Using IBM LTO Ultrium with Open Systems - RS/6000 Home

Front cover 

Using IBM LTO Ultrium 

with Open Systems 

How to setup Ultrium tape drives in 

multiple open environments 

Integration with popular backup 

applications 

Direct SCSI and SAN 

attachment 

Charlotte Brooks 

Anthony Abete 

Marco Ceresoli 

Josef Weingand 

ibm.com/redbooks

International Technical Support Organization 

Using IBM LTO Ultrium with Open Systems 

January 2002 

SG24-6502-00

Take Note! Before using this information and the product it supports, be sure to read the 

general information in “Special notices” on page 317. 

First Edition (January 2002) 

This edition applies to the IBM Ultrium LTO tape drives and libraries and associated SAN products 

and adapters. 

This document was created or updated on June 12, 2002. 

Comments may be addressed to: 

IBM Corporation, International Technical Support Organization 

Dept. QXXE Building 80-E2 

650 Harry Road 

San Jose, California 95120-6099 

When you send information to IBM, you grant IBM a non-exclusive right to use or distribute the 

information in any way it believes appropriate without incurring any obligation to you. 

© Copyright International Business Machines Corporation 2002. All rights reserved. 

Note to U.S Government Users – Documentation related to restricted rights – Use, duplication or disclosure is subject to 

restrictions set forth in GSA ADP Schedule Contract with IBM Corp.

Contents 

Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 

The team that wrote this redbook. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 

Special notice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 

IBM trademarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 

Comments welcome. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 

Part 1. Setting up LTO Ultrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 

Chapter 1. Introduction to LTO Ultrium with UNIX . . . . . . . . . . . . . . . . . . . . 3 

1.1 LTO overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 

1.2 LTO development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 

1.2.1 LTO enhancements since introduction . . . . . . . . . . . . . . . . . . . . . . . . 5 

1.2.2 LTO Ultrium models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 

1.2.3 IBM 3580 Ultrium Tape Drive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 

1.2.4 IBM 3581 Ultrium Tape Autoloader. . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

1.2.5 IBM 3583 Ultrium Scalable Tape Library . . . . . . . . . . . . . . . . . . . . . 10 

1.2.6 IBM 3584 Ultrium UltraScalable Tape Library. . . . . . . . . . . . . . . . . . 12 

1.3 Hardware and operating system platforms . . . . . . . . . . . . . . . . . . . . . 16 

1.3.1 Hardware server platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 

1.3.2 ISV storage management software. . . . . . . . . . . . . . . . . . . . . . . . . . 18 

1.4 Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 

1.4.1 SCSI direct attach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 

1.4.2 SAN Data Gateway attach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 

1.4.3 The Integrated SAN Data Gateway Module . . . . . . . . . . . . . . . . . . . 23 

1.4.4 Determining the number of drives on a SCSI bus . . . . . . . . . . . . . . . 23 

1.5 Host bus adapters and device drivers. . . . . . . . . . . . . . . . . . . . . . . . . 24 

1.6 LVD versus HVD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 

1.6.1 What is LVD and HVD?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 

1.6.2 Why is this important?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 

1.7 SCSI HD68 and VHDCI cable connectors . . . . . . . . . . . . . . . . . . . . . 27 

Chapter 2. Basic LTO setup for UNIX systems . . . . . . . . . . . . . . . . . . . . . . 29 

2.1 Installing library and device drivers . . . . . . . . . . . . . . . . . . . . . . . . . . 30 

2.2 LTO driver installation for AIX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 

2.2.1 Atape driver installation using the command line interface . . . . . . . . 33 

© Copyright IBM Corp. 2002 3

2.2.2 Installation using the SMIT interface . . . . . . . . . . . . . . . . . . . . . . . . . 34 

2.2.3 Configuring tape and medium changer devices . . . . . . . . . . . . . . . . 35 

2.2.4 Verifying the Atape driver installation . . . . . . . . . . . . . . . . . . . . . . . . 36 

2.2.5 Configuring the LTO device parameters . . . . . . . . . . . . . . . . . . . . . . 37 

2.2.6 Deleting LTO devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 

2.2.7 Removing Atape driver from the system (uninstallation). . . . . . . . . . 39 

2.3 LTO driver installation for Solaris . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 

2.3.1 IBMtape driver installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 

2.3.2 Configuring tape and medium changer devices . . . . . . . . . . . . . . . . 43 

2.3.3 Verifying the IBMtape driver installation . . . . . . . . . . . . . . . . . . . . . . 46 

2.3.4 Deleting LTO devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 

2.3.5 Removing IBMtape driver from the system (uninstallation). . . . . . . . 47 

2.4 LTO driver installation for HP-UX . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 

2.4.1 Atdd driver installation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 

2.4.2 LTO medium changer configuration . . . . . . . . . . . . . . . . . . . . . . . . . 50 

2.4.3 Verifying the Atdd driver installation . . . . . . . . . . . . . . . . . . . . . . . . . 54 

2.4.4 Configuring tape devices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 

2.4.5 Deleting the LTO medium changer . . . . . . . . . . . . . . . . . . . . . . . . . . 56 

2.5 Testing the library with tapeutil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

2.6 Fibre Channel host bus adapter driver installation . . . . . . . . . . . . . . . 59 

2.7 IBM RS/6000 Fibre Channel HBA installation . . . . . . . . . . . . . . . . . . . 61 

2.7.1 IBM RS/6000 HBA microcode installation. . . . . . . . . . . . . . . . . . . . . 62 

2.8 SUN Solaris Fibre Channel HBA driver installation . . . . . . . . . . . . . . . 64 

2.8.1 QLogic QLA2200 HBA driver installation . . . . . . . . . . . . . . . . . . . . . 64 

2.8.2 QLogic QLA2200 HBA driver configuration. . . . . . . . . . . . . . . . . . . . 67 

2.8.3 QLogic QLA2200 HBA driver removal . . . . . . . . . . . . . . . . . . . . . . . 68 

2.8.4 QLogic HBA FCode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 

2.8.5 Emulex LP8000 HBA driver installation . . . . . . . . . . . . . . . . . . . . . . 69 

2.8.6 Emulex LP8000 HBA driver configuration. . . . . . . . . . . . . . . . . . . . . 73 

2.8.7 Emulex LP8000 HBA driver removal. . . . . . . . . . . . . . . . . . . . . . . . . 73 

2.8.8 Emulex HBA firmware and bootcode . . . . . . . . . . . . . . . . . . . . . . . . 75 

2.9 StorWatch Tape Library Specialist . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 

2.9.1 IBM 3583 StorWatch Specialist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 

2.9.2 IBM 3584 StorWatch Specialist. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 

2.10 Updating library, drive and RMU firmware . . . . . . . . . . . . . . . . . . . . 88 

2.11 How to update the firmware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88 

2.11.1 Upgrading library firmware using tapeutil . . . . . . . . . . . . . . . . . . . . 89 

2.11.2 Upgrading drive firmware using tapeutil . . . . . . . . . . . . . . . . . . . . . 90 

2.11.3 Upgrading RMU firmware using the StorWatch Specialist . . . . . . . 91 

Chapter 3. SAN setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 

3.1 SAN design considerations for LTO libraries . . . . . . . . . . . . . . . . . . . 94 

3.1.1 How many tape drives can I attach to a FC HBA? . . . . . . . . . . . . . . 94 

4 Using IBM LTO Ultrium with Open Systems

3.1.2 Connecting tape drives to a SDG . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 

3.1.3 Multiple paths to tape drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 

3.2 Zoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 

3.2.1 Types of zoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 

3.2.2 Zoning setup with IBM 2109 Switch StorWatch Specialist . . . . . . . 103 

3.2.3 Another zoning example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 

3.3 Persistent binding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 

3.3.1 The need for persistent binding. . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 

3.3.2 Persistent binding with QLogic HBA on Solaris . . . . . . . . . . . . . . . 118 

3.3.3 Persisting binding with Emulex HBA on Solaris . . . . . . . . . . . . . . . 121 

3.3.4 Persistent binding not applicable to AIX . . . . . . . . . . . . . . . . . . . . . 123 

3.4 SAN Data Gateway setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 

3.4.1 Basic setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 

3.4.2 Install and use the StorWatch SDG Specialist . . . . . . . . . . . . . . . . 129 

3.4.3 LUN mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 

3.4.4 Access control by channel zoning. . . . . . . . . . . . . . . . . . . . . . . . . . 141 

3.4.5 Access control by Virtual Private SAN (VPS) . . . . . . . . . . . . . . . . . 142 

3.5 Operating system device names. . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 

Part 2. LTO Ultrium with applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 

Chapter 4. Implementing Tivoli Storage Manager with LTO . . . . . . . . . . 147 

4.1 Tivoli Storage Manager overview . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 

4.1.1 TSM commonly used terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 

4.2 Configuring non-shared LTO with TSM. . . . . . . . . . . . . . . . . . . . . . . 152 

4.2.1 Installing Tivoli Storage Manager . . . . . . . . . . . . . . . . . . . . . . . . . . 153 

4.2.2 Configure non-shared tape library and drives for AIX . . . . . . . . . . . 154 

4.2.3 Configure non-shared tape library and drives for Solaris . . . . . . . . 158 

4.2.4 Configure non-shared tape library and drives for HP-UX . . . . . . . . 162 

4.2.5 Define device class and storage pool . . . . . . . . . . . . . . . . . . . . . . . 165 

4.2.6 Inserting data and cleaner cartridges . . . . . . . . . . . . . . . . . . . . . . . 168 

4.2.7 Performance hints for TSM and LTO . . . . . . . . . . . . . . . . . . . . . . . 172 

4.3 TSM library sharing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 

4.3.1 Configuring the Library Manager to share libraries . . . . . . . . . . . . . 175 

4.3.2 Configuring the Library Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176 

4.3.3 Define library and drives on Library Client . . . . . . . . . . . . . . . . . . . 177 

4.3.4 Administering shared libraries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 

4.4 LAN-free data transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 

4.4.1 LAN-free client setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 

4.4.2 Mapped drive considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184 

4.4.3 Mapped drive configuration for AIX LAN-free client . . . . . . . . . . . . 185 

4.4.4 Mapped drive configuration for Solaris LAN-free client. . . . . . . . . . 185 

4.4.5 Test if data transfer goes over the SAN . . . . . . . . . . . . . . . . . . . . . 186 

Contents 5

4.5 TSM and 3584 redundant library control paths . . . . . . . . . . . . . . . . . 189 

Chapter 5. Implementing VERITAS NetBackup DataCenter with LTO . . 191 

5.1 VERITAS NetBackup DataCenter overview . . . . . . . . . . . . . . . . . . . 192 

5.2 Installing NetBackup DataCenter on Solaris . . . . . . . . . . . . . . . . . . . 193 

5.2.1 Test environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 

5.2.2 Device drivers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194 

5.2.3 Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 

5.3 Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 

5.3.1 Detect tape drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 

5.3.2 Scan for tape devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 

5.3.3 Device configuration wizard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206 

5.3.4 Importing media . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 

5.3.5 Storage units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 

5.3.6 Define backup policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 

5.4 Testing the configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 

5.5 Miscellaneous tape topics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 

5.5.1 Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 

5.5.2 Cartridge labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 

Chapter 6. Implementing Legato NetWorker with LTO. . . . . . . . . . . . . . . 235 

6.1 Legato NetWorker overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 

6.2 Legato NetWorker installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 

6.3 Installation in an AIX environment . . . . . . . . . . . . . . . . . . . . . . . . . . 237 

6.3.1 Software installation (AIX). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 

6.4 Installation in a Solaris environment . . . . . . . . . . . . . . . . . . . . . . . . . 240 

6.4.1 Software installation (Solaris) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 

6.5 Installation in a HP-UX environment . . . . . . . . . . . . . . . . . . . . . . . . . 246 

6.5.1 Software installation (HP-UX) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 

6.6 Legato NetWorker implementation and use . . . . . . . . . . . . . . . . . . . 249 

6.7 Implementation in AIX environment . . . . . . . . . . . . . . . . . . . . . . . . . 250 

6.7.1 Autochanger configuration (AIX) . . . . . . . . . . . . . . . . . . . . . . . . . . . 250 

6.8 Implementation in a Solaris environment . . . . . . . . . . . . . . . . . . . . . 253 

6.8.1 Autochanger configuration (SUN) . . . . . . . . . . . . . . . . . . . . . . . . . . 253 

6.9 Implementation in HP-UX environment . . . . . . . . . . . . . . . . . . . . . . . 257 

6.9.1 Autochanger configuration (HP-UX) . . . . . . . . . . . . . . . . . . . . . . . . 257 

6.10 Installation verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261 

6.11 Adding a new LTO device to Legato NetWorker . . . . . . . . . . . . . . . 263 

6.12 Legato NetWorker exploitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 

6.12.1 Dynamic drive sharing (DDS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265 

6.12.2 Sharing autochangers between NetWorker hosts. . . . . . . . . . . . . 267 

6.12.3 Auto media management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268 

6.12.4 Inventorying and labeling volumes with barcode . . . . . . . . . . . . . 268 


Appendix A. Useful commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 

AIX commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 

SUN Solaris commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272 

Tivoli Storage Manager tape and tape library commands . . . . . . . . . . . . . . . 273 

Legato NetWorker autochanger maintenance commands . . . . . . . . . . . . . . . 273 

Verifying proper device attachment with tapeutil . . . . . . . . . . . . . . . . . . . . . . 274 

AIX tape and medium changer device attachment test. . . . . . . . . . . . . . . 274 

Solaris and HP tape and medium changer device attachment test . . . . . 275 

Using tapeutil element inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 

Appendix B. Additional information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 

Simple Network Management Protocol (SNMP) alerts . . . . . . . . . . . . . . . 280 

Configuring SNMP for the 3583 tape library. . . . . . . . . . . . . . . . . . . . . . . . . . 281 

Configuring SNMP for the 3584 tape library. . . . . . . . . . . . . . . . . . . . . . . . . . 283 

Guidelines for booting SAN Data Gateway and FC hosts . . . . . . . . . . . . . . . 287 

Performance considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288 

Appendix C. IBM LTO Tape Library sharing and partitioning . . . . . . . . . 291 

IBM LTO tape libraries attached to multiple servers . . . . . . . . . . . . . . . . 292 

Tape Library sharing and partitioning definitions . . . . . . . . . . . . . . . . . . . . . . 292 

Sharing the 3583 Ultrium Scalable Tape Library . . . . . . . . . . . . . . . . . . . . . . 292 

Sharing the 3583 tape library using SAN and TSM . . . . . . . . . . . . . . . . . 294 

Sharing the 3583 tape library without a SAN . . . . . . . . . . . . . . . . . . . . . . 295 

Sharing the 3583 tape library using multiple initiators. . . . . . . . . . . . . . . . 295 

Sharing the 3584 Ultrium UltraScalable Tape Library . . . . . . . . . . . . . . . . . . 296 

Configuring the 3584 with multiple logical libraries . . . . . . . . . . . . . . . . . . 298 

Configuring the 3584 tape library using labels . . . . . . . . . . . . . . . . . . . . . 298 

Configuring the 3584 tape library using menus. . . . . . . . . . . . . . . . . . . . . 304 

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310 

Related publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 

IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 

Other resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313 

Referenced Web sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314 

How to get IBM Redbooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 

IBM Redbooks collections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315 

Special notices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317 

Abbreviations and acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 

Contents 7


Figures 

1-1 LTO Ultrium roadmap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 

1-2 The LTO Ultrium product family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 

1-3 3580 Ultrium Tape Autoloader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 

1-4 3581 Ultrium Tape Autoloader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 

1-5 3583 Scalable Tape Library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 

1-6 3584 Ultrium UltraScalable Tape Library . . . . . . . . . . . . . . . . . . . . . . . . 13 

1-7 Example of a 6-frame 3584 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 

1-8 3584 frame capacity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 

1-9 RS/6000 LTO support matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 

1-10 Extract from SAN SDG 2108 attachment Web site . . . . . . . . . . . . . . . . 22 

1-11 RS/6000 microcode Web site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 

1-12 Device data flow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 

1-13 SCSI connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 

2-1 System Administration Manager (SAM) main panel . . . . . . . . . . . . . . . 52 

2-2 Enabling the medium changer in the kernel configuration . . . . . . . . . . . 53 

2-3 Disabling the medium changer in the kernel configuration . . . . . . . . . . 57 

2-4 Tapeutil main menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 

2-5 Where to look for the supported HBA in the SDG matrix. . . . . . . . . . . . 60 

2-6 3583 operator panel: configuring RMU . . . . . . . . . . . . . . . . . . . . . . . . . 76 

2-7 3583 Specialist: welcome panel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 

2-8 3583 Specialist: configuration panel . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 

2-9 3583 Specialist: diagnostics file panel . . . . . . . . . . . . . . . . . . . . . . . . . . 80 

2-10 3583 Specialist: operator panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 

2-11 Change Ethernet parameters on 3584. . . . . . . . . . . . . . . . . . . . . . . . . . 82 

2-12 3854 Specialist: Welcome window. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 

2-13 3584 Specialist: Physical Library entry panel . . . . . . . . . . . . . . . . . . . . 85 

2-14 3854 Specialist: Logical Libraries drives panel . . . . . . . . . . . . . . . . . . . 86 

2-15 3584 Specialist: library settings panel . . . . . . . . . . . . . . . . . . . . . . . . . . 87 

2-16 The 3583 tape library panel with the microcode level information . . . . . 90 

2-17 Tapeutil inquiry on tape drive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 

2-18 Ultrium Tape Library Specialist: updating RMU firmware . . . . . . . . . . . 92 

3-1 IBM 2109 StorWatch - fabric view . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 

3-2 Switch Management window . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 

3-3 Performance/data transfer rate from a SAN switch/FC . . . . . . . . . . . . . 97 

3-4 Example of a connection with SDG . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 

3-5 Our lab configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 

3-6 Zoning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 

3-7 Zone to restrict the tapes to one HBA only . . . . . . . . . . . . . . . . . . . . . 101 


3-8 First zone with both HBA in the zone. . . . . . . . . . . . . . . . . . . . . . . . . . 104 

3-9 WWN of the SDG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 

3-10 Start Zone Admin panel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 

3-11 Create alias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 

3-12 Define zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 

3-13 Define zone configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 

3-14 3584 with 8 FC LTO drives, one server with 2 HBA and 2 zones . . . . 111 

3-15 Lab configuration with Sun server and 3583 . . . . . . . . . . . . . . . . . . . . 112 

3-16 Lab configuration with one 3583 and one 3570 . . . . . . . . . . . . . . . . . . 113 

3-17 One SDG is missing during boot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 

3-18 Binding with QLogic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 

3-19 Binding with Emulex. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 

3-20 Connected to the SDG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 

3-21 Install of the SDG StorWatch Specialist. . . . . . . . . . . . . . . . . . . . . . . . 130 

3-22 SDG StorWatch Specialist - server application . . . . . . . . . . . . . . . . . . 131 

3-23 SDG StorWatch connect to server. . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 

3-24 SDG StorWatch logon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 

3-25 Connect to SDG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 

3-26 Install firmware on the SDG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 

3-27 Select the SCSI option. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 

3-28 SCSI Channel parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 

3-29 Select the Fibre Channel options. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 

3-30 Fibre Channel parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 

3-31 Basic SCSI connection to a system . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 

3-32 SAN Data Gateway attached through Fibre Channel — host view . . . 138 

3-33 Select Device Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 

3-34 Device Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 

3-35 Device Mapping required a reboot . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 

3-36 Select Channel Zoning. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 

3-37 Channel Zoning settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 

4-1 Tivoli Storage Manager supported platforms . . . . . . . . . . . . . . . . . . . . 148 

4-2 TSM library sharing overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 

4-3 TSM LAN-free data transfer overview . . . . . . . . . . . . . . . . . . . . . . . . . 152 

4-4 SCSI element number of tape drives in the 3584 . . . . . . . . . . . . . . . . 156 

4-5 SCSI element number of tape drives in the 3583 . . . . . . . . . . . . . . . . 156 

4-6 Our lab library sharing environment. . . . . . . . . . . . . . . . . . . . . . . . . . . 174 

4-7 LAN-free backup lab example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 

4-8 3584 with multiple control paths enabled. . . . . . . . . . . . . . . . . . . . . . . 189 

5-1 System administration login . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207 

5-2 NetBackup DataCenter assistant. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 

5-3 System administration main window . . . . . . . . . . . . . . . . . . . . . . . . . . 209 

5-4 Media and device management. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 

5-5 Select hosts to scan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 


5-6 Display detected drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 

5-7 Display discovered devices configuration . . . . . . . . . . . . . . . . . . . . . . 211 

5-8 Updating device configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 

5-9 Configure storage units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212 

5-10 Show defined drives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213 

5-11 Setup robot inventory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214 

5-12 Media Manager with configured volumes . . . . . . . . . . . . . . . . . . . . . . 216 

5-13 Storage unit configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217 

5-14 Storage unit properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 

5-15 Add a new backup class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 

5-16 Enter backup class name. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 

5-17 Backup policy configuration wizard . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 

5-18 Select clients and OS for backup class . . . . . . . . . . . . . . . . . . . . . . . . 220 

5-19 Display client list . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 

5-20 Configure backup directories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222 

5-21 Configure backup type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223 

5-22 Configure schedule rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224 

5-23 Define backup start windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 

5-24 Display backup class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 

5-25 Backup class attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 

5-26 Initiate client backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227 

5-27 Configure client backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 

5-28 Monitor backup job details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 

5-29 End of backup job status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230 

5-30 Restoring files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 

5-31 Display drives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 

5-32 Change drive attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233 

6-1 Specify source window of the software installation GUI interface . . . . 247 

6-2 Select Legato NetWorker package . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 

6-3 Subproducts selection window of the software install dialog . . . . . . . . 248 

6-4 Install window after the successful package installation . . . . . . . . . . . 248 

6-5 Legato NetWorker administration GUI . . . . . . . . . . . . . . . . . . . . . . . . . 262 

6-6 Jukeboxes window for the LTO tape library . . . . . . . . . . . . . . . . . . . . . 263 

6-7 Configuring a new tape device. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 

6-8 Legato NetWorker Dynamic Drive Sharing (DDS) example. . . . . . . . . 266 

6-9 How Legato NetWorker autochanger sharing works . . . . . . . . . . . . . . 267 

6-10 Jukebox labeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 

6-11 Inventorying the library. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270 

B-1 3583 Storwatch Specialist SNMP MIB download panel. . . . . . . . . . . . 281 

B-2 3583 Storwatch Specialist: configuring SNMP. . . . . . . . . . . . . . . . . . . 282 

B-3 Tivoli NetView SNMP monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283 

B-4 3584 Operator Panel: SNMP settings . . . . . . . . . . . . . . . . . . . . . . . . . 284 

B-5 3584 Operator Panel: enable/disable SNMP . . . . . . . . . . . . . . . . . . . . 284 

Figures 11

B-6 3584 Operator Panel: set destination IP address . . . . . . . . . . . . . . . . 285 

B-7 3584 Operator Panel: set remote port . . . . . . . . . . . . . . . . . . . . . . . . . 286 

B-8 3584 Operator Panel: set community name . . . . . . . . . . . . . . . . . . . . 287 

C-1 3583 Ultrium Scalable Tape Library single path design. . . . . . . . . . . . 293 

C-2 3583 Ultrium Scalable Tape Library sharing with SAN and TSM . . . . 294 

C-3 3583 Ultrium Scalable Tape Library sharing without a SAN . . . . . . . . 295 

C-4 3583 Ultrium Scalable Tape Library sharing with multiple initiators. . . 296 

C-5 3584 Ultrium UltraScalable Tape Library multipath design . . . . . . . . . 297 

C-6 3854 StorWatch Specialist Logical Libraries entry panel. . . . . . . . . . . 299 

C-7 3854 StorWatch Specialist automated configuration wizard . . . . . . . . 300 

C-8 3854 StorWatch Specialist configuration review . . . . . . . . . . . . . . . . . 301 

C-9 3854 StorWatch Specialist configuration final window. . . . . . . . . . . . . 302 

C-10 3584 Operator Panel Main Menu. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303 

C-11 3854 Operator Panel Configuration Summary. . . . . . . . . . . . . . . . . . . 304 

C-12 3854 StorWatch Specialist: using menus for LTO range . . . . . . . . . . . 305 

C-13 3854 StorWatch Specialist: configure drives and slots for LTO range. 306 

C-14 3854 StorWatch Specialist: reviewing the configuration . . . . . . . . . . . 307 

C-15 3584 Operator Panel: configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 

C-16 3584 Operator Panel: set logical libraries . . . . . . . . . . . . . . . . . . . . . . 309 

C-17 3584 Operator Panel: set storage slots . . . . . . . . . . . . . . . . . . . . . . . . 309 

C-18 3584 Operator Panel: set drives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310 

C-19 Summary of the IBM LTO Tape Libraries sharing capabilities. . . . . . . 311 


Tables 

1-1 3583 SCSI and FC supported server and operating systems . . . . . . . . 17 

1-2 3583 supported ISV applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 

1-3 HVD and LVD characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 

2-1 Special device file names used with UNIX operating systems . . . . . . . 30 

2-2 AIX device software for FC HBAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 

2-3 AIX software for FC devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 

3-1 Target ID and device mapping — native SCSI . . . . . . . . . . . . . . . . . . 137 

3-2 Device map from host perspective — with SAN Data Gateway . . . . . 138 

3-3 Library with three drives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 

4-1 Configuring LAN-free backups. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 

4-2 Performance panel during a LAN-free backup. . . . . . . . . . . . . . . . . . . 188 

6-1 Legato NetWorker daemons for AIX . . . . . . . . . . . . . . . . . . . . . . . . . . 240 

6-2 Legato NetWorker daemons for Solaris. . . . . . . . . . . . . . . . . . . . . . . . 245 

6-3 Legato NetWorker daemons for HP-UX. . . . . . . . . . . . . . . . . . . . . . . . 249 

A-1 AIX commands. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 

A-2 Commonly used SUN Solaris commands . . . . . . . . . . . . . . . . . . . . . . 272 

A-3 Commonly used TSM tape commands . . . . . . . . . . . . . . . . . . . . . . . . 273 

A-4 Legato NetWorker autochangers maintenance commands. . . . . . . . . 273 



Preface 

This IBM Redbook is a follow-on to The IBM Ultrium Tape Libraries Guide, 

SG24-5946, and will help you plan, install and configure IBM Ultrium LTO tape 

drives and libraries in UNIX environments. It focuses on the setup and 

customization of these drives and libraries in both direct-attached SCSI and SAN 

configurations for the AIX, Solaris and HP-UX operating systems. 

The first part of the book describes how to attach and configure the drives and 

libraries and covers basic installation and administration. The second part 

documents how to use the hardware with popular data backup applications — 

Tivoli Storage Manager, VERITAS NetBackup DataCenter and Legato 

NetWorker. 

This redbook will help IBM personnel, Business Partners and customers to better 

understand and implement the IBM Ultrium LTO product line in UNIX 

environments. 

We assume that the reader is familiar with tape drives and libraries and 

understands basic SAN concepts and technologies. 

The team that wrote this redbook 

This redbook was produced by a team of specialists from around the world 

working at the International Technical Support Organization, San Jose Center. 

Charlotte Brooks is a Project Leader for open tape and storage management 

solutions. She has 11 years of experience with IBM in the fields of RISC 

System/6000 and Storage Management. She has written six redbooks and 

teaches IBM classes worldwide on all areas of Storage Management. Before 

joining the ITSO at the beginning of 2000, Charlotte was the Technical Support 

Manager for Tivoli Storage Management in the Asia Pacific region. 

Anthony Abete is a member of the IBM Storage Systems Group's Advanced 

Technical Support Team where he specializes in LTO, SAN/Tape and open 

systems tape performance. He has worked at IBM for 24 years. His areas of 

expertise include storage hardware, software and management. 


Marco Ceresoli is a Consulting I/T Specialist working for EMEA Advanced 

Technical Support (ATS). He has 12 years of experience in the Storage Solutions 

field. He has worked at IBM for 15 years. His areas of expertise include 

enterprise tape storage products like IBM3590, IBM3494 and Virtual Tape 

Server, as well as Storage Area Networks and open systems storage products. 

He has written several redbooks on tape storage products including 

DFSMSrmm, VTS and Peer-to-Peer VTS. 

Josef Weingand is an IT Specialist within the IBM Storage Subsystems Group in 

Germany. He has four years of experience providing technical support in IBM. He 

started as a customer engineer for Large Systems and worked 2 1/2 years as a 

Team Leader for the ITS European Support Center for Optical, Tape and Tape 

Libraries. Now he is providing technical sales support for tape and backup 

solutions. He holds a Diplom-Ingenieur (FH) degree in Electronic Engineering 

from the Fachhochschule Kempten. 

