You are on page 1of 256

EMC® NetWorker®

Module for Oracle


Release 5.0

Administration Guide
P/N 300-006-990
REV A02

EMC Corporation
Corporate Headquarters:
Hopkinton, MA 01748-9103
1-508-435-1000
www.EMC.com
Copyright © 2003 - 2009 EMC Corporation. All rights reserved.

Published June, 2009

EMC believes the information in this publication is accurate as of its publication date. The information is subject to change
without notice.

THE INFORMATION IN THIS PUBLICATION IS PROVIDED “AS IS.” EMC CORPORATION MAKES NO
REPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE INFORMATION IN THIS PUBLICATION,
AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR
PURPOSE.

Use, copying, and distribution of any EMC software described in this publication requires an applicable software license.

For the most up-to-date regulatory document for your product line, go to the Technical Documentation and Advisories section
on EMC Powerlink.

For the most up-to-date listing of EMC product names, see EMC Corporation Trademarks on EMC.com.

All other trademarks used herein are the property of their respective owners.

2 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Contents

Preface

Chapter 1 Introduction
Naming conventions used in this guide............................................................... 20
Importance of backups............................................................................................ 20
Product features ....................................................................................................... 20
Scheduled compared to manual backups ..................................................... 21
Deduplication backups and restores ............................................................. 21
Probe-based backups ....................................................................................... 23
Proxy backups and restores ............................................................................ 24
VMware support............................................................................................... 25
Archived redo log backups ............................................................................. 25
Control file autobackup ................................................................................... 25
Automatic channel allocation ......................................................................... 26
Backup and restore optimization ................................................................... 26
Backup copies.................................................................................................... 27
Backup of backup sets...................................................................................... 29
Cluster and RAC backups and restores......................................................... 29
Configuration wizards..................................................................................... 30
Internationalization (I18N) and localization (L10N) ................................... 33
Oracle Data Guard support............................................................................. 36
Parallelism ......................................................................................................... 36
Policy uniformity .............................................................................................. 37
Restartable backups.......................................................................................... 37
Retention policies ............................................................................................. 38
Save set bundling.............................................................................................. 39
Other Oracle features ....................................................................................... 44
Software used in the NMO environment ............................................................. 47
NetWorker software ......................................................................................... 47
NMO software................................................................................................... 48
NetWorker PowerSnap Module software ..................................................... 48
Oracle backup and recovery system .............................................................. 49
NMO components ................................................................................................... 51
NMO backup and restore processes ..................................................................... 52
Regular scheduled backup processes ............................................................ 53
Regular manual backup processes................................................................. 54
Deduplication backup processes.................................................................... 55
Regular restore processes ................................................................................ 57
Deduplication restore processes..................................................................... 58

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 3


Contents

Chapter 2 Software Configuration


Configuration roadmap.......................................................................................... 62
Verify the Oracle Server configuration................................................................. 63
Verify the NetWorker resources ............................................................................ 63
NetWorker Server resource.............................................................................. 63
NetWorker user group privileges ................................................................... 64
NetWorker Schedule resource......................................................................... 66
NetWorker Device resources ........................................................................... 66
NetWorker volume pools................................................................................. 66
Firewall support ................................................................................................ 67
Configuring I18N support ..................................................................................... 67
Requirements for I18N support ...................................................................... 68
Configure I18N support ................................................................................... 68
Configure L10N support ........................................................................................ 69
Migrating a legacy configuration with the nsrnmoadmin command ..............69
Requirements for using the nsrnmoadmin command for migration ........ 70
Migration command syntax and options ...................................................... 71
Configuring a backup with the NMC wizard......................................................72
About the backup configuration wizard ....................................................... 72
Requirements for using the backup configuration wizard ......................... 73
Configure a backup with the wizard ............................................................. 73
Configuring a backup with the NMC legacy method ........................................75
About backup configuration with the legacy method................................. 75
Create RMAN scripts for backups.................................................................. 75
Customize the nsrnmo script .......................................................................... 79
Configure a Group resource with NMC ........................................................ 86
Configure a Schedule resource with NMC.................................................... 87
Configure a Client resource with NMC ......................................................... 87
Configure save set bundling............................................................................ 91
Configure policy uniformity ........................................................................... 92
Configuring a deduplication backup ................................................................... 93
Requirements for a deduplication backup .................................................... 94
Best practices for a deduplication backup..................................................... 94
Configure a deduplication backup................................................................. 95
Configuring a probe-based backup ...................................................................... 97
Requirements for a probe-based backup ....................................................... 97
Configure a probe-based backup.................................................................... 97

Chapter 3 Backup Procedures


Manual backup procedures ..................................................................................104
Requirements for a manual backup ............................................................. 104
Manual backup through the RMAN command line interface ................. 105
Manual backup through Oracle Enterprise Manager Backup Management
Tools .................................................................................................................. 105
NetWorker server bootstrap backup ............................................................ 106
Cancel a manual backup ................................................................................ 107
Monitor a manual backup.............................................................................. 108
Scheduled backup procedures .............................................................................110
Test a scheduled backup ................................................................................ 110
Cancel a scheduled backup ........................................................................... 111
Monitor a scheduled backup......................................................................... 112
Automation of NSR_SERVER and NSR_GROUP parameter settings .... 113
Regular backup information in NetWorker indexes .........................................114
Deduplication backup information in NetWorker indexes..............................115

4 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Contents

Chapter 4 Data Restore and Recovery


About restore and recovery.................................................................................. 118
NetWorker indexes and policies used for restores .................................... 118
Recovery configuration wizard .................................................................... 118
RMAN scripts for restore and recovery ............................................................. 121
Performing an Oracle data restore and recovery .............................................. 123
Requirements for an Oracle data restore..................................................... 123
Restore through the RMAN command line interface ............................... 125
Restore with Oracle Enterprise Manager Backup Management Tools ... 126
Perform an Oracle data recovery ................................................................. 127

Chapter 5 Disaster Recovery


About disaster recovery........................................................................................ 130
Preparing an Oracle database for disaster recovery ......................................... 130
Create the DBID text file................................................................................ 131
Set up a postcommand script for backup of Oracle-related files............. 131
Set up RMAN backups of the database and related files ......................... 132
Set up RMAN backups of Recovery Catalog.............................................. 132
Performing a disaster recovery............................................................................ 133
NetWorker server recovery ........................................................................... 133
Oracle disaster recovery to a new host........................................................ 133
Sample postcommand script................................................................................ 134

Chapter 6 Cluster and RAC Systems


Considerations for cluster and RAC systems .................................................... 138
Parameters in the RMAN script or session................................................. 138
NetWorker software configuration .............................................................. 139
Backup failover ............................................................................................... 139
Cluster systems ...................................................................................................... 139
Roadmap for backup/restore configuration in a cluster system............. 140
Backup failover ............................................................................................... 140
RAC systems........................................................................................................... 140
RAC terminology............................................................................................ 140
RAC backups and restores ............................................................................ 141
Roadmap for backup/restore configuration in a RAC system................ 141
Setting up RAC nodes to back up to a local storage node........................ 142
Connect-time failover..................................................................................... 144
Creating RMAN backup scripts ................................................................... 146
Creating RMAN restore scripts .................................................................... 146
Archived redo logs ......................................................................................... 147

Chapter 7 Proxy Backups and Restores


Overview of proxy backups and restores........................................................... 150
Types of proxy backups ................................................................................. 150
Types of proxy restores .................................................................................. 151
Proxy backup and restore system........................................................................ 151
Proxy backup processes ........................................................................................ 153
Scheduled proxy backup ............................................................................... 154
Proxy restore processes ......................................................................................... 159
Instant restore or rollback.............................................................................. 160
Restore from secondary storage ................................................................... 161
Software configuration roadmap........................................................................ 163

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 5


Contents

Basic configurations........................................................................................ 163


Additional configurations for proxy operations ........................................ 163
Proxy backup requirements..................................................................................171
Checking configuration consistency ............................................................ 171
Performing proxy backups ............................................................................ 172
Proxy backup information in the NetWorker indexes ......................................174
Entries in the client file index........................................................................ 175
Entries in the media database ....................................................................... 175
NWORA resource file backup in the NetWorker indexes......................... 176
Proxy restore requirements.................................................................................. 177
Creating RMAN scripts for proxy restores.................................................. 177
Performing proxy restores ............................................................................. 178
Relocating files during proxy restores ......................................................... 179
Restoring to a different host .......................................................................... 180
Point-in-time recoveries without a Recovery Catalog............................... 181
Catalog synchronization for proxy backups ......................................................181
Extra entries in the catalogs........................................................................... 181
Control file versus recovery catalog............................................................. 183
The NWORA resource file ............................................................................. 183
Automatic catalog synchronization with the nsroraclecat program ....... 191
Proxy backups and restores on cluster systems.................................................194
Proxy backup failover .................................................................................... 194
Proxy backups from a virtual cluster client ................................................ 195
Proxy backups from a physical cluster client.............................................. 196
Restores from proxy backups on a cluster system ..................................... 197

Appendix A Parameters in an RMAN Session


Setting the NSR* parameters ............................................................................... 200
NSR* parameter definitions................................................................................. 201

Appendix B RMAN Commands


The delete expired backup command ................................................................ 210
The change...crosscheck and crosscheck commands ....................................... 210
The pool option of the backup command.......................................................... 210
The send command............................................................................................... 211
Syntax rules...................................................................................................... 211
Two ways to run the send command........................................................... 213
Precedence rules.............................................................................................. 215
The set duplex command..................................................................................... 215
The trace option of the backup command ......................................................... 217

6 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Contents

Appendix C Troubleshooting and Error Messages


Troubleshooting tips............................................................................................. 220
RMAN error messages ......................................................................................... 221
NMO error messages............................................................................................ 222
Error messages from the libnwora library .................................................. 223
Error messages from the nsrnmoadmin program ..................................... 231
Error messages from the nsrnmoinfo program.......................................... 233
Error messages from the nsrnmoprobe program ...................................... 235
Error messages from the nsrnmostart program......................................... 236

Glossary

Index

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 7


Contents

8 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Figures

Title Page

1 Regular Oracle backup and recovery system.................................................................... 50


2 Regular scheduled NMO backup ....................................................................................... 54
3 Regular manual NMO backup ............................................................................................ 55
4 Deduplication NMO backup ............................................................................................... 57
5 Regular NMO restore ........................................................................................................... 58
6 UNIX Client resource for scheduled backups................................................................... 90
7 Windows Client resource for scheduled backups............................................................ 91
8 Oracle backup messages in Sessions tab of Monitoring window ................................ 108
9 Oracle backup messages in Devices tab of Monitoring window ................................. 109
10 Oracle backup messages in Log tab of Monitoring window ........................................ 109
11 Group details for regular scheduled backups................................................................. 113
12 Proxy backup and restore system..................................................................................... 153
13 Scheduled instant backup .................................................................................................. 156
14 Scheduled immediate live backup.................................................................................... 158
15 Deferred live backup .......................................................................................................... 159
16 Instant restore or rollback .................................................................................................. 161
17 Restore from secondary storage........................................................................................ 162

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 9


Figures

10 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Tables

Title Page

1 NMO components................................................................................................................. 51
2 NetWorker Server resource attributes ............................................................................... 63
3 User group privileges required for NMO operations...................................................... 65
4 Options of the nsrnmoadmin command for migration................................................... 72
5 NetWorker Client resource attributes ................................................................................ 88
6 NetWorker Probe resource attributes ................................................................................ 98
7 PowerSnap parameters ...................................................................................................... 167
8 NWORA parameter resources .......................................................................................... 184
9 NWORA SID resource components ................................................................................. 186
10 Parameters in the RMAN command or script ................................................................ 201
11 Option values in the send command ............................................................................... 213
12 Set duplex command values.............................................................................................. 216
13 Trace option values and conditions traced...................................................................... 217
14 Error messages from the libnwora library....................................................................... 223
15 Error messages from the nsrnmoadmin program.......................................................... 231
16 Error messages from the nsrnmoinfo program .............................................................. 234
17 Error messages from the nsrnmoprobe program ........................................................... 235
18 Error messages from the nsrnmostart program ............................................................. 236

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 11


Tables

12 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Examples

Title Page

1 Using the configure channel command with parms option for automatic channels .. 26
2 Specifying parameter values per automatic channel ....................................................... 26
3 Using the set backup copies command in the RMAN script .......................................... 28
4 Using automatic channels for backup copies .................................................................... 28
5 Expiring a backup ................................................................................................................. 38
6 Save set bundling for a one-week scheduled backup cycle of a tablespace ................. 42
7 Save set bundle join .............................................................................................................. 43
8 Splitting a save set bundle across volumes ....................................................................... 43
9 Using save set consolidation to re-unite a save set bundle ............................................. 44
10 RMAN script for a manual backup .................................................................................... 75
11 RMAN script for AES encryption during an Oracle backup .......................................... 76
12 RMAN script for a scheduled backup ................................................................................ 77
13 RMAN script for a manual deduplication backup ........................................................... 96
14 Possible Command Options settings for the nsrnmoprobe program ........................... 99
15 Multiple probes for a probe-based backup ..................................................................... 100
16 RMAN script to restore a tablespace ................................................................................ 121
17 RMAN script to restore an AES encrypted backup ....................................................... 122
18 RMAN script to restore from a specified pool ................................................................ 122
19 Sample nsrnmoinfo commands for Oracle restores ....................................................... 125
20 Volume information displayed by the nsrnmoinfo command ..................................... 125
21 Sample postcommand script on UNIX ............................................................................ 134
22 Sample postcommand script on Windows ...................................................................... 135
23 Setting up RAC nodes as storage nodes .......................................................................... 142
24 RMAN script for a manual Oracle backup on a RAC system ...................................... 146
25 RMAN script for an Oracle restore on a RAC system ................................................... 146
26 RMAN scripts with multiple channels ............................................................................ 165
27 PowerSnap parameter settings .......................................................................................... 168
28 PowerSnap parameter settings for a Celerra NAS device ............................................ 168
29 Proxy backup failure ........................................................................................................... 172
30 Proxy backup entries in the client file index ................................................................... 175
31 Proxy backup entries in the media database ................................................................... 176
32 Resource file backup entry in the client file index ......................................................... 176
33 Resource file backup entry in the media database ......................................................... 176
34 RESTORE_TYPE_ORDER parameter settings ................................................................ 177
35 Symbolic link specified in the set newname command ................................................. 180
36 Relocation of a raw volume ............................................................................................... 180
37 Default NWORA parameter resources ............................................................................ 186
38 NWORA SID resource ........................................................................................................ 187
39 Connection file contents ..................................................................................................... 188
40 Proxy backup entries in the index of a physical cluster client ...................................... 196

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 13


Examples

Title Page

41 Proxy backup entries in the index of a virtual cluster client ......................................... 197
42 A send command sets the parameters for a specified channel ..................................... 212
43 An rman send command sets a parameter for all channels .......................................... 214
44 Order of parameters set according to the precedence rules .......................................... 215

14 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Preface

As part of an effort to improve and enhance the performance and capabilities of its product
lines, EMC periodically releases revisions of its hardware and software. Therefore, some
functions described in this document may not be supported by all versions of the software or
hardware currently in use. For the most up-to-date information on product features, refer to
your product release notes.
If a product does not function properly or does not function as described in this document,
please contact your EMC representative.

Audience This document is part of the EMC NetWorker Module for Oracle (NMO)
documentation set, and 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.

Related Documentation related to the use of this product can be found at the EMC website,
documentation http://Powerlink.EMC.com, including:
◆ The NetWorker Module for Oracle release 5.0 documentation set:
• Administration guide
• Installation guide
• Release notes
• Command reference guide
◆ The NetWorker documentation set:
• Administration guide
• Installation guide
• Release notes
• Command reference guide
• Disaster recovery guide
◆ Other EMC documentation:
• NetWorker PowerSnap Module documentation
• Software compatibility guide
• UNIX man pages

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 15


Preface

The following additional documentation may be useful:


◆ Oracle Server documentation
◆ Oracle database backup and recovery documentation

Conventions used in EMC uses the following conventions for special notices.
this document
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 conventions
EMC uses the following type style conventions 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

16 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Preface

| 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

Where to get help EMC support, product, and licensing information can be obtained as follows.
Product information — For documentation, release notes, software updates, or for
information about EMC products, licensing, and service, go to the EMC Powerlink
website (registration required) at:
http://Powerlink.EMC.com
Technical support — For technical support, go to EMC Customer Service on
Powerlink. To open a service request through Powerlink, you must have a valid
support agreement. Please contact your EMC sales representative for details about
obtaining a valid support agreement or to answer any questions about your account.

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:
SSGdocumentation@EMC.com
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.

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 17


Preface

18 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


1

Introduction

This chapter includes the following major sections:


◆ Naming conventions used in this guide .................................................................... 20
◆ Importance of backups ................................................................................................. 20
◆ Product features............................................................................................................. 20
◆ Software used in the NMO environment .................................................................. 47
◆ NMO components......................................................................................................... 51
◆ NMO backup and restore processes........................................................................... 52

Introduction 19
Introduction

Naming conventions used in this guide


Consider the naming conventions used throughout this guide:
◆ “Administrator Microsoft user” and “a member of the Microsoft Windows
Administrators group” are used interchangeably.
◆ “Oracle Server host” refers to the host where both the Oracle Server and EMC®
NetWorker® Module for Oracle (NMO) 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 NMO 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.
◆ "Windows" refers to all the supported Microsoft Windows operating systems,
unless specified otherwise.

Importance of backups
The 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 features
The following sections describe major features of the NMO, 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 NMO backup, as described in “The send command” on
page 211.
The Oracle backup and recovery documentation provides more details on RMAN
backups and restores and Oracle product features.

20 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

Scheduled compared to manual backups


An NMO backup can be either a scheduled or manual (unscheduled) backup:
◆ A scheduled NMO 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 23.
A proxy backup is only supported by using a scheduled backup, as described in
“Proxy backups and restores” on page 24.
◆ A manual NMO backup includes the following features:
• The backup is initiated by a user on the NMO 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 NMO 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.
Chapter 3, “Backup Procedures,” provides details on scheduled and manual backup
procedures.

Deduplication backups and restores


NMO 5.0 software provides new support for deduplication backups and restores.
The NetWorker Module for Oracle release notes 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 NMO
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 NMO 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 NMO 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 NMO
backups.

Product features 21
Introduction

The Avamar client-side binary, nsravtar, is installed on the NMO client with the
NetWorker client software. During a deduplication backup, the nsravtar process
sends the deduplicated client data to the Avamar server. A deduplication NMO
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 55
◆ “Deduplication restore processes” on page 58
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 NMO 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 administration guide provides more
information.
You must configure a scheduled or manual deduplication backup according to
“Configuring a deduplication backup” on page 93.
“Deduplication backup information in NetWorker indexes” on page 115 describes
the backup information stored in the NetWorker indexes, and how to delete the
backups.

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, NMO 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 EMC
Professional Services. Avamar documentation and the NetWorker administration
guide provide details on replication of deduplication backups.

22 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

Probe-based backups
NMO 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.
The NetWorker Module for Oracle release notes provides details on the NetWorker
requirements for the support of probe-based backups.
Probe-based backups are not supported for proxy backups or cluster/RAC
environments.

Workflow of probe-based backups


A probe-based NMO 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.
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.

Product features 23
Introduction

Types of probes
There are two different types of probes:
◆ An NMO probe.
The NMO probe is implemented through the NMO 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 97.
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 97 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 NMO software and a specific PowerSnap Module. The EMC
Information Protection Software Compatibility Guide on the EMC Powerlink® website
provides a list of supported PowerSnap Modules.
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.
Chapter 7, “Proxy Backups and Restores,” provides more information on proxy
backup and restore operations.

24 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

VMware support
NMO 4.5 provided support for regular backups and restores of an Oracle database
installed on a VMware Virtual Machine (VM) on an ESX server.
NMO 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 a user does not experience any
disconnection. If a migration occurs during an NMO 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 NMO backup or restore,
the backup or restore is not interrupted.
The NetWorker Module for Oracle release notes 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 NMO backups of archived redo logs are enabled.
“Backing up all archived logs from each node” on page 147 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:
configure controlfile autobackup on
configure controlfile autobackup format for device type ’sbt_tape’ to
’/NMO_%f/’

Product features 25
Introduction

If the control file autobackup is set to on and the RMAN backup is performed with
NMO, the control file autobackup will also be performed with NMO. 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
NMO 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
NMO. 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 NMO 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 201 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 NMO backups with the NetWorker server
server1.emc.com by typing the following configure channel...parms... command:
configure channel device type ’sbt_tape’ parms
’ENV=(NSR_SERVER=server1.emc.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 NMO 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)’

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.

26 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

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 NMO 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 NMO 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, NMO does not
support the use of RMAN backup copies commands during scheduled backups.
NMO 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 NMO 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 215 provides more information on the set duplex command and
setting up NetWorker pools for each copy.

Product features 27
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 NMO.
For example, specify persistent settings for duplex copies of datafiles and
archived redo logs (respectively) in NMO backups with the following types of
configure commands:
configure datafile backup copies for device type ’sbt_tape’ to 2
configure 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)’;

28 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


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, NMO 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 NMO 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, NMO 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, NMO 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.
NMO 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 NMO software.

Product features 29
Introduction

Configuration wizards
NMO 5.0 software supports new backup and recovery configuration wizards that are
integrated with the NetWorker Management Console (NMC). The configuration
wizards used in NMO release 4.5 or earlier are replaced by the NMC-based wizards
in NMO release 5.0.
Configuration wizard plug-in components are installed with the NMO software on
the NMO client host. NMC loads the NMO wizard plug-in at run time. You can run
the NMO 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.
The NetWorker Module for Oracle release notes provides details on the NetWorker
requirements for the support of the NMC-based configuration wizards.

Main features of the wizards


The configuration wizards can be used to configure the following for an NMO client:
◆ Scheduled Oracle backups (either typical or customized).
“Configuring a backup with the NMC wizard” on page 72 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 118 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 NMO release 4.5 or
earlier.

Features of the backup configuration wizard


The backup configuration wizard can perform the following:
◆ Configure a new NetWorker Client resource for an NMO 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 NMO
configuration wizard.
◆ Modify a backup configuration that was created either with the wizard from
NMO 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 32.

30 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

Note: NMO 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 75 describes the
legacy configuration method.

“Configuring a backup with the NMC wizard” on page 72 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 119 provides information on
the features of the recovery configuration wizard.
Backup configuration storage with the wizard
The 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 NMO 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 be
backed 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.

Product features 31
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
NMO 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 32.

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 NMO 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 NMO
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 69.

Note: Migration of a proxy backup configuration is not supported. “Requirements for using the
nsrnmoadmin command for migration” on page 70 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.

32 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


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) and localization (L10N)


The following sections describe internationalization (I18N) and localization (L10N)
features that NMO release 5.0 supports in a non-English locale.

I18N features
NMO I18N is the capability of the NMO software to operate in a non-English
environment or locale without itself generating non-ASCII data. After you set up
NMO I18N as described in “Configuring I18N support” on page 67, NMO 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 EMC Information Protection Software Compatibility Guide on the Powerlink website
provides details on the different languages supported and the operating system,
Oracle Server, and NetWorker software requirements for NMO I18N.
NMO 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 NetWorker Module for Oracle release notes
As a separate NMO feature, NMO localization (L10N) is the capability of NMO to
itself generate non-ASCII data and print the data to the user interface. “L10N
features” on page 35 provides more information on NMO L10N support.
The term internalization is used differently in the NetWorker documentation (as
opposed to this NMO 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.
The extent of the NMO I18N support is dependent on the following:
◆ I18N support that is provided by the operating system on the NMO client host.
◆ I18N support that is provided by the NetWorker client and server software.
◆ National Language Support (NLS) or globalization support that is provided by
the Oracle software.
For example, if NetWorker software does not support non-ASCII data in a specific
NetWorker resource attribute (such as the group name in the Group resource), NMO
cannot support non-ASCII data in that resource attribute. The EMC Information
Protection Software Compatibility Guide and the NetWorker documentation include
more information on the I18N support provided by NetWorker.

Product features 33
Introduction

When NMO I18N support is set up as described in “Configuring I18N support” on


page 67, NMO 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 nsrnmo(.bat),
nsrnmoadmin(.exe), nsrnmoinfo(.exe), and nsroraclecat(.exe) commands
◆ Tablespace names and datafile paths
◆ The format string of the RMAN backup command (unless the nsrnmoadmin -M
or nsrnmoadmin -P command is used for migration of a legacy backup
configuration on Windows)
◆ 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

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 NMO 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 NMO wizard:
• Review and Edit the RMAN Script screen
• Review and Accept the Client Configuration screen
◆ Content of output files created by the NMO wizard, such as nsrnmo scripts (for
example, nsrnmo.bat or nsrnmo.sh file) 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 NMO client generates messages
in English only.

34 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

When NMO I18N support is set up, NMO generates debug messages in English only.
NMO 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 NMO
releases. (The log file does not contain Oracle Server or RMAN errors.)
The NetWorker administration 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 67 describes how to set up NMO I18N support.

L10N features
NMO L10N is the capability of the NMO software to generate non-ASCII character
and numeric data based on the locale of the user, and output the data to the user
interface. This capability requires the installation of the appropriate NMO language
packs in non-English locales. The non-ASCII data generated by NMO is printed to the
command line or to specific log files or wizard screens.
◆ When NMO I18N is set up, NMO can process and display non-ASCII data that is
passed to it by the operating system, NetWorker software, and Oracle software.
◆ When NMO L10N is set up by also installing the required NMO language packs,
NMO can also generate non-ASCII data, localized according to the user locale. The
non-ASCII data can include text messages, dates, times, and numeric values
displayed in the locale-dependent format.
The EMC Information Protection Software Compatibility Guide on the Powerlink website
provides details on the different languages supported and the operating system,
Oracle Server, and NetWorker software requirements for NMO L10N.
NMO L10N is supported for the following:
◆ Regular backups and restores
◆ Proxy backups and restores with a supported PowerSnap Module release, as
described in the NetWorker Module for Oracle release notes
The extent of the NMO L10N support is limited by the following:
◆ L10N support that is provided by the operating system on the NMO client host.
◆ L10N support that is provided by the NetWorker client and server software.
◆ National Language Support (NLS) or globalization support that is provided by
the Oracle software.
When NMO L10N support is set up as described in “Configure L10N support” on
page 69, NMO generates messages in localized form and displays localized text in the
wizard, except for the following:
◆ Comments (lines starting with the symbol #) in the nwora.res file and RMAN
scripts
◆ All debug messages, including the messages in the configuration wizard debug
log files and in the log files specified by NSR_DEBUG_FILE and
NSR_SB_DEBUG_FILE
◆ Output to the Oracle trace file
“Configure L10N support” on page 69 describes how to set up NMO L10N support.

Product features 35
Introduction

Oracle Data Guard support


NMO 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.
To configure NMO 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 NMO 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 NMO 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
NMO 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 NMO backup.

36 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

◆ 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 NMO software on multiple nodes of a cluster. Chapter 6, “Cluster and
RAC Systems,” provides more details.

Policy uniformity
If policy uniformity is enabled, NMO 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.
After NMO 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 NMO program nsrnmostart
changes the policies of all save sets in the cycle to match the longest policy of the new
incremental save sets. NMO 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 NMO 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 92 provides information on how to configure
policy uniformity for NMO 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 NMO backups:
◆ “Cancel a manual backup” on page 107
◆ “Cancel a scheduled backup” on page 111

Product features 37
Introduction

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.
NMO 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.

Example 5 Expiring a backup

A NetWorker server might expire an NMO backup that is not yet obsolete in RMAN,
depending on the NetWorker browse and retention policies. Use of NMO 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 87 provides more information on how to set NetWorker policies
for NMO backups.

! IMPORTANT
Run the crosscheck command on the NMO 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;

38 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

Save set bundling


If NMO save set bundling is configured, NMO 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: NMO does not support save set bundling for proxy backups. NMO performs save set
bundling for regular Oracle backups only.

During staging operations with NetWorker server release 7.4 and later, if the staging
criteria determine that a particular NMO 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 administration guide provides details on how to work with staging
policies and perform automatic and manual staging operations through the
NetWorker server.
“NMO scheduled backups and save set bundling” on page 39 describes NMO save
set bundling during regular scheduled backups, and how to configure save set
bundling.
If policy uniformity is configured, NMO 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 44 provides more information on how policy
uniformity relates to save set bundling. Save set bundling and policy uniformity can
be enabled and disabled independently.

NMO 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 NMO scheduled backups. Both are performed to
prepare for staging operations with NetWorker server 7.4 and later.
“Configure save set bundling” on page 91 provides information on how to configure
save set bundling for NMO scheduled backups.

RMAN backup levels and dependencies


An 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.)

Product features 39
Introduction

◆ 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.
– 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 NMO save sets


NMO save set bundling is performed at the backup set level.
An NMO 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 NMO 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 NMO save set bundles


If 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 NMO save sets from a
backup set are placed into the same save set bundle.)

40 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

At the end of a full or level 0 scheduled backup, the NMO 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, NMO 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.
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 NMO cannot find a preceding dependent
backup in any existing bundles, NMO 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 NMO 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 NMO 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 91 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 91 provides
more details on the requirements of save set bundling.

Product features 41
Introduction

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.
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. NMO 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 43.

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 restrictions


When planning the strategy for NMO 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 43 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 NMO backup cycle
to be staged before the cycle is complete. For example, if a one-week NMO 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 NMO 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 administration guide provides details on how to work with staging
policies and perform automatic and manual staging operations through the
NetWorker server.

42 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

Save set bundling information in the media database


The NMO software stores information about each save set bundle in the NetWorker
media database. “NetWorker software” on page 47 provides more information about
the media database.
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 staging


The 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, NMO combines all
three backups into the same save set bundle.
If the original file A backup has the oldest backup time, NMO places the new backup
set (from the level 1 backup) into the save set bundle of the (level 0) file A backup.
NMO 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.

Product features 43
Introduction

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.

Other Oracle features


This section describes supported features of the Oracle Server software. The NMO
software supports the Oracle releases on specific platforms, as outlined in the EMC
Information Protection Software Compatibility Guide on the Powerlink website.
The NetWorker Module for Oracle release notes describes known NMO limitations
related to specific Oracle releases.
Examples of the Oracle RMAN features that NMO 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 NMO, 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 area
backup recovery files

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

44 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

For example, the following sequence of RMAN commands can be used to


configure an automatic channel for NMO 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:
• 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 NMO 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


NMO 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

Product features 45
Introduction

◆ Backup of read-only transportable tablespaces


◆ Oracle Enterprise Manager enhancements, with new interfaces for the Data
Recovery Advisor
◆ Oracle Globalization Support enhancements
To enable NMO 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 46 and “Archival backup feature” on
page 46.
The appropriate Oracle documentation provides more information on the Oracle11g
features.

Data Recovery Advisor


The 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.
Before you can use the Data Recovery Advisor to invoke an RMAN restore script that
involves NMO 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 NMO parameters may also be specified for the automatic channel
configuration.

To enable the use of Data Recovery Advisor with Oracle11g and NMO:
◆ If automatic channels have not been configured for NMO 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=NMO_client_name)’;
configure channel device type 'sbt_tape' parallelism
number_of_restore_channels;
◆ If automatic channels are already configured for NMO backups, no additional
configuration steps are required.

Archival backup feature


With 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 NMO:
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.

46 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

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 NMO.

The NetWorker administration guide provides more information on how to


configure resources and specify pool selection criteria through the NetWorker server.

Software used in the NMO environment


The following sections describe the different software components involved in the
NMO 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.

Software used in the NMO environment 47


Introduction

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 installation guide provides information on how to install
NetWorker software.
◆ The NetWorker administration guide provides information on how to configure
and use NetWorker software.

NMO software
NetWorker Module for Oracle (NMO) 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
NMO 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, NMO 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
NMO, with the NetWorker server and the appropriate NetWorker PowerSnap
Module software.
The NetWorker Module for Oracle installation guide provides information on how to
install the NMO software.

NetWorker PowerSnap Module software


The NetWorker PowerSnap Module software works with the NetWorker server and
NMO 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.

48 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

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 NMO and PowerSnap Module software:
◆ Chapter 7, “Proxy Backups and Restores”
◆ NetWorker PowerSnap Module documentation
The EMC Information Protection Software Compatibility Guide on the Powerlink website
provides a complete list of supported PowerSnap Modules.

Oracle backup and recovery system


This section describes the system used for regular Oracle backup and recovery.
The NMO 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
◆ NetWorker Module for Oracle
Figure 1 on page 50 illustrates the architecture of the regular Oracle backup and
recovery system, and shows the functional relationship between the NetWorker and
Oracle software components. The NetWorker Module for Oracle media management
library (MML) is integrated with the Oracle Server during the NMO 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 NMO environment 49


Introduction

GUI console Oracle Server system NetWorker


server
Oracle
database files
OEM
Backup RMAN
Manager

Oracle
Recovery Server
Catalog
Storage
NetWorker medium
Module
for Oracle
MML

NetWorker
client

NetWorker
NetWorker server
Management Console
GEN-000170

Figure 1 Regular Oracle backup and recovery system

Oracle Recovery Manager


During the Oracle backup and restore operations, NMO 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.

50 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

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
◆ 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.

NMO components
Table 1 on page 51 lists the components that are installed on the NMO client host
during the NMO software installation. Unless specified otherwise, the files are
located in the same directory as the NetWorker client software.

Table 1 NMO components (page 1 of 2)

Name on UNIX Name on Windows Description

fdiff.jar fdiff.jar Plug-in jar files for the NMO backup and recovery configuration wizards.
nwora.jar nwora.jar
nwora_res.jar nwora_res.jar

libcommonssl.7.5.build#.so In a subdirectory under /usr/lib/nsr/nmo or /opt/networker/lib/nmo on UNIX only. An NMO


library that is required for communication with NetWorker.

libnwora.xx In /usr/lib on UNIX only. The main NMO library (known as Media Management Library in
Oracle documentation) that is loaded by the Oracle backup or restore process.

nsrnmo nsrnmo.bat 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 /etc on UNIX only. A sample template for the nsrnmo script. You should copy and customize
this template, and save the original file in /etc.

nsrnmoadmin nsrnmoadmin.exe 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 nsrnmodrpostcmd.exe 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 nsrnmoinfo.exe The program that determines the NetWorker volumes required to restore specified Oracle
backup pieces from NMO backups.

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

NMO components 51
Introduction

Table 1 NMO components (page 2 of 2)

Name on UNIX Name on Windows Description

nsrnmostart nsrnmostart.exe 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 nsroraclecat.exe Not available on Linux Itanium, Solaris AMD64/EM64T, or Windows Itanium (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 nsrorara.exe The program that performs operations on behalf of the NMO configuration wizard on the local
or remote NMO host.

nsrsbtcn.exe On Windows only. The orasbt.dll file is the main NMO library (known as Media Management
orasbt.dll Library in Oracle documentation) that is loaded by the Oracle backup or restore thread, and it
uses nsrsbtcn.exe to perform any corresponding NetWorker operations.

NMO backup and restore processes


This section describes the processes involved in regular backups and restores.
To provide storage management services for Oracle Server data, NMO 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 NMO installation on the Oracle Server host.
Oracle loads the NMO MML at run time, and the Oracle Server backup sessions call
the NMO 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.

52 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

Regular scheduled backup processes

Note: “Deduplication backup processes” on page 55 provides information specifically for a


deduplication backup.

A scheduled NMO 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:
– 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 NMO 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 54.

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 NMO backup.

Figure 2 on page 54 shows how the Oracle Server, NetWorker server, and NMO
processes interact during a regular scheduled NMO backup that was configured
through the legacy method (without the wizard).

NMO backup and restore processes 53


Introduction

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.

Oracle Server NetWorker server Storage


(NetWorker client) medium

Backup pieces Client file index Media database

tracking
information
Oracle
process nsrindexd nsrmmdbd
(channel)

NetWorker NetWorker Module


Module elements
for Oracle
MML data
nsrmmd
rman
interprocess
communication
savefs nsrnmostart

nsrnmo savegrp nsrd

nsrexecd
GEN-001110

Figure 2 Regular scheduled NMO backup

Regular manual backup processes

Note: “Deduplication backup processes” on page 55 provides information specifically for a


deduplication backup.

A manual NMO backup is a user-initiated backup through an Oracle backup utility,


RMAN or Oracle Enterprise Manager. NMO 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 NMO media management library (MML) software routines to initiate the backup.
A manual NMO backup includes the following interactions:
1. The Oracle Server backup session that loads the NMO 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.

54 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

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 55 shows how the Oracle Server, NetWorker server, and NMO
processes interact during a regular manual NMO backup.

Oracle Server NetWorker server Storage medium


(NetWorker client)

Backup pieces Client file index Media database

RMAN
tracking
information
Oracle nsrindexd nsrmmdbd
process
(channel) data
nsrmmd
NetWorker
Module NetWorker
for Oracle Module
MML element

interprocess
nsrexecd communication
nsrd
GEN-000171

Figure 3 Regular manual NMO backup

Deduplication backup processes


“Deduplication backups and restores” on page 21 describes the features of a
deduplication NMO backup.
For scheduled deduplication backups, the steps from “Regular scheduled backup
processes” on page 53 still apply.
A deduplication NMO backup includes the following interactions:
1. The Oracle Server backup session that loads the NMO 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.

NMO backup and restore processes 55


Introduction

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.
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 115.
• The Oracle backup session sends tracking information to the NetWorker client
file index by using the nsrindexd service.
Figure 4 on page 57 shows how the Avamar server, NetWorker, and NMO processes
interact during a deduplication NMO 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 NMO save set is passed between the Oracle Server
backup session and NetWorker storage node.

56 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

NetWorker NetWorker
server storage node

Hash ID

NMO client NetWorker


deduplication node
(Avamar server)
Oracle Server
backup session
NetWorker client /
with NMO MML
utility node

acle
Oracle
data /
hash ID
GSAN

Node 1
nsravtar
Node 2
Deduplicated
data / hash ID
GEN-001111

Figure 4 Deduplication NMO backup

Regular restore processes


An NMO restore can be performed in either of two ways:
◆ By issuing the appropriate commands through the RMAN command line
interface.
Or
◆ By using the Oracle Enterprise Manager Backup Management Tools, which
include an optional graphical user interface to the RMAN utility.
The RMAN utility starts Oracle Server sessions on the target database. These Oracle
Server sessions initiate the restore by calling NMO media management library
(MML) software routines.

Note: “Deduplication restore processes” on page 58 provides information specifically for a


deduplication restore.

A regular NMO restore includes the following interactions:


1. The NMO 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.

NMO backup and restore processes 57


Introduction

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 NMO MML.
6. The Oracle Server sessions write the data to the disk.
Figure 5 on page 58 shows how the Oracle Server, NetWorker server, and NMO
processes interact during a regular NMO restore.

Oracle Server NetWorker server Storage


(NetWorker client) medium

Backup pieces Client file index Media database

RMAN
tracking
information
Oracle
process nsrindexd nsrmmdbd
(channel)

NetWorker
Module NetWorker Module
element
for Oracle
MML
data
nsrmmd
interprocess
nsrexecd communication
nsrd
GEN-000173

Figure 5 Regular NMO 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 NMO restore includes the following interactions:
1. The NMO 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 NMO client.
4. When the NetWorker server receives a restore request from the client, the main
NetWorker service, nsrd, contacts the media service, nsrmmd.

58 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Introduction

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.
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.

NMO backup and restore processes 59


Introduction

60 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


2
Software
Configuration

This chapter includes the following major sections:


◆ Configuration roadmap................................................................................................ 62
◆ Verify the Oracle Server configuration....................................................................... 63
◆ Verify the NetWorker resources .................................................................................. 63
◆ Configuring I18N support ........................................................................................... 67
◆ Configure L10N support .............................................................................................. 69
◆ Migrating a legacy configuration with the nsrnmoadmin command ................... 69
◆ Configuring a backup with the NMC wizard........................................................... 72
◆ Configuring a backup with the NMC legacy method ............................................. 75
◆ Configuring a deduplication backup ......................................................................... 93
◆ Configuring a probe-based backup ............................................................................ 97

Software Configuration 61
Software Configuration

Configuration roadmap
The Oracle Server and NetWorker server must be properly configured before the
NetWorker Module for Oracle (NMO) software can be used for backup and restore
operations.
Before configuring the Oracle Server and NetWorker server, ensure that the NMO
software is installed on the Oracle Server host according to the instructions in the
NetWorker Module for Oracle installation guide.
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 63.
2. Verify the NetWorker resources according to “Verify the NetWorker resources”
on page 63.
3. If required, configure internationalization (I18N) support according to
“Configuring I18N support” on page 67.
4. If required, configure localization (L10N) support according to “Configure L10N
support” on page 69.
5. 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 69.
6. Complete the backup configuration by using either the wizard or NMC:
• “Configuring a backup with the NMC wizard” on page 72
• “Configuring a backup with the NMC legacy method” on page 75 (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 93
◆ “Configuring a probe-based backup” on page 97
Chapter 6, “Cluster and RAC Systems,” provides information on configuring any
required cluster or Real Application Cluster (RAC) systems.

62 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

Verify the Oracle Server configuration


The Oracle Server system must be properly installed and configured before the
NetWorker server and NMO 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 resources


Verify 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 63 describes the main NetWorker Server resource attributes. Verify
that the attribute settings in the Server resource are valid for the NMO backup
environment. Modify the settings as required.

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 on UNIX and Administrators@hostname on Windows, where
hostname is the NetWorker server hostname. “Specify the required NetWorker privileges” on
page 64 provides more information.

Datazone pass Specifies the key or pass phrase to use for AES encryption of Oracle data during an NMO
phrase backup. The pass phrase is required to restore the Oracle data from the backup.
“NSR_AES_ENCRYPTION” on page 201 provides more information.

Verify the Oracle Server configuration 63


Software Configuration

The NetWorker server online help and the NetWorker administration guide provide
more information on how to configure a NetWorker Server resource and its
attributes.

Specify the required NetWorker privileges


Certain NMO operations require specific NetWorker privileges. The NetWorker
privileges are specified through the User Group resource. “NetWorker user group
privileges” on page 64 provides more information.
One NMO operation that requires special privileges is deletion. NMO 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 107
• “Cancel a scheduled backup” on page 111

Note: If the Oracle user is not granted the required NetWorker privileges in these cases, NMO
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 NMO 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 64.

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 UNIX, and members of the Microsoft
Windows Administrators group, are always members 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

64 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

The default user group configurations are sufficient for RMAN backup, restore, and
crosscheck operations that use NMO. If the default user group configurations are
changed, ensure that the required privileges are assigned for the operations.
Verify that the required user group privileges exist for the NMO operations, as
described in Table 3 on page 65. The NetWorker administration 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.

Table 3 User group privileges required for NMO operations

Operation Operating system user that performs operation Required user group privileges

Migration of a legacy backup configuration with Root user, or a member of the Microsoft Windows Configure NetWorker
the nsrnmoadmin command Administrators group, on the Oracle Server

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, or a member of the Microsoft Windows Monitor NetWorker, Backup Local Data
Administrators group, on the Oracle Server (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 Root user, or a member of the Microsoft Windows Recover Local Data
Oracle Server Administrators group, on the Oracle Server (This privilege is set by default)

Save set bundling Root user, or a member of the Microsoft Windows Operate NetWorker, and all its prerequisite
Administrators group, on the Oracle Server privileges

The Oracle user is defined as the following:


◆ On UNIX:
• 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
and system on Microsoft Windows.
◆ On Windows, the operating system user that runs the Oracle service
(OracleServiceoracle_sid).

! 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 administration guide provides more information on NetWorker user
groups and setting user group privileges.

Verify the NetWorker resources 65


Software Configuration

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 NMO client by using either
backup configuration wizard or the legacy configuration method.
A NetWorker Schedule resource specifies the attributes of the backup schedule. The
NetWorker administration guide provides more information on the features of a
Schedule resource.
The NMO software does not honor the NetWorker backup levels (except skip) that are
specified in the NetWorker Schedule resource.

! IMPORTANT
For scheduled NMO backups, the NetWorker backup levels that appear in the
Schedule resource do not determine the NMO backup levels. The level of the
NMO 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 87.

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 administration guide provides more information on storage devices, the
NMC program, and how to configure Device resources.
The EMC NetWorker Hardware Compatibility Guide on the Powerlink website provides
a complete list of the storage devices that the NetWorker server supports.
The NetWorker administration 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 NMO, the only valid pool types are backup and backup clone.

66 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

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.

Configure any required NetWorker Pool resources and corresponding Label


Template resources with the NMC program. The NetWorker administration 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 169.

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 27 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 NMO software provides firewall support. The ports that the NMO software uses
for the firewall depend on the corresponding ports configured for the NetWorker
server.
To configure the firewall that the NMO software uses, follow the firewall
configuration instructions in the NetWorker administration guide for the particular
NetWorker server platform.

Configuring I18N support


“Internationalization (I18N) and localization (L10N)” on page 33 describes the
features of NMO internationalization (I18N) support.
To configure I18N support:
1. Ensure that you meet the “Requirements for I18N support” on page 68.
2. Follow the configuration steps in “Configure I18N support” on page 68.

Configuring I18N support 67


Software Configuration

Requirements for I18N support


Ensure that all of the following I18N requirements are met:
◆ The NMO 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 7.4 or later is installed:
• Internationalized NetWorker server software is installed, either on the NMO
client or on a remote host.
• If the NetWorker server is located on a remote host, internationalized
NetWorker client or storage node software is installed on the NMO client.
The NetWorker installation guide provides details on installation of the
NetWorker software.
◆ For I18N support during proxy operations, a supported release of the PowerSnap
Module is installed and configured, as described in the NetWorker Module for
Oracle release notes.
The NetWorker documentation provides details on any other I18N requirements.

Configure I18N support

Note: Configuration of proxy backups or restores with the NMO 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 NMO client host:


1. On UNIX only, 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.
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 /nolog
SQL*Plus: Release 10.1.0.2.0 - Production on Thu Apr 26 15:12:03
Copyright (c) 1982, 2004, Oracle. All rights reserved.
SQL> connect sys/oracle as sysdba;

68 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

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 188 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 183.

Configure L10N support


“L10N features” on page 35 describes the features of NMO localization (L10N)
support.

Note: L10N support for proxy operations requires PowerSnap Module 2.4 SP2 or later.

To configure NMO L10N support, enable the locale by installing the appropriate
NMO language pack for the user locale on the NMO client host. Each supported
language has its own language pack.
The NetWorker Module for Oracle installation guide provides details on how to
install a supported NMO language pack on a UNIX or Windows system.

Migrating a legacy configuration with the nsrnmoadmin command


You can use the nsrnmoadmin command to migrate a scheduled backup
configuration that was created with a legacy method (without the configuration
wizard in NMO 5.0) to the configuration storage framework that is supported by the
new NMO 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 NMO wizard to
retrieve an existing configuration from an RMAN script.

Configure L10N support 69


Software Configuration

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 70.
2. Use the proper nsrnmoadmin command and options, according to the
“Migration command syntax and options” on page 71.

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:
◆ “Windows 2008 and Windows Vista requirements for the nsrnmoadmin
command” on page 189 are met on Windows 2008 or Windows Vista.
◆ The required NetWorker releases are installed, as described in the NetWorker
Module for Oracle release notes.
◆ 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.

◆ The RMAN script on Microsoft Windows does not include non-ASCII characters.
◆ 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

70 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

for the client by using NMC to edit the resource. The NetWorker
administration 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.

! 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 system@physical_hostname
(Windows) or root@physical_hostname (UNIX).
- After the migration, use NMC to edit the Lockbox resource for the cluster virtual
client, and add the name system@physical_hostname (Windows) or
root@physical_hostname (UNIX) to the resource.

Migration command syntax and options


You must type the nsrnmoadmin command for migration on the operating system
command line, as one of the following users:
◆ Root user on UNIX
◆ A member of the Microsoft Windows Administrators group
The command user must meet the migration requirements listed in Table 3 on
page 65.
“Windows 2008 and Windows Vista requirements for the nsrnmoadmin command”
on page 189 provides details for Windows 2008 and Windows Vista systems.
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.

Migrating a legacy configuration with the nsrnmoadmin command 71


Software Configuration

Table 4 on page 72 describes the nsrnmoadmin command options.

Table 4 Options of the nsrnmoadmin command for migration

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 NMO 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.

-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.

Configuring a backup with the NMC wizard


To configure a backup with the wizard:
1. Review the information in “About the backup configuration wizard” on page 72.
2. Ensure that you meet the “Requirements for using the backup configuration
wizard” on page 73.
3. Follow the configuration steps in “Configure a backup with the wizard” on
page 73.

About the backup configuration wizard


NMO 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 30 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

72 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

must be configured manually (without the wizard) through NMC, as described in


“Configuring a backup with the NMC legacy method” on page 75.
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 NMO 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 69.
The following sources provide more information on the configuration wizard:
◆ NetWorker Module for Oracle installation guide
◆ NetWorker administration guide or release notes
◆ 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.
◆ Communication between the NMC server, NetWorker server, and NMO 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 NMO client hosts, as described in the NetWorker Module for Oracle
release notes.

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.

Configuring a backup with the NMC wizard 73


Software Configuration

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

74 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

Configuring a backup with the NMC legacy method


The following sections describe how to configure backups with the legacy method
(without the configuration wizard).

About backup configuration with the legacy method


NMO release 5.0 still supports the legacy method of backup configuration, as used in
previous NMO 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 72. 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 200. The use of the send command is recommended
where possible. “The send command” on page 211 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.
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 26 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.emc.com:
run {
allocate channel t1 type ’SBT_TAPE’;
allocate channel t2 type ’SBT_TAPE’;
send ’NSR_ENV=(NSR_SERVER=mars.emc.com,
NSR_DATA_VOLUME_POOL=MondayFulls)’;
backup full filesperset 4 format ’FULL_%d_%U’ (database);

Configuring a backup with the NMC legacy method 75


Software Configuration

release channel t1;


release channel t2;
}
To specify a Media Management (in this case, NMO) 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 NMO 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 114 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.emc.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.emc.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 201 provides more information.

76 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

! 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 NMO
backup, the encrypted data cannot be recovered. The NetWorker administration
guide provides more information on pass phrases.

Example 17 on page 122 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, NMO does not support the following for scheduled backups:
allocate channel t1 type ’SBT_TAPE’ send
’NSR_ENV=(NSR_SERVER=mars.emc.com)’;
The following is the correct form of the commands:
allocate channel t1 type ’SBT_TAPE’;
send channel t1 ’NSR_ENV=(NSR_SERVER=mars.emc.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 26 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 4
format ’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

Configuring a backup with the NMC legacy method 77


Software Configuration

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 111 provides more
information on how to cancel a scheduled backup.
The remainder of the scheduled backup script in Example 12 on page 77, starting
with the first allocate channel command, is similar to the manual backup script in
Example 10 on page 75 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. NMO 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

78 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

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 backups
When 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 NMO installation on the Oracle Server host stores the following files:
◆ A sample nsrnmo script in the same directory as the NetWorker executables. The
sample script filenames are:
• nsrnmo on UNIX
• nsrnmo.bat on Windows
◆ A template of the script in the /etc/nsrnmo.sh file on UNIX. Do not modify this
file.

Note: On Windows, save a copy of the original nsrnmo.bat file before modifying it.
“Editing the nsrnmo.bat file on Windows 2008 or Windows Vista” on page 80 provides
specific requirements for Windows 2008 and Windows Vista systems.

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 save. On Windows, the name must also
end in .bat or .cmd.

Configuring a backup with the NMC legacy method 79


Software Configuration

To create a new version of the nsrnmo script:


1. Copy the sample script file (nsrnmo on UNIX or nsrnmo.bat on Windows) 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
or nsrnmo_prodDB.bat on Windows.
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 NMO release works with NMO release 5.0.
However, the nsrnmo script installed with NMO 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 NMO
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

Editing the nsrnmo.bat file on Windows 2008 or Windows Vista


Only a member of the Microsoft Windows Administrators group can edit the
nsrnmo.bat file or any file located in the NetWorker_installation_dir\bin directory on
Windows.
On Windows 2008 or Windows Vista, you must follow these steps to edit the
nsrnmo.bat file in the NetWorker_installation_dir\bin directory:
1. Click Start.
2. Right-click Notepad or any other text editor listed as a menu item.
3. Select Run as administrator.
4. Open the NetWorker_installation_dir\bin\nsrnmo.bat file in the text editor.
5. Edit and save the nsrnmo.bat file.

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 NMO is installed on a sparse root zone and /usr is one of the shared
directories, the nsrnmo script in /usr/sbin 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 /usr/sbin. 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 87 provides more information on the Client resource.

Parameters in the nsrnmo script


The sample nsrnmo script installed with NMO contains parameters that must be
customized for a particular scheduled Oracle backup. The parameters in the sample
nsrnmo script are all undefined by default.

80 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

The nsrnmo script contains only the parameters required on the particular system
where the script is installed. The following parameters appear in the nsrnmo script
on specific UNIX systems only:
◆ LD_LIBRARY_PATH appears in the script on:
• HP-UX (64-bit Itanium)
• Linux
• Solaris
◆ LIBPATH appears in the script on AIX.
◆ SHLIB_PATH appears in the script on HP-UX (32-bit and 64-bit PA-RISC).
Follow the instructions in the Oracle installation guide to determine which of the
parameters (LD_LIBRARY_PATH, LIBPATH, SHLIB_PATH) to set on a particular
operating system and to what value.
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 either the UNIX setenv
command or Windows set 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 81
◆ “LD_LIBRARY_PATH” on page 82
◆ “LIBPATH” on page 82
◆ “SHLIB_PATH” on page 82
◆ “NSR_RMAN_ARGUMENTS” on page 82
◆ “NSR_SB_DEBUG_FILE” on page 83
◆ “ORACLE_SID” on page 83
◆ “ORACLE_USER” on page 83
◆ “PRECMD” on page 84
◆ “POSTCMD” on page 85
◆ “TNS_ADMIN” on page 86

ORACLE_HOME
This 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

Configuring a backup with the NMC legacy method 81


Software Configuration

LD_LIBRARY_PATH
This 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_HOME
export NSR_RMAN_ARGUMENTS
:
:
export LD_LIBRARY_PATH
}

LIBPATH
This parameter is optional for a scheduled backup. It appears in the nsrnmo script on
AIX only. Set LIBPATH in the nsrnmo script file to the pathname of the directory that
contains the Oracle libraries, typically $ORACLE_HOME/lib.
To set LIBPATH in the nsrnmo script, follow the example shown in
“LD_LIBRARY_PATH” on page 82. Replace LD_LIBRARY_PATH with LIBPATH.
Uncomment the line “export LIBPATH” in the script.

SHLIB_PATH
This parameter is optional for a scheduled backup. It appears in the nsrnmo script on
HP-UX only. Set SHLIB_PATH in the nsrnmo script file to the pathname of the
directory that contains the Oracle libraries, typically $ORACLE_HOME/lib.
To set SHLIB_PATH in the nsrnmo script, follow the example shown in
“LD_LIBRARY_PATH” on page 82. Replace LD_LIBRARY_PATH with
SHLIB_PATH. Uncomment the line “export SHLIB_PATH” in the script.

NSR_RMAN_ARGUMENTS
This 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"

82 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

NSR_SB_DEBUG_FILE
This 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
“NMO error messages” on page 222 provides more information on how the debug
information is written if the parameter NSR_SB_DEBUG_FILE is set to an invalid
pathname.

ORACLE_SID
This 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_USER
This parameter is optional for a scheduled backup configured through the legacy
method (without the wizard) on UNIX only.
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 NMO backup through operating system


authentication is not supported for the following:
- A scheduled backup configured through the new configuration wizard.
- A scheduled backup on Microsoft Windows.
- A probe-based backup.
- A scheduled proxy backup.

Configuring a backup with the NMC legacy method 83


Software Configuration

For example, to perform a scheduled NMO 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

PRECMD
This parameter is optional. Set PRECMD to the complete pathname of a file containing
a preprocessing script to be executed before the RMAN backup script.

Note: The pathname value of PRECMD must not contain any spaces. For example, instead of
setting PRECMD to C:\Program Files\Legato\nsr\precmd.bat, set the parameter to
C:\Progra~1\Legato\nsr\precmd.bat.

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 Here
TARGET_DATABASE=proddb
ERRFILE=/tmp/precmd.log
ORACLE_USER=proddb
#
# Scheduled backup launches as "root" run PRECMD as oracle user
su - $ORACLE_USER -c "{
# Shutdown target database and start in maintenace mode
export ORACLE_SID=$TARGET_DATABASE
svrmgrl <<EOF1 > $ERRFILE
connect internal
shutdown
startup mount
exit
EOF1
}"
# Do checking as root since "su" will always return its status
export status
# Define status otherwise it is local to if [] statement
status=0
# Scan file for errors and only return true or false
grep -e error -e warning -e fatal $ERRFILE > /dev/null 2>&1
if [ $? -eq 0 ]; then
status=1
else

84 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

status=0
fi
exit $status

POSTCMD
This 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.

Note: The pathname value of POSTCMD must not contain any spaces. For example, instead of
setting POSTCMD to C:\Program Files\Legato\nsr\postcmd.bat, set the parameter to
C:\Progra~1\Legato\nsr\postcmd.bat.

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 Here
TARGET_DATABASE=proddb
ERRFILE=/tmp/postcmd.log
ORACLE_USER=proddb
#
# Scheduled backup launches as "root" run POSTCMD as user proddb
su - $ORACLE_USER -c "{
# Mount target database as ready for use
export ORACLE_SID=$TARGET_DATABASE
svrmgrl <<EOF > $ERRFILE
connect internal
alter database open;
exit
EOF
}"
# Do checking as root since "su" will always return its status
export status
# Define status otherwise it is local to if [] statement
status=0
# Scan file for errors and only return true or false
grep -e ORA- -e error -e warning -e fatal $ERRFILE > /dev/null 2>&1
if [ $? -eq 0 ]; then
status=1
else
status=0
fi
exit $status

Configuring a backup with the NMC legacy method 85


Software Configuration

TNS_ADMIN
This 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 82. 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 72.
“Configuring a probe-based backup” on page 97 provides details on configuration
requirements for a probe-based backup.

For a scheduled NMO 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
administration guide and NMC online help provide more information.

86 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

Configure a Schedule resource with NMC


Review the summary information in “NetWorker Schedule resource” on page 66.
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 administration 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 87.

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 72.
“Configuring a probe-based backup” on page 97 provides details on configuration
requirements for a probe-based backup.

A NetWorker Client resource is a set of attributes assigned to the NMO client host
and stored on the NetWorker server. Before NMO software can be used for backups
or restores, a Client resource must be configured for the NMO client host.
If the NetWorker server software is installed on the NMO client host, a basic Client
resource for the NMO client is created automatically during the NetWorker
installation. The Client resource must be customized for an NMO backup.
“NetWorker indexes and policies used for restores” on page 118 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 170 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 88.
Leave the following attributes blank:
◆ Directive
◆ Archive Users
◆ Remote User
◆ Password

Configuring a backup with the NMC legacy method 87


Software Configuration

Note: On a Solaris system with Solaris zones, ensure that the security fields (such as Remote
Access and Privileges) of NetWorker resources used during NMO backups and restores refer
to the hostname of the zone in which NMO 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 90 shows a sample Client resource for a UNIX NMO client.
Figure 7 on page 91 shows a sample Client resource for a Windows NMO client.

Table 5 NetWorker Client resource attributes (page 1 of 2)

Attribute Description

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, or nsrnmo.bat or nsrnmo2.bat on Windows.
On a Solaris system with Solaris zones where NMO 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 86
provides more information on NetWorker backup groups.

88 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

Table 5 NetWorker Client resource attributes (page 2 of 2)

Attribute Description

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 NMO
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. On Windows, the pathname can include forward slashes,
for example, RMAN:F:/scripts/incr_1_bkup.
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/archlogbkup
RMAN:/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 31.

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 114
• “Deduplication backup information in NetWorker indexes” on page 115

Schedule Specifies the NetWorker backup schedule to use for a scheduled backup. “Configure a Schedule resource with NMC” on
page 87 provides more information on NetWorker backup schedules.

Configuring a backup with the NMC legacy method 89


Software Configuration

Figure 6 UNIX Client resource for scheduled backups

90 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

Figure 7 Windows 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 administration 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 NMO
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 72 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 92.

Configuring a backup with the NMC legacy method 91


Software Configuration

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.
“Configuring the NWORA resource file with the nsrnmoadmin program” on
page 188 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 183.
◆ Ensure that NMO scheduled backups are configured properly according to the
“Configuration roadmap” on page 62.
◆ Ensure that the user group privileges for the root or administrative user on the
NMO 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 64.
◆ 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 79.
• 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 186.
NMO 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, NMO
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
NMO save set bundles.
◆ In a RAC system, ensure that all channels are allocated on the same NMO 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 NMO
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 72 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 91, policy uniformity should also be enabled.

92 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

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.
“Configuring the NWORA resource file with the nsrnmoadmin program” on
page 188 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 183.
◆ Ensure that NMO scheduled backups are configured properly according to the
“Configuration roadmap” on page 62.
◆ Ensure that the user group privileges for the root or administrative user on the
NMO 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 64.
◆ 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 79.
• 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 186.
NMO 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, NMO
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 NMO 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 backup


To configure a deduplication NMO backup:
1. Review the following sections:
• “Deduplication backups and restores” on page 21
• “Requirements for a deduplication backup” on page 94
• “Best practices for a deduplication backup” on page 94
2. Follow the configuration steps in “Configure a deduplication backup” on
page 95.

Configuring a deduplication backup 93


Software Configuration

Requirements for a deduplication backup


Before you configure a deduplication backup, ensure that all of the following
requirements are met:
◆ The NMO client platform and operating system support deduplication. The EMC
Information Protection Software Compatibility Guide on the Powerlink website
provides details on the platforms and operating systems that support
deduplication.
◆ The NetWorker client and server releases support NMO deduplication. The
NetWorker Module for Oracle release notes 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 NMO deduplication backup) is configured as an advanced file type
device (AFTD), as described in the NetWorker administration guide.

Best practices for a deduplication backup


This section provides recommendations on when to use NMO 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.

94 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

◆ 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 NMO 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 203
• “NSR_DEDUP_CACHE_TAG” on page 204
◆ 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.emc.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 72.

◆ To use the configuration wizard (for a scheduled backup only), follow the
instructions in “Configuring a backup with the NMC wizard” on page 72.
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 95


Software Configuration

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 75.

Note: For a manual deduplication backup, the Client resource needs to include only the
two attribute settings in step a on page 96.

Apply these additional settings for a deduplication backup:


a. Set the following attributes in the Client resource for the NMO 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 203 (for a manual backup only)
– “NSR_DEDUP_CACHE_TAG” on page 204
– “NSR_DEDUP_NODE” on page 204 (for a manual backup only)
Example 13 on page 96 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 96, 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.emc.com, NSR_CLIENT=oracle.emc.com,
NSR_DEDUP_BACKUP=TRUE, NSR_DEDUP_NODE=node3.emc.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;
}

96 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

Configuring a probe-based backup


To configure a probe-based NMO backup:
1. Review the following sections:
• “Probe-based backups” on page 23
• “Requirements for a probe-based backup” on page 97
2. Follow the configuration steps in “Configure a probe-based backup” on page 97.

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 NetWorker
Module for Oracle release notes.
◆ If you want to use the nsrnmoprobe program provided with NMO 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 97.
◆ 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 97.
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 /usr/sbin 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 administration 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 NMO 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 98.

Configuring a probe-based backup 97


Software Configuration

Table 6 NetWorker Probe resource attributes

Attribute Description

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 NMO 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 99 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 99 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 99 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 99 provides a sample setting of this parameter.

98 EMC NetWorker Module for Oracle Release 5.0 Administration Guide


Software Configuration

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
administration 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 NMO client according to the
instructions in the appropriate section:
• “Configuring a backup with the NMC wizard” on page 72
• “Configure a Client resource with NMC” on page 87
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 100 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 NMO software, the Probe
resource must be set up properly, as described in “Configure a probe-based backup”
on page 97.

Configuring a probe-based backup 99


Software Configuration

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.)
For example, Command Options is set as follows:
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 188 provides important details on the nsrnmoadmin command.
For example, Command Options is set as follows:
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 NMO 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 NMO client and storage node are Solaris machines.

100 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Software Configuration

The nsrnmoprobe program is installed with the NMO software in /usr/sbin on the
NMO client. The nsrnmoprobe program checks for the number of Oracle log files
generated on the NMO client.
A script named nsrjukeboxprobe is created with “execute” permissions and stored in
the /usr/sbin 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 — NMO 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 NMO client through the configuration wizard.
The Client resource includes the following attribute settings:
• Name — mars
• Backup Command — nsrnmostart -C
• Group — probe_group
• Probe — NMO 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

Configuring a probe-based backup 101


Software Configuration

• 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.

102 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
3

Backup Procedures

This chapter includes the following major sections:


◆ Manual backup procedures ....................................................................................... 104
◆ Scheduled backup procedures ................................................................................... 110
◆ Regular backup information in NetWorker indexes ............................................... 114
◆ Deduplication backup information in NetWorker indexes.................................... 115

Backup Procedures 103


Backup Procedures

Manual backup procedures


To perform a manual NMO backup of Oracle data:
1. Review information in “Product features” on page 20 about the features that
apply to your particular backup.
2. Review “Requirements for a manual backup” on page 104.
3. Review the information on how to cancel a manual backup in “Cancel a manual
backup” on page 107.
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 105.
• Use the Oracle Enterprise Manager Backup Management Tools, according to
“Manual backup through Oracle Enterprise Manager Backup Management
Tools” on page 105.

! 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 106.
Regular NetWorker server bootstrap backups help to ensure adequate preparation
for disaster recovery.

“Monitor a manual backup” on page 108 provides information on how to monitor the
status of a manual NMO backup.
“Scheduled backup procedures” on page 110 provides information on scheduled
NMO 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 75.

104 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
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 75 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\’
On Microsoft Windows, the command to run the RMAN script is rman.exe. The
appropriate Oracle backup and recovery documentation provides more information
on the rman or rman.exe command line options.

! 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 106.
Regular NetWorker server bootstrap backups help to ensure adequate preparation
for disaster recovery.

“Cancel a manual backup” on page 107 provides information on how to cancel a


manual NMO 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 NMO 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.

Manual backup procedures 105


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. Log in as follows:
• On a UNIX NetWorker server as the root user.
• On a Windows NetWorker server as the Windows system administrator.
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 administration guide for the particular NetWorker server platform
◆ savegrp entry in the NetWorker command reference guide
◆ savegrp man page on UNIX
The NetWorker administration guide provides information on bootstrap backups.

106 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Backup Procedures

The NetWorker disaster recovery guide provides information on how to use the
bootstrap backup during a disaster recovery.

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 64 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 111.

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, NMO 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 77 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.
To cancel a nonresponding Oracle backup on Windows, stop the nsrsbtcn.exe
process in Task Manager.

Manual backup procedures 107


Backup Procedures

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 administration 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 NMO diagnostic and error messages.
The following figures show the types of Oracle backup messages displayed in the
Monitoring Window of the NMC interface:
◆ Figure 8 on page 108 shows messages displayed in the Sessions tab.
◆ Figure 9 on page 109 shows messages displayed in the Devices tab.
◆ Figure 10 on page 109 show messages displayed in the Log tab.
The NetWorker administration guide and NMC online help provide more
information.

Figure 8 Oracle backup messages in Sessions tab of Monitoring window

108 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Backup Procedures

Figure 9 Oracle backup messages in Devices tab of Monitoring window

Figure 10 Oracle backup messages in Log tab of Monitoring window

Manual backup procedures 109


Backup Procedures

Scheduled backup procedures


To prepare for a scheduled NMO backup of Oracle data:
1. Review information in “Product features” on page 20 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 62.
3. Run a test scheduled Oracle backup according to “Test a scheduled backup” on
page 110.
4. Review the steps for canceling a scheduled backup in “Cancel a scheduled
backup” on page 111.
At the end of a successful scheduled backup, NMO automatically backs up the
NWORA resource file if it exists, as described in “NWORA resource file backup” on
page 173.
“Monitor a scheduled backup” on page 112 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 114
◆ “Deduplication backup information in NetWorker indexes” on page 115

Test a scheduled backup


After the NMO 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, or as the administrator
Microsoft user on Windows.
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 administration 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.

110 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Backup Procedures

Note: “Cancel a scheduled backup” on page 111 provides information on how to cancel the
scheduled Oracle backup.

Scheduled backup error messages


If the scheduled backup fails, an error message is produced. The NetWorker
administration 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 83 provides details on the parameter.
If the scheduling portion of a scheduled NMO 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 221.

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 64 provides more information.
NMO 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.
◆ If the NetWorker server system is Windows, then the Stop button is not
supported.

Note: Although the progress message might indicate that the backup has exited
prematurely (before finishing), the NMO 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 NMO
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.)

Scheduled backup procedures 111


Backup Procedures

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;
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 107 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 108 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 NMO 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 NMO 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 NMO for the backup. Figure 11 on
page 113 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 administration guide provides details on viewing information about


scheduled backups in the NMC program.

112 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Backup Procedures

Figure 11 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 NMO 5.0 wizard.

The parameters NSR_SERVER and NSR_GROUP must not be set in a scheduled


backup RMAN script. NMO automatically passes server and group information
(received from the NetWorker server that started the backup) to Oracle Server
processes, created for each allocated channel.
NMO 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, NMO invokes the RMAN
executable with send as a command line option. “The send command” on page 211
provides more information.

Scheduled backup procedures 113


Backup Procedures

Regular backup information in NetWorker indexes


The 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 115 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 NMO 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 NMO
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 size
NMO.001 ca-oracle1 12/11/08 16:05:59 145 MB
ssid fl level name
4212032038 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.

114 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Backup Procedures

Cross-check the client file index and media database by using the save time. For
example:
mminfo -c ca-oracle1 -t 1192133159
nsrinfo -n oracle -t ’12/11/08 16:05:59’ ca-oracle1

Deduplication backup information in NetWorker indexes


The 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.emc.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

Deduplication backup information in NetWorker indexes 115


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.

116 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
4
Data Restore and
Recovery

This chapter includes the following sections:


◆ About restore and recovery ........................................................................................ 118
◆ RMAN scripts for restore and recovery ................................................................... 121
◆ Performing an Oracle data restore and recovery.................................................... 123

Data Restore and Recovery 117


Data Restore and Recovery

About restore and recovery


Only 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 NMO.

Review the following sections in preparation for the Oracle data restore and recovery
procedures:
◆ “NetWorker indexes and policies used for restores” on page 118
◆ “Recovery configuration wizard” on page 118
◆ “RMAN scripts for restore and recovery” on page 121

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.
NetWorker Module for Oracle (NMO) 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 75.

The NetWorker administration 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


NMO 5.0 software supports a new recovery configuration wizard that is integrated
with the NetWorker Management Console (NMC). The recovery configuration
wizard used in NMO release 4.5 or earlier is replaced by the NMC-based wizard in
NMO release 5.0.

118 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Data Restore and Recovery

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 NetWorker Module for Oracle release notes 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 119.
2. Ensure that you meet the “Requirements for using the recovery configuration
wizard” on page 120.
3. Follow the steps in “Configure a restore with the wizard” on page 120.

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 121 provides more information
about RMAN scripts for restore and recovery.

About restore and recovery 119


Data 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:
◆ NetWorker Module for Oracle installation guide
◆ 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 NMO client
configuration is stored.
◆ Communication between the NMC server, NetWorker server, and NMO 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 NMO client hosts, as described in the NetWorker Module for Oracle
release notes.
◆ The NetWorker Client resource for the NMO client was created through one of
the following:
• Backup configuration wizard in NMO 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.

120 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Data Restore and Recovery

5. To start the wizard, right-click the NMO 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.

! 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 recovery


An 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 118 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 200. The use of the send command is recommended
where possible. “The send command” on page 211 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.emc.com. The Oracle data is restored to the
NetWorker client server1.emc.com. This RMAN script also includes the recovery
step, which is explained in “Perform an Oracle data recovery” on page 127:
run {
allocate channel t1 type ’SBT_TAPE’;
allocate channel t2 type ’SBT_TAPE’;
send ’NSR_ENV=(NSR_SERVER=mars.emc.com,
NSR_CLIENT=server1.emc.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 75 provides more information on setting
NSR* parameters in an RMAN script.

RMAN scripts for restore and recovery 121


Data Restore and Recovery

Example 17 RMAN script to restore an AES encrypted backup

By default, if the NSR_ENCRYPTION_PHRASES parameter is not set, NMO 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 76 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 205 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.emc.com, NSR_CLIENT=server1.emc.com,
NSR_ENCRYPTION_PHRASES="original_encryption_phrase")’;
restore database;
release channel t1;
}

Example 18 RMAN script to restore from a specified pool

By default, NMO and NetWorker use configuration settings and information in the
media database to determine the backup volume to use for an NMO 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 206 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;
}

122 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Data Restore and Recovery

Performing an Oracle data restore and recovery


To perform a restore and recovery of Oracle data:
1. Review “Requirements for an Oracle data restore” on page 123.
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 125.
• Use the Oracle Enterprise Manager Backup Management Tools, according to
“Restore with Oracle Enterprise Manager Backup Management Tools” on
page 126.

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 127.

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 121.
◆ 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 123.

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 NMO software in the same directory as the NetWorker client
software.

Note: On Windows, the command name is nsrnmoinfo.exe.

Performing an Oracle data restore and recovery 123


Data Restore and Recovery

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.

124 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Data Restore and Recovery

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 126.
If the RMAN restore script on page 121 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\’

Performing an Oracle data restore and recovery 125


Data Restore and Recovery

On Microsoft Windows, the command to run the RMAN script is rman.exe. The
appropriate Oracle backup and recovery documentation provides more information
on the rman or rman.exe command options.

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 NMO 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 NMO 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.

126 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Data Restore and Recovery

Perform an Oracle data recovery


After restoring the NMO 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 121.
◆ 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.

Performing an Oracle data restore and recovery 127


Data Restore and Recovery

128 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
5

Disaster Recovery

This chapter includes the following sections:


◆ About disaster recovery ............................................................................................. 130
◆ Preparing an Oracle database for disaster recovery .............................................. 130
◆ Performing a disaster recovery ................................................................................. 133
◆ Sample postcommand script ..................................................................................... 134

Disaster Recovery 129


Disaster Recovery

About disaster recovery


It 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 recovery


To prepare an Oracle database environment for disaster recovery, you must back up
the following minimum list of files:
◆ Oracle database (all the datafiles)
◆ Archived redo logs
◆ 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:
• On UNIX, $ORACLE_HOME/dbs/orapw$ORACLE_SID
• On Windows, %ORACLE_HOME%\database\PWD%ORACLE_SID%.ora
◆ Registry files:
• On UNIX, oratab is typically in /var/opt/oracle or /etc
• On Windows, My Computer\HKEY_LOCAL_MACHINE\SOFTWARE\Oracle
◆ 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.

130 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Disaster Recovery

◆ Back up the archived redo logs frequently between database backups.


◆ Back up the Recovery Catalog after every target database backup.
◆ 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 106.
To back up the required files in preparation for disaster recovery, perform the
following:
1. “Create the DBID text file” on page 131
2. “Set up a postcommand script for backup of Oracle-related files” on page 131
3. “Set up RMAN backups of the database and related files” on page 132
4. “Set up RMAN backups of Recovery Catalog” on page 132

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 131.
In the sample postcommand script provided with the NMO software, the DBID text
file is assumed to be dbid.txt, located in the $ORACLE_HOME directory. “Sample
postcommand script” on page 134 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 131
◆ Password file in the following location by default:
• On UNIX, $ORACLE_HOME/dbs/orapw$ORACLE_SID
• On Windows, %ORACLE_HOME%\database\PWD%ORACLE_SID%.ora
◆ Registry files:
• On UNIX, oratab is typically in /var/opt/oracle or /etc
• On Windows, My Computer\HKEY_LOCAL_MACHINE\SOFTWARE\Oracle
◆ RMAN scripts, if applicable
You can either create a postcommand script from scratch, or modify the
postcommand script that is provided with the NMO software.

Preparing an Oracle database for disaster recovery 131


Disaster Recovery

On UNIX, you can use any name for the postcommand script. On Windows, the
script name must end in .bat.

Note: During a scheduled NMO backup, the nsrnmostart process passes the options
-s server_name -g group_name to the postcommand script if the script name begins with
nsrnmodr.

In a scheduled RMAN backup, include the postcommand script by one of the


following methods:
◆ If you use the NMO 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 134 provides details on the postcommand
script that is provided with the NMO software.

Set up RMAN backups of the database and related files


Set up an RMAN backup with NMO to back up the following files:
◆ Oracle database (all the datafiles)
◆ Archived redo logs
◆ 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 NMO.
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 131 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 132.
Oracle documentation provides more information on setting up and running
Recovery Catalog backups.

132 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Disaster Recovery

Performing a disaster recovery


The following sections provide guidelines for different types of disaster recovery
operations:
◆ “NetWorker server recovery” on page 133
◆ “Oracle disaster recovery to a new host” on page 133

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 106 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 15 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 NMO 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

Performing a disaster recovery 133


Disaster Recovery

Note: On Windows, you may need to reinsert the oracle.reg file into the registry after
recovering it, for example, with the following command:
regedit /S C:\temp\oracle.reg
The Oracle documentation provides more details.

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.

Sample postcommand script


You 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 131. The NMO software includes a sample postcommand script that you can
modify for your environment.
The NMO installation provides a sample postcommand script that is specific to UNIX
or Windows, depending on the platform where NMO is installed. The sample script
is installed in the bin subdirectory under the NetWorker software directory, for
example, under /usr/sbin.
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 NMO 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 NMO as follows:


◆ Example 21 on page 134
◆ Example 22 on page 135

Example 21 Sample postcommand script on UNIX

The sample postcommand script named nsrnmodrpostcmd is installed with the


NMO 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=no
SRV=no
complete=0
shift
shift

134 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Disaster Recovery

while [ "$#" -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
shift
done

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
save -s $SRV -g $GRP /var/opt/oracle/oratab
save -s $SRV -g $GRP $ORACLE_HOME/dbid.txt
fi

Example 22 Sample postcommand script on Windows

The sample postcommand script named nsrnmodrpostcmd.bat is installed with the


NMO software on Windows.
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.bat script provided on Windows is as follows:
echo off
set SRV=no
set GRP=no
set ORACLE_HOME=
shift
set ORACLE_SID=
shift

:start
if %1==-g goto assignG
if %1==-s goto assignS
if not exist %1 goto fail
shift
goto start

:assignG
set GRP=%2
shift
shift
if %SRV%==no goto start
goto end

:assignS
set SRV=%2
shift
shift
if %GRP%==no goto start

:end

Sample postcommand script 135


Disaster Recovery

save -s %SRV% -g %GRP% %ORACLE_HOME%\network\admin


save -s %SRV% -g %GRP% %ORACLE_HOME%\database\PWD%ORACLE_SID%.ora
regedit -E C:\temp\oracle.reg HKEY_LOCAL_MACHINE\SOFTWARE\Oracle
save -s %SRV% -g %GRP% C:\temp\oracle.reg
save -s %SRV% -g %GRP% %ORACLE_HOME%\dbid.txt
:fail

136 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
6
Cluster and RAC
Systems

This chapter includes the following sections:


◆ Considerations for cluster and RAC systems ......................................................... 138
◆ Cluster systems............................................................................................................ 139
◆ RAC systems ................................................................................................................ 140

Cluster and RAC Systems 137


Cluster and RAC Systems

Considerations for cluster and RAC systems


To prepare for successful NetWorker Module for Oracle (NMO) operations on cluster
and RAC systems, consider the following:
◆ “Parameters in the RMAN script or session” on page 138
◆ “NetWorker software configuration” on page 139
◆ “Backup failover” on page 139

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 administration 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.

138 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Cluster and RAC Systems

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 installation guide 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 87 provides information on
how to configure a Client resource for NMO 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 NMO 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 144 provides more information.

Cluster systems
A 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 NMO 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, NMO 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.

Cluster systems 139


Cluster and RAC Systems

To configure a cluster system for Oracle backup and restore operations that use
NMO, follow the “Roadmap for backup/restore configuration in a cluster system” on
page 140.

Roadmap for backup/restore configuration in a cluster system


The Oracle Server and NetWorker server systems must be configured before NMO
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 NMO software is properly installed according to the instructions in the
NetWorker Module for Oracle installation guide.
◆ The Oracle Server and NetWorker server system configurations are completed
according to the instructions in the “Configuration roadmap” on page 62.
To properly configure Oracle backup and restore operations with NMO 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 138 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
The NMO 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.emc.com. An Oracle instance is a memory structure and a group of Oracle
Server processes running on a node.

140 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Cluster and RAC Systems

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, NMO 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 138 provides more information on setting the
parameter.
NMO 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 141.

Roadmap for backup/restore configuration in a RAC system


To properly configure Oracle backup and restore operations with NMO 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 144
provides guidelines. The appropriate Oracle documentation provides more
information.
4. Install NMO on each node of the cluster to be used for backup and recover
operations.
The NetWorker Module for Oracle installation guide 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 146
• “Creating RMAN restore scripts” on page 146

RAC systems 141


Cluster and RAC Systems

7. Review the additional issues concerning Oracle recover operations in “Archived


redo logs” on page 147.
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 142.

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 NMO backup.
2. On the NetWorker server, create a NetWorker Storage Node resource for each
RAC node to be used for the NMO backup. The NetWorker administration 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 NMO backup. The NetWorker administration 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 NMO 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 NMO backup
and to initiate the backup.
7. For the RAC node that will initiate the NMO backup, create a NetWorker Client
resource with the attribute settings required for the backup, as described in
“NetWorker software configuration” on page 139:
• 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:
curphyhost
nsrserverhost
8. For each of the other RAC nodes that will not initiate the NMO backup, create a
basic Client resource. (Settings in these other Client resources do not affect the
NMO backup.)
9. On the RAC node that will initiate the NMO 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 138.
Example 23 on page 142 shows how to set up three RAC nodes as storage nodes for
NMO backups.

Example 23 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 NMO 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.

142 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Cluster and RAC Systems

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 NMO 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:
curphyhost
nsrserverhost
◆ 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)’;

RAC systems 143


Cluster and RAC Systems

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.emc.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.
◆ 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.

144 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Cluster and RAC Systems

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 NMO 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)
• NMO
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 NMO backup.

RAC systems 145


Cluster and RAC Systems

Creating RMAN backup scripts


A single RMAN backup script can be used to run a parallel Oracle backup with NMO
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 24 RMAN script for a manual Oracle backup on a RAC system

Suppose a RAC system consists of two nodes named node1.emc.com and


node2.emc.com. The Oracle instances named instance1 and instance2 are running on
node1.emc.com and node2.emc.com, respectively. The NetWorker server is located
on a separate node, server1.emc.com.
The following RMAN script for a manual backup is intended to run on
node1.emc.com by using the NOCATALOG mode of RMAN. The script sets
NSR_CLIENT to node1.emc.com and NSR_SERVER to server1.emc.com. As a result,
the NetWorker server stores the backup information in the node1.emc.com client file
index, as described in “NSR_CLIENT” on page 138. Two channels are allocated to
each of the nodes, node1.emc.com and node2.emc.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.emc.com,
NSR_SERVER=server1.emc.com);
backup filesperset 1
format ’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 147 provides a sample script
to back up all the archive log files in a RAC system.

Creating RMAN restore scripts


A single RMAN restore script can be used to run a parallel Oracle restore with NMO
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: NMO 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 25 RMAN script for an Oracle restore on a RAC system

Refer to Example 24 on page 146. A RAC system consists of two nodes named
node1.emc.com and node2.emc.com. The Oracle instances named instance1 and
instance2 are running on node1.emc.com and node2.emc.com, respectively. The
NetWorker server is located on a separate node, server1.emc.com.

146 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Cluster and RAC Systems

The following RMAN restore script is to be run on node2.emc.com. The script


restores the backup that was created by the backup script in Example 24 on page 146.
This restore script sets NSR_CLIENT to node1.emc.com and NSR_SERVER to the
remote NetWorker server name. As a result, the NetWorker server will obtain the
backup information from the node1.emc.com client file index. Two channels are
allocated to each of the nodes, node1.emc.com and node2.emc.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.emc.com,
NSR_SERVER=server1.emc.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.emc.com for each channel in the RMAN backup script. In order to restore the
backup data to node2.emc.com, add node2.emc.com to the Remote Access attribute in
the NetWorker Client resource for node1.emc.com.
The NetWorker administration guide provides more information on how to modify
the Client resource.
“Restoring all archived logs from each node” on page 148 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 147
◆ “Restoring all archived logs from each node” on page 148

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.emc.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.emc.com)’;

RAC systems 147


Cluster and RAC Systems

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.emc.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.emc.com, NSR_CLIENT=ops1.emc.com)’;
restore (archive log all);
release t1;
release t2;
}

148 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
7
Proxy Backups and
Restores

This chapter includes the following sections:


◆ Overview of proxy backups and restores ................................................................ 150
◆ Proxy backup and restore system ............................................................................. 151
◆ Proxy backup processes ............................................................................................. 153
◆ Proxy restore processes .............................................................................................. 159
◆ Software configuration roadmap .............................................................................. 163
◆ Proxy backup requirements....................................................................................... 171
◆ Proxy backup information in the NetWorker indexes ........................................... 174
◆ Proxy restore requirements........................................................................................ 177
◆ Catalog synchronization for proxy backups ........................................................... 181
◆ Proxy backups and restores on cluster systems...................................................... 194

Proxy Backups and Restores 149


Proxy Backups and Restores

Overview of proxy backups and restores


Proxy backups and restores provide continuous snapshot-based protection and
availability of Oracle data on specific types of primary storage. The NetWorker
Module for Oracle (NMO) 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 NMO.
The EMC Information Protection Software Compatibility Guide on the Powerlink
website provides a complete list of supported PowerSnap Modules.
◆ 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 Module for Oracle installation guide
◆ NetWorker PowerSnap Module documentation
(Refer to the PowerSnap Module version for the primary storage system.)

Types of proxy backups


NMO supports the following types of proxy backups, in cooperation with the
appropriate PowerSnap Module:
◆ “Instant backups” on page 150
◆ “Live backups” on page 150

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 163 provides more
information.

Note: NMO supports scheduled instant backups only. Manual (unscheduled) instant backups
are not supported.

Live backups
NMO 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.
◆ 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.

150 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

A snapshot policy must be configured to enable live backups. “Software


configuration roadmap” on page 163 provides more information.

! 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.

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


NMO supports the following types of proxy restores, in cooperation with the
appropriate PowerSnap Module:
◆ “Instant restores” on page 151
◆ “Rollbacks” on page 151
◆ “Restores from secondary storage” on page 151

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 system


For proxy backups and restores, the NMO, 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 EMC software components
work together.

Proxy backup and restore system 151


Proxy Backups and Restores

The Oracle backup and restore system consists of the following:


◆ Oracle Server
◆ Recovery Manager (RMAN)
◆ Recovery Catalog (optional)
◆ Oracle Enterprise Manager Backup Management Tools (optional)
The following sections provide more information on RMAN and the Recovery
Catalog:
◆ “Oracle Recovery Manager” on page 50
◆ “Oracle Recovery Catalog” on page 51
The EMC software components required for proxy backups are the following:
◆ NetWorker server
◆ NetWorker client (installed on the Oracle Server and proxy client host)
◆ NMO (installed on the Oracle Server)
◆ NetWorker PowerSnap Module (installed on the Oracle Server and proxy client
host)

Note: NetWorker storage node software can be optionally installed on the proxy client host.

Figure 12 on page 153 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 153
◆ “Proxy restore processes” on page 159

152 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Oracle Server host PowerSnap proxy client NetWorker


(NetWorker storage node) server

RMAN NetWorker
client

Oracle
Server

NetWorker
Module
for Oracle NetWorker
Recovery
MML client
Catalog

NetWorker
server
PowerSnap PowerSnap
Module Module

Storage
Oracle Point-in-time copy medium
data files of Oracle data files
Primary storage
GEN-000174

Figure 12 Proxy backup and restore system

Proxy backup processes


During a proxy backup of Oracle data, Oracle Server processes interact with the
NetWorker server through the NMO and PowerSnap Module software. The NMO
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.

Proxy backup processes 153


Proxy Backups and Restores

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 administration
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 154
◆ “Scheduled immediate live backup” on page 156
◆ “Deferred live backup” on page 158

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:
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,
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 191
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 NMO program,
nsrnmostart.

154 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

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 183.
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 158.

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 13 on page 156 shows how the Oracle Server and NetWorker processes
interact during a scheduled instant backup.

Proxy backup processes 155


Proxy Backups and Restores

Oracle Server PowerSnap proxy client NetWorker


(NetWorker client) (NetWorker storage node) server
Primary storage
Oracle Point-in-time copy
data files of Oracle data files
Client file index

data

tracking
information
Oracle PowerSnap PowerSnap nsrindexd
process processes processes

NetWorker Media database


Module
for Oracle nsrpsd nsrexecd nsrmmd
MML
PowerSnap
RMAN elements
nsrmmdbd
nsrnmostart NetWorker
Module
nsrnmo elements Snapshot
pool media

nsrsnap nsrd
interprocess communication

nsrsnapck nsrexecd savegrp

savefs
nsroraclecat Start
(optional) Here

GEN-000175

Figure 13 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 154.

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,
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.

156 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

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 191
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 NMO 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 183.
10. The savegrp program backs up the NetWorker server bootstrap and Oracle client
file indexes.
Figure 14 on page 158 shows how the Oracle Server and NetWorker processes
interact during a scheduled immediate live backup.

Proxy backup processes 157


Proxy Backups and Restores

Oracle Server PowerSnap proxy client NetWorker


(NetWorker client) (NetWorker storage node) server
Primary storage
Oracle Point-in-time copy
data files of Oracle data files
Client file index

data

tracking
information
Oracle PowerSnap PowerSnap nsrindexd
process processes processes

NetWorker Media database


Module
for Oracle nsrpsd nsrexecd nsrmmd
MML
PowerSnap
RMAN elements
nsrmmdbd
nsrnmostart NetWorker
Module
nsrnmo elements Secondary
storage medium

nsrsnap nsrd
interprocess communication

nsrsnapck nsrexecd savegrp

savefs
nsroraclecat Start
(optional) Here

GEN-000176

Figure 14 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.

158 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Figure 15 on page 159 shows how the Oracle Server and NetWorker processes
interact during a scheduled deferred live backup.

Oracle Server PowerSnap proxy client NetWorker


(NetWorker client) (NetWorker storage node) server
Primary storage

Oracle Point-in-time copy


data files of Oracle data files

Client file index

tracking
information
Start
PowerSnap nsrindexd
Here
interprocess processes
communication data
Media database
nsrsnap (at end
nsrexecd nsrmmd
of savegrp) or
nsrsnapadmin

PowerSnap nsrmmdbd
elements
Secondary
storage medium
GEN-000177

Figure 15 Deferred live backup

Proxy restore processes


A proxy restore can be performed in either of two ways:
◆ By issuing the appropriate commands through the RMAN command line
interface.
Or
◆ By using the Oracle Enterprise Manager Backup Management Tools, which
include an optional graphical user interface to the RMAN utility.
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 160
◆ “Restore from secondary storage” on page 161

Proxy restore processes 159


Proxy Backups and Restores

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.
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 NMO.
– 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, NMO 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 16 on page 161 shows how the Oracle Server and NetWorker processes
interact during an instant restore or rollback.

160 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Oracle Server PowerSnap proxy client NetWorker


(NetWorker client) (NetWorker storage node) server
Primary storage
Oracle Point-in-time copy
data files of Oracle data files

data Client file index


(for details on
data flow, see
description)

RMAN
tracking
information
Oracle PowerSnap PowerSnap nsrindexd
process module module

NetWorker
Module
for Oracle Media database
nsrpsd nsrexecd nsrmmd
MML

PowerSnap
nsrexecd elements

NetWorker nsrmmdbd
Module Snapshot
element pool media
nsrd
interprocess communication
GEN-000178

Figure 16 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.

Proxy restore processes 161


Proxy Backups and Restores

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
NMO. The processes use the nsrmmd and nsrmmdbd processes to determine
which media contain the requested save sets and to read the backup volumes.
4. If operating system files (not raw devices or volumes) are restored, NMO 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 17 on page 162 shows how the Oracle Server and NetWorker processes
interact during a restore from secondary storage.

Oracle Server PowerSnap proxy client NetWorker


(NetWorker client) (NetWorker storage node) server
Primary storage
Oracle
data files

Client file index

data
RMAN

tracking information
Oracle PowerSnap nsrindexd
process module

NetWorker
Module
for Oracle nsrpsd Media database
nsrmmd
MML

PowerSnap
nsrexecd elements

NetWorker nsrmmdbd
Module Secondary
element storage

nsrd
interprocess communication
GEN-000179

Figure 17 Restore from secondary storage

162 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Software configuration roadmap


This section describes the configuration requirements for proxy backups and restores.
Before performing the required configuration procedures, ensure that both the NMO
and required PowerSnap Module software are installed according to the instructions
in the following:
◆ NetWorker Module for Oracle installation guide
◆ NetWorker PowerSnap Module documentation
(Refer to the PowerSnap Module version for the primary storage system.)

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 62. The NetWorker Server, Client,
Device, and other resources must be configured:
• For live Oracle backups, a Device resource must be configured for each
secondary storage device, such as a tape drive, to be used for the backups. The
devices must be mounted prior to the backups.
• For instant Oracle backups, a snapshot Pool resource must be configured, as
described in “Additional configurations for proxy operations” on page 163.
◆ The required scheduled backup configurations are performed according to
“Configuring a backup with the NMC legacy method” on page 75.

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 164
◆ “Configuring the required Oracle settings” on page 164
◆ “Configuring the NWORA resource file” on page 164
◆ “Creating RMAN scripts for proxy backups” on page 165
◆ “Configuring the NetWorker Pool resources” on page 169
◆ “Configuring the NetWorker Snapshot Policy resource” on page 169
◆ “Configuring the NetWorker Group resource” on page 169
◆ “Configuring the NetWorker Client resource” on page 170
◆ “Testing a scheduled proxy backup” on page 170
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 171
◆ “Proxy backup information in the NetWorker indexes” on page 174
◆ “Proxy restore requirements” on page 177

Software configuration roadmap 163


Proxy Backups and Restores

Configuring internationalization (I18N) support


In a non-English environment, NMO supports internationalization (I18N) of proxy
backups and restores with a supported release of the PowerSnap Module, as
described in “Internationalization (I18N) and localization (L10N)” on page 33.
To configure I18N support for proxy backups, follow the instructions in
“Configuring I18N support” on page 67.

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.
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, specify the following values in
the Registry and Initialization Parameter file:
◆ In the Registry, specify the following parameters under
HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE:
• ORA_oracle_sid_WORKINGSETMAX
• ORA_oracle_sid_WORKINGSETMIN
Possible values to set for these parameters are as follows:
• ORA_oracle_sid_WORKINGSETMAX = 1600
• ORA_oracle_sid_WORKINGSETMIN = 1200
More information on these parameters and Oracle memory management on
Windows is available in the Oracle document number 46001.1, “Oracle Database
and the Windows NT Memory Architecture, Technical Bulletin.”
◆ In the Initialization Parameter file (such as initoracle_sid.ora), increase the value of
LARGE_POOL_SIZE to a large value that is appropriate for the particular system.

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 183. 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 181.
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.

164 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Creating RMAN scripts for proxy backups


The basic information about RMAN backup scripts in “Create RMAN scripts for
backups” on page 75 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 166.
◆ 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.
◆ Allocate only one channel in the RMAN script. Do not 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 167 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 only
format ’FULL_%d_%U’
(database);
release channel t1;
}
NSR_PROXY_PFILE is an optional NMO parameter used for proxy backups. “Setting
the parameters” on page 166 provides details.

Multiple channels in RMAN scripts


The 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 26 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)’;

Software configuration roadmap 165


Proxy Backups and Restores

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 parameters


Two types of parameters can be set for the proxy backup and restore operations:
◆ NMO 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 200.
◆ PowerSnap Module parameters, as described in “PowerSnap parameter settings”
on page 166.

PowerSnap parameter settings


The 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 211 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 206.

166 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

The configuration file consists of a separate line such as the following for each
parameter 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 27 on page 168 and Example 28 on page 168 provide examples of
PowerSnap parameter settings.
Table 7 on page 167 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.

Table 7 PowerSnap parameters (page 1 of 2)

Parameter Description Default and valid values

NSR_DATA_MOVER Mandatory for a proxy backup that uses a proxy • Local host (default).
client host. Specifies the hostname of the proxy • The valid hostname of the proxy client host.
client host.

NSR_PS_SAVE_PARALLELISM Optional. Specifies the number of concurrent save • 16 (default).


streams on the proxy client host. • An integer value less than or equal to the
Parallelism attribute value in the NetWorker Client
resource.

Software configuration roadmap 167


Proxy Backups and Restores

Table 7 PowerSnap parameters (page 2 of 2)

Parameter Description Default and valid values

NSR_MAX_STREAMS Optional. Specifies the maximum number of • 16 (default).


restore streams. • An integer value.

RESTORE_TYPE_ORDER Optional. Specifies the type of proxy restore to be • pit:conventional (default).


performed. • One or more of the following values, each value
delimited from the others by a colon(:):
Note: If multiple values are specified, each type of • pit — Specifies an instant restore.
restore is attempted (n the order specified) until a
• conventional — Specifies a proxy restore from
restore operation is successful. secondary storage media.
• rollback — Specifies a rollback restore from a
point-in-time proxy copy.
“Proxy restore requirements” on page 177 provides
more information.

Example 27 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 NMO 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 28 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 NMO 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.
◆ 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.

168 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

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=9
NSR_DEBUG_LEVEL=9
NSR_DATA_MOVER=datamover.emc.com
NSR_SNAP_NAS_CEL_CS_HOST=controlstn
NAS_SNAP_SUBTYPE=CEL_SNAPSURE
NSR_SNAP_TYPE=nas
NSR_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=9
NSR_DEBUG_LEVEL=9
NSR_DATA_MOVER=datamover.emc.com
NSR_SNAP_NAS_CEL_CS_HOST=controlstn
NAS_SNAP_SUBTYPE=CEL_SNAPSURE
NSR_SNAP_TYPE=nas
NSR_SNAP_NAS_CLIENT=11.222.333.44
RESTORE_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 NMO 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 169.

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.

Software configuration roadmap 169


Proxy Backups and Restores

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 87.
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 resource
To 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 169 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.

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 110.
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

170 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

backups” on page 79. 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 requirements


Review the following information, specific to proxy backups:
◆ “Checking configuration consistency” on page 171
◆ “Performing proxy backups” on page 172
“Proxy backups and restores on cluster systems” on page 194 provides information
on proxy backups in a cluster environment.

Checking configuration consistency


During a scheduled backup, NMO checks for consistency between the NetWorker
Group resource configuration and the RMAN backup session.
If NMO 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 171
◆ “With a group configured for regular backups” on page 172

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.
If RMAN performs only regular Oracle backups due to one of these conditions, NMO
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.
The current EMC compatibility guides provide details on the primary storage devices
supported for proxy backups with the PowerSnap Modules.
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

Proxy backup requirements 171


Proxy Backups and Restores

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 29 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.

172 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Directory for temporary files


NMO creates temporary files for processing purposes in the following directory:
◆ On UNIX, the directory is /nsr/tmp.
◆ On Microsoft Windows, the directory is NetWorker_install_path\tmp, where
NetWorker_install_path is the root directory of the NetWorker installation path.

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 153 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 171 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.

NWORA resource file backup


If a scheduled proxy backup completes successfully, NMO automatically backs up
the NWORA resource file, as described in “The NWORA resource file” on page 183.
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.

Proxy backup requirements 173


Proxy Backups and Restores

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 176 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 administration 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 107
◆ “Cancel a scheduled backup” on page 111

Proxy backup information in the NetWorker indexes


The NetWorker server maintains information about each backup in its online
indexes. “NetWorker software” on page 47 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
The NetWorker command reference guide and the UNIX man pages provide more
information on these NetWorker commands.

174 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

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 30 on page 175.
◆ The second entry is generated for the point-in-time metadata, such as
/brcmeta.1/ in Example 30 on page 175. 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 30 on page 175. This entry is created for a live backup only.

Example 30 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 2007
Physical 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.
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 31 on page 176.

Proxy backup information in the NetWorker indexes 175


Proxy Backups and Restores

Example 31 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 size
nmo.002 marmaris 05/10/07 13:18:39 102 MB
snap.001 marmaris 05/10/07 13:18:41 2 KB
ssid fl lvl name
4064690015 cb full /space1/home/oracle/bp1
4098244417 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 32 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 33 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 size
nmo.002 marmaris 05/10/07 13:18:39 4 KB
ssid fl lvl name
3863367791 cb full /nsr/res/nwora.res
The NetWorker command reference guide and the UNIX man pages provide more
information on these NetWorker commands.

176 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Proxy restore requirements


Chapter 4, “Data Restore and Recovery,” provides basic information on how to
configure and run Oracle data restore and recovery by using NMO.
Review the following information, specific to proxy restores:
◆ “Creating RMAN scripts for proxy restores” on page 177
◆ “Performing proxy restores” on page 178
◆ “Relocating files during proxy restores” on page 179
◆ “Restoring to a different host” on page 180
◆ “Point-in-time recoveries without a Recovery Catalog” on page 181
“Proxy backups and restores on cluster systems” on page 194 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 166.

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 NMO.
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 34 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 178 provides more information on setting up a rollback
operation.

Proxy restore requirements 177


Proxy Backups and Restores

! IMPORTANT
For the RESTORE_TYPE_ORDER parameter, NMO 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 180 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 179.
◆ 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 179.
◆ 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 167 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, NMO automatically cleans this subdirectory.

178 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Rollback restore
For a rollback restore, the psrollback.res file must contain the directory name
.nworapc. The file is located as follows:
◆ On UNIX: /nsr/res/psrollback.res
◆ On Microsoft Windows: NetWorker_install_path\res\psrollback.res, where
NetWorker_install_path is the root directory of the NetWorker installation path
Add the directory name to the file by using a text editor as either the root user on
UNIX or a member of the Microsoft Windows Administrators group.
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 devices


NMO 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 NMO 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 appropriate file:
◆ /etc/vfstab on Solaris
◆ /etc/fstab on HP-UX
◆ /etc/filesystems on IBM AIX
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 NMO.

! 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, NMO 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.

Proxy restore requirements 179


Proxy Backups and Restores

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.

Example 35 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 36 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 126. Task 2 of that section explains that NMO
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.

180 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

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 backups


During Oracle backups, RMAN stores information about each backup piece in the
RMAN repository, also known as the "RMAN catalog". Similarly, NMO 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 NMO 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 NMO, 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.

Catalog synchronization for proxy backups 181


Proxy Backups and Restores

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.

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 195
- “Proxy backups from a physical cluster client” on page 196

◆ 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 183 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 79 provides details.
- An NWORA resource file must include the required resources, as described in
“The NWORA resource file” on page 183.

The NMO 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.

182 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

The following sections provide complete information on how to configure and


perform catalog synchronization:
◆ “The NWORA resource file” on page 183
◆ “Automatic catalog synchronization with the nsroraclecat program” on page 191

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
◆ On Microsoft Windows: NetWorker_install_path\res\nwora.res, where
NetWorker_install_path is the root directory of the NetWorker installation path
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 either the root user on UNIX or a member of the Microsoft Windows
Administrators group.

“Configuring the NWORA resource file with the nsrnmoadmin program” on


page 188 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 184
◆ “NWORA SID resources” on page 186

Catalog synchronization for proxy backups 183


Proxy Backups and Restores

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 184
◆ “NSR_ORACLECAT_DEBUG_FILE” on page 184
◆ “NSR_ORACLECAT_LOG_FILE” on page 184
◆ “NSR_ORACLECAT_MODE” on page 185
◆ “NSR_REMOVE_ON_FAILURE” on page 185

Note: The parameter resources listed in Table 8 on page 184 are the only ones supported. Do
not attempt to add other parameter resources to the NWORA resource file.

Table 8 NWORA parameter resources (page 1 of 2)

Parameter resource Description Default and valid values

NSR_NWPATH Specifies the directory location of the NetWorker binary • Directory pathname for the location of
nsrsnapck. nsrsnapck (default).
• Valid directory pathname for the location of
Note: If you use NMO with Sun-branded NetWorker, you must the NetWorker binary nsrsnapck.
set NSR_NWPATH by using the following nsrnmoadmin
command:
nsrnmoadmin -r update NSR_NWPATH=/usr/sbin/nsr

NSR_ORACLECAT_DEBUG_FILE Specifies the debug file used by the nsroraclecat program. • Undefined (default).
Set this parameter only for the purpose of debugging the • Valid pathname of the nsroraclecat debug
nsroraclecat program. file.
Note: The nsroraclecat debug file must be created in a Note: If undefined, debug information is not
secure location since it includes a copy of the strings from the generated.
RMAN connection file.

NSR_ORACLECAT_LOG_FILE Specifies the operations log file used by the nsroraclecat • Undefined (default).
program. The logged information includes the backup pieces • Valid pathname of the nsroraclecat log file.
successfully removed from the RMAN catalog, and those that
failed to be removed during automatic catalog Note: If undefined, logging information is written
synchronization. to the /nsr/applogs/nsroraclecat.log file by
default.

184 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Table 8 NWORA parameter resources (page 2 of 2)

Parameter resource Description Default and valid values

NSR_ORACLECAT_MODE Specifies whether automatic catalog synchronization is • Undetermined (default).


enabled or disabled during proxy backups. • 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 • Undetermined (default).


non-English environment only. Specifies the non-English • Valid locale value, same as set in the
locale value, as set in the NLS_LANG environment variable. NLS_LANG environment variable.
“Configure I18N support” on page 68 provides more
information. 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 • FALSE (default).
are removed when the nsroraclecat program fails to remove • TRUE.
one or more RMAN catalog entries during automatic catalog
synchronization. “Automatic catalog synchronization with the
nsroraclecat program” on page 191 provides more
information.

Using the nsrnmoadmin command to set parameter resources


When the nsrnmoadmin command (with any options) is used for the first time after
the NMO installation, the NWORA resource file is automatically populated with five
parameter resources from Table 8 on page 184: 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.

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 188 provides details on how to use the nsrnmoadmin command.

Catalog synchronization for proxy backups 185


Proxy Backups and Restores

Example 37 Default NWORA parameter resources

After the NMO 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_path
NSR_ORACLECAT_MODE=undetermined
NSR_REMOVE_ON_FAILURE=FALSE
NSR_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 186.

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 186.
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 191 provides more information.

Note: Each NWORA SID resource must have a unique NSR_ORACLE_SID value.

Table 9 NWORA SID resource components (page 1 of 2)

Parameter Description Default and valid values

NSR_ORACLE_CONNECT_FILE Mandatory. Specifies the location of the file containing the • Undefined (default).
connection strings required to create an RMAN session. “The • Valid pathname of the RMAN connection file.
connection file” on page 188 provides more information.

NSR_ORACLE_HOME Mandatory. Specifies the home directory of the Oracle • Undefined (default).
installation. The RMAN executable must be located in • Valid pathname of the Oracle home directory.
subdirectory bin of this directory.
Note: The value must be equal to the Oracle
parameter $ORACLE_HOME value.

186 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Table 9 NWORA SID resource components (page 2 of 2)

Parameter Description Default and valid values

NSR_ORACLE_LIB_PATH Optional. Specifies the pathname of the directory containing • Undefined (default).
the Oracle shared libraries on UNIX, typically • Valid pathname of the Oracle shared library
$ORACLE_HOME/lib. directory on UNIX.

Note: This parameter is not required on


Windows.

NSR_ORACLE_SID Mandatory. Specifies the SID value of the Oracle database • Undefined (default).
whose RMAN catalog is to be synchronized. • 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 83 provides more
information.

NSR_ORACLE_TNS_ADMIN Optional. Specifies the pathname of the directory containing • Undefined (default).
the Oracle Net configuration files. • Valid pathname of Oracle network
configuration directory.

Note: The value must be equal to the Oracle


parameter $TNS_ADMIN value.

Using the nsrnmoadmin command to set SID resources


To 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 188 provides details on how to use the nsrnmoadmin command.

Example 38 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.
◆ 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.file
NSR_ORACLE_HOME=/dbapps/proddb/app/oracle/product/10.2.0/Db_1
NSR_ORACLE_LIB_PATH=/usr/lib
NSR_ORACLE_SID=proddb
NSR_ORACLE_TNS_ADMIN=/dbapps/proddb/tns

Catalog synchronization for proxy backups 187


Proxy Backups and Restores

In this sample, 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 186 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 39 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 on UNIX or as a member of the Microsoft Windows
Administrators group.
“The nsrnmoadmin command syntax and options related to proxy backups” on
page 189 provides details on the command syntax and options.

188 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Windows 2008 and Windows Vista requirements for the nsrnmoadmin


command
Note: Whenever you make changes to the NWORA resource file on Windows 2008 or
Windows Vista, back up the file manually.

On Windows 2008 and Windows Vista, you must run the nsrnmoadmin command in
the Command Prompt window as an administrator:
1. Click Start.
2. Right-click Command Prompt.
3. Select Run as administrator.
4. Run the nsrnmoadmin command in the open Command Prompt window.

The nsrnmoadmin command syntax and options related to proxy backups


The 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.

Catalog synchronization for proxy backups 189


Proxy Backups and Restores

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 190
◆ “Add the NWORA resources” on page 190
◆ “Update the NWORA resources” on page 191
◆ “Delete the NWORA SID resources” on page 191
The following sources provide more information on the nsrnmoadmin command:
◆ The nsrnmoadmin man page on a UNIX Oracle Server that contains the NMO
software.
◆ The nsrnmoadmin entry in the NetWorker Module for Oracle command
reference guide on the Powerlink website.
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 38 on page 187, 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.

190 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

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
NMO 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 210. 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 191
◆ “Failure of the nsroraclecat program” on page 192
◆ “NetWorker index entry removals with nsrsnapck” on page 193

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.

Catalog synchronization for proxy backups 191


Proxy Backups and Restores

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:
◆ On UNIX:
/nsr/tmp/.nworapc/nsroraclecat_date_pid
◆ On Microsoft Windows:
NetWorker_install_path\tmp\.nworapc\nsroraclecat_date_pid
where:
• NetWorker_install_path is the root directory of the NetWorker installation path.
• date is the current date.
• pid is the nsroraclecat process ID.
The 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 184
provide more information on these files.

The following sources provide more information on the nsroraclecat program:


◆ The nsroraclecat man page on a UNIX Oracle Server that contains the NMO
software.
◆ The nsroraclecat entry in the NetWorker Module for Oracle command reference
guide on the Powerlink website.

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 193.

192 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

The following files (if they exist) need to be removed:


◆ Files in one of these directories:
• On UNIX: /nsr/tmp/.nworapc
• On Microsoft Windows: NetWorker_install_path\tmp\.nworapc, where
NetWorker_install_path is the root directory of the NetWorker installation path
◆ Files in either the temporary directory /tmp on UNIX or the temporary directory
specified by the TEMP system variable on Microsoft Windows, 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.
- 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.

Catalog synchronization for proxy backups 193


Proxy Backups and Restores

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 following file:
• On UNIX: /nsr/logs/daemon.raw
• On Microsoft Windows: NetWorker_install_path\logs\daemon.raw, where
NetWorker_install_path is the root directory of the NetWorker installation path
The NetWorker administration guide provides more information on the
daemon.raw log file and how to view its contents.

Proxy backups and restores on cluster systems


NMO 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 194
◆ “Proxy backups from a virtual cluster client” on page 195
◆ “Proxy backups from a physical cluster client” on page 196
◆ “Restores from proxy backups on a cluster system” on page 197

Proxy backup failover


During a scheduled proxy backup where the Oracle software is configured to fail
over (for example, by using Oracle Fail Save with MSCS on Microsoft Windows), 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 37 provides more information.

194 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

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 NMO 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 170:
• 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 163:
• 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 administration
guide.
4. Configure the NWORA resource file on each node of the cluster, as described in
“Configuring the NWORA resource file” on page 164.
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 follows:
• On UNIX: NetWorker_install_path/bin/pathownerignore
• On Microsoft Windows: 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 182 provides details on the expiration and removal of backups specified
with nsrnmo -c client_name.

Proxy backups and restores on cluster systems 195


Proxy Backups and Restores

Notes:
• The host specified with -c client_name must have access to instant backups.
• NMO 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 40 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.emc.com, where the script name is nsrnmo, specify the following in the Backup
Command attribute:
nsrnmo -c mars.emc.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 171
◆ NetWorker administration 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 182
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.
◆ NMO and the PowerSnap Module must be installed and configured on the host
specified with -c client_name.

196 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Proxy Backups and Restores

Example 41 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.emc.com, where the script name is nsrnmo, specify the following in the
Backup Command attribute:
nsrnmo -c monalisa.emc.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 200:
• 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.

Proxy backups and restores on cluster systems 197


Proxy Backups and Restores

198 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
A
Parameters in an RMAN
Session

This appendix includes the following sections:


◆ Setting the NSR* parameters ..................................................................................... 200
◆ NSR* parameter definitions....................................................................................... 201

Parameters in an RMAN Session 199


Parameters in an RMAN Session

Setting the NSR* parameters


This appendix describes the NSR* parameters that can be set in the RMAN command
or script for NetWorker Module for Oracle (NMO) backups and restores.

Note: Unless noted otherwise, the parameters are supported for both regular and proxy
backups and restores. “Setting the parameters” on page 166 provides information specific to
proxy operations.

The NSR* parameters described in Table 10 on page 201 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 NMO backup or
restore, set them in one of the following ways unless specified otherwise in Table 10
on page 201:
◆ If using automatic channels, set the parameters with the parms option in the
configure channel command.
“Automatic channel allocation” on page 26 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 UNIX setenv command or Windows set command on the
operating system command line to set the NSR* parameters has no effect.
On Windows, when an NSR* parameter is set through the parms option, the value of
that NSR* parameter remains in effect for all subsequent allocated channels, and for
all subsequent RMAN sessions until one of the following occurs:
◆ The Oracle database is shut down.
◆ The NSR* parameter is unset for the channel by using the parms option, as in the
following example:
run {
allocate channel t1 type ’SBT_TAPE’
parms ’ENV=(NSR_SERVER=,NSR_DATA_VOLUME_POOL=)’;
:
:
release channel t1;
}

Note: On Windows, this does not occur if the parameters are set through the send command in
all RMAN sessions.

200 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Parameters in an RMAN Session

NSR* parameter definitions


Table 10 on page 201 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)

Parameter Description Default and valid values

NSR_AES_ENCRYPTION Optional. Specifies whether the Oracle data is • FALSE (default) = The Oracle data is not encrypted
encrypted by NMO by using 256-bit AES encryption through 256-bit AES encryption during the backup.
during the backup before being sent over the network. • TRUE = The Oracle data is encrypted through
The AES encryption uses the key that is set in the 256-bit AES encryption during the backup.
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 NMO backup, the
encrypted data cannot be recovered. The NetWorker
administration guide provides more information on pass
phrases.

NSR_CHECKSUM Optional. Specifies whether or not the NetWorker • FALSE (default) = The NetWorker software
software performs checksumming on the backup data. performs no checksumming.
• TRUE = The NetWorker software performs
checksumming.

NSR_CLIENT Recommended for a backup of Oracle running on a • The physical host on which the session runs
cluster or RAC system. (default).
Recommended for all restores. • A valid NetWorker client hostname.
Specifies the NetWorker Client resource to use for a
backup or restore session.
Might be needed for disaster recovery.

NSR_COMPRESSION Optional. Specifies whether or not the NetWorker • FALSE (default) = The NetWorker software
software performs compression on the backup data. performs no compression.
• TRUE = The NetWorker software performs
compression.

NSR* parameter definitions 201


Parameters in an RMAN Session

Table 10 Parameters in the RMAN command or script (page 2 of 7)

Parameter Description Default and valid values

NSR_DATA_VOLUME_POOL Mandatory if using the set duplex command (with • The most appropriate pool, selected by the
duplex set to 1, 2, 3, or 4) or other RMAN commands to NetWorker server (default).
generate backup copies during a manual backup. • A valid NetWorker pool name.
Specifies the name of the volume pool to use for the • The name must be different from the name used by
Oracle backup, including a duplexed Oracle backup. the parameter NSR_DATA_VOLUME_POOL1,
In the case of proxy backups, specifies the volume pool NSR_DATA_VOLUME_POOL2, or
for live backups only (backups to secondary storage NSR_DATA_VOLUME_POOL3.
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.

NSR_DATA_VOLUME_POOL1 Mandatory if using the set duplex command (with • Undefined (default).
duplex set to 2, 3, or 4) or other RMAN commands to • A valid NetWorker pool name.
generate two or more backup copies during a manual • The name must be different from the name used by
backup. the parameter NSR_DATA_VOLUME_POOL,
Specifies the name of the volume pool to use for a NSR_DATA_VOLUME_POOL2, or
duplexed Oracle backup. NSR_DATA_VOLUME_POOL3.
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.

NSR_DATA_VOLUME_POOL2 Mandatory if using the set duplex command (with • Undefined (default).
duplex set to 3 or 4) or other RMAN commands to • A valid NetWorker pool name.
generate three or more backup copies during a manual • The name must be different from the name used by
backup. the parameter NSR_DATA_VOLUME_POOL,
Specifies the name of the volume pool to use for a NSR_DATA_VOLUME_POOL1, or
duplexed Oracle backup. NSR_DATA_VOLUME_POOL3.
Note: For a manual backup that generates backup
copies, NSR_DATA_VOLUME_POOL2 must be set
through the parms option, not through the send
command or option.

NSR_DATA_VOLUME_POOL3 Mandatory if using the set duplex command (with • Undefined (default).
duplex set to 4) or other RMAN commands to generate • A valid NetWorker pool name.
four backup copies during a manual backup. • The name must be different from the name used by
Specifies the name of the volume pool to use for a the parameter NSR_DATA_VOLUME_POOL,
duplexed Oracle backup. NSR_DATA_VOLUME_POOL1, or
NSR_DATA_VOLUME_POOL2.
Note: For a manual backup that generates backup
copies, NSR_DATA_VOLUME_POOL3 must be set
through the parms option, not through the send
command or option.

202 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Parameters in an RMAN Session

Table 10 Parameters in the RMAN command or script (page 3 of 7)

Parameter Description Default and valid values

NSR_DEBUG_FILE Optional. Enables debugging by specifying the • Undefined (default).


complete pathname of the log file on the Oracle Server • The valid pathname of the debug file to use on the
that receives debug information during a proxy or Oracle Server.
regular backup or restore.
Only used for debugging of the NMO media Note: If undefined, no debug information is generated.
management library (MML). If a file with the specified pathname cannot be created,
debug information is either directed to the following
Note: NSR_DEBUG_FILE must be set through the location or not generated:
parms option. Do not set it through the send option or - /nsr/applogs/nmosbt.log (UNIX)
command. - NetWorker_install_path\applogs\nmosbt.log
(Windows)

NSR_DEBUG_LEVEL Optional. Controls the amount of debug information that • 0 (default) = Only operational messages are written
NMO writes to the debug log file, which is specified by for the sbtread(), sbtwrite(), and sbtpcstatus()
NSR_DEBUG_FILE. functions. No entry or exit messages are written for
the functions.
Note: NSR_DEBUG_LEVEL must be set through the • 1 = Only entry and exit messages are written for the
parms option. Do not set it through the send option or sbtread(), sbtwrite(), and sbtpcstatus() functions. No
command. operational messages are written for the functions.
• 2 = All debug messages for all the functions are
The debug level limits only the amount of debug written to the log file.
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_DEDUP_BACKUP Mandatory for a manual deduplication backup. Do not • FALSE (default) = Deduplication is not performed
set for a scheduled deduplication backup. Specifies during a manual NMO backup. Other NSR_DEDUP*
whether or not deduplication is performed during a parameters are ignored.
manual NMO backup. • TRUE = Deduplication is performed during a manual
NMO backup. If NSR_DEDUP_NODE is not set, the
Note: For a scheduled deduplication backup, set the backup fails.
De-duplication Backup attribute in the NetWorker Client
resource, instead of setting this parameter.

NSR_DEDUP_CACHE_ENABLED Optional for a deduplication backup. Specifies whether • TRUE (default) = A hash cache is used to increase
a hash cache is used during a deduplication backup. performance during a deduplication backup. This
The nsravtar process creates the cache in the value is recommended in most cases.
/nsr/dedup/cache or Windows equivalent directory. Use • FALSE = A hash cache is not used during a
of the cache increases both the deduplication backup deduplication backup. The parameter
performance and disk usage in the cache directory. NSR_DEDUP_CACHE_TAG is ignored.

Note: Setting of this parameter requires knowledge of


the potential effects on Avamar server operations.

NSR* parameter definitions 203


Parameters in an RMAN Session

Table 10 Parameters in the RMAN command or script (page 4 of 7)

Parameter Description Default and valid values

NSR_DEDUP_CACHE_TAG Mandatory for a deduplication backup if the parameter • / (default; signifies the root directory).
NSR_DEDUP_CACHE_ENABLED is set to TRUE. • String value of the tag to use to generate a
Specifies the tag for the nsravtar process to use to deduplication cache name. Do not include the client
generate the hash cache name for a deduplication name in the value.
backup: • Recommended value is:
• Set to a different value for each channel; if the same <ORACLE_SID or Net_service_name>_<channel_ID>
tag value is used for more than one channel, the For example: ORCL102_t1
deduplication backup fails.
• Set with the send command (not the parms option) Note: The nsravtar process uses the tag value to
if multiple channels are used on Windows. generate the cache name through hashing. The actual
• Set with the parms option (not the send command) cache name does not contain this parameter value.
if automatic channel allocation is used (and multiple
channels are not used on Windows).

NSR_DEDUP_CHUNK_SIZE Optional for a deduplication backup. Specifies the size • 0 (default; signifies that variable sizes are used for
in bytes that the Avamar server uses for data chunks in the data chunks, as determined by the Avamar
a deduplication backup. server). This value is recommended in most cases.
If a nonzero value is specified, the Avamar server uses • Size (greater than zero) in bytes to use for all of the
the fixed size for all of the data chunks saved in the data chunks in a deduplication backup; for example,
deduplication backup. 1024, 2048, 5096, 8194, or a value recommended in
the Avamar documentarion.
Note: Setting of this parameter requires knowledge of
the potential effects on Avamar server operations.

NSR_DEDUP_NODE Mandatory for a manual deduplication backup. Do not • Undefined (default).


set for a scheduled deduplication backup. Specifies the • Avamar server hostname set in the NetWorker
hostname of the Avamar server to perform a manual De-duplication Node resource.
NMO 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.

NSR_DPRINTF Optional, for use with Tech Support only. Specifies • FALSE (default) = NetWorker core debug messages
whether NetWorker core debug messages are printed are not printed to NSR_DEBUG_FILE.
to the file specified by NSR_DEBUG_FILE. • TRUE = NetWorker core debug messages are
printed to NSR_DEBUG_FILE.
Note: NSR_DPRINTF must be set through the parms - If NSR_DEBUG_FILE is not set, the messages are
option, not through the send command or option. printed to the default location.
Recommended value for NSR_DEBUG_LEVEL is 9. - If NSR_DEBUG_LEVEL is set to 0, few debug
messages are printed.

204 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Parameters in an RMAN Session

Table 10 Parameters in the RMAN command or script (page 5 of 7)

Parameter Description Default and valid values

NSR_ENCRYPTION_PHRASES Optional. Specifies one or more encryption phrases to • Undefined (default).


use for decrypting Oracle data during an NMO restore. • One or more encryption phrases to use during an
If this parameter is not set, the NMO restore obtains the Oracle restore with NMO. Each phrase must be a
encryption phrase from the NetWorker server. string enclosed in quotes. Multiple phrases must be
If both of the following are true, set this parameter to the separated by commas, and the entire group of
phrase used to originally back up the data: phrases surrounded by outer quotes that are
• The Oracle data that is being restored was backed different from the inner quotes. For example:
up with 256-bit AES encryption. NSR_ENCRYPTION_PHRASES="’key1’,’key2’"
• The encryption phrase on the NetWorker server has - NMO itself supports double ("), single (’), and
changed since the Oracle data was backed up. backward (‘) quotes.
- Certain shells might not support certain types of
Note: NSR_ENCRYPTION_PHRASES must be set quotes, for example, when the parameter is set
with the send command on the command line.
through the RMAN send command only.
- Oracle does not support the use of single quotes
within an RMAN script.

Note: If undefined, the NMO restore obtains the


encryption phrase from the Datazone pass phrase
attribute of the NetWorker Server resource, as
described in “Datazone pass phrase” on page 63.

NSR_GROUP Optional. Use for a manual backup only. Specifies the • The predefined NetWorker group named Default
NetWorker group to use for a backup session. Used by (default).
the NetWorker server to select the volume pool for the • A valid NetWorker backup group name.
manual backup.
Note: A NetWorker Group resource must be configured
for the specified backup group.

NSR_MMDB_RETRY_TIME Optional. Specifies the number of minutes that NMO • 0 (default). NMO does not try to reconnect to the
should try to connect to the NetWorker media database media database if the first attempt fails.
before terminating the operation (backup, restore, or • A valid number of minutes.
RMAN maintenance commands). When the media
database is busy, NMO tries to reconnect after sleeping
for five seconds between attempts.

NSR_NO_BUSY_ERRORS Optional. Specifies whether a scheduled backup fails • FALSE (default) = The scheduled backup waits for
immediately when the NetWorker server is busy or the NetWorker server to accept the connection.
waits for the NetWorker server to accept the • TRUE = The scheduled backup fails immediately
connection. when the NetWorker server is busy.
Note: For proxy backups, this parameter is always set
to FALSE.

NSR_NO_MULTIPLEX Optional. When set for a specific RMAN channel, • FALSE (default) = Multiplexing is enabled on the
specifies whether multiplexing is disabled during a device that the RMAN channel is using.
backup on the NetWorker device that the RMAN • TRUE = Multiplexing is disabled on the device that
channel is using. If multiplexing is disabled, no other the RMAN channel is using.
save sets can be written to the device.
To optimize restore operations, RMAN requires NMO Note: If a random access NetWorker device such as an
backups to not be multiplexed. advanced file device is used, do not set the parameter
Setting the parameter to TRUE may affect the backup to TRUE.
performance. For example, the device may sit idle
during part of the backup. If the performance is
adversely affected, reset the parameter to FALSE.

NSR* parameter definitions 205


Parameters in an RMAN Session

Table 10 Parameters in the RMAN command or script (page 6 of 7)

Parameter Description Default and valid values

NSR_NWPATH Mandatory when removing a backup from the • The platform-specific default location of the
NetWorker server if both of the following are true: NetWorker client binaries (default).
• The nsrsnapck binary is in a nondefault location. • A valid pathname of the directory that contains the
• The NWORA resource file does not contain the nsrsnapck binary, for example, /usr/local/bin.
correct location of the NetWorker client binaries.
“The NWORA resource file” on page 183 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 NMO client host.
Specifies the complete pathname of the directory that
contains nsrsnapck or NetWorker binaries.

Note: If you use NMO with Sun-branded NetWorker,


you must set NSR_NWPATH by using the following
nsrnmoadmin command:
nsrnmoadmin -r update
NSR_NWPATH=/usr/sbin/nsr
“The NWORA resource file” on page 183 provides
details on using the nsrnmoadmin command.

NSR_PROXY_PFILE Supported only for a proxy backup or restore. • Undefined (default).


Mandatory if setting PowerSnap parameters in a • A valid pathname of the configuration file.
configuration file.
Specifies the complete pathname of a configuration file Note: If undefined or an invalid pathname, parameter
that contains PowerSnap parameter settings for a proxy settings in the preferred configuration file are ignored.
backup or restore.

NSR_RECOVER_POOL Optional for restore only. Specifies the name of the • Undefined (default).
NetWorker volume pool to use for an NMO restore. You • A valid name of a NetWorker volume pool that
can use this option to restore data from a specified contains a cloned backup to use for a restore.
volume pool if there are multiple copies (clones) of the
backup on different volume pools.

Note: This parameter cannot be used for proxy


restores.

NSR_RETENTION_DISABLED Optional. Specifies whether the NetWorker browse and • FALSE (default) = NetWorker browse and retention
retention policies are disabled. policies are enabled, and used to manage the
Set to TRUE to use Oracle policies only (not NetWorker lifecycle of the NMO backup data.
policies) to manage the backup data lifecycle. Then the • TRUE = NetWorker browse and retention policies
RMAN catalog and NetWorker indexes cannot become are disabled. Only Oracle policies are used to
unsynchronized, for example, when a NetWorker index manage the lifecycle of the NMO backup data.
entry is expired but the corresponding RMAN catalog
entry is not expired.

NSR_SAVESET_BROWSE Mandatory if setting a nondefault browse policy for a • The most appropriate policy for the client, as
manual Oracle backup. selected by the NetWorker server (default).
Specifies the date when the entry for the backup is to • A valid date in nsr_getdate(3) format.
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.

206 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Parameters in an RMAN Session

Table 10 Parameters in the RMAN command or script (page 7 of 7)

Parameter Description Default and valid values

NSR_SAVESET_RETENTION Mandatory if setting a nondefault retention policy for a • The most appropriate policy for the client, as
manual Oracle backup. selected by the NetWorker server (default).
Specifies the date when the save set becomes • A valid date in nsr_getdate(3) format.
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.

NSR_SERVER Mandatory for a manual backup if the NetWorker server • The local physical host (default).
host is different from the Oracle Server host. • A valid NetWorker server name.
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.

NSR_SERVER_NIC Optional. Specifies the name of a network interface • Undefined (default).


card (NIC) on a NetWorker server. • A valid name of a NetWorker server NIC.
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.

NSR* parameter definitions 207


Parameters in an RMAN Session

208 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
B

RMAN Commands

This appendix includes the following sections:


◆ The delete expired backup command ...................................................................... 210
◆ The change...crosscheck and crosscheck commands ............................................. 210
◆ The pool option of the backup command................................................................ 210
◆ The send command...................................................................................................... 211
◆ The set duplex command........................................................................................... 215
◆ The trace option of the backup command ............................................................... 217

RMAN Commands 209


RMAN Commands

The delete expired backup command


For 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 64.
If the user does not have the required NetWorker privileges, or there is an
authorization problem when the delete expired backup command runs, NetWorker
Module for Oracle (NMO) 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 UNIX man pages 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 NMO installation, the
parameter NSR_NWPATH might need to be set in the NWORA resource file or in the RMAN
script. “NSR_NWPATH” on page 206 provides more information.

The change...crosscheck and crosscheck commands


For 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
NMO 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 NMO returns the following error message:
sbtbackup: Oracle pools are not supported
“NMO error messages” on page 222 provides more information on this error
message.
To specify the NetWorker volume pool to be used by NMO, set the parameter
NSR_DATA_VOLUME_POOL in the RMAN script. Appendix A, “Parameters in an
RMAN Session,” provides more information.

210 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
RMAN Commands

The send command


The NMO 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 200. 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 211
◆ “Two ways to run the send command” on page 213
◆ “Precedence rules” on page 215

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 211
◆ “The send command options” on page 212

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 NMO 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.

The send command 211


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 213 provides more information.

Example 42 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)’;
:
}
This sample RMAN script is referenced in Table 11 on page 213.

212 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
RMAN Commands

Table 11 on page 213 lists the values for options used with the send command. The
example referred to is Example 42 on page 212.

Table 11 Option values in the send command

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 NMO parameter name. In the example, the first parameter name is
NSR_SERVER.

value1 The value assigned to the first parameter. In the example, the first value is server1.

name2 The second NMO 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.

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 213.
◆ In the run job of the RMAN script, as described in “The send command in the
RMAN script” on page 214.

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 211, 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 212 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 send command 213


RMAN Commands

◆ 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 215 provides more information.

Example 43 An rman send command sets a parameter for all channels

In the following example, the NSR_SERVER parameter value (mars.emc.com) is


applied to all three channels (t1, t2, t3) allocated in the RMAN script:
rman send ”’NSR_ENV=(NSR_SERVER=mars.emc.com)’”
(RMAN script:)
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 211.
◆ 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 212.
◆ Specify the correct option values in the send command, as described in “The send
command options” on page 212.
◆ 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.emc.com:
run {
allocate channel t1 type ’SBT_TAPE’;
allocate channel t2 type ’SBT_TAPE’;
send ’NSR_ENV=(NSR_SERVER=mars.emc.com,
NSR_DATA_VOLUME_POOL=MondayFulls)’;
backup full filesperset 4
format ’FULL_%d_%U’
(database);
release channel t1;
release channel t2;
}
This script is the same as the sample script on page 75. The single send command sets
the parameters for both channels.

214 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
RMAN Commands

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.

! 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 44 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 command


For 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: NMO supports the generation of backup copies for manual backups only, not for
scheduled backups.

The set duplex command 215


RMAN Commands

Table 12 on page 216 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=2 Two copies of the backup set are directed to the separate pools specified by
or NSR_DATA_VOLUME_POOL and NSR_DATA_VOLUME_POOL1.
set duplex=on These two pools must be different.

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.

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 27 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
NMO, 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 38 provides a definition of expired) and
to use the duplexed copy for the restore operation.

216 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
RMAN Commands

The appropriate Oracle backup and recovery documentation provides more


information.

The trace option of the backup command


Set the trace option of the RMAN backup command to the value 0, 1, or 2. The
default value of trace is 0.
The output of trace is written to the Oracle sbtio.log file. The output is also written to
the following default log file on the Oracle Server host:
◆ On UNIX: /nsr/applogs/nmo.messages.raw
◆ On Windows: NetWorker_install_path\applogs\nmo.messages.raw, where
NetWorker_install_path is the root directory of the NetWorker installation path
These log files do not contain Oracle Server or RMAN errors.
NMO 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 NMO
releases.
The NetWorker administration guide provides information on how to use the
nsr_render_log program to read any language-independent binary file, such as
nmo.messages.raw.
Table 13 on page 217 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 NMO implementation of
the Oracle SBT interface).

2 • All error conditions.


• Entry and exit for each SBT function (the NMO implementation of the Oracle SBT interface).
• Values of all function parameters.
• First 32 bytes of each read/write buffer.

The trace option of the backup command 217


RMAN Commands

218 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
C
Troubleshooting and
Error Messages

This appendix includes the following sections:


◆ Troubleshooting tips ................................................................................................... 220
◆ RMAN error messages ............................................................................................... 221
◆ NMO error messages .................................................................................................. 222

Troubleshooting and Error Messages 219


Troubleshooting and Error Messages

Troubleshooting tips
The following list of troubleshooting tips refers to sections of this administration
guide and the NetWorker Module for Oracle installation guide. Use the following list
to troubleshoot any problems in running Oracle backup and restore operations with
NetWorker Module for Oracle (NMO).
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 NetWorker Module for Oracle installation guide 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 63 provides more information.
Without NMO 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 installation guide
• NetWorker administration guide
• Chapter 2, “Software Configuration”
4. Install and enable the NMO software according to the instructions in the
NetWorker Module for Oracle installation guide.
To relink the library file on UNIX, use the linking commands listed in the
installation guide. Use the correct libnwora.* library file.
To verify the version of NMO 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 filenm
strings filenm | grep “@(#)”
pkginfo -l LGTOnmo (Solaris only)
• On Windows:
a. Find the nsrnmostart.exe or libwora.dll file.
b. Right-click the file icon and select Properties.
c. In the Properties window, click the Version tab and view the file version
information.
Compare the library file with the libnwora.* file in the NMO software package;
the two files should be identical. Ensure that Oracle is not linked to another
libnwora.*, libobk.* (NetWorker BusinesSuite Module for Oracle version of
libnwora.*), or liblsm.* (from Legato Storage Manager or Legato Single Server
Version) file.

220 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

5. Perform a manual Oracle backup by using NMO 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 75 provides a simple
startup RMAN script.
If the manual backup fails, check the debug files for NMO and the NetWorker
server. For NMO debug information, set the parameters NSR_DEBUG_FILE and
NSR_DEBUG_LEVEL. “NMO error messages” on page 222 provides more
information.
If the backup fails with the following error, ensure that both NMO 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 Library
Additional information: 25
6. Perform a scheduled Oracle backup by using NMO 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 77. Customize the nsrnmo script, as described in “Customize the nsrnmo
script” on page 79.
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 221 provides more information.

RMAN error messages


RMAN 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 82 provides more information.
The appropriate Oracle error messages guide provides more information on specific
RMAN error messages and recommended courses of action.

Note: During a backup on AIX or Windows, if an NMO parameter is set to an invalid value, the
resulting error message might be truncated in the RMAN output. This is due to an Oracle
RMAN limitation.

RMAN error messages 221


Troubleshooting and Error Messages

NMO error messages


During a backup or restore, the NMO software records the NMO error messages in
both the Oracle sbtio.log file and the following default log file on the Oracle Server
host:
◆ On UNIX: /nsr/applogs/nmo.messages.raw
◆ On Windows: NetWorker_install_path\applogs\nmo.messages.raw, where
NetWorker_install_path is the root directory of the NetWorker installation path
These log files do not contain Oracle Server or RMAN errors.
NMO 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 NMO
releases.
The NetWorker administration guide provides information on how to use the
nsr_render_log program to read any language-independent binary file, such as
nmo.messages.raw.
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 as follows:
◆ On UNIX, in the /nsr/applogs/nmosbt.log file.
Otherwise, the information is not written.
◆ On Windows, in the NetWorker_install_path\applogs\nmosbt.log file, where
NetWorker_install_path is the root directory of the NetWorker installation path.
Otherwise, the information is not written.
“NSR_DEBUG_FILE” on page 203 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 203 provides more
information.
During a scheduled backup, the NMO 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 as follows:
◆ On UNIX, in the /nsr/applogs/nsrnmostart.log file.
Otherwise, the information is not written.
◆ On Windows, in the NetWorker_install_path\applogs\nsrnmostart.log file, where
NetWorker_install_path is the root directory of the NetWorker installation path.
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 83 provides more
information.

222 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

The NetWorker server also writes diagnostic information from the manual or
scheduled Oracle backup to specific log files on the NetWorker server. The
NetWorker administration guide provides more information on these log files.
NMO 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 223
◆ “Error messages from the nsrnmoadmin program” on page 231
◆ “Error messages from the nsrnmoinfo program” on page 233
◆ “Error messages from the nsrnmoprobe program” on page 235
◆ “Error messages from the nsrnmostart program” on page 236

Error messages from the libnwora library


Table 14 on page 223 lists error messages generated by the libnwora library, in
alphabetical order.

Note: The library name libnwora applies to UNIX. On Windows, the corresponding library is
named orasbt.dll.

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 NMO 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 9)

Error message Description Resolution

A connection to NW server 'server' could not NMO could not connect to the NetWorker client Take the corrective action suggested by the error
be established because 'reason'. file index due to the given reason. The client message.
might not be configured as a client on the server.

Attempted to restore file 'filename' to raw A proxy restore of a regular file to a raw device Do not attempt to restore a regular file to a raw
device 'device_name'. was attempted. This type of restore is not device.
supported.

Attempted to restore raw device A proxy restore of a raw device file to a regular file Do not attempt to restore a raw device file to a
'device_name' to file 'filename'. was attempted. This type of restore is not regular file.
supported.

Cannot back up object object_name with The RMAN backup command included the proxy When the backup command includes the proxy
proxy copy. only option, but the object object_name did not only option, ensure that the object object_name
reside on a primary storage device that the resides on a primary storage device that the
PowerSnap Module supports. PowerSnap Module supports.

Could not create the LNM index lock file NMO failed to create the lock file required for an Report the error number (errno) to Technical
'filename' (errno) index deletion operation. Support.

NMO error messages 223


Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 2 of 9)

Error message Description Resolution

Could not decode the 'sf_check' value: xdrs = This is an internal XDR error caused by a network Report the error to Technical Support.
0xvalue read or write operation.

Could not decode the 'sf_magic' value: xdrs = This is an internal XDR error caused by a network Report the error to Technical Support.
0x%value read or write operation.

Could not decode the 'sf_more' flag: xdrs = This is an internal XDR error caused by a network Report the error to Technical Support.
0xvalue read or write operation.

Could not find the nsrsnapck binary. During an index removal for a proxy backup, NMO Ensure that the parameter NSR_NWPATH is set
could not locate the nsrsnapck binary, which is correctly.
probably in a nondefault location.

Could not locate the LNM save file NMO could not locate an index record for the Use the mminfo and nsrinfo commands to verify
'backup_piece_name' on server 'server'. backup piece. The index record is probably the status of the index record.
missing.

Could not locate the LNM save time NMO could not locate an index record for the save Use the mminfo and nsrinfo commands to verify
'save_time' on server 'server'. time in the client file index. The index record is the status of the index record.
probably missing.

Could not lock 'filename' for index deletion. NMO was able to create the lock file required for Report the error number (errno) to Technical
There were number attempts. (errno) an index deletion operation, but could not lock the Support.
file after the given number of attempts.

Could not lstat - filename The lstat() system call failed. The file filename Ensure that the file is an existing file with valid
either did not exist or had invalid permissions. permissions.

Could not lstat secondary link - filename The lstat() system call failed. The file filename Ensure that the symbolic link points to an existing
was a symbolic link that pointed to a file that either file with valid permissions.
did not exist or had invalid permissions.

Could not obtain NSR_ORACLECAT_MODE The error was caused by one of the following Based on the condition, perform one of the
from NWORA resource file. conditions: following:
• The NWORA resource file does not exist. • If the NWORA resource file does not exist,
• The NWORA resource file has incorrect create the file.
permissions. • Ensure that the NWORA resource file has
• The NWORA resource file is corrupted. correct permissions.
• If the NWORA resource file is corrupted,
re-create the file.
“The NWORA resource file” on page 183 provides
more information.

Could not read link - pathname A proxy backup failed due to the pathname that Before a proxy backup, ensure that any symbolic
was an invalid symbolic link. link is a valid link.

Direct saves are not enabled. NMO attempted to connect to an old release of Update the NetWorker server software to a release
NetWorker server software that is no longer supported by NMO 5.0. The EMC Information
supported. Protection Software Compatibility Guide on the
Powerlink website provides details on the
supported server releases.

Error creating staging directory 'directory'. During a proxy restore of a regular file, the Ensure that the destination directory has valid
permissions of the destination directory were permissions for a proxy restore.
possibly invalid. NMO was not able to create the
required staging subdirectory, .nworapc.

Error in mmdb lookup by time: reason A lookup in the media database failed for the Use the mminfo command to verify the status of
given reason. the media database record. Take the corrective
action suggested by the error message.

224 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 3 of 9)

Error message Description Resolution

Exceeded the number of retries. The NMO could not contact the NetWorker index Restart the NetWorker services on the server, as
NetWorker server may be down or service nsrindexd. This was probably caused by required.
unreachable. the NetWorker services being shutdown.

Exceeded the number of retries for nsr_init(). After a maximum of five attempts, NMO failed to Restart the NetWorker services on the server, as
The NetWorker server may be down or call the NetWorker core function, nsr_init(). This required.
unreachable. was probably caused by the NetWorker services
being shut down.

Exceeded the number of retries for nsr_start(). After a maximum of five attempts, NMO failed to Restart the NetWorker services on the server, as
The NetWorker server may be down or call the NetWorker core function, nsr_start(). This required.
unreachable. was probably caused by the NetWorker services
being shut down.

Invalid browse and retention policies. Values The NSR_SAVESET_BROWSE and Ensure that the parameters
Ignored. NSR_SAVESET_RETENTION parameters both NSR_SAVESET_BROWSE and
had invalid time values. NSR_SAVESET_RETENTION in the RMAN script
both have valid values in the NetWorker date
format.

Invalid browse policy browse_time. Value The NSR_SAVESET_BROWSE parameter had Ensure that the parameter
Ignored. an invalid time value, browse_time. 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 “The send command” on page 211 provides the
command was incorrect. correct send command syntax.

Invalid retention policy: retention_time. Value The NSR_SAVESET_RETENTION parameter Ensure that the parameter
Ignored. had an invalid time value, retention_time. 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 Perform a proxy backup with a valid source
pathname. pathname only.

NSR_DATA_VOLUME_POOLn is not set. Multiple copies of the backup data were When multiple copies of backup data are
requested, but the required requested, set the required
NSR_DATA_VOLUME_POOL parameters were NSR_DATA_VOLUME_POOL parameters.
not set. In the message, n was replaced by a “NSR_DATA_VOLUME_POOL” on page 202
number corresponding to the missing pool provides more information.
parameter.

nsrsnapck_binary_name process failed with During an index removal for a proxy backup, the Report the error to Technical Support.
error - reason nsrsnapck binary failed. The binary name is
nsrsnapck on UNIX and nsrsnapck.exe on
Windows.

ORA-19511: Error received from media An NMO backup failed because a valid NWORA If you do not use the wizard to configure a
manager layer, error text: resource file does not exist or is not available. scheduled backup with save set bundling, use the
Could not create the NWORA resource lock nsrnmoadmin command to create a valid
file (13) (103:105:13) NWORA resource file, according to instructions in
Chapter 2 or 7 of this administration guide.

Oracle pools are not supported NMO does not support Oracle pools. NMO Remove the pool option of the backup command
supports NetWorker pools only. in the RMAN script or set the pool option to zero.
“The pool option of the backup command” on
page 210 provides more information.

Path pathname is too long. A proxy backup failed because the given Ensure that any pathname involved in a proxy
pathname exceeded the limit of 1,024 bytes. backup does not exceed 1,024 bytes.

pb_init() failed with (reason): invalid BRCAPI The version number of the BRC API that was Report the error to Technical Support.
version reported by the PowerSnap Module was
corrupted.

NMO error messages 225


Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 4 of 9)

Error message Description Resolution

Proxy copy is not supported. A proxy operation was attempted on a platform Do not attempt a proxy operation on an
that NMO does not support for proxy operations. unsupported platform. The EMC Information
Protection Software Compatibility Guide on the
Powerlink website provides details on supported
platforms.

'string' should be in format: KEY=(xxxxx) The syntax of the string in the RMAN send “The send command” on page 211 provides the
command was incorrect. correct send command syntax.

The ASDF body could not be unwrapped. The incoming recover stream of data could not be Report the error to Technical Support.
decoded due to a possible network error or data
corruption.

The backup file already exists: NMO could not complete the backup because the Change the format option string of the RMAN
backup_piece_name backup piece name already existed in the command to produce a unique backup piece
NetWorker client file index. name, or remove obsolete backup pieces. Then
restart the backup operation.

The BRC API did not return an error string for An unknown error occurred during a BRC API Report the error to Technical Support.
the SBTPC object: object_name function call by the PowerSnap Module.

The BRC status of logical object 'filename' The PowerSnap Module reported a failure during Report the error to Technical Support.
was failure: file_status a proxy backup of the file filename.

The call to nsr_init() failed with the message: A call of the NetWorker core function, nsr_init(), Take the corrective action suggested by the error
reason failed due to the given reason. message. If required, report the error to Technical
Support.

The call to nsr_start() failed with the message: A call of the NetWorker core function, nsr_rtart(), Take the corrective action suggested by the error
reason failed due to the given reason. message. If required, report the error to Technical
Support.

The call to pb_environment() failed with error: During a proxy operation, a pb_environment() Take the corrective action suggested by the error
reason function call failed due to the given reason. message. If required, report the error to Technical
Support.

The call to pb_open() failed with error: During a proxy operation, a pb_open() function Take the corrective action suggested by the error
reason call failed due to the given reason. message. If required, report the error to Technical
Support.

The call to pb_prepare() failed with error: During a proxy operation, a pb_prepare() function Take the corrective action suggested by the error
reason call failed due to the given reason. message. If required, report the error to Technical
Support.

The call to pb_status() failed for object During a proxy operation, a pb_status() function Take the corrective action suggested by the error
'object_name' with the error: call failed due to the given reason. message. If required, report the error to Technical
reason Support.

The call to pb_status() for object During a proxy operation, a pb_status() function Take the corrective action suggested by the error
'object_name' failed with error: call failed due to the given reason. message. If required, report the error to Technical
reason Support.

The canonical OS file name path is invalid: The operating system filename specified for a Ensure that the file pathname specified for a proxy
filename proxy operation was not a valid pathname. operation is a valid full pathname that is not a
directory.

The current time could not be obtained (errno). NMO could not obtain the current time due to an Report the operating system error (errno) to the
operating system error. appropriate vendor.

The data could not be XDR'd from the stream. The incoming recover stream of data could not be Report the error to Technical Support.
decoded due to a possible network error or data
corruption.

226 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 5 of 9)

Error message Description Resolution

The data source is neither a file or a RAW The file filename involved in a proxy backup was Ensure that filename is either a regular file or raw
volume - filename not recognized as a regular file or raw volume. For volume, as required for proxy backups.
proxy backups, NMO supports only regular files
and raw volumes.

The destination does not have the same A proxy restore of a raw device was attempted to Perform a proxy restore of the raw device to a
terminating name as the source a location with a different basename from the location with the same basename as the
'device_name'. backed-up source. For example, c1t2d0s2 is the backed-up source.
basename (or terminating name) of
/dev/rdsk/c1t2d0s2.

The file being recovered could not be found in During a proxy restore, an error occurred at the Ensure that there are no permission or other
its staging location: filename point where the file filename was to be moved problems with the destination directory and the
from the staging directory .nworapc to the staging directory .nworapc, and then restart the
destination directory. proxy restore. If the error recurs, report it to
Technical Support.

The file 'filename' cannot be removed from the During a proxy restore of the file filename, a file Remove the file file_name from the .nworapc
staging directory (errno). with the same name was found in the .nworapc subdirectory, and restart the proxy restore.
subdirectory, probably left there by a previous
failed restore. The errno is the error number from
the failed attempt to remove the existing file.

The function mm_retrieve() failed with the During a restore, a call of the NetWorker core Take the corrective action suggested by the error
error: reason function, mm_retrieve(), failed due to the given message. If required, report the error to Technical
reason. Support.

The function nsr_bind_recov_mm() failed with During a restore, a call of the NetWorker core Take the corrective action suggested by the error
the error: reason function, nsr_bind_recov_mm(), failed due to the message. If required, report the error to Technical
given reason. Support.

The function nsr_end() failed with the error A call of the NetWorker core function, nsr_end(), Take the corrective action suggested by the error
message: reason failed due to the given reason. message. If required, report the error to Technical
Support.

The function nsr_rstart() failed with the error: During a restore, a call of the NetWorker core Take the corrective action suggested by the error
reason function, nsr_rstart(), failed due to the given message. If required, report the error to Technical
reason. Support.

The function sbtinit2() has already been This is an internal error caused by Oracle calling Report the error to Technical Support.
called. the function sbinit2() twice.

The functions sbtinit() or sbtinit2() have not This is an internal error caused by Oracle not Report the error to Technical Support.
been called. calling the two SBT initialization routines.

The index entry failed the cross check: During an index lookup, the entry was located in Restart the NetWorker services, and use the
cfx_name(backup_piece_name) the client file index but not in the media database. mminfo and nsrinfo commands to verify the
save_time(save_time) backup information in the indexes. Run the nsrck
program to resolve any corruption of the indexes.

The lookup of 'backup_piece_name' on server NMO could not locate backup_piece_name in the Run the nsrck program to resolve any corruption
'server' failed - 'reason' indexes due to the reason. The indexes might be of the indexes.
corrupted.

The name of the NSR client could not be The name of the NetWorker client could not be Set the parameter NSR_CLIENT to the NetWorker
determined. determined. client name by using the send command.

The name of the NSR server could not be The name of the NetWorker server could not be Set the parameter NSR_SERVER to the
determined. determined. NetWorker server name by using the send
command.

NMO error messages 227


Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 6 of 9)

Error message Description Resolution

The NMO BRCAPI version version is outside NMO release 5.0 does not support the Ensure that a supported release of the PowerSnap
the range supported by the BRC service: PowerSnap Module release that was used for a Module is installed. The EMC Information
earliest_version - latest_version proxy operation. Protection Software Compatibility Guide on the
Powerlink website provides details on supported
releases.

The NSR client name could not be The name of the NetWorker client could not be Set the parameter NSR_CLIENT to the NetWorker
determined. determined. client name by using the send command.

The NSR server name could not be The name of the NetWorker server could not be Set the parameter NSR_SERVER to the
determined. determined. NetWorker server name by using the send
command.

The NSR_CLIENT parameter was not set. The name of the NetWorker client could not be Set the parameter NSR_CLIENT to the NetWorker
determined. client name by using the send command.

The NSR_DMO_BENCHMARK_1 parameter is The undocumented parameter Do not set the unsupported
no longer supported. NSR_DMO_BENCHMARK_1 was specified, but it parameter NSR_DMO_BENCHMARK_1.
is no longer supported.

The NSR_SERVER parameter was not set. The name of the NetWorker server could not be Set the parameter NSR_SERVER to the
determined. NetWorker server name by using the send
command.

The NW authentication for client 'client' was NMO could not obtain the required authentication Take the corrective action suggested by the error
refused by server 'server' because 'reason'. to connect to the NetWorker client file index due message.
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 Set the parameter NSR_CLIENT to the NetWorker
determined. client name by using the send command.

The NW server does not have a valid NMO The NetWorker server attempted a proxy Ensure that the NetWorker server has the required
proxy copy license. operation without the required license. license for the proxy operation.

The NW server has not been set. The name of the NetWorker server could not be Set the parameter NSR_SERVER to the
determined. NetWorker server name by using the send
command.

The NWORA file ID could not be XDR'd. This is an internal XDR error caused by a network Report the error to Technical Support.
xdrm: 0xvalue read or write operation.
NWORA fid: 0xvalue
ssid: 0xvalue
ssoff: 0xvalue

The NWORA resource file does not exist. A proxy backup failed because the NWORA Create the NWORA resource file by using the
Please create it with nsrnmoadmin. resource file did not exist. nsrnmoadmin command, and restart the proxy
backup. “The NWORA resource file” on page 183
provides details.

The NWORA resource lock file does not exist. A proxy backup failed because the NWORA Create the NWORA resource lock file by using the
Please create it by running 'nsrnmoadmin -r resource lock file did not exist. nsrnmoadmin -r list command, and restart the
list' proxy backup. “The NWORA resource file” on
page 183 provides details.

The NWORA resource In the NWORA resource file, Set the value of NSR_ORACLECAT_MODE to
NSR_ORACLECAT_MODE is in the NSR_ORACLECAT_MODE was set to the default either enabled or disabled (as required) by using
'undetermined' state. value of undetermined. the nsrnmoadmin command.

The object 'filename' is not a file. A proxy backup failed because the file filename is Perform a proxy backup of a supported type of
not a data file—neither a raw file nor a regular file. data file only.

228 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 7 of 9)

Error message Description Resolution

The ORACLE_SID must be set when During a scheduled proxy backup, the parameter In the nsrnmo script, set the parameter
performing proxy copy backups. ORACLE_SID was not set in the nsrnmo script. ORACLE_SID to the SID value of the Oracle
database.

The OS file name has been specified multiple This is an internal Oracle error caused by Oracle Report the error to Technical Support.
times by Oracle: filename specifying the same filename twice during a proxy
operation.

The parameter file cannot be open: filename The configuration file specified by the parameter Ensure that the value specified by the parameter
NSR_PROXY_PFILE could not be opened. The NSR_PROXY_PFILE is a valid pathname of the
file should contain PowerSnap parameter settings configuration file.
for a proxy backup or restore.

The pb_cancel() call for object 'object_name' The pb_cancel() function call failed during a proxy Report the error to Technical Support.
returned the error message: operation.
error

The pb_inquiry() call failed for object The pb_inquiry() function call failed during a proxy Report the error to Technical Support.
'object_name': operation.
error

The pb_inquiry() for object 'object_name' The pb_inquiry() function call failed during a proxy Report the error to Technical Support.
failed because: operation.
error

The pb_inquiry() of object 'object_name' The pb_inquiry() function call failed during a proxy Report the error to Technical Support.
returned error: operation.
error

The pb_restore() for object 'object_name' The pb_restore() function call failed during a Report the error to Technical Support.
failed with error: proxy operation.
error

The pb_save() of object 'object_name' The pb_save() function call failed during a proxy Report the error to Technical Support.
returned error: operation.
error

The pb_snapshot() call for object The pb_snapshot() function call failed during a Report the error to Technical Support.
'object_name' failed with error: proxy operation.
error

The record obtained has the wrong save time NMO located an index record in the client file Restart the NetWorker services, and run the nsrck
'save_time1'. The save time queried was index, but it had an unexpected save time. The program to resolve any corruption of the indexes.
'save_time2'. indexes might be corrupted.

The removal of SSID 'save_set_id' failed with An index deletion operation failed for the given Use the mminfo and nsrinfo commands to verify
error: reason reason. the status of the index record. If required, report
the error to Technical Support.

The restore destination path is not valid: During a proxy restore operation, NMO found the Ensure that the specified restore destination is a
filename specified restore destination, filename, to be valid pathname.
invalid.

The restore operation for the file failed for an During a proxy restore, an error occurred at the Ensure that there are no permission or other
unknown reason: filename point where the file filename was to be moved problems with the destination directory and the
from the staging directory .nworapc to the staging directory .nworapc, and then retry the
destination directory. 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 Take the corrective action suggested by the error
function, savefile_fini(), failed due to the given message. If required, report the error to Technical
reason. Support.

NMO error messages 229


Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 8 of 9)

Error message Description Resolution

The SBTPC object could not determine the During a proxy restore operation, NMO was Report the error to Technical Support.
destination of the restore. unable to determine where to restore the file.

The SBTPC object is not in the During a proxy operation, NMO and the Report the error to Technical Support.
PB_TYPE_PREPARE state: object_name PowerSnap Module became unsynchronized as
to the status of the object object_name.

The SBTPC object is not in the During a proxy operation, NMO and the Report the error to Technical Support.
SBTPCSTATUS_NOTREADY state: PowerSnap Module became unsynchronized as
object_name to the status of the object object_name.

The SBTPC object 'object_name' failed with The proxy backup or restore of a file failed during Report the error to Technical Support.
the error message: a PowerSnap Module operation, for the given
reason reason.

The SBTPC object 'object_name' is entering During a proxy operation, NMO and the Report the error to Technical Support.
the SBTPCSTART backup state but its BRC PowerSnap Module became unsynchronized as
type is: type to the status of the object object_name.

The SBTPC object 'object_name' is entering During a proxy operation, NMO and the Report the error to Technical Support.
the SBTPCSTART restore state but its BRC PowerSnap Module became unsynchronized as
type is: type to the status of the object object_name.

The SBTPC object 'object_name' is entering During a proxy operation, NMO and the Report the error to Technical Support.
the SBTPCSTART state but its status is: PowerSnap Module became unsynchronized as
status to the status of the object object_name.

The SBTPC object 'object_name' is leaving During a proxy operation, NMO and the Report the error to Technical Support.
the BRC prepare state but its status is: status PowerSnap Module became unsynchronized as
to the status of the object object_name.

The SBTPC object 'object_name' is leaving During a proxy operation, NMO and the Report the error to Technical Support.
the BRC save state but its status is: status PowerSnap Module became unsynchronized as
to the status of the object object_name.

The SBTPC object 'object_name' is leaving During a proxy operation, NMO and the Report the error to Technical Support.
the BRC snapshot state but its status is: status PowerSnap Module became unsynchronized as
to the status of the object object_name.

The SBTPC object 'object_name' was aborted The PowerSnap Module terminated the proxy Examine the PowerSnap Module logs for a
by the BRC service. Please check the operation. possible reason for the termination.
PowerSnap logs for an explanation.

The sbtpccommit() function was called during This is an internal Oracle error that occurred Report the error to Technical Support.
restore. during a proxy restore.

The sfhead could not be XDR'd. This is an internal XDR error caused by a network Report the error to Technical Support.
write operation.

The SS browse time is not in the future: The specified browse policy time was in the past. Ensure that the browse policy time is set correctly.
current time: current_time This might be due to a problem with the operating If required, ensure that the operating system time
browse: browse_time system time setting. is set correctly.

The SS retention time is not in the future: The specified retention policy time was in the Ensure that the retention policy time is set
current time: current_time past. This might be due to a problem with the correctly. If required, ensure that the operating
retention: rentention_time operating system time setting. system time is set correctly.

The staging directory 'directory' has invalid During a proxy restore, NMO was unable to write Ensure that the staging directory has valid
permissions (errno). to the staging directory, directory. The errno is the permissions for a proxy restore.
error number from the function call that failed.

230 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

Table 14 Error messages from the libnwora library (page 9 of 9)

Error message Description Resolution

The UNIX attributes could not be XDR'd. This is an internal XDR error caused by a network Report the error to Technical Support.
xdrm: 0xvalue read or write operation.
ua: 0xvalue

There are no SBTPC objects that have not This is an internal error during a proxy operation Report the error to Technical Support.
returned their status. caused by Oracle expecting more files to be
processed whereas NMO has completed its file
processing.

This backup piece name is already used in the This is an Oracle error caused by Oracle Report the error to Technical Support.
SBTPC session: backup_piece_name specifying the same backup piece name twice
during a proxy operation.

Error messages from the nsrnmoadmin program


Table 15 on page 231 lists error messages generated by the nsrnmoadmin program,
in alphabetical order.
The error messages appear in the following format:
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)

Error message Description Resolution

Command line arguments are not understood. The nsrnmoadmin command included one or Use the nsrnmoadmin command with the correct
more invalid options. options. “Configuring the NWORA resource file
with the nsrnmoadmin program” on page 188
provides details.

Could not create the NWORA resource file The nsrnmoadmin command could not create Ensure that valid permissions exist to allow the
(errno) the NWORA resource file, possibly due invalid nsrnmoadmin command to create the NWORA
permissions. resource file. “The NWORA resource file” on
page 183 provides details.

Could not create the NWORA resource lock The nsrnmoadmin command could not obtain Report the error to Technical Support.
file (errno) 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 Verify that the NWORA resource file exists and
NWORA resource file, possibly due invalid has valid permissions. If required, create or repair
permissions. the file by using the nsrnmoadmin command, or
modify the file permissions.

No command line parameters are set. The nsrnmoadmin command options were Use the nsrnmoadmin command with the correct
missing. options.

NSR_ORACLECAT_MODE can only be set to In the nsrnmoadmin command, the In the nsrnmoadmin command, set the
'enabled', 'disabled' or 'undetermined'. NSR_ORACLE_CAT_MODE parameter resource NSR_ORACLE_CAT_MODE parameter resource
was set to a value other than enabled, disabled, to enabled or disabled for instant backups.
or undetermined. “Configuring the NWORA resource file with the
nsrnmoadmin program” on page 188 provides
details.

NMO error messages 231


Troubleshooting and Error Messages

Table 15 Error messages from the nsrnmoadmin program (page 2 of 3)

Error message Description Resolution

NSR_REMOVE_ON_FAILURE can only be In the nsrnmoadmin command, the In the nsrnmoadmin command, set the
set to 'TRUE' or 'FALSE'. NSR_REMOVE_ON_FAILURE parameter NSR_REMOVE_ON_FAILURE parameter
resource was set to a value other than TRUE or resource to either TRUE or FALSE only.
FALSE.

NWORA parameter resources must be In the nsrnmoadmin command, an NWORA In the nsrnmoadmin command, specify the
specified in the 'ResourceName parameter resource name and value were not NWORA parameter resource name and value in
ResourceValue' format. specified in the correct format. the correct format.

NWORA SID resource must be specified In the nsrnmoadmin command with the -r delete In the nsrnmoadmin command with the -r delete
when doing deletion. option, the SID value of an Oracle database was option, specify the correct SID value.
not specified.

The '-r' flag cannot be set multiple times. The nsrnmoadmin command contained more Use the nsrnmoadmin command with only one -r
than one -r option. option.

The '-r' option requires an NWORA resource The nsrnmoadmin command with the -r option In the nsrnmoadmin command with the -r option,
specification. did not include the required resource specify the required resource name and value.
specification.

The '-r' option requires either an 'add', In the nsrnmoadmin command, the -r option did In the nsrnmoadmin command, include one of the
'update', 'list' or 'delete' option. not include one of the required keywords: add, required keywords with the -r option. “Configuring
update, list, or delete. the NWORA resource file with the nsrnmoadmin
program” on page 188 provides details.

The first NWORA resource is not a header The NWORA resource file is probably corrupted. Verify the contents of the NWORA resource file. If
(errno). required, repair the resource file by using the
nsrnmoadmin command.

The NWORA resource file does not contain The NWORA resource file does not contain the Verify the contents of the NWORA resource file. If
the NSR_NWPATH resource. mandatory NSR_NWPATH parameter resource. required, repair the resource file by using the
The file might be corrupted. 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 The nsrnmoadmin command specified the name In the nsrnmoadmin command, specify a valid
'resource_name' is not found. of a resource that does not exist in the NWORA resource name from the NWORA resource file.
resource file.

The NWORA resource parameter list can only The NWORA resource file includes multiple Repair the NWORA resource file by using the
contain one entry. values for a resource, which is not supported. The nsrnmoadmin command.
file is probably corrupted. The file might have
been edited manually, which is not supported.

The NWORA resource parameter list contains The NWORA resource file contains an invalid Repair the NWORA resource file by using the
the invalid element 'resource_name'. resource name. The file is probably corrupted. nsrnmoadmin command. “Configuring the
The file might have been edited manually, which NWORA resource file with the nsrnmoadmin
is not supported. program” on page 188 provides details.

The NWORA resource parameter list for a SID The nsrnmoadmin command for creating or In the nsrnmoadmin command for creating or
requires the item1, item2 and item3 updating an NWORA SID resource was missing updating an NWORA SID resource, include the
information. the required items. required items.

The NWORA resource 'resource_name' is not The nsrnmoadmin command with the -r delete In the nsrnmoadmin command with the -r delete
a SID resource. option did not include a valid name of an NWORA option, specify a valid name of an NWORA SID
SID resource. resource.

The NWORA resource specified is not In the nsrnmoadmin command, an invalid name In the nsrnmoadmin command, specify a valid
supported: resource_name = resource_value or value were specified for an NWORA parameter name and value for an NWORA parameter
resource. resource. “NWORA parameter resources” on
page 184 provides details.

232 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

Table 15 Error messages from the nsrnmoadmin program (page 3 of 3)

Error message Description Resolution

The NWORA SID resource for 'sid_value' The nsrnmoadmin command attempted to add In the nsrnmoadmin command, specify the
already exists. an NWORA SID resource that already existed. values for a new NWORA SID resource.

The SID token 'connect' is an empty string. The nsrnmoadmin command did not include the In the nsrnmoadmin command, specify a valid
required pathname of the RMAN connection file pathname of the RMAN connection file with the
with the connect keyword. connect keyword. “Configuring the NWORA
resource file with the nsrnmoadmin program” on
page 188 provides details.

The SID token 'home' is an empty string. The nsrnmoadmin command did not include the In the nsrnmoadmin command, specify a valid
required pathname of the Oracle home directory pathname of the Oracle home directory with the
with the home keyword. home keyword.

The SID token 'ORACLE_SID' is invalid. In the nsrnmoadmin command with the sid In the nsrnmoadmin command, specify a valid
keyword, the specified SID value of the Oracle SID value with the sid keyword.
database was invalid.

The SID token 'sid' is an empty string. The nsrnmoadmin command did not include the In the nsrnmoadmin command, specify a valid
required SID value of the Oracle database with SID value with the sid keyword.
the sid keyword.

The tokens 'sid', 'home' and 'connect' must be The nsrnmoadmin command to add an NWORA In the nsrnmoadmin command to add an
set when adding a SID. SID resource did not include the settings of the NWORA SID resource, include the settings of the
mandatory sid, home, and connect keywords. sid, home, and connect keywords.

The value of the NWORA resource is missing. In the nsrnmoadmin command with the -r In the nsrnmoadmin command with the -r update
update option, the NWORA resource value was option, specify the NWORA resource value with
not specified with the resource name. the resource name.

Unrecognized argument 'option'. The nsrnmoadmin command included the Use the nsrnmoadmin command with the correct
unrecognized option option. options. “Configuring the NWORA resource file
with the nsrnmoadmin program” on page 188
provides details.

You must be the super-user to update the The nsrnmoadmin command was typed by the Type the nsrnmoadmin command as the root
NWORA resource file. wrong user. user on UNIX, or as a member of the Microsoft
Windows Administrators group.

Error messages from the nsrnmoinfo program


Table 16 on page 234 lists error messages generated by the nsrnmoinfo program, in
alphabetical order.
The error messages appear in the following format:
The NW volume information lookup failed:
error_message
where error_message is the text of the error message, as shown in the table.

NMO error messages 233


Troubleshooting and Error Messages

Table 16 Error messages from the nsrnmoinfo program

Error message Description Resolution

A connection to NW server 'server' could not NMO could not connect to the NetWorker client Take the corrective action suggested by the error
be established because 'reason'. file index due to the given reason. The client message.
might not be configured as a client on the server.

Could not locate the LNM save file NMO could not locate an index record for the Use the mminfo and nsrinfo commands to verify
'backup_piece_name' on server 'server'. backup piece. The index record is probably the status of the index record.
missing.

Could not locate the LNM save time NMO could not locate an index record for the save Use the mminfo and nsrinfo commands to verify
'save_time' on server 'server'. time in the client file index. The index record is the status of the index record.
probably missing.

Error in mmdb lookup by time: reason A lookup in the media database failed for the Use the mminfo command to verify the status of
given reason. the media database record. Take the corrective
action suggested by the error message.

Exceeded the number of retries. The NMO could not contact the NetWorker index Restart the NetWorker services on the server, as
NetWorker server may be down or service nsrindexd. This was probably caused by required.
unreachable. the NetWorker services being shutdown.

The file 'filename' could not be opened. The file specified with the -f option of the Ensure that the specified file exists, and then type
nsrnmoinfo command could not be accessed. the nsrnmoinfo command again with the -f option.

The file name provided is NULL. In the nsrnmoinfo command, the -f option did not In the nsrnmoinfo command, include the required
include the required filename. filename with the -f option.

The index entry failed the cross check: During an index lookup, the entry was located in Restart the NetWorker services, and use the
cfx_name(backup_piece_name) the client file index but not in the media database. mminfo and nsrinfo commands to verify the
save_time(save_time) backup information in the indexes. Run the nsrck
program to resolve any corruption of the indexes.

The lookup of 'backup_piece_name' on server NMO could not locate backup_piece_name in the Run the nsrck program to resolve any corruption
'server' failed - 'reason' indexes due to the reason. The indexes might be of the indexes.
corrupted.

The NW authentication for client 'client' was NMO could not obtain the required authentication Take the corrective action suggested by the error
refused by server 'server' because 'reason'. to connect to the NetWorker client file index due message.
to the given reason. The client might not
configured as a client on the server.

The record obtained has the wrong save time NMO located an index record in the client file Restart the NetWorker services, and run the nsrck
'save_time1'. The save time queried was index, but it had an unexpected save time. The program to resolve any corruption of the indexes.
'save_time2'. indexes might be corrupted.

234 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Troubleshooting and Error Messages

Error messages from the nsrnmoprobe program


Table 17 on page 235 lists error messages generated by the nsrnmoprobe program, in
alphabetical order.

Table 17 Error messages from the nsrnmoprobe program

Error message Description Resolution

Could not connect to the Oracle database The nsrnmoprobe program could not connect to Ensure the following:
Oracle_service at Oracle_home. the specified Oracle database. • 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 97
provides more information.

The configuration information is not valid: The Command Options attribute in the NetWorker Correct the Command Options attribute setting in
string Probe resource was not configured properly. the NetWorker Probe resource, according to
“Configure a probe-based backup” on page 97.

The connect string was missing the user or the The RMAN connection file (specified through Edit the connection file to correct the problem.
password. NSR_ORACLE_CONNECT_FILE in the NWORA
resource file) used for the NMO probe did not
contain the username or password.

The nsrnmoprobe command does not support The nsrnmoprobe program does not support a Do not use the nsrnmoprobe program with a RAC
databases with more than one sthread. RAC database, where the database has multiple database.
threads (instances).

The nwora.res file has not been created. ORACLE_SERVICE was set in the Command Create the NWORA resource file according to the
Options attribute in the NetWorker Probe information in “Configure a probe-based backup”
resource, but the NWORA resource file did not on page 97.
exist.

The v$database_incarnation SQL statement The nsrnmoprobe program could not determine This error usually indicates that the Oracle
could not be run because: reason. the database incarnation information due to the database version is not supported. The EMC
reason given in the error message. Information Protection Software Compatibility
Guide on Powerlink provides details on the
supported Oracle database versions.

NMO error messages 235


Troubleshooting and Error Messages

Error messages from the nsrnmostart program


Table 18 on page 236 lists error messages generated by the nsrnmostart program, in
alphabetical order.

Table 18 Error messages from the nsrnmostart program

Error message Description Resolution

Authentication denied The lockbox that stored the database connection Use the NetWorker Management Console (NMC)
or credentials was not accessible by the super-user to ensure that the Lockbox resource is created for
Authentication denied for user username on the client where the backup failed with this the given client and the Users attribute contains
error message. the super-user of the client.

client: WARNING! The NWORA resource file The NWORA resource file could not be backed up Analyze the log_file and if it includes an error
'save' process output error messages. after a successful RMAN backup. message, take the corrective action suggested by
client: Please check the save log file for more the error message.
information: log_file

ORACLE_HOME is not defined. Cannot start ORACLE_HOME was not set properly in the Set ORACLE_HOME properly in the nsrnmo
RMAN. nsrnmo script. script.

The backup config did not contain a string. The nsrnmostart program was run with the -C Remove this Client resource, and re-create the
option, but the Backup Config attribute was not Client resource by using the backup configuration
set properly in the Client resource. wizard.

The NSR client resource for client_name does The nsrnmostart program was run with the -C Remove this Client resource, and re-create the
not contain any backup configuration. option, but the Backup Config attribute was not Client resource by using the backup configuration
set properly in the Client resource. wizard.

The temporary file 'rman_script_path' could The scheduled backup binary, nsrnmostart, Ensure that the root user on UNIX or the Windows
not be created (errno). could not create the file rman_script_path to write Administrator has "write" permissions on the
the RMAN script generated by the backup directory path of the rman_script_path file.
configuration wizard.

236 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Glossary

This glossary contains the definitions of terms found in this manual. Most of the
terms are specific to the NetWorker Module for Oracle (NMO) software. For terms
specific to the NetWorker software, refer to the latest NetWorker administration
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.

Administrators group Microsoft Windows user group whose members have the rights and privileges of
users in other groups, plus the ability to create, modify, and manage the users and
groups in the domain.

API (application Agreed-upon set of computer library routines, protocols, and tools used to
programming communicate and accomplish tasks within software applications.
interface)

ASM (application Program that, when used in a directive, specifies the way a set of files or directories is
specific module) 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 media Feature that enables the storage device controlled by the NetWorker server to
management automatically label, mount, and overwrite a volume it considers unlabeled. Volumes
that are eligible for reuse are also automatically recycled.

autochanger See “library.”

B
backup Operation that saves data to a volume. An Oracle backup of several datafiles may
include several backup sets.

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 237
Glossary

backup cycle See “scheduled backup cycle.”

backup group See “group.”

backup host See “proxy client host.”

backup level See “level.”

Backup Operators Microsoft Windows user group whose members have the capability to log in to a
group domain from a workstation or a server, back it up, and restore the data. Backup
Operators can also shut down servers or workstations.

backup piece Binary file created during an NMO 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 NMO 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.

C
catalog Process that removes a proxy Oracle backup entry from the RMAN catalog when the
synchronization 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.

238 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Glossary

D
daemon 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. NMO 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.

DBMS Abbreviation for database management system, which refers to the primary
architecture of an Oracle database.

deduplication Type of backup from the client to an Avamar server (NetWorker deduplication
backup 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.”

E
enabler code Special code that activates the software. The enabler code that unlocks the base
features for software is called a base enabler. Enabler codes for additional features or
products (for example, library support) are called add-on enablers.

event-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.

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 239
Glossary

F
failover 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
are several types of firewall techniques. The NetWorker software supports client
backups from computers that are protected by packet filtering.

full backup See “level.”

G
group Client or group of client computers that are configured to back up files at a
designated time of day.

H
high-availability System of multiple computers configured as cluster nodes on a network that ensures
system 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.”

I
I18N Capability of the NMO software to operate in a non-English environment or locale
(internationalization) without itself generating non-ASCII data. After I18N is set up, NMO can process and
display non-ASCII data that is passed to it by the operating system, NetWorker
software, and Oracle software. See also “L10N (localization).”

immediate live Type of proxy Oracle backup where a point-in-time copy (snapshot) is created during
backup 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.

240 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Glossary

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.”

L
L10N (localization) Capability of the NMO software to generate non-ASCII character and numeric data
based on the locale of the user, and output the data to the user interface. This
capability requires the installation of the appropriate NMO language packs in
non-English locales. See also “I18N (internationalization).”

level Backup configuration option that specifies how much data is saved during a
scheduled or manual backup. An NMO backup level is specified by an RMAN
command in the RMAN backup script only:
◆ A full NMO backup backs up all of the data blocks in the database, regardless of
when they last changed.
◆ An incremental NMO backup backs up only data blocks that have changed since
the last backup.

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.

license enabler Enabler code that is required to run a feature or product.

live backup One of the following two types of proxy Oracle backup:
◆ “deferred live backup”
◆ “immediate live backup”

M
manual 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 (media Library of media management routines provided by the NMO software and linked
management library) with the Oracle kernel software during the installation of NMO on the Oracle Server
host.

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 241
Glossary

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 to Set time required to perform an instance or media recovery for an Oracle database.
recover) 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.

N
NetWorker Network-based EMC software product that backs up and restores file systems.

NetWorker client See “client.”

NetWorker Module for NetWorker add-on module for the NetWorker server software that enables backups
Oracle and restores of an Oracle DBMS. See also “NMO.”

NetWorker resource See “resource.”

NetWorker server See “server.”

NetWorker storage See “storage node.”


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 NMO installation, which must be extended to
include specific additional resources to enable proxy Oracle backups and (optionally)
catalog synchronization.

O
offline 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 Enterprise Oracle Enterprise Manager Backup Management Tools, which include an optional
Manager graphical user interface to the RMAN utility.

Oracle10g Server Computer running an Oracle10g release 10.x DBMS. See “DBMS.”

242 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Glossary

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.

P
parallelism 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
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 NMO 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. NMO 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.

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 243
Glossary

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.”

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.

R
RDBMS (relational Type of “DBMS” that stores data in the form of related tables.
database
management system)

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.

244 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Glossary

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 Type of proxy restore that restores a proxy backup from a secondary storage
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 (Recovery Oracle utility that acts as an intelligent interface to Oracle databases and works with
Manager) third-party media management products, such as NMO, to back up and restore
Oracle database objects.

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 NMO 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, the superuser account.


◆ On Microsoft Windows and UNIX, the highest level of the system directory
structure.

RPC (remote Protocol that the NetWorker server uses to perform client requests over a network.
procedure call)
S
save 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 NMO into a bundle according to configuration
settings.

save set bundle join Creation, during an incremental scheduled NMO backup, of a combined save set
bundle from co-dependent save sets in different save set bundles.

save set bundling Process whereby NMO 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.

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 245
Glossary

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.

scheduled backup Full or level 0 backup of an Oracle database object and all the subsequent incremental
cycle 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.

246 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Glossary

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.

T
tablespace 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 NMO software for an additional 45 days beyond
the evaluation period.

TNS (Transparent Oracle networking technology that provides a single interface to all standard
Network Substrate) 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.”

U
unscheduled backup See “manual backup.”

user Person who uses NetWorker software from a computer to back up and restore files.

V
versions 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.”

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 247
Glossary

248 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Index

A Schedule resource 87
ADDRESS_LIST parameter 144 sessions 54
Administrator attribute backup and recovery system 49, 50, 152
in Server resource 63 Backup Command attribute
Administrator program, NetWorker in Client resource 88, 170, 195, 196
Stop button 111 backup copies during manual backups 27
Administrator, NetWorker backup current control file command 132
Client resource 87, 170 Backup Snapshots attribute
Device resource 66 in Snapshot Policy resource 158
Group resource 86, 169 backup spfile command 132
Label Template resource 67 backup types
Pool resource 66, 169 archived redo log 20, 131, 147
Schedule resource 87 control file 130, 132
Server resource 63 deduplication 21, 93
Snapshot Policy resource 150, 154, 169 deferred live 150, 158, 173
User Group resource 64 immediate live 150, 156
AES encryption 122, 201 instant 150, 154, 173, 182
allocate channel command manual 21, 48, 54, 104
parms option 215 NetWorker bootstrap 48, 53, 106, 155
API, SBT 52, 57, 58 NWORA resource file 173, 176
archived redo log password file 130
backup 20, 131, 147 probe-based 23, 97
restore 148 proxy 24, 150, 154, 171
sharing across RAC nodes 147 registry files 130
attributes for resources scheduled 21, 48, 110, 154
Client resource 87 BACKUP_TAPE_IO_SLAVES parameter 216
Server resource 63 bootstrap, NetWorker 48, 53, 106, 155
autochanger 66, 105, 123 Browse Policy attribute
automatic catalog synchronization for proxy backups 181, in Client resource 88, 170
191 browse policy uniformity 37
automatic channel allocation 26, 75, 77 bundling, save set 39

B C
backup canceling
Client resource 87, 170 manual backup 107
command 165, 214 nonresponding backup 107
command (pool option) 210 proxy backup 174
command (trace option) 217 scheduled backup 78, 111
copies 27 catalog synchronization for proxy backups 181, 191
devices, multiple 47 change backuppiece...unavailable command 216
error messages 221, 222 change...crosscheck command 210, 216
Group resource 86, 169 change...delete command 192
level (full or incremental in RMAN script) 66 channel option, send command 212

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 249
Index

client file index 47, 175, 176 Server resource 63


Client resource 87, 170 Snapshot Policy resource 150, 154, 169
Client resource attributes User Group resource 64
Backup Command 88 configuration wizard 62
Browse Policy 88 configure channel command
Group 88 parms option 167, 200
Name 88 connection file, for catalog synchronization 184, 186, 188
Remote Access 89 control file
Retention Policy 89 backup 130, 132
Save Set 89 control files, mirrored 130
Schedule 89 crosscheck command 210, 216
Client Retries attribute
in Group resource 194
D
Client Retries attribute in Group resource 195
database
cluster systems 29, 138, 194
manual backup 48, 54, 104
command
scheduled backup 48, 110, 154
allocate channel, parms option 215
Datazone pass phrase attribute
backup 165, 214
in Server resource 63
backup (pool option) 210
DBID 130, 131
backup (trace option) 217
deduplication
backup current control file 132
backup 21, 93
backup spfile 132
restore 21
change backuppiece...unavailable 216
deferred live backup 150, 158, 173
change...crosscheck 210, 216
delete expired backup command 210
change...delete 192
destructive restores
configure channel, parms option 167, 200
overview 151
crosscheck 210, 216
Device resource 66
delete expired backup 210
device_type option, send command 212
nsrnmoinfo 123
devices, multiple 47
restore 214
disaster recovery
rman 82, 105, 126
preparation 130
rman send 200, 213, 215
types 133
rman.exe 105, 126
savefs 53, 154, 156
savegrp 53, 106, 154, 156 E
send 166, 200, 211 email notification 91
send (channel option) 212 enabling
send (device_type option) 212 policy uniformity 93
send (NSR_ENV keyword) 211 save set bundling 92
send (precedence rules) 215 encryption
set 81, 200 AES encryption 122, 201
set duplex 215, 216 environment variable
setenv 81, 200 LC_ALL 68
su 84, 85 NLS_LANG 68
Command attribute error messages
in Probe resource 98 NetWorker Module for Oracle 222
Command Options attribute RMAN 221
in Probe resource 98
configuration
Client resource 87, 170
F
Device resource 66 failover
Group resource 86, 169 backup 139, 140, 145
Label Template resource 67 connect-time 139, 144
manual backup 104 proxy backup 194
Oracle system 63 FAILOVER parameter 144
Pool resource 66, 169 force_rollback option, not supported 178
proxy backup 163
roadmap 62 G
Schedule resource 87 Group attribute
scheduled backup 110 in Client resource 88, 170

250 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Index

Group resource 86, 169 listener.ora 130, 131, 144


GUI, NetWorker sqlnet.ora 130, 131
Stop button 111 tnsnames.ora 130, 131, 144
NetWorker
bootstrap backup 48, 53, 106, 155
I client file index 47, 175, 176
I18N (internationalization) 33, 35, 67, 68
configuration
immediate live backup 150, 156
Client resource 170
initialization parameter file
media database 47, 175, 176
initoracle_sid.ora 145
media database service, nsrmmdbd 55, 58, 155
PFILE 130, 131, 132
media service, nsrmmd 54, 57
SPFILE 130, 132
processes
INSTANCE_NAME parameter 145
restore 163
instant backup 150, 154, 173, 182
software features 47
instant restore 151, 160, 177
staging 39
internationalization (I18N) 33, 35, 67, 68
NetWorker Administrator program
Stop button 111
L NetWorker configuration
L10N (localization) 35, 69 Client resource 87
Label Template resource 67 Device resource 66
LC_ALL environment variable 68 Group resource 86, 169
LD_LIBRARY_PATH parameter 81, 82 Label Template resource 67
LIBPATH parameter 82 Pool resource 66, 169
live backup roadmap 62
deferred 150, 158, 173 Schedule resource 87
immediate 150, 156 Server resource 63
LOCAL_LISTENER parameter 145 Snapshot Policy resource 150, 154, 169
localization (L10N) 35, 69 NetWorker Module for Oracle
LOG_THRESHOLD parameter 98 API, SBT 52, 57, 58
error messages 222
features 48
M MML 153, 160, 161
mandatory parameters 81, 98, 202, 203, 204, 206, 207 troubleshooting 220, 221
manual backup 48, 54, 104 NetWorker PowerSnap Modules 24, 48, 65, 150, 163
canceling 107 NLS_LANG environment variable 68
monitoring 108 NMO internationalization (I18N) 33, 35, 67, 68
procedure for running 104 NMO localization (L10N) 35, 69
roadmap 104 NMO wizard 62
manual catalog synchronization for proxy backups 191 nmo.messages.raw file 217, 222
media database 47, 175, 176 NOCATALOG mode 146
media management nonresponding Oracle backup, canceling 107
storage devices 66 notification by email 91
volume pools 66 NSR_AES_ENCRYPTION parameter 76, 201
media management library (MML) 52, 153, 160, 161 NSR_BUNDLING parameter 92
mirrored NSR_CHECKSUM parameter 201
control files 130 NSR_CLIENT parameter 126, 178, 201
online redo logs 130 NSR_COMPRESSION parameter 201
MML 52, 153, 160, 161 NSR_DATA_MOVER parameter 167
MML catalog 181 NSR_DATA_VOLUME_POOL* parameters 67, 78, 202,
monitoring 216
manual backup 108 NSR_DEBUG_FILE parameter 83, 98, 203
scheduled backup 112 NSR_DEBUG_LEVEL parameter 203
multiple backup devices 47 NSR_DEDUP_BACKUP parameter 203
NSR_DEDUP_CACHE_ENABLED parameter 203
N NSR_DEDUP_CACHE_TAG parameter 204
Name attribute NSR_DEDUP_CHUNK_SIZE parameter 204
in Client resource 88 NSR_DEDUP_NODE parameter 204
in Probe resource 98 NSR_DPRINTF parameter 204
in Server resource 63 NSR_ENCRYPTION_PHRASES parameter 122, 205
network files NSR_ENV keyword in send command 211

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 251
Index

NSR_GROUP parameter 78, 205 media management library (MML) 52


NSR_INCR_EXPIRATION parameter 93 mirrored
NSR_MAX_STREAMS parameter 168 control files 130
NSR_MMDB_RETRY_TIME parameter 205 online redo logs 130
NSR_NO_BUSY_ERRORS parameter 205 password file 78
NSR_NO_MULTIPLEX parameter 205 recover 20, 127
NSR_NWPATH parameter 184, 206 Recovery Catalog 49, 50, 51, 152
NSR_ORACLE_CONNECT_FILE parameter 98, 186 Recovery Catalog database
NSR_ORACLE_HOME parameter 186 connection to 78
NSR_ORACLE_LIB_PATH parameter 187 Recovery Manager (RMAN) 49, 57, 152, 159
NSR_ORACLE_NLS_LANG parameter 69 restore 20, 48, 121, 123, 148
NSR_ORACLE_NLS_LANG parameter resource 185 sbtio.log file 217, 222
NSR_ORACLE_SID parameter 187 scheduled backup script 77
NSR_ORACLE_TNS_ADMIN parameter 187 Server sessions 54
NSR_ORACLECAT_DEBUG_FILE parameter resource System Backup to Tape (SBT) API 52, 57, 58
184 system configuration 63
NSR_ORACLECAT_LOG_FILE parameter resource 184 Oracle ASM 44
NSR_ORACLECAT_MODE parameter resource 185, 193 See also Oracle Automated Storage
NSR_PROXY_PFILE parameter 206 Oracle Automated Storage 44
NSR_PS_SAVE_PARALLELISM parameter 167 Oracle Enterprise Manager
NSR_RECOVER_POOL parameter 206 See Oracle Enterprise Manager Backup Management
NSR_REMOVE_ON_FAILURE parameter resource 185 Tools
NSR_RETENTION parameter 89 Oracle Enterprise Manager Backup Management Tools 52,
NSR_RETENTION_DISABLED parameter 206 57, 105, 126, 159
NSR_RMAN_ARGUMENTS parameter 82 Oracle Net 63, 86, 140, 141, 144
NSR_SAVESET_BROWSE parameter 88, 206 ORACLE_HOME parameter 81, 83, 98
NSR_SAVESET_RETENTION parameter 207 ORACLE_SERVICE parameter 98
NSR_SB_DEBUG_FILE parameter 83 ORACLE_SID parameter 83
NSR_SERVER parameter 78, 126, 207 ORACLER_USER parameter 83
NSR_SERVER_NIC parameter 207
nsrexecd service 53, 54, 57, 154
P
nsrindexd service 55, 57, 155
parallel save stream 47
nsrmmd service 54, 57
Parallelism attribute
nsrmmdbd service 55, 58, 155
in Client resource 170
nsrnmo script 53, 81, 154, 157
in Server resource 63
parameters 79, 80
parameter
nsrnmoadmin program 185, 187, 188, 190
ADDRESS_LIST 144
nsrnmoinfo command 123
FAILOVER 144
nsrnmoprobe program 24, 98, 99
INSTANCE_NAME 145
nsrnmostart program 53, 154, 157
LOCAL_LISTENER 145
nsroraclecat program 182, 191, 194
SERVICE_NAME 144, 145
nsrsnapck program 154, 157, 191, 192, 194
SID_LIST_LISTENER 144, 145
NWORA parameter resources
parameter resources
NSR_ORACLE_NLS_LANG 185
NSR_ORACLE_NLS_LANG 185
NSR_ORACLECAT_DEBUG_FILE 184
NSR_ORACLECAT_DEBUG_FILE 184
NSR_ORACLECAT_LOG_FILE 184
NSR_ORACLECAT_LOG_FILE 184
NSR_ORACLECAT_MODE 185, 193
NSR_ORACLECAT_MODE 185, 193
NSR_REMOVE_ON_FAILURE 185
NSR_REMOVE_ON_FAILURE 185
NWORA resource file 183, 185, 186, 187
parameters
backup 173, 176
BACKUP_TAPE_IO_SLAVES 216
NWORA SID resources 186, 187
in nsrnmo script 79, 80
nworapc directory 178
LD_LIBRARY_PATH 81, 82
LIBPATH 82
O LOG_THRESHOLD 98
online NetWorker indexes 47 mandatory 81, 98, 202, 203, 204, 206, 207
online redo logs, mirrored 130 NSR_AES_ENCRYPTION 76, 201
Oracle NSR_BUNDLING 92
backup and recovery system 49, 50, 152 NSR_CHECKSUM 201
DBID 130, 131 NSR_CLIENT 126, 178, 201
manual backup script 75, 165, 214 NSR_COMPRESSION 201

252 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Index

NSR_DATA_MOVER 167 policy uniformity 37


NSR_DATA_VOLUME_POOL* 67, 78, 202, 216 pool option, backup command 210
NSR_DEBUG_FILE 83, 98, 203 Pool resource 66, 169
NSR_DEBUG_LEVEL 203 pool, volume
NSR_DEDUP_BACKUP 203 defined 66
NSR_DEDUP_CACHE_ENABLED 203 label template for 66
NSR_DEDUP_CACHE_TAG 204 pool types 66
NSR_DEDUP_CHUNK_SIZE 204 POSTCMD parameter 85
NSR_DEDUP_NODE 204 postcommand script 85, 131, 134
NSR_DPRINTF 204 postprocessing scripts 85
NSR_ENCRYPTION_PHRASES 122, 205 PowerSnap Modules 24, 48, 65, 150, 163
NSR_GROUP 78, 205 PowerSnap parameters 165, 166, 167
NSR_INCR_EXPIRATION 93 precedence rules for send command 215
NSR_MAX_STREAMS 168 PRECMD parameter 84
NSR_MMDB_RETRY_TIME 205 preprocessing scripts 84
NSR_NO_BUSY_ERRORS 205 primary storage 24, 150, 153, 163, 171
NSR_NO_MULTIPLEX 205 Probe resource
NSR_NWPATH 184, 206 attributes
NSR_ORACLE_CONNECT_FILE 98, 186 Command 98
NSR_ORACLE_HOME 186 Command Options 98
NSR_ORACLE_LIB_PATH 187 Name 98
NSR_ORACLE_NLS_LANG 69 probe-based
NSR_ORACLE_SID 187 backup 23, 97
NSR_ORACLE_TNS_ADMIN 187 processes
NSR_PROXY_PFILE 206 restore 163
NSR_PS_SAVE_PARALLELISM 167 program
NSR_RECOVER_POOL 206 nsrd 53, 54, 57, 154, 156
NSR_RETENTION 89 nsrexecd 53, 154
NSR_RETENTION_DISABLED 206 nsrindexd 55, 57, 155
NSR_RMAN_ARGUMENTS 82 nsrmmd 54, 57
NSR_SAVESET_BROWSE 88, 206 nsrmmdbd 55, 58, 155
NSR_SAVESET_RETENTION 207 nsrnmo 53, 154, 157
NSR_SB_DEBUG_FILE 83 nsrnmoadmin 185, 187, 188, 190
NSR_SERVER 78, 126, 207 nsrnmoprobe 24, 98, 99
NSR_SERVER_NIC 207 nsrnmostart 53, 154, 157
ORACLE_HOME 81, 83, 98 nsroraclecat 182, 191, 194
ORACLE_SERVICE 98 nsrsnapck 154, 157, 191, 192, 194
ORACLE_SID 83 savefs 53, 154, 156
ORACLE_USER 83 savegrp 53, 154, 156
PATH 81 proxy
POSTCMD 85 backup 24, 150, 154, 171
PowerSnap 165, 166, 167 restore 151, 159, 177
PRECMD 84 psrollback.res file 179
RESTORE_TYPE_ORDER 168, 177, 178, 179
scheduled backup 110
R
SHLIB_PATH 82
RAC nodes as storage nodes 142
TNS_ADMIN 86
RAC systems 29, 138, 140
parms option
recover 20, 127
allocate channel command 215
Recovery Catalog 49, 50, 51, 152
configure channel command 167, 200
Recovery Catalog database
password file 78
connection to 78
backup 130
Recovery Catalog, backup 130, 132
PATH parameter 81
Recovery Manager (RMAN) 49, 52, 57, 152, 159
persistent settings 25, 26, 77
backup scripts 75
PFILE 130, 131, 132
commands 209
physical cluster client, proxy backups from 196
error messages 221
point-in-time copy 24, 150, 151, 160, 169, 170, 175
features 50
policy
manual backup scripts 76
browse 88
proxy backup scripts 165
retention 89
proxy restore scripts 177

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 253
Index

Recovery Catalog 49, 50, 51, 152 sbtio.log file 217, 222
restore scripts 121 Schedule attribute
scheduled backup scripts 77, 78 in Client resource 89, 170
redo logs, mirrored 130 Schedule resource 87
registry files scheduled backup 48, 110, 154
backup 130 automatic backup of client indexes and server
relocating files during proxy restores 179, 180 bootstrap 106
Remote Access attribute canceling 78, 111
in Client resource 89, 170, 195, 196 configuring Group resource 86
remote procedure call (RPC) protocol 47 configuring Schedule resource 87
resource monitoring 112
types of parameters 79, 110
Client 87, 170 postprocessing scripts 85
Device 66 preprocessing scripts 84
Group 86, 169 roadmap 110
Label Template 67 script
Pool 66, 169 nsrnmo 53, 154, 157
Schedule 87 secondary storage 150, 151
Server 63 send command 166, 200, 211
Snapshot Policy 150, 154, 169 channel option 212
User Group 64 device_type option 212
restore NSR_ENV keyword 211
archived redo log 148 precedence rules 215
command 214 Server resource
deduplication 21 attributes
determining required volumes 123 Datazone pass phrase 63
from secondary storage 151 Name 63
instant 151, 160, 177 Parallelism 63
Oracle data 20, 48, 121, 123 Server resource attributes
processes 163 Administrator 63
proxy 151, 159, 177 service
rollback 151, 160, 177, 179 nsrd 53, 54, 57, 154, 156
to a different host 126, 180 nsrexecd 53, 154
RESTORE_TYPE_ORDER parameter 168, 177, 178, 179 nsrindexd 55, 57, 155
Retain Snapshots attribute nsrmmd 54, 57
in Snapshot Policy resource 182 nsrmmdbd 55, 58, 155
Retention Policy attribute SERVICE_NAME parameter 144, 145
in Client resource 89, 170 set command 81, 200
retention policy uniformity 37 set duplex command 215, 216
RMAN setenv command 81, 200
See Recovery Manager (RMAN) SHLIB_PATH parameter 82
RMAN catalog 181 SID_LIST_LISTENER parameter 144, 145
rman command 82, 105, 126 silo 66
RMAN repository 181 snapshot 24, 150, 151, 160, 169, 170, 175
rman send command 200, 213, 215 Snapshot attribute in Group resource 171, 172
rman.exe command 105, 126 Snapshot Policy resource 150, 154, 169
roadmap Snapshot Pool attribute in Group resource 169
manual backup 104 SPFILE 130, 132
scheduled backup 110 staging 39
rollback restore 151, 160, 177, 179 Stop button
RPC (remote procedure call) 47 in NetWorker Administrator program 111
stopping
manual backup 107
S nonresponding backup 107
Save Set attribute
scheduled backup 78, 111
in Client resource 89, 170, 195
storage devices
save set bundling 39
Device resource 66
savefs command 53, 154, 156
volume pool 66, 67
savegroup completion report for proxy backup 173
Storage Nodes attribute
savegrp command 53, 106, 154, 156
in Client resource 170
SBT API 52, 57, 58

254 EMC NetWorker Module for Oracle Release 5.0 Administration Guide
Index

su command 84, 85
synchronization
automatic catalog 181, 191
manual catalog 191
System Backup to Tape (SBT) API 52, 57, 58

T
tablespace
manual backup 48, 54, 104
scheduled backup 48, 110, 154
target database
connection to 78
TNS_ADMIN parameter 86
trace option, backup command 217
troubleshooting
manual backup 221
NetWorker Module for Oracle 220, 221
scheduled backup 79, 221
types of backup
archived redo log 20, 131, 147
manual 48, 54, 104
NetWorker bootstrap 48, 53, 106, 155
scheduled 48, 110, 154

U
uniformity, policy 37
User Group resource 64

V
virtual cluster client, proxy backups from 195
volume pool
Default pool 78
defined 66
resource 66, 169
specifying 78
types 66
volumes, determining for restore 123

W
wizard, configuration 62

EMC NetWorker Module for Oracle Release 5.0 Administration Guide 255
Index

256 EMC NetWorker Module for Oracle Release 5.0 Administration Guide