacts the nsrexecd service, which starts the nsrnmo script for the scheduled backup.

6. The nsrnmo script sets the required parameters and invokes the NSR-DBM-ORA program, nsrnmostart.

7. The nsrnmostart program starts the RMAN utility to run the required RMAN backup script.

8. RMAN starts Oracle processes that call the MML, which contacts the PowerSnap master program, nsrpsd, to perform the immediate live backup:

• PowerSnap processes create the point-in-time copy, move the data to secondary storage, and then delete the point-in-time copy from the primary storage.

• The nsrmmd service records the tracking information about the secondary storage save sets in the NetWorker media database by using the nsrmmdbd service.

• The nsrindexd service records the tracking information for each backup piece name and for each stored operating system datafile in the NetWorker client file index. The client index contains two entries per datafile.

• The Oracle Server records the backup entry in the RMAN catalog.

9. After the RMAN script completes successfully, the nsrnmostart program backs up the NWORA resource file, described in “The NWORA resource file” on page 176.

10. The savegrp program backs up the NetWorker server bootstrap and Oracle client file indexes.

Figure 13 on page 152 shows how the Oracle Server and NetWorker processes interact during a scheduled immediate live backup.



Figure 13 Scheduled immediate live backup

Deferred live backup

A deferred live backup runs automatically as part of a scheduled backup, as specified by the Backup Snapshots attribute of the Snapshot Policy resource. The NetWorker PowerSnap Module documentation provides more information.

A deferred live backup involves the following steps:

1. The nsrsnap or nsrsnapadmin program invokes the PowerSnap processes to copy the point-in-time copy to secondary storage.

Note: The point-in-time copy is not deleted from the primary storage.

2. The nsrmmd service records the tracking information about the deferred live backup in the NetWorker media database by using the nsrmmdbd service.

3. The nsrindexd service records the tracking information for each stored operating system datafile in the NetWorker client file index.

nsrnmo



NetWorker Server

Secondary Storage Medium

nsrmmdbd

Media Database

Client File Index

nsroraclecat (optional)


Oracle process

NSR-DBM-ORA

MML

nsrindexd

nsrexecd nsrmmd


RMAN

savegrp

nsrsnap nsrd



data

PowerSnap processes

PowerSnap processes

nsrpsd

nsrexecd nsrsnapck

savefs

nsrnmostart

GEN-000176

PowerSnap elements

NetWorker Module

elements

Start Here

Proxy restore processes 153


Figure 14 on page 153 shows how the Oracle Server and NetWorker processes interact during a scheduled deferred live backup.

Figure 14 Deferred live backup

Proxy restore processesA proxy restore can be performed in either of two ways:


Or


The RMAN utility starts Oracle Server processes on the target database. These Oracle Server processes start the restore by calling MML routines.

The following sections describe the process steps in the three types of proxy restore:

◆ “Instant restore or rollback” on page 153

◆ “Restore from secondary storage” on page 155

Instant restore or rollback

An instant restore or rollback involves the following steps:

1. The MML performs the following:

a. Translates the instant backup object names requested by RMAN into a format that the NetWorker server understands.

Oracle Server(NetWorker Client)

PowerSnap Proxy Client(NetWorker Storage Node)

NetWorkerServer

SecondaryStorage Medium

nsrmmdbd

Media Database

Client File Index

interprocesscommunication

nsrindexd

nsrexecd nsrmmd

trackinginformation

Primary Storage

OracleData Files

Point-in-Time Copyof Oracle Data Files

data

GEN-000177

StartHere

nsrsnap (at endof savegrp) ornsrsnapadmin

PowerSnapprocesses

PowerSnapelements



b. Verifies the existence of the instant backup objects through the nsrindexd service.

c. Verifies that the restore destination directory exists and is accessible, and creates the .nworapc directory.

2. The MML contacts the PowerSnap master program, nsrpsd, on the Oracle Server host.

3. The nsrpsd program works with other PowerSnap and NetWorker programs to perform the restore operation:

• PowerSnap processes retrieve the point-in-time metadata stored on the device associated with the snapshot pool by using the nsrmmd and nsrmmdbd processes.

• PowerSnap processes perform one of the following:

– During an instant restore, mount the point-in-time copy and copy the files into a destination requested by NSR-DBM-ORA.

– During a rollback, execute a command (specific to the primary storage) to move the whole point-in-time copy (snapshot unit) into the source destination.

4. If operating system files (not raw devices or volumes) are restored during an instant restore, NSR-DBM-ORA moves the restored files from the .nworapc directory to the parent directory.

Once the required Oracle files are restored, a database administrator can complete the standard Oracle database recovery.

Figure 15 on page 155 shows how the Oracle Server and NetWorker processes interact during an instant restore or rollback.

Proxy restore processes 155


Figure 15 Instant restore or rollback

Restore from secondary storage

A restore of a proxy backup from secondary storage involves the following steps:

1. The MML performs the following:

a. Translates the instant backup object names requested by RMAN into a format that the NetWorker server understands.

b. Verifies the existence of the instant backup objects through the nsrindexd service.

c. Verifies that the restore destination directory exists, and creates the .nworapc directory.

2. The MML contacts the PowerSnap master program, nsrpsd, on the Oracle Server host.

3. The nsrpsd program works with other PowerSnap and NetWorker programs to retrieve the data from secondary storage, and perform the restore operation.

PowerSnap processes restore the files (save sets) into a destination requested by NSR-DBM-ORA. The processes use the nsrmmd and nsrmmdbd processes to determine which media contain the requested save sets and to read the backup volumes.



NetWorker Server

SnapShot Pool Media

nsrmmdbd

Media Database

Client File Index


nsrindexd

nsrmmd


nsrd



data (for details on data flow, see description)

nsrexecd

Oracle process

nsrexecd

RMAN

GEN-000178

p

NSR-DBM-ORA

MML

PowerSnap module

nsrpsd

PowerSnap module

PowerSnap elements




4. If operating system files (not raw devices or volumes) are restored, NSR-DBM-ORA moves the restored files from the .nworapc directory to the parent directory.

Once the required Oracle files are restored from the backup volumes, a database administrator can complete the standard Oracle database recovery.

Figure 16 on page 156 shows how the Oracle Server and NetWorker processes interact during a restore from secondary storage.

Figure 16 Restore from secondary storage

Software configuration roadmapThis section describes the configuration requirements for proxy backups and restores.

Before performing the required configuration procedures, ensure that both the NSR-DBM-ORA and required PowerSnap Module software are installed according to the instructions in the following:


◆ NetWorker PowerSnap Module documentation(Refer to the PowerSnap Module version for the primary storage system.)



NetWorker Server


Data Files

Secondary Storage

nsrmmdbd

Media Database

Client File Index


nsrindexd

nsrmmd


nsrd

data

nsrexecd

Oracle process

RMAN

p

NSR-DBM-ORA

MML

PowerSnap module

nsrpsd

GEN-000179

PowerSnap elements


Software configuration roadmap 157


Basic configurations

Ensure that the following basic configurations are performed:

◆ The basic Oracle Server and NetWorker configurations are performed according to the “Configuration roadmap” on page 58. The NetWorker Server, Client, Device, and other resources must be configured:

• For live Oracle backups, a secondary storage device, such as a tape drive, must be configured.

• For instant Oracle backups, a snapshot Pool resource must be configured, as described in “Additional configurations for proxy operations” on page 157.

◆ The required scheduled backup configurations are performed according to “Configuring a backup with the NMC legacy method” on page 70.

Additional configurations for proxy operations

For proxy operations, you must configure the Oracle database layout to position the datafiles on primary storage that is supported by the specific PowerSnap Module.

Ensure that the additional configuration procedures described in the following sections are performed:

◆ “Configuring internationalization (I18N) support” on page 157

◆ “Configuring the required Oracle settings” on page 157

◆ “Configuring the NWORA resource file” on page 158

◆ “Creating RMAN scripts for proxy backups” on page 158

◆ “Configuring the NetWorker Pool resources” on page 162

◆ “Configuring the NetWorker Snapshot Policy resource” on page 162

◆ “Configuring the NetWorker Group resource” on page 163

◆ “Configuring the NetWorker Client resource” on page 163

◆ “Testing a scheduled proxy backup” on page 164

Once all the configuration requirements are met, review the information in the following sections prior to performing proxy backups and restores:

◆ “Proxy backup requirements” on page 164

◆ “Proxy backup information in the NetWorker indexes” on page 167

◆ “Proxy restore requirements” on page 170

Configuring internationalization (I18N) support

In a non-English environment, NSR-DBM-ORA supports internationalization (I18N) of proxy backups and restores with a supported release of the PowerSnap Module, as described in “Internationalization (I18N)” on page 31.

To configure I18N support for proxy backups, follow the instructions in “Configuring I18N support” on page 64.

Configuring the required Oracle settings

Do not locate the database control files and online redo log files on the same volume (snapshot unit) as the datafiles that will be backed up through proxy backups. Otherwise point-in-time recover would not be able after proxy rollbacks



If the Oracle database is expected to have a lot of read or write activity, or an error, such as “skgfdisp: async read/write failed” appears, increase the value of LARGE_POOL_SIZE to a large value that is appropriate for the particular system in the Initialization Parameter file(such as initoracle_sid.ora).

Configuring the NWORA resource file

To enable proxy backups, the NSR_ORACLECAT_MODE parameter resource must be set to either enabled or disabled in the NWORA resource file, as described in “The NWORA resource file” on page 176. If the resource value is left unset, proxy backups fail.

To enable catalog synchronization, perform the configuration procedures in “Catalog synchronization for proxy backups” on page 174.

Notes:

◆ If catalog synchronization is enabled for instant backups, the NWORA resource file must contain an NWORA SID resource for each Oracle database to be backed up during instant backups.

◆ Catalog synchronization must be configured before any proxy backups of a database are performed.

◆ The parameter ORACLE_SID in the nsrnmo script (used for the proxy backup) must be set to the same value as the parameter NSR_ORACLE_SID in the NWORA SID resource.

Creating RMAN scripts for proxy backups

The basic information about RMAN backup scripts in “Create RMAN scripts for backups” on page 70 also applies to RMAN scripts for proxy backups:

These added requirements apply to RMAN scripts for proxy backups:

◆ The appropriate parameters must be set, as described in “Setting the parameters” on page 160.

◆ The proxy or proxy only option must be specified with each RMAN backup command.

Note: Certain options of the RMAN backup command, such as maxsetsize and diskratio, are not supported with the proxy option. Contact Oracle Corporation for more information on the RMAN options that are not supported.

◆ As required by Oracle for proxy backups, the %p variable must be included in the format string, either explicitly or implicitly within %U. The appropriate Oracle backup and recovery documentation provides more information.

◆ It is recommended to allocate only one channel in the RMAN script. It is not recommended to allocate more than one channel in the RMAN script, in an attempt to distribute the proxy backup over more than one channel.

Note: Proxy backup parallelism is defined by the PowerSnap parameter NSR_PS_SAVE_PARALLELISM. Table 7 on page 161 provides more information.



The following sample RMAN script performs a proxy backup of an entire Oracle database that resides on one or more primary storage devices:

run {allocate channel t1 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_PROXY_PFILE=/oracle/rman/proxy.cfg)’;backup full proxy onlyformat ’FULL_%d_%U’(database);release channel t1;

}

NSR_PROXY_PFILE is an optional NSR-DBM-ORA parameter used for proxy backups. “Setting the parameters” on page 160 provides details.

Multiple channels in RMAN scriptsThe allocation of multiple channels in an RMAN script does not control the degree of backup or restore parallelism. Oracle uses only one of the allocated channels for the proxy backup or restore, unless specific backup options are used.

Example 25 RMAN scripts with multiple channels

The proxy backup performed with the following RMAN script is written to either the OracleVolume1 or OracleVolume2 volume pool (not to both volume pools) because Oracle uses only one of the allocated channels for the proxy backup:

run {allocate channel c1 type ’SBT_TAPE’;allocate channel c2 type ’SBT_TAPE’;send channel c1 ’NSR_ENV=(NSR_DATA_VOLUME_POOL=OracleVolume1)’;send channel c2 ’NSR_ENV=(NSR_DATA_VOLUME_POOL=OracleVolume2)’;backup proxy only tablespace tbs1, tbs2, tbs3, tbs4;release channel c1;release channel c2;

}

The following RMAN script uses specific backup options to enforce distribution of the proxy backup over two channels:

run {allocate channel c1 type ’SBT_TAPE’;allocate channel c2 type ’SBT_TAPE’;send channel c1 ’NSR_ENV=(NSR_DATA_VOLUME_POOL=OracleVolume1)’;send channel c2 ’NSR_ENV=(NSR_DATA_VOLUME_POOL=OracleVolume2)’;backup proxy(tablespace tbs1, tbs2 channel c1)(tablespace tbs3, tbs4 channel c2);release channel c1;release channel c2;

}

Use the following RMAN script to replace both of the preceding two backup scripts:

run {allocate channel c1 type ’SBT_TAPE’;send channel c1 ’NSR_ENV=(NSR_DATA_VOLUME_POOL=OracleVolume1)’;backup proxy tablespace tbs1, tbs2, tbs3, tbs4;release channel c1;

}

You might want to allocate more than one channel if you know that some of the data does not reside on supported primary storage devices. In this case, one channel is used for proxy backups and all the others are used for regular backups.



Setting the parametersTwo types of parameters can be set for the proxy backup and restore operations:

◆ NSR-DBM-ORA parameters, as described in Appendix A, “Parameters in an RMAN Session.”

The parameters must be set by using one of the methods in “Setting the NSR* parameters” on page 190.

◆ PowerSnap Module parameters, as described in “PowerSnap parameter settings” on page 160.

PowerSnap parameter settingsThe PowerSnap parameters must be set by using one of the following methods:

◆ By setting the parameters in the send command in one of these ways:

• With the rman command on the operating system command line.

• In the RMAN backup or restore script.

“The send command” on page 201 provides more information on how to use the send command.

◆ By setting the parameters in a user-defined configuration file. The complete pathname of the file must be specified in the parameter NSR_PROXY_PFILE, as described in “NSR_PROXY_PFILE” on page 195.

The configuration file consists of a separate line such as the following for each parameter value setting:

parameter_name=parameter_value

where:

• parameter_name is the parameter name, such as RESTORE_TYPE_ORDER.

• parameter_value is the parameter value, such as pit.

Use the following guidelines to set PowerSnap parameters:

◆ A parameter setting in the configuration file takes precedence over a parameter setting in the send command.

If the same PowerSnap parameter is set to different values in the configuration file and send command, the value in the configuration file is the one used for the proxy operation.

◆ In the configuration file, the first valid occurrence of a PowerSnap parameter takes precedence over any other occurrences of the same parameter in the same file.

◆ The following are not supported:

• The use of the parms option in the configure channel command to set PowerSnap parameters.

• The use of the setenv command on the operating system command line to set PowerSnap parameters.

Example 26 on page 161 and Example 27 on page 161 provide examples of PowerSnap parameter settings.

Table 7 on page 161 provides a basic list of supported PowerSnap parameters. The list is not exhaustive. The NetWorker PowerSnap Module documentation provides a complete list of PowerSnap parameters.

For each parameter, the table includes:



◆ A description of the parameter.

◆ The default value of the parameter.

◆ The valid values that can be assigned to the parameter.

Example 26 PowerSnap parameter settings

To set the PowerSnap parameter RESTORE_TYPE_ORDER for a proxy restore, a configuration file named /oracle/rman/proxy.cfg can be created, consisting of the following line:

RESTORE_TYPE_ORDER=rollback:pit:conventional

In this case, the NSR-DBM-ORA parameter NSR_PROXY_PFILE must be set to /oracle/rman/proxy.cfg by using the send command. For example, the following command sets the parameter correctly:

allocate channel t1 device type ’SBT_TAPE’;send ’NSR_ENV=(NSR_PROXY_PFILE=/oracle/rman/proxy.cfg)’;

Example 27 PowerSnap parameter settings for a Celerra NAS device

To enable proxy backup and restore operations with Celerra® NAS devices, ensure that the following PowerSnap parameters are set in the user-defined configuration file that you specify with the NSR-DBM-ORA parameter NSR_PROXY_PFILE:

◆ NSR_DATA_MOVER=name or IP of NetWorker data mover

Identifies the NetWorker data mover to use for rollovers.

◆ NSR_SNAP_NAS_CEL_CS_HOST=name or IP of Celerra control station

Identifies the Celerra control station.

◆ NAS_SNAP_SUBTYPE=CEL_SNAPSURE

Identifies the NAS SCM subtype to use.

Table 7 PowerSnap parameters

Parameter Description Default and valid values

NSR_DATA_MOVER Mandatory for a proxy backup that uses a proxy client host. Specifies the hostname of the proxy client host.

• Local host (default).• The valid hostname of the proxy client host.

NSR_PS_SAVE_PARALLELISM Optional. Specifies the number of concurrent save streams on the proxy client host.

• 16 (default).• An integer value less than or equal to the

Parallelism attribute value in the NetWorker Client resource.

NSR_MAX_STREAMS Optional. Specifies the maximum number of restore streams.

• 16 (default).• An integer value.

RESTORE_TYPE_ORDER Optional. Specifies the type of proxy restore to be performed.

Note: If multiple values are specified, each type of restore is attempted (n the order specified) until a restore operation is successful.

• pit:conventional (default).• One or more of the following values, each value

delimited from the others by a colon(:):• pit — Specifies an instant restore.• conventional — Specifies a proxy restore from

secondary storage media.• rollback — Specifies a rollback restore from a

point-in-time proxy copy.“Proxy restore requirements” on page 170 provides more information.



◆ NSR_SNAP_TYPE=nas

Specifies that this is a NAS save object.

Note: The value of NSR_SNAP_TYPE must be lowercase nas.

◆ NSR_SNAP_NAS_CLIENT=name or IP address of NAS filer with the NFS file system

Identifies the NFS server for the specified mount point.

The PowerSnap Module documentation provides more details on these PowerSnap parameters.

For example, the following PowerSnap parameters are included in the /nsr/res/nas_backup.cfg file (specified with NSR_PROXY_PFILE) for a proxy backup with a Celerra NAS device:

cat /nsr/res/nas_backup.cfg

NSR_PS_DEBUG_LEVEL=9NSR_DEBUG_LEVEL=9NSR_DATA_MOVER=datamover.fujitsu.com NSR_SNAP_NAS_CEL_CS_HOST=controlstnNAS_SNAP_SUBTYPE=CEL_SNAPSURENSR_SNAP_TYPE=nasNSR_SNAP_NAS_CLIENT=11.222.333.44

For example, the following PowerSnap parameters are included in the /nsr/res/nas_restore.cfg file (specified with NSR_PROXY_PFILE) for a proxy restore with a Celerra NAS device:

cat /nsr/res/nas_restore.cfg

NSR_PS_DEBUG_LEVEL=9NSR_DEBUG_LEVEL=9NSR_DATA_MOVER=datamover.fujitsu.com NSR_SNAP_NAS_CEL_CS_HOST=controlstnNAS_SNAP_SUBTYPE=CEL_SNAPSURENSR_SNAP_TYPE=nasNSR_SNAP_NAS_CLIENT=11.222.333.44RESTORE_TYPE_ORDER=conventional

Configuring the NetWorker Pool resources

A separate pool must be configured to support proxy backups. The PowerSnap Module stores the metadata from the point-in-time copy (snapshot) in this pool. The pool is configured by using the same method as for a regular NSR-DBM-ORA backup. However, the specified backup device should be a file or advanced file type.

Note: Specify the pool name in the Snapshot Pool attribute of the NetWorker Group resource, as described in “Configuring the NetWorker Group resource” on page 163.

The NetWorker PowerSnap Module documentation provides more information on configuring this extra pool.

Configuring the NetWorker Snapshot Policy resource

A special NetWorker snapshot policy is required to perform proxy backups. You can either specify a preconfigured policy or create a new snapshot policy.

Configure a NetWorker Snapshot Policy resource by using the instructions in the NetWorker PowerSnap Module documentation.



Configuring the NetWorker Group resource

For proxy backups, configure a NetWorker Group resource by using the instructions in the NetWorker PowerSnap Module documentation.

Configuring the NetWorker Client resource

The basic NetWorker Client resource for the Oracle Server is configured according to “Configure a Client resource with NMC” on page 81.

Notes:

◆ Only members of a NetWorker user group with the Change Security Settings privilege can set the Remote Access attribute in the Client Resource.

◆ The Browse Policy attribute in the Client resource applies only to the NetWorker client file index entries for backups residing on secondary storage (live backups).

◆ The Retention Policy attribute in the Client resource applies only to the NetWorker media database entries for live backups.

◆ The lifecycle of a point-in-time copy (instant backup) is governed by the snapshot policy specified in the Group resource to which the given client belongs.

How to configure the NetWorker Client resourceTo configure the NetWorker Client resource for the Oracle Server:

1. For the Save Set attribute, specify the complete pathname of each RMAN script to be used for the backups.

2. For the Group attribute, specify the name of the NetWorker Group resource created for the proxy backups. “Configuring the NetWorker Group resource” on page 163 provides more information.

3. For the Schedule attribute, specify a NetWorker backup schedule.

4. For the Backup Command attribute, specify the name of the nsrnmo script to use for the proxy backups.

5. For the Parallelism attribute (a hidden attribute), specify the number of data streams that the Oracle Server is allowed to send in parallel to the NetWorker server or storage node.

6. For the Storage Nodes attribute, specify the name of each storage node to which the Oracle Server can back up data.

The Oracle Server backs up to the first active, enabled storage node in the order listed in the attribute. The default storage node name, nsrserverhost, represents the NetWorker server.

7. For the Remote Access attribute, specify the user ID or hostnames of other clients that are allowed to back up or restore this client’s files.

For proxy backups that use a proxy client host, the Remote Access attribute must include the proxy client hostname.

8. In the Application Information attribute, specify all PowerSnap module specific parameters that you did not already set in the NSR_PROXY_PFILE. For a complete list of PowerSnap parameters refer to the appropriate NetWorker PowerSnap Module Installation and Administrator's Guide.



Testing a scheduled proxy backup

IMPORTANT!A proxy backup can be started only by automatic or manual invocation of the scheduled NetWorker backup group. A proxy backup cannot be scheduled through Oracle Enterprise Manager, or started by invoking RMAN from the operating system command line.

To verify the scheduled backup setup, follow the instructions for regular backups in “Test a scheduled backup” on page 104.

Since manual (unscheduled) proxy backups are not supported, you cannot test a proxy RMAN script by using the information in “Test RMAN scripts for scheduled backups” on page 74. To determine if the script contains any errors, log the RMAN output into a file by setting the parameter NSR_RMAN_ARGUMENTS in the nsrnmo script used for the backup.

Proxy backup requirementsReview the following information, specific to proxy backups:

◆ “Checking configuration consistency” on page 164

◆ “Performing proxy backups” on page 166

“Proxy backups and restores on cluster systems” on page 185 provides information on proxy backups in a cluster environment.

Checking configuration consistency

During a scheduled backup, NSR-DBM-ORA checks for consistency between the NetWorker Group resource configuration and the RMAN backup session.

If NSR-DBM-ORA finds a discrepancy between the Group resource configuration and the RMAN session, warning messages are generated or the backup fails, as described in the following sections:

◆ “With a group configured for proxy backups” on page 164

◆ “With a group configured for regular backups” on page 165

With a group configured for proxy backups

If the Snapshot attribute in the NetWorker Group resource is set to True, the resource is configured for proxy backups. However, this configuration does not guarantee that a proxy backup is executed. RMAN might still perform only regular Oracle backups if either of the following exists:

◆ None of the backup commands in the RMAN script include the proxy or proxy only option.

◆ The backup commands in the RMAN script include the proxy or proxy only option, but none of the Oracle database objects (tablespaces or datafiles) specified in the backup commands reside on a primary storage device that the PowerSnap Module supports.

Proxy backup requirements 165


If RMAN performs only regular Oracle backups due to one of these conditions, NSR-DBM-ORA generates the following warnings in the savegroup completion report:

WARNING: Snapshot savegrp is completed but no Oracle proxy backup is detected.WARNING: Either fix your RMAN script or reconfigure the group resource without snapshot flag.

While the resulting backups are valid regular (nonproxy) backups, correct the RMAN script or relocate the Oracle datafiles to a supported primary storage device, as required to enable proxy backups.

If a backup command in the RMAN script includes the proxy only option and the Oracle data objects reside on volumes that do not support snapshots, the scheduled backup fails since RMAN cannot perform a regular backup of the objects. The Oracle documentation provides a detailed description of the difference between the proxy and proxy only options.

Note: If the PowerSnap Module software involved in a proxy backup cannot determine if a file is snapshotable, the proxy backup fails.

With a group configured for regular backups

If the Snapshot attribute in the NetWorker Group resource is set to False, the resource is configured for regular backups. In this case, the use of the proxy or proxy only option with a backup command in the RMAN script is not supported. Any proxy backup specified in the RMAN script will fail. If there are regular and proxy backups in the same RMAN script, RMAN might complete one or more regular backups before a proxy backup fails.

Notes:

◆ If RMAN terminates any of the proxy backups in an RMAN script, the savegroup completion report lists failure of the scheduled backup.

◆ If any proxy backups in an RMAN script fail, RMAN still performs a regular backup of the corresponding archived redo logs.

Example 28 Proxy backup failure

A scheduled backup includes the following RMAN script, with the database files residing on volumes that support snapshots. However, the Snapshot attribute in the Group resource is set to False. As a result, the proxy database backup fails:

run {allocate channel ch1 type ’SBT_TAPE’;allocate channel ch2 type ’SBT_TAPE’;backup proxy database plus archivelog;

}

Despite the proxy backup failure, RMAN performs a regular backup of the archived redo logs. The savegroup completion report lists failure of the scheduled backup.



Performing proxy backups

A proxy backup can be started only by automatic or manual invocation of the scheduled NetWorker backup group.

IMPORTANT!A proxy backup cannot be scheduled through Oracle Enterprise Manager or started by invoking RMAN from the operating system command line. The NetWorker PowerSnap Module documentation provides information on how to manually invoke a scheduled backup.

Specific types of Oracle files, such as control files, cannot be backed up through a proxy backup. This is an Oracle constraint. The Oracle documentation for the particular Oracle Server release provides more information on the Oracle file types that do not support proxy backups.

Directory for temporary files

NSR-DBM-ORA creates temporary files for processing purposes in the /nsr/tmp directory.

Note: During RMAN operations, do not touch any files in this directory.

Backup summary line in savegroup report

The savegroup completion report for a proxy backup contains a summary line that includes the backup size and number of files. The summary line refers to backup data written to NetWorker devices only.

The summary line for an instant backup includes the size of only the metadata stored for the backup, not the size of the files stored on the primary storage as a point-in-time copy. The number of files includes the number of entries generated for the metadata plus the number of entries generated for the backup pieces.

Savegroup completion status

When a deferred live backup is run as part of a scheduled group, the backup process involves two steps:

1. An instant backup is performed. At the end of the instant backup, the backup entries for the point-in-time copy are recorded in the NetWorker indexes and RMAN catalog.

2. The deferred live backup is performed. At the end of the deferred live backup, the backup entries for data stored on the secondary storage are recorded in the NetWorker indexes.

“Proxy backup processes” on page 147 provides more details.

If the instant backup succeeds but the deferred live backup fails, the entire scheduled backup is reported as failed. However, the point-in-time copy created during the instant backup is a valid backup and can be used for instant or rollback restore.

Note: If RMAN performs only a regular Oracle backup during the instant backup step (“Checking configuration consistency” on page 164 provides details on when this can happen), the deferred live backup fails because there is no point-in-time copy to be moved to secondary storage. The entire backup is reported as failed, but the data is stored on tape and can be used for restore.

Proxy backup information in the NetWorker indexes 167


NWORA resource file backup

If a scheduled proxy backup completes successfully, NSR-DBM-ORA automatically backs up the NWORA resource file, as described in “The NWORA resource file” on page 176.

The NWORA resource file backup is performed at the backup level specified in the Schedule resource (for example, incremental). Oracle backups are always performed at the full level. The NetWorker server selects the pool for the NWORA resource file backup based on existing resource configurations. The setting of the parameter NSR_DATA_VOLUME_POOL does not affect the pool selection.

The savegroup completion report contains a summary line for the backup that includes the phrase "NWORA Resource Backup." The information is also written to the scheduled backup debug file, specified by the parameter NSR_SB_DEBUG_FILE in the nsrnmo script.

In the NetWorker indexes, the save set name for the NWORA resource file backup is the same as the file pathname. You can use the NetWorker mminfo command to display the save set name. “NWORA resource file backup in the NetWorker indexes” on page 169 provides information on how the backup is represented in the NetWorker indexes.

The NWORA resource file backup can be restored by using the NetWorker recover command or nwrecover GUI program. The NetWorker Administrator’s Guide provides more information.

Note: The file is stored under the "backup" namespace, not the "oracle" namespace.

The browse and retention policies applied to the NWORA resource file backup are the most conservative policies associated with the given NetWorker client, not the policies that are applied to the Oracle backups. As a result, you may see a difference between the policies assigned to the NWORA resource file backup and the Oracle backups.

Canceling proxy backups

Proxy backups can be canceled by using the same methods as for regular Oracle backups. The following sections provide more information:

◆ “Cancel a manual backup” on page 101

◆ “Cancel a scheduled backup” on page 105

Proxy backup information in the NetWorker indexesThe NetWorker server maintains information about each backup in its online indexes. “NetWorker software” on page 44 provides more information.

The index entry for a proxy backup is stored in the NetWorker client file index for the Oracle Server host, under the "oracle" namespace (as is the case for a regular backup). The NetWorker client file index and media database each contain a different value for the name of the save set for a proxy backup (as is the case for a regular scheduled backup).

Query the online NetWorker indexes by using the NetWorker commands, nsrinfo and mminfo:

◆ Type the nsrinfo command to query the NetWorker client file index. For example:

nsrinfo -n oracle -s NetWorker_server Oracle_Server_hostname



◆ Type the mminfo command to query the NetWorker media database. For example:

mminfo -v -s NetWorker_server -c Oracle_Server_hostname


Entries in the client file index

For a backup piece created through a proxy backup, the client file index contains three types of backup entries under the "oracle" namespace:

◆ One entry is generated for the backup piece name assigned by RMAN, such as /PROXY_O901JB_811_1/ in Example 29 on page 168.

◆ The second entry is generated for the point-in-time metadata, such as /brcmeta.1/ in Example 29 on page 168. This entry is created for an instant backup only.

◆ The third entry is generated for the Oracle datafile that is backed up to secondary storage, for example, /JBOD13_NMO41_MVOL3/tbspc4_data1.dbf in Example 29 on page 168. This entry is created for a live backup only.

Example 29 Proxy backup entries in the client file index

The nsrinfo command provides information on the proxy backup entries in the NetWorker client file index:

nsrinfo -n oracle marmaris

scanning client ’marmaris’ for all savetimes from the oracle namespace/PROXY_O901JB_811_1/, date=1178916449 Fri May 11 13:47:28 2007/brcmeta.1/, data=1178916446 Fri May 11 13:47:25 2007Physical files to rollover:/JBOD13_NMO41_MVOL3/tbspc4_data1.dbf/JBOD13_NMO41_MVOL3/tbspc4_data1.dbf, date=1178916453 Fri May 11 13:47:31 2007

Entries in the media database

For a backup piece created through a proxy backup, the media database contains two types of entries:

◆ One entry is generated for the point-in-time metadata. This entry is created for an instant backup only.

In the mminfo command output for this entry:

• The Size field contains the size of the metadata stored on the NetWorker device.

• The Flag field (fl) includes the letter P, representing the point-in-time copy.

To list the entries for an instant backup only, type the following mminfo command:

mminfo -v -c Oracle_Server_hostname -q ’snap’

The NetWorker PowerSnap Module documentation provides more information.

◆ The other entry is generated for the Oracle datafile that is backed up to secondary storage. This entry is created for a live backup only.

Proxy backup information in the NetWorker indexes 169


Both entries in the media database include the name of the RMAN backup script used for the proxy backup, such as /space1/home/oracle/bp1 in Example 30 on page 169.

Example 30 Proxy backup entries in the media database

The mminfo command provides information on the proxy backup entries in the NetWorker media database:

mminfo -v -c marmaris

volume client date time sizenmo.002 marmaris 05/10/07 13:18:39 102 MBsnap.001 marmaris 05/10/07 13:18:41 2 KB

ssid fl lvl name4064690015 cb full /space1/home/oracle/bp14098244417 cbP full /space1/home/oracle/bp1

NWORA resource file backup in the NetWorker indexes

In the NetWorker indexes, the NWORA resource file backup is stored under the "backup" namespace. As a result, the NetWorker recover or nwrecover program can be used to restore the backup. The save set name for the backup is the same as the file pathname.

Query the NetWorker indexes for information about the NWORA resource file backup by using the NetWorker commands nsrinfo and mminfo.

Example 31 Resource file backup entry in the client file index

The nsrinfo Oracle_Server_hostname command provides information on the NWORA resource file backup entry in the NetWorker client file index:

nsrinfo marmaris

scanning client ’marmaris’ for all savetimes from the backup namespace

/nsr/res/nwora.res, date=1178808677 Thu May 10 13:18:39 2007/nsr/res/, date=1178808677 Thu May 10 13:18:39 2007/nsr/, date=1178808677 Thu May 10 13:18:39 2007/, date=1178808677 Thu May 10 13:18:39 2007

Note: This entry is not displayed with the nsrinfo -n oracle command because it is stored under the "backup" namespace, not the "oracle" namespace. The "backup" namespace is the default namespace for the nsrinfo command.

Example 32 Resource file backup entry in the media database

The mminfo -v -c Oracle_Server_hostname command provides information on the NWORA resource file backup entry in the NetWorker media database:

mminfo -v -c marmaris

volume client date time sizenmo.002 marmaris 05/10/07 13:18:39 4 KB

ssid fl lvl name3863367791 cb full /nsr/res/nwora.res




Proxy restore requirementsChapter 4, “Data Restore and Recovery,” provides basic information on how to configure and run Oracle data restore and recovery by using NSR-DBM-ORA.

Review the following information, specific to proxy restores:

◆ “Creating RMAN scripts for proxy restores” on page 170

◆ “Performing proxy restores” on page 171

◆ “Relocating files during proxy restores” on page 172

◆ “Restoring to a different host” on page 173

◆ “Point-in-time recoveries without a Recovery Catalog” on page 173

“Proxy backups and restores on cluster systems” on page 185 provides information on proxy restores in a cluster environment.

Creating RMAN scripts for proxy restores

The same RMAN script used for a regular Oracle restore can also be used for a proxy restore.

Note: The RMAN restore command does not include a proxy option.

To create an RMAN script for a proxy restore, follow the instructions in Chapter 4, “Data Restore and Recovery.”

To perform a proxy restore, the appropriate parameters must be set, as described in “Setting the parameters” on page 160.

The RESTORE_TYPE_ORDER parameter

The RESTORE_TYPE_ORDER parameter setting determines the type of proxy restore that is performed:

1. RMAN determines which backup needs to be restored and passes the required backup piece name to NSR-DBM-ORA.

2. The RESTORE_TYPE_ORDER parameter specifies whether the backup piece is to be restored by using one of the following:

• The point-in-time copy

• The copy stored on secondary storage

Example 33 RESTORE_TYPE_ORDER parameter settings

If the RESTORE_TYPE_ORDER parameter is set to the value rollback:pit, a rollback restore is attempted first. If it fails, an instant restore (indicated by pit) is attempted. If the parameter is not set, the default order pit:conventional is used, where conventional represents a restore from secondary storage.

If the rollback option is not set explicitly, a rollback is not attempted. “Performing proxy restores” on page 171 provides more information on setting up a rollback operation.

Proxy restore requirements 171


IMPORTANT!For the RESTORE_TYPE_ORDER parameter, NSR-DBM-ORA does not support the force_rollback option, which is supported by PowerSnap Modules. If the option is specified, the restore fails, even if other valid restore options are also specified.

The NSR_CLIENT parameter

To restore Oracle data to a different host, the parameter NSR_CLIENT must be set to the required hostname. “Restoring to a different host” on page 173 provides more information.

Performing proxy restores

The following requirements apply to proxy restores:

◆ The PowerSnap Module software must be installed, according to the instructions in the NetWorker PowerSnap Module documentation (refer to the PowerSnap Module version for the primary storage system).

◆ Each element of the restore path must exist. Otherwise, the restore fails. For example, to restore a file backup to /space1/oradata/file.dbf, the path /space1/oradata must exist.

◆ A proxy restore of a symbolic link restores the Oracle file to the location pointed to by the symbolic link. Both the symbolic link and the restore path must exist. Otherwise, the restore fails.

◆ For a rollback restore, the psrollback.res file must be set up properly, as described in “Rollback restore” on page 172.

◆ For user-specified relocation of files during a proxy restore, the relocation path must be specified as described in “Relocating files during proxy restores” on page 172.

◆ After an Oracle restore is complete, a database administrator must recover the database by using the standard Oracle recover command.

Concurrent restore streams

During a proxy restore, the PowerSnap Module creates concurrent restore streams to optimize the restore.

The maximum number of concurrent restore streams is defined by the PowerSnap parameter NSR_MAX_STREAMS. Table 7 on page 161 provides more information.

Directory created for file system data restore

During a proxy restore of regular file system data, a .nworapc subdirectory (with 0700 permissions) is created under the restore directory for the temporary relocation of the files being restored. (This relocation is independent of user-specified relocation.) The empty .nworapc subdirectory persists after the restore and can be deleted manually, if required.

If a proxy restore of file system data fails, the non-empty .nworapc subdirectory persists after the restore, and can be deleted manually, if required. Do not use any datafiles from this subdirectory for Oracle recovery, or database corruption might occur. If you restart the failed restore, NSR-DBM-ORA automatically cleans this subdirectory.



Rollback restore

For a rollback restore, the psrollback.res file must contain the directory name .nworapc. The file is located in the /nsr/res/psrollback.res directory.

Add the directory name to the file by using a text editor as the root user.

The following sources provide more information on the psrollback.res file:

◆ NetWorker PowerSnap Module documentation for the primary storage system.

◆ Comments within the psrollback.res file itself.

Note: The NetWorker PowerSnap Module documentation provides information on whether rollback is supported on a particular storage platform.

Rollback restore on Celerra NAS devicesNSR-DBM-ORA software supports the rollback safety check feature for rollback restores on Celerra NAS devices. The psrollback.res file lists all the files, directories, partitions, and volumes that are to be excluded from the rollback safety check. The items excluded from the safety check will be overwritten during a rollback operation.

Note: For NSR-DBM-ORA systems, “.etc” must be added to the psrollback.res file.

To enable remount of the NAS file system at the end of a rollback operation, place an entry for the target file system in the /etc/vfstab file on Solaris.

If this is not done, the remount at the end of the rollback fails. The data is recovered, but the file system must be remounted manually and the tablespace brought back online.

Relocating files during proxy restores

This section describes the user-specified relocation of a proxy restore with NSR-DBM-ORA.

IMPORTANT!Relocation is not supported during a rollback restore. If the RESTORE_TYPE_ORDER parameter includes the rollback value and the RMAN restore script specifies relocation, the restore fails, even if the parameter includes other values.

During a proxy restore, NSR-DBM-ORA supports and controls relocation, which is the restore of datafiles (regular files or raw volumes) to a new location. The new location can be specified by using the RMAN set newname command.

Note: During a regular Oracle restore, relocation is also supported, but it is controlled by the Oracle Server.

To relocate a regular file or raw volume during a proxy restore, the set newname command must specify the name of the relocated file as one of the following:

◆ The complete pathname of the relocated file.

◆ The complete pathname of a symbolic link that points to the location where the file will be restored.

Proxy restore requirements 173


Example 34 Symbolic link specified in the set newname command

If the symbolic link /tmp/file1 points to /dbapps/proddb/file2 and the symbolic link /tmp/file1 is specified in the set newname command, the backed-up file will be restored to /dbapps/proddb/file2.

IMPORTANT!The procedure to relocate a raw volume includes a restriction that does not apply when relocating a regular file.

To relocate a raw volume, the base filename (the filename without the directory path) of the original backed-up raw volume must be identical to one of the following:

◆ The base filename of the relocation path specified in the set newname command.

◆ If the set newname command specifies a symbolic link, the base filename in the symbolic link.

Example 35 Relocation of a raw volume

If a backed-up raw volume is named /dev/volume_one/rvol1, the /dev/volume_two/rvol1 relocation path can be specified in the set newname command. This can occur because the original and relocation paths have the same base filename, rvol1. However, specifying the /dev/volume_one/rvol2 path in the set newname command would cause the proxy restore to fail, since the original and relocation paths have different base filenames.

The following procedure is one way to relocate /dev/volume_one/rvol1 to /dev/volume_one/rvol2:

1. Create a symbolic link named /tmp/rvol1 that points to /dev/volume_one/rvol2.

2. Specify /tmp/rvol1 in the set newname command in the RMAN restore script.

In this case, the relocation succeeds because both the original path and symbolic link name have the same base filename, rvol1.

Restoring to a different host

To restore proxy database backups to a different host, follow the guidelines in “Restore to a different host” on page 120. Task 2 of that section explains that NSR-DBM-ORA must be installed and configured on the system where the Oracle data is to be restored. For proxy restores to a different host, the required PowerSnap Module must also be installed and configured on that system.

Point-in-time recoveries without a Recovery Catalog

Note: If point-in-time recoveries are performed with an RMAN Recovery Catalog, the information in this section does not apply.

During a proxy database backup, Oracle backs up the control file after the proxy backup of the datafiles is complete. In a large database production environment, there might be a delay between the end time of the datafile backup and the start time of the control file backup.



During this time delay, if the physical structure of the database is changed (for example, a new datafile is added), the control file must be backed up in a separate RMAN session before the changes occur. This is due to the fact that the control file backup from the proxy database backup session will include information on the new database structure.

Catalog synchronization for proxy backupsDuring Oracle backups, RMAN stores information about each backup piece in the RMAN repository, also known as the "RMAN catalog". Similarly, NSR-DBM-ORA stores information about each backup piece in the NetWorker indexes, or what Oracle documentation refers to as the "MML catalog."

During Oracle restores, the following occurs:

◆ The RMAN catalog determines the data to be restored.

◆ The NetWorker indexes provide information that NSR-DBM-ORA requires to perform the restore.

It is important to keep the RMAN catalog and NetWorker indexes synchronized, especially when performing instant backups.

The catalogs are unsynchronized when one of the following exists:

◆ The RMAN catalog contains backup piece entries that do not have corresponding NetWorker index entries.

◆ The NetWorker indexes contain backup piece entries that do not have corresponding RMAN catalog entries.

Note: If catalog synchronization is enabled for NSR-DBM-ORA, proxy backup entries in the catalogs are synchronized automatically.

Extra entries in the catalogs

Extra entries in the NetWorker indexes do not cause problems as long as the extra entries contain unique backup piece names that RMAN does not attempt to reuse for backups.

However, extra entries in the RMAN catalog can cause serious problems. These extra entries can occur when corresponding NetWorker index entries are removed through either expiration or NetWorker commands such as nsrmm. For example, instant backups are often configured to expire quickly (within hours), causing the NetWorker index entries to be removed.

Removing instant backup entries from the NetWorker indexes

Instant backup entries in the NetWorker indexes are removed in one of the following ways:

◆ At the start of an instant backup, if the number of existing instant backups equals the value of the Retain Snapshots attribute in the NetWorker Snapshot Policy resource, the oldest instant backup is automatically expired and its NetWorker index entries are removed.

Catalog synchronization for proxy backups 175


Note: This automatic expiration and index entry removal does not apply to instant backups specified with the nsrnmo -c client_name command.The following sections provide more information on using this command:- “Proxy backups from a virtual cluster client” on page 186- “Proxy backups from a physical cluster client” on page 187

◆ When the expiration policy for an instant backup expires, the NetWorker process nsrim prunes the backup entries from the NetWorker indexes.

◆ The NetWorker process nsrexecd performs consistency checking that can remove backup entries from the NetWorker indexes.

◆ The DBA uses a NetWorker command, such as nsrmm, to remove a save set that includes an instant backup.

Problems with extra entries in the RMAN catalog

When the RMAN catalog contains extra entries (without corresponding entries in the NetWorker indexes), the following types of problems can occur:

◆ When RMAN backup optimization is enabled, RMAN might skip backing up certain files.

◆ The RMAN catalog might expire backups that are required for restores.

◆ RMAN restores might fail when RMAN attempts to restore backup pieces that have no corresponding NetWorker index entries.

The RMAN repository can be stored in either the control file of the target database or in an RMAN recovery catalog. An RMAN recovery catalog is an Oracle database.

Note: For proxy instant backups, use an RMAN recovery catalog instead of a control file. “Control file versus recovery catalog” on page 175 provides more information.

IMPORTANT!To enable automatic catalog synchronization for proxy backups:- The parameter ORACLE_SID must be properly set in the nsrnmo scheduled backup script at the time of the proxy backup. “Customize the nsrnmo script” on page 75 provides details.- An NWORA resource file must include the required resources, as described in “The NWORA resource file” on page 176.

The NSR-DBM-ORA program nsroraclecat uses the NWORA resources in the file to perform automatic synchronization of the RMAN catalog and NetWorker indexes. DBAs can also synchronize the catalogs manually by using RMAN commands.

The following sections provide complete information on how to configure and perform catalog synchronization:

◆ “The NWORA resource file” on page 176

◆ “Automatic catalog synchronization with the nsroraclecat program” on page 183

Control file versus recovery catalog

The control file of an Oracle database can store only a limited number of backup entries. When the maximum number of entries is exceeded, old entries in the control file are overwritten by new ones. You can determine the number of entries in a control



file from the appropriate Oracle dynamic view. The Oracle documentation provides more information.

Instant backups use control file entries of type PROXY COPY. For instant backups, an RMAN recovery catalog (instead of a control file) can be used, since there is no limit on the number of entries a recovery catalog can contain.

IMPORTANT!If you use a control file as the RMAN catalog during an instant backup, ensure that the control file contains enough free entries for the backup. RMAN creates a new entry in the control file for each file backed up in an instant backup. The backup of a large database with many files can quickly use all the free entries in the control file and start overwriting old entries. When entries are overwritten, the corresponding backups cannot be restored.

The NWORA resource file

Proxy backups require the NWORA resource file to exist in the following location on UNIX: /nsr/res/nwora.res

The NWORA resource file is created by the nsrnmoadmin program when it is run for the first time.

To enable instant backups and catalog synchronization, specific NWORA resources must be added to the file with the nsrnmoadmin program.

Note: The NWORA resource file must not be edited manually. All resources in the file must be added, modified, or deleted by using the nsrnmoadmin program only. The nsrnmoadmin program must be run by the root user .

“Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details on the nsrnmoadmin program.

The NWORA resource file for instant backups must contain two types of resources, NWORA parameter resources and NWORA SID resources. The following sections provide more information:

◆ “NWORA parameter resources” on page 176

◆ “NWORA SID resources” on page 178

NWORA parameter resources

An NWORA parameter resource comprises one specific parameter setting, such as for the parameter NSR_ORACLECAT_MODE. For instant backups, the NWORA resource file must contain at least the following parameter resources:

◆ “NSR_NWPATH” on page 177

◆ “NSR_ORACLECAT_DEBUG_FILE” on page 177

◆ “NSR_ORACLECAT_LOG_FILE” on page 177

◆ “NSR_ORACLECAT_MODE” on page 177

◆ “NSR_REMOVE_ON_FAILURE” on page 177



Note: The parameter resources listed in Table 8 on page 177 are the only ones supported. Do not attempt to add other parameter resources to the NWORA resource file.

Using the nsrnmoadmin command to set parameter resourcesWhen the nsrnmoadmin command (with any options) is used for the first time after the NSR-DBM-ORA installation, the NWORA resource file is automatically populated with five parameter resources from Table 8 on page 177: NSR_NWPATH, NSR_ORACLECAT_DEBUG_FILE, NSR_ORACLECAT_LOG_FILE, NSR_ORACLECAT_MODE, NSR_REMOVE_ON_FAILURE. Depending on the nsrnmoadmin command options used, the parameter resources are set to either default or customized values.

Table 8 NWORA parameter resources

Parameter resource Description Default and valid values

NSR_NWPATH Specifies the directory location of the NetWorker binary nsrsnapck.

• Directory pathname for the location of nsrsnapck (default).

• Valid directory pathname for the location of the NetWorker binary nsrsnapck.

NSR_ORACLECAT_DEBUG_FILE Specifies the debug file used by the nsroraclecat program. Set this parameter only for the purpose of debugging the nsroraclecat program.

Note: The nsroraclecat debug file must be created in a secure location since it includes a copy of the strings from the RMAN connection file.

• Undefined (default).• Valid pathname of the nsroraclecat debug

file.

Note: If undefined, debug information is not generated.

NSR_ORACLECAT_LOG_FILE Specifies the operations log file used by the nsroraclecat program. The logged information includes the backup pieces successfully removed from the RMAN catalog, and those that failed to be removed during automatic catalog synchronization.

• Undefined (default).• Valid pathname of the nsroraclecat log file.

Note: If undefined, logging information is written to the /nsr/applogs/nsroraclecat.log file by default.

NSR_ORACLECAT_MODE Specifies whether automatic catalog synchronization is enabled or disabled during proxy backups.

• Undetermined (default).• Enabled.• Disabled.

Note: Instant backups require the resource value to be set to either "enabled" or "disabled". If the value is unset, instant backups fail.

NSR_ORACLE_NLS_LANG Required to enable proxy catalog synchronization in a non-English environment only. Specifies the non-English locale value, as set in the NLS_LANG environment variable. “Configure I18N support” on page 64 provides more information.

• Undetermined (default).• Valid locale value, same as set in the

NLS_LANG environment variable.

Note: If the value is not set to the same value as the NLS_LANG variable in a non-English environment, proxy catalog synchronization fails.

NSR_REMOVE_ON_FAILURE Specifies whether the corresponding NetWorker index entries are removed when the nsroraclecat program fails to remove one or more RMAN catalog entries during automatic catalog synchronization. “Automatic catalog synchronization with the nsroraclecat program” on page 183 provides more information.

• FALSE (default).• TRUE.



Note: Once an NWORA parameter resource is added to the resource file, it cannot be deleted. However, its value can be modified.

To view the NWORA parameter resources in the resource file, use the nsrnmoadmin -r list command.

To modify NWORA parameter resource settings, use the nsrnmoadmin -r update command.

“Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details on how to use the nsrnmoadmin command.

Example 36 Default NWORA parameter resources

After the NSR-DBM-ORA installation, if the first nsrnmoadmin command used is nsrnmoadmin -r list (to list the NWORA resource file contents), the command adds the following default settings for the NWORA parameter resources to the resource file:

NSR_NWPATH=NetWorker_binary_pathNSR_ORACLECAT_MODE=undeterminedNSR_REMOVE_ON_FAILURE=FALSENSR_ORACLE_LOG_FILE=NSR_ORACLECAT_DEBUG_FILE=

NetWorker_binary_path is the pathname of the directory that contains the NetWorker binary nsrsnapck.

To enable instant backups, NSR_ORACLECAT_MODE must be set to either enabled or disabled by using the nsrnmoadmin -r update command.

This default NWORA resource file does not yet contain any NWORA SID resources, as described in “NWORA SID resources” on page 178.

NWORA SID resources

An NWORA SID resource comprises a specific group of parameters for a single Oracle database. If automatic catalog synchronization is enabled (NSR_ORACLECAT_MODE is set to enabled), the NWORA resource file must contain an NWORA SID resource for each Oracle database (ORACLE_SID). The NWORA SID resource can include only the parameters described in Table 9 on page 179.

However, an unlimited number of NWORA SID resources can be added to the resource file.

IMPORTANT!If automatic catalog synchronization is enabled, but you do not create an NWORA SID resource for an Oracle database, the catalogs will not be synchronized during instant backups of that database. As a result, the catalogs can become unsynchronized unless you synchronize them manually by using RMAN commands. “Automatic catalog synchronization with the nsroraclecat program” on page 183 provides more information.

Note: Each NWORA SID resource must have a unique NSR_ORACLE_SID value.



Using the nsrnmoadmin command to set SID resourcesTo add an NWORA SID resource to the resource file, use the nsrnmoadmin -r add command.

To modify NWORA SID resource settings, use the nsrnmoadmin -r update command.

“Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details on how to use the nsrnmoadmin command.

Example 37 NWORA SID resource

Automatic catalog synchronization is enabled for instant backups when the NSR_ORACLECAT_MODE parameter resource is set to enabled. Prior to performing instant backups of an Oracle database with an ORACLE_SID value of proddb, add an NWORA SID resource to the resource file by using the nsrnmoadmin -r add command. The SID resource must include the following:

◆ NSR_ORACLE_SID set to proddb.

◆ NSR_ORACLE_CONNECT_FILE and NSR_ORACLE_HOME set to suitable values.

Table 9 NWORA SID resource components


NSR_ORACLE_CONNECT_FILE Mandatory. Specifies the location of the file containing the connection strings required to create an RMAN session. “The connection file” on page 180 provides more information.

• Undefined (default).• Valid pathname of the RMAN connection file.

NSR_ORACLE_HOME Mandatory. Specifies the home directory of the Oracle installation. The RMAN executable must be located in subdirectory bin of this directory.

• Undefined (default).• Valid pathname of the Oracle home directory.

Note: The value must be equal to the Oracle parameter $ORACLE_HOME value.

NSR_ORACLE_LIB_PATH Optional. Specifies the pathname of the directory containing the Oracle shared libraries on UNIX systems, typically $ORACLE_HOME/lib.

• Undefined (default).• Valid pathname of the Oracle shared library

directory on UNIX.

NSR_ORACLE_SID Mandatory. Specifies the SID value of the Oracle database whose RMAN catalog is to be synchronized.

• Undefined (default).• Valid SID value of the Oracle database.

Note: The value must be equal to the ORACLE_SID value in the particular nsrnmo script used for the database backup. “ORACLE_SID” on page 77 provides more information.

NSR_ORACLE_TNS_ADMIN Optional. Specifies the pathname of the directory containing the Oracle Net configuration files.

• Undefined (default).• Valid pathname of Oracle network

configuration directory.

Note: The value must be equal to the Oracle parameter $TNS_ADMIN value.



◆ NSR_ORACLE_LIB_PATH and NSR_ORACLE_TNS_ADMIN (optional) set to suitable values.

The following NWORA SID resource can be added for the Oracle database:

NSR_ORACLE_CONNECT_FILE=/dbapps/proddb/connect.fileNSR_ORACLE_HOME=/dbapps/proddb/app/oracle/product/10.2.0/Db_1NSR_ORACLE_LIB_PATH=/usr/libNSR_ORACLE_SID=proddbNSR_ORACLE_TNS_ADMIN=/dbapps/proddb/tns

In this example, the RMAN connection file is /dbapps/proddb/connect.file and the Oracle home directory is /dbapps/proddb/app/oracle/product/10.2.0/Db_1.

The connection file

Catalog synchronization requires the connection file for an Oracle database. The nsroraclecat program uses the information in the connection file to delete RMAN catalog entries.

In the NWORA SID resource for the target database, the parameter NSR_ORACLE_CONNECT_FILE must be set to the pathname of the connection file. “NWORA SID resources” on page 178 provides more information.

IMPORTANT!A DBA must create the connection file in a secure location.

The connection file must include the following:

◆ The connection string that is required to connect to the target database.

◆ If an RMAN recovery catalog is used, the connection string that is required to connect to the RMAN recovery catalog.

Note: The connection file must not include any lines starting with the # symbol.

If the connection file does not contain a connection string for an RMAN recovery catalog, the nsroraclecat program assumes that a control file is used as the RMAN repository during instant backups.

Example 38 Connection file contents

If the following lines exist in the connection file, an RMAN recovery catalog is used as the RMAN repository:

connect target sys/oracle@proddb;connect rcvcat rman/rman@oracat;

Note: RMAN catalog deletions fail if the connection file for a backup piece does not exist or does not contain valid connection strings.

Configuring the NWORA resource file with the nsrnmoadmin program

All resources in the NWORA resource file must be added, modified, or deleted by using the nsrnmoadmin program only.

To run the program, type the nsrnmoadmin command at the operating system command line, as the root user .

“The nsrnmoadmin command syntax and options related to proxy backups” on page 181 provides details on the command syntax and options.



The nsrnmoadmin command syntax and options related to proxy backupsThe nsrnmoadmin command syntax and options used to configure proxy backup settings are as follows:

nsrnmoadmin [-D debug_level] -r list [ResourceName | SidName]

nsrnmoadmin [-D debug_level] -r add ResourceName ResourceValue

nsrnmoadmin [-D debug_level] -r add sid=SidName home=OracleHome connect=ConnectFilePath [lib=LibraryPath] [tns=TNSPath]

nsrnmoadmin [-D debug_level] -r update ResourceName ResourceValue

nsrnmoadmin [-D debug_level] -r update sid=SidName [home=OracleHome] [connect=ConnectFilePath] [lib=LibraryPath] [tns=TNSPath]

nsrnmoadmin [-D debug_level] -r delete SidName

where:

◆ debug_level is the level of debug information generated.

◆ ResourceName is the name of an NWORA parameter resource.

◆ SidName is the value of the NSR_ORACLE_SID parameter of an NWORA SID resource.

◆ ResourceValue is the value of the NWORA parameter resource.

◆ OracleHome is the value of the NSR_ORACLE_HOME parameter of the NWORA SID resource.

◆ ConnectFilePath is the value of the NSR_ORACLE_CONNECT_FILE parameter of the NWORA SID resource.

◆ LibraryPath is the value of the NSR_ORACLE_LIB_PATH parameter of the NWORA SID resource.

◆ TNSPath is the value of the NSR_ORACLE_TNS_ADMIN parameter of the NWORA SID resource.

Only the -D and -r options are supported:

◆ The -D option causes the nsrnmoadmin command to print debug information.

◆ The -r option must be followed by the appropriate keywords, which determine the NWORA resource operation to be performed.

Command options and settings in brackets ([ ]) are optional. Do not include the brackets when typing the command.

The following sections provide examples of how to use the nsrnmoadmin command to list, add, update, and delete NWORA resources:

◆ “List the NWORA resources” on page 181

◆ “Add the NWORA resources” on page 182

◆ “Update the NWORA resources” on page 182

◆ “Delete the NWORA SID resources” on page 182

For more information on the nsrnmoadmin command, refer to the nsrnmoadmin man page or the nsrnmoadmin entry in Appendix D.

List the NWORA resources

◆ To display the entire NWORA resource file contents, type the following:

nsrnmoadmin -r list

◆ To display the NSR_ORACLECAT_MODE parameter resource only, type the following:



nsrnmoadmin -r list NSR_ORACLECAT_MODE

◆ To display an NWORA SID resource with the NSR_ORACLE_SID value of proddb, type the following:

nsrnmoadmin -r list proddb

Add the NWORA resources

◆ To add the NSR_ORACLECAT_MODE parameter resource with the value of enabled, type one of the following:

nsrnmoadmin -r add NSR_ORACLECAT_MODE enabled

nsrnmoadmin -r add NSR_ORACLECAT_MODE=enabled

Note: If the NWORA parameter resource already exists in the resource file, use of the add keyword causes the resource value to be updated.

◆ To add a new NWORA SID resource with the NSR_ORACLE_SID value of proddb and other values as specified in Example 37 on page 179, type the following:

nsrnmoadmin -r add sid=proddb home=/dbapps/proddb/app/oracle/product/10.2.0/Db_1 connect=/dbapps/proddb/connect.file lib=/usr/lib tns=/dbapps/proddb/tns

Note: - When adding an NWORA SID resource, the keywords sid, home, and connect are mandatory; the keywords lib and tns are optional.- If an NWORA SID resource with the same NSR_ORACLE_SID value already exists, the command updates the values of the existing resource.

Update the NWORA resources

◆ To update the value of the NSR_ORACLECAT_MODE parameter resource to enabled, type one of the following:

nsrnmoadmin -r update NSR_ORACLECAT_MODE enabled

nsrnmoadmin -r update NSR_ORACLECAT_MODE=enabled

◆ To update the values of the parameters NSR_ORACLE_HOME and NSR_ORACLE_CONNECT_FILE in an NWORA SID resource with the NSR_ORACLE_SID value of proddb, type the following:

nsrnmoadmin -r update sid=proddb home=/dbapps/proddb/10.2.0/Db_1 connect=/dbapps/oracle/connect/proddb.connect

Note: When updating an NWORA SID resource, the keyword sid is mandatory. The keywords home, connect, lib, and tns are optional.

Delete the NWORA SID resources

To delete an NWORA SID resource with the NSR_ORACLE_SID value of proddb, type the following:

nsrnmoadmin -r delete proddb

Note: Only NWORA SID resources can be deleted from the resource file. NWORA parameter resources cannot be deleted.



Automatic catalog synchronization with the nsroraclecat program

Automatic catalog synchronization is managed jointly by NetWorker server and NSR-DBM-ORA programs. To remove instant Oracle backup entries from the NetWorker indexes, the NetWorker server invokes the nsrsnapck program. Prior to removing the index entries, nsrsnapck invokes the nsroraclecat program to remove the corresponding RMAN catalog entries.

Note: To perform manual catalog synchronization, you can use specific RMAN commands, as described in “The change...crosscheck and crosscheck commands” on page 200. The appropriate Oracle documentation provides more information on RMAN commands.

Review the following information on automatic catalog synchronization:

◆ “RMAN catalog entry removals with nsroraclecat” on page 183

◆ “Failure of the nsroraclecat program” on page 183

◆ “NetWorker index entry removals with nsrsnapck” on page 184

RMAN catalog entry removals with nsroraclecat

The nsroraclecat program runs on the Oracle Server host that performed the instant backup:

◆ Do not attempt to run the nsroraclecat program manually.

◆ The nsroraclecat program is run automatically by the nsrsnapck program.

◆ Only one nsroraclecat program can run at a time. If two nsroraclecat programs are started, the one started first completes its operation before the second one proceeds.

To remove the RMAN catalog entries, nsroraclecat obtains information from the NWORA resource file and generates temporary RMAN scripts that include an RMAN change...delete command for each backup piece to be removed.

A separate script is created for all the backup pieces from the same database (or ORACLE_SID).

The nsroraclecat program names each RMAN script as follows:

/nsr/tmp/.nworapc/nsroraclecat_date_pidThe nsroraclecat program runs each script in an RMAN session. After the scripts have finished running, the program removes them.

Note: The nsroraclecat program generates information about the backup piece entries removed from the RMAN catalog. The information is written to the nsroraclecat log and debug files. “NSR_ORACLECAT_LOG_FILE” and “NSR_ORACLECAT_DEBUG_FILE” on page 177 provide more information on these files.

For more information on the nsrnmoadmin command, refer to the nsrnmoadmin man page or the nsrnmoadmin entry in Appendix D.

Failure of the nsroraclecat program

A fatal error that causes nsroraclecat to fail can be produced by the following:

◆ The nsrsnapck program passes invalid information to nsroraclecat, for example, an invalid NetWorker client name or an invalid save time of a backup piece.

◆ The nsroraclecat program cannot connect to the NetWorker server to query the NetWorker indexes.



◆ The nsroraclecat program cannot locate the required backup pieces in the NetWorker indexes.

To diagnose the cause of a nsroraclecat program failure, review the nsroraclecat log files specified by NSR_ORACLECAT_DEBUG_FILE and NSR_ORACLECAT_LOG_FILE. The operations log file is /nsr/applogs/nsroraclecat.log by default.

If the nsroraclecat program fails, the nsrsnapck program removes the corresponding NetWorker index entries by using the procedures described in “NetWorker index entry removals with nsrsnapck” on page 184.

The following files (if they exist) need to be removed:

◆ Files in the directory: /nsr/tmp/.nworapc

◆ Files in the temporary directory /tmp, where the files have the name nwora_bp_sid_pid:

• sid is an ORACLE_SID value.

• pid is a nsroraclecat process ID.

Note: If nsroraclecat fails continuously, disable catalog synchronization (by setting NSR_ORACLECAT_MODE to disabled) until the cause of the problem is determined.

IMPORTANT!After a nsroraclecat program failure occurs or while catalog synchronization is disabled, the DBA must synchronize the catalogs manually by using specific RMAN commands. The appropriate Oracle documentation provides more information.

NetWorker index entry removals with nsrsnapck

Once the nsroraclecat program has finished the RMAN catalog operations, the nsrsnapck program removes the NetWorker index entries for all the backups that were successfully removed from the RMAN catalog.

If some of the backup entries failed to be removed from the RMAN catalog, the nsrsnapck program does the following:

◆ Removes the corresponding NetWorker index entries when NSR_REMOVE_ON_FAILURE is set to TRUE.

◆ Does not remove the corresponding NetWorker index entries when NSR_REMOVE_ON_FAILURE is set to FALSE.

Note: When NSR_REMOVE_ON_FAILURE is set to FALSE, nsrsnapck removes only those NetWorker index entries that correspond to removed RMAN catalog entries.

IMPORTANT!The NSR_REMOVE_ON_FAILURE setting controls the result of the nsroraclecat program failure to remove RMAN catalog entries.- In general, NSR_REMOVE_ON_FAILURE should be set to TRUE, to enable NetWorker index entries to be removed, even if the RMAN catalog entries are not removed. Otherwise, if entries are not removed from the NetWorker indexes, the snapshot resources are not freed and subsequent backups might fail.

Proxy backups and restores on cluster systems 185


- If RMAN backup optimization is enabled, NSR_REMOVE_ON_FAILURE should be set to FALSE, to prevent the removal of NetWorker index entries. Otherwise, RMAN might skip backing up certain files.

When the nsroraclecat program fails to run properly, the nsrsnapck program’s actions depend on whether the instant backup on the primary storage is intact:

◆ If the instant backup on the primary storage is destroyed or invalid, the nsrsnapck program removes the corresponding entry from the NetWorker indexes.

◆ If the instant backup on the primary storage is intact, the nsrsnapck program does not remove any entries from the NetWorker indexes and generates an error message about the failure in the /nsr/logs/daemon.raw file:

The NetWorker Administrator’s Guide provides more information on the daemon.raw log file and how to view its contents.

Proxy backups and restores on cluster systemsNSR-DBM-ORA can perform proxy backups and restores of an Oracle database configured on a cluster system. The software also enables cluster failover for proxy backups.

IMPORTANT!The parameter NSR_CLIENT is not supported for proxy backups on a cluster system. This parameter is used for restores and regular Oracle backups on a cluster system, as described in Chapter 6, “Cluster and RAC Systems.”

Review the following information, specific to proxy operations on a cluster system:

◆ “Proxy backup failover” on page 185

◆ “Proxy backups from a virtual cluster client” on page 186

◆ “Proxy backups from a physical cluster client” on page 187

◆ “Restores from proxy backups on a cluster system” on page 188

Proxy backup failover

During a scheduled proxy backup where the Oracle software is configured to fail over, the NetWorker server retries the backup on the failover node if the Client Retries attribute is set to a nonzero value in the Group resource. The retry occurs at the RMAN script level, whereby the RMAN script restarts from the beginning.

Note: To avoid restarting the backups of all objects in the RMAN script during the NetWorker retry, you can use the Oracle restartable backups feature. This feature enables you to back up only the files that have not been backed up since a specified time, for example, by using the ’sysdate -1’ option. “Restartable backups” on page 34 provides more information.



Proxy backups from a virtual cluster client

A proxy backup from a virtual cluster client (virtual host) protects Oracle data on shared cluster disks.

To set up a proxy backup from a virtual cluster client:

1. Install the NSR-DBM-ORA software on each physical node of the cluster, along with the NetWorker client and appropriate PowerSnap Module software.

2. Create a NetWorker Client resource for the virtual host and each physical host, as described in “Configuring the NetWorker Client resource” on page 163:

• In the Remote Access attribute in the Client resource for a virtual cluster client, specify the Oracle user from each physical client that can store and retrieve backups.

• In the Save Set attribute, specify the complete pathname of the RMAN script to back up the Oracle data on the shared disk.

3. Configure the other NetWorker resources required for proxy backups, as described in “Additional configurations for proxy operations” on page 157:

• To enable backup failover, specify a nonzero value in the Client Retries attribute in the NetWorker Group resource for the scheduled backup. This value causes the NetWorker server to restart the backup of uncompleted Oracle save sets on the failover node.

• Specify other recommended attribute settings in the Group resource, as described in the cluster support information of the NetWorker Administrator’s Guide.

4. Configure the NWORA resource file on each node of the cluster, as described in “Configuring the NWORA resource file” on page 158.

5. If the Save Set attribute in the Client resource for a virtual client specifies an RMAN script that resides on a private disk, ensure that an empty pathownerignore file exists on the Oracle Server host as NetWorker_install_path/bin/pathownerignore

Here, NetWorker_install_path is the root directory of the NetWorker installation path.

Without the pathownerignore file on the Oracle Server, the scheduled backup fails due to the path-ownership rules used to distinguish between data associated with a physical host and data managed by a virtual client.

6. To specify that the proxy backup entries be stored in a NetWorker client file index other than the virtual client index, for example, in a physical client index:

• Add -c client_name to the nsrnmo script name in the Backup Command attribute in the Client resource.

• Specify the Oracle user from the virtual host in the Remote Access attribute in the Client resource for client_name.

The expiration of instant backups created with the nsrnmo -c client_name command differs from the expiration of instant backups created without the -c option. “Removing instant backup entries from the NetWorker indexes” on page 174 provides details on the expiration and removal of backups specified with nsrnmo -c client_name.

Proxy backups and restores on cluster systems 187


Notes:

• The host specified with -c client_name must have access to instant backups.

• NSR-DBM-ORA and the PowerSnap Module must be installed and configured on the host specified with -c client_name.

• When the backup is started from the virtual cluster client, the backup entries are stored in the NetWorker client file index of the virtual client by default.

• The entries for the NWORA resource file backup are always stored in the NetWorker index of the physical client.

Example 39 Proxy backup entries in the index of a physical cluster client

To specify that the backup entries be stored in the index of the physical cluster client mars.fujitsu.com , where the script name is nsrnmo, specify the following in the Backup Command attribute:

nsrnmo -c mars.fujitsu.com

Proxy backups from a physical cluster client

A proxy backup from a physical cluster client protects Oracle data on private disks. This type of backup is similar to a regular scheduled Oracle backup on a non-cluster system.

The following sources provide information on how to set up a proxy backup from a physical cluster client:

◆ “Proxy backup requirements” on page 164

◆ NetWorker Administrator’s Guide (chapter on cluster support)

When the backup is started from the physical client, the backup entries are stored in the NetWorker index of the physical client by default.

Note: The entries for the NWORA resource file backup are always stored in the NetWorker index of the physical client.

To specify that the proxy backup entries be stored in a NetWorker client file index other than the physical client index, for example, in a virtual client index:

◆ Add -c client_name to the nsrnmo script name in the Backup Command attribute in the Client resource.

◆ Specify the Oracle user from the physical host in the Remote Access attribute in the Client resource for client_name.

The expiration of instant backups created with the nsrnmo -c client_name command differs from the expiration of instant backups created without the -c option. “Removing instant backup entries from the NetWorker indexes” on page 174 provides details on the expiration and removal of backups specified with nsrnmo -c client_name.

Notes:

◆ The host specified with -c client_name must have access to instant backups.

◆ NSR-DBM-ORA and the PowerSnap Module must be installed and configured on the host specified with -c client_name.



Example 40 Proxy backup entries in the index of a virtual cluster client

To specify that the backup entries be stored in the index of the virtual client monalisa.fujitsu.com , where the script name is nsrnmo, specify the following in the Backup Command attribute:

nsrnmo -c monalisa.fujitsu.com

Restores from proxy backups on a cluster system

To set up a restore from a proxy backup on a cluster system, perform the following:

1. Set the parameter NSR_CLIENT to the correct value by using one of the methods in “Setting the NSR* parameters” on page 190:

• To restore a backup from a virtual cluster client, set NSR_CLIENT to the name of the virtual cluster client.

• To restore a backup from a physical cluster client, set NSR_CLIENT to the name of the physical cluster client.

2. In the Remote Access attribute of the Client resource, specify the hostname of the client on which the restore is to be started.

Note: When a failover occurs during a restore, the restore must be restarted manually on the failover node.

Parameters in an RMAN Session 189

A

This appendix includes the following sections:

◆ Setting the NSR* parameters ..................................................................................... 190◆ NSR* parameter definitions....................................................................................... 191

Parameters in an RMANSession


Parameters in an RMAN Session

Setting the NSR* parametersThis appendix describes the NSR* parameters that can be set in the RMAN command or script for NSR-DBM-ORA backups and restores.

Note: Unless noted otherwise, the parameters are supported for both regular and proxy backups and restores. “Setting the parameters” on page 160 provides information specific to proxy operations.

The NSR* parameters described in Table 10 on page 191 are parameters specific to NetWorker that can be set for an RMAN session of:

◆ A manual or scheduled Oracle backup.

◆ An Oracle restore.

To ensure that the required NSR* parameters are set during an NSR-DBM-ORA backup or restore, set them in one of the following ways unless specified otherwise in Table 10 on page 191:

◆ If using automatic channels, set the parameters with the parms option in the configure channel command.

“Automatic channel allocation” on page 24 provides more information on automatic channels.

◆ If not using automatic channels, set the parameters with the RMAN send command (recommended), as one of the following:

• The rman send command on the operating system command line.

• The send command in the RMAN session or script.

Do not mix these different ways of setting the NSR* parameters in the same RMAN session. The use of a setenv command on the operating system command line to set the NSR* parameters has no effect.

NSR* parameter definitions 191


NSR* parameter definitionsTable 10 on page 191 lists all the NSR* parameters to set in the RMAN session. For each parameter, the table includes:

◆ A description of the parameter.

◆ The default value of the parameter.

◆ The valid values that can be assigned to the parameter.

Table 10 Parameters in the RMAN command or script (page 1 of 7)


NSR_AES_ENCRYPTION Optional. Specifies whether the Oracle data is encrypted by NSR-DBM-ORA by using 256-bit AES encryption during the backup before being sent over the network. The AES encryption uses the key that is set in the Datazone pass phrase attribute of the NetWorker Server resource. Keep a record of this key or pass phrase because it is required for later restore of the backup.

Note: NSR_AES_ENCRYPTION must be set through the send command only.

Record all keys (pass phrases) used for 256-bit AES encryption. Be careful when you change the pass phrase on the NetWorker server. If the pass phrase on the server is changed and you cannot remember the pass phrase originally used for an NSR-DBM-ORA backup, the encrypted data cannot be recovered. The NetWorker Administrator’s Guide provides more information on pass phrases.

• FALSE (default) = The Oracle data is not encrypted through 256-bit AES encryption during the backup.

• TRUE = The Oracle data is encrypted through 256-bit AES encryption during the backup.

NSR_CHECKSUM Optional. Specifies whether or not the NetWorker software performs checksumming on the backup data.

• FALSE (default) = The NetWorker software performs no checksumming.

• TRUE = The NetWorker software performs checksumming.

NSR_CLIENT Recommended for a backup of Oracle running on a cluster or RAC system.Recommended for all restores.Specifies the NetWorker Client resource to use for a backup or restore session.Might be needed for disaster recovery.

• The physical host on which the session runs (default).

• A valid NetWorker client hostname.

NSR_COMPRESSION Optional. Specifies whether or not the NetWorker software performs compression on the backup data.

• FALSE (default) = The NetWorker software performs no compression.

• TRUE = The NetWorker software performs compression.

• gzipn = The NetWorker software performs compression with gzip. Compression level is n. n can be in the range of 1 to 9.

• bzipn = The NetWorker software performs compression with bzip. Compression level is n. n can be in the range of 1 to 255.



NSR_DATA_VOLUME_POOL Mandatory if using the set duplex command (with duplex set to 1, 2, 3, or 4) or other RMAN commands to generate backup copies during a manual backup.Specifies the name of the volume pool to use for the Oracle backup, including a duplexed Oracle backup.In the case of proxy backups, specifies the volume pool for live backups only (backups to secondary storage only).

Note: For a scheduled backup, the NSR_DATA_VOLUME_POOL setting takes precedence over any pool associated with the scheduled backup group. For a manual backup that generates backup copies, NSR_DATA_VOLUME_POOL must be set through the parms option, not through the send command or option.

• The most appropriate pool, selected by the NetWorker server (default).

• A valid NetWorker pool name.• The name must be different from the name used by

the parameter NSR_DATA_VOLUME_POOL1, NSR_DATA_VOLUME_POOL2, or NSR_DATA_VOLUME_POOL3.

NSR_DATA_VOLUME_POOL1 Mandatory if using the set duplex command (with duplex set to 2, 3, or 4) or other RMAN commands to generate two or more backup copies during a manual backup.Specifies the name of the volume pool to use for a duplexed Oracle backup.

Note: For a manual backup that generates backup copies, NSR_DATA_VOLUME_POOL1 must be set through the parms option, not through the send command or option.

• Undefined (default).• A valid NetWorker pool name.• The name must be different from the name used by

the parameter NSR_DATA_VOLUME_POOL, NSR_DATA_VOLUME_POOL2, or NSR_DATA_VOLUME_POOL3.

NSR_DATA_VOLUME_POOL2 Mandatory if using the set duplex command (with duplex set to 3 or 4) or other RMAN commands to generate three or more backup copies during a manual backup.Specifies the name of the volume pool to use for a duplexed Oracle backup.




NSR_DATA_VOLUME_POOL3 Mandatory if using the set duplex command (with duplex set to 4) or other RMAN commands to generate four backup copies during a manual backup.Specifies the name of the volume pool to use for a duplexed Oracle backup.








NSR_DEBUG_FILE Optional. Enables debugging by specifying the complete pathname of the log file on the Oracle Server that receives debug information during a proxy or regular backup or restore.Only used for debugging of the NSR-DBM-ORA media management library (MML).

Note: NSR_DEBUG_FILE must be set through the parms option. Do not set it through the send option or command.

• Undefined (default).• The valid pathname of the debug file to use on the

Oracle Server.

Note: If undefined, no debug information is generated. If a file with the specified pathname cannot be created, debug information is either directed to /nsr/applogs/nmosbt.log or not generated.

NSR_DEBUG_LEVEL Optional. Controls the amount of debug information that NSR-DBM-ORA writes to the debug log file, which is specified by NSR_DEBUG_FILE.

Note: NSR_DEBUG_LEVEL must be set through the parms option. Do not set it through the send option or command.

The debug level limits only the amount of debug information generated by the three functions sbtread(), sbtwrite(), and sbtpcstatus(), which are called most often during backups and restores. All the debug messages for other functions are always written to the log file, regardless of the debug level.

• 0 (default) = Only operational messages are written for the sbtread(), sbtwrite(), and sbtpcstatus() functions. No entry or exit messages are written for the functions.

• 1 = Only entry and exit messages are written for the sbtread(), sbtwrite(), and sbtpcstatus() functions. No operational messages are written for the functions.

• 2 - 9 = All debug messages for all the functions are written to the log file.

NSR_DEDUP_BACKUP Mandatory for a manual deduplication backup. Do not set for a scheduled deduplication backup. Specifies whether or not deduplication is performed during a manual NSR-DBM-ORA backup.

Note: For a scheduled deduplication backup, set the De-duplication Backup attribute in the NetWorker Client resource, instead of setting this parameter.

• FALSE (default) = Deduplication is not performed during a manual NSR-DBM-ORA backup. Other NSR_DEDUP* parameters are ignored.

• TRUE = Deduplication is performed during a manual NSR-DBM-ORA backup. If NSR_DEDUP_NODE is not set, the backup fails.

NSR_DEDUP_CACHE_ENABLED Optional for a deduplication backup. Specifies whether a hash cache is used during a deduplication backup.The nsravtar process creates the cache in /nsr/dedup/cache. Use of the cache increases both the deduplication backup performance and disk usage in the cache directory.

Note: Setting of this parameter requires knowledge of the potential effects on Avamar server operations.

• TRUE (default) = A hash cache is used to increase performance during a deduplication backup. This value is recommended in most cases.

• FALSE = A hash cache is not used during a deduplication backup. The parameter NSR_DEDUP_CACHE_TAG is ignored.

NSR_DEDUP_CACHE_TAG Mandatory for a deduplication backup if the parameter NSR_DEDUP_CACHE_ENABLED is set to TRUE. Specifies the tag for the nsravtar process to use to generate the hash cache name for a deduplication backup:• Set to a different value for each channel; if the same

tag value is used for more than one channel, the deduplication backup fails.

•

• / (default; signifies the root directory).• String value of the tag to use to generate a

deduplication cache name. Do not include the client name in the value.

• Recommended value is:<ORACLE_SID or Net_service_name>_<channel_ID> For example: ORCL102_t1

Note: The nsravtar process uses the tag value to generate the cache name through hashing. The actual cache name does not contain this parameter value.





NSR_DEDUP_CHUNK_SIZE Optional for a deduplication backup. Specifies the size in bytes that the Avamar server uses for data chunks in a deduplication backup.If a nonzero value is specified, the Avamar server uses the fixed size for all of the data chunks saved in the deduplication backup.

Note: Setting of this parameter requires knowledge of the potential effects on Avamar server operations.

• 0 (default; signifies that variable sizes are used for the data chunks, as determined by the Avamar server). This value is recommended in most cases.

• Size (greater than zero) in bytes to use for all of the data chunks in a deduplication backup; for example, 1024, 2048, 5096, 8194, or a value recommended in the Avamar documentarion.

NSR_DEDUP_NODE Mandatory for a manual deduplication backup. Do not set for a scheduled deduplication backup. Specifies the hostname of the Avamar server to perform a manual NSR-DBM-ORA backup and store the deduplicated client data. The hostname must be the same as the Avamar server hostname set in the De-duplication Node resource.

Note: For a scheduled deduplication backup, set the De-duplication Node attribute in the NetWorker Client resource, instead of setting this parameter.

• Undefined (default).• Avamar server hostname set in the NetWorker

De-duplication Node resource.

NSR_DPRINTF Optional, for use with Tech Support only. Specifies whether NetWorker core debug messages are printed to the file specified by NSR_DEBUG_FILE.

Note: NSR_DPRINTF must be set through the parms option, not through the send command or option.Recommended value for NSR_DEBUG_LEVEL is 9.

• FALSE (default) = NetWorker core debug messages are not printed to NSR_DEBUG_FILE.

• TRUE = NetWorker core debug messages are printed to NSR_DEBUG_FILE.- If NSR_DEBUG_FILE is not set, the messages are printed to the default location.- If NSR_DEBUG_LEVEL is set to 0, few debug messages are printed.

NSR_ENCRYPTION_PHRASES Optional. Specifies one or more encryption phrases to use for decrypting Oracle data during an NSR-DBM-ORA restore. If this parameter is not set, the NSR-DBM-ORA restore obtains the encryption phrase from the NetWorker server.If both of the following are true, set this parameter to the phrase used to originally back up the data:• The Oracle data that is being restored was backed

up with 256-bit AES encryption.• The encryption phrase on the NetWorker server has

changed since the Oracle data was backed up.

Note: NSR_ENCRYPTION_PHRASES must be set through the RMAN send command only.

• Undefined (default).• One or more encryption phrases to use during an

Oracle restore with NSR-DBM-ORA. Each phrase must be a string enclosed in quotes. Multiple phrases must be separated by commas, and the entire group of phrases surrounded by outer quotes that are different from the inner quotes. For example:NSR_ENCRYPTION_PHRASES="’key1’,’key2’"- NSR-DBM-ORA itself supports double ("), single (’), andbackward (‘) quotes.

- Certain shells might not support certain types of quotes, for example, when the parameter is setwith the send command on the command line.

- Oracle does not support the use of single quoteswithin an RMAN script.

Note: If undefined, the NSR-DBM-ORA restore obtains the encryption phrase from the Datazone pass phrase attribute of the NetWorker Server resource, as described in “Datazone pass phrase” on page 59.





NSR_GROUP Optional. Use for a manual backup only. Specifies the NetWorker group to use for a backup session. Used by the NetWorker server to select the volume pool for the manual backup.

• The predefined NetWorker group named Default (default).

• A valid NetWorker backup group name.

Note: A NetWorker Group resource must be configured for the specified backup group.

NSR_MMDB_RETRY_TIME Optional. Specifies the number of minutes that NSR-DBM-ORA should try to connect to the NetWorker media database before terminating the operation (backup, restore, or RMAN maintenance commands). When the media database is busy, NSR-DBM-ORA tries to reconnect after sleeping for five seconds between attempts.

• 0 (default). NSR-DBM-ORA does not try to reconnect to the media database if the first attempt fails.

• A valid number of minutes.

NSR_NO_BUSY_ERRORS Optional. Specifies whether a scheduled backup fails immediately when the NetWorker server is busy or waits for the NetWorker server to accept the connection.

Note: For proxy backups, this parameter is always set to FALSE.

• FALSE (default) = The scheduled backup waits for the NetWorker server to accept the connection.

• TRUE = The scheduled backup fails immediately when the NetWorker server is busy.

NSR_NO_MULTIPLEX Optional. When set for a specific RMAN channel, specifies whether multiplexing is disabled during a backup on the NetWorker device that the RMAN channel is using. If multiplexing is disabled, no other save sets can be written to the device.To optimize restore operations, RMAN requires NSR-DBM-ORA backups to not be multiplexed.Setting the parameter to TRUE may affect the backup performance. For example, the device may sit idle during part of the backup. If the performance is adversely affected, reset the parameter to FALSE.

• FALSE (default) = Multiplexing is enabled on the device that the RMAN channel is using.

• TRUE = Multiplexing is disabled on the device that the RMAN channel is using.

Note: If a random access NetWorker device such as an advanced file device is used, do not set the parameter to TRUE.

NSR_NWPATH Mandatory when removing a backup from the NetWorker server if both of the following are true:• The nsrsnapck binary is in a nondefault location.• The NWORA resource file does not contain the

correct location of the NetWorker client binaries. “The NWORA resource file” on page 176 provides details on the resource file.

Mandatory for a manual deduplication backup or restore and for a proxy backup or restore when NetWorker software is installed in a nondefault location on a UNIX NSR-DBM-ORA client host.Specifies the complete pathname of the directory that contains nsrsnapck or NetWorker binaries.

• The platform-specific default location of the NetWorker client binaries (default).

• A valid pathname of the directory that contains the nsrsnapck binary, for example, /usr/local/bin.

NSR_PROXY_PFILE Supported only for a proxy backup or restore.Mandatory if setting PowerSnap parameters in a configuration file.Specifies the complete pathname of a configuration file that contains PowerSnap parameter settings for a proxy backup or restore.

• Undefined (default).• A valid pathname of the configuration file.

Note: If undefined or an invalid pathname, parameter settings in the preferred configuration file are ignored.





NSR_RECOVER_POOL Optional for restore only. Specifies the name of the NetWorker volume pool to use for an NSR-DBM-ORA restore. You can use this option to restore data from a specified volume pool if there are multiple copies (clones) of the backup on different volume pools.

Note: This parameter cannot be used for proxy restores.

• Undefined (default).• A valid name of a NetWorker volume pool that

contains a cloned backup to use for a restore.

NSR_RETENTION_DISABLED Optional. Specifies whether the NetWorker browse and retention policies are disabled.Set to TRUE to use Oracle policies only (not NetWorker policies) to manage the backup data lifecycle. Then the RMAN catalog and NetWorker indexes cannot become unsynchronized, for example, when a NetWorker index entry is expired but the corresponding RMAN catalog entry is not expired.

• FALSE (default) = NetWorker browse and retention policies are enabled, and used to manage the lifecycle of the NSR-DBM-ORA backup data.

• TRUE = NetWorker browse and retention policies are disabled. Only Oracle policies are used to manage the lifecycle of the NSR-DBM-ORA backup data.

NSR_SAVESET_BROWSE Mandatory if setting a nondefault browse policy for a manual Oracle backup.Specifies the date when the entry for the backup is to be removed from the NetWorker client index.

Note: NSR_SAVESET_BROWSE must be set through the send command only. For a scheduled Oracle backup, this parameter overrides the Browse Policy attribute setting in the NetWorker Client resource.

• The most appropriate policy for the client, as selected by the NetWorker server (default).

• A valid date in nsr_getdate(3) format.

NSR_SAVESET_RETENTION Mandatory if setting a nondefault retention policy for a manual Oracle backup.Specifies the date when the save set becomes recyclable.

Note: NSR_SAVESET_RETENTION must be set through the send command only. For a scheduled Oracle backup, this parameter overrides the Retention Policy attribute setting in the NetWorker Client resource.

• The most appropriate policy for the client, as selected by the NetWorker server (default).

• A valid date in nsr_getdate(3) format.

NSR_SERVER Mandatory for a manual backup if the NetWorker server host is different from the Oracle Server host.Recommended for a local manual backup.Specifies the hostname of the NetWorker server to perform the Oracle backup.

Note: For a manual backup that generates backup copies, NSR_SERVER must be set through the parms option, not through the send command or option.

• The local physical host (default).• A valid NetWorker server name.





NSR_SERVER_NIC Optional. Specifies the name of a network interface card (NIC) on a NetWorker server.When this parameter is set through the RMAN send command for an allocated channel, its value overrides the NSR_SERVER setting for that channel only.

Note: This parameter must be explicitly set for each channel to which it applies. Setting this parameter is the only supported way to override the NSR_SERVER value for a scheduled backup.

• Undefined (default).• A valid name of a NetWorker server NIC.





RMAN Commands 199

B


◆ The delete expired backup command ...................................................................... 200◆ The change...crosscheck and crosscheck commands ............................................. 200◆ The pool option of the backup command................................................................ 200◆ The send command..................................................................................................... 201◆ The set duplex command........................................................................................... 205◆ The trace option of the backup command............................................................... 206

RMAN Commands


RMAN Commands

The delete expired backup commandFor a particular user to be able to use the delete expired backup command with a NetWorker server, the user must have the required NetWorker privileges, as described in “Specify the required NetWorker privileges” on page 60.

If the user does not have the required NetWorker privileges, or there is an authorization problem when the delete expired backup command runs, NSR-DBM-ORA fails to remove the required entries in the NetWorker client file index and media database.

Despite this failure, the delete expired backup command removes the corresponding backup set or backup piece entries in the Oracle Recovery Catalog. In this case, the appropriate NetWorker media management command can be used to manually remove the required save set entries from the NetWorker indexes.

The NetWorker Command Reference Guide and man pages on UNIX systems provide more information on the NetWorker media management commands.

Note: If the NetWorker client binaries are located in a nondefault directory on the Oracle Server host and the NWORA resource file was not created during the NSR-DBM-ORA installation, the parameter NSR_NWPATH might need to be set in the NWORA resource file or in the RMAN script. “NSR_NWPATH” on page 195 provides more information.

The change...crosscheck and crosscheck commandsFor all NetWorker client file index entries that are not browsable, the running of the change...crosscheck or crosscheck command causes the status of the corresponding backup pieces to change to expired in the RMAN catalog.

In the RMAN catalog, an expired status for a backup piece indicates that the NetWorker browse policy specified for that backup piece has expired.

The pool option of the backup command

IMPORTANT!NSR-DBM-ORA does not support the pool option of the RMAN backup command, with the exception of pool=0.

If any nonzero value is specified for the pool option of the RMAN backup command, the RMAN session terminates and NSR-DBM-ORA returns the following error message:

sbtbackup: Oracle pools are not supported

“NSR-DBM-ORA error messages” on page 212 provides more information on this error message.

To specify the NetWorker volume pool to be used by NSR-DBM-ORA, set the parameter NSR_DATA_VOLUME_POOL in the RMAN script. Appendix A, “Parameters in an RMAN Session,” provides more information.

The send command 201

RMAN Commands

The send commandThe NSR-DBM-ORA implementation of the send command enables the NSR* parameters to be set, as described in Appendix A, “Parameters in an RMAN Session.”

Set the NSR* parameter values by using the methods described in “Setting the NSR* parameters” on page 190. The use of the send command is recommended where possible.

The following sections describe the send command syntax and precedence rules and how to use the send command to set the parameters.

◆ “Syntax rules” on page 201

◆ “Two ways to run the send command” on page 203

◆ “Precedence rules” on page 204

Note: In the following sections, brackets ([]) are used to denote the optional portions of a command, for example, command options and corresponding settings. When typing the command, do not include the brackets.

Syntax rules

The send command must have the following format:

send [ device_type ’device_specifier’ | channel channel_id ] ’NSR_ENV=(name1=value1 [, name2=value2, ...])’

These sections describe the syntax rules for the two main parts of the send command:

◆ “The send command string” on page 201

◆ “The send command options” on page 202

The send command string

The command string in the send command is the string inside the quotes, ’NSR_ENV=(name1=value1...)’.

Follow these syntax rules for the send command string:

◆ The maximum length of the command string is restricted by Oracle to 512 bytes, including the terminating NULL.

◆ The NSR_ENV keyword and the parameter names must be all uppercase.

◆ Between the NSR_ENV keyword and left parenthesis, an equal sign and spaces are optional. For example, these commands are all correct:

send ’NSR_ENV = (NSR_SERVER=server1)’send ’NSR_ENV=(NSR_SERVER=server1)’send ’NSR_ENV (NSR_SERVER=server1)’send ’NSR_ENV(NSR_SERVER=server1)’

◆ The parentheses in the command string are mandatory.

◆ Inside the parentheses, there must be one or more NSR-DBM-ORA parameter names and the corresponding parameter values.

◆ Inside the parentheses, spaces are not allowed around the equal signs. A space before an equal sign becomes part of the parameter name. A space after an equal sign becomes part of the parameter’s value.

◆ Commas separating the name=value entries are mandatory.


RMAN Commands

◆ Comments are not allowed inside the quotes. In the following example,“# NSR_SERVER” is considered the first parameter’s name:

run {allocate channel t1 type ’SBT_TAPE’;send ’NSR_ENV=(

# NSR_SERVER=server1,NSR_CLIENT=oracle)’;

:

◆ A send command in an RMAN script can span multiple lines. For example:

send ’NSR_ENV=(NSR_SERVER=server1,NSR_CLIENT=oracle)’;

The send command options

Run the send command with only one of the following:

◆ send with no option (only the quoted command string) sets the parameters for all allocated channels.

◆ send device_type ‘SBT_TAPE’ sets the parameters for all channels of the backup tape device.

Note: The send command has no effect with device type disk.

◆ send channel sets the parameters for the specified channels only.

IMPORTANT!The device_type or the channel option can be used in the send command in an RMAN script only. Neither option can be used in the send command on the operating system command line. “The send command on the operating system command line” on page 203 provides more information.

Example 41 A send command sets the parameters for a specified channel

In the following sample script, the parameters are set for channel t1 only, not for channel t2:

run {allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;send channel t1 ’NSR_ENV=(NSR_SERVER=server1,NSR_DATA_VOLUME_POOL=MondayFulls)’;

:}

Table 11 on page 202 lists the values for options used with the send command. The example referred to is Example 41 on page 202.

Table 11 Option values in the send command (page 1 of 2)

Option value Describes

device_specifier The device type as specified in an allocate channel command in the RMAN script. For a backup tape device, use SBT_TAPE.

channel_id The channel identifier as specified in an allocate channel command in the RMAN script. In the example, the identifier is t1.

name1 The first NSR-DBM-ORA parameter name. In the example, the first parameter name is NSR_SERVER.

The send command 203

RMAN Commands

Two ways to run the send command

There are two different ways to run the send command:

◆ As an option of the rman command on the operating system command line, as described in “The send command on the operating system command line” on page 203.

◆ In the run job of the RMAN script, as described in “The send command in the RMAN script” on page 204.

The send command on the operating system command line

To run the send command as an option of the rman invocation on the operating system command line, type the command in the following format:

rman send ”’NSR_ENV=(name1=value1[, name2=value2, ...])’”

◆ If more than one send option appears in the rman command, only the last send command is executed.

◆ Follow all the send command syntax rules listed in “The send command string” on page 201, except for the last rule, which applies only to a send command in an RMAN script.

◆ Do not use either the device_type or channel option. “The send command options” on page 202 provides more information.

◆ Use two sets of quotes around the command string, each set consisting of a single and double quote. The single quote can be either before or after the double quote, but the second set of quotes must be opposite to the first set. For example, this command is also correct:

rman send ’”NSR_ENV=(name1=value1[, name2=value2, ...])”’

Two sets of quotes are required to prevent some operating system shells (for example, ksh) from treating spaces inside the quotes as meta (special) characters and attempting to tokenize the string.

◆ The parameter values in the quoted string are applied to all channels allocated during the RMAN session. These values are applied before any parameter values specified in send commands within the RMAN script itself. “Precedence rules” on page 204 provides more information.

Example 42 An rman send command sets a parameter for all channels

In the following example, the NSR_SERVER parameter value (mars.fujitsu.com ) is applied to all three channels (t1, t2, t3) allocated in the RMAN script:

rman send ”’NSR_ENV=(NSR_SERVER=mars.fujitsu.com )’”

(RMAN script:)

value1 The value assigned to the first parameter. In the example, the first value is server1.

name2 The second NSR-DBM-ORA parameter name. In the example, the second parameter name is NSR_DATA_VOLUME_POOL.

value2 The value assigned to the second parameter. In the example, the second value is MondayFulls.

Table 11 Option values in the send command (page 2 of 2)

Option value Describes


RMAN Commands

run {allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;allocate channel t3 type ’SBT_TAPE’;

:}

The send command in the RMAN script

To run the send command in the run job of the RMAN script, type the command in the following format, at the required point within the run command brackets:

send [ device_type ’device_specifier’ | channel channel_id ] ’NSR_ENV=(name1=value1 [, name2=value2, ...])’

◆ Follow all the send command syntax rules listed in “The send command string” on page 201.

◆ Use either the device_type or channel option (if required) with the send command in an RMAN script, as described in “The send command options” on page 202.

◆ Specify the correct option values in the send command, as described in “The send command options” on page 202.

◆ RMAN commands are run in the order that they appear in the backup or restore script. For a parameter value to be in effect during a backup or restore, put the send command (setting the value) before the backup or restore command in the script, but after the allocate channel commands for those channels to which the parameter value applies.

◆ If no channel is allocated when the send command runs, an RMAN error is produced.

The following sample RMAN script performs an Oracle backup of the entire database to the volume pool MondayFulls of the (remote) NetWorker server mars.fujitsu.com :

run {allocate channel t1 type ’SBT_TAPE’;allocate channel t2 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_SERVER=mars.fujitsu.com , NSR_DATA_VOLUME_POOL=MondayFulls)’;backup full filesperset 4format ’FULL_%d_%U’(database);release channel t1;release channel t2;

}

This script is the same as the sample script on page 71. The single send command sets the parameters for both channels.

Precedence rules

Parameters are set for channels allocated during the RMAN session in the following order:

1. In the parms option in the allocate channel or configure channel command (configure channel is used only for automatic channel allocation).

2. In the rman send command on the operating system command line.

3. In the send command in the run job of the RMAN script.

The set duplex command 205

RMAN Commands

IMPORTANT!If both the send command on the operating system command line and the send option in the configure channel command are used at the same time, only the send option in the configure channel command is executed by Oracle.

To prevent confusion and simplify the task of setting parameters in a specific order, do not mix these different ways of setting parameters in the same RMAN session.

Example 43 Order of parameters set according to the precedence rules

In the following example, the parameters NSR_SERVER and NSR_CLIENT are set in this order:

◆ NSR_SERVER is set to server1 (by rman send), changed to server2 (by the first send command), and finally changed to server3 (by send channel).

◆ NSR_CLIENT is set to client1 (by rman send), changed to client2 (by the first send command), and finally changed to client3 (by send channel):

rman send ”’NSR_ENV=(NSR_SERVER=server1, NSR_CLIENT=client1)’”

(RMAN script:)

run {allocate channel t1 type ’SBT_TAPE’;send ’NSR_ENV=(NSR_SERVER=server2, NSR_CLIENT=client2)’;send channel t1 ’NSR_ENV=(NSR_SERVER=server3, NSR_CLIENT=client3)’;

:}

The set duplex commandFor a manual backup only, the set duplex command can be set in the RMAN backup script to generate up to four copies of an Oracle backup and then store those copies on separate media. Set duplex to the value 1, 2 (or instead of 2, set it to on), 3, or 4 to produce 1, 2, 3, or 4 copies, respectively, of every Oracle backup set generated by subsequent backup commands.

Note: NSR-DBM-ORA supports the generation of backup copies for manual backups only, not for scheduled backups.

Table 12 on page 205 describes the results of setting duplex to each of the valid values.

Table 12 Set duplex command values

Set duplex command Oracle backup results

set duplex=1 The backup set is directed to NSR_DATA_VOLUME_POOL.

set duplex=2orset duplex=on

Two copies of the backup set are directed to the separate pools specified by NSR_DATA_VOLUME_POOL and NSR_DATA_VOLUME_POOL1.These two pools must be different.


RMAN Commands

There are no default values for the parameters NSR_DATA_VOLUME_POOL, NSR_DATA_VOLUME_POOL1, NSR_DATA_VOLUME_POOL2, and NSR_DATA_VOLUME_POOL3. For a manual backup with backup copies, the values of these parameters must be defined with the parms option, not with the send command or option. Appendix A, “Parameters in an RMAN Session,” provides more information on how to set parameters.

Each pool specified by one of these NSR_DATA_VOLUME_POOL* parameters must be properly configured, and each pool must be different from the other pools used. If a pool is not properly defined or configured, the Oracle backup will be suspended, waiting for the proper configuration of that pool.

To enable use of the set duplex command, set the parameter BACKUP_TAPE_IO_SLAVES to TRUE in the initoracle_sid.ora file. The Oracle backup and recovery documentation provides more information.

If the current control file is included in the backup, RMAN duplexes the backup pieces of the control file in the same backup set. If the control file autobackup is enabled, RMAN also duplexes the backup pieces that belong to the control file autobackup.

Note: The set duplex command is deprecated by Oracle. “Backup copies” on page 25 provides information on additional Oracle commands to use for backup set duplexing during manual backups.

During an Oracle restore, RMAN selects only one of the copies to use, and if it fails for some reason, the restore fails. If the first copy of a backup piece cannot be found in NSR-DBM-ORA, RMAN issues the following type of error message:

RMAN-10035: exception raised in RPC: ORA-19507: failed to retrieve sequential file, handle="CH2_bkup3_1_1"

To force RMAN to use the duplexed copy of this missing backup piece, run the change...crosscheck, crosscheck, or change backuppiece...unavailable command and retry the restore. These commands cause RMAN to mark the missing backup piece as expired (“Retention policies” on page 34 provides a definition of expired) and to use the duplexed copy for the restore operation.

The appropriate Oracle backup and recovery documentation provides more information.

The trace option of the backup commandSet the trace option of the RMAN backup command to the value 0, 1, or 2. The default value of trace is 0.

set duplex=3 Three copies of the backup set are directed to the separate pools specified by NSR_DATA_VOLUME_POOL, NSR_DATA_VOLUME_POOL1, and NSR_DATA_VOLUME_POOL2.These three pools must be different.

set duplex=4 Four copies of the backup set are directed to the separate pools specified by NSR_DATA_VOLUME_POOL, NSR_DATA_VOLUME_POOL1, NSR_DATA_VOLUME_POOL2, and NSR_DATA_VOLUME_POOL3.These four pools must be different.

Table 12 Set duplex command values

Set duplex command Oracle backup results

The trace option of the backup command 207

RMAN Commands

You can set the trace option with the following command:

allocate channel t1 type ’SBT-TAPE’ trace=1;

The output of trace is written to the Oracle sbtio.log file and to the /nsr/applogs/nmo.messages.raw file.

These log files do not contain Oracle Server or RMAN errors.

NSR-DBM-ORA generates error messages in the nmo.messages.raw file in a language-independent binary form, readable by the nsr_render_log program only. The nmo.messages.raw file replaces the nmo.messages file used in previous NSR-DBM-ORA releases.


Table 13 on page 207 outlines the conditions that are traced when the trace option is set to each of the three valid values.

Table 13 Trace option values and conditions traced

Trace value Conditions traced

0 (default) All error conditions.

1 • All error conditions.• Entry and exit for each System Backup to Tape (SBT) function (the NSR-DBM-ORA

implementation of the Oracle SBT interface).

2 • All error conditions.• Entry and exit for each SBT function (the NSR-DBM-ORA implementation of the Oracle

SBT interface).• Values of all function parameters.• First 32 bytes of each read/write buffer.


RMAN Commands

Troubleshooting and Error Messages 209

C


◆ Troubleshooting tips ................................................................................................... 210◆ RMAN error messages ................................................................................................ 211◆ NSR-DBM-ORA error messages ............................................................................... 212

Troubleshooting andError Messages


Troubleshooting and Error Messages

Troubleshooting tipsThe following list of troubleshooting tips refers to sections of this Administrator’s Guide and the NSR-DBM-ORA Release Notice. Use the following list to troubleshoot any problems in running Oracle backup and restore operations with NSR-DBM-ORA.

To set up an Oracle backup and restore system correctly:

1. Verify that the combination of the operating system, Oracle Server, NetWorker server, and NetWorker client is supported.

The NSR-DBM-ORA Release Notice provides information on installation requirements.

2. Configure the Oracle Server and Net service (if used) according to the instructions in the Oracle documentation. “Verify the Oracle Server configuration” on page 59 provides more information.

Without NSR-DBM-ORA installed on the Oracle Server host, it should be possible to perform a backup and restore by using the allocate channel t1 type disk command.

3. Install and configure the NetWorker server and client software.

The following sources provide more information:

• NetWorker Administrator’s Guide

• Chapter 2, “Software Configuration”

4. Install and license NSR-DBM-ORA according to the instructions in the NSR-DBM-ORA Release Notice.

To relink the library file, use the linking commands listed in the NSR-DBM-ORA Release Notice. Use the correct libnwora.* library file.

To verify the version of NSR-DBM-ORA installed, check the version of the nsrnmostart program file or libnwora.* library file:

• On UNIX, use one of the following commands, where filenm is either the nsrnmostart or libnwora.* filename:

what filenmstrings filenm | grep “@(#)”pkginfo -l SMAWnwnmo (Solaris only)

Compare the library file with the libnwora.* file in the NSR-DBM-ORA software package; the two files should be identical. Ensure that Oracle is not linked to another libnwora.*, libobk.* (NSR-DBM-ORAversion of libnwora.*), or liblsm.*.

5. Perform a manual Oracle backup by using NSR-DBM-ORA and the proper RMAN script.

Set the required NSR* parameters in either the RMAN backup script or the rman send command on the operating system command line. Appendix A, “Parameters in an RMAN Session,” provides information on how to set the NSR* parameters. “RMAN scripts for manual backups” on page 71 provides a simple startup RMAN script.

If the manual backup fails, check the debug files for NSR-DBM-ORA and the NetWorker server. For NSR-DBM-ORA debug information, set the parameters NSR_DEBUG_FILE and NSR_DEBUG_LEVEL. “NSR-DBM-ORA error messages” on page 212 provides more information.

RMAN error messages 211


If the backup fails with the following error, ensure that both NSR-DBM-ORA and Oracle have the same bitness, and refer to the RMAN user guide for details on how to test that the media management library is integrated correctly:

ORA-19554: error allocating device, device type: SBT_TAPE, device name:ORA-27211: Failed to load Media Management LibraryAdditional information: 25

6. Perform a scheduled Oracle backup by using NSR-DBM-ORA and the proper RMAN and nsrnmo scripts.

In the working RMAN manual backup script, add the connect target and connect rcvcat commands, as described in “RMAN scripts for scheduled backups” on page 72. Customize the nsrnmo script, as described in “Customize the nsrnmo script” on page 75.

If the manual backup succeeds but the scheduled backup fails, set the parameter NSR_SB_DEBUG_FILE to generate debug information for the scheduled backup. “RMAN error messages” on page 211 provides more information.

RMAN error messagesRMAN stores information and RMAN-specific error messages in the log file specified by using the msglog option. Review the RMAN information in this log file after each backup.

To specify the name of the RMAN log file:

◆ For a manual Oracle backup, specify the msglog option in the rman command on the command line:

rman target ... rcvcat ... msglog filename

◆ For a scheduled Oracle backup, specify the msglog option in the parameter NSR_RMAN_ARGUMENTS in the nsrnmo script. “NSR_RMAN_ARGUMENTS” on page 77 provides more information.

The appropriate Oracle error messages guide provides more information on specific RMAN error messages and recommended courses of action.



NSR-DBM-ORA error messagesDuring a backup or restore, the NSR-DBM-ORA software records the NSR-DBM-ORA error messages in both the Oracle sbtio.log file and the following default log file on the Oracle Server host nsr/applogs/nmo.messages.raw

This log file does not contain Oracle Server or RMAN errors.

NSR-DBM-ORA generates error messages in the nmo.messages.raw file in a language-independent binary form, readable by the nsr_render_log program only. The nmo.messages.raw file replaces the nmo.messages file used in previous NSR-DBM-ORA releases.


To obtain more debug information for a backup or restore, set the parameter NSR_DEBUG_FILE to the valid pathname of a log file to store the information. If the parameter is set to an invalid pathname or the specified file does not have write permissions, the debug information is written to the /nsr/applogs/nmosbt.log file. Otherwise, the information is not written.

“NSR_DEBUG_FILE” on page 193 provides more information.

The parameter NSR_DEBUG_LEVEL controls the amount of debug information written to the log file specified by NSR_DEBUG_FILE. The debug level setting limits only the amount of debug information generated by the three functions sbtread(), sbtwrite(), and sbtpcstatus(), which are called most often during backups and restores. All the debug messages for other functions are always written to the log file, regardless of the debug level. “NSR_DEBUG_LEVEL” on page 193 provides more information.

During a scheduled backup, the NSR-DBM-ORA software records additional information and error messages in the file specified by the parameter NSR_SB_DEBUG_FILE in the nsrnmo script. If the parameter is set to an invalid pathname or the specified file does not have write permissions, the debug information is written to the /nsr/applogs/nsrnmostart.log file. Otherwise, the information is not written.

If the parameter NSR_SB_DEBUG_FILE is not set in the nsrnmo script file, the debug information is not written. “NSR_SB_DEBUG_FILE” on page 77 provides more information.

The NetWorker server also writes diagnostic information from the manual or scheduled Oracle backup to specific log files on the NetWorker server. The NetWorker Administrator’s Guide provides more information on these log files.

NSR-DBM-ORA error messages can be grouped into categories, according to the program that generates the message and the message format:

◆ “Error messages from the libnwora library” on page 213

◆ “Error messages from the nsrnmoadmin program” on page 221

◆ “Error messages from the nsrnmoinfo program” on page 223

◆ “Error messages from the nsrnmoprobe program” on page 224

◆ “Error messages from the nsrnmostart program” on page 225

NSR-DBM-ORA error messages 213


Error messages from the libnwora library

Table 14 on page 213 lists error messages generated by the libnwora library, in alphabetical order.

The error messages appear in the following format:

function_name: error_message (error_type:error_code:error_number)

where:

◆ function_name is the name of the NSR-DBM-ORA function that produced the error.

◆ error_message is the text of the error message, as shown in the table.

◆ error_type, error_code, error_number are internal numbers that represent an error type or code. Their significance for the user is as follows:

• If error_code is 1, the system is out of memory.

• If error_code is 3, 13, or 17, a code-level error has occurred. Report the error message to Technical Support.

Table 14 Error messages from the libnwora library (page 1 of 8)

Error message Description Resolution

A connection to NW server 'server' could not be established because 'reason'.

NSR-DBM-ORA could not connect to the NetWorker client file index due to the given reason. The client might not be configured as a client on the server.

Take the corrective action suggested by the error message.

Attempted to restore file 'filename' to raw device 'device_name'.

A proxy restore of a regular file to a raw device was attempted. This type of restore is not supported.

Do not attempt to restore a regular file to a raw device.

Attempted to restore raw device 'device_name' to file 'filename'.

A proxy restore of a raw device file to a regular file was attempted. This type of restore is not supported.

Do not attempt to restore a raw device file to a regular file.

Cannot back up object object_name with proxy copy.

The RMAN backup command included the proxy only option, but the object object_name did not reside on a primary storage device that the PowerSnap Module supports.

When the backup command includes the proxy only option, ensure that the object object_name resides on a primary storage device that the PowerSnap Module supports.

Could not create the LNM index lock file 'filename' (errno)

NSR-DBM-ORA failed to create the lock file required for an index deletion operation.

Report the error number (errno) to Technical Support.

Could not decode the 'sf_check' value: xdrs = 0xvalue

This is an internal XDR error caused by a network read or write operation.

Report the error to Technical Support.

Could not decode the 'sf_magic' value: xdrs = 0x%value



Could not decode the 'sf_more' flag: xdrs = 0xvalue



Could not find the nsrsnapck binary. During an index removal for a proxy backup, NSR-DBM-ORA could not locate the nsrsnapck binary, which is probably in a nondefault location.

Ensure that the parameter NSR_NWPATH is set correctly.

Could not locate the LNM save file 'backup_piece_name' on server 'server'.

NSR-DBM-ORA could not locate an index record for the backup piece. The index record is probably missing.

Use the mminfo and nsrinfo commands to verify the status of the index record.



Could not locate the LNM save time 'save_time' on server 'server'.

NSR-DBM-ORA could not locate an index record for the save time in the client file index. The index record is probably missing.


Could not lock 'filename' for index deletion. There were number attempts. (errno)

NSR-DBM-ORA was able to create the lock file required for an index deletion operation, but could not lock the file after the given number of attempts.

Report the error number (errno) to Technical Support.

Could not lstat - filename The lstat() system call failed. The file filename either did not exist or had invalid permissions.

Ensure that the file is an existing file with valid permissions.

Could not lstat secondary link - filename The lstat() system call failed. The file filename was a symbolic link that pointed to a file that either did not exist or had invalid permissions.

Ensure that the symbolic link points to an existing file with valid permissions.

Could not obtain NSR_ORACLECAT_MODE from NWORA resource file.

The error was caused by one of the following conditions:• The NWORA resource file does not exist.• The NWORA resource file has incorrect

permissions.• The NWORA resource file is corrupted.

Based on the condition, perform one of the following:• If the NWORA resource file does not exist,

create the file.• Ensure that the NWORA resource file has

correct permissions.• If the NWORA resource file is corrupted,

re-create the file.“The NWORA resource file” on page 176 provides more information.

Could not read link - pathname A proxy backup failed due to the pathname that was an invalid symbolic link.

Before a proxy backup, ensure that any symbolic link is a valid link.

Direct saves are not enabled. NSR-DBM-ORA attempted to connect to an old release of NetWorker server software that is no longer supported.

Update the NetWorker server software to a release supported by NSR-DBM-ORA 5.0. The NSR-DBM-ORA Release Notice provides details on the supported server releases.

Error creating staging directory 'directory'. During a proxy restore of a regular file, the permissions of the destination directory were possibly invalid. NSR-DBM-ORA was not able to create the required staging subdirectory, .nworapc.

Ensure that the destination directory has valid permissions for a proxy restore.

Error in mmdb lookup by time: reason A lookup in the media database failed for the given reason.

Use the mminfo command to verify the status of the media database record. Take the corrective action suggested by the error message.

Exceeded the number of retries. The NetWorker server may be down or unreachable.

NSR-DBM-ORA could not contact the NetWorker index service nsrindexd. This was probably caused by the NetWorker services being shutdown.

Restart the NetWorker services on the server, as required.

Exceeded the number of retries for nsr_init(). The NetWorker server may be down or unreachable.

After a maximum of five attempts, NSR-DBM-ORA failed to call the NetWorker core function, nsr_init(). This was probably caused by the NetWorker services being shut down.


Exceeded the number of retries for nsr_start(). The NetWorker server may be down or unreachable.

After a maximum of five attempts, NSR-DBM-ORA failed to call the NetWorker core function, nsr_start(). This was probably caused by the NetWorker services being shut down.






Invalid browse and retention policies. Values Ignored.

The NSR_SAVESET_BROWSE and NSR_SAVESET_RETENTION parameters both had invalid time values.

Ensure that the parameters NSR_SAVESET_BROWSE and NSR_SAVESET_RETENTION in the RMAN script both have valid values in the NetWorker date format.

Invalid browse policy browse_time. Value Ignored.

The NSR_SAVESET_BROWSE parameter had an invalid time value, browse_time.

Ensure that the parameter NSR_SAVESET_BROWSE in the RMAN script has a valid value in the NetWorker date format.

Invalid KEY word The syntax of the string in the RMAN send command was incorrect.

“The send command” on page 201 provides the correct send command syntax.

Invalid retention policy: retention_time. Value Ignored.

The NSR_SAVESET_RETENTION parameter had an invalid time value, retention_time.

Ensure that the parameter NSR_SAVESET_RETENTION in the RMAN script has a valid value in the NetWorker date format.

Invalid source path argument A proxy backup failed due to an invalid source pathname.

Perform a proxy backup with a valid source pathname only.

NSR_DATA_VOLUME_POOLn is not set. Multiple copies of the backup data were requested, but the required NSR_DATA_VOLUME_POOL parameters were not set. In the message, n was replaced by a number corresponding to the missing pool parameter.

When multiple copies of backup data are requested, set the required NSR_DATA_VOLUME_POOL parameters. “NSR_DATA_VOLUME_POOL” on page 192 provides more information.

nsrsnapck_binary_name process failed with error - reason

During an index removal for a proxy backup, the nsrsnapck binary failed.


ORA-19511: Error received from media manager layer, error text:Could not create the NWORA resource lock file (13) (103:105:13)

An NSR-DBM-ORA backup failed because a valid NWORA resource file does not exist or is not available.

If you do not use the wizard to configure a scheduled backup with save set bundling, use the nsrnmoadmin command to create a valid NWORA resource file, according to instructions in Chapter 2 or Chapter 7.

Oracle pools are not supported NSR-DBM-ORA does not support Oracle pools. NSR-DBM-ORA supports NetWorker pools only.

Remove the pool option of the backup command in the RMAN script or set the pool option to zero. “The pool option of the backup command” on page 200 provides more information.

Path pathname is too long. A proxy backup failed because the given pathname exceeded the limit of 1,024 bytes.

Ensure that any pathname involved in a proxy backup does not exceed 1,024 bytes.

pb_init() failed with (reason): invalid BRCAPI version

The version number of the BRC API that was reported by the PowerSnap Module was corrupted.


Proxy copy is not supported. A proxy operation was attempted on a platform that NSR-DBM-ORA does not support for proxy operations.

Do not attempt a proxy operation on an unsupported platform. The NSR-DBM-ORA Release Notice provides details on the supported platforms.

'string' should be in format: KEY=(xxxxx) The syntax of the string in the RMAN send command was incorrect.

“The send command” on page 201 provides the correct send command syntax.

The ASDF body could not be unwrapped. The incoming recover stream of data could not be decoded due to a possible network error or data corruption.






The backup file already exists: backup_piece_name

NSR-DBM-ORA could not complete the backup because the backup piece name already existed in the NetWorker client file index.

Change the format option string of the RMAN command to produce a unique backup piece name, or remove obsolete backup pieces. Then restart the backup operation.

The BRC API did not return an error string for the SBTPC object: object_name

An unknown error occurred during a BRC API function call by the PowerSnap Module.


The BRC status of logical object 'filename' was failure: file_status

The PowerSnap Module reported a failure during a proxy backup of the file filename.


The call to nsr_init() failed with the message:reason

A call of the NetWorker core function, nsr_init(), failed due to the given reason.

Take the corrective action suggested by the error message. If required, report the error to Technical Support.

The call to nsr_start() failed with the message:reason

A call of the NetWorker core function, nsr_rtart(), failed due to the given reason.


The call to pb_environment() failed with error:reason

During a proxy operation, a pb_environment() function call failed due to the given reason.


The call to pb_open() failed with error:reason

During a proxy operation, a pb_open() function call failed due to the given reason.


The call to pb_prepare() failed with error:reason

During a proxy operation, a pb_prepare() function call failed due to the given reason.


The call to pb_status() failed for object 'object_name' with the error:reason

During a proxy operation, a pb_status() function call failed due to the given reason.


The call to pb_status() for object 'object_name' failed with error:reason

During a proxy operation, a pb_status() function call failed due to the given reason.


The canonical OS file name path is invalid: filename

The operating system filename specified for a proxy operation was not a valid pathname.

Ensure that the file pathname specified for a proxy operation is a valid full pathname that is not a directory.

The current time could not be obtained (errno). NSR-DBM-ORA could not obtain the current time due to an operating system error.

Report the operating system error (errno) to the appropriate vendor.

The data could not be XDR'd from the stream. The incoming recover stream of data could not be decoded due to a possible network error or data corruption.


The data source is neither a file or a RAW volume - filename

The file filename involved in a proxy backup was not recognized as a regular file or raw volume. For proxy backups, NSR-DBM-ORA supports only regular files and raw volumes.

Ensure that filename is either a regular file or raw volume, as required for proxy backups.

The destination does not have the same terminating name as the source 'device_name'.

A proxy restore of a raw device was attempted to a location with a different basename from the backed-up source. For example, c1t2d0s2 is the basename (or terminating name) of /dev/rdsk/c1t2d0s2.

Perform a proxy restore of the raw device to a location with the same basename as the backed-up source.





The file being recovered could not be found in its staging location: filename

During a proxy restore, an error occurred at the point where the file filename was to be moved from the staging directory .nworapc to the destination directory.

Ensure that there are no permission or other problems with the destination directory and the staging directory .nworapc, and then restart the proxy restore. If the error recurs, report it to Technical Support.

The file 'filename' cannot be removed from the staging directory (errno).

During a proxy restore of the file filename, a file with the same name was found in the .nworapc subdirectory, probably left there by a previous failed restore. The errno is the error number from the failed attempt to remove the existing file.

Remove the file file_name from the .nworapc subdirectory, and restart the proxy restore.

The function mm_retrieve() failed with the error: reason

During a restore, a call of the NetWorker core function, mm_retrieve(), failed due to the given reason.


The function nsr_bind_recov_mm() failed with the error: reason

During a restore, a call of the NetWorker core function, nsr_bind_recov_mm(), failed due to the given reason.


The function nsr_end() failed with the error message: reason

A call of the NetWorker core function, nsr_end(), failed due to the given reason.


The function nsr_rstart() failed with the error: reason

During a restore, a call of the NetWorker core function, nsr_rstart(), failed due to the given reason.


The function sbtinit2() has already been called.

This is an internal error caused by Oracle calling the function sbinit2() twice.


The functions sbtinit() or sbtinit2() have not been called.

This is an internal error caused by Oracle not calling the two SBT initialization routines.


The index entry failed the cross check: cfx_name(backup_piece_name) save_time(save_time)

During an index lookup, the entry was located in the client file index but not in the media database.

Restart the NetWorker services, and use the mminfo and nsrinfo commands to verify the backup information in the indexes. Run the nsrck program to resolve any corruption of the indexes.

The lookup of 'backup_piece_name' on server 'server' failed - 'reason'

NSR-DBM-ORA could not locate backup_piece_name in the indexes due to the reason. The indexes might be corrupted.

Run the nsrck program to resolve any corruption of the indexes.

The name of the NSR client could not be determined.

The name of the NetWorker client could not be determined.

Set the parameter NSR_CLIENT to the NetWorker client name by using the send command.

The name of the NSR server could not be determined.

The name of the NetWorker server could not be determined.

Set the parameter NSR_SERVER to the NetWorker server name by using the send command.

The NSR-DBM-ORA BRCAPI version version is outside the range supported by the BRC service: earliest_version - latest_version

NSR-DBM-ORA release 5.0 does not support the PowerSnap Module release that was used for a proxy operation.

Ensure that a supported release of the PowerSnap Module is installed. The NSR-DBM-ORA Release Notice provides details on the supported platforms.

The NSR client name could not be determined.

The name of the NetWorker client could not be determined.


The NSR server name could not be determined.

The name of the NetWorker server could not be determined.


The NSR_CLIENT parameter was not set. The name of the NetWorker client could not be determined.






The NSR_DMO_BENCHMARK_1 parameter is no longer supported.

The undocumented parameter NSR_DMO_BENCHMARK_1 was specified, but it is no longer supported.

Do not set the unsupportedparameter NSR_DMO_BENCHMARK_1.

The NSR_SERVER parameter was not set. The name of the NetWorker server could not be determined.


The NW authentication for client 'client' was refused by server 'server' because 'reason'.

NSR-DBM-ORA could not obtain the required authentication to connect to the NetWorker client file index due to the given reason. The client might not be configured as a client on the server.


The NW client has not been set. The name of the NetWorker client could not be determined.


The NW server does not have a valid NSR-DBM-ORA proxy copy license.

The NetWorker server attempted a proxy operation without the required license.

Ensure that the NetWorker server has the required license for the proxy operation.

The NW server has not been set. The name of the NetWorker server could not be determined.


The NWORA file ID could not be XDR'd.xdrm: 0xvalueNWORA fid: 0xvaluessid: 0xvaluessoff: 0xvalue



The NWORA resource file does not exist. Please create it with nsrnmoadmin.

A proxy backup failed because the NWORA resource file did not exist.

Create the NWORA resource file by using the nsrnmoadmin command, and restart the proxy backup. “The NWORA resource file” on page 176 provides details.

The NWORA resource lock file does not exist. Please create it by running 'nsrnmoadmin -r list'

A proxy backup failed because the NWORA resource lock file did not exist.

Create the NWORA resource lock file by using the nsrnmoadmin -r list command, and restart the proxy backup. “The NWORA resource file” on page 176 provides details.

The NWORA resource NSR_ORACLECAT_MODE is in the 'undetermined' state.

In the NWORA resource file, NSR_ORACLECAT_MODE was set to the default value of undetermined.

Set the value of NSR_ORACLECAT_MODE to either enabled or disabled (as required) by using the nsrnmoadmin command.

The object 'filename' is not a file. A proxy backup failed because the file filename is not a data file—neither a raw file nor a regular file.

Perform a proxy backup of a supported type of data file only.

The ORACLE_SID must be set when performing proxy copy backups.

During a scheduled proxy backup, the parameter ORACLE_SID was not set in the nsrnmo script.

In the nsrnmo script, set the parameter ORACLE_SID to the SID value of the Oracle database.

The OS file name has been specified multiple times by Oracle: filename

This is an internal Oracle error caused by Oracle specifying the same filename twice during a proxy operation.


The parameter file cannot be open: filename The configuration file specified by the parameter NSR_PROXY_PFILE could not be opened. The file should contain PowerSnap parameter settings for a proxy backup or restore.

Ensure that the value specified by the parameter NSR_PROXY_PFILE is a valid pathname of the configuration file.

The pb_cancel() call for object 'object_name' returned the error message:error

The pb_cancel() function call failed during a proxy operation.






The pb_inquiry() call failed for object 'object_name':error

The pb_inquiry() function call failed during a proxy operation.


The pb_inquiry() for object 'object_name' failed because:error



The pb_inquiry() of object 'object_name' returned error:error



The pb_restore() for object 'object_name' failed with error:error

The pb_restore() function call failed during a proxy operation.


The pb_save() of object 'object_name' returned error:error

The pb_save() function call failed during a proxy operation.


The pb_snapshot() call for object 'object_name' failed with error:error

The pb_snapshot() function call failed during a proxy operation.


The record obtained has the wrong save time 'save_time1'. The save time queried was 'save_time2'.

NSR-DBM-ORA located an index record in the client file index, but it had an unexpected save time. The indexes might be corrupted.

Restart the NetWorker services, and run the nsrck program to resolve any corruption of the indexes.

The removal of SSID 'save_set_id' failed with error: reason

An index deletion operation failed for the given reason.

Use the mminfo and nsrinfo commands to verify the status of the index record. If required, report the error to Technical Support.

The restore destination path is not valid: filename

During a proxy restore operation, NSR-DBM-ORA found the specified restore destination, filename, to be invalid.

Ensure that the specified restore destination is a valid pathname.

The restore operation for the file failed for an unknown reason: filename

During a proxy restore, an error occurred at the point where the file filename was to be moved from the staging directory .nworapc to the destination directory.

Ensure that there are no permission or other problems with the destination directory and the staging directory .nworapc, and then retry the proxy restore. If the error occurs again, report it to Technical Support.

The savefile_fini() call failed. reason During a restore, a call of the NetWorker core function, savefile_fini(), failed due to the given reason.


The SBTPC object could not determine the destination of the restore.

During a proxy restore operation, NSR-DBM-ORA was unable to determine where to restore the file.


The SBTPC object is not in the PB_TYPE_PREPARE state: object_name

During a proxy operation, NSR-DBM-ORA and the PowerSnap Module became unsynchronized as to the status of the object object_name.


The SBTPC object is not in the SBTPCSTATUS_NOTREADY state: object_name



The SBTPC object 'object_name' failed with the error message:reason

The proxy backup or restore of a file failed during a PowerSnap Module operation, for the given reason.






The SBTPC object 'object_name' is entering the SBTPCSTART backup state but its BRC type is: type



The SBTPC object 'object_name' is entering the SBTPCSTART restore state but its BRC type is: type



The SBTPC object 'object_name' is entering the SBTPCSTART state but its status is: status



The SBTPC object 'object_name' is leaving the BRC prepare state but its status is: status



The SBTPC object 'object_name' is leaving the BRC save state but its status is: status



The SBTPC object 'object_name' is leaving the BRC snapshot state but its status is: status



The SBTPC object 'object_name' was aborted by the BRC service. Please check the PowerSnap logs for an explanation.

The PowerSnap Module terminated the proxy operation.

Examine the PowerSnap Module logs for a possible reason for the termination.

The sbtpccommit() function was called during restore.

This is an internal Oracle error that occurred during a proxy restore.


The sfhead could not be XDR'd. This is an internal XDR error caused by a network write operation.


The SS browse time is not in the future:current time: current_timebrowse: browse_time

The specified browse policy time was in the past. This might be due to a problem with the operating system time setting.

Ensure that the browse policy time is set correctly. If required, ensure that the operating system time is set correctly.

The SS retention time is not in the future:current time: current_timeretention: rentention_time

The specified retention policy time was in the past. This might be due to a problem with the operating system time setting.

Ensure that the retention policy time is set correctly. If required, ensure that the operating system time is set correctly.

The staging directory 'directory' has invalid permissions (errno).

During a proxy restore, NSR-DBM-ORA was unable to write to the staging directory, directory. The errno is the error number from the function call that failed.

Ensure that the staging directory has valid permissions for a proxy restore.

The UNIX attributes could not be XDR'd.xdrm: 0xvalueua: 0xvalue



There are no SBTPC objects that have not returned their status.

This is an internal error during a proxy operation caused by Oracle expecting more files to be processed whereas NSR-DBM-ORA has completed its file processing.


This backup piece name is already used in the SBTPC session: backup_piece_name

This is an Oracle error caused by Oracle specifying the same backup piece name twice during a proxy operation.






Error messages from the nsrnmoadmin program

Table 15 on page 221 lists error messages generated by the nsrnmoadmin program, in alphabetical order.


nsrnmoadmin: error_message

where error_message is the text of the error message, as shown in the table.

Table 15 Error messages from the nsrnmoadmin program (page 1 of 3)


Command line arguments are not understood. The nsrnmoadmin command included one or more invalid options.

Use the nsrnmoadmin command with the correct options. “Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details.

Could not create the NWORA resource file (errno)

The nsrnmoadmin command could not create the NWORA resource file, possibly due invalid permissions.

Ensure that valid permissions exist to allow the nsrnmoadmin command to create the NWORA resource file. “The NWORA resource file” on page 176 provides details.

Could not create the NWORA resource lock file (errno)

The nsrnmoadmin command could not obtain the required lock file in the /nsr/tmp or NetWorker_install_path\tmp directory. The lock file is required for accessing the NWORA resource file.


Could not open resource file 'filename' (errno). The nsrnmoadmin command could not open the NWORA resource file, possibly due invalid permissions.

Verify that the NWORA resource file exists and has valid permissions. If required, create or repair the file by using the nsrnmoadmin command, or modify the file permissions.

No command line parameters are set. The nsrnmoadmin command options were missing.

Use the nsrnmoadmin command with the correct options.

NSR_ORACLECAT_MODE can only be set to 'enabled', 'disabled' or 'undetermined'.

In the nsrnmoadmin command, the NSR_ORACLE_CAT_MODE parameter resource was set to a value other than enabled, disabled, or undetermined.

In the nsrnmoadmin command, set the NSR_ORACLE_CAT_MODE parameter resource to enabled or disabled for instant backups. “Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details.

NSR_REMOVE_ON_FAILURE can only be set to 'TRUE' or 'FALSE'.

In the nsrnmoadmin command, the NSR_REMOVE_ON_FAILURE parameter resource was set to a value other than TRUE or FALSE.

In the nsrnmoadmin command, set the NSR_REMOVE_ON_FAILURE parameter resource to either TRUE or FALSE only.

NWORA parameter resources must be specified in the 'ResourceName ResourceValue' format.

In the nsrnmoadmin command, an NWORA parameter resource name and value were not specified in the correct format.

In the nsrnmoadmin command, specify the NWORA parameter resource name and value in the correct format.

NWORA SID resource must be specified when doing deletion.

In the nsrnmoadmin command with the -r delete option, the SID value of an Oracle database was not specified.

In the nsrnmoadmin command with the -r delete option, specify the correct SID value.

The '-r' flag cannot be set multiple times. The nsrnmoadmin command contained more than one -r option.

Use the nsrnmoadmin command with only one -r option.

The '-r' option requires an NWORA resource specification.

The nsrnmoadmin command with the -r option did not include the required resource specification.

In the nsrnmoadmin command with the -r option, specify the required resource name and value.



The '-r' option requires either an 'add', 'update', 'list' or 'delete' option.

In the nsrnmoadmin command, the -r option did not include one of the required keywords: add, update, list, or delete.

In the nsrnmoadmin command, include one of the required keywords with the -r option. “Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details.

The first NWORA resource is not a header (errno).

The NWORA resource file is probably corrupted. Verify the contents of the NWORA resource file. If required, repair the resource file by using the nsrnmoadmin command.

The NWORA resource file does not contain the NSR_NWPATH resource.

The NWORA resource file does not contain the mandatory NSR_NWPATH parameter resource. The file might be corrupted.

Verify the contents of the NWORA resource file. If required, repair the resource file by using the nsrnmoadmin command.

The NWORA resource file does not exist. The NWORA resource file does not yet exist. Create the NWORA resource file by using the nsrnmoadmin command.

The NWORA resource named 'resource_name' is not found.

The nsrnmoadmin command specified the name of a resource that does not exist in the NWORA resource file.

In the nsrnmoadmin command, specify a valid resource name from the NWORA resource file.

The NWORA resource parameter list can only contain one entry.

The NWORA resource file includes multiple values for a resource, which is not supported. The file is probably corrupted. The file might have been edited manually, which is not supported.

Repair the NWORA resource file by using the nsrnmoadmin command.

The NWORA resource parameter list contains the invalid element 'resource_name'.

The NWORA resource file contains an invalid resource name. The file is probably corrupted. The file might have been edited manually, which is not supported.

Repair the NWORA resource file by using the nsrnmoadmin command. “Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details.

The NWORA resource parameter list for a SID requires the item1, item2 and item3 information.

The nsrnmoadmin command for creating or updating an NWORA SID resource was missing the required items.

In the nsrnmoadmin command for creating or updating an NWORA SID resource, include the required items.

The NWORA resource 'resource_name' is not a SID resource.

The nsrnmoadmin command with the -r delete option did not include a valid name of an NWORA SID resource.

In the nsrnmoadmin command with the -r delete option, specify a valid name of an NWORA SID resource.

The NWORA resource specified is not supported: resource_name = resource_value

In the nsrnmoadmin command, an invalid name or value were specified for an NWORA parameter resource.

In the nsrnmoadmin command, specify a valid name and value for an NWORA parameter resource. “NWORA parameter resources” on page 176 provides details.

The NWORA SID resource for 'sid_value' already exists.

The nsrnmoadmin command attempted to add an NWORA SID resource that already existed.

In the nsrnmoadmin command, specify the values for a new NWORA SID resource.

The SID token 'connect' is an empty string. The nsrnmoadmin command did not include the required pathname of the RMAN connection file with the connect keyword.

In the nsrnmoadmin command, specify a valid pathname of the RMAN connection file with the connect keyword. “Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details.

The SID token 'home' is an empty string. The nsrnmoadmin command did not include the required pathname of the Oracle home directory with the home keyword.

In the nsrnmoadmin command, specify a valid pathname of the Oracle home directory with the home keyword.

The SID token 'ORACLE_SID' is invalid. In the nsrnmoadmin command with the sid keyword, the specified SID value of the Oracle database was invalid.

In the nsrnmoadmin command, specify a valid SID value with the sid keyword.

The SID token 'sid' is an empty string. The nsrnmoadmin command did not include the required SID value of the Oracle database with the sid keyword.

In the nsrnmoadmin command, specify a valid SID value with the sid keyword.





Error messages from the nsrnmoinfo program

Table 16 on page 223 lists error messages generated by the nsrnmoinfo program, in alphabetical order.


The NW volume information lookup failed:error_message

where error_message is the text of the error message, as shown in the table.

The tokens 'sid', 'home' and 'connect' must be set when adding a SID.

The nsrnmoadmin command to add an NWORA SID resource did not include the settings of the mandatory sid, home, and connect keywords.

In the nsrnmoadmin command to add an NWORA SID resource, include the settings of the sid, home, and connect keywords.

The value of the NWORA resource is missing. In the nsrnmoadmin command with the -r update option, the NWORA resource value was not specified with the resource name.

In the nsrnmoadmin command with the -r update option, specify the NWORA resource value with the resource name.

Unrecognized argument 'option'. The nsrnmoadmin command included the unrecognized option option.

Use the nsrnmoadmin command with the correct options. “Configuring the NWORA resource file with the nsrnmoadmin program” on page 180 provides details.

You must be the super-user to update the NWORA resource file.

The nsrnmoadmin command was typed by the wrong user.

Type the nsrnmoadmin command as the root user.



Table 16 Error messages from the nsrnmoinfo program (page 1 of 2)


A connection to NW server 'server' could not be established because 'reason'.

NSR-DBM-ORA could not connect to the NetWorker client file index due to the given reason. The client might not be configured as a client on the server.


Could not locate the LNM save file 'backup_piece_name' on server 'server'.

NSR-DBM-ORA could not locate an index record for the backup piece. The index record is probably missing.


Could not locate the LNM save time 'save_time' on server 'server'.

NSR-DBM-ORA could not locate an index record for the save time in the client file index. The index record is probably missing.


Error in mmdb lookup by time: reason A lookup in the media database failed for the given reason.

Use the mminfo command to verify the status of the media database record. Take the corrective action suggested by the error message.

Exceeded the number of retries. The NetWorker server may be down or unreachable.

NSR-DBM-ORA could not contact the NetWorker index service nsrindexd. This was probably caused by the NetWorker services being shutdown.


The file 'filename' could not be opened. The file specified with the -f option of the nsrnmoinfo command could not be accessed.

Ensure that the specified file exists, and then type the nsrnmoinfo command again with the -f option.

The file name provided is NULL. In the nsrnmoinfo command, the -f option did not include the required filename.

In the nsrnmoinfo command, include the required filename with the -f option.



Error messages from the nsrnmoprobe program

Table 17 on page 224 lists error messages generated by the nsrnmoprobe program, in alphabetical order.

The index entry failed the cross check: cfx_name(backup_piece_name) save_time(save_time)

During an index lookup, the entry was located in the client file index but not in the media database.

Restart the NetWorker services, and use the mminfo and nsrinfo commands to verify the backup information in the indexes. Run the nsrck program to resolve any corruption of the indexes.

The lookup of 'backup_piece_name' on server 'server' failed - 'reason'

NSR-DBM-ORA could not locate backup_piece_name in the indexes due to the reason. The indexes might be corrupted.

Run the nsrck program to resolve any corruption of the indexes.

The NW authentication for client 'client' was refused by server 'server' because 'reason'.

NSR-DBM-ORA could not obtain the required authentication to connect to the NetWorker client file index due to the given reason. The client might not configured as a client on the server.


The record obtained has the wrong save time 'save_time1'. The save time queried was 'save_time2'.

NSR-DBM-ORA located an index record in the client file index, but it had an unexpected save time. The indexes might be corrupted.

Restart the NetWorker services, and run the nsrck program to resolve any corruption of the indexes.

Table 16 Error messages from the nsrnmoinfo program (page 2 of 2)


Table 17 Error messages from the nsrnmoprobe program


Could not connect to the Oracle database Oracle_service at Oracle_home.

The nsrnmoprobe program could not connect to the specified Oracle database.

Ensure the following:• The connection strings (database username

and password), Oracle Net service name, and ORACLE_HOME value that nsrnmoprobe uses are correct.

• You can manually connect to that Net service name by using Oracle client tools, such as sqlplus.

“Configure a probe-based backup” on page 90 provides more information.

The configuration information is not valid: string

The Command Options attribute in the NetWorker Probe resource was not configured properly.

Correct the Command Options attribute setting in the NetWorker Probe resource, according to “Configure a probe-based backup” on page 90.

The connect string was missing the user or the password.

The RMAN connection file (specified through NSR_ORACLE_CONNECT_FILE in the NWORA resource file) used for the NSR-DBM-ORA probe did not contain the username or password.

Edit the connection file to correct the problem.

The nsrnmoprobe command does not support databases with more than one sthread.

The nsrnmoprobe program does not support a RAC database, where the database has multiple threads (instances).

Do not use the nsrnmoprobe program with a RAC database.

The nwora.res file has not been created. ORACLE_SERVICE was set in the Command Options attribute in the NetWorker Probe resource, but the NWORA resource file did not exist.

Create the NWORA resource file according to the information in “Configure a probe-based backup” on page 90.

The v$database_incarnation SQL statement could not be run because: reason.

The nsrnmoprobe program could not determine the database incarnation information due to the reason given in the error message.

This error usually indicates that the Oracle database version is not supported. The NSR-DBM-ORA Release Notice provides details on the supported Oracle database versions.



Error messages from the nsrnmostart program

Table 18 on page 225 lists error messages generated by the nsrnmostart program, in alphabetical order.

Table 18 Error messages from the nsrnmostart program


Authentication deniedorAuthentication denied for user username

The lockbox that stored the database connection credentials was not accessible by the super-user on the client where the backup failed with this error message.

Use the NetWorker Management Console (NMC) to ensure that the Lockbox resource is created for the given client and the Users attribute contains the super-user of the client.

client: WARNING! The NWORA resource file 'save' process output error messages.client: Please check the save log file for more information: log_file

The NWORA resource file could not be backed up after a successful RMAN backup.

Analyze the log_file and if it includes an error message, take the corrective action suggested by the error message.

ORACLE_HOME is not defined. Cannot start RMAN.

ORACLE_HOME was not set properly in the nsrnmo script.

Set ORACLE_HOME properly in the nsrnmo script.

The backup config did not contain a string. The nsrnmostart program was run with the -C option, but the Backup Config attribute was not set properly in the Client resource.

Remove this Client resource, and re-create the Client resource by using the backup configuration wizard.

The NSR client resource for client_name does not contain any backup configuration.

The nsrnmostart program was run with the -C option, but the Backup Config attribute was not set properly in the Client resource.

Remove this Client resource, and re-create the Client resource by using the backup configuration wizard.

The temporary file 'rman_script_path' could not be created (errno).

The scheduled backup binary, nsrnmostart, could not create the file rman_script_path to write the RMAN script generated by the backup configuration wizard.

Ensure that the root user on UNIX has "write" permissions on the directory path of the rman_script_path file.



Command Reference 227

D

This appendix includes the man pages that accompany NSR-DBM-ORA:

◆ nsrnmo - NSR-DBM-ORA scheduled backup command script .......................... 228◆ nsrnmoadmin - NSR-DBM-ORA resource database management command.. 229◆ nsrnmoinfo - Lists the NetWorker volumes required for a restore. ................. 233◆ nsrnmostart - NSR-DBM-ORA scheduled backup executable program........ 234◆ nsroraclecat - Synchronizes the RMAN repository with the NetWorker

indexes. ......................................................................................................................... 235◆ nsrorara - NSR-DBM-ORA Scheduled Backup and Recovery Configuration

Wizard Remote Agent. ............................................................................................... 237

Command Reference


Command Reference

Maintenance Commands nsrnmo(1m)

NAMEnsrnmo - NSR-DBM-ORA scheduled backup command script

SYNOPSISnsrnmo <save options>

DESCRIPTIONThe nsrnmo shell script is used by the NetWorker server program savegrp(1m) to trigger scheduled backups on an Oracle database server. The nsrnmo script sets specific environment variables for the scheduled Oracle backup and invokes the Oracle Recovery Manager (RMAN), the backup interface provided by Oracle.

The nsrnmo script is usually run by savegrp only. It is not run manually. To use the nsrnmo script for a scheduled Oracle backup, you must specify nsrnmo in the Backup Backup Command attribute of the Client resource (see nsr_client(5)) for the scheduled Oracle backup.

To modify the nsrnmo command script, you should copy the template file /opt/nsr/nsrnmo.sh to the nsrnmo file and modify the nsrnmo file. Do not modify the template file /opt/nsr/nsrnmo.sh. Keep the template file unchanged, as a reference to the original nsrnmo script.

For information on how to create a separate nsrnmo script file for each Oracle instance, refer to Chapter 3, “Backup Procedures,”.

There are several environment variables in the nsrnmo script that you can customize for a particular scheduled Oracle backup. The variables ORACLE_HOME and PATH are mandatory for each scheduled Oracle backup. The variable ORACLE_SID is only required by proxy copy backups when catalog synchronization is enabled. The other variables are optional, and you can leave them undefined in the script, if desired. Note that the variables in the nsrnmo file are all initially undefined. Please refer to the /opt/nsr/nsrnmo.sh template, your site-specific nsrnmo script, and this guide for descriptions of the various environment variables.

OPTIONSPlease see the save(1m) man page for a description of options supported by nsrnmo.

FILES/opt/nsr/nsrnmo.sh - A template of this command is provided in /opt/nsr.

SEE ALSOnsrnmostart(1m), nsrnmoadmin(1m), mminfo(1m), nsr_client(5), nsr_pool(5), save(1m), savegrp(1m)

229

Command Reference

Maintenance Commands nsrnmoadmin(1m)

NAMEnsrnmoadmin - NSR-DBM-ORA resource database management command

SYNOPSIS

nsrnmoadmin [ -D ] -r list [ ResourceName ] | [ SidName ]

nsrnmoadmin [ -D ] -r add ResourceName ResourceValue

nsrnmoadmin [ -D ] -r add sid=SidName home=OracleHome connect=ConnectFilePath[ tns=TNSpath ] [ lib=LibraryPath ]

nsrnmoadmin [ -D ] -r update ResourceName ResourceValue

nsrnmoadmin [ -D ] -r update sid=SidName [ home=OracleHome ][ connect=ConnectFilePath ] [ tns=TNSpath ] [ lib=LibraryPath ]

nsrnmoadmin [ -D ] -r delete SidName

nsrnmoadmin -M [ -Y ] [ -D debug_level ] -s server [ -g group ] [ -c client_name ][ -N save_set_name ]

nsrnmoadmin -P [ -D debug_level ] -s server [ -g group ] [ -c client_name ][ -N save_set_name ]

DESCRIPTIONnsrnmoadmin creates, updates and deletes resources from the NSR-DBM-ORA resource file. It also migrates traditional NSR-DBM-ORA configurations to the new NSR-DBM-ORA wizard configuration format which is stored in the RAP database. nsrnmoadmin can only be run by the root user.

During migration operations, nsrnmoadmin reads the nsrnmo.sh, nwora.res and the RMAN script and stores the configurations they contain in the NSR-DBM-ORA client's NSR client RAP resource. After the migration operation is performed, the nsrnmo.sh, nwora.res and RMAN script are no longer read by the client. Instead all the configuration information is read from the RAP resource. When migrating, nsrnmoadmin migrates all the NSR client resources that match the group, client name and save set name specified on the command line. The save set name must begin with the prefix ‘RMAN:’, or nsrnmoadmin will assume that the NSR client is not for an NSR-DBM-ORA backup.

The nsrnmoadmin creates and updates the NSR-DBM-ORA resource file which is located at /nsr/res/nwora.res on the host where NSR-DBM-ORA is installed. An NSR-DBM-ORA resource file is automatically created during the NSR-DBM-ORA installation process when NSR-DBM-ORA is not installed to the default location. The NSR-DBM-ORA resource file created during installation only contains the NSR-DBM-ORA resources required for a regular (nonproxy) NSR-DBM-ORA backup. The NSR-DBM-ORA resource file must be updated or created before performing proxy copy backups.

When nsrnmoadmin accesses the NSR-DBM-ORA resource file for the first time, the NSR-DBM-ORA resources required for proxy copy backups are automatically added to the NSR-DBM-ORA resource file. The added NSR-DBM-ORA resources are initialized with


Command Reference

default values. If the file does not exist it is created and the NSR-DBM-ORA resources are initialized with default values.

There are two types of NSR-DBM-ORA resources. The first type is the NSR-DBM-ORA parameter resource. An NSR-DBM-ORA parameter resource is a single line that follows the NAME=VALUE format, where NAME is the name of the parameter, and VALUE is what the parameter is set to.

The second type of NSR-DBM-ORA resource is the NSR-DBM-ORA SID resource. An NSR-DBM-ORA SID resource contains all the information nsroraclecat(1m) requires to synchronize an RMAN repository. An NSR-DBM-ORA SID resource contains several NSR-DBM-ORA parameters which are described in the nsroraclecat(1m) man page. Each parameter is represented by a keyword on the nsrnmoadmin command line:

sid: NSR_ORACLE_SID

home: NSR_ORACLE_HOME

connect: NSR_ORACLE_CONNECT_FILE

tns: NSR_ORACLE_TNS_ADMIN

lib: NSR_ORACLE_LIB_PATH

The NSR-DBM-ORA resource file is automatically backed up by nsrnmostart(1m) during scheduled backups.

OPTIONS

-c

Specifies the name of the NSR-DBM-ORA clients to migrate.

-D

Specifies that debug information should be printed to stdout.

-g

Specifies the group of the NSR-DBM-ORA NSR client resource that is being migrated.

-M

Specifies that the configuration migration operation should be performed.

-N

Specifies the save set of the NSR-DBM-ORA NSR client resource that is being migrated.

-P

Specifies that the configuration migration operation should run in probe mode. The program does not actually migrate any legacy configurations.

-r

Several keywords specify the operation the -r option will perform:list

The list keyword causes nsrnmoadmin to list the contents of the NSR-DBM-ORA resource file. If ResourceName is specified only the NSR-DBM-ORA parameter resource that matches ResourceName will be printed. If SidName is specified only the NSR-DBM-ORA SID resource for the specified Oracle SID will be

231

Command Reference

printed.

add

The add keyword causes nsrnmoadmin to add a new NSR-DBM-ORA resource to the NSR-DBM-ORA resource file. If the 'sid', 'home' and 'connect' keywords appears in the parameters following add an NWORA SID resource is added to the NSR-DBM-ORA resource file. The 'tns' and 'lib' keywords are optional. If the parameters following add do not contain the 'sid', 'home' and 'connect' keywords it is assumed that the parameters are the ResourceName and ResourceValue of an NSR-DBM-ORA parameter resource.

update

The update keyword causes nsrnmoadmin to change the value of the specified NSR-DBM-ORA resource. An NSR-DBM-ORA parameter resource is specified with the same semantics used by the add keyword. When updating NSR-DBM-ORA SID resources only the 'sid' keyword is mandatory. The 'sid' keyword specifies which NSR-DBM-ORA SID resource will be updated. The 'sid' of an NSR-DBM-ORA SID resource cannot be changed. The 'home', 'connect', 'tns' and 'lib' keywords specify which parameters will be updated. See the EXAMPLES section for more information.

delete

The delete keyword causes nsrnmoadmin to delete an NSR-DBM-ORA SID resource from the NSR-DBM-ORA resource file. The parameter after the delete keyword is the name of the NSR-DBM-ORA SID resource without the 'sid=' prefix. Only NSR-DBM-ORA SID resources can be deleted.

-Y

Specifies to turn off all prompting during migration operations. When this parameter is specified, all questions are automatically answered Yes.

EXAMPLESThe following lists the entire contents of the NSR-DBM-ORA resource file:

nsrnmoadmin -r list

The following lists the NSR-DBM-ORA parameter resource NSR_ORACLECAT_MODE:

nsrnmoadmin -r list NSR_ORACLECAT_MODE

The following lists the NSR-DBM-ORA SID parameter resource for the ORACLE_SID 'orcl815':

nsrnmoadmin -r list orcl815

The following adds the NSR-DBM-ORA parameter resource NSR_ORACLECAT_MODE:

nsrnmoadmin -r add NSR_ORACLECAT_MODE enabled

The following adds the NSR-DBM-ORA SID resource 'orcl815':

nsrnmoadmin -r add sid=orcl815 home=/dbapps/orcl815 connect=/dbapps/orcl815/connect.file


Command Reference

The following updates the NSR-DBM-ORA parameter resource NSR_ORACLECAT_MODE:

nsrnmoadmin -r update NSR_ORACLECAT_MODE disabled

The following updates the NSR_ORACLE_TNS_ADMIN parameter of the NSR-DBM-ORA SID resource 'orcl815':

nsrnmoadmin -r update sid=orcl815 tns=/dbapps/orcl815/orcl815net8

The following deletes the NSR-DBM-ORA SID resource 'orcl815':

nsrnmoadmin -r delete orcl815

FILES/nsr/res/nwora.res - The NSR-DBM-ORA resource file.

SEE ALSOnsrnmo(1m), nsrnmostart(1m), nsroraclecat(1m)

233

Command Reference

Maintenance Comands nsrnmoinfo(1m)

NAMEnsrnmoinfo - Lists the NetWorker volumes required for a

restore.

SYNOPSISnsrnmoinfo [ -s server ] [ -c client ] [ -f file ] [ backup_piece ... ]

DESCRIPTIONThe nsrnmoinfo binary lists the NetWorker volumes required to restore the backup pieces specified on the command line or in a file. The volumes listed are the ones the NetWorker server intends to use at the time the nsrnmoinfo command is run. If volumes are removed from NetWorker devices or if volumes are deleted after the nsrnmoinfo command was run, the NetWorker server may use different volumes than those specified by nsrnmoinfo when the restore is performed.

The nsrnmoinfo command will list clones of volumes if the original volumes are not available.

OPTIONS

-s server

The NetWorker server to query.

-c client

The NetWorker client index that contains the backup pieces.

-f file

A file containing a list of backup pieces to query.

EXAMPLESDisplay the volumes for backup piece 'bckupc_1' for the client mars on the NetWorker server jupiter:

nsrnmoinfo -s jupiter -c mars bckupc_1

Display the volumes for backup pieces 'bckupc_1', 'bckupc_2', and 'bckupc_3' for the client mars on the NetWorker server jupiter:

nsrnmoinfo -s jupiter -c mars bckupc_1 bckupc_2 bckupc_3

Display the volumes for the backup pieces listed in file bckupc.txt for the client mars on the NetWorker server jupiter:

nsrnmoinfo -s jupiter -c mars -f bckupc.txt

Display the volumes for the backup pieces listed in file bckupc.txt and for 'bckupc_1' for the client mars on the NetWorker server jupiter:

nsrnmoinfo -s jupiter -c mars -f bckupc.txt bckupc_1

SEE ALSOnsrnmo(1m), nsrnmoadmin(1m), save(1m), savegrp(1m)


Command Reference

Maintenance Comands nsrnmostart(1m)

NAMEnsrnmostart - NSR-DBM-ORA scheduled backup executable program

SYNOPSISnsrnmostart <backup options>

DESCRIPTIONThe nsrnmostart executable program is used by the nsrnmo script to invoke scheduled backups on an Oracle database server. The nsrnmostart program launches Oracle Recovery Manager (RMAN) by passing the appropriate arguments.

The nsrnmo script runs the nsrnmostart program. Do not run it manually. nsrnmostart uses environment variables set in the nsrnmo script and backup options passed by nsrnmo (see nsrnmo(1m)). If the mandatory environment variables are not set in the nsrnmo script, nsrnmostart returns a nonzero status code. For details on the environment variables to set in the nsrnmo script, see Chapter 3, “Backup Procedures,”.

The directory where nsrnmostart is located must be included in the PATH environment variable in the nsrnmo script.

The nsrnmostart program writes debugging information into the debug file specified by the NSR_SB_DEBUG_FILE environment variable in the nsrnmo script.

The nsrnmostart program runs a pre-command script specified by the PRECMD environment variable--if the variable is set in the nsrnmo script. If the pre-command does not run successfully, nsrnmostart returns a nonzero status code to nsrnmo without performing the scheduled Oracle backup.

The nsrnmostart program invokes RMAN for Oracle database backups based on the inputs from nsrnmo and savegrp (see nsrnmo(1m) and savegrp(1m)).

When RMAN execution is finished, the nsrnmostart program runs a post-command script specified by the POSTCMD environment variable--if the variable is set in the nsrnmo script. If the post-command does not run successfully, nsrnmostart returns a nonzero status code to nsrnmo. Even if RMAN fails, the post-command is executed.

At the end of each scheduled Oracle proxy backup, the nsrnmostart program also backs up the NSR-DBM-ORA resource file.

SEE ALSOnsrnmo(1m), nsrnmoadmin(1m), save(1m), savegrp(1m)

235

Command Reference

Maintenance Comands nsroraclecat(1m)

NAMEnsroraclecat - Synchronizes the RMAN repository with the

NetWorker indexes.

SYNOPSISnsroraclecat

DESCRIPTIONNSR-DBM-ORA proxy copy backups are periodically pruned from the NetWorker indexes when the snapshots they are stored on expire. The nsroraclecat binary keeps RMAN repositories and the NetWorker indexes synchronized during the pruning operations by removing the backup piece entries of the snapshot backup from the RMAN repository. nsroraclecat is started automatically by nsrsnapck(1m) during pruning. The RMAN repository entries are removed by forking an RMAN session. The nsroraclecat binary should never be run manually.

The configuration of nsroraclecat is stored in the NSR-DBM-ORA resource file which is administered by nsrnmoadmin(1m). The following NSR-DBM-ORA resources specify the nsroraclecat configuration:

NSR_ORACLECAT_MODE

The default setting of NSR_ORACLECAT_MODE is 'undetermined'. When NSR_ORACLECAT_MODE is set to 'undetermined', NSR-DBM-ORA will not perform proxy copy backups. To use proxy copy backups, the NSR-DBM-ORA administrator must set NSR_ORACLECAT_MODE to 'enabled' or 'disabled'. If NSR_ORACLECAT_MODE is set to 'disabled', catalog synchronization will not be performed.

NSR_REMOVE_ON_FAILURE

When NSR_REMOVE_ON_FAILURE is set to 'FALSE', nsrsnapck(1m) will not remove the NetWorker index entries for backup pieces that were not successfully synchronized.

NSR_ORACLECAT_LOG_FILE

Designates the location of the nsroraclecat operations log file. All operational messages are written to this file. If NSR_ORACLECAT_LOG_FILE is not set, the operations messages go to the default log file /nsr/applogs/nsroraclecat.log.

NSR_ORACLECAT_DEBUG_FILE

Designates the location of the nsroraclecat debug log file. All debugging messages are written to this file. If NSR_ORACLECAT_DEBUG_FILE is not set, no debug logging is done.

An NSR-DBM-ORA SID resource must be configured for each Oracle database that has an RMAN repository that will be synchronized. An NSR-DBM-ORA SID resource has the following parameters:

NSR_ORACLE_CONNECT_FILE

Specifies the location of the file that contains the connection strings for the RMAN repository.


Command Reference

NSR_ORACLE_HOME

The ORACLE_HOME of the Oracle Server installation. It is used to locate the copy of RMAN that is used to prune the RMAN repository entries.

NSR_ORACLE_LIB_PATH

Some Oracle Server installations require the system shared library path to be set when using NSR-DBM-ORA. The name of the environment variable used to set the shared library path varies from system to system (it is LD_LIBRARY_PATH on Solaris, etc). NSR_ORACLE_LIB_PATH must be set to the value of the shared library path environment variable. If the Oracle Server installation does not require the shared library path, NSR_ORACLE_LIB_PATH does not have to be set.

NSR_ORACLE_SID

The ORACLE_SID of the Oracle database instance that performed the NSR-DBM-ORA proxy copy backup. NSR-DBM-ORA caches the ORACLE_SID in the NetWorker online indexes when NSR-DBM-ORA proxy copy backups are performed. The cached copy of the ORACLE_SID is then used by nsroraclecat as a key to locate the required NSR-DBM-ORA SID resource during catalog synchronizations.

NSR_ORACLE_TNS_ADMIN

The location of the Oracle Net configuration files required by the Oracle Server installation whose RMAN will perform the catalog synchronization. If the Oracle Net configuration files are in the default location, NSR_ORACLE_TNS_ADMIN does not have to be set.

FILES/nsr/applogs/nsroraclecat.log - The default nsroraclecat log file.

SEE ALSOnsrnmoadmin(1m), nsrsnapck(1m)

237

Command Reference

Maintenance Comands nsrorara(1m)

NAMEnsrorara - NSR-DBM-ORA Scheduled Backup and Recovery

Configuration Wizard Remote Agent.

SYNOPSISnsrorara [ -i ] [ -u ]

DESCRIPTIONnsrorara is part of NSR-DBM-ORA Scheduled Backup Configuration Wizard and NSR-DBM-ORA Recover Configuration Wizard, available starting with NSR-DBM-ORA release 5.0. The primary function of the nsrorara command is to perform operations on behalf of the wizard on the local or remote NSR-DBM-ORA host. Those operations include, but are not limited to, the database queries to the Oracle server. The secondary function is to register (-i) or unregister (-u) the NSR-DBM-ORA wizard as a plug-in on the client host during installation.

The nsrorara command is invoked only by:

(1) The NetWorker Management Console (NMC) during the creation or modification of an NSR-DBM-ORA scheduled backup configuration.

(2) The packaging subsystem during NSR-DBM-ORA installation or uninstallation.

The nsrorara command should not be run on the command line without instructions from Fujitsu Technology Solutions Technical Support.


Command Reference

Glossary 239

This glossary contains the definitions of terms found in this manual. Most of the terms are specific to the NSR-DBM-ORA software. For terms specific to the NetWorker software, refer to the latest NetWorker Administrator’s Guide.

A? Oracle placeholder for the main directory of the Oracle database instance identified

as $ORACLE_HOME.

administrator Person who normally installs, configures, and maintains software on network computers, and who adds users and defines user privileges.

API (applicationprogramming

interface)

Agreed-upon set of computer library routines, protocols, and tools used to communicate and accomplish tasks within software applications.

ASM (applicationspecific module)

Program that, when used in a directive, specifies the way a set of files or directories is to be backed up and recovered.

archived redo log Archived copy of a filled online Oracle redo log that preserves older redo log data for recovery operations. See also “redo log.”

attribute Feature of a NetWorker resource. It is a service or information that the “resource” provides.

auto mediamanagement

Feature that enables the storage device controlled by the NetWorker server to automatically label, mount, and overwrite a volume it considers unlabeled. Volumes that are eligible for reuse are also automatically recycled.

autochanger See “library.”

Bbackup Operation that saves data to a volume. An Oracle backup of several datafiles may

include several backup sets.

backup cycle See “scheduled backup cycle.”

backup group See “group.”

backup host See “proxy client host.”

Glossary


Glossary

backup level See “level.”

backup piece Binary file created during an NSR-DBM-ORA backup that corresponds to one save set and contains Oracle backup data in an RMAN-specific format from one or more database files. See “backup set.”

backup set Group of one or more backup pieces, created through the RMAN backup command during an NSR-DBM-ORA backup.

backup volume See “volume.”

bootstrap Save set that is essential for NetWorker disaster recovery procedures. The bootstrap consists of three components that reside on the NetWorker server: the media database, the resource database, and a server index.

browse policy NetWorker policy that specifies the time period during which backup entries for regular Oracle backups and proxy live backups are stored in the online NetWorker client file index and the associated backup files are readily accessible to users.

Ccatalog

synchronizationProcess that removes a proxy Oracle backup entry from the RMAN catalog when the corresponding backup piece is removed from the NetWorker indexes. See “NWORA resource file.”

client Computer, workstation, or fileserver whose data can be backed up and restored.

client file index Database maintained by the NetWorker server that tracks every data object, file, or “file system” backed up. The NetWorker server maintains a single index file for each client computer.

cluster Two or more independent network servers that operate and appear to clients as if they are a single unit. The cluster configuration enables work to be shifted from one server to another, providing "high availability" that allows application services to continue despite most hardware or software failures. See also “high-availability system.”

cold Oracle backup See “offline backup.”

command line Line on a display screen, also known as a command prompt or shell prompt, where you type software commands.

connection port Port that NetWorker processes use to perform backup and restore sessions through a firewall.

Ddaemon Process on UNIX systems that runs in the background and performs a specified

operation at predefined times or in response to certain events.

database Database instance of a third-party “DBMS” vendor. NSR-DBM-ORA backs up and restores Oracle database files.

DBA Abbreviation for database administrator, the person that is typically responsible for installing, configuring, and maintaining Oracle database systems.

Glossary 241

Glossary

DBMS Abbreviation for database management system, which refers to the primary architecture of an Oracle database.

deduplication backup Type of backup from the client to an Avamar server (NetWorker deduplication node), where the server identifies redundant data blocks on the client and backs up only the unique blocks (not entire files) that contain changes. Only a single instance of any unique data block is maintained on the server.

deferred live backup Type of proxy Oracle backup where an existing point-in-time copy (snapshot), created during an instant backup, is backed up to secondary storage such as tape. The snapshot is retained on the primary storage.

device ◆ Storage unit that reads from and writes to backup volumes (see “volume”) during backups and restores. The storage unit can be a tape device, optical drive, “autochanger,” or file connected to the “server” or “storage node.”

◆ When dynamic drive sharing (DDS) is enabled, refers to the access path to the physical drive.

directive Instruction that directs NetWorker software to take special actions on a given set of files for a specified client during a backup or recovery operation. Directives are ignored in manual (unscheduled) backups.

drive Hardware device through which media can be read or written to. See also “device.”

Eevent-based backup See “probe-based backup.”

exit code Indicator that specifies whether a backup or restore session succeeded.◆ Exit code of zero (0) indicates the session completed successfully.◆ Nonzero exit code indicates the session did not complete successfully.

expiration date Date when the status of a volume changes from read/write to read-only.

expired save set Save set whose browse time has been reached. The save set can no longer be browsed because it has been removed from the client file index.

Ffailover Safeguard capability that automatically switches activity from a failed or abnormally

terminated computer server, disk drive, or network to a redundant standby server, drive, or network, with little or no disruption of service. Failover is a feature of systems that require high reliability and continuous availability.

file index See “client file index.”

file system ◆ Software interface used to save, retrieve, and manage files on storage media by providing directory structures, data transfer methods, and file association.

◆ Entire set of all files.

fileserver Computer with disks that provides services to other computers on the network.

firewall System designed to prevent unauthorized access to or from a private network. All messages entering or leaving the intranet pass through the firewall, which examines each message and blocks those that do not meet the specified security criteria. There


Glossary

are several types of firewall techniques. The NetWorker software supports client backups from computers that are protected by packet filtering.

full backup See “level.”

Ggroup Client or group of client computers that are configured to back up files at a

designated time of day.

Hhigh-availability

systemSystem of multiple computers configured as cluster nodes on a network that ensures the application services continue despite a hardware or software failure. Each cluster node has its own IP address with private resources or disks that are available only to that computer.

host Computer on a network.

hot Oracle backup See “online backup.”

II18N

(internationalization)Capability of the NSR-DBM-ORA software to operate in a non-English environment or locale without itself generating non-ASCII data. After I18N is set up, NSR-DBM-ORA can process and display non-ASCII data that is passed to it by the operating system, NetWorker software, and Oracle software.

immediate livebackup

Type of proxy Oracle backup where a point-in-time copy (snapshot) is created during an instant backup and immediately backed up to secondary storage, such as tape. The snapshot is automatically deleted from the primary storage.

incremental backup See “level.”

index policy Policy that specifies how long file and data index entries should remain valid. Clients specify which index policies they wish to use.

instance Combination of processes that runs each time a database starts up.

instant backup Type of proxy backup that creates a snapshot of Oracle data as a point-in-time copy on a primary storage unit. See “proxy backup.”

instant restore Type of proxy restore that restores Oracle data from a mounted point-in-time copy that was created during an instant backup. See “proxy restore.”

Llevel Backup configuration option that specifies how much data is saved during a

scheduled or manual backup. An NSR-DBM-ORA backup level is specified by an RMAN command in the RMAN backup script only:◆ A full NSR-DBM-ORA backup backs up all of the data blocks in the database,

regardless of when they last changed.◆ An incremental NSR-DBM-ORA backup backs up only data blocks that have

changed since the last backup.

Glossary 243

Glossary

library Hardware device containing one or more removable media drives, as well as slots for pieces of media, media access ports, and a robotic mechanism for moving pieces of media between these components. Libraries automate media loading and mounting functions during backup and recovery. The term library is synonymous with autochanger, autoloader, carousel, datawheel, jukebox, and near-line storage.

live backup One of the following two types of proxy Oracle backup:◆ “deferred live backup”◆ “immediate live backup”

Mmanual backup Backup that a user performs from the client, also known as an unscheduled backup.

The user specifies the files, file systems, and directories to be backed up. A manual backup does not generate a “bootstrap” save set.

media Physical storage, such as magnetic tape, optical disk, or file system, to which backup data is written. See also “volume.”

media database Database that contains indexed entries about the storage volume location and the lifecycle status of all data and volumes the NetWorker server manages. See also “volume.”

media index See “media database.”

media manager NetWorker database that tracks save sets stored on backup volumes.

media pool Feature to sort backup data to selected storage volumes.

MML (mediamanagement library)

Library of media management routines provided by the NSR-DBM-ORA software and linked with the Oracle kernel software during the installation of NSR-DBM-ORA on the Oracle Server host.

mount To make a database available for use, or to place a removable tape or disk volume into a drive for reading or writing.

MTTR (mean time torecover)

Set time required to perform an instance or media recovery for an Oracle database. For example, you might set 10 minutes as the goal for media recovery from a disk failure.

multiplexing NetWorker feature that permits data from more than one save set to be simultaneously written to the same storage device.

NNetWorker Network-based software product that backs up and restores file systems.

NetWorkerAdministrator program

The following NetWorker Administrator programs are available:◆ Command Line Interface: nsradmin program◆ Graphical User Interface: NetWorker Management Console (NMC) as of

NetWorker Release 7.3 or networker for earlier releases.

NetWorker client See “client.”

NetWorker Module forOracle

Synonym for NSR-DBM-ORA.


Glossary

NetWorker resource See “resource.”

NetWorker server See “server.”

NetWorker storagenode

See “storage node.”

NMO Abbreviation for “NetWorker Module for Oracle.”

notification Message generated and sent to the NetWorker administrator about important NetWorker events.

nsrhost Logical hostname of the computer that is the NetWorker server.

NWORA resource file Resource file installed during the NSR-DBM-ORA installation, which must be extended to include specific additional resources to enable proxy Oracle backups and (optionally) catalog synchronization.

Ooffline backup Offline Oracle backup performed while the Oracle instance is shut down and

unavailable to users.

online backup Online Oracle backup performed while the Oracle instance is running and available to users.

online indexes Databases located on the NetWorker server that contain all the information pertaining to the client backups (“client file index”) and backup volumes (“media database”).

operator Person who monitors the server status, loads backup volumes into storage devices, and executes the day-to-day NetWorker tasks.

Oracle EnterpriseManager

Oracle Enterprise Manager Backup Management Tools, which include an optional graphical user interface to the RMAN utility.

Oracle10g Server Computer running an Oracle10g release 10.x DBMS. See “DBMS.”

Oracle11g Server Computer running an Oracle11g release 11.x DBMS. See “DBMS.”

override Different backup level that is used in place of the regularly scheduled backup level.

Pparallelism Method that backs up or restores data for multiple clients, or multiple save sets for

one client, at the same time.

pathname Set of instructions to the operating system for accessing a file:◆ An absolute pathname indicates how to find a file starting from the root directory

and working down the directory tree.◆ A relative pathname indicates how to find a file starting from the current location.

physical host Node or host that forms part of a cluster.

point-in-time copy Fully usable copy of a defined collection of data, such as a consistent file system, database, or volume, which contains an image of the data as it appeared at a single

Glossary 245

Glossary

point in time. A point-in-time (PiT) copy is also called a shadow copy or a snapshot. A snapshot of Oracle data is created on a supported type of primary storage during an instant backup.

policy Set of constraints that specify how long the save sets for a client are available for recovery:◆ For a regular Oracle backup and proxy live backup, the NetWorker Client

resource specifies a browse policy and a retention policy. The nsr_policy(5) man page and NetWorker Command Reference Guide provide more information.

◆ For a proxy instant backup, the NetWorker Group resource specifies a snapshot policy that is configured with a Snapshot Policy resource. The NetWorker PowerSnap Module documentation provides more information.

policy uniformity Consistency of the browse and retention policies in a group of co-dependent save sets from the same scheduled backup cycle or save set bundle, enforced by NSR-DBM-ORA to ensure that incremental backups do not persist after other backups they depend on have expired.

pool Feature to sort backup data to selected storage volumes.

PowerSnap EMC technology that provides point-in-time snapshots of data to be backed up. Applications that are running on the host system continue to write data during the snapshot operation, and data from open files is included in the snapshots.

PowerSnap Module EMC software module that exports services of a storage subsystem by interfacing with vendor-specific APIs. The module is independent of applications and backup and recovery interfaces. NSR-DBM-ORA operates with a PowerSnap Module to perform proxy Oracle backups.

preconfigured Initial default selections or configurations for software features.

primary storage Server storage subsystem that contains the Oracle source data and any persistent snapshot backups of the data. The NetWorker PowerSnap Module documentation provides information on the supported types of primary storage.

probe Query operation to determine if a specified condition is met on a client.

probe-based backup Type of scheduled backup, also known as an event-based backup, where the NetWorker server initiates the backup only when specified conditions are met, as determined by one or more probes.

proxy client Surrogate client that performs the NetWorker save operation for the client that requests the backup. A proxy client is required to perform a serverless backup.

proxy client host Host used in proxy Oracle backups that is separate from the Oracle Server host, with access to the primary storage unit. During a proxy live backup, either the Oracle Server host or proxy client host backs up an Oracle database point-in-time copy (snapshot) to secondary storage.

proxy backup Backup of Oracle data that creates a point-in-time copy (snapshot) on primary storage through the PowerSnap Module. The snapshot is optionally backed up to secondary storage, with or without deletion of the snapshot on primary storage. Two types of proxy Oracle backup are the “instant backup” and “live backup.”


Glossary

proxy restore Restore of Oracle data from a proxy Oracle backup through the PowerSnap Module. Three types of proxy Oracle restore are the “instant restore,” “rollback restore,” and “restore from secondary storage.”

purging Process of deleting all entries for files on a volume from the client file index, but allowing entries for the save sets to remain in the media database.

RRDBMS (relational

databasemanagement system)

Type of “DBMS” that stores data in the form of related tables.

recover To apply archived redo logs and online redo logs to an Oracle database to make the database consistent with a given point in time.

Recovery Catalog Collection of Oracle database tables maintained by RMAN, including information about Oracle backup sets and pieces, image and proxy copies, archived redo logs, stored scripts, and the “target database” schema.

recyclable save set Save set whose browse and retention policies have expired. Recyclable save sets are removed from the media database.

recyclable volume Volume whose data has exceeded both its browse and retention policies and is now available to be relabeled and reused.

redo log Online log of an Oracle database, consisting of at least two redo log files (separate from the datafiles) that record all the most current changes made in the database instance. See also “archived redo log.”

remote device Storage device that is attached to a NetWorker storage node.

resource Component of either the NetWorker server configuration or the NWORA resource file:◆ A NetWorker resource describes the NetWorker server or its clients. Devices,

schedules, clients, groups, and policies are examples of NetWorker resources. Each resource has attributes that define its properties.

◆ The NWORA resource file contains resources that enable proxy backups and (optionally) catalog synchronization.

resource database Database of information about each of the configured NetWorker resources.

restore Process of retrieving individual datafiles from backup storage and copying the files to disk.

restore from secondarystorage

Type of proxy restore that restores a proxy backup from a secondary storage medium, such as tape. See “proxy restore.”

retention policy NetWorker policy setting that determines how long save set entries for a regular Oracle backup or proxy live backup are retained in the NetWorker media database and the corresponding backup data is recoverable.

RMAN (RecoveryManager)

Oracle utility that acts as an intelligent interface to Oracle databases and works with third-party media management products, such as NSR-DBM-ORA, to back up and restore Oracle database objects.

Glossary 247

Glossary

RMAN catalog RMAN repository that stores information about each Oracle backup piece in either a control file of the target database or an RMAN Recovery Catalog.

RMAN script Script of RMAN commands used to perform an NSR-DBM-ORA backup or restore or an Oracle database duplication.

rollback restore Type of proxy Oracle restore that restores a specific point-in-time copy (snapshot) of Oracle data to the source location by using the hardware’s particular capabilities. A rollback is a destructive save set restore. See “proxy restore.”

rollover save set Save set that is backed up to tape from a snapshot. Whether this snapshot is retained or not depends on a snapshot policy. When a snapshot is rolled over to tape, entries are made in the client file index and media database, which enable the save set to be browsed for restore.

root On UNIX systems. The superuser account. The highest level of the system directory structure.

RPC (remoteprocedure call)

Protocol that the NetWorker server uses to perform client requests over a network.

Ssave NetWorker command that backs up client files to backup volumes and makes data

entries in the online index.

save set Group of files or a file system from a single client computer, which is backed up on storage media.

save set bundle Group of co-dependent save sets from the same scheduled backup cycle of an Oracle database object, assembled by NSR-DBM-ORA into a bundle according to configuration settings.

save set bundle join Creation, during an incremental scheduled NSR-DBM-ORA backup, of a combined save set bundle from co-dependent save sets in different save set bundles.

save set bundling Process whereby NSR-DBM-ORA automatically creates a save set bundle for each scheduled backup cycle of an Oracle database object, by grouping all the dependent save sets from the same backup cycle into a save set bundle.

save set ID Internal identification number that NetWorker software assigns to a save set.

save set status NetWorker attribute that indicates whether a save set is browsable, recoverable, or recyclable. The save set status also indicates whether the save set was successfully backed up.

save stream Data and save set information that is written to a storage volume during a backup. A save stream originates from a single save set.

scanner NetWorker command used to read a backup volume when the online indexes are not available.

scheduled backup Type of backup that is configured to start automatically at a specified time for a group of one or more NetWorker clients. A scheduled backup generates a “bootstrap” save set.


Glossary

scheduled backupcycle

Full or level 0 backup of an Oracle database object and all the subsequent incremental backups that are dependent on the level 0 backup. If “save set bundling” is enabled, a separate save set bundle is created for each scheduled backup cycle.

secondary storage Storage library attached to the NetWorker server or storage node, used to store traditional or snapshot backups.

server Computer on a network that runs the NetWorker server software, contains the online indexes, and provides backup and restore services to the clients and storage nodes on the same network.

server index File that lists all the server files backed up during a scheduled backup.

service port Port used by a server or storage node to listen for backup and restore requests from clients through a firewall.

shared disk Storage disk that is connected to multiple nodes in a cluster.

shell prompt Cue for input in a shell window where you type a command.

SID (system identifier) Unique name for an Oracle database instance. This value is typically set in an ORACLE_SID parameter.

silo Repository for holding hundreds or thousands of volumes. Silo volumes are identified by barcodes, not by slot numbers.

skip Backup level in which designated files are skipped and not backed up. See also “level.”

snap set Group of files, volumes, or file systems from a single client that describes the collection of data for which a point-in-time copy is created on an external disk subsystem, such as a storage array.

snapshot Point-in-time copy of Oracle data created on a supported type of primary storage during an instant backup.

snapshot policy Policy configured through a NetWorker Snapshot Policy resource, to control the lifecycle of snapshots created during instant backups. The snapshot policy specifies the frequency of instant backups, and how long snapshots are retained before recycling.

ssid See “save set ID.”

staging Moving data from one storage medium to a less costly medium, and later removing the data from its original location.

stand-alone device Type of “storage device” that contains a single drive for backing up data. Stand-alone devices cannot store or automatically load backup volumes.

storage device Hardware that reads and writes data during backup, restore, or other NetWorker operations.

storage node Storage device physically attached to a computer other than the NetWorker server, whose backup operations are administered from the controlling NetWorker server.

system administrator Person normally responsible for installing, configuring, and maintaining NetWorker software.

Glossary 249

Glossary

Ttablespace Oracle database structure that consists of one or more datafiles.

target database Database that the NetWorker server backs up as a safeguard against data loss.

temporary enabler Code that enables operation of the NSR-DBM-ORA software for an additional 45 days beyond the evaluation period.

TNS (TransparentNetwork Substrate)

Oracle networking technology that provides a single interface to all standard network protocols.

traditional restore Type of proxy Oracle restore, performed as a regular Oracle restore that restores a point-in-time copy from a secondary storage medium, such as tape. See “proxy restore.”

traditional storage See “secondary storage.”

Uunscheduled backup See “manual backup.”

user Person who uses NetWorker software from a computer to back up and restore files.

Vversions Date-stamped collection of available backups for any single file.

volume Backup volume used to store backup data. Backup data cannot be stored on an archive volume or a clone volume.

volume ID Internal identification that the NetWorker software assigns to a backup volume.

volume name Name assigned to a backup volume when it is labeled.

volume pool See “pool.”


Glossary


AADDRESS_LIST parameter 138

Administrator attributein Server resource 59

Administrator program, NetWorkerStop button 105

Administrator, NetWorkerClient resource 81, 163

Device resource 62

Group resource 80, 163

Label Template resource 63

Pool resource 62, 162

Schedule resource 81

Server resource 59

Snapshot Policy resource 144, 148, 162

User Group resource 60

AES encryption 116, 191

allocate channel commandparms option 204

API, SBT 49, 54

archived redo logbackup 19, 124, 142

restore 142

sharing across RAC nodes 142

attributes for resourcesClient resource 81

Server resource 59

autochanger 62, 99, 117

automatic catalog synchronization for proxybackups 174, 183

automatic channel allocation 24, 71, 73

Bbackup

Client resource 81, 163

command 158, 204

command (pool option) 200

command (trace option) 206

copies 25

devices, multiple 44

error messages 211, 212


level (full or incremental in RMAN script) 62


sessions 51

backup and recovery system 46, 47, 146

Backup Command attributein Client resource 82, 163, 186, 187

backup copies during manual backups 25

backup current control file command 126

Backup Snapshots attributein Snapshot Policy resource 152

backup spfile command 126

backup typesarchived redo log 19, 124, 142

control file 124, 126

deduplication 20, 86

deferred live 144, 152, 166

immediate live 145, 150

instant 144, 148, 166, 174

manual 19, 45, 51, 98

NetWorker bootstrap 44, 50, 100, 149

NWORA resource file 167, 169

password file 124

probe-based 21, 90

proxy 22, 144, 148, 164

Index


Index

backup types (Cont.)registry files 124

scheduled 19, 45, 104, 148

BACKUP_TAPE_IO_SLAVES parameter 206

bootstrap, NetWorker 44, 50, 100, 149

Browse Policy attributein Client resource 82, 163

browse policy uniformity 33

bundling, save set 35

Ccanceling

manual backup 101

nonresponding backup 101

proxy backup 167

scheduled backup 73, 105

catalog synchronization for proxy backups 174,183

change backuppiece...unavailable command 206

change...crosscheck command 200, 206

change...delete command 183

channel option, send command 202

client file index 44, 168, 169

Client resource 81, 163

Client resource attributesBackup Command 82

Browse Policy 82

Group 82

Name 82

Remote Access 83

Retention Policy 83

Save Set 83

Schedule 83

Client Retries attributein Group resource 185

Client Retries attribute in Group resource 186

cluster systems 27, 132, 185

commandallocate channel, parms option 204

backup 158, 204

backup (pool option) 200

backup (trace option) 206

backup current control file 126

backup spfile 126

change backuppiece...unavailable 206

change...crosscheck 200, 206

change...delete 183

configure channel, parms option 160, 190

crosscheck 200, 206

delete expired backup 200

nsrnmoinfo 117

restore 204

rman 77

rman send 190, 203, 204

savefs 49, 148, 150

savegrp 49, 100, 148, 150

send 160, 190, 201

send (channel option) 202

send (device_type option) 202

send (NSR_ENV keyword) 201

send (precedence rules) 204

set duplex 205, 206

setenv 76, 190

su 78, 79

Command attributein Probe resource 91

Command Options attributein Probe resource 91

configurationClient resource 81, 163

Device resource 62



manual backup 98

Oracle system 59


proxy backup 157

roadmap 58


scheduled backup 104

Server resource 59



configuration wizard 58

configure channel commandparms option 160, 190

connection file, for catalog synchronization 177,179, 180

control filebackup 124, 126

control files, mirrored 124

crosscheck command 200, 206


Index

Ddatabase

manual backup 45, 51, 98

scheduled backup 45, 104, 148

Datazone pass phrase attributein Server resource 59

DBID 124, 125

deduplicationbackup 20, 86

restore 20

deferred live backup 144, 152, 166

delete expired backup command 200

destructive restoresoverview 145

Device resource 62

device_type option, send command 202

devices, multiple 44

disaster recoverypreparation 124

types 127

documentationrelated 9

Eelectronic access to Fujitsu information 11

email notification 84

enablingpolicy uniformity 86

save set bundling 85

encryptionAES encryption 116, 191

environment variableLC_ALL 64

NLS_LANG 64

error messagesNSR-DBM-ORA 212

RMAN 211

Ffailover

backup 133, 134, 139

connect-time 133, 138

proxy backup 185

FAILOVER parameter 138

force_rollback option, not supported 171

Fujitsuservice 11

GGroup attribute

in Client resource 82, 163


GUI, NetWorkerStop button 105

II18N (internationalization) 32, 64

immediate live backup 145, 150

initialization parameter fileinitoracle_sid.ora 139

PFILE 124, 125, 126

SPFILE 124, 126

INSTANCE_NAME parameter 139

instant backup 144, 148, 166, 174

instant restore 145, 153, 170

internationalization (I18N) 32, 64

LLabel Template resource 63

LC_ALL environment variable 64

LD_LIBRARY_PATH parameter 75, 76

live backupdeferred 144, 152, 166

immediate 145, 150

LOCAL_LISTENER parameter 139

LOG_THRESHOLD parameter 91

Mmandatory parameters 76, 91, 192, 193, 194, 195


canceling 101

monitoring 101

procedure for running 98

roadmap 98

manual catalog synchronization for proxy bak-kups 183

media database 44, 168, 169

media managementstorage devices 62

volume pools 62

media management library (MML) 49, 147, 153,155

mirroredcontrol files 124

online redo logs 124


Index

MML 49, 147, 153, 155

MML catalog 174

monitoringmanual backup 101


multiple backup devices 44

NName attribute

in Client resource 82

in Probe resource 91

in Server resource 59

network fileslistener.ora 124, 125, 139

sqlnet.ora 124, 125

tnsnames.ora 124, 125, 138

NetWorkerbootstrap backup 44, 50, 100, 149

client file index 44, 168, 169

configurationClient resource 163

media database 44, 168, 169

media database service, nsrmmdbd 51, 54, 149

media service, nsrmmd 51, 54

processesrestore 156

software features 44

staging 35

NetWorker Administrator programStop button 105

NetWorker configurationClient resource 81

Device resource 62




roadmap 58


Server resource 59


NetWorker PowerSnap Modules 22, 45, 61, 144,156

NLS_LANG environment variable 64

nmo.messages.raw file 207, 212

NOCATALOG mode 140

nonresponding Oracle backup, canceling 101

notification by email 84

NSR_AES_ENCRYPTION parameter 72, 191

NSR_BUNDLING parameter 85

NSR_CHECKSUM parameter 191

NSR_CLIENT parameter 120, 171, 191

NSR_COMPRESSION parameter 191

NSR_DATA_MOVER parameter 161

NSR_DATA_VOLUME_POOL* parameters63, 74, 192, 206

NSR_DEBUG_FILE parameter 77, 91, 193

NSR_DEBUG_LEVEL parameter 193

NSR_DEDUP_BACKUP parameter 193

NSR_DEDUP_CACHE_ENABLED parameter193

NSR_DEDUP_CACHE_TAG parameter 193

NSR_DEDUP_CHUNK_SIZE parameter 194

NSR_DEDUP_NODE parameter 194

NSR_DPRINTF parameter 194

NSR_ENCRYPTION_PHRASES parameter116, 194

NSR_ENV keyword in send command 201

NSR_GROUP parameter 74, 195

NSR_INCR_EXPIRATION parameter 86

NSR_MAX_STREAMS parameter 161

NSR_MMDB_RETRY_TIME parameter 195

NSR_NO_BUSY_ERRORS parameter 195

NSR_NO_MULTIPLEX parameter 195

NSR_NWPATH parameter 177, 195

NSR_ORACLE_CONNECT_FILE parameter91, 179

NSR_ORACLE_HOME parameter 179

NSR_ORACLE_LIB_PATH parameter 179

NSR_ORACLE_NLS_LANG parameter 65

NSR_ORACLE_NLS_LANG parameter resour-ce 177

NSR_ORACLE_SID parameter 179

NSR_ORACLE_TNS_ADMIN parameter 179

NSR_ORACLECAT_DEBUG_FILE parameterresource 177

NSR_ORACLECAT_LOG_FILE parameter re-source 177

NSR_ORACLECAT_MODE parameter resour-ce 177, 184

NSR_PROXY_PFILE parameter 195

NSR_PS_SAVE_PARALLELISM parameter161

NSR_RECOVER_POOL parameter 196


Index

NSR_REMOVE_ON_FAILURE parameter re-source 177

NSR_RETENTION parameter 83

NSR_RETENTION_DISABLED parameter 196

NSR_RMAN_ARGUMENTS parameter 77

NSR_SAVESET_BROWSE parameter 82, 196

NSR_SAVESET_RETENTION parameter 196

NSR_SB_DEBUG_FILE parameter 77

NSR_SERVER parameter 74, 120, 196

NSR_SERVER_NIC parameter 197

NSR-DBM-ORAAPI, SBT 49, 54

error messages 212

features 45

MML 147, 153, 155

troubleshooting 210

NSR-DBM-ORA API, SBT 54

NSR-DBM-ORA internationalization (I18N) 32, 64

NSR-DBM-ORA wizard 58

nsrexecd service 49, 51, 54, 148

nsrindexd service 51, 54, 149

nsrmmd service 51, 54

nsrmmdbd service 51, 54, 149

nsrnmo script 50, 76, 148, 151

parameters 75

nsrnmoadmin program 177, 179, 180, 182

nsrnmoinfo command 117

nsrnmoprobe program 22, 91, 92

nsrnmostart program 50, 148, 151

nsroraclecat program 175, 183, 185

nsrsnapck program 148, 151, 183, 184, 185

NWORA parameter resourcesNSR_ORACLE_NLS_LANG 177

NSR_ORACLECAT_DEBUG_FILE 177

NSR_ORACLECAT_LOG_FILE 177

NSR_ORACLECAT_MODE 177, 184

NSR_REMOVE_ON_FAILURE 177

NWORA resource file 176, 178, 179

backup 167, 169

NWORA SID resources 178, 179

nworapc directory 171

Oonline NetWorker indexes 44

online redo logs, mirrored 124

Oracle

backup and recovery system 46, 47, 146

DBID 124, 125

manual backup script 71, 159, 204

media management library (MML) 49

mirroredcontrol files 124

online redo logs 124

password file 73

recover 15, 121

Recovery Catalog 46, 47, 146

Recovery Catalog databaseconnection to 73

Recovery Manager (RMAN) 46, 54, 146, 153

restore 15, 45, 115, 117, 142

sbtio.log file 207, 212

scheduled backup script 73

Server sessions 51

System Backup to Tape (SBT) API 49, 54

system configuration 59

Oracle ASM 41

See also Oracle Automated StorageOracle Automated Storage 41

Oracle Enterprise ManagerSee Oracle Enterprise Manager Backup Ma-

nagement ToolsOracle Enterprise Manager Backup ManagementTools 49, 54, 99, 120, 153

Oracle Net 59, 80, 134, 135, 138

ORACLE_HOME parameter 76, 77, 91

ORACLE_SERVICE parameter 91

ORACLE_SID parameter 77

ORACLER_USER parameter 78

Pparallel save stream 44

Parallelism attributein Client resource 163

in Server resource 59

parameterADDRESS_LIST 138

FAILOVER 138

INSTANCE_NAME 139

LOCAL_LISTENER 139

SERVICE_NAME 139

SID_LIST_LISTENER 139


Index

parameter resourcesNSR_ORACLE_NLS_LANG 177

NSR_ORACLECAT_DEBUG_FILE 177

NSR_ORACLECAT_LOG_FILE 177

NSR_ORACLECAT_MODE 177, 184

NSR_REMOVE_ON_FAILURE 177

parametersBACKUP_TAPE_IO_SLAVES 206

in nsrnmo script 75

LD_LIBRARY_PATH 75, 76

LOG_THRESHOLD 91

mandatory 76, 91, 192, 193, 194, 195, 196

NSR_AES_ENCRYPTION 72, 191

NSR_BUNDLING 85

NSR_CHECKSUM 191

NSR_CLIENT 120, 171, 191

NSR_COMPRESSION 191

NSR_DATA_MOVER 161

NSR_DATA_VOLUME_POOL* 63, 74,192, 206

NSR_DEBUG_FILE 77, 91, 193

NSR_DEBUG_LEVEL 193

NSR_DEDUP_BACKUP 193

NSR_DEDUP_CACHE_ENABLED 193

NSR_DEDUP_CACHE_TAG 193

NSR_DEDUP_CHUNK_SIZE 194

NSR_DEDUP_NODE 194

NSR_DPRINTF 194

NSR_ENCRYPTION_PHRASES 116, 194

NSR_GROUP 74, 195

NSR_INCR_EXPIRATION 86

NSR_MAX_STREAMS 161

NSR_MMDB_RETRY_TIME 195

NSR_NO_BUSY_ERRORS 195

NSR_NO_MULTIPLEX 195

NSR_NWPATH 177, 195

NSR_ORACLE_CONNECT_FILE 91, 179

NSR_ORACLE_HOME 179

NSR_ORACLE_LIB_PATH 179

NSR_ORACLE_NLS_LANG 65

NSR_ORACLE_SID 179

NSR_ORACLE_TNS_ADMIN 179

NSR_PROXY_PFILE 195

NSR_PS_SAVE_PARALLELISM 161

NSR_RECOVER_POOL 196

NSR_RETENTION 83

NSR_RETENTION_DISABLED 196

NSR_RMAN_ARGUMENTS 77

NSR_SAVESET_BROWSE 82, 196

NSR_SAVESET_RETENTION 196

NSR_SB_DEBUG_FILE 77

NSR_SERVER 74, 120, 196

NSR_SERVER_NIC 197

ORACLE_HOME 76, 77, 91

ORACLE_SERVICE 91

ORACLE_SID 77

ORACLE_USER 78

PATH 76

POSTCMD 79

PowerSnap 158, 160, 161

PRECMD 78

RESTORE_TYPE_ORDER 161, 170, 171, 172


TNS_ADMIN 80

parms optionallocate channel command 204

configure channel command 160, 190

password file 73

backup 124

PATH parameter 76

persistent settings 23, 24, 73

PFILE 124, 125, 126

physical cluster client, proxy backups from 187

point-in-time copy 22, 144, 145, 154, 162, 163, 168

policybrowse 82

retention 83

policy uniformity 33

pool option, backup command 200


pool, volumedefined 62

label template for 62

pool types 62

POSTCMD parameter 79

postcommand script 79, 125, 128

postprocessing scripts 79

PowerSnap Modules 22, 45, 61, 144, 156

PowerSnap parameters 158, 160, 161

precedence rules for send command 204

PRECMD parameter 78

preprocessing scripts 78

primary storage 22, 144, 147, 157, 164, 165


Index

Probe resource attributes

Command 91

Command Options 91

Name 91

probe-basedbackup 21, 90

processesrestore 156

programnsrd 49, 51, 54, 148, 150

nsrexecd 49, 148

nsrindexd 51, 54, 149

nsrmmd 51, 54

nsrmmdbd 51, 54, 149

nsrnmo 50, 148, 151

nsrnmoadmin 177, 179, 180, 182

nsrnmoprobe 22, 91, 92

nsrnmostart 50, 148, 151

nsroraclecat 175, 183, 185

nsrsnapck 148, 151, 183, 184, 185

savefs 49, 148, 150

savegrp 49, 148, 150

proxybackup 22, 144, 148, 164

restore 145, 153, 170

psrollback.res file 172

RRAC nodes as storage nodes 137

RAC systems 27, 132, 135

recover 15, 121


Recovery Catalog databaseconnection to 73

Recovery Catalog, backup 124, 126

Recovery Manager (RMAN) 46, 49, 54, 146, 153

backup scripts 70

commands 199

error messages 211

features 47

manual backup scripts 71

proxy backup scripts 158, 159

proxy restore scripts 170


restore scripts 115

scheduled backup scripts 73, 74

redo logs, mirrored 124

registry filesbackup 124

relocating files during proxy restores 172, 173

Remote Access attributein Client resource 83, 163, 186, 187

remote procedure call (RPC) protocol 44

resourcetypes of

Client 81, 163

Device 62

Group 80, 163

Label Template 63

Pool 62, 162

Schedule 81

Server 59

Snapshot Policy 144, 148, 162

User Group 60

restorearchived redo log 142

command 204

deduplication 20

determining required volumes 117

from secondary storage 145

instant 145, 153, 170

Oracle data 15, 45, 115, 117

processes 156

proxy 145, 153, 170

rollback 145, 153, 170, 172

to a different host 120, 173

RESTORE_TYPE_ORDER parameter 161, 170,171, 172

Retain Snapshots attributein Snapshot Policy resource 174

Retention Policy attributein Client resource 83, 163

retention policy uniformity 33

RMANSee Recovery Manager (RMAN)

RMAN catalog 174

rman command 77

RMAN repository 174

rman send command 190, 203, 204

roadmapmanual backup 98


rollback restore 145, 153, 170, 172


Index

RPC (remote procedure call) 44

SSave Set attribute

in Client resource 83, 163, 186

save set bundling 35

savefs command 49, 148, 150

savegroup completion report for proxy backup166

savegrp command 49, 100, 148, 150

SBT API 49, 54

sbtio.log file 207, 212

Schedule attributein Client resource 83, 163



automatic backup of client indexes and ser-ver bootstrap 100

canceling 73, 105

configuring Group resource 80

configuring Schedule resource 81

monitoring 106

parameters 75, 104

postprocessing scripts 79

preprocessing scripts 78

roadmap 104

troubleshooting 211

scriptnsrnmo 50, 148, 151

secondary storage 144, 145

send command 160, 190, 201

channel option 202

device_type option 202

NSR_ENV keyword 201

precedence rules 204

Server resourceattributes

Datazone pass phrase 59

Name 59

Parallelism 59

Server resource attributesAdministrator 59

service 11

nsrd 49, 51, 54, 148, 150

nsrexecd 49, 148

nsrindexd 51, 54, 149

nsrmmd 51, 54

nsrmmdbd 51, 54, 149

service for customers 11

SERVICE_NAME parameter 139

set duplex command 205, 206

setenv command 76, 190

SID_LIST_LISTENER parameter 139

silo 62

snapshot 22, 144, 145, 154, 162, 163, 168

Snapshot attribute in Group resource 164, 165


Snapshot Pool attribute in Group resource 162

SPFILE 124, 126

staging 35

Stop buttonin NetWorker Administrator program 105

stoppingmanual backup 101

nonresponding backup 101


storage devicesDevice resource 62

volume pool 62

Storage Nodes attributein Client resource 163

su command 78, 79

support 11

synchronizationautomatic catalog 174, 183

manual catalog 183

System Backup to Tape (SBT) API 49, 54

Ttablespace



target databaseconnection to 73

TNS_ADMIN parameter 80

trace option, backup command 206

troubleshootingmanual backup 210

NSR-DBM-ORA 210



Index

types of backuparchived redo log 19, 124, 142

manual 45, 51, 98

NetWorker bootstrap 44, 50, 100, 149

scheduled 45, 104, 148

Uuniformity, policy 33


Vvirtual cluster client, proxy backups from 186

volume poolDefault pool 74

defined 62

resource 62, 162

specifying 74

types 62

volumes, determining for restore 117

Wweb address 11

wizard, configuration 58


Index

Date post:	13-Jan-2020
Category:	Documents
Upload:	others
View:	12 times
Download:	0 times

NSR-DBM-ORA V5.0 Administration Guidemanuals.ts.fujitsu.com/file/8960/nmoag50.pdf · 2011-10-06 ·...

Documents