Thanks to the following people for their contributions to this project: 

Jon Tate 

International Technical Support Organization, San Jose Center 

Emma Jacobs 


Yvonne Lyon 


Nancy Coon 

LTO Marketing, IBM Tucson 

Greg Kovacs 

Tape Development, IBM Tucson 

Steve Solewin 

Tape Development, IBM Tucson 

Dan Watanabe 

Tape Product Marketing, IBM Tucson 

Jeff Zgonc 

Tape Product Marketing, IBM Tucson 

Hagen Laufersweiler 

Tape Advanced Technical Support, IBM Mainz 

Jeff Ziehm 

ATS, IBM Dallas 


Tom King 

ATS, IBM Gaithersburg 

Ron Henkhaus 

ATS, IBM Kansas City 

Tom Hepner 

ATS, IBM San Jose 

Tricia Jiang 

Tivoli Storage Manager Marketing, IBM San Jose 

JB Burke 

Jeskell, San Jose 

Akbar Syed 

Jeskell, San Jose 

Dan Raynes 

Legato Systems 

Edward Ho 

Legato Systems 

Bill Roth 

VERITAS 

George Winter 

VERITAS 

Special notice 

This publication is intended to help systems and storage administrators to plan, 

install, configure, setup, customize, and use IBM LTO Ultrium with Open 

Systems. The information in this publication is not intended as the specification of 

any programming interfaces. See the PUBLICATIONS section of the IBM 

Programming Announcement for the products we mention for more information 

about what publications are considered to be product documentation. 

Preface 17

IBM trademarks 

The following terms are trademarks of the International Business Machines 

Corporation in the United States and/or other countries: 

e (logo)® 

IBM ® 

AIX® 

AS/400® 

DFS 

DFSMSrmm 

Enterprise Storage Server 

iSeries 

Magstar® 

Netfinity® 

NetView® 

OS/390® 

OS/400® 

Redbooks 

Redbooks Logo 

Perform 

pSeries 

RS/6000® 

S/390® 

SANergy 

Sequent® 

SP 

StorWatch 

Tivoli® 

xSeries 

z/OS 

Comments welcome 

Your comments are important to us! 

We want our IBM Redbooks to be as helpful as possible. Send us your 

comments about this or other Redbooks in one of the following ways: 

► Use the online Contact us review redbook form found at: 

► 

► 

ibm.com/redbooks 

Send your comments in an Internet note to: 

redbook@us.ibm.com 

Mail your comments to the address on page ii. 


Part 1 

up LTO 

Part 1 Setting 

Ultrium 

In this part we introduce the LTO Ultrium products and describe how to set them 

up in various UNIX environments. Both native SCSI and SAN (Fibre Channel) 

attachments are presented. We also show how to use the administration tools for 

the LTO, such as the StorWatch Specialist. 



1 

Chapter 1. 

Introduction to LTO Ultrium 

with UNIX 

This chapter provides the following: 

► The development of LTO 

► An overview of the four IBM LTO models available 

► Server and OS platforms explained 

► Storage Management software 

► Connectivity 

– Direct SCSI attach 

– SAN Data Gateway attach 

– Integrated SDG attach 

– SCSI bus performance 

► HBAs and drivers discussion 

► LVD versus HVD discussion 

► HD68 versus VHDCI discussion 


1.1 LTO overview 

This section will cover the following: 

► LTO development 

► Enhancements since announcement 

► LTO model overview 

► LTO detail on individual models 

1.2 LTO development 

The Linear Tape Open (LTO) program is a joint initiative of Hewlett-Packard, IBM 

and Seagate Technology. In 1997, the three companies set out to enable the 

development of best-of-breed tape storage products by consolidating 

state-of-the-art technologies from numerous sources. The three companies also 

took steps to protect customer investment by providing a four generation 

roadmap and establishing an infrastructure to enable compatibility between 

competitive products. 

The LTO technology objective was to establish new open-format specifications 

for high capacity, high performance tape storage products for use in the midrange 

and network server computing environments, and to enable superior tape 

product options. 

LTO program cooperation goes beyond the initial three companies. LTO format 

specifications have been made available to all who want to participate through 

standard licensing provisions. LTO program technology has already attracted a 

number of other industry leaders, so that LTO specified products (tape drives and 

tape storage cartridges) will reach the market from multiple manufacturers, not 

just the Technology Provider Companies. This is critical to meeting an open 

market objective, and is accomplished through open licensing of the technology. 

Cooperation is also evident in the LTO program requirement that all products 

produced by licensees be technically certified annually. The primary objective of 

this certification is to help determine whether LTO format cartridges will be 

exchangeable across drives produced by different Ultrium manufacturers. Simply 

restated as, "LTO compliant media from any vendor can be read and written in 

LTO compliant drives from any vendor ". 

All three consortium members (IBM, HP and Seagate) are now shipping LTO 

products. This redbook will discuss only the IBM LTO Ultrium product line. 

The Linear Tape-Open organization homepage is at http://www.lto.org 


1.2.1 LTO enhancements since introduction 

Since their introduction in 2000, there have been several enhancements to the 

IBM LTO family in the areas of feature, function and connectivity. 

August 23, 2000 - Initial announcements for all products 

On August 23, 2000, the LTO family was announced. It comprised the following 

products: 

► 3580 Ultrium Tape Drive 

► 3581 Ultrium Tape Autoloader 

► 3583 Ultrium Scalable Tape Library 

► 3584 Ultrium UltraScalable Tape Library 

Included with this announcement, there were already several previews of future 

function planned for the LTO product line. The IBM definition of preview is: 

Previews provide insight into IBM plans and direction. Specific availability 

dates, ordering information and terms and conditions will be provided when the 

product is announced. 

Preview for 3580, 3581 & 3583 

► Fibre Channel attachment through SAN Data Gateway 

► Linux operating system support 

Preview for 3584 

► Fibre Channel — Arbitrated Loop (FC-AL) attachment 

► Expanded Input/Output Station 

► Support for Digital Linear Tape (DLT) technology 

► Remote management and support features 

► Certification in SAN solutions including LAN-free backup 

► LINUX operating system support for open systems servers 

February 20, 2001 

3584 announcement information 

► Native FC-AL attach 

► SAN Lan free backup capability 

3584 preview 

► Linux server attachment 

Chapter 1. Introduction to LTO Ultrium with UNIX 5

► 

► 

HP-UX Fibre Channel attachment 

Fibre Channel attach via Emulex family of adapters 

June 12, 2001 

3580 announcement 

► Rack Mount 

► Linux Support 


► Linux Support 


► Native Fibre Channel via SAN Data Gateway Integrated Module 

► Remote Management Unit (RMU) 

► StorWatch Specialist 

3583 preview 

► SCSI Extended copy command (data mover function) 

► Attachment to McDATA and INRANGE switches 

► Attachment to Linux servers 

► HP-UX Fibre Channel attachment 

► 2 Gigabit Fibre Channel attachment via Emulex and QLogic adapters 


► StorWatch Specialist 

► DLT support via D42 frame 

► Optional 20 LTO or 18 DLT cartridge I/O Station 

► Hot swappable drives (LVD, HVD and FC-AL) 

► Redundant Drive power supplies 

June 26, 2001 


► Non hot swappable, sled mounted drives (LVD, HVD and FC-AL) 

For more information consult The IBM LTO Ultrium Tape Libraries Guide, 

SG24-5946. 


The IBM LTO homepage is located at: 

http://www.ibm.com/storage/lto 

LTO Ultrium Road Map 

Generation 1 Generation 2 Generation 3 Generation 4 

Capacity 

(Native) 

100GB 200GB 400GB 800GB 

Transfer Rate 

(Native) 

10-20MB/s 20-40MB/s 40-80MB/s 80-160MB/s 

Media Metal Particle Metal Particle Metal Particle Thin Film 

Figure 1-1 LTO Ultrium roadmap 

Important: Hewlett-Packard, IBM, and Seagate reserve the right to change 

the information in this migration path without notice. 

1.2.2 LTO Ultrium models 

There are four IBM LTO Ultrium models available. They are the 3580 Ultrium 

Tape Drive, the 3581 Ultrium Tape Autoloader, the 3583 Ultrium Scalable Tape 

library, and the 3584 UltraScalable tape library. 

Ultrium tape capacity is 100 GB per cartridge in native format, and using (2:1) 

compression is 200 GB. The Ultrium drives have sustained data rates of 15 MB/s 

native and 30 MB/s at 2:1 compression. 

Note that some suppliers may offer alternative cartridge capacities either as 

standard or on demand. Check with your supplier for details. 


Figure 1-2 The LTO Ultrium product family 

1.2.3 IBM 3580 Ultrium Tape Drive 

The IBM 3580 Ultrium Tape Drive (abbreviated to 3580), is an external, 

stand-alone, SCSI-attached tape drive that attaches to iSeries, pSeries, xSeries, 

Intel, AS/400, RS/6000, Netfinity, RS/6000 SP, and other UNIX and PC servers 

supporting OS/400, IBM AIX, Sun Solaris, HP-UX, and Microsoft Windows NT 

open systems using a suitable SCSI adapter. 

The IBM 3580 Ultrium tape drive can also connect to Fibre Channel server host 

bus adapters through the IBM SAN Data Gateway Routers models 2108-R03 or 

2108-G07. 

The IBM 3580 is available as two separate model types, depending on which 

type of SCSI interface is required. 


The two model types are: 

► IBM 3580 -L11 has a Low-Voltage Differential (LVD) Ultra2 SCSI attachment 

that connects to LVD fast/wide adapters. 

► IBM 3580 -H11 has a High-Voltage Differential (HVD) Ultra SCSI attachment 

that connects to HVD fast/wide adapters. 

Figure 1-3 is a picture of the 3580 Ultrium Tape Drive. 

Figure 1-3 3580 Ultrium Tape Autoloader 

1.2.4 IBM 3581 Ultrium Tape Autoloader 

The 3581 Ultrium Tape Autoloader (abbreviated to 3581) is an external, 

stand-alone or rack-mounted autoloader that incorporates an IBM Ultrium Tape 

Drive. It attaches to iSeries, pSeries, xSeries, Intel, AS/400, RS/6000, Netfinity, 

RS/6000 SP, and other UNIX and PC servers supporting OS/400, IBM AIX, Sun 

Solaris, HP-UX, and Microsoft Windows NT open systems using a suitable SCSI 

adapter. 

The 3581 Ultrium Tape Autoloader capacity is seven tape cartridges, providing a 

media capacity of up to 700 GB (1.4 TB with 2:1 compression) data storage per 

library and a sustained data rate of up to 15 MB per second (uncompressed). 

The two model types are: 

► IBM 3581-L17 has a Low-Voltage Differential (LVD) Ultra2 SCSI attachment 

that connects to LVD fast/wide adapters. 


► 

IBM 3581-H17 has a High-Voltage Differential (HVD) Ultra SCSI attachment 

that connects to HVD fast/wide adapters. 

Figure 1-4 is a picture of the 3581 Ultrium Tape Autoloader. 

Figure 1-4 3581 Ultrium Tape Autoloader 

1.2.5 IBM 3583 Ultrium Scalable Tape Library 

The IBM 3583 Ultrium Scalable Tape Library (abbreviated to 3583) is a 

high-performance, reliable, scalable tape subsystem. Designed for tape 

automation, the IBM 3583 Scalable Tape Library can be attached to iSeries, 

pSeries, xSeries, Intel, AS/400, RS/6000, Netfinity, RS/6000 SP, and other UNIX 

and PC servers supporting OS/400, IBM AIX, Sun Solaris, HP-UX, and Microsoft 

Windows NT open systems using SCSI attachment or Fibre Channel attachment. 

The libraries use the IBM Ultrium Tape Drives for faster data transfer and 

reliability in automated library service. Each aspect of the library subsystem has 

been designed for repeated, reliable unattended tape handling. 

The tape handling mechanism is designed to reliably move cartridges to IBM 

Ultrium Tape Drives within the library. The IBM Ultrium tape cartridges have been 

refined using the many years of IBM tape experience to provide a tape cartridge 

that provides fast data transfer, but stands up to the rigors of automated handling. 


The Tape capacity is 100 GB per cartridge in native format, and using (2:1) 

compression is 200 GB. The Ultrium drives have maximum sustained data rates 

of 15 MB/s native and 30 MB/s at 2:1 compression. Of course, the real achieved 

performance depends on many other system variables. 

The tape library is designed for easy expansion. It can accommodate from one to 

six tape drives and from 18 to 72 cartridges. There are three Library Models: 

L18, L36, and L72. The cartridge capacity of the Model L18 is 18 cartridges, the 

Model L36 is 36 cartridges, and the Model 72 is 72 cartridges. Models 18 and 36 

can be field upgraded to hold 72 cartridges. Cartridge expansion is in increments 

of 18 cartridges. One IBM Ultrium drive is required and five additional drives are 

available for factory or field installation. Upgrade features are modular units that 

are easy to install. If you are NOT using the integrated SAN Data Gateway, then 

direct attached SCSI drives may be any mixture of LVD or HVD up to a total of 

six. The tape library input/output (I/O) station enables cartridges to be inserted 

and removed without disrupting library operation. There are two I/O station 

options: a single slot option and a 12 slot option. The 12 slot I/O station is 

required to achieve the 72 cartridge maximum library configuration. With the 12 

slot I/O station feature the library can be configured as 72 storage slots, or 60 

storage slots and 12 I/O slots. 

The Remote Management Unit (RMU) comes standard in every library shipped 

after July 27, 2001. The unit can be added to pre-existing libraries by ordering the 

chargeable upgrade Feature Code 1660. The RMU provides an Ethernet port, so 

that the library can be configured as a TCP/IP device on the network. Library 

status can be sent to the network as Simple Network Management Protocol 

(SNMP) traps. The StorWatch Specialist enables network access (via Web 

browser) to the library for more detailed status and control. All library operator 

panel functions can be accessed using the StorWatch Specialist. We provide 

detailed information on the StorWatch Tape Library Specialist for the 3583 in 

2.9.1, “IBM 3583 StorWatch Specialist” on page 76. 

The SAN Data Gateway Integrated Module is another available, chargeable 

library feature. The gateway provides the ability for SCSI devices in the library to 

connect into a SAN infrastructure. With 2 Gigabit port speeds, this library is ready 

for the next generation of 2 Gigabit Fibre Channel devices. It can also run at 1 

Gigabit port speeds if attached to 1 Gigabit Fibre Channel devices. Two Fibre 

Channel ports make multiple attachments easy and support failover redundancy. 

The Gateway has two Fibre Channel interfaces or ports and four SCSI interfaces 

or ports. The Fibre Channel ports are equipped with 2 Gigabit, SC-style 

shortwave multimode Gigabit Interface Converter (GBIC) modules. The Fibre 

Channel ports are capable of communicating reliably at distances of up to 300 m 

over 50 µm multimode-optical fiber cables. The four SCSI ports are Ultra2, low 

voltage differential (LVD), with VHDCI-style connectors (Very High Density Cable 

Interconnect), so be sure to order the LVD version of the LTO drive. 


You can specify the SDG feature when ordering a new 3583 or order it as an 

upgrade to an existing model. You cannot install an existing separate SAN Data 

Gateway (2108) into the library. However, it may be used standalone to provide 

FC attachment. 

Note: Only shortwave (SW) GBICs and LVD drives are supported when using 

the Integrated SAN Data Gateway module. 

Figure 1-5 is a picture of the 3583 Ultrium Scalable tape library. 

Figure 1-5 3583 Scalable Tape Library 

1.2.6 IBM 3584 Ultrium UltraScalable Tape Library 

The IBM 3584 UltraScalable Tape Library (abbreviated to 3584), as shown in 

Figure 1-6, is designed for the large, unattended storage requirements from 

today's mid-range systems up to high-end open systems. Each aspect of the 

subsystem is designed to optimize access to data and reliability. The 3584 

UltraScalable Tape Library is designed to connect to host systems using any 

combination of Fibre Channel, Ultra2/Wide Low Voltage Differential (LVD) SCSI, 

or Ultra/Wide High Voltage Differential (HVD) SCSI interfaces. 


Figure 1-6 3584 Ultrium UltraScalable Tape Library 

All 3584 libraries shipped after August 31, 2001 come equipped with standard no 

charge Feature Code 9660 which provides the ability to connect the library to a 

10/100 Ethernet LAN. The unit can be added to pre-existing libraries by ordering 

the chargeable upgrade Feature Code 1660. The StorWatch Specialist enables 

network access (via Web browser) to the library for more detailed status and 

control. All library operator panel functions can be accessed using the StorWatch 

Specialist. We provide detailed information on the StorWatch Tape Library 

Specialist for the 3584 in 2.9.2, “IBM 3584 StorWatch Specialist” on page 81. 

Base Unit 3584-L32 

The IBM 3584 base library, the Model L32, has 141 to 281 cartridge slots and 

support for up to twelve IBM LTO Ultrium tape drives with an incremental 

reduction of storage slots for more than four drives. Data capacity for the Model 

L32 is 14 to 28 TB native and 28 to 56 TB using LTO-DC (LTO Data 

Compression) (2:1) compression. The aggregate sustained data rate is from 108 

GB to 1.3 TB/hour with 2:1 compression. Up to twelve logical libraries and/or up 

to twelve control paths can be configured for each L32 frame. Each 3584-L32 

library has a standard 10-slot cartridge input/output station for importing or 

exporting cartridges from the library without requiring a re-inventory. 


Alternatively, a 30 I/O station can be selected. For bulk-loading of IBM LTO 

Ultrium tape cartridges the library door can be opened. Each time the library 

door is closed, a bar code reader mounted on the autochanger scans the 

cartridge labels enabling a re-inventory of the cartridges in the library frame in 

less than 60 seconds. A door lock is included to restrict physical access to 

cartridges in the library. Customers can expand library capacity and number of 

drives to meet their changing needs. 

The base 3584-L32 can be expanded by adding up to five additional frames - 

D32, D42 or a combination of the two. 

LTO expansion frame 3584-D32 

Up to five expansion frames, Model 3584-D32, may be added to the base frame 

(Model 3584-L32) to add storage and/or drive capacity. Each 3584-D32 frame 

supports up to 440 storage slots and up to twelve drives, with incremental 

reduction of storage slots for each set of four of drives installed. A fully configured 

3584 with one Model L32 frame and five Model D32 frames will support up to 

seventy-two drives with an aggregate data rate of 7.8 TB per hour (2:1 

compression), or up to 2481 storage slots with a total capacity of 248.1 TB native 

or 496.2 TB with 2:1 compression. Each frame can have up to 12 logical libraries 

and 12 control paths due to the multipath capability of the 3584. 

Figure 1-7 Example of a 6-frame 3584 


DLT expansion frame 3584-D42 

Each 3584 Model D42 expansion frame contains from 1 to 12 DLT 8000 tape 

drives, each having a native data transfer rate of 6 MB/s and a cartridge capacity 

of 40 GB. Using data compression, the DLT 8000 drives have an effective data 

rate of up to 12 MB/s and a cartridge capacity of up to 80 GB (with 2:1 

compression) on DLTtape IV media. Each 3584-D42 frame supports up to 360 

storage slots and up to twelve drives, with incremental reduction of storage slots 

for each set of four of drives installed. A 3584 UltraScalable Tape Library with five 

D42 frames has a capacity of up to 1784 DLTtape slots and 229 LTO Ultrium 

slots, resulting in a maximum native data capacity of 22.9 TB on LTO Ultrium 

media and 71.4 TB on DLT tape. Each frame can have up to six logical libraries 

and six control paths due to the multipath capability of the 3584. 

Figure 1-8 shows the different slot and drive capacities contained in each type of 

3584 frame. 

Drives in Model 

L32 

Frame 

0-4 

5-8 

9-12 


D32 Frame 

0 

1-4 

5-8 

9-12 


D42 Frame 

0 

1-4 

5-8 

9-12 

Slots in Model L32 Frame 

(without Capacity Expansion 

Feature 

Native Comp. 

Quantity 

141 

113 

87 

Capacity 

(in TB) 

14.1 

11.3 

8.7 

Quantity 

440 

423 

409 

396 

Quantity 

360 

346 

333 

324 

Figure 1-8 3584 frame capacity 

Capacity 

(in TB) 

28.2 

22.6 

17.4 


(with Capacity Expansion 

Feature and 30 I/O Slots) 

Native Comp. 

Quantity Capacity Capacity 

(in TB) (in TB) 

229 22.9 45.8 

201 20.1 40.2 

175 17.5 35.0 

Slots in Model D32 Frame 

Native Capacity (in TB) 

44.0 

42.3 

40.9 

39.6 

Slots in Model D42 Frame 

Native Capacity (in TB) 

14.4 

13.8 

13.3 

12.9 


(with Capacity Expansion 

Feature and 10 I/O Slots) 

Native Comp. 

Quantity Capacity Capacity 

(in TB) (in TB) 

281 28.1 56.2 

253 25.3 50.6 

227 22.7 45.4 

Compressed Capacity (in TB) 

88.0 

84.6 

81.8 

79.2 

Compressed Capacity (in TB) 

28.8 

27.6 

26.6 

25.8 


Note: The quantity depends on whether the Capacity Expansion Feature is 

installed, whether the upper and lower I/O stations are used, and the number 

of drives installed in a frame. 

1.3 Hardware and operating system platforms 

This section will document the hardware platforms, and the operating systems 

platforms that support the LTO family. 

1.3.1 Hardware server platforms 

The LTO products are supported on the following hardware server platforms: 

► IBM iSeries, AS/400 

► IBM pSeries, RS/6000, SP 

► IBM xSeries and other, Intel Based (Netfinity, Windows NT, Windows 2000) 

► Sun SPARC 

► Hewlett-Packard 

► Linux 

► And other open systems using SCSI attachment or Fibre Channel attachment 

For specifics of which Ultrium models support which attachment, see the 

following Web sites. 

http://www.storage.ibm.com/hardsoft/tape/3580/3580opn.html 




Attention: Always check the Web site for the most current detailed and 

accurate information. 

As an extract from the supported Web sites, Table 1-1 shows the supported 

attachments for the 3583 Ultrium Scalable Tape Library on AIX, Solaris and 

HP-UX. There are a number of provisos associated with various combinations 

and HBAs this support which are listed on the Web site. 


Table 1-1 3583 SCSI and FC supported server and operating systems 

Servers Operating Systems Host Adapters Fabric Support 

HP systems 

HP 9000 Series 

A, L, N Class 

HP-UX 

11.0 

HP 

HP4800A HVD Ult 

HP5149A LVD Ult2 

HP5150A LVD Ult2 

Previewed 

pSeries and RS/6000 

Servers that support the 

listed host bus adapters 

AIX 

4.3.2 and higher 


IBM 

FC 6204 HVD Ult 

FC 6207 HVD Ult 

FC 6205 LVD Ult2 

Integrated LVD 


2108 R03 

FC 2840 (LVD) 

FC 2830 (HVD) 

2108 G07 

FC 2214 (HVD) 

pSeries and RS/6000 

with SAN data Gateway Module Feature 8005 


listed host adapters 

AIX 



FC 6227 1 

FC 6228 1 


2109 Models S08 and 

S16 

(1) These Gigabit Fibre Channel Adapters are only supported by the LTO Ultrium Fibre Drive feature 8005 

in 3583 Scalable Tape Library, and requires a minimum of AIX 4.3 

Sun Systems 

Sun SPARC, UltraSparc, 

and Ultra/Enterprise 



Solaris 

2.6, 7 and 8 

Sun 

x1065A HVD Ult (Sbus) 

X6541A HVD Ult (PCI) 


2108 R03 

FC 2840 (LVD) 

FC 2830 (HVD) 

2108 G07 

FC 2214 (HVD 

Sun Systems 

with SAN data Gateway Module Feature 8005 

Sun SPARC, UltraSparc, 

and Ultra/Enterprise 



Solaris 

2.6, 7 and 8 

Qlogic QLA2200F 1 

Emulex LP8000 

Emulex LP8000S 


2109 Models S08 and 

S16 

(1) The Qlogic adapter is only supported by the LTO Ultrium Fabric Drive feature 

8005 in the 3583 scalable tape library. 


1.3.2 ISV storage management software 

Although operating systems provide utilities such as dd, tar and cpio to perform 

basic read/write operations for tape drives, in most cases a particular storage 

management software package is used. These packages provided specific 

functions and much more sophisticated capabilities than the native operating 

system utilities. They are available from companies known as ISVs or 

Independent Software Vendors. We discuss some of the most popular 

applications for UNIX in Part 2, “LTO Ultrium with applications” on page 145. 

IBM publishes information about ISV storage management applications which 

are certified with LTO Ultrium devices on which operating system platforms. 

These URLs are: 

http://www.storage.ibm.com/hardsoft/tape/conntrix/isv3580.html 




You should also check with the ISV for support information. 

Table 1-2 is an excerpt from the 3583 Ultrium Scalable Tape Library Independent 

Software Vendor (ISV) Matrix Chart, however always check the Web site for the 

most up to date information: 

Table 1-2 3583 supported ISV applications 

Software Application 

Version Level 

(Vendor) 

OS Platform and Level 

Channel 

Connection 

Device 

Driver 

Native AIX AIX 4.3.3 SCSI IBM 

Tivoli Storage 

Manager 

V 4.1.1.0, 4.1.1.1(IBM) 

AIX 4.3.3 SCSI, Fibre IBM 

Solaris 2.6, 2.7, 2.8 SCSI, Fibre IBM 

HP-UX 11.0 SCSI, Fibre IBM 

NetWorker 

6.1 

(Legato) 

AIX 4.3.3 

Solaris 2.6, 2.7, 2.8 

SCSI 

SCSI 


Software Application 

Version Level 

(Vendor) 

OS Platform and Level 

Channel 

Connection 

Device 

Driver 

VERITAS NetBackup 

DataCenter 3.4.1 

AIX 4.3.3 

Solaris 2.6, 2.7, 2.8 

SCSI 

SCSI 

HP-UX 10.2, 11.0, 

11.11 

SCSI 


Attention: This table is an example only and was accurate on the day it was 

created. Always check the Web site referenced for the most current up-to-date 

information. 

1.4 Connectivity 

In this section we’ll cover methods of connecting LTO models to servers, using 

the 3583 as an example. 

1.4.1 SCSI direct attach 

Figure 1-9 is an example of the IBM Web site that details server, operating 

system levels, host bus adapters and fabric support for the 3583 in the RS/6000 

environment. We gave the URLs for each LTO model in 1.3.1, “Hardware server 

platforms” on page 16. 


Figure 1-9 RS/6000 LTO support matrix 

Using this figure as an example, the following is true. In order to attach a 3583 to 

a RS/6000, you need to be running 4.3.2 or 4.3.3 or higher of the operating 

system. Let’s assume for our example that you want to SCSI direct attach the 

3583. The matrix reflects that there is a choice of three SCSI HBAs that you can 

order via feature code. Feature codes 6204 and 6207 are for a HVD interface and 

feature 6205 is the LVD version. Also note that some models of the RS/6000 

have an integrated LVD adapter which is also supported. Let’s assume that you 

decide on the feature number 6205 LVD adapter due to the 80 MB/s data rate 


and the fact that it’s a dual port and you’re running short on PCI slots on your 

RS/6000 model 7017-S80. Due to the number of RS/6000 models, the next step 

is to check if that adapter feature number 6205 is supported on your RS/6000 

model. 

A method to determine if an adapter is supported on your model of RS/6000 is to 

check the IBM announcement letters or the IBM Sales manual. These are 

excellent sources of information. You can find them at the following URL: 

http://www.ibmlink.ibm.com/ 

Select your country or region, then Announcements, then Announcements 

search. You can search from here, using the following keywords: 

3583 or 7017-S80 

The data in Example 1-1 is from the 3583 announcement letter 100-238, dated 

August 23, 2000. This example reflects the models supported with that adapter, 

and as you can see, feature number 6205 PCI Dual Channel Ultra2 SCSI 

Adapter is supported on the 7017-S80. 

Example 1-1 RS/6000 announcement material 

Feature number 6205 PCI Dual Channel Ultra2 SCSI Adapter (LVD/SE with VHDCI 

connector) (requires AIX 4.3.3, or later). This adapter is used in the 

following RS/6000 models: 

7013-S7A- 7015-S7A 

7017-S7A, 7017-S80 

7025-F40, 7025-F50, 7025-F80, 7025-H70 

7026-H50, 7026-H70, 7026-H80, 7026-M80 

7043-150, 7043-240, 7043-260, 7043-270 

7044-170, 7044-270 

7046-B50 

1.4.2 SAN Data Gateway attach 

Still using Figure 1-9 on page 20 as an example, suppose you want to attach the 

3583 with its six drives more than 12 meters away, which means that now you’re 

looking at attaching via fibre. Since the cost of the 2108 R03 is less than the G07, 

and since the G07 doesn’t support LVD, we’ll concentrate on the R03. You can 

see from the example that you’ll need feature number 2840 on the 2108 R03, but 

what adapter will you need? For that, let’s look at Figure 1-10 which is a part of 

the table you will see at the following Web site. We are showing just the piece 

relating to attaching to pSeries and RS/6000 servers. 

http://www.storage.ibm.com/hardsoft/products/tape/ro3superserver.htm 


Figure 1-10 Extract from SAN SDG 2108 attachment Web site 

Figure 1-10 indicates that with feature number 6227 on your RS/6000, you can 

attach the LVD version of the 3583 through the 2108-R03. Alternatively, you can 

add a switch or director therefore extending the distance out to 11 km. Do you 

need microcode for this adapter? For this information, you need to go the 

RS/6000 support pages. The direct link is shown in Figure 1-11 which is found on 

the following Web site: 

http://www.rs6000.ibm.com/support/micro/download.html#adapter 

Figure 1-11 RS/6000 microcode Web site 


Using this as an example, under the Files to Read/Download heading, select 

Description. 

This will take you to a link for the 6227 adapter (specifically, 

http://www.rs6000.ibm.com/support/micro/df1000f7.html). This site gives details 

of: 

► Microcode overview 

► New microcode availability 

► How to determine your adapter microcode level 

► How to obtain and update the adapter microcode 

We give more information on using SCSI and SAN adapters to attach the LTO 

drives and libraries in Chapter 2, “Basic LTO setup for UNIX systems” on page 29 

and Chapter 3, “SAN setup” on page 93. 

1.4.3 The Integrated SAN Data Gateway Module 

The SAN Data Gateway Module, feature code 8005, is an optional feature 

available for the 3583 Ultrium Scalable Tape Library. If you have this feature 

installed, then the drives are internally converted from SCSI to FC, enabling them 

to be connected into a SAN without an external gateway. 

Refer again to Figure 1-9 on page 20. What if the 3583 has feature number 8005, 

the integrated SAN data gateway module? How do you attach your RS/6000 

now? Looking at the Web site you’ll see that it calls for either of two adapters, 

feature code 6227 or feature 6228. To confirm microcode levels for 6228 use the 

same procedure that you used for feature number 6227. Go to the RS/6000 

microcode Web site as shown in Figure 1-11. 

1.4.4 Determining the number of drives on a SCSI bus 

How do you determine how many SCSI drives to place on a host adapter? There 

are several factors to take into consideration. Here are a few: 

► Tape drive speed capability 

► 

► 

► 

► 

Disk drive speed capability 

Application 

– Application needs 

– Application characteristics - read/write mix, amount of data transferred, 

streaming or stop/start, data block size 

Adapter slot availability and capability 

Cost of more adapters 


To ensure best performance, if possible, avoid daisy-chaining (connecting more 

than one drive to an adapter). If cost or slot availability considerations make 

daisy-chaining unavoidable, then connect no more than two drives per HVD and 

two to four drives per LVD. Another situation where daisy-chaining may be 

required is when connecting a multi-drive library through a SAN Data Gateway. 

The SDG has a total of four ports and one is recommended for use by the SCSI 

medium changer. Therefore, if more than three drives are installed in the library, 

daisy-chaining will be required. You can see a sample configuration in 3.1.2, 

“Connecting tape drives to a SDG” on page 97. The type of application is also 

very significant. If you are mainly doing smaller volume reads or writes, then 

more drives can be attached without saturating the adapter. A more typical use of 

tape is in high-volume backup, usually writing large blocks of data (for example, 

backing up large database files). In this scenario, a single drive per adapter will 

give the best performance. 

1.5 Host bus adapters and device drivers 

This section will cover the following: 

► What are HBAs? 

► Why are they needed, and which one do you need? 

► Are drivers needed for HBAs? 

Restriction: At this time, sharing a HBA with Disk and Tape is NOT 

recommended. In many instances, the microcode or device drivers HBAs 

required to support the different devices makes this impossible in any case. In 

addition, disk and tape usually generate very different types of I/O accesses, 

and using the same HBA will likely give poor performance. You should check 

with your local support organization for details of specific recommended and 

supported combinations. 

What are HBAs? 

HBA stands for host bus adapter. An HBA is an I/O adapter (or card) that sits 

between the host computer's bus and the SCSI or Fibre Channel loop and 

manages the transfer of information between the two channels. In order to 

minimize the impact on host processor performance, the host bus adapter 

performs many low-level interface functions automatically or with minimal 

processor involvement. 


A p p l i c a t i o n 

P r o g r a m 

( s o f t w a r e ) 

L T O d e v i c e 

d r i v e r 


H B A d e v i c e 

d r i v e r 


H B A d e v i c e 

a d a p t e r 

( h a r d w a r e ) 

L T O t a p e 

d e v i c e 

What are device drivers? 

A device driver is a program that controls a device. Every device, whether it be a 

printer, disk drive, or keyboard, must have a driver program. Many drivers, such 

as the keyboard driver, come with the operating system. For other devices, you 

may need to load a new driver either from the operating system CD or provided 

by the manufacturer when you first install the device on your computer. 

A driver acts like a translator between the device and programs that use the 

device. Each device has its own set of specialized commands that only its driver 

knows. In contrast, most programs access devices by using generic commands, 

that is, Read, Write, Put, Get. The driver, therefore, accepts generic commands 

from a program and then translates them into specialized commands for the 

device. 

Figure 1-12 is an example of the data flow between an application program and a 

LTO tape device. Note the different pieces of code or microcode involved. 

Note that the HBA device driver is not always a separate piece of software; 

sometimes it’s part of the standard operating system. This is true for some SCSI 

adapters — you should check the documentation provided with the card to see if 

this applies. For Fibre Channel HBA’s, extra software will almost always need to 

be installed. 

Figure 1-12 Device data flow 

Data flow 

The simplified list below reflects the components involved in the data path for 

moving data at a file level between disk storage devices and tape. 

Problem or performance analysis must be approached by determining which 

component of the data path impacts performance. 

► Disk device 

► SCSI device adapter 

► Adapter device driver 

► Disk device driver 

► Logical Volume Manager 

► File system 

► Application program 

► Atape device driver - tape driver 

► HBA adapter device driver 

► HBA (SCSI or FC-AL) device adapter 


► 

Tape device 

IBM device driver overview 

The IBM Ultrium tape and medium changer device drivers are designed 

specifically to take advantage of the features provided by the IBM Ultrium tape 

drives and medium changer devices. The goal is to give applications access to 

the functions required for basic tape functions (such as, backup and restore) and 

medium changer operations (such as, cartridge mount and dismount), as well as 

to the advanced functions needed by full tape management systems. Whenever 

possible, the driver will take advantage of the device features transparently to the 

application. 

1.6 LVD versus HVD 

What is it? Why choose one over the other? Why should you care? 

1.6.1 What is LVD and HVD? 

LTO tape drives have two types of SCSI attachment interfaces, LVD (Low Voltage 

Differential), and HVD (High Voltage Differential). LVD and HVD, as the names 

imply, use different voltage levels while sending the signal down the SCSI bus. 

LVD uses 3.3 Volts DC and HVD uses 5 volts DC. LTO tape drives are either LVD 

or HVD and cannot be upgraded or converted from one to the other. In order to 

change from LVD to HVD, a LVD drive would need to be replaced with a HVD 

drive. Therefore, when configuring an LTO product, consider carefully whether 

you should choose HVD or LVD drives. 

1.6.2 Why is this important? 

Why choose one over the other and why should you care? The most overriding 

reason is of connectivity. If you to try to plug in an HVD LTO drive into an LVD 

adapter, it will not work. Placing an LVD device on an HVD bus will cause the 

LVD device to shut down. Placing an HVD device on an LVD bus will cause the 

entire LVD bus to shut down. Therefore, check to see what type of adapter you 

have available in the hosts which will be attaching and choose the correct model 

of drive accordingly. 

There are also differences in the speed characteristics of HVD versus LVD. LVD 

is available either as Ultra2 SCSI which has an 80 MB/s speed or the new 

Ultra160 SCSI at 160 MB/s, while HVD is Ultra SCSI which has a maximum 

speed of 40 MBs/sec. Does this mean that the LTO drive will run faster with LVD 

versus HVD? Not necessarily. The speed seen here is the SCSI bandwidth or 

bus speed. All Ultrium drives (whether LVD, HVD or native Fibre Channel) have a 


nominal speed of 15 MB/second. If the speed of the bus were the only limiting 

factor in your configuration, then yes, switching to a faster bus should improve 

your performance. Usually, the speed capability of the bus will determine the 

number of LTO drives that can be attached on the same bus. But again, this isn’t 

the only factor in the equation. 

In addition to the speed differences, there are connection restrictions associated 

with each. Both HVD and LVD can span 25 meters in a point to point 

configuration. In a multidrop (daisy chain) configuration HVD can span 25 

meters, while LVD is limited to 12 meters. 

So which should you choose? 

If you have a situation where the 12 meter multidrop limitation of LVD isn't an 

issue, then you would purchase LVD, right? You would think so, but the correct 

answer is, “it depends”. It depends on the HBA that you have installed in your 

server. If you have an existing adapter and you don’t want to change it, then 

whether the adapter is HVD or LVD will dictate which LTO drive you order and the 

configuration capabilities. 

What if you don’t already have an adapter. Order LVD, right? Again, not 

necessarily. At this time, not all supported SCSI HBAs are available in both 

interface types. For example, AS/400 and SUN support only HVD drives when 

directly SCSI attached. See the following URL for current information: 

http://ssddom01.storage.ibm.com/techsup/swtechsup.nsf/support/ultriumcompinfo 

Table 1-3 HVD and LVD characteristics 

HVD 

LVD 

Point to Point 25 meters 25 meters 

Multidrop 25 meters 12 meters 

Speed 40 MB/s 80 or 160 MB/s 

1.7 SCSI HD68 and VHDCI cable connectors 

There are three types of SCSI cables, HD68-HD68, HD68-VHDCI and 

VHDCI-VHDCI. The HD68 connector is the normal 68 pin SCSI connector. The 

VHDCI (Very High Density) connector is a mini SCSI connector, about half the 

width of the HD68 connector. The server SCSI adapter can be either HD68 or 

VHDCI. For RS/6000, AS/400, and HP servers, LVD SCSI adapters are always 

VHDCI, and HVD adapters are always HD68. But Sun and Netfinity/NT servers 

have some adapters that are reversed. The connector type for a given adapter 


can be found in either of two ways. You can look at the adapter and the plug will 

be either about 3 cm wide (VHDCI) or 6cm wide (HD68). The other method is to 

look up the adapter characteristics to determine the plug type. One source of this 

information is the 3584 planning and operator’s guide available online at: 

http://www.storage.ibm.com/hardsoft/tape/pubs/pubs3584.html 

Prior to the June 12, 2001 announcement, all the LTO tape drives used HD68 

connectors. The new 3584 hot swappable canister LTO drive uses the VHDCI 

connectors. Check to make sure you order the correct cable with the right end 

plug to attach to your HBA. 

VHDCI 

HD68 

Figure 1-13 SCSI connectors 


2 

Chapter 2. 

Basic LTO setup for UNIX 

systems 

In this chapter we give you detailed information on the basic setup of the LTO 

drives and libraries, including: 

► 

► 

► 

► 

Device and library drivers installation and update 

– AIX environment 

– Solaris environment 

– HP-UX environment 

Host Bus Adapter drivers installation and update 


– Sun Solaris environment 

IBM StorWatch Specialist for the Ultrium Tape Libraries 

– 3583 StorWatch Specialist 

– 3584 StorWatch Specialist 

Device, library and RMU firmware upgrade 

At the end of this chapter you should be able to understand and implement the 

tasks required to: 

► Install the required device drivers for the platform you are using 

► Install and configure the required Host Bus Adapter drivers 

► Understand the use of the StorWatch Specialist product 

► Upgrade the library and drive microcode 


2.1 Installing library and device drivers 

After physically installing and connecting your LTO drive or library to your 

environment, the next step in the basic implementation of the LTO library and 

drives, is to install the driver. 

The IBM Ultrium tape and medium changer device drivers are designed 

specifically to take advantage of the features provided by the IBM Ultrium tape 

drives and medium changer devices. 

The objective is to give applications access to the functions required for basic 

tape operations (such as, backup and restore) and medium changer operations 

(such as, cartridge mount and dismount), as well as to the advanced functions 

needed by full tape management systems. 

Note: It may be necessary, for some application software, to install and use 

their own device driver in order to be able to access the library and drives. 

Follow the specific implementation instructions for the software that you are 

installing to determine if the IBM-supplied device drivers are suitable for your 

installation. 

Later in this book, in Part 2, “LTO Ultrium with applications” on page 145, we 

describe the LTO implementation for some of the most common application 

software in UNIX environments. 

For more information and specific installation instructions for every platform, refer 

to IBM Ultrium Device Drivers: Installation and User’s Guide, GA32-0430. 

The device drivers allow the operating system and the application software to 

manage tape devices and medium changer to automate the use of the tape 

media. The tape devices and medium changer are managed using a special 

device file name, that is the name used to address each tape operation (mount, 

dismount, write, read) to the required device or robotic changer. 

This special device file name, is specific for each operating system platform. In 

Table 2-1 you find the most common used names for the UNIX platforms. 

Table 2-1 Special device file names used with UNIX operating systems 

AIX Solaris HP-UX 

Medium changer smc[0-255] [0-255]smc [0-255]chng 


Tape device 

AIX Solaris HP-UX 

rmt[0-255] 

rmt[0-255].1 1 

[0-255]st 

[0-255]stb 2 

[0-255]stc 3 

[0-255]stn 4 

[0-255]stbn 2,4 

[0-255]stcb 2,3 

[0-255]stcn 3,4 

[0-255]stcbn 2,3,4 

[0-255]m 

[0-255]mb 2 

[0-255]mn 4 

[0-255]mnb 2,4 

[tTGTdLUN 5 ]BEST 

[tTGTdLUN 5 ]BESTb 2 

[tTGTdLUN 5 ]BESTn 4 

[tTGTdLUN 5 ]BESTnb 2,4 

1. Adding (.1) to the device name, means that it is a Rewind on Close device. 

Under certain conditions, before rewinding, the device writes filemarks. This option 

is used by some application software, such Legato NetWorker — see 6.3.1, 

“Software installation (AIX)” on page 238. 

2. A (b) stands for Berkeley Software Distribution (BSD) compatibility. The BSD 

device special file modifies close behavior for non-rewind devices. If the device is 

opened for no rewind on close, in non-BSD mode, if the last command before 

closing the device was a read, then the tape is positioned after the file mark immediately 

following the last block read. If the device is opened for no rewind on 

close, in BSD mode, if the last command before closing the device was a read, 

the tape is left positioned exactly where it was following the last block read. If the 

device is opened for rewind on close, the BSD mode is not relevant. 

3. A (c) stands for Compression: the compression device special file determines 

whether the tape device will use built-in hardware compression while storing data 

on the tape. 

4. An (n) stands for No rewind on close: the no rewind on close device special file 

does not rewind the tape during a close operation. Otherwise, the tape is rewound 

when the device is closed. 

5. The TGT is the SCSI target ID, and LUN is the Logical Unit Number associated 

with the device. 

In the following section, we give detailed examples of installing the drivers in AIX, 

SUN Solaris and HP-UX environments. 

2.2 LTO driver installation for AIX 

The device driver used for IBM pSeries or RS/6000 AIX platforms is called Atape 

and it includes both the tape drive and medium changer drivers. 

To check if your environment has the correct operating system level, combined 

with the appropriate server and the supported host bus adapter and Storage 

Area Network fabric components, check the IBM pSeries, RS/6000 and SP 

Servers section of the following link: 


Chapter 2. Basic LTO setup for UNIX systems 31

This link is specific for the 3583 Ultrium Scalable Tape Library, if you have a 

different LTO Ultrium tape drive or library, you can use the same link, changing 

only the 3583opn keyword to 3580opn, 3581opn or 3584opn, as appropriate for your 

LTO model. 

You must have root authority to proceed with the installation of the driver. 

Attention: A reboot of the host system is required to complete the installation 

so you should schedule this operation accordingly. 

Enter the following command to determine if the Atape.driver is installed and to 

determine the version. 

Example 2-1 Checking if Atape.driver is installed on AIX 

# lslpp -l Atape.driver 

Fileset Level State Description 

---------------------------------------------------------------------------- 

Path: /usr/lib/objrepos 

Atape.driver 

6.1.8.0 COMMITTED IBM AIX Enhanced Tape and 

Medium Changer Device Driver 

You can see we do have the driver installed at level 6.1.8.0. If the driver is not 

installed, or you want to check if there is a later version available, look at: 

ftp://ftp.software.ibm.com/storage/devdrvr/AIX/ 

Or 

http://www.storage.ibm.com/storagesmart/lto/support/lto_ftp.htm 

Be sure to check the README file for the latest installation notes and 

prerequisites. 

In our example, we are installing the Atape driver in the following environment: 

► IBM RISC/6000 F50 server with IBM AIX operating system release 4.3.3 

► LVD SCSI adapter 

► IBM 2108-R03 SAN Data Gateway 

► 3583 Ultrium Scalable Tape Library with LVD drives 

You can use either the command line or the SMIT interface to install the driver. If 

the Atape device driver is already installed in your system and you just want to 

update it, first uninstall the previous Atape driver following the instructions in 

2.2.7, “Removing Atape driver from the system (uninstallation)” on page 39. 


2.2.1 Atape driver installation using the command line interface 

To install from a diskette, put it in the diskette drive. Assuming the default system 

diskette drive, use the following command to install: 

installp -acXd /dev/rfd0 Atape.driver 

Otherwise, if you have downloaded the driver to your local system, assuming it is 

located in /marco/Atape.driver, use this command: 

installp -acXd /marco Atape.driver 

This will install and commit the Atape driver in your system. Here is an example 

of the installp command output: 

Example 2-2 AIX Atape installation output 

# installp -acXd /marco Atape.driver 

+-----------------------------------------------------------------------------+ 

Pre-installation Verification... 

+-----------------------------------------------------------------------------+ 

Verifying selections...done 

Verifying requisites...done 

Results...SUCCESSES--------- 

Filesets listed in this section passed pre-installation verification and will 

be installed. 

Selected Filesets ----------------- 

Atape.driver 6.1.8.0 

# IBM AIX Enhanced Tape and Me... 

> 

FILESET STATISTICS------------------ 

1 Selected to be installed, of which: 

1 Passed pre-installation verification ---- 

1 Total to be installed 

0503-409 installp: bosboot verification starting... 

installp: bosboot verification completed. 

+-----------------------------------------------------------------------------+ 

Installing Software... 

+-----------------------------------------------------------------------------+ 

installp: APPLYING software for: Atape.driver 6.1.8.0 

. . . . . > . . . . . . . IBM AIX Enhanced Tape 

and Medium Changer Device Driver 

(C) COPYRIGHT International Business Machines Corp. 1993 1997 All Rights 

Reserved Licensed Materials - Property of IBM 

US Government Users Restricted Rights - Use, duplication or disclosure 

restricted by GSA ADP Schedule Contract with IBM Corp. 

. . . . . >. . . . 

Checking for existing Atape devices... 

Installing AIX Version 4.3 Atape.driver... 

Adding device prototype... 

Adding odm and smit entries... 


Adding catalogs... 

Adding trace template... 

Adding error template... 

5 entries added. 

0 entries deleted. 

0 entries updated. 

Adding utility programs... 

Finished processing all filesets. (Total time: 14 secs). 



0503-408 installp: bosboot process starting... 

bosboot: Boot image is 8469 512 byte blocks. 

0503-292 This update will not fully take effect until after a system reboot. 

installp: bosboot process completed. 

+-----------------------------------------------------------------------------+ 

Summaries: 

+-----------------------------------------------------------------------------+ 

Installation Summary-------------------- 

Name Level Part Event Result 

------------------------------------------------------------------------------- 

Atape.driver 6.1.8.0 USR APPLY SUCCESS 

After the successful installation of the driver, go to section 2.2.3, “Configuring 

tape and medium changer devices” on page 35. 

2.2.2 Installation using the SMIT interface 

Start SMIT (Systems Management Interface Tool) by typing smit at the 

command line. Take the following options, Software Installation and 

Maintenance -> Install and Update Software-> Install and Update Software 

by Package Name (includes devices and printers). 

Note that we show the ASCII SMIT interface, however if you have an XWindows 

capable display, you will see the GUI version. Both have identical function. Your 

screen will look like this. Enter in the directory where you have downloaded the 

driver to be installed (or /dev/fd0 if the driver is on a diskette) and press Enter. 

Example 2-3 ATape installation using SMIT 

Install and Update Software by Package Name (includes devices and printers) 

Type or select a value for the entry field. 

Press Enter AFTER making all desired changes. 

[Entry Fields] 

* INPUT device / directory for software [/marco] + 


Press Enter. A popup list will display. Select Atape and press Enter again. 

Another popup will show a line similar to: 

@ 6.1.8.0 IBM AIX Enhanced Tape and Medium Changer Device Driver 

Select this by pressing F7 and Enter. Complete the installation options on the 

next screen like this: 

Example 2-4 AIX Atape SMIT installation options 

Install and Update Software by Package Name (includes devices and printers) 

Type or select values in entry fields. 



* INPUT device / directory for software /marco 

* SOFTWARE to install [@ 6.1.8.0 IBM AIX Enh> 

PREVIEW only? (install operation will NOT occur) no + 

COMMIT software updates? yes + 

SAVE replaced files? no + 

AUTOMATICALLY install requisite software? yes + 

EXTEND file systems if space needed? yes + 

OVERWRITE same or newer versions? no + 

VERIFY install and check file sizes? no + 

Include corresponding LANGUAGE filesets? yes + 

DETAILED output? no + 

Process multiple volumes? yes + 

F1=Help F2=Refresh F3=Cancel F4=List 

Esc+5=Reset Esc+6=Command Esc+7=Edit Esc+8=Image 

Esc+9=Shell Esc+0=Exit Enter=Do 

After this selection, the installation process begins. You see messages like in 

Example 2-2 on page 33. 

You are now ready to configure the devices. 

2.2.3 Configuring tape and medium changer devices 

After the driver software is installed and an LTO tape device or library is 

connected to the server, the device can be configured and made available for 

use. You cannot access the devices before completing this configuration step. 

Configure a tape device by using either of the following procedures. 

► Enter the following command without parameters: 

cfgmgr 


► 

This command configures all devices automatically (including any new tape or 

medium changer device). Since a bosboot verification is run during the 

installp command processing, you may need a reboot of the server in order 

for the changes to take place. Look for warning messages after submitting the 

command. 

Or 

Power off your system and reboot to configure the LTO automatically during 

the startup and make available any new tape or medium changer devices in 

the system. 

2.2.4 Verifying the Atape driver installation 

To verify the correctness of the installation and configuration of the Atape device 

driver, issue the following command: 

Example 2-5 Configured tape devices 

# lsdev -Cctape 

smc0 Available 14-08-00-6,0 IBM 3583 Library Medium Changer 

rmt0 Available 14-08-00-0,0 IBM 3580 Ultrium Tape Drive 


In the output you can see: 

► smc0, rmt0, rmt1 are the device special file names for the medium changer 

and the tape drives. Refer to Table 2-1 on page 30 for a list of the device 

special file names for every platform. If you have installed a 3580 manual 

drive, you will not see any entry for the medium changer as it does not have 

one. 

► Available means that the device is installed correctly and ready. 

► We are using a SCSI. 14-08-0 is the ID of the SCSI adapter, 6 is the SCSI 

address and 0 is the LUN (for the medium changer. The drives use LUN 0 on 

SCSI ids 0 and 1, the same apply for the devices). 

You can now use the tapeutil utility to test the library. Go to section 2.5, “Testing 

the library with tapeutil” on page 58 to learn how to use this utility. Open the 

device, issue inquiry commands, read and write, and move media across the 

library to check the driver functionality. 

The drives can also be used with operating system commands such as tar and 

cpio. 


2.2.5 Configuring the LTO device parameters 

You can change the default operating parameters for the tape drive and device 

drivers using the smit interface. 

Type smit at the AIX command prompt, then select Devices ---> Tape Drive ---> 

Change / Show Characteristics of a Tape Drive. Select the device you want to 

change from the list. The output in Figure 2-6 appears. 

Example 2-6 SMIT characteristics of a tape drive 

Change / Show Characteristics of a Tape Drive 




Tape Drive 

rmt0 

Tape Drive type 3580 

Tape Drive interface 

scsi 

Description IBM 3580 Ultrium Tape > 

Status 

Available 

Location 14-08-00-0,0 

Parent adapter 

scsi1 

Connection address 0,0 

Block Size (0=Variable Length) [0] 

Use Hardware Compression on Tape 

yes 

Activate volume information logging 

no 

Maximum size of log file (in # of entries) [500] 

Backward Space/Forward Space Record Mode 

SCSI 

Use Immediate Bit in Rewind Commands 

no 

Trailer Label Processing 

no 

Select and possibly change the device options and click OK. A command will be 

issued and smit prompts for the results. Check for the output message rmtx 

Changed. The default for the hardware compression on tape is yes for every LTO 

Ultrium device, and it is recommended for performance reasons to have the 

device compression activated. Blocksize=0 means that the device uses variable 

block size. Some application software may override this value. Refer to Part 2, 

“LTO Ultrium with applications” on page 145 for specific application software 

related information. 


2.2.6 Deleting LTO devices 

In some cases, it could be necessary to remove the tape device or the medium 

changer from the AIX device configuration. This is a required step for some 

application software that must have a specific device driver for medium changer 

and the IBM-provided Atape driver for the tape drive. 

Refer to Part 2, “LTO Ultrium with applications” on page 145, for more information 

on specific application software driver requirements. 

Remove the devices from the configuration using either of the following 

procedures: 

► 

► 

The first method leaves the device defined in the configuration database. It is 

similar to bringing the device offline. Enter the following command to bring the 

/dev/rmtn device offline, but leave it defined in the device database: 

rmdev -l rmtn 

In this case, the removed device will appear as: 

rmt0 Defined 14-08-00-0,0 IBM 3580 Ultrium Tape Drive 

Or 

The second method brings the device offline and also removes its definition 

from the device database. Enter the following command: 

rmdev -l rmtn -d 

After this, the removed device will not show up in the device listing (lsdev) at 

all. 

Use the actual device number (for example, /dev/rmt0) and repeat for each drive 

to be removed. If you are removing a medium changer as well, repeat with 

/dev/smcx. 

Alternatively, you can use the smit interface. From the main menu, select 

Devices-> Tape Drive -> Remove a Tape Drive. From the list of devices, select 

the device that you want to remove from the configuration as shown in the 

example. 

Example 2-7 Select tape device to delete 

lqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqk 

x Tape Drive x 

x 

x 

x Move cursor to desired item and press Enter. 

x 

x 

x 

x smc0 Available 20-58-01-6,0 IBM 3583 Library Medium Changer (FCP) x 

x rmt0 Available 20-58-01-0,0 IBM 3580 Ultrium Tape Drive (FCP) x 

x rmt1 Available 20-58-01-1,0 IBM 3580 Ultrium Tape Drive (FCP) x 


Select no in the KEEP definition in database field, then press OK. The device will 

then be deleted. 

Now you can uninstall the Atape driver. You cannot unload the device driver from 

the kernel until the last device used by this driver has been deconfigured. 

2.2.7 Removing Atape driver from the system (uninstallation) 

You can uninstall the Atape device driver using the smit command, by going to 

the Uninstall software menu and selecting Atape.driver. 

Otherwise, you can use the installp command: 

installp -u Atape.driver 

Attention: All tape devices that use the Atape driver must be closed and idle 

when uninstalling Atape, or the uninstall will fail. 

Here is the sample output of the installp -u removal command: 

Example 2-8 AIX uninstall Atape driver 

+-----------------------------------------------------------------------------+ 

Pre-deinstall Verification... 

+-----------------------------------------------------------------------------+ 

Verifying selections...done 

Verifying requisites...done 

Results...SUCCESSES--------- 

Filesets listed in this section passed pre-deinstall verification and will be 

removed. 

Selected Filesets ----------------- 

Atape.driver 6.1.8.0 

# IBM AIX Enhanced Tape and Me.. 

> 

FILESET STATISTICS------------------ 

1 Selected to be deinstalled, of which: 

1 Passed pre-deinstall verification ---- 

1 Total to be deinstalled 



+-----------------------------------------------------------------------------+ 

Deinstalling Software... 

+-----------------------------------------------------------------------------+ 

installp: DEINSTALLING software for: Atape.driver 6.1.8.0 

De-installing Atape.driver... 

Atape.driver de-installed 

Finished processing all filesets. (Total time: 10 secs). 



0503-408 installp: bosboot process starting... 


osboot: Boot image is 8437 512 byte blocks. 

0503-292 This update will not fully take effect until after a system reboot. 

installp: bosboot process completed. 

+-----------------------------------------------------------------------------+ 

Summaries: 

+-----------------------------------------------------------------------------+ 

Installation Summary-------------------- 

Name Level Part Event Result 

------------------------------------------------------------------------------- 

Atape.driver 6.1.8.0 USR DEINSTALL SUCCESS 

2.3 LTO driver installation for Solaris 

The device driver used for SUN Solaris platforms is called IBMtape and it 

includes both the tape drive and medium changer drivers. 

This driver provides SCSI and FC-AL attachment for IBM magnetic tape and 

library subsystem products for Sun Microsystems SPARC and UltraSPARC 

platforms running the Solaris operating system, including the Ultra/Enterprise 

family of servers. 

To cross-check if your environment has the correct operating system level, 

combined with the appropriate server and the supported host bus adapter and 

Storage Area Network fabric components, check the SUN servers section of the 

following link: 


This link is specific for 3583 Ultrium Scalable Tape Library, if you have a different 

LTO Ultrium tape drive or library, you can use the same link changing only the 

3583opn keyword to 3580opn, 3581opn or 3584opn, as appropriate for your LTO 

model. 

Download the latest IBMtape driver version from: 

ftp://ftp.software.ibm.com/storage/devdrvr/Solaris/ 

Select the desired file with the .bin extension and read the IBMtape.README for 

latest release information, fixes and prerequisites. 

The following sections describe the installation, verification, configuration and 

removal procedures of the IBMtape device driver for the 3583 Ultrium Scalable 

Tape Library in a SUN Solaris environment. 


2.3.1 IBMtape driver installation 



IBMtape attempts to claim and operate only the supported IBM tape devices 

(including Ultrium and Magstar drives, among others). However, the Solaris 

operating system includes its own SCSI tape device driver, named st, which will 

automatically claim any SCSI-compliant tape drive it detects, including devices 

that IBMtape should be managing. 

In order to avoid conflicts between IBMtape and st, you must prevent the st driver 

from claiming and attempting to operate IBMtape-owned devices. Likewise, other 

supplier’s SCSI tape device drivers that you have installed must be prevented 

from claiming IBMtape-owned devices. 

To prevent such conflicts, read carefully the suggestions detailed in IBM Ultrium 

Device Drivers: Installation and User’s Guide, GA32-0430, before continuing the 

driver installation. 

The following steps are required to complete the process. As an example, we 

detail the installation for a Solaris system without any previous tape subsystem 

configured, installing the IBMtape driver for a 3583 Ultrium Scalable Tape 

Library. 

In this example, we are installing the IBMtape driver in the following environment: 

► SUN Microsystems Enterprise250 SPARC server with SUN Solaris operating 

system release 2.7 

► QLogic QLA2200F PCI Fibre Channel Host Bus Adapter (we tested also 

Emulex LP8000 PCI FC HBA) 

► IBM 2108-R03 SAN Data Gateway 


This example assumes we have downloaded a package file named 

IBMtape.4.2.2.5 into the /tmp directory. 

1. Log on to the target system as root. 

2. Ensure that all user and tape drive activity on the system has halted. 

3. Use pkgadd to install the driver: 

pkgadd -d /tmp/IBMtape.4.0.4.5.bin 

4. After submitting the command, you should see the following messages: 


Example 2-9 IBMtape install output on Solaris 

The following packages are available: 

1 IBMtape IBM 32-bit and 64-bit SCSI Tape & Medium Changer Device 

Driver 4.0.4.5 (sparc) 4.0.4.5 

Select package(s) you wish to process (or 'all' to process all packages). 

(default: all) [?,??,q]: 1 

select [1] and press Enter. The following messages should follow: 

Processing package instance from 

IBM 32-bit and 64-bit SCSI Tape & Medium Changer Device Driver 4.0.4.5 

(sparc) 4.0.4.5 

IBM SCSI Tape & Medium Changer Device Driver for Sun Solaris 2.x Version 

4.0.4.5 

Contains Licensed Internal Code Licensed Materials - Property of IBM(C) 

Copyright 1994-2000. IBM Corporation All Rights Reserved 

Portions (C) 1994 Sun Microsystems 

US Government Users Restricted Rights - Use, duplication, or disclosure 

restricted by GSA ADP Schedule Contract with IBM 

(R) IBM is a registered trademark of IBM Corporation(R) Solaris is a 

registered trademark of Sun Microsystems 

Using as the package base directory. 

## Processing package information. 

## Processing system information. 

## Verifying disk space requirements. 

## Checking for conflicts with packages already installed. 

## Checking for setuid/setgid programs. 

This package contains scripts which will be executed with super-user 

permission during the process of installing this package. 

Do you want to continue with the installation of [y,n,?] y 

Installing IBM 32-bit and 64-bit SCSI Tape & Medium Changer Device Driver 

4.0.4.5 as 

## Installing part 1 of 1. 

/opt/IBMtape/IBMtape.conf 

/opt/IBMtape/tapeutil 

/opt/IBMtape/tapeutil.c 

/usr/include/sys/oldtape.h 

/usr/include/sys/smc.h 

/usr/include/sys/st.h 

/usr/include/sys/svc.h 

/usr/kernel/drv/IBMtape 

/usr/kernel/drv/sparcv9/IBMtape 

[ verifying class ] 

## Executing postinstall script. 

## The /usr/kernel/drv/IBMtape.conf file already exists. 

## It will be preserved as the current configuration file 

## for the IBM SCSI Tape & Medium Changer Device Driver. 

## Loading IBM SCSI Tape & Medium Changer Device Driver. 

Aug 8 14:10:38 sol-e unix: 


Aug 8 14:10:38 sol-e 

Aug 8 14:10:38 sol-e IBM SCSI Tape & Medium Changer Device Driver 

Aug 8 14:10:38 sol-e for Solaris 2.x 

Aug 8 14:10:38 sol-e 

Aug 8 14:10:38 sol-e Version 4.0.4.5 

Aug 8 14:10:38 sol-e 

Aug 8 14:10:38 sol-e (C) COPYRIGHT IBM Corporation, 1994-1998 

Aug 8 14:10:38 sol-e Licensed Materials - Property of IBM 

Aug 8 14:10:38 sol-e All Rights Reserved 

Aug 8 14:10:38 sol-e 

Aug 8 14:10:38 sol-e US Government Users Restricted Rights - 

Aug 8 14:10:38 sol-e Use, duplication, or disclosure restricted 

Aug 8 14:10:38 sol-e by GSA ADP Schedule Contract with IBM Corp. 

Aug 8 14:10:38 sol-e 

Aug 8 14:10:38 sol-e last message repeated 1 time 

Aug 8 14:10:38 sol-e 

Aug 8 14:10:43 sol-e unix: /kernel/drv//sparcv9/fca-pci symbol 

Aug 8 14:10:43 sol-e unix: ddi_model_convert_from multiply defined 

Aug 8 14:10:43 sol-e unix: NOTICE: IBMtape _probe: Found Medium 

Changer ULT3583-TL , Inst 193, Tgt 2, Lun 1 

Installation of was successful. 

Before using the tape devices and the library, configure the device drivers 

parameters, as detailed in the following section. 

2.3.2 Configuring tape and medium changer devices 

Edit IBMtape.conf, located in /kernel/drv, be sure that all the required target and 

LUN definitions for the drives and medium changer are in place. 

In this example, we are using an IBM SAN Data Gateway to connect the 3583 

Ultrium Scalable Tape Library to a Storage Area Network. When devices are 

attached via the IBM SAN Data Gateway, SCSI target IDs and LUNs on the 

gateway's SCSI channels are remapped to other values which are then 

presented to the host operating system. 

When the server boots, the devices attached to the first gateway encountered by 

the boot I/O scan process, are assigned SCSI target ID 0 (with LUNs between 

0-127), devices attached to the second gateway have SCSI target ID 1 and so 

on. 


Important: For every SDG attached to the server, LUN 0 is used by the 

gateway itself, and LUNs 2-126 are used for tape devices attached to the 

gateway. Tape drives are always assigned an even number LUN and start at 2. 

The medium changer will have an ODD number. (See 3.4.3, “LUN mapping” 

on page 136.) 

Then for the tape devices, the gateway assigns LUNs in increasing sequence, as 

the devices are discovered on its attached SCSI channels. 

This means that adding and removing devices, cabling changes, and 

readdressing of devices may cause gaps in the sequence of assigned LUNs, 

because the LUN number assignment is made every time the server reboots. 

Important: Refer to 3.3, “Persistent binding” on page 112 for more information 

and recommendations for ensuring a consistent assignment of target/LUN 

IDs. 

Therefore, in this configuration example, as discussed before, LUN addresses 0 

should not be included, so the IBMtape.conf file should be configured like in the 

following example: 

Example 2-10 IBMtape.conf example for smc and two drives 

name="IBMtape"class="scsi" 

target=0 lun=1 

block_size=0 

buffering=1 

immediate=0 

trailer=0 

sili=0; 



block_size=0 

buffering=1 

immediate=0 

trailer=0 

sili=0; 



block_size=0 

buffering=1 

immediate=0 

trailer=0 

sili=0; 


Block_size=0 means that the device uses variable block size. Some application 

software may override this value. Refer to Part 2, “LTO Ultrium with applications” 

on page 145 for specific application software related information. 

If you have multiple SCSI connected devices attached to the same server, it 

could be difficult to know what SCSI target ID and LUN number will be assigned 

by the OS boot scan process to the LTO drives. 

In this situation, if you are working on a test system and you can reboot without 

impacting production operation, try first to include in IBMtape.conf all possible 

target and LUN numbers, like in the following example: 

Example 2-11 Including all LUN/target combinations in IBMtape.conf 



.............. 



.............. 



.............. 



.............. 



.............. 



.............. 

You can use every reasonable combination of target and LUN numbers to include 

in the IBMtape.conf file. Then, after rebooting, you can see which target and LUN 

IDs were actually assigned for the devices by using the dmesg command. Look for 

the entries: 

Example 2-12 Tape entries in the dmesg command output 

Aug 22 18:22:52 sol-e unix: NOTICE: IBMtape _probe: Found Medium Changer 

ULT3583-TL, Inst 234, Tgt 0, Lun 1 

Aug 22 18:22:52 sol-e unix: NOTICE: IBMtape _probe: Found Tape Drive 

ULT3580-TD1, Inst 235, Tgt 0, Lun 2 

Aug 22 18:22:52 sol-e unix: NOTICE: IBMtape _probe: Found Tape Drive 

ULT3580-TD1, Inst 237, Tgt 0, Lun 4 

The SCSI target and LUN number are the addresses assigned to the tape drives 

and the medium changer during the boot process. 


Now you can easily modify again IBMtape.conf excluding the unused entries. 

Don’t leave the unused entries in the file, because if you have too many entries, it 

can slow the boot process. 

Reboot the system again after removing the unused entries in IBMtape.conf. One 

common method to perform a reboot is shown here but use your installation 

normal procedures. 

reboot -- -r 

Verify operation of the newly installed or readdressed equipment (see the next 

section). 

2.3.3 Verifying the IBMtape driver installation 

To verify the correctness of the installation and configuration of the IBMtape 

device driver, issue the following command: 

ls -l /dev/rmt/*stbn /dev/rmt/*smc 

lrwxrwxrwx 1 root root 47 Aug 22 18:12 /dev/rmt/0smc -> 

../../devices/pci@1f,4000/scsi@2/IBMtape@0,1:st 

lrwxrwxrwx 1 root root 47 Aug 22 18:12 /dev/rmt/1stcbn -> 


lrwxrwxrwx 1 root root 47 Aug 22 18:12 /dev/rmt/2stcbn -> 


You should see the above output, where: 

0smc 

is the device special file name for the 3583 medium changer 

1stcbn 

is the device special file name for the first 3580 tape drive 

2stcbn 

is the device special file name for the second 3580 tape drive 

IBMtape is listed correctly as the owner of these devices. Refer to Table 2-1 on 

page 30 for a list of the device special file names for every platform. For 

performance reasons, it is suggested to use a special device file name that has 

compression turned on. If you have installed a 3580, you will not see any entry 

for the medium changer as it does not have one. 

You can now use the tapeutil utility to test the library. Go to section 2.5, “Testing 

the library with tapeutil” on page 58 to learn how to use this utility. Open the 

device, issue inquiry commands, read and write, and move media across the 

library to check the driver functionality. 

The drives can also be used with operating system commands such as tar and 

cpio. 


2.3.4 Deleting LTO devices 

In some cases, it could be necessary to remove the tape device or the medium 

changer from the device configuration. This is a required step for some 

application software, that must have a specific device driver for medium changer, 

and the IBMtape device driver for the tape drive. 

Refer to Part 2, “LTO Ultrium with applications” on page 145 for more information 


To remove the IBMtape driver support for a specific device, edit the IBMtape.conf 

file in the /kernel/drv directory and remove (comment out) the statement for the 

device that is no longer to be configured with this driver. 

Then reboot the system in order for the changes to take effect. 

2.3.5 Removing IBMtape driver from the system (uninstallation) 

Use the pkgrm command to remove the IBMtape package from the system: 

/usr/sbin/pkgrm IBMtape 

All active processes using any IBM device supported by the IBMtape driver must 

be stopped in order for the removal procedure to complete successfully. 

Here is the sample output of the pkgrm command: 

Example 2-13 Uninstall IBMtape drive on Solaris 

The following package is currently installed: 

IBMtape IBM 32-bit and 64-bit SCSI Tape & Medium Changer Device 

Driver 4.0.3.5 (sparc) 4.0.3.5 

Do you want to remove this package? y 

## Removing installed package instance 

This package contains scripts which will be executed with super-user permission 

during the process of removing this package. 

Do you want to continue with the removal of this package [y,n,?,q] y 

## Verifying package dependencies. 


## Executing preremove script. 

## Unloading IBM SCSI Tape & Medium Changer Device Driver. 

## Deleting device special file nodes. 

## The /usr/kernel/drv/IBMtape.conf file will not be removed 

## to preserve previously established configurations. 

## Removing pathnames in class 

/usr/kernel/drv/sparcv9/IBMtape 

/usr/kernel/drv/IBMtape 

/usr/include/sys/svc.h 

/usr/include/sys/st.h 


usr/include/sys/smc.h 

/usr/include/sys/oldtape.h 

/opt/IBMtape/tapeutil.c 

/opt/IBMtape/tapeutil 

/opt/IBMtape/IBMtape.conf 

/opt/IBMtape 

# Updating system information. 

Removal of was successful. 

2.4 LTO driver installation for HP-UX 

The device driver used for HP-UX UNIX platforms is called atdd and it includes 

both the tape drive and medium changer drivers. 

Restriction: At this time, this device is available only for SCSI connections. 

To check if your environment has the correct operating system level, combined 

with the appropriate server and the supported host bus adapter, check the 

Hewlett-Packard Servers section of the following link: 


This link is specific for 3583 Ultrium Scalable Tape Library, if you have a different 

LTO Ultrium tape drive or library, you can use the same link changing only the 

3583opn keyword to 3580opn, 3581opn and 3584opn, as appropriate for your LTO 

model. 

Download the latest Atdd driver version from: 

ftp://ftp.software.ibm.com/storage/devdrvr/HPUX/ 

To find the right driver, look in the README file for the correct version 

corresponding to your HP server’s bus (either Precision or PCI) and operating 

system. In our case, we have an HP PCI bus system with HP-UX Version 11.0, 

therefore we look for files called atdd.1.x.x.x.bin: the versions of this driver (for 

LTO drives and library) are identified with this name. 

At the same link, we recommend that you also download the tapeutil utility. This 

file is called tapeutil.hpux.x.x.x.x.bin and is packaged separately, unlike for 

Solaris or AIX. You use this utility to address tape and medium changer 

commands, such as mount, dismount, move media, write and read files. 

Be sure to check the README for the information on fixes and prerequisites. 


The following sections describe the installation, configuration, verification and 

removal procedures of the IBMtape device driver for the 3583 Ultrium Scalable 

Tape Library in HP-UX environment. 

2.4.1 Atdd driver installation 



For detailed installation instructions, refer also to IBM Ultrium Device Drivers 

Installation and User’s Guide, GA32-0430. 

In this example, we are installing the Atdd driver in the following environment: 

► HP9000 server with HP-UX operating system release 11.00 

► SCSI Ultra-2 LVD adapter HP A5150A PCI 


Next, we give the detailed steps of the installation process. You must have root 

authority to proceed with the installation of the driver. We assume the driver 

atdd.1.7.8.2.bin has been downloaded to the directory /marco. 

1. Copy the installation file to the software depot using the command: 

swcopy -p -s /marco/atdd.1.7.8.2.bin atdd 

swcopy -s /marco/atdd.1.7.8.2.bin atdd 

The first command performs a preview copy only. Repeat without the -p option 

to actually execute the copy. Use the following command to verify that the 

Atdd software has been copied in the depot: 

swlist -d atdd 

You should see these messages: 

atdd 1.7.8.2 GES IBM Magstar and Ultrium tape device driver 

atdd.driver 

Advanced Tape Device Driver 

2. Use the swlist command to view the product’s README file: 

swlist -d -a readme atdd | more 

3. The following command installs Atdd from the depot to the default root file 

system: 

swinstall atdd 

If you receive an error message saying that a reboot of the system is needed, 

then reissue the swinstall command with the autoreboot option as follows: 

swinstall -x autoreboot=true atdd 


Note: If an earlier version of the product is already installed on the 

target root filesystem, the existing version will be replaced. This is true 

even if the version already installed is more recent than the version 

being installed. 

You see the following messages during the installation flow: 

Example 2-14 Installing atdd on HP-UX 

* Agent session started for user "root@easter.almaden.ibm.com". (pid=13597) 

* Beginning Analysis Phase. 

* Source: easter.almaden.ibm.com:/marco/atdd.1.7.8.2.bin 

* Target: easter.almaden.ibm.com:/var/spool/sw 

* Target logfile: easter.almaden.ibm.com:/var/spool/sw/swagent.log 

* Reading source for product information. 

* Reading source for file information. 

NOTE: The used disk space on filesystem "/var" is estimated to increase 

by 1918 Kbytes. 

This will leave 808263 Kbytes of available user disk space after 

the installation. 

* Summary of Analysis Phase: 

* 1 of 1 filesets had no Errors or Warnings. 

* The Analysis Phase succeeded. 

* Beginning the Copy Execution Phase. 

* Filesets: 1 

* Files: 17 

* Kbytes: 1888 

* Copying fileset "atdd.driver,r=1.7.8.2" (1 of 1). 

* Summary of Execution Phase: 

* 1 of 1 filesets had no Errors or Warnings. 

* The Execution Phase succeeded. 

4. To verify the correct installation, use the following command: 

swverify atdd 

2.4.2 LTO medium changer configuration 

This step applies only if you are installing a 3581 Tape Autoloader, a 3583 

Ultrium Scalable Tape Library or a 3584 Ultrium UltraScalable Tape Library. The 

3580 LTO tape drive does not include a medium changer, so this step is not 

necessary. 


Attention: by default, the medium changer is NOT configured in the Atdd 

device driver. This step is REQUIRED for LTO libraries that have a medium 

changer installed. 

After having successfully installed the Atdd device driver, if you issue the 

following command: 

ioscan -fn 

You will see the following lines of output: 

Example 2-15 Output of ioscan before configuring medium changer 

ext_bus 5 0/7/0/0 c720 CLAIMED INTERFACE SCSI C896 ... 

tape 0 0/7/0/0.0.0 atdd CLAIMED DEVICE IBM ULT3580-TD1 


unknown -1 0/7/0/0.6.0 UNCLAIMED UNKNOWN IBM ULT3583-TL 

ctl 5 0/7/0/0.7.0 sctl CLAIMED DEVICE Initiator/... 

The tape devices are CLAIMED by the device driver, but the medium changer 

(IBMULT3583-TL) is in unknown state and UNCLAIMED (line in bold). It is not 

recognized by the device driver. 

To allow the Atdd driver to CLAIM all attached IBM medium changer targets, the 

atdd_autoch parameter must be enabled: 

1. Start the System Administration Manager (SAM) by typing sam at the 

command line. If you are using an XWindows capable display you will see the 

GUI version, otherwise you will see an ASCII equivalent. Both offer the same 

function. Figure 2-1 shows the first panel of the GUI version. 


Figure 2-1 System Administration Manager (SAM) main panel 

2. Select Kernel Configuration. 

3. Select Configurable Parameters. 

4. Scroll until you find the configuration parameter atdd_autoch then press 

Enter. 

5. In the new window, change Formula/Value to 1, then select OK. 

6. The Pending Value is now 1, as shown in Figure 2-2. 


Figure 2-2 Enabling the medium changer in the kernel configuration 

7. Exit SAM — a window opens asking to Create a New Kernel Now. Select this 

option and the system will build a new kernel. This may take a few minutes. 

8. A second window will open that allows you to Move Kernel Into Place and 

Shutdown/Reboot System Now. Select this option. The system reboots. 

When the atdd_autoch parameter is enabled (value=1), the Atdd driver claims all 

IBM medium changer targets that respond during the boot process. 

Attention: The library must be online during the boot process in order to be 

recognized by the system. 

After rebooting, the output of the ioscan command should now look like this: 

Example 2-16 Output of ioscan after configuring medium changer 

ext_bus 5 0/7/0/0 c720 CLAIMED INTERFACE SCSI C896 ... 



autoch 2 0/7/0/0.6.0 atdd CLAIMED DEVICE IBM ULT3583-TL 

ctl 5 0/7/0/0.7.0 sctl CLAIMED DEVICE Initiator/... 

The LTO medium changer is now CLAIMED and therefore controlled by the Atdd 

driver. 


2.4.3 Verifying the Atdd driver installation 

To verify the correctness of the installation and configuration of the Atdd device 

driver, issue the following command: 

Example 2-17 Verifying atdd driver is correctly installed 

# ls -la /dev/rmt 

crw-rw-rw- 1 bin bin 245 0x050000 Aug 16 11:13 0m 

crw-rw-rw- 1 bin bin 245 0x050080 Aug 16 11:13 0mb 

crw-rw-rw- 1 bin bin 245 0x050040 Aug 16 11:13 0mn 

crw-rw-rw- 1 bin bin 245 0x0500c0 Aug 17 17:14 0mnb 





crw-rw-rw- 1 bin bin 245 0x056000 Aug 16 11:14 2chng 

crw-rw-rw- 1 bin bin 245 0xfffffc Aug 16 11:13 atdd.cfg 

crw-rw-rw- 1 bin bin 245 0xfffffd Aug 16 11:13 atdd.dbg 

crw-r--r-- 1 root root 245 0xfffffe Aug 15 11:24 atdd_config 

crw-rw-rw- 1 bin bin 245 0x050000 Aug 16 11:13 c5t0d0BEST 

crw-rw-rw- 1 bin bin 245 0x050080 Aug 16 11:13 c5t0d0BESTb 

crw-rw-rw- 1 bin bin 245 0x050040 Aug 16 11:13 c5t0d0BESTn 

crw-rw-rw- 1 bin bin 245 0x0500c0 Aug 16 11:13 c5t0d0BESTnb 





You should see the above output, where: 

2chng This is the device special file name for the 3583 medium changer. 

0m This is one of the device special file names used for the first 3580 

tape drive. 

1m This is one of the device special file names used for the second 3580 

tape drive. 

And, the same for the other tape drives installed. Refer to Table 2-1 on page 30 

to interpret all the device names. 

You can now use the tapeutil utility to test the library. If you downloaded the 

tapeutil installation file as described in 2.4, “LTO driver installation for HP-UX” on 

page 48, install the package with the following commands: 

swcopy /marco/tapeutil.hpux.4.0.0.0.bin tapeutil 

swinstall tapeutil 


Then go to section 2.5, “Testing the library with tapeutil” on page 58 to learn how 

to use tapeutil. 

Open the device, issue inquiry commands, read and write, move media across 

the library to check the driver functionality. 

2.4.4 Configuring tape devices 

If you want to change the default configuration parameters for the tape devices, 

like compression, blocksize, buffering, use the following procedure. 

You can examine the current configuration using the atdd_cfg program located 

in /opt/OMImag/bin directory. Every tape device is associated with a specific 

instance. To get the instances used in your environment, you first need to know 

the hardware path (SCSI path) to the device. Use the output of the ioscan 

command in the H/W Path column as shown in Example 2-15 on page 51 to 

determine the SCSI path of your device. In our case, the two drives are 

configured at addresses 0/0.0.0 and 0/0.1.0. Issue the following command: 

Example 2-18 atdd_cfg command to get the instance number 

/opt/OMImag/bin/atdd_cfg -g INSTANCE 0/0.0.0 

INSTANCE: 1 

/opt/OMImag/bin/atdd_cfg -g INSTANCE 0/0.1.0 

INSTANCE: 2 

Our two devices correspond to instance number 1 and 2. Having the instance 

number of the device, you can query the device parameters using the following 

commands: 

Example 2-19 Querying the device parameters 

/opt/OMImag/bin/atdd_cfg -g DENSITY 

/opt/OMImag/bin/atdd_cfg -g SILI 

/opt/OMImag/bin/atdd_cfg -g DEVICES 

/opt/OMImag/bin/atdd_cfg -g BLOCKSIZE 

/opt/OMImag/bin/atdd_cfg -g COMPRESSION 

/opt/OMImag/bin/atdd_cfg -g BUFFERING 

/opt/OMImag/bin/atdd_cfg -g IMMEDIATE 

/opt/OMImag/bin/atdd_cfg -g TRAILER 

/opt/OMImag/bin/atdd_cfg -g ERRNO_LEOT 

And if needed, change the device parameters using the command: 

Example 2-20 Changing the device parameters 

/opt/OMImag/bin/atdd_cfg -s DENSITY 

/opt/OMImag/bin/atdd_cfg -s SILI 

/opt/OMImag/bin/atdd_cfg -s BLOCKSIZE 


opt/OMImag/bin/atdd_cfg -s COMPRESSION 

/opt/OMImag/bin/atdd_cfg -s BUFFERING 

/opt/OMImag/bin/atdd_cfg -s IMMEDIATE 

/opt/OMImag/bin/atdd_cfg -s TRAILER 

/opt/OMImag/bin/atdd_cfg -s ERRNO_LEOT 

This change has effect only for the time until the server is up. If you reboot the 

server any changes made with this command is lost. If you want to set a new 

boot default value for the configuration parameter, modify entries in the atdd.cfg 

file located in /etc/rc.config.d directory. 

The example file, named atdd.cfg.ex can be used to tailor the configuration 

parameters for the selected device that you want to modify. In this example, we 

change the compression from off [0] to on [1] for the device at 

ATDD_HWPATH[0]: 

Example 2-21 Example of the atdd.cfg device configuration file 

# Sample configuration file for IBM SCSI and Ultrium Tape Drive and Medium 

Changer 

#------------------------------------------------------# 

# IBM 3583 Ultrium Scalable Tape Library # 

# SCSI ID: Tape drive 1 (3580): 0 # 

# Tape drive 2 (3580): 1 # 

# Medium Changer (3583): 6 # 

# HP-UX 11.00 A, L, N Class Server # 

#------------------------------------------------------# 

# HW Path for Tape drive 1 

ATDD_HWPATH[0]=0/7/0/0.0.0 

ATDD_COMPRESSION[0]=1 

# HW Path for Tape drive 2 

ATDD_HWPATH[1]=0/7/0/0.1.0 

# HW Path for Medium Changer 

ATDD_HWPATH[2]=0/7/0/0.6.0 

The default for compression is on for every LTO Ultrium device, and it is 

recommended for performance reasons to have the device compression 

activated. Blocksize=0 means that the device uses variable block size. Some 

application software may override this value. Refer to Part 2, “LTO Ultrium with 

applications” on page 145 for specific application software related information. 

2.4.5 Deleting the LTO medium changer 

In some cases, it could be necessary to remove the medium changer from the 

HP-UX kernel configuration. This is a required step for some application 

software that must have a specific device driver for medium changer and the 

IBM-provided Atdd driver for the tape drive. 


Refer to Part 2, “LTO Ultrium with applications” on page 145 for more information 


To remove the medium changer: 

1. Start the System Administration Manager (SAM) by typing sam at the 

command line prompt. 

2. Select Kernel Configuration. 

3. Select Configurable Parameters. 

4. Scroll until you find the configuration parameter atdd_autoch, then press 

Enter. 

5. In the new window, change Formula/Value to 0, then select OK. 

The Pending Value is now 0, as shown in Figure 2-3. 

Figure 2-3 Disabling the medium changer in the kernel configuration 

6. Exit SAM - A window opens asking to Create a New Kernel Now. Select this 

option, and the system will build a new kernel. This may take a few minutes. 

7. A second window opens that allows you to Move Kernel Into Place and 

Shutdown/Reboot System Now. Select this option. The system reboots. 


2.5 Testing the library with tapeutil 

Tapeutil is a tape utility that is installed with the LTO drivers on Solaris and AIX, or 

as a separate package for HP-UX. This utility exercises or tests the functions of 

the tape device and the device driver. It also performs basic tape and medium 

changer operations. The tape utility program provides two versions: the 

interactive menu and the UNIX command line. 

Invoke the tapeutil interactive menu typing tapeutil without any parameters. 

This displays the menu in Figure 2-4. 

General Commands: 

1. Open a Device 5. Inquiry 9.LogSensePage 

2. Close a Device 6. Test Unit Ready 10.ModeSensePage 

3. Device Info 7. Reserve Device 11.ReleaseDevice 

4. Tape Drive Service Aids 8. Request Sense Q. Quit Program 

Medium Changer Commands: 

12. Element Information 16. Move Medium 

13. Position To Element 17. Load/Unload Medium 

14. Element Inventory 18. Initialize Element Status 

15. Exchange Medium 19. Prevent/Allow Medium Removal 

Tape Commands: 

20. Query/Set Parameters 30. Read and Write Tests 

21. Prevent/Allow Medium Removal 31. Unload Tape 

22. Rewind 32. Erase 

23. Erase Gap 33. Write Filemarks 

24. Forward Space Filemarks 34. Backward Space Filemarks 

25. Forward Space Records 35. Backward Space Records 

26. Space to End of Data 36. Query/Set Tape Position 

27. Log Sense 37. Read or Write Files 

28. Display Message 38. Query/Create/Change Partition 

29. Synchronize Buffers 39. Report Density Support 

Enter Selection: 

Figure 2-4 Tapeutil main menu 

A list of general subcommands, medium changer subcommands, and tape 

subcommands is displayed. You must open a device before using these 

commands and operations. To open a device: 

1. Select 1 (Open a Device) from General Commands. 

2. Enter the name of the device special file. Use any special file that exists for 

the device, for example, /dev/rmt0, /dev/rmt0.1, /dev/rmt1.smc,or /dev/smc0. 


This applies to AIX, for the correct device special file name to use with the 

other platforms, refer to Table 2-1 on page 30. 

3. Enter the Read/Write, Read Only, Write Only, or Append mode to open a 

device. These modes apply to the tape devices only, not for the medium 

changer. 

After you open a device, select a command by using the appropriate number for 

the command from the menu. Some commands require additional information 

after they are selected from the menu. 

You can also use the command-line interface to tapeutil by entering commands in 

the format: 

tapeutil -f DeviceName Subcommand [Subcommand ...] 

We give more examples of using the tapeutil command in “Verifying proper 

device attachment with tapeutil” on page 274. Detailed tapeutil reference 

information is in the manual IBM Ultrium Device Drivers Installation and User’s 

Guide, GA32-0430. 

2.6 Fibre Channel host bus adapter driver installation 

Up to now, we have assumed direct SCSI attachment. Alternatively, you can 

connect LTO Ultrium tape libraries and drives through a Storage Area Network, 

using either direct Fibre Channel attach (using native FC drives where available 

or with the 3583 Ultrium Scalable Tape Library with internal SAN Data Gateway 

module), or by using an external SAN Data Gateway with supported Ultrium 

models. 

The tape device driver installation process is the same for both SCSI attached 

and Fibre Channel drives. From the server side, you must have a Fibre Channel 

Host Bus Adapter (HBA) to connect the server to a Storage Area Network fabric. 

The same applies if you have a direct point-to-point Fibre Channel connection 

between a server and an IBM LTO Fibre Channel library. 

The Fibre Channel host bus adapter must be supported and certified by IBM for 

the LTO model that you are connecting to the server. Check the following Web 

sites to verify that your HBA is supported, selecting the appropriate LTO model 

and the corresponding server: 



For LTO models 3580 and 3581, since they are only available through SDG, 

check the following Web site: 


http://www.storage.ibm.com/hardsoft/products/tape/ro3superserver.htm#358x LTO 

The above link points to the support matrix for the IBM SDG 2108-R03 model. 

The following link is for the IBM SDG 2108-G07 model: 

http://www.storage.ibm.com/hardsoft/products/tape/tapesupport.htm 

You can easily find the supported Fibre Channel HBA adapters at the top of the 

picture describing the SDG support, as shown in Figure 2-5. 

Figure 2-5 Where to look for the supported HBA in the SDG matrix 

Once you have carefully verified that your HBA is supported, check the driver 

level and, if necessary, download an updated driver. 

Since the links to check and download the HBA driver are specific for each HBA 

type, you can find this information in the following sections: 

► 2.7.1, “IBM RS/6000 HBA microcode installation” on page 62, 

► 2.8.1, “QLogic QLA2200 HBA driver installation” on page 64 

► 2.8.5, “Emulex LP8000 HBA driver installation” on page 69 


2.7 IBM RS/6000 Fibre Channel HBA installation 

The AIX operating system CD’s provided with your pSeries or RS/6000 include 

the necessary packages to support the RS/6000 Fibre Channel HBA. You should 

check that the following filesets have been installed. If not, then use SMIT to 

install them. If you have the Feature Code 6227 adapter, you only need the 

devices.pci.df1000f7 filesets. If you have the Feature Code 6228 adapter, you 

need both the devices.pci.df1000f7 and devices.pci.df1000f9 filesets. 

Table 2-2 AIX device software for FC HBAs 

Fileset name 

Function 

devices.pci.df1000f7.com 

devices.pci.df1000f7.diag 

devices.pci.df1000f7.rte 

devices.pci.df1000f9.diag 

devices.pci.df1000f9.rte 

FC 6227 Common PCI FC Adapter Device Software 

FC 6227 PCI FC Adapter Device Diagnostics 

FC 6227 PCI FC Adapter Device Software 

FC 6228 PCI FC Adapter Device Diagnostics 

FC 6228 PCI FC Adapter Device Software 

You should also install the FC device filesets to enable support of FC attached 

tape drives. 

Table 2-3 AIX software for FC devices 

Fileset name 

Function 

devices.common.IBM.fc.rte Common IBM FC Software 

devices.fcp.tape.rte FC SCSI Tape Device Software 

Once the base filesets are installed check for the latest fix versions at this URL: 

http://techsupport.services.ibm.com/rs6k/fixdb.html 

Download and install any required fixes. Now you should be able to see your FC 

HBA devices using the following command: 

Example 2-22 Displaying AIX FC adapters 

# lsdev -C|grep fc 

fcs0 Available 10-68 FC Adapter 

fcs1 Available 20-58 FC Adapter 

In our system, we have two FC HBAs. 

Next, check the microcode levels for your HBA at the following link: 

http://www.rs6000.ibm.com/support/micro/download.html#adapter 


Here you find a list of supported HBAs for the RS/6000 server. It does not mean 

that every adapter listed is supported by LTO. For a list of the supported 

adapters, refer to the previous section, 2.6, “Fibre Channel host bus adapter 

driver installation” on page 59. 

Once you have identified the feature code and the related adapter to use, check if 

your adapter has the latest microcode level. Issue the following AIX command: 

lscfg -vl fcs0 

Where, fcs0 is the adapter device name. The output of the command should look 

like this: 

Example 2-23 Output of lscfg command on AIX HBA 

DEVICE LOCATION DESCRIPTION 

fcs0 3A-08 FC Adapter 

Part Number.................09P1162 

EC Level....................D 

Serial Number...............KT05110105 

Manufacturer................0010 

FRU Number..................09P1173 

Network Address.............10000000C9258E54 

ROS Level and ID............02903290 

Device Specific.(Z0)........4002206D 

Device Specific.(Z1)........10020193 



Device Specific.(Z4)........FF101450 




Device Specific.(Z8)........20000000C9258E54 

Device Specific.(Z9)........SS3.22A0 

Device Specific.(ZA)........S1F3.22A0 

ROS Level and ID shows the microcode level, corresponding to the Last Updated 

column of the table in the adapter Web site. 

If the adapter microcode needs to be updated, download the file from the Web 

site with the .bin extension and follow the installation instructions. 

Next, we give an example of the installation process. 

2.7.1 IBM RS/6000 HBA microcode installation 

In this example, we are installing the file df1000f7.bin, corresponding to the 

microcode level 3.22.A0 for adapter FC6227. 


With the server powered on and booted to AIX, perform the following steps to 

install the FLASH code into your adapter: 

1. Log in as root. 

2. Download the file df1000f7.bin into the /tmp directory. This is an AIX zipped 

file that contains the required installation files and instructions. 

3. Copy the file df1000f7.bin from /tmp to the root (/) directory. 

4. Run the following command to make it executable: 

chmod 777 df1000f7.bin 

5. Be careful to follow the RS/6000 microcode download procedure. There is a 

link in the microcode Web site that points to: 

http://www.rs6000.ibm.com/support/micro/downproc.html 

This is where you can obtain the password needed to decompress the zip file. 

6. Execute the file. It prompts you for the unzip password. The self-extracting zip 

file uncompresses the README and the microcode files into the 

/etc/microcode directory. Carefully, read the README file before continuing 

with the installation. 

7. Ensure there is no activity running on the adapter. Vary off all volume groups 

associated with the adapter. 

8. You are now ready to FLASH the EEPROM in the adapter using the single 

command (assuming that fcs0 is the device corresponding to the adapter that 

we are updating): 

diag -c -d fcs0 -T "download -s /etc/microcode -f -l latest" 

This starts the microcode download process. 

9. If the process completes successfully, you see a message indicating success 

and the new level of microcode. Vary on the volume groups associated with 

the adapter. 

The HBA does not need to be configured in most cases. The only important 

parameter that you must check is the Maximum Transfer Size, which must be 

set to 0X100000. 

To check if this parameter is set correctly, use smit. Select Devices -> FC 

Adapter -> FC Adapter -> Change / Show Characteristics of an FC Adapter, 

then select the adapter where the LTO library is connected. Check the Maximum 

Transfer Size parameter and reset if necessary. 


Example 2-24 Smit AIX HBA characteristics 

Change / Show Characteristics of a FC Adapter 




FC Adapter 

fcs0 

Description 

FC Adapter 

Status 

Available 

Location 3A-08 

Maximum number of COMMANDS to queue to the adapter [200] 

Maximum Transfer Size [0x100000] + 

Preferred AL_PA [0x1] + 

Apply change to DATABASE only no + 

2.8 SUN Solaris Fibre Channel HBA driver installation 

For SUN Solaris servers, you need a supported HBA adapter listed in the LTO 

matrix, for more information refer to section 2.6, “Fibre Channel host bus adapter 

driver installation” on page 59, where you find the Web page link. 

The currently supported Fibre Channel HBAs for SUN Solaris are manufactured 

by QLogic and Emulex. The following sections detail the driver installation for 

these host bus adapters. 

2.8.1 QLogic QLA2200 HBA driver installation 

The currently supported QLogic adapter for LTO library and tape drives, is the 

QLA2200. Here is how to check the latest driver level and to install the driver for 

Solaris. 

First, login to the Solaris server as root. Use the following command to display 

the installed adapter driver level: 

pkginfo -l QLA2200-2 

Where, QLA2200-2 is the default name of the QLogic adapter driver package. 

Note: In this example, we assume that you are installing the driver on a SUN 

Solaris SPARC server with operating system version 2.7 and PCI bus. 

If the driver is not installed, you will see the message: 


ERROR: information for "QLA2200-2" was not found 

Otherwise, if the driver is already installed, the output will be like this: 

Example 2-25 Check QLogic device driver level on Solaris 

PKGINST: QLA2200-2 

NAME: QLogic QLA2200 driver 

CATEGORY: system 

ARCH: sparc 

VERSION: Solaris 2.7, Rev=3.07 

BASEDIR: / 

PSTAMP: dvt20010220184351 

INSTDATE: Aug 08 2001 09:44 

STATUS: completely installed 

FILES: 3 installed pathnames 

2 executables 

863 blocks used (approx) 

The currently installed driver level is found in the VERSION field. 

To download the driver or to check if there is a later update available, go to this 

URL: 

http://www.QLogic.com/bbs-html/csg_web/adapter_pages/driver_pages/22xx/22solari 

s.html 

If necessary, download the installation file and the README instructions. Be sure 

to read the instructions file before continuing with the adapter driver installation 

or update. 

If you have an old version of the driver installed, be sure to remove the old 

package before installing the new version. Refer to 2.8.3, “QLogic QLA2200 HBA 

driver removal” on page 68 for instructions. Once you have removed the previous 

driver, or if you are installing the driver from scratch, follow these steps to install 

the new version: 

1. Copy the downloaded compressed package file to the /tmp directory: 

cp /qla2200.Z 

Where, qla2200.Z is the compressed installation package file name. 

2. Uncompress the package file: 

uncompress qla2200.Z 

This command replaces the compressed package file with its uncompressed 

version, qla2200. 


Note: The .Z file extension must be an upper case letter for the 

uncompress command to work correctly. Rename the file to use the .Z 

extension if necessary. 

3. If the Solaris volume manager is running, stop it with the following command: 

/etc/init.d/volmgt stop 

4. Use the pkgadd command to install the adapter driver package: 

pkgadd -d qla2200 

You see the following prompt, where you have to choose the selection 

appropriate to your operating system release (in this example [2]). 

Example 2-26 Install Solaris QLogic driver - select package 


1 QLA2200-1 QLogic QLA2200 driver (sparc) Solaris 2.6, Rev=3.07 




(default: all) [?,??,q]: 2 

After making the appropriate selection, you see the following messages 

during the package installation: 

Example 2-27 Install Solaris QLogic driver - verification messages 


QLogic QLA2200 driver (sparc) Solaris 2.7, Rev=3.07 

Copyright (c) 1996-2000, by QLogic Corporation. All rights reserved. 

Where do you want the driver object installed (default=/kernel/drv): 

## Executing checkinstall script. 










Answer yes, and installation continues with these messages: 

Example 2-28 Install Solaris QLogic driver - installation messages 

Installing QLogic QLA2200 driver as 


/kernel/drv/qla2200 


kernel/drv/qla2200.conf 

/kernel/drv/sparcv9/qla2200 




5. The installation of the QLogic QLA2200 adapter driver is successful. You 

need to reboot the server to complete the installation. Use the following 

command: 

reboot -- -r 

After the reboot, you next configure the HBA driver. 

2.8.2 QLogic QLA2200 HBA driver configuration 

Locate the adapter driver configuration file, named qla2200.conf, in the 

/kernel/drv directory. 

Edit the file with your favorite editor, and check the setting of the 

hba0-connection-options parameter: 

Example 2-29 QLogic Solaris HBA connection options parameter 

# Connection options 

# 0 = loop only 

# 1 = point-to-point only 

# 2 = loop preferred, otherwise point-to-point 

hba0-connection-options=1; 

We are connecting the LTO library to the Sun server in a Storage Area Network 

through an IBM SAN Data Gateway. Point to point (value 1) is the correct option. 

Next check that hba0-fc-tape is set on (1) to enable support of Fibre Channel 

tapes. 

Example 2-30 QLogic Solaris HBA FC tape support parameter 

# Enable/disable adapter support of fibre channel tape. 

# 0 = disable, 1 = enabled 

hba0-fc-tape=1; 

If you are connecting the adapter to a fabric switch in a Storage Area Network 

environment, you may need to change other adapter configuration settings. Refer 

to 3.3.2, “Persistent binding with QLogic HBA on Solaris” on page 118 for more 

information. 


2.8.3 QLogic QLA2200 HBA driver removal 

If you need to remove a previous version of the driver, use the following 

command: 

pkgrm QLA2200-2 

Where, QLA2200-2 is the currently installed package name. 

You will see the following output: 

Example 2-31 Deinstalling QLogic Solaris HBA driver 


QLA2200-2 QLogic QLA2200 driver(sparc) Solaris 2.7, Rev=3.07 









/kernel/drv/sparcv9/qla2200 

/kernel/drv/qla2200.conf 

/kernel/drv/qla2200 

## Executing postremove script. 

Device busy 

Cannot unload module: qla2200 

Will be unloaded upon reboot. 

## Updating system information. 

Removal of was successful. 

After removing the driver, you may have to reboot the system if indicated. 

However, you do not need to reboot at this time if you are only removing the 

driver before installing a new version. 

2.8.4 QLogic HBA FCode 

QLogic HBAs (QLA2200 and higher) also have FCode (like microcode) which 

can be downloaded from the QLogic Web site and installed into flash ROM from 

the Solaris operating system. You need to install the FCode to provide the ability 

to boot the Solaris systems from devices attached to the HBA. Since this is only 

relevant for disk devices, we do not document how to flash the FCode. 

Instructions and the FCode for the QLA2200 HBA are available at: 

http://www.qlogic.com/bbs-html/csg_web/adapter_pages/driver_pages/22xx/22solari 

s.html 


2.8.5 Emulex LP8000 HBA driver installation 

The currently supported Emulex adapter for LTO library and tape drives, is the 

LP8000. Depending on the bus type installed on the server, you need the 

following device driver: 

► LP8000 for PCI bus adapter 

► LP8000S for Sbus adapter 

First, login to Solaris as root. Use the following command to display the installed 

adapter driver level: 

pkginfo -l lpfc 

Where, lpfc is the default name of the Emulex adapter driver package. 

Note: In this example, we assume that you are installing the driver on a SUN 

Solaris SPARC server with operating system version 2.7 and PCI bus. 

If the driver is not installed, you will see the message: 

ERROR: information for "lpfc" was not found 

Otherwise, if the driver is already installed, the output will be like this: 

Example 2-32 Check Emulex device driver level on Solaris 

PKGINST: lpfc 

NAME: Emulex LightPulse FC SCSI/IP Host Bus Adapter driver 

CATEGORY: system 

ARCH: sun4u 

VERSION: Release 4.20k 

BASEDIR: / 

PSTAMP: ultrapci20010531121050 

INSTDATE: Aug 10 2001 10:37 

STATUS: completely installed 

FILES: 44 installed pathnames 

14 shared pathnames 

12 directories 

13 executables 

8013 blocks used (approx) 

The currently installed driver level is found in the VERSION field. 

To download the driver or to check if there is a later update available, go to this 

URL for the LP8000 adapter (PCI): 

http://www.emulex.com/ts/fc/docs/sol/420k/rm420k.htm 


And for LP8000S adapter (Sbus) check: 

http://www.emulex.com/ts/fc/docs/oboot/sbus/rm_210.htm 

If an upgrade of the driver is required, or if you are installing the driver for the first 

time, download the installation file. 

You find detailed installation instructions for the LP8000 adapter (PCI) at the 

following link: 

http://www.emulex.com/ts/fc/docs/frame8k.htm 

And for the LP8000S adapter (Sbus), find the instructions at this link: 

http://www.emulex.com/ts/fc/docs/frame8ks.htm 

Click on the Drivers for Solaris link and you will find the installation information in 

the Documentation column. Be sure to read this information before continuing 

the driver installation. 

If you have an old version of the driver installed, remove the old package before 

installing the newer version. Refer to 2.8.7, “Emulex LP8000 HBA driver removal” 

on page 73 for instructions. Once you have removed the previous driver, or if you 

are installing the driver from scratch, follow these steps to install the new version: 

1. Create a temporary directory (for example, emlxtemp): 

mkdir emlxtemp 

2. Change directory to the temporary directory: 

cd emlxtemp 

3. Copy or download the device driver file to the temporary directory then untar 

it: 

tar xvf lpfc-sparc.tar 

Where, lpfc-sparc.tar is the compressed adapter driver package file. This 

command creates subdirectories and unpacks the installation files. 

4. Install the package using the command: 

pkgadd -d `pwd` 

And, the installation script will prompt you to respond: 

Example 2-33 Install Solaris Emulex driver - select package 


1 lpfc.1 Emulex LightPulse FC SCSI/IP Host Bus Adapter driver 

(sun4u) Release 4.20k 


(default: all) [?,??,q]: 


Simply, press Enter. These messages follow: 

Example 2-34 Install Solaris Emulex driver - install man pages 


Emulex LightPulse FC SCSI/IP Host Bus Adapter driver(sun4u) Release 4.20k 

lpfc (FCP) / lpfn (IP) combo driver 

Copyright (c) 1997, 1998, 1999, 2000 Emulex Corporation 

3535 Harbor Boulevard, Costa Mesa, CA 92626 

All rights reserved. This product and related documentation is protected 

by copyright and distributed under licenses restricting its use, copying, 

distribution and decompilation. No part of this product or related 

documentation may be reproduced in any form by any means without prior 

written authorization of Emulex Corporation and its licensors, if any. 

CAUTION: The lpfc driver will not work with any release of Solaris/SunOS 

earlier than 2.6/5.6. However, even for later releases, lpfc may have to 

be compiled specifically for that release of Solaris. This version of lpfc 

has been built under Solaris/SunOS 5.7 for sun4u. 

Rebuild manual pages database for section 7d [y,n,?] y 

Answer yes [y] if you want to include the driver information in the man pages. 

Use IP networking over Fibre Channel [y,n,?] n 

Answer no [n] if you do not want to activate the IP protocol over Fibre 

Channel. This protocol is not used by LTO. 

Example 2-35 Install Solaris Emulex driver - installation messages 




6 package pathnames are already properly installed. 







Installing Emulex LightPulse FC SCSI/IP Host Bus Adapter driver as 



/usr/share/man/man7d/lpfc.7d 

/usr/share/man/man7d/lpfn.7d 


/kernel/drv/lpfc 

/kernel/drv/lpfc.conf 

/kernel/drv/lpfc.hotplug 

/kernel/drv/sparcv9/lpfc 

/kernel/drv/sparcv9/lpfc.hotplug 



usr/include/fcdiag.h 

/usr/lib/libHBAAPI.so 

/usr/lib/libHBAAPIV6.so 

/usr/lib/libdfc.a 

/usr/lib/libemulexhbaapi.so 

/usr/lib/libemulexhbaapiV6.so 

/usr/lib/sparcv9/libHBAAPI.so 

/usr/lib/sparcv9/libdfc.a 

/usr/lib/sparcv9/libemulexhbaapi.so 

/usr/sbin/lpfc/CT_fmw 

/usr/sbin/lpfc/DD_fmw 

/usr/sbin/lpfc/REV_fmw 

/usr/sbin/lpfc/SF_fmw 

/usr/sbin/lpfc/convert_path_lpfc 

/usr/sbin/lpfc/dfc 

/usr/sbin/lpfc/dfc32 


/usr/sbin/lpfc/download_fmw_lpfc 

/usr/sbin/lpfc/lputil 

/usr/sbin/lpfc/lputil32 


/usr/sbin/lpfc/resetqdepth 


Modifying /etc/path_to_inst 

Modifying /etc/system 

Modifying /kernel/drv/sd.conf 



cat: cannot open //etc/hba.conf 

Updating /kernel/drv/lpfc.conf 

Updating /etc/system (moddir) 

catman -M //usr/share/man -w 7d 

This may take a while ... 

Driver lpfc added to system. 

Adding emulexapilibrary from //etc/hba.conf************ 

IMPORTANT: You must reboot the system to start this driver. 

SCSI: If you are using lpfc to access disks, be sure to check the 

configuration file of your SCSI target driver (presumably sd.conf) to 

ensure that the driver will probe for all of the targets/luns in your 

environment. 

HOTPLUG: If you wish to run with HOTPLUG enabled you must manually: 

cp //kernel/drv/sparcv9/lpfc.hotplug //kernel/drv/sparcv9/lpfc 

NOTE: lpfc supports FCP and IP but wasn't configured for networking. 

ALL: Examine the contents of /kernel/drv/lpfc.conf for 

driver-specific variables and behavior.************ 


5. The installation of the Emulex LP8000 adapter driver is successful. You need 

to reboot the server to complete the installation. Use the following command: 


eboot -- -r 

Now you need to configure the driver. 

2.8.6 Emulex LP8000 HBA driver configuration 

Locate the adapter driver configuration file, named lpfc.conf, in the /kernel/drv 

directory. 

Edit the file with your favorite editor, and check the topology parameter. 

Example 2-36 Emulex Solaris HBA topology parameter 

# topology: link topology for initializing the Fibre Channel connection. 

# 0 = attempt loop mode, if it fails attempt point-to-point mode 

# 2 = attempt point-to-point mode only 

# 4 = attempt loop mode only 

# 6 = attempt point-to-point mode, if it fails attempt loop mode 

# Set point-to-point mode if you want to run as an N_Port. 

# Set loop mode if you want to run as an NL_Port. 

topology=2; 

We are connecting the LTO library to the SUN server in a Storage Area Network 

environment through an IBM SAN Data Gateway. Point to point (2) is the correct 

option. 

If you are connecting the adapter to a fabric switch in a Storage Area Network 

environment, you may need to change other configuration settings. Refer to 

3.3.3, “Persisting binding with Emulex HBA on Solaris” on page 121 for more 

information. 

2.8.7 Emulex LP8000 HBA driver removal 

If you need to remove a previous version of the driver, use the following 

command: 

pkgrm lpfc 

Where, lpfc is the currently installed package name. Answer yes [y] when 

prompted and you will see the following output: 

Example 2-37 Deinstalling Emulex Solaris HBA driver 


lpfc 

Emulex LightPulse FC SCSI/IP Host Bus Adapter driver (sun4u) 

Release 4.20k 









## Executing preremove script. 

NOTE: Saving sd.conf lpfc.conf in //usr/tmp as file.pkgrm 


Modifying /kernel/drv/sd.conf 

Modifying /etc/system 

Modifying /etc/path_to_inst 


/usr/sbin/lpfc/resetqdepth 



/usr/sbin/lpfc/lputil 

/usr/sbin/lpfc/download_fmw_lpfc 



/usr/sbin/lpfc/dfc 

/usr/sbin/lpfc/convert_path_lpfc 

/usr/sbin/lpfc/SF_fmw 

/usr/sbin/lpfc/REV_fmw 

/usr/sbin/lpfc/DD_fmw 

/usr/sbin/lpfc/CT_fmw 

/usr/lib/sparcv9/libemulexhbaapi.so 

/usr/lib/sparcv9/libdfc.a 

/usr/lib/sparcv9/libHBAAPI.so 

/usr/lib/libemulexhbaapiV6.so 

/usr/lib/libemulexhbaapi.so 

/usr/lib/libdfc.a 

/usr/lib/libHBAAPIV6.so 

/usr/lib/libHBAAPI.so 

/usr/include/fcdiag.h 


/kernel/drv/sparcv9/lpfc.hotplug 

/kernel/drv/sparcv9/lpfc 

/kernel/drv/lpfc.hotplug 

/kernel/drv/lpfc.conf 

/kernel/drv/lpfc 


/usr/share/man/man7d/lpfn.7d 

/usr/share/man/man7d/lpfc.7d 


/usr/share/man/man7d 

/usr/share/man 

/usr/share 

/usr/sbin/lpfc 

/usr/sbin 


usr/lib 

/usr/include 

/usr 

/kernel/drv/sparcv9 

/kernel/drv 

/kernel 

/etc 

## Executing postremove script. 

/etc/hostname.lpfn0 

Removing emulexapilibrary from //etc/hba.conf 

NOTE: forceload: drv/clone has been left in //etc/system 

## Updating system information. 

Removal of was successful 

After removing driver you may have to reboot the system if indicated. However, 

you do not need to reboot at this time if you are only removing the driver before 

installing a new version. 

2.8.8 Emulex HBA firmware and bootcode 

You may need to update the firmware on the Emulex HBA. Download the 

firmware image from the Web site (URL below) and copy it to an accessible 

directory on the Solaris system. Execute the lputil command located in the 

/usr/bin/lpfc directory. You must have already installed the Emulex device driver 

to have this command available. Select option 3. Firmware Maintenance, then 1. 

Load Firmware Image. This will transfer the firmware to flash ROM. 

Bootcode is also available for this adapter which enables you to designate a 

Fibre Channel attached disk drive as a system boot device. Since this is only 

relevant for disk devices, we do not document how to load the bootcode. 

Instructions plus the firmware and bootcode for the Emulex LP8000 HBA are 

available at: 

http://www.emulex.com/ts/fc/docs/frame8k.htm 

2.9 StorWatch Tape Library Specialist 

The StorWatch Ultrium Tape Library Specialist is part of the IBM StorWatch 

products family. 

StorWatch — IBM's Enterprise Storage Resource Management (ESRM) solution 

— is a growing software family whose goal is to enable storage administrators to 

efficiently manage storage resources from any location within an enterprise. 


It is available at no charge for currently shipping models of the 3583 Ultrium 

Scalable Tape Library and the 3584 Ultrium UltraScalable Tape Library. 

This product consists of a Web user interface accessible from any 

browser-equipped host on the network. You also require the RMU or Ethernet 

connection on the 3583 Ultrium Scalable Tape Library and the 3584 Ultrium 

UltraScalable Tape Library, respectively. As described in 1.2.5, “IBM 3583 

Ultrium Scalable Tape Library” on page 10 and 1.2.6, “IBM 3584 Ultrium 

UltraScalable Tape Library” on page 12, these are available as priced upgrades 

for previously shipped libraries which do not have them. 

In the next sections we describe in more detail the StorWatch Specialist 

applications for the 3583 and 3584. 

2.9.1 IBM 3583 StorWatch Specialist 

The StorWatch Ultrium Tape Library Specialist for the 3583 library is included 

with the RMU (Remote Management Unit). The RMU may have been shipped 

with your library - if not, then it is available as an upgrade - see 1.2.5, “IBM 3583 

Ultrium Scalable Tape Library” on page 10 for more information. 

The RMU provides remote access to the library over a network. You can attach 

the library to your network through a 10/100 Ethernet port on the RMU and 

access the StorWatch functions through the browser interface. 

If you have already installed the RMU, then you need to connect the Ethernet 

port of the RMU unit to your network. 

Using the 3583 operator panel, select More -> Setup -> Library -> RMU. The 

panel in Figure 2-6 is displayed. 

Figure 2-6 3583 operator panel: configuring RMU 

Use the push buttons Up and Down to enter the IP address, subnet address, 

gateway (if present) and hostname. When finished, press OK. 


You should now be able to access the Specialist from your browser by entering in 

the IP address (for example, http://193.1.1.70). Figure 2-7 shows the welcome 

window of the 3583 Tape Library Specialist. 

Figure 2-7 3583 Specialist: welcome panel 

The first time you use the Specialist, login as admin with the password secure. 

Then you can access the Configuration menu and add the users that you want 

to grant access to the Specialist. Remember that every user, authorized to 

access the Specialist panels, can potentially access every option available on the 

3583 operator panel. 

Use the menu bar at the top of the panel, right below the title bar, to navigate 

between the menus. You can choose between the following options: 

Status 

Library Status 

Drive Status 

RMU User 

Hostname 

IP address 

MAC address 


Configuration 

Firmware 

Diagnostic file 

Operator panel 

Logs 

Library Serial # 

SNMP Alerts 

Library Firmware level 

RMU Firmware level 

Network Configuration 

SNMP Configuration 

User Configuration 

Date and Time 

Update Library Firmware 

Update RMU Firmware 

Update Drive Firmware 

Library Command log 

Library Error log 

RMU Support log 

RMU Error log 

access every option available on the 3583 operator panel 

(this is a graphical interface updated every few seconds) 

view the current command log 

In the black column at the left hand side of the Specialist window, you find the 

help and documentation links that allow you to directly access the 3583 

publications. 

Next, we show some example of the 3583 Ultrium Tape Library Specialist panels. 

In the configuration panel shown in Figure 2-8, you can set the network 

configuration parameters, as well as Simple Network Management Protocol 

(SNMP) settings to send the alerts generated by the RMU to an SNMP server in 

your private network. We will describe setting up SNMP for the 3583 Ultrium 

Scalable Tape Library in “Simple Network Management Protocol (SNMP) alerts” 

on page 280. The configuration panel also provides user management for the 

Specialist interface. 


Figure 2-8 3583 Specialist: configuration panel 

In the Diagnostics file panel shown in Figure 2-9 you can view and download the 

library and RMU logs, to possibly debug error situations and send the required 

documentation to the IBM Support Center. 


Figure 2-9 3583 Specialist: diagnostics file panel 

In the Operator panel shown in Figure 2-10, you can operate the library exactly 

the same way as if you were at the real 3583 operator panel. The graphic display 

showing the image of the 3583 LCD panel has exactly the same functions as the 

physical library display. This feature allows you to view or update the 

configuration, reboot the library, vary drives offline and all other tasks from any 

browser with TCP/IP access to the library. 


Figure 2-10 3583 Specialist: operator panel 

2.9.2 IBM 3584 StorWatch Specialist 

The StorWatch Ultrium Tape Library Specialist for the 3584 library is included 

with the Ethernet support. Ethernet support may have been included with your 

library. If not, then it is available as an upgrade. See 1.2.6, “IBM 3584 Ultrium 

UltraScalable Tape Library” on page 12 for more information. 

You will need to attach the Ethernet port to your network with a suitable cable and 

configure the TCP/IP addressing information. Choose Settings -> Network 

Settings -> Ethernet. You will see the current interface MAC address (which 

cannot be changed), and the assigned TCP/IP address, subnet mask and 

gateway. If the library has more than one frame, each frame requires a separate 

address. Use the UP and DOWN buttons to access the panels for the additional 

frames. The current configuration window is shown in Figure 2-11. 


Panel0175 

Ethernet 

______________________________ 

CurrentSettingsFrame1: 

MACAddress: 18:36:F3:98:4F:9A 

IPAddresses:10.1.1.1 

SubnetMask: 255.255.255.0 

Gateway: 10.1.1.254 

[ChangeSettings] 

______________________________ 

BACK UP DOWN ENTER 

Figure 2-11 Change Ethernet parameters on 3584 

Press Enter to make changes. You can select to disable the Ethernet interface, 

use DHCP to automatically assign an address (if supported in your network) or 

manually configure the parameters. 

Once you have configured the network connection, to use the 3584 StorWatch 

Specialist, enter the TCP/IP address of the library in your browser (for example, 

http://10.1.1.1). 

The introductory window of the 3584 Specialist Web interface displays as shown 

in Figure 2-12. The window lets you access library functions and provides 

information about using Help. It also offers a Getting Started link that gives an 

introduction to the interface and a Reference Information link that describes the 

library. 


Figure 2-12 3854 Specialist: Welcome window 

Use the left hand side of the panel to navigate between the menus. You can 

choose between the following options: 

Physical Library 

Logical Libraries 

Manual Operations 

Cartridges 

I/O Station 

Drives 

Accessor 

Node cards 

Storage Slots 

Control Ports 

Cartridges 

Drives 

Storage Slots 

Inventory Library 

Clean Drive 

Move Cartridge 


Settings 

Service 

Insert Cleaning Cartridge 

Remove Cartridge 

Security 

Library Configuration 

World Wide Names 

Control Paths 

Cleaning Mode 

Date and Time 

Drive SCSI/Loop IDs 

Control Port SCSI IDs 

Vital Product Data 

Download Logs 

Firmware Update 

The StorWatch Ultrium Tape Library Specialist is a complete interface that allows 

you to easily monitor the library operations, with graphics and tables like those 

shown in Figure 2-13. 


Figure 2-13 3584 Specialist: Physical Library entry panel 

Other options are available to manage the library, as seen in Figure 2-14. Here, 

after selecting a tape drive you can clean it, change its SCSI ID or view usage 

statistics. 


Figure 2-14 3854 Specialist: Logical Libraries drives panel 

In addition, you can configure the library directly from the Specialist panels, 

specifying the SAN parameters, or the SCSI control paths, control ports as well 

as security options, as shown in Figure 2-15. Note that the TCP/IP and SNMP 

configuration functions are not available from the StorWatch Specialist. 


Figure 2-15 3584 Specialist: library settings panel 

A password is required to access some functions of the Specialist, such as 

firmware update and the modification of the library configuration. 


2.10 Updating library, drive and RMU firmware 

Once you have completed the physical installation of the LTO library or drive, we 

recommend that you check for any higher level of firmware for the LTO drive, the 

LTO library and the optional Remote Management Unit (RMU), before beginning 

to use the devices. 

It is not mandatory to upgrade the microcode, but it is a good habit to check from 

time to time for available updates at the IBM Web site, to see if any new code has 

more support functions or has solved problems. To upgrade the installed 

microcode level of your library, refer to the appropriate Operator Guide for your 

library or drive. These are listed in “Other resources” on page 313. In our 

example, we describe the procedures to check and update the library and drive 

microcode for a 3583 Ultrium Scalable Tape Library. 

First, check the installed microcode level from the 3583 operator panel: 

► For the library code, select More -> About. The Version field displays the 

microcode level. 

► For the drive code, select Status -> Drives. The Version field displays the 

microcode level. 

To check for microcode updates, use the following links: 

http://SSDDOM01.storage.ibm.com/techsup/swtechsup.nsf/support/ultriumfmr_ftp 

Or 

ftp://ftp.software.ibm.com/storage/358x/ 

Here you can find the latest microcode releases for the LTO library, drives and 

RMU. 

For the library microcode, use the compressed file appcode.exe for Windows 

systems, or appcode.tar for UNIX systems. It contains the Vn_n_n.lif file that is 

the microcode image file to upload in the library. For the drive code, download the 

latest .fmr file. 

2.11 How to update the firmware 

You can use the following methods to update the firmware: 

► Using the Remote Management Unit (RMU) through the StorWatch Specialist 

► Using the SCSI bus 

► Using the Field Microcode Replacement (FMR) tape (drive firmware only) 


► 

► 

Using the library serial port (library firmware only - this is only done by the 

IBM Customer Engineer) 

Using the 3583 Integrated SAN Data Gateway module (optional) 

You can find detailed instructions on how to update the firmware at the following 

link (or in the relevant library Setup and Operator Guide): 

http://SSDDOM01.storage.ibm.com/techsup/swtechsup.nsf/support/ultriumfirmware 

If the 3583 library has the optional SAN Data Gateway module installed you can 

update the firmware using the gateway. For more information, see IBM Storage 

Area Network Data Gateway Module Setup, Operator and Service Guide, 

GA32-0436. 

Note: Before updating firmware on the drives, the library, or the RMU, vary the 

library OFFLINE to ALL attached hosts. 

As an example, next we describe a practical procedure that shows the detailed 

steps required to upgrade the library and drives microcode for the 3583 Ultrium 

Scalable Tape Library, through the SCSI bus interface using the tapeutil utility 

for AIX. Then we document the steps required to update the RMU code from the 

StorWatch Specialist panels (for more information on the StorWatch Specialist, 

see 2.9, “StorWatch Tape Library Specialist” on page 75). 

The process is similar in the other UNIX environments like HP-UX and SUN 

Solaris. 

2.11.1 Upgrading library firmware using tapeutil 

The following procedure describes how to update the library firmware in the 3583 

Ultrium Scalable Tape Library, over the SCSI bus using the tapeutil utility. In 

brackets [ ] you find the tapeutil selection to use during the process. 

1. Download the latest library firmware from: 

ftp.software.ibm.com/storage/358x,to a directory 

2. Open a tapeutil session from the directory, where the library firmware image 

was saved, by typing tapeutil at a command prompt. 

3. Open the /dev/rmt/X device you want to upgrade [1]. 

4. Verify connectivity by initializing an inquiry [5]. 

5. Set the 3583 Tape Library to FIRMWAREUPDATE to accept new firmware at 

the operator panel. To do this, select More -> Service -> Library -> 

Firmware Update. Press OK to reboot the 3583 Tape Library and continue. 


6. The 3583 Tape Library operator panel shows the message One moment, the 

library is booting.... This procedure could take up to 2 or 3 minutes to switch 

over. 

7. On tapeutil select Tape drive service aids [4]. 

8. On tapeutil select Microcode Load. 

9. Choose the smcx device you want to upgrade and press Enter. 

10.The next window ask you what directory and microcode name to use. Once 

entered, press F7 to commit. 

11.This process takes about 5 to 10 minutes. The 3583 Library then reboots and 

continues its normal startup process. You then need to close the tapeutil 

session, because the firmware update process terminates communications to 

the drive. 

To verify that the firmware has installed correctly, do the following: 

1. On the library panel, press More -> About. 

2. Verify that the version information displayed matches the information that you 

have in the README file of the microcode package that you downloaded. 

In Figure 2-16 you can see an example of the 3583 Ultrium Scalable Tape Library 

display panel showing the library microcode version. 

Figure 2-16 The 3583 tape library panel with the microcode level information 

2.11.2 Upgrading drive firmware using tapeutil 

The following procedure describes how to update the drive firmware, over the 

SCSI bus using the tapeutil utility. In brackets [ ] you find the tapeutil selection to 

use during the process. 

1. Download the latest drive firmware (xxxx.fmr) to a local directory from: 

ftp://ftp.software.ibm.com/storage/358x 

2. Open a tapeutil session from the directory, where the library firmware image 

was saved, by typing tapeutil at a command prompt. 

3. Open the /dev/rmtx device you want to upgrade [1]. 


4. Verify connectivity by initializing an inquiry [5]. 

5. On tapeutil select Tape drive service aids [4]. 

6. Select Microcode load. 

7. Choose the /dev/rmtx device you want to upgrade. 

Once completed, verify that the drive is at the new firmware level by issuing a 

inquiry the drive [5]. The output of the inquiry command should look like 

Figure 2-17. 

I n q u i r y D a t a , L e n g t h 3 8 

0 2 3 4 1 5 6 7 8 9 A B C D E F 0 1 2 3 4 5 6 7 8 9 A B C D E F 

1 8 0 3 0 2 1 0 0 1 3 0 9 4 2 D 2 0 0 2 0 0 2 0 . € . . ! . . 0 I B M 0 0 0 0 - 0 0 2 0 4 4 2 2 [ ] 

0 0 1 0 - 5 5 4 C 5 4 3 3 3 5 3 8 3 0 2 D 5 4 4 4 3 1 2 0 2 0 2 0 2 0 2 0 [ U L T 3 5 8 0 - T D 1 ] 

0 0 2 0 - 3 1 3 6 4 5 3 0 0 0 0 0 [ 1 6 E 0 . . ] 

D r iv e 

m i c r o c o d e 

le v e l 

Figure 2-17 Tapeutil inquiry on tape drive 

2.11.3 Upgrading RMU firmware using the StorWatch Specialist 

To update the RMU firmware, use the StorWatch Specialist panels provided with 

the RMU unit installed. 

1. Visit ftp.software.ibm.com/storage/358x, to download the latest drive firmware 

(xxxx.fmr) to a directory (it is usually called rmucode.exe). 

2. Open the RMU Web page and select the firmware tab (login if necessary). 

Put the library offline. 

3. Select Update RMU Firmware. Then enter the path and firmware filename or 

browse for the downloaded firmware file. Then select Update Firmware, as in 

Figure 2-18. 


Figure 2-18 Ultrium Tape Library Specialist: updating RMU firmware 

4. Check the StorWatch Ultrium Tape Library Specialist main panel for the 

updated RMU firmware level information after the successful completion of 

the operation. 


3 

Chapter 3. 

SAN setup 

In this chapter we describe considerations for LTO libraries which are connected 

to a SAN. 

We assume the reader is already familiar with the basic concepts of Storage 

Area Networks. We will list some good resources for acquiring this information in 

3.1, “SAN design considerations for LTO libraries” on page 94. In this chapter we 

are emphasizing specific issues relevant to SAN tape attachment including: 

► Design considerations for SAN-attached tape libraries 

► Switch fabric zoning 

► Persistent binding 

► The IBM SAN Data Gateway 

In addition we provide detailed information on SAN setup for the 3583 Ultrium 

Scalable Tape Library and 3584 Ultrium UltraScalable Tape Library. 

Note: In our redbook, we document SAN attachment for the 3583 using the 

IBM SAN Data Gateway 2108-R03. The setup when using the new SAN Data 

Gateway Module, Feature Code 8005 for the 3583 is very similar. More 

information on this configuration is available in IBM Storage Area Network 

Data Gateway Module, GA32-0436. 


3.1 SAN design considerations for LTO libraries 

In this section we cover the following LTO SAN design considerations: 

► How many tape drives can I connect to a FC HBA? 

► How should I connect tape drives to a SAN Data Gateway (SDG)? 

► How to handle multiple paths to tape drives 

For more Information about general SAN design, please refer to: 

► Designing an IBM Storage Area Network, SG24-5758 

► IBM SAN Survival Guide, SG24-6143 

Currently all tape drives with native FC connections use FC-AL protocol. If you 

connect one tape drive to a port of the SAN then this port has to support the 

FC-AL protocol also. Therefore not all possible connections will be supported. 

3.1.1 How many tape drives can I attach to a FC HBA? 

With FC you cannot daisy-chain drives as you can with a SCSI connection. Each 

drive is connected individually to the SAN, either directly if you are using native 

FC drives or via a SCSI to FC gateway. If you want to attach one drive to a single 

server, you can cable it directly to the HBA - this is called a point-to-point 

connection. Otherwise, if you have more than one drive and/or more than one 

server, you need additional components such as a SAN switch or hub. 

Using a SAN fabric configuration you can potentially attach and access many 

tape drives through one or more HBA’s. But how will this affect performance? 

Restriction: The IBM Fibre Channel Storage Hub (2103-H07) is supported for 

distance solutions beyond 500m and up to 10Km. The hub is used for 

conversion between short wave and long wave optical only. Multiple initiators 

or targets on the same loop are not supported. 

The theoretical maximum data transfer rate for one FC connection in a SAN is 

100 MB/s. In reality we typically see an effective data transfer rate of about 80 

MB/s. LTO tape drives have a sustained data transfer rate of 15 MB/s native, and 

30 MB/s using compression. This means you should therefore be able to access 

two to five tape drives concurrently through one HBA. However, it is quite 

dependent on the type of data. If you write very large files with a good 

compression ratio you should get a transfer rate of about 30 MB/s. In that case 

you should only access 2 or 3 drives on one HBA. On the other hand, if you write 


small files or files with lower compression then you will see a transfer rate of less 

than 15 MB/s, which will enable you to connect more drives to one HBA. 

Nevertheless, we do not recommend for you to try to access more than eight 

drives with a single HBA. 

You can check the real data transfer rate achieved on the SAN using software 

tools provided with your SAN fabric hardware. For example, if you have an IBM 

SAN Fibre Channel Switch, the 2109, you can use the StorWatch Fibre Channel 

Switch Specialist. Enter the TCP/IP address of one of the switches in your Web 

browser and click on the switch you want to see (Figure 3-1). 

Figure 3-1 IBM 2109 StorWatch - fabric view 

On the Switch Management Application (Figure 3-2), click on the performance 

icon in order to the see the performance window. 

Chapter 3. SAN setup 95

Tip: If you have problems connecting to the switch then check the settings 

from your browser about proxy. Disable your proxy settings (use direct 

connection to the internet) and try it again. 

Figure 3-2 Switch Management window 

In Figure 3-3 you see the performance panel from a switch in our ITSO lab. The 

display shows there is some traffic on Ports 0 and 7 of around 15 MB/s. If you 

find that one port is showing high traffic from around 80 MB/s or more then you 

should consider reconfiguring your fabric by adding additional HBA’s or extra FC 

attachments to the SAN Data Gateway. 


Figure 3-3 Performance/data transfer rate from a SAN switch/FC 

3.1.2 Connecting tape drives to a SDG 

The effective data transfer throughput for an IBM SAN Data Gateway 2108 G07 

is about 100 MB/s. For a SAN Data Gateway 2108 R03, the throughput is about 

50 MB/s. 


Our suggestion for achieving the best effective data rate on a SDG is: 

► Do not daisy-chain drives. If this is necessary (for example, if there are more 

than three drives to be connected), then do not connect more than two drives 

to one SCSI Bus. 

► Install a low number of drives to one SDG. 

► Attach the SCSI medium changer to its own separate port if possible. 

In Figure 3-4 you see an example of a connection with SDG and a 3583. The 

3583 contains six LTO drives. Two drives share each SCSI connection. The SCSI 

Medium Changer attaches to its own SCSI bus. The server is connected to the 

SDG using two HBAs and FC ports. If you have more drives to attach you may 

choose not to dedicate a SCSI port to the SCSI medium changer. 

3583 Lxx 

Server 

Medium Changer 

F 

E 

D 

C 

B 

A 

Fibre 

Channel 

Aapter 

SDG Module 

Tape Library 

with 6 drives 

Figure 3-4 Example of a connection with SDG 

3.1.3 Multiple paths to tape drives 

We have discussed the reasons for using more than one HBA on the server to 

give good performance for all drives. However, if you implement this configuration 

you will probably see duplicate tape drive definitions on your server. Let us 

discuss this in the context of our lab configuration. 

We have one RS/6000 with two FC HBAs. These HBAs are connected to a SAN 

Fabric consisting of IBM 2109 switches. We have a 3583 Ultrium Scalable Tape 

Library with two drives installed. The 3583 is connected to a SAN Data Gateway. 

The SDG is connected to the fabric with one FC port. See Figure 3-5. 


3583 Lxx 

RS/6K 

Server 

Fibre 

Channel 

Aapter 

2109 FC 

Switches 


B 

Tape Library 


SDG 

WWN: 

2001006045161ff5 

A 

Figure 3-5 Our lab configuration 

If you boot the RS/6000 or run cfgmgr to detect the new devices, when you list 

the tapes found (use the lsdev command) you will get output as in Example 3-1. 

Example 3-1 Duplicate device names because of dual pathing 

# lsdev -Cc tape 

smc0 Available 20-58-01-6,0 IBM 3583 Library Medium Changer (FCP) 

rmt0 Available 20-58-01-0,0 IBM 3580 Ultrium Tape Drive (FCP) 





You can see that the medium changer and two tape drives are each detected and 

defined twice — once on each HBA in the RS/6000, giving a total of four tape 

drives and two SCSI Medium Changers. This gives the impression there are 

more devices available than are actually physically present. This is not a good 

situation. There are at least three ways to solve this problem: 

1. You can enable zoning which allows one HBA to see only certain tape drives. 

2. You can use persistent binding for the HBAs to see only certain WWN. 

3. You can use the configuration options of the SAN Data Gateway, such as 

channel zoning. 

Note: Option 3 is only available if drives are connected to a SAN Data 

Gateway. Options 1 and 2 are available for both native FC drives and for SCSI 

drives which are fabric connected via a SAN Data Gateway. 

The different options are described in the following section. 


3.2 Zoning 

Zoning allows you to partition your SAN into logical groupings of devices so that 

each group is isolated from each other and can only access the devices in their 

own group. Using zoning, you can arrange fabric-connected devices into logical 

groups, or zones, over the physical configuration of the fabric. Zones can be 

configured dynamically. They can vary in size depending on the number of fabric 

connected devices, and devices can belong to more than one zone. Because 

zone members can access only other members of the same zone, a device not 

included in a zone is not available to members of that zone. Therefore, you can 

use zones to: 

► 

► 

► 

Create closed user groups, for example between test and development 

environments 

Create a barrier between different environments such as Windows and UNIX 

operating systems 

Secure fabric areas 

Figure 3-6 below illustrates two slightly overlapping zones. 

Unix 

Windows NT/2000 

Zone green 

Zone red 

Figure 3-6 Zoning 

In addition to the general reasons for zoning, as we have seen, with tape drives, 

there is one more valid argument for using them. Unlike with disk devices such as 

the IBM Enterprise Storage Server, most tape device drivers, (including the one 

for the LTO Ultrium drives), do not support dual paths. This means that, although 


each server can use two separate HBAs and two connections to the SAN fabric 

for redundancy, there is no automatic fail-over. This is because the device driver 

has no intelligence that the dual paths are in fact for the same physical devices 

and simply configures multiple device names instead. To solve this, we must 

“hide” the alternate path to the library and tapes using zoning, LUN masking, 

persistent binding or software. Figure 3-7 shows how we could setup a zone to 

eliminate dual pathing so that there is only one connection from the host to the 

tape devices. 

Host 

FC 

ISL 

FC 

FC 

Host 

FC 

2109 FC Switch 

FC 

2108 SDG 

SCSI 

Figure 3-7 Zone to restrict the tapes to one HBA only 

Note: Currently only the AIX device driver for the IBM 3590 drive supports 

alternate paths. 

Some common information about zoning: 

► 

► 

► 

► 

► 

► 

Zones can be configured dynamically. 

Configuring new zones does not interrupt traffic on unaffected ports or 

devices. Also new zones do not affect data traffic across interswitch links 

(ISLs) in cascaded switch configurations. 

Zoning can be configured and administered by telneting to the switch or using 

the optional GUI IBM StorWatch Specialist. 

Devices can belong to more than one zone. 

Zoning can be administered from any switch in the fabric. 

Changes that are configured to one switch automatically replicate to all 

switches in the fabric; if a new switch is added to an existing fabric, all zone 


characteristics are automatically applied to the new switch. Because each 

switch stores zoning information, zoning ensures a high level of reliability and 

redundancy. 

3.2.1 Types of zoning 

There are two types of zoning used: hardware zoning and software zoning. 

Hardware zoning 

Hardware zoning is based on the physical fabric port number. It can be 

implemented in the following configurations: 

► 

► 

► 

One to one 

One to many 

Many to many 

Disadvantage: 

A zoning configuration can become unusable if the device is connected to a 

different port, since hardware zoning relies on each device being connected 

to a specific port. 

Advantage: 

Hardware zoning works independently of influence from other sources, such 

as HBA firmware. The switch hardware ensures that there is no data transfer 

between unauthorized zone members. However, devices can transfer data 

between ports within the same zone. Consequently, hard zoning provides the 

greatest security possible. Use it where security must be rigidly enforced. 

Hardware zoning also has a very low impact on the routing performance. 

If you use hardware zoning you must carefully document and update your SAN 

Fabric - but then, documentation is a necessary task anyway for managing a 

SAN. 

Software Zoning 

Software zoning is implemented within the Simple Name Server (SNS) running 

inside the fabric switch. In a soft zone, at least one zone member is specified by 

WWN. When a device logs in, it queries the name server for devices within the 

fabric. If zoning is in effect, only the devices in the same zone(s) are returned. 

Other devices are hidden from the name server query reply. 

The members of a zone can be defined with: 

► Node WWN (WWNN) 

► Port WWN (WWPN) 


Advantage: 

Software zoning is not affected by moving devices to different physical switch 

ports. If you use WWNs for the zone members, even if a device is connected 

to another physical ports, it will still remain in the same zoning definition 

because the device’s WWN remains the same. 

Disadvantage: 

Each component of the SAN fabric must support soft zoning or be specially 

configured. Here is an example for the Emulex HBA device driver. You need to 

make sure the zone-rscn option is set to 1 in the lpfc.conf file. More 

information on configuring this device driver is at 2.8.6, “Emulex LP8000 HBA 

driver configuration” on page 73: 

Example 3-2 Enabling software zoning with Emulex HBA 

# Setting zone-rscn to 1 causes the driver to check with the 

# NameServer to see if an N_Port ID received from an RSCN applies. 

# If "Soft Zoning" is used, with Brocade Fabrics, this 

# should be set to 1. 

zone-rscn=1; 

The switch does not control data transfer so there is no guarantee against 

data transfer from unauthorized zone members. 

You can intermix hardware and software zoning. 

3.2.2 Zoning setup with IBM 2109 Switch StorWatch Specialist 

Now, we want to create a zone so that the dual pathing issue is eliminated. We 

are using the lab system configuration shown in Figure 3-5 on page 99 and want 

to configure a zone as in Figure 3-7 on page 101. Remember, currently we are 

seeing duplicate device names for our Ultrium drives and medium changer 

(shown in Example 3-1 on page 99), because we have two HBAs on the same 

host accessing the SDG. 

The StorWatch Specialist in Figure 3-8 shows we are using hardware zoning 

since the members of the zone LTOUnix are the switch port numbers. Currently, 

both HBAs are in the same zone as the SDG which causes the problem of dual 

pathing. The SDG is connected to port 0 of switch 3. One HBA is connected to 

port 8 of switch 3 and the other one is connected to port 5 of switch 2. All ports 

are in the same zone. 


Figure 3-8 First zone with both HBA in the zone 

Now we want to make sure that only one HBA will see the SDG. We want to do 

this with software zoning so the first step is to create alias names. For this we 

need to know the WWN of our HBAs and SDG, since these will be the members 

of the zone. 


Tip: About aliases 

► An alias is a name assigned to a device or group of devices. 

► 

► 

► 

An alias can be assigned either to a switch port or to a WWN. 

An alias must be a unique alphanumeric string beginning with an alpha 

character. The underscore character (_) is allowed and alias names are 

case sensitive. 

Aliases are optional, but are recommended as they greatly simplify zone 

administration. 

Get WWN of HBA on AIX 

In order to get the WWN of the HBA, you need to login as root. Type lscfg -vl 

fcsx, where x is the adapter number (Example 3-3). 

The network address value is the adapter WWN. 

Example 3-3 WWN of HBA on AIX 

# lscfg -vl fcs1 


fcs1 20-58 FC Adapter 

Part Number.................03N4167 

EC Level....................D 

Serial Number...............KT90601339 

Manufacturer................0010 

FRU Number..................03N4167 

Network Address.............10000000C920D21F 

ROS Level and ID............02903219 





Device Specific.(Z4)........FF101410 




Device Specific.(Z8)........10000000C920D21F 

Device Specific.(Z9)........SS3.20A9 

Device Specific.(ZA)........S1F3.20A9 

Device Specific.(ZB)........S2F3.20A9 

Device Specific.(YL)........P2-I1 


Get WWN of HBA on Solaris 

Our Solaris system here had both QLogic 2200 and Emulex LP8000 HBAs 

installed. Login as root. Type dmesg and scan down till you find the entries related 

to the HBA. They will look similar to this: 

Example 3-4 WWN of HBA on Solaris 

Aug 10 16:10:28 sol-e unix: QLogic qla2200 Fibre Channel Driver v3.07 Instance: 1. Firmware 

v2.1.31. 

Aug 10 16:10:36 sol-e unix: qla2200(1): Verification complete. Hardware initialized. 

Aug 10 16:10:37 sol-e unix: qla2200-hba1-adapter-node-name="200000e08b0190ef"; 

Aug 10 16:10:37 sol-e unix: qla2200-hba1-adapter-port-name="210000e08b0190ef"; 

Aug 10 16:10:37 sol-e unix: PCI-device: scsi@2, qla22001 

Aug 10 16:10:37 sol-e unix: qla22001 is /pci@1f,4000/scsi@2 

Aug 10 16:10:37 sol-e unix: NOTICE: 

Aug 10 16:10:37 sol-e Emulex LightPulse FC SCSI/IP 4.20k 

Aug 10 16:10:37 sol-e unix: NOTICE: lpfc0:045:Vital Product Data Data: 82 23 0 36 

Aug 10 16:10:37 sol-e unix: NOTICE: lpfc0:031:Link Up Event received Data: 1 1 0 0 

Aug 10 16:10:39 sol-e unix: NOTICE: lpfc0: Firmware Rev 3.81A1 (D2D3.81A1) 

Aug 10 16:10:39 sol-e unix: NOTICE: lpfc0: WWPN:10:00:00:00:c9:23:c3:05 

WWNN:20:00:00:00:c9:23:c3:05 DID 0x31700 

The entries beginning with qla2200-hba1-adapter-node-name and 

qla2200-hba1-adapter-port-name give the WWN and WWPN of the QLogic 220 

HBA. The entry beginning with lpfc0: WWPH gives the same information for the 

Emulex LP8000 HBA. 

Get WWN of the SDG 

Telnet to the SDG and use the command fcShowNames (Example 3-5). 

Example 3-5 WWN of SDG 

Router > fcShowNames 

----------------------------------------------------------------- 

Ctlr : PCI Addr : ISP : Node : Port 

Id : Bs Dv Fn : Type : Name : Name 

----------------------------------------------------------------- 

1 : 00 06 00 : 2200 : 10000060.45161ff5 : 20010060.45161ff5 

----------------------------------------------------------------- 

You can also use the SAN Data Gateway StorWatch Specialist (Figure 3-9) to 

display this information. 


Figure 3-9 WWN of the SDG 

We describe these interfaces to the SDG in detail in 3.4, “SAN Data Gateway 

setup” on page 125. 

Create alias using the IBM 2109 StorWatch 

From the initial Fabric View window, select the Zone Admin panel. See 

Figure 3-10. 

Figure 3-10 Start Zone Admin panel 


We will create an alias for the WWNN of our RS/6000, brazil. We retrieved the 

WWNN in Example 3-3 on page 105. Enter an alias name (in our case, BRAZIL) 

and press the button Create Alias. Select either the WWN or port number and 

press the button Add Member. See Figure 3-11. 

Figure 3-11 Create alias 

Using the same process, we then created another alias, SDG_LTO for the SAN 

Data Gateway, using the WWNN we discovered in Figure 3-5. 


Define Zone 

From the Zone Administration panel, select the Zone Tag. Click the Create 

Zone button and enter the name of the zone (LTO_Risc_only in our example). 

Select now from the left side your pre-defined alias, switch port or WWN and 

press the Add Member button. After you added all members to your zone click 

on Apply(Figure 3-12). Our zone contains the SDG and one of the RS/6000 

HBAs. 

Figure 3-12 Define zone 

Now your zone is configured but not yet active. We have to make the zone 

available through the zone configuration. 

Define and apply zone configuration 

A zone configuration is a group of zones that are enforced whenever that zone 

configuration is enabled. A zone can be included in more than one zone 

configuration. To define a zone configuration select the Config tag on the Zone 

Admin window. 


If needed, create a new zone configuration — ours is called NEW. 

Select the Config Name where you want to put your new zone. Then select the 

zone to be added and press Add Member. If you want to enable this 

configuration now, then select Enable Config. Click Apply to activate your 

changes. See Figure 3-13. 

Figure 3-13 Define zone configuration 

For more information about zoning and usage of the IBM 2109 Switch we refer 

you to: 

► 

► 

► 

IBM SAN Survival Guide Featuring the IBM 2109, SG24-6127 

2109 Model S08 User’s Guide, SC26-7349 

2109 Model S16 User’s Guide, SC26-7351 

Now we have created a zone and applied it to the SAN fabric. Only one HBA 

should now see the tape drives. 


If we delete the configuration from RS/6000 (remove the tape and medium 

changers using SMIT or rmdev as described in 2.2.6, “Deleting LTO devices” on 

page 38) and run cfgmgr again, then we will see only the devices shown in 

Example 3-6 available. We now see the correct number of physical devices, 

which has been achieved with switch zoning. 

Example 3-6 Tape drives with dual pathing eliminated 





3.2.3 Another zoning example 

In this example, shown in Figure 3-14, there are multiple tape drives and HBAs. 

You might use zones as shown so that the drives are split between the HBAs for 

performance or security reasons. Without zoning or other configurations, you 

would see the same dual pathing issue as shown in the previous section. 

Unix 

Server 

Fibre 

Channel 

Aapter 

Zone B 

3584 with FC 

LTO Drives 

2109 FC Switch 

Zone A 

Figure 3-14 3584 with 8 FC LTO drives, one server with 2 HBA and 2 zones 

But you may also manage this scenario by using persistent binding as the next 

section will show. 


Note: For performance reasons, we strongly recommend that you do not mix 

tape and disk devices on the same HBA. 

3.3 Persistent binding 

This function allows a subset of discovered targets to be bound between a server 

and device. Some operating systems do not guarantee that devices will always 

be allocated the same SCSI target ID after rebooting. This can cause a problem 

as some software expects that tape devices will always have the same SCSI 

target ID. Persistent binding allows a tape device’s WWN to be bound to a 

specific SCSI target ID. Another reason to use persistent binding, is to ensure 

that the drive names are configured in the operating system in their physically 

installed order. We will discuss this in more detail in Section 3.5, “Operating 

system device names” on page 142. Once a configuration has been set it will 

survive reboots and any hardware configuration changes as the information will 

be held in the registry of the server. Binding also solves the dual pathing 

problem. Binding can be implemented by WWNN or WWPN. 

Persistent binding is mainly useful in a Solaris environment and we highly 

recommend that you use it. It is not required for AIX since it is not so dependent 

on the LUN and target ID for accessing SAN devices. This is explained in 3.3.4, 

“Persistent binding not applicable to AIX” on page 123. 

3.3.1 The need for persistent binding 

This example from our lab shows what happens without any binding. First, we 

connected the 3583 to a SDG 2108. This SDG is connected over SAN fabric to a 

Sun server. The Sun server is using a QLogic HBA (Figure 3-15). 

3583 Lxx 

Sun 

Server 

Fibre 

Channel 

Aapter 

2109 FC 

Switches 


B 

A 

Tape Library 


SDG 

WWN: 

2001006045161ff5 

Figure 3-15 Lab configuration with Sun server and 3583 


With this configuration, the operating system configures two drives and one SCSI 

Medium Changer. The device names and target/LUN IDs are: 

► 

► 

► 

0smc with target 0 and LUN 1 - the medium changer 

1stbn with target 0 and LUN 2 - first tape drive 

2stbn with target 0 and LUN 4 - second tape drive 

We get this information by going to directory /dev/rmt and listing the device files 

with ls -la (see Example 3-7). The target/LUN IDs are listed at the end of each 

line (for example, IBMtape@0,1:smc for the first medium changer). 

Example 3-7 Output of ls -la showing tape device names 

ls -la 

lrwxrwxrwx 1 root root 48 Aug 8 19:53 0smc -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,1:smc 

lrwxrwxrwx 1 root root 47 Aug 8 19:53 1st -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:st 

lrwxrwxrwx 1 root root 48 Aug 8 19:53 1stb -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:stb 

lrwxrwxrwx 1 root root 49 Aug 8 19:53 1stbn -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:stbn 

lrwxrwxrwx 1 root root 48 Aug 8 19:53 1stc -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:stc 

lrwxrwxrwx 1 root root 49 Aug 8 19:53 1stcb -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:stcb 

lrwxrwxrwx 1 root root 50 Aug 8 19:53 1stcbn -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:stcbn 

lrwxrwxrwx 1 root root 49 Aug 8 19:53 1stcn -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:stcn 

lrwxrwxrwx 1 root root 48 Aug 8 19:53 1stn -> ../../devices/pci@1f,4000/scsi@2/IBMtape@0,2:stn 









Next we connected another tape library — an IBM Magstar MP 3570 with two 

drives. We used a separate model SDG to attach the 3570, because it requires a 

model with HVD ports. The SDG already being used for the 3583 had LVD ports 

to support the 3583’s LVD drives (Example 3-16). 

SUN 

Server 


Fibre 

Channel 

Aapter 

2109 FC 

Switches 

SDG 

WWN: 

2004006045160d2e 

SDG 

WWN: 

2001006045161ff5 

B 

A 

Tape Library 


B 

A 

3570 

Figure 3-16 Lab configuration with one 3583 and one 3570 


After executing either drvconfig or reboot -- -r we see the newly configured 

tape drives on the directory /dev/rmt. An Output of ls -la is shown in 

Example 3-8. 

Example 3-8 Output of ls -la 



































We can see the following configured devices with target/LUN IDs: 

► 

0smc with target 0 and LUN 1 - a medium changer 

► 1stbn with target 0 and LUN 2 - first tape drive for changer at target 0 

► 2stbn with target 0 and LUN 4 - second tape drive for changer at target 0 

► 3stbn with target 4 and LUN 2 - first tape drive for changer at target 4 

► 

4smc with target 4 and LUN 3 - a medium changer 

► 5stbn with target 4 and LUN 4 - second tape drive for changer at target 4 

The output of the ls command does not differentiate between the 3570 and the 

3583 drives. It also does not tell us which library and drives are associated with 

each SDG. 

To get this information, we need to look in the boot-log-file using the command 

dmesg. Example 3-9 shows us an output of this command after we connected and 

rebooted the Sun server. 


Example 3-9 Output of dmesg showing device associations 

Aug 9 17:05:02 sol-e unix: QLogic qla2200 Fibre Channel Driver v3.07 Instance: 1. Firmware v2.1.31. 




Aug 9 17:05:11 sol-e unix: qla2200-hba1-SCSI-target-id-0-fibre-channel-name="2001006045161ff5"; 

Aug 9 17:05:11 sol-e unix: qla2200-hba1-SCSI-target-id-1-fibre-channel-name="210000e08b023629"; 

Aug 9 17:05:11 sol-e unix: qla2200-hba1-SCSI-target-id-2-fibre-channel-name="10000000c9258e54"; 

Aug 9 17:05:11 sol-e unix: qla2200-hba1-SCSI-target-id-3-fibre-channel-name="50001fe1000ae401"; 

Aug 9 17:05:11 sol-e unix: qla2200-hba1-SCSI-target-id-4-fibre-channel-name="2004006045160d2e"; 

Aug 9 17:05:11 sol-e unix: qla2200-hba1-SCSI-target-id-5-fibre-channel-name="10000000c920d21f"; 

Aug 9 17:06:22 sol-e unix: NOTICE: IBMtape _probe: Found Medium Changer ULT3583-TL , Inst 209, Tgt 0, Lun 1 

Aug 9 17:06:23 sol-e unix: IBMtape209 at qla22001: 

Aug 9 17:06:23 sol-e unix: target 0 lun 1 

Aug 9 17:06:23 sol-e unix: IBMtape209 is /pci@1f,4000/scsi@2/IBMtape@0,1 

Aug 9 17:06:23 sol-e unix: NOTICE: IBMtape _probe: Found Tape Drive ULT3580-TD1 , Inst 210, Tgt 0, Lun 2 








Aug 9 17:06:23 sol-e unix: NOTICE: IBMtape _probe: Found Tape Drive 03570C12 , Inst 254, Tgt 4, Lun 2 




Aug 9 17:06:23 sol-e unix: NOTICE: IBMtape _probe: Found Medium Changer 03570C12 , Inst 255, Tgt 4, Lun 3 








In the dmesg output we can see more descriptive names for the discovered 

devices. It also shows how the FC adapter assigns SCSI IDs to the device 

WWPNs. The first part of the listing shows that the SDG for the LTO Drives 

(WWPN 2001006045161ff5) is assigned SCSI ID 0 and the SDG for the 3570 

(WWPN 2004006045160d2e) is assigned SCSI ID 4. Now we can figure out 

which stbn device is in which library. 

► 0smc with target 0 and LUN 1 is the medium changer for the 3583 

► 1stbn with target 0 and LUN 2 is the first drive in the 3583 

► 2stbn with target 0 and LUN 4 is the second drive in the 3583 

► 3stbn with target 4 and LUN 2 is the first drive in the 3570 

► 4smc with target 4 and LUN 3 is the medium changer for the 3570 

► 5stbn with target 4 and LUN 4 is the second drive in the 3570 

This is all logical up to now, however the problem is that, by default, the 

target/LUN assignment is not deterministic, and can change. Let’s simulate a 

reconfiguration by turning off the 3583’s SDG — causing its devices to be no 

longer visible when we reboot or run drvconfig. Figure 3-17 shows how we take 

out the SDG. 


SUN 

Server 


3583 Lxx 

Fibre 

Channel 

Aapter 

2109 FC 

Switches 

SDG 

WWN: 

2004006045160d2e 

SDG 

WWN: 

2001006045161ff5 

B 

A 

Tape Library 


B 

A 

3570 

Figure 3-17 One SDG is missing during boot 

After rebooting we check the dmesg output again (Example 3-10). The SDG of the 

3570 (which is still visible) has been assigned a different SCSI target ID - three 

this time. The drives will also be assigned to LUNs on SCSI ID 3. This will create 

a problem with backup or other software which expects to access the drives or 

medium changer at their old addresses on target ID 4 - they will fail with I/O 

errors. 

Example 3-10 Output of dmesg with changed SCSI ID 








Now let’s see how we can fix this problem of non-definitive target/LUN IDs using 

persistent binding. We will bind the SDG for the 3570 to target ID 0 and the SDG 

for the 3583 to target ID 1. Detailed instructions for doing this for different HBAs 

are in section 3.3.2, “Persistent binding with QLogic HBA on Solaris” on 

page 118 and on the section 3.3.3, “Persisting binding with Emulex HBA on 

Solaris” on page 121: 

hba1-SCSI-target-id-0-fibre-channel-name="2004006045160d2e"; 

hba1-SCSI-target-id-1-fibre-channel-name="2001006045161ff5"; 

We force a new configuration by deleting all files in the directory /dev/rmt and 

reboot with device configuration (reboot -- -r). By looking at the new dmesg and 

ls -la output (Example 3-11 and Example 3-12) we can see that our devices 

will be permanently set at: 


► 0stbn with target 0 and LUN 2 - first drive in the 3570 

► 

1smc with target 0 and LUN 3 - 3570 medium changer 

► 2stbn with target 0 and LUN 4 - second drive in the 3570 

► 

3smc with target 1 and LUN 1 - 3583 medium changer 

► 4stbn with target 1 and LUN 2 - first drive in the 3583 

► 5stbn with target 1 and LUN 4 - second drive in the 3583 

Example 3-11 Output of dmesg after binding 

Aug 9 18:07:04 sol-e unix: QLogic qla2200 Fibre Channel Driver v3.07 Instance: 1. Firmware v2.1.31. 





Aug 9 18:07:13 sol-e unix: qla2200-hba1-SCSI-target-id-1-fibre-channel-name="2001006045161ff5"; 





. 


Aug 9 18:08:20 sol-e unix: NOTICE: IBMtape _probe: Found Medium Changer 03570C12 , Inst 211, Tgt 0, Lun 3 


Aug 9 18:08:21 sol-e unix: NOTICE: IBMtape _probe: Found Medium Changer ULT3583-TL , Inst 234, Tgt 1, Lun 1 




Example 3-12 Output of ls -la after binding 



































By using persistent binding the 3570 drives will always use target ID 0 and the 

3583 drives will always use target ID 1 even if we power off one of the SDG’s or if 

we add additional devices to the SAN. 

Tip: This example is also valid for 3584 with native FC drives. Every drive has 

a unique WWN. Therefore every drive gets a separate SCSI ID. Without 

persistent binding, if drives are missing at server boot time or new drives are 

added, the SCSI IDs will probably change. Therefore, you should always 

implement persisting binding for a 3584 with native FC drives! 

Tip: If you install a 3584 with native FC Drives then you could not expect that 

the drives defined on the server will show up in the same order as they are 

installed in the library. Persistent binding will solve this problem too. 

3.3.2 Persistent binding with QLogic HBA on Solaris 

Persistent binding (binding a WWN to a specific SCSI target ID) is implemented 

by editing the QLogic adapter’s configuration file — qla2200.conf in the directory 

/kernel/drv/. This file contains documentation and examples for binding. The 

entries use the required SCSI target ID and WWNN (World Wide Node Name) 


following the format as shown in Figure 3-18. You should add your own required 

definitions to the file, using the SCSI target ID and the WWN using Example 3-13 

as a guide. You can use either the World Wide Node Name (WWNN) or the 

World Wide Port Name (WWN). You also need to choose the correct adapter 

number — hba1 is the first QLogic adapter in your system. Our example uses the 

World Wide Port Name (WWN). 



SCSI ID 

WWN 

Figure 3-18 Binding with QLogic 

The WWN 2001006045161ff5 was bound to SCSI target ID 1. This is the SDG for 

the 3583 in our lab configuration (see Figure 3-16 on page 113). The WWN 

2004006045160d2e, the SDG for the 3570 was bound to SCSI target ID 0. 

Example 3-13 Binding with QLogic adapter 

# ======================= 

# Persistent Name Binding 

# ======================= 

# 

# The Persistent Name Binding support for adapter devices associates a 

# user specified World Wide Node Name to a specified adapter instance. 

# In other words, this entry changes the node name of the specified 

# adapter instance to the node name specified in the entry. 

# To enable it, entries of the following format should be added in the 

# qla2200.conf file: 

# 

# hba-adapter-node-name=""; 

# 

# For example, to associate the following node name to adapter 

# instance 2, the entry will be: 

# 

# hba2-adapter-node-name="200000e08b023227"; 

# 

# The Persistent Name Binding support for target devices associates a 

# SCSI target ID to a specified device World Wide Name. In other 

# words, this entry changes the target ID of the specified device 

# to the number specified in the entry. To enable it, lines of 

# the following format should be added in the qla2200.conf file: 

# 

# hba-SCSI-target-id--fibre-channel-name=""; 

# 

# For example, to associate the following WWN with target ID 2 on 


# adapter instance 1, the entry will be: 

# 

# hba1-SCSI-target-id-2-fibre-channel-name="2000002037004ac9"; 

# 

# ********* 

# IMPORTANT 

# ********* 

# 

# - Unlike other parameter entries in the qla2200.conf file, values 

# indicated in persistent name binding entries for instance 0 does 

# not apply to other instances when there are no entries for the 

# other instances. 

# 

# - Persistent Name Binding entries are optional entries in the 

# qla2200.conf file. These entries do not dictate which devices 

# are scanned by the driver, and do not perform any device masking. 

# Absence of this entry for any given device name or SCSI ID does 

# not direct the driver to perform any type of device masking. 

# 

# - For adapter devices, if no Persistent Name Binding entries are 

# specified, the default WWNN used by the adapters will be the ones 

# stored in the NVRAM. 

# 

# - For target devices, if no Persistent Name Binding entries are 

# specified, the default SCSI target IDs assigned will be the loop 

# ID of the devices. Please refer to the FAQ section for which 

# default SCSI target IDs are assigned to fabric devices. 

# 

# ******* 

# CAUTION 

# ******* 

# 

# When the Persistent Name Binding entries are used in combination with 

# the device entries in the /kernel/drv/sd.conf or /kernel/drv/st.conf 

# file, it is possible to purposely mask off certain devices of certain 

# adapter instances from the operating system, by purposely commenting 

# out/deleting entries in the sd.conf/st.conf file for those SCSI target 

# IDs/LUNs bound to those certain adapter instances. However, when the 

# devices are shared among multiple adapters, doing so requires extreme 

# caution, and it is strongly recommeded that all target devices be bound 

# with all adapter instances so that the devices will show up in a pre- 

# determinable way. A shared device which is bound on one adapter 

# instance but unbound on another adapter instance may show up at a SCSI 

# target ID which is unexpected by the user for the unbound adapter 

# instance, which may lead to possible data corruption on the device 

# if the user unknowingly writes to the unexpected device. 

# 

# The QLogic Solaris SCSI HBA driver does not filter out the OS scanning 

# of SCSI devices. The target IDs/LUNs scanned by the OS is based on the 


# entries specified in the sd.conf and st.conf files. When the HBA 

# driver first initializes during the system boot process, it assigns a 

# target ID internally to all of the devices it detected. During the OS 

# device scan, the driver ignores the 'class' value indicated in the 

# sd.conf/st.conf file and processes the IOs addressed to a device whose 

# target ID/LUN matches that of a device previously detected by the 

# driver. 

#Binding 



After you edit the configuration file, reboot the system to apply the changes. 

3.3.3 Persisting binding with Emulex HBA on Solaris 

Persistent binding (binding a WWN to a specific SCSI target ID) is implemented 

by editing the Emulex adapter’s configuration file. The configuration files are 

located in the directory /kernel/drv/ and are called lpfc.conf (for LP8000) or 

lpfs.conf (for LP8000S). This file contains documentation and examples for 

binding. The entries use the required SCSI target ID and World Wide names. 

Follow the format as shown in Figure 3-19. You should add your own required 

definitions to the file, using the SCSI target ID and either the WWNN or WWPN, 

using Example 3-14 as a guide (in our example we used WWNN). You also need 

to choose the correct adapter number — lpfc0 is the first adapter in your system 

and lpfc1 would be the second one. For LP8000 you have to use device name 

lpfc and for LP8000S you have to use lpfs. 

fcp-bind-WWNN="1000006045161ff5:lpfc0t4", 

"1004006045160d2e:lpfc0t3"; 

WWNN 

SCSI ID 

fcp-bind-WWPN="2100123456789abc:lpfc0t0", 

"21000020370c2855:lpfc0t1"; 

WWPN 

SCSI ID 

Figure 3-19 Binding with Emulex 


Example 3-14 Configuration file for the Emulex Adapter - example of binding 

# Setup FCP persistent bindings, 

# fcp-bind-WWPN binds a specific WorldWide PortName to a target id, 

# fcp-bind-WWNN binds a specific WorldWide NodeName to a target id, 

# fcp-bind-DID binds a specific DID to a target id. 

# Only one binding method can be used. 

# WWNN, WWPN and DID are hexadecimal values. 

# WWNN must be 16 digits with leading 0s. 

# WWPN must be 16 digits with leading 0s. 

# DID must be 6 digits with leading 0s. 

# The SCSI ID to bind to consists of two parts, the lpfc interface 

# to bind to, and the target number for that interface. 

# Thus lpfc0t2 specifies target 2 on interface lpfc0. 

# NOTE: Target ids, with all luns supported, must also be in sd.conf. 

# scan-down must be set to 0 or 1, not 2 which is the default!! 

# 

# Here are some examples: 

# WWNN SCSI ID 

# fcp-bind-WWNN="2000123456789abc:lpfc1t0", 

# "20000020370c27f7:lpfc0t2"; 

# 

# WWPN SCSI ID 

# fcp-bind-WWPN="2100123456789abc:lpfc0t0", 

# "21000020370c2855:lpfc0t1", 

# "2100122222222222:lpfc2t2"; 

# 

# DID SCSI ID 

# fcp-bind-DID="0000ef:lpfc0t3"; 

# BEGIN: LPUTIL-managed Persistent Bindings 

fcp-bind-WWNN="1000006045161ff5:lpfc0t4", 

"1004006045160d2e:lpfc0t3"; 

# If automap is set, SCSI IDs for all FCP nodes without 

# persistent bindings will be automatically generated. 

# If new FCP devices are added to the network when the system is down, 

# there is no guarantee that these SCSI IDs will remain the same 

# when the system is booted again. 

# If one of the above fcp binding methods is specified, then automap 

# devices will use the same mapping method to preserve 

# SCSI IDs between link down and link up. 

# If no bindings are specified above, a value of 1 will force WWNN 

# binding, 2 for WWPN binding, and 3 for DID binding. 

# If automap is 0, only devices with persistent bindings will be 

# recognized by the system. 

automap=1; 

We bind the SDG for the 3583 (WWN 2001006045161ff5) to target ID 4 and the 

SDG for the 3570 (WWN 2004006045160d2e) to target ID 3. 

Example 3-15 shows you an output of the dmesg after the binding. During the 

boot only the 3583 was connected to the Sun server. You see that WWN 

2001006045161ff5 was bound to target ID 4. 


Example 3-15 dmesg after binding with Emulex 

Aug 13 14:04:34 sol-e Emulex LightPulse FC SCSI/IP 4.20k 

Aug 13 14:04:34 sol-e unix: /pci@1f,2000/fibre-channel@1 (lpfc0): 

Aug 13 14:04:34 sol-e Binding WWNN 10:00:00:60:45:16:1f:f5 to FCP/SCSI Target 4 

Aug 13 14:04:34 sol-e unix: NOTICE: lpfc0:045:Vital Product Data Data: 82 23 0 36 

Aug 13 14:04:34 sol-e unix: NOTICE: lpfc0:031:Link Up Event received Data: 1 1 0 0 

Aug 13 14:04:36 sol-e unix: NOTICE: lpfc0: Firmware Rev 3.81A1 (D2D3.81A1) 

Aug 13 14:04:36 sol-e unix: NOTICE: lpfc0: WWPN:10:00:00:00:c9:23:c3:05 WWNN:20:00:00:00:c9:23:c3:05 DID 0x31700 

Aug 13 14:04:36 sol-e unix: NOTICE: Device Path for interface lpfc0: 

Aug 13 14:04:36 sol-e unix: PCI-device: fibre-channel@1, lpfc0 

Aug 13 14:04:36 sol-e unix: lpfc0 is /pci@1f,2000/fibre-channel@1 

Aug 13 14:04:36 sol-e unix: NOTICE: lpfc0: Acquired FCP/SCSI Target 4 LUN 0 

Aug 13 14:04:36 sol-e D_ID 0x31000 WWPN:20:01:00:60:45:16:1f:f5 WWNN:10:00:00:60:45:16:1f:f5 

3.3.4 Persistent binding not applicable to AIX 

The supported FC adapters for AIX do not have persisting binding capability. This 

is because it is not needed for AIX. 

Unlike Solaris, AIX does not communicate with the devices over the SCSI ID and 

LUN. AIX talks to the devices over the WWN. See Example 3-16 which shows 

you the result of lsattr -El with a FC connected tape. Notice the fields scsi_id 

and lun_id. These are not a real SCSI ID or LUN ID. Also you recognize in the 

ww_name field, the value 2001006045161ff5 is the WWN of the SDG for the 

3583. But the location field looks like a typical SCSI ID. This is done by the Atape 

device driver which gets this information from the SDG and puts it in the location 

field used also by lsdev (Example 3-17) as a convenience. 

Example 3-16 AIX tape drive attributes 

# lsattr -El rmt0 

block_size 0 Block Size (0=Variable Length) True 

compress yes Use Hardware Compression on Tape True 

logging no Activate volume information logging True 

max_log_size 500 Maximum size of log file (in # of entries) True 

space_mode SCSI Backward Space/Forward Space Record Mode True 

rew_immediate no Use Immediate Bit in Rewind Commands True 

trailer_labels no Trailer Label Processing True 

retain_reserve no Retain Reservation False 

devtype ULT3580- Device Type False 

scsi_id 0x31000 SCSI Target ID True 

lun_id 0x2000000000000 Logical Unit Number True 

location 20-58-01-0,0 Location True 

ww_name 0x2001006045161ff5 World Wide Name False 

new_name New Logical Name True 

Example 3-17 Output of lsdev for FC attached drives 






You see the indication (FCP) at the end of the lines of lsdev output, indicating 

these devices are FC connected. This designation is also added by the AIX 

Atape device Driver. 

If the SAN configuration was changed, after rebooting the AIX server or running 

cfgmgr, the device to target/LUN mapping will be preserved because of this 

connection to the WWN of the SDG. New devices would get any free device 

names. Therefore, it is not relevant or necessary to use persistent binding with 

AIX. 

If you want to rename the device to reflect a different logical order (give tape 

drives a new name), you can do this using smit devices. Go to Tape drives -> 

Change / Show Characteristics of a Tape Drive and select the tape drives 

which you want change. Select the name which you want to use in the line New 

Logical Name. In Example 3-18 we change the name from rmt1 to rmt4. 

Example 3-18 SMIT Change / Show Characteristics of a Tape Drive 





Tape Drive 

rmt1 



fcp 

Description 

IBM 3580 Ultrium Tape 

Drive (FCP) 

Status 

Available 

Location 20-58-01-1,0 


fscsi1 

Connection address 3 

SCSI ID 

0x31000 

Logical Unit ID 

0x4000000000000 

World Wide Name 

0x2001006045161ff5 

New Logical Name 

[rmt4] 

Block Size (0=Variable Length) [0] +# 

Use Hardware Compression on Tape yes + 

Activate volume information logging no + 

Maximum size of log file (in # of entries) [500] +# 

Backward Space/Forward Space Record Mode SCSI + 

Use Immediate Bit in Rewind Commands no + 

Trailer Label Processing 

no 

F1=Help F2=Refresh F3=Cancel 

F4=List 


F5=Reset F6=Command F7=Edit 

F8=Image 

F9=Shell F10=Exit Enter=Do 

If you have natively attached FC Drives or if you have more than one SDG 

attached to one library then you will find that normally the drives on AIX are not 

configured in the same order as they are physically in the library. This smit option 

allows you to easily change the order to reflect your configuration. 

3.4 SAN Data Gateway setup 

The IBM SAN Data Gateway (SDG) allows you to connect supported SCSI 

devices to a SAN by providing the interface function. 

The SDG maps addresses across and among the different interfaces and 

preserves the persistency of the address maps across startups of systems, 

devices, and the SDG. 

Attention: Sharing the gateway between disk and tape products is not 

supported! 

The SDG Router allows you to connect more than one host to it. In its default 

configuration, all hosts can access all target devices. In practice, you want to 

restrict this access, unless you are sure that the different host operating systems 

and applications will correctly handle simultaneous access to the same device by 

multiple servers. Otherwise, device operation failures and potential data integrity 

problems can occur. To avoid these, you can use zoning (as discussed in 3.2, 

“Zoning” on page 100), use intelligent applications which can handle the device 

sharing (such as tape library sharing on Tivoli Storage Manager), or use these 

options available with the SDG: 

► 

► 

Channel zoning option. You can set the desired access between SAN 

connections and individual SCSI channels. 

Virtual Private SAN (VPS) or LUN masking. You can set the desired access 

permissions between hosts on the SAN connections and individual LUNs. 

Note: VPS is only available as an option for the IBM 2108 G07 and not for the 

IBM 2108 R03 or the Integrated SAN Data Gateway Module for the 3583. 

Special software is also require. 

In the following section we give you an overview of: 

► 

Basic setup of the SDG 


► 

► 

► 

► 

Installation and usage of the StorWatch SDG Specialist 

How LUN mapping works 

How to use channel zoning 

Virtual Private SAN 

3.4.1 Basic setup 

For more Information on the SAN Data Gateway, we refer you to: 

► 2108 Model G07 Installation and User's Guide, SC26-7304 

► 2108 Model R03 Installation and User's Guide, SC26-7355 

We will briefly summarize the main steps for setting up the SAN Data Gateway — 

detailed information is provided in the Installation and User’s Guides listed in the 

previous section. 

1. After installing the hardware set up a connection between the SDG’s serial 

port and a serial port on a PC, using the supplied null modem cable and a PC 

terminal program, such as Netterm or Hyperterm. Use the following 

parameters to configure the terminal session: 

– Bits per second: 19 200 

– Data bits: 8 

– Parity: None 

– Stop bits: 1 

– Flow Control: Xon/Xoff 

If you are connected, press Enter and you will get a response from the SDG as 

shown in Figure 3-20. 

Figure 3-20 Connected to the SDG 

2. We suggest following basic commands sequence to install a new or an 

already used SDG: 


a. initializeBox 

This restores the factory defaults by deleting all configuration files 

including persistent address map and VPS databases, and then a 

restart is needed. 

b. ethAddrSet 

Use the ethAddrSet command to set or change the IP address and 

network mask (if required) of the SAN Data Gateway. 

Router > ethAddrSet “192.168.1.54” “255.255.255.0” 

c. ethEnable 

The ethEnable command alters the start parameters of the SDG Router, 

enabling the Ethernet port. This command does not take effect until the 

SDG Router is restarted. 

d. userAdd "admin","password" 

This adds a user and password to the password file. The user name 

variable must be 3 to 80 characters. The password variable must be 8 to 

40 characters. Here you create your own userID. 

Router > userAdd "nancy","password" 

e. disableCC or enableCC 

The disableCC command disables the command and control interface 

(LUN 0). The SDG is addressable as a SCSI target device for command 

and control support. On a Fibre Channel interface, this device is seen as 

logical unit number 0, (LUN 0). If the command and control support 

enabled then you might see problems on AIX machine during a cfgmgr 

run: 

cfgmgr: 0514-621 WARNING: The following device packages are required for 

device support but are not currently installed. devices.fcp.array 

Therefore we suggest to disableCC if the SDG is connected to an AIX 

server. For other operating system you can enableCC. 

f. setHost X,YYY (where port X can be 1,4,2,5,3,6 and YYY can be "solaris", 

"nt", "hpux", "aix",”switch”, ...) 

The setHost command sets the operating system type for the specified 

SAN interface. This provides some customization in the way the SDG is 

presented to the particular operating system. If the port is 0, the change 

applies to all SAN connections; otherwise, the host type is applied only to 

the SAN connection on the specified interface. The default setting is NT. 

Currently, OS can be specified as NT, AIX, Netware, HP-UX, or Solaris. 

You have to put the hosttype in double quotes as shown in Example 3-19. 


Example 3-19 setHost for FC 4 to AIX 

Router > setHost 4,"aix" 

value = 0 = 0x0 

Note: Port 1 & 4 belongs to 1st FC Card, 2 & 5 to the 2nd FC Card,3 & 6 to the 

3rd FC Card inside the SGW. 

3. reboot 

A reboot makes the above changes effective. You can now connect to the 

SDG via a telnet session, using the Ethernet address you defined. 

4. fcShowDevs and fcShowNames 

The fcShowDevs command displays information about the devices that are 

accessible from each Fibre Channel interface. The display shows the LUN 

that the SDG has assigned to each device, the SCSI Channel that the device 

is attached to, the actual SCSI ID and LUN of the device, the vendor, product, 

revision and serial number of the device. The fcShowNames command displays 

the node and port names (addresses) of the Fibre Channels. If the output 

does not meet your physical installed devices then execute scsiRescan and 

reboot. Collect the output from fcShowDevs and fcShowNames for further 

use (fcShowNames shows you the WWN, the assigned LUN and the serial 

number of the tape drives; see Example 3-20). 

Example 3-20 Output of fcShowDevs and fcShowNames 

Router > fcShowDevs 

FC 1: 

LUN Chan Id Lun Vendor Product Rev SN 

----------------------------------------------------- 

0 0 0 0 PATHLGHT SAN Router 32aC 21081341573 

1 2 6 0 IBM ULT3583-TL 2.50 IBM7801954 

2 2 0 0 IBM ULT3580-TD1 16E0 6811020764 

4 2 1 0 IBM ULT3580-TD1 16E0 6811007030 

value = 6 = 0x6 

Router > fcShowNames 

----------------------------------------------------------------- 

Ctlr : PCI Addr : ISP : Node : Port 

Id : Bs Dv Fn : Type : Name : Name 

----------------------------------------------------------------- 

1 : 00 06 00 : 2200 : 10000060.45161ff5 : 20010060.45161ff5 


3.4.2 Install and use the StorWatch SDG Specialist 

The IBM StorWatch SAN Data Gateway Specialist software allows you to 

manage multiple gateways from any location on your network. The system 

consists of two Java applications but is a client-server application, not 

Web-based like the StorWatch for the IBM 2109 Switch. You can install the server 

on a single system on your network and manage connections between multiple 

clients and multiple SAN Data Gateways. Also you can install the server and 

client on one single system. The installation program allows you to install the 

appropriate components on your target host system. The Java runtime 

environment is also installed when you install the application components. 

The server does not need to be used as a client but it can. The sever needs 

network access to the SDGs to be managed. 

The client system can be located anywhere as long as it can connect to the 

server system using TCP/IP. 

The SDG StorWatch Specialist is available for AIX, Windows and Sun Solaris. 

There is only one install file available which contains the server and the client 

code. During installation you can choose which of them you want to install. 

Client and server do not need to be on the same operating system. This means 

you can install the server for instance on AIX or Sun Solaris and use the client on 

Windows 2000 or on Windows NT. 

You find the latest version of the SDG StorWatch Specialist (in the Downloads 

section): 

http://www.ibm.com/storage/sangateway/ 

Install on AIX 

Download the latest version and the readme file for AIX from the Web site above. 

If needed decompress and/or untar the files. The install images will be called 

something similar to storwsdg_bff. Install server and the client code with: 

installp -a -d storwsdg_bff all 

If you want to install only the server code type: 

installp -a -d storwsdg_bff storwsdg.server 

Or for install the client code only: 

installp -a -d storwsdg_bff storwsdg.client 

As always, check the readme file for the latest information on fixes and 



Install on Sun Solaris 

Download the latest version and the readme file for Sun Solaris from the Web 

site above. If needed, decompress and/or untar the files. For installation of the 

files enter: 

pkgadd -d . IBMswsdg 

During the installation you were asked if you want to install the client and if you 

want to install the server: 

Install StorWatch SAN Data Gateway Specialist Client [y/n] ? y 

Install StorWatch SAN Data Gateway Specialist Server [y/n] ? y 

As always, check the readme file for the latest information on fixes and 


Install on Windows 

Download the latest version and the readme file for Windows from the Web site 

above. Execute the downloaded file for installation. During installation you can 

choose to install the server and/or the client. See Figure 3-21. 

Figure 3-21 Install of the SDG StorWatch Specialist 

As always, check the readme file for the latest information. 


Start the server 

In order to connect to a SDG, you must first start the server application. The 

server application can run in the background. 

Start the server on AIX and on Sun Solaris in the background with: 

/usr/bin/ServerLaunch -background & 

To start the server on Windows, go to Start -> Programs -> IBM StorWatch SAN 

Data Gateway Specialist -> Server. This opens a DOS-window (Figure 3-22). 

Figure 3-22 SDG StorWatch Specialist - server application 

Next you can start the client. 

Start the client 

Start the client on AIX and on Sun Solaris with: 

/usr/bin/ClientLaunch 

To start the client on Windows, go to Start -> Programs -> IBM StorWatch SAN 

Datagateway Specialist -> Client. 

Enter the TCP/IP Address of the server where the SDG StorWatch Specialist 

Server application is running (Figure 3-25). 


Figure 3-23 SDG StorWatch connect to server 

Now you will be asked for User Name and Password (Figure 3-24). The default 

entries are: 

► 

► 

User Name: StorWatch 

Password: StorWatch 

Both values are key sensitive. 

Figure 3-24 SDG StorWatch logon 

Next, enter the TCP/IP address of the SDG which you want to manage 

(Figure 3-25). 

Figure 3-25 Connect to SDG 

Install firmware on the SDG 

After you are connected to the SDG you should check the firmware level which is 

printed on the right window. Check on the following Web site in the Downloads 

section to see if there is a newer version available.: 

http://www.ibm.com/storage/sangateway/ 


If so, download the file the StorWatch client system and update the firmware on 

the SDG. 

Attention: All I/Os must be halted while updating the firmware. 

Select first the SDG to be updated in the left window. Then select from the menu 

Controls -> Update Firmware. See Figure 3-26. You will be asked for the 

filename of the new firmware. After you have selected the downloaded file you 

are asked to confirm that you want to update now. Answer with OK if all I/O 

activity to the SDG is stopped. 

Firmware Level 

Figure 3-26 Install firmware on the SDG 

SCSI settings 

If you want to view or change the SCSI settings, first select the desired SCSI 

Channel on the left side of the client display. Then select Controls -> SCSI 

Channel from the menu. See Figure 3-27. 


Figure 3-27 Select the SCSI option 

If you update SCSI device firmware, the StorWatch application does not display 

the new firmware version until the SDG Router has issued a SCSI inquiry. The 

SCSI inquiry occurs when it rescans the SCSI buses. If you want to rescan the 

SCSI bus select the button Re-scan SCSI Bus (Figure 3-28). 

Figure 3-28 SCSI Channel parameters 


FC settings 

If you want to view or change the FC options, first select the desired FC Channel 

on the left side of the client display. Then select Controls -> Fibre Channel. See 

Figure 3-29. 

Figure 3-29 Select the Fibre Channel options 

On the Fibre Channel Parameter window you can change settings valid for the 

selected FC. You can change the host type to match it for your configuration. The 

connection options should be point-to-point or point-to-point preferred if the 

SDG is connected to a switch. If the SDG is directly connected to a server, both 

server HBA and SDG need to have the same connection options. See 

Figure 3-30. 


Figure 3-30 Fibre Channel parameters 

3.4.3 LUN mapping 

In SCSI terminology, a tape drive or a disk drive is attached to a "bus" and has a 

unique address on that bus. There are three parts of the address in a 

conventional SCSI ID: 

► 

► 

► 

Bus (or channel) 

Target ID 

LUN 

A simple case of two tape drives attached to a single bus is shown in 

Figure 3-31. 


Server 

SCSI 

Aapter 

SCSI 

Drive 0 

SCSI ID 0 LUN 0 

Drive 1 


3584 Lxx 

Control path 

enabled smc 

A 

SCSI ID 0 

LUN 1 

B 


Figure 3-31 Basic SCSI connection to a system 

This configuration has a device map (from the host point of view) that looks as 

shown in Table 3-1. Some tape library devices, like the 3584, support LUNs for 

the devices; the tape drive is addressed as LUN 0, the SCSI Medium Changer 

(smc) is available as LUN 1. 

Table 3-1 Target ID and device mapping — native SCSI 

Target ID LUN Device 

0 0 Drive 0 

0 1 smc 

1 0 Drive 1 

If a product such as a SAN Data Gateway is placed between the system and 

target devices, the addressing has another layer. This is because the targets (the 

tape drives) are not directly attached to the host but are connected to a SCSI 

adapter installed in the gateway instead. Figure 3-32 shows the device mapping 

with the additional layer due to the SAN Data Gateway. 


Server 

FC 

Aapter 

FC 

SCSI 

Drive 0 


Drive 1 


SDG 

SCSI Id 0 LUN 0 

3584 Lxx 

Control path 

enabled 

B 

smc 

SCSI ID 0 

A 

LUN 3 


Figure 3-32 SAN Data Gateway attached through Fibre Channel — host view 

The system now only has a single target ID (target 0 in this case) directly 

attached to the SAN Data Gateway. The gateway forwards commands to and 

from its targets, the tape drives. However, there is a need to map the devices 

(tapes) so the host system can use them. This is achieved by using another layer 

of mapping, LUNs. The device map might now look like Table 3-2 from the host 

perspective. 

Table 3-2 Device map from host perspective — with SAN Data Gateway 

Target ID LUN Device 

0 0 SAN Data Gateway 

0 2 Drive 0 

0 3 smc 

0 4 Drive 1 

Note that LUN 0 points to the SAN Data Gateway. This allows you to send 

commands to control the gateway. This is referred to as the Command and 

Control Interface. 

Tape drives are always assigned an even LUN number. If a control path is 

enabled (3584) for this drive, its LUN is one higher than the drives. The 

odd-assigned LUN number that follows the tape-drive even number is reserved 

for the smc. This algorithm has been chosen for the best compatibility with 

existing applications and operating systems. The smc in a 3583 has its own SCSI 

ID, but will still be assigned an odd LUN number. 


If the resulting map is not suitable for your environment then you can edit the 

mapping done by the SDG with StorWatch Specialist. First select your desired 

SDG, then select Controls -> Device Mapping as shown in Figure 3-33. 

Figure 3-33 Select Device Mapping 

Figure 3-34 shows the current configuration if you use instead the command line 

interface (telnet or serial connection) with the command fcShowDevs. Devices that 

have already been entered into the persistent device map or devices which were 

automatically assigned by the SDG are shown in black type on the left. The 

assigned LUN for each device is shown in the left-most column. On the left side 

you see devices which are currently not assigned. You can delete a drive from 

the right side by dragging it down to the Recycle Bin icon. 


Figure 3-34 Device Mapping 

You can easily change this configuration by drag and drop. Select the device 

which you want to move to a different LUN. Hold your left mouse button and 

move the device to an empty LUN. After you have made any changes you need 

to reboot the SDG. See Figure 3-35. 

Figure 3-35 Device Mapping required a reboot 


3.4.4 Access control by channel zoning 

Click Controls -> Access Options-> Channel Zoning (see Figure 3-36) to 

configure zones to restrict access between SAN connections and SCSI 

channels. The default settings allow all SAN connections to access all SCSI 

channels. 

Figure 3-36 Select Channel Zoning 

When you select this menu option, a pop-up window displays the current channel 

zoning settings. Figure 3-37 shows the settings for a gateway that has two SAN 

connections and four SCSI channels. Currently only SCSI Channel 3 and SCSI 

Channel 4 are assigned to FC 1 and FC 2. Clear the check marks or put 

additional check marks by clicking in the boxes to create restricted access zones 

for the desired SAN connections and SCSI channels. All combinations are 

possible. 


Figure 3-37 Channel Zoning settings 

3.4.5 Access control by Virtual Private SAN (VPS) 

Channel zoning provides access control between ports. While channel zoning 

provides control of paths between host adapters and SCSI storage ports, it does 

not limit access to specific devices (LUNs) within the storage system. Virtual 

Private SAN (VP SAN) provides “LUN masking” to limit access between host 

adapters and LUNs attached to SAN Data Gateway SCSI ports. The IBM 

StorWatch SAN Gateway Specialist, an easy to use graphical user interface, 

provides the tools to define SAN Data Gateway channel zoning, the VP SAN 

LUN-masking, and control which host systems have access to specific storage 

devices. 

VPS is applicable mainly in disk environments, so we do not discuss it further 

here. It is also not a capability of the Integrated SAN Data Gateway Module for 

the 3583. You can find more information on VPS in Planning and Implementing 

an IBM SAN, SG24-6116. 

3.5 Operating system device names 

We want to remind you that the operating system device names for FC attached 

tapes are not necessarily presented in the same order as they are installed 

physically in the library. This could cause problems with backup applications 

which are managing the tape library. 


Let’s assume we have one library with three drives on it. The table lists the 

element address which is a number indicating the physical location of a drive 

within an automatic library. Applications use the drive’s element address to 

connect the physical location of the drive to to the drive’s SCSI address. We 

show how to display element addresses for drives as well as the slots in 

“Verifying proper device attachment with tapeutil” on page 274. The drives shown 

are configured in the operating system as rmt0, rmt1 and rmt2, shown in 

Table 3-3. 

Table 3-3 Library with three drives 

Drive# on the library Element address Drive on the operating 

system 

0 256 rmt1 

1 257 rmt0 

2 258 rmt2 

Many backup applications (for example, Legato NetWorker and VERITAS 

NetBackup DataCenter) will assume by default that the devices have been 

assigned their drive names in the same order as the element numbers. That is, 

the lowest drive name has 256, the next, 257, and so on. However, in our 

example this is not true, and so the application will not correctly access the 

drives. If the application tries to mount a cartridge on rmt0, it will issue move 

medium command with the target element address 256, because of the 

application assumes that rmt0 corresponds to the drive at element address 256 

which is the default. But since our drives were configured in a different order by 

the operating system, when the library moves the cartridge to element 256, it is 

actually loaded in drive rmt1. The application now waits indefinitely for the media 

to be mounted on rmt0, or until you kill the request. 

Therefore, check your drive configuration carefully before you configure further 

software. One way to check the drive order is to compare the drive serial 

numbers. Use lscfg -vl rmtx in AIX to get serial numbers (as shown in 

Example 3-21), and compare with the serial numbers physically written on the 

drives. If you need to change them, use smit as described in Section 3.3.4, 

“Persistent binding not applicable to AIX” on page 123 to change the /dev/rmtx 

mapping). 

Example 3-21 lscfg -vl rmtx shows you the serial number 

# lscfg -vl rmt1 


rmt1 20-58-01-1,0 IBM 3580 Ultrium Tape Drive (FCP) 

Manufacturer................IBM 


Machine Type and Model......ULT3580-TD1 

Serial Number...............6811007030 

Device Specific.(FW)........16E0 

On Solaris, we strongly recommend using persistent binding to ensure the drives 

are numbered by the OS in their real physical order (as described in 3.3.1, “The 

need for persistent binding” on page 112). In any case, when configuring your 

library with the backup application, make sure that the defined application drive 

names correspond to the correct operating system device name and element 

number. We will address specific examples of how to do this with different backup 

applications in the following chapters. 


Part 2 

Part 2 LTO 

Ultrium 

with 

applications 

In this part we describe how to implement the LTO Ultrium libraries with some of 

the most popular backup software packages — Tivoli Storage Manager, 

VERITAS NetBackup DataCenter, and Legato NetWorker. 



4 

Chapter 4. 

Implementing Tivoli Storage 

Manager with LTO 

This chapter describes the installation and implementation of Tivoli Storage 

Manager (TSM) 4.2 in the AIX, Sun Solaris and HP-UX environments. In each 

environment we tested the interaction of Tivoli Storage Manager with the 3583 

Ultrium Scalable Tape Library. In particular, we confirmed TSM’s ability to: 

► 

► 

► 

Achieve normal backup/restore processing 

Allow multiple servers to share the 3583 library 

Provide LAN-free backup services 

We begin with a brief overview of Tivoli Storage Manager. For more in depth 

information of this product, refer to these redbooks: 

► 

► 

► 

Tivoli Storage Management Concepts Guide, SG24-4877 

Getting Started with Tivoli Storage Manager: Implementation Guide, 

SG24-5416 

Using Tivoli Storage Manager in a SAN Environment, SG24-6132 


4.1 Tivoli Storage Manager overview 

What is Tivoli Storage Manager? Tivoli Storage Manager is part of the Tivoli 

Storage Management product set — an enterprise-wide solution integrating 

automated network backup, archive and restore, storage management and 

disaster recovery. Tivoli Storage Manager is ideal for heterogeneous, 

data-intensive environments; supporting over 35 platforms and over 250 storage 

devices across LANs, WANs and SANs, plus providing protection for leading 

databases and e-mail applications. See Figure 4-1. 

DB2 UDB 

APPLE 

Macintosh 

AUSPEX** 

DATA 

GENERAL 

DG/UX 

DIGITAL 

OpenVMS 

(SSSI)*** 

UNIX 

Tru64 UNIX 

FUJITSU*** 

HEWLETT- 

PACKARD 

HP-UX 

IBM* 

AIX OS/2 Lan Server 

NUMA-Q OS/2 Warp 

AS/400 OS/390 UNIX 

OS/2 System Services 

MICROSOFT 

Windows 95 

Windows 98 

Windows NT 

Windows NT DEC Alpha 

Windows 2000 

INFORMIX 

Linux 

LOTUS 

DOMINO 

LOTUS 

NOTES 

Tivoli Storage Manager client platforms 

NCR UNIX SVR4 

NEC EWS-UX/V 

MICROSOFT 

Exchange Server 

SQL Server 

ORACLE 

Oracle7 EBU 

Oracle8 RMAN 

R/3 

SYBASE 

TANDEM 

GUARDIAN 

(ETI)*** 

SUN 

MICROSYSTEMS 

Solaris 

Tivoli Data Protection for application 

Family: 

Lotus Domino 

Lotus Notes 

Oracle 

Microsoft SQL Server 

Microsoft Exchange 

Informix 

R/3 

Tivoli Storage Manager also supports: 

IBM DB2 UDB 

Sybase 

SILICON 

GRAPHICS 

IRIX 

SIEMENS NIXDORF 

SINIX 

SINIX Reliant UNIX 

SINIX 386/486 

Tivoli 

Storage 

Manager 

server 

platforms 

SEQUENT PTX 

SCO PYRAMID 

Unix 386 NILE 

Open Desktop 

Open Server 

Windows 

NT 

Windows 

2000 

NOVELL 

NETWARE 

Solaris 

HP-UX 

AIX 

Storage 

Hierarchy 

VM 

OS/2 

OS/400 

MVS 

Disk 

Optical 

Tape 

NSM 

Figure 4-1 Tivoli Storage Manager supported platforms 

Tivoli Storage Manager (TSM) allows users to confidently protect and manage 

information; it integrates unattended network backup and archive capabilities 

with centralized storage management and powerful disaster recovery functions. 

Tivoli Storage Manager is intended for companies with homogeneous or 

heterogeneous platforms and complex environments that include both traditional 


LANs as well as SANs. It is a best-of-breed, scalable storage management 

solution that helps provide consistent and reliable protection and management of 

mission-critical data that is spread across your company's enterprise. It protects 

a broad range of data across the enterprise from the laptop to the data center. 

Tivoli Storage Manager is an industrial-strength centralized storage management 

product for your enterprise. Tivoli Storage Manager can protect the following 

backup-archive clients: Windows 98/NT/2000, NetWare, Macintosh, as well as 

AIX, Sun Solaris, HP-UX, Linux and other UNIX variants, as reflected in 

Figure 4-1. A Tivoli Storage Manager server is provided for OS/390, z/OS, 

Windows NT/2000, AIX, Solaris, HP-UX, and OS/400. This breadth of platform 

coverage affords you the choice in selecting the storage management platform 

that suits your environment and leverages your hardware and software 

investments. Tivoli Storage Manager can help control the cost of distributed 

storage management by leveraging storage resources, helping to reduce the cost 

of downtime and lost data, and helping to increase the productivity of storage 

administrators and end users. 

Tivoli Storage Manager exploits the numerous advantages of SANs with its 

LAN-free and library sharing functions. These help to remove traffic from the 

LAN, allow for multiple Tivoli Storage Manger servers to share a library, and off 

load backup processing from mission-critical servers. 

The latest release of Tivoli Storage Manager is Version 4.2 expands the LAN-free 

backup/restore functionality with additional Tivoli Data Protection platforms and 

databases. It introduces LAN-free backup/restore and archive/retrieve of 

Windows NT/2000, AIX and Solaris files. It exploits Tivoli SANergy to allow for 

LAN-free transfers to file-based storage pools on disk. With this version, HP-UX 

servers can share tape and library resources and S/390 Linux is added as a 

supported TSM client. 

For more information about Tivoli Storage Management, visit this homepage: 

http://www.tivoli.com/products/index/storage_mgr 

4.1.1 TSM commonly used terms 

Let’s discuss a few commonly used TSM terms. 

Server 

A server is a computer system that provides services to one or more clients, or 

other devices over a network. A Tivoli Storage Manager server is the repository 

and manager of all the backed up client data. Administrative policies defined at 

the server control the types of backup performed and retention policies for the 

data. The server also manages the physical media and devices where the 

backed up data is stored. 

Chapter 4. Implementing Tivoli Storage Manager with LTO 149

Client 

A client is a computer system that requests a service of another computer 

system that is typically referred to as a server. Multiple clients may share access 

to a common server. In Tivoli Storage Manager terms, a client is a computer 

system which has data assets requiring protection by the TSM server. The client 

decides what data will be backed up and is subject to the server’s defined 

administrative policies for data retention. Typically, a client’s data is backed up 

automatically by a server scheduled operation. 

Tape library 

A tape library consists of the physical robotics that move cartridges, one or more 

tape drives, and slots for tape storage. It must also have a mechanism for 

controlling the robotics (a library controller), and may also have a library manager 

which maintains inventory and mediates sharing. In most cases, a library does 

not have a built-in library manager, so server-based software has to provide the 

library management function. As an example, the IBM 3494 has a built-in library 

manager, whereas the IBM 3583 does not. 

Tape library sharing 

Tape library sharing is two or more servers sharing the robotics of a tape library. 

The tape drives and slots within the library may or may not be shared among the 

attached servers. 

TSM and tape library sharing 

Now let’s look at the TSM implementation of Tape library sharing. This feature 

allows multiple TSM servers to use the same tape library and drives on a SAN to 

improve both backup/restore performance and tape hardware asset utilization. 

When two or more TSM servers share a library, one server is defined as the 

library manager, and controls the library operations. Other servers, known as 

library clients, use server-to-server communications to contact the library 

manager and request library services. This process is shown in Figure 4-2. 


LAN 

Read Data 

clients 

Mount a tape 

Write Data 

TSM Server 

Library client 

SCSI I/O 

FC 

Tape Lib 

FC 

SCSI I/O 

TSM server 

Library manager 

select drive 

mount, dismount 

volumes 

release, query 

volumes 

Figure 4-2 TSM library sharing overview 

Library manager 

The task of the library manager is to physically control the library. All the 

communication with the library is done by this server. The library manager will 

also serialize the access to the tape drives, so that only one server is using a 

tape drive at the same time. The checkin and checkout of volumes will be 

performed by the library manager. This server is the only server with a library 

inventory table for that library. The ownership is added to the inventory table to 

keep track of which tape belongs to a certain library client. This is the only server 

in the environment that knows all the volumes. The library manager can also use 

the library for its own purposes, without being a library client at the same time (no 

library client configuration is necessary on the manager). 

Library client 

The library client uses server-to-server communications to contact the library 

manager for the handling of the physical hardware. The client send the requests 

to the library manager, and afterwards reads or writes to the tape drive it was 

given access to. To send requests from the library client to the library manager, 

and to send the response from the manager to the client, the server-to-server 

communication protocol is used. That means that a TCP/IP connection is 

necessary and the two Tivoli Storage Manager servers must be configured to 

each other. In an environment with multiple library clients, there is no need to 

have each Tivoli Storage Manager server configured to every other Tivoli Storage 

Manager server, because the communication has always been done directly with 

the library manager. 


LAN-free data transfer 

LAN-free data transfer with TSM allows the SAN to be used as an alternative 

path for moving data between the TSM clients and the server. LAN-free data 

transfer exploits this SAN path by enabling the TSM client to backup and restore 

data directory to and from SAN-attached storage (disk and tape), which is shared 

between the TSM server and client, and managed by the client. This means that 

the data is not transferred over the LAN to the server as in traditional TSM 

backup, but transfers directly from the client to the SAN-attached storage 

devices. TSM currently provides LAN-free data transfer for both normal files as 

well as database applications through the Tivoli Data Protection products. The 

clients require an additional Storage Agent to be installed which directs the 

movement of the data from the client disk to the SAN-attached storage. This 

process is shown in Figure 4-3. For a complete list of operating systems and 

applications supported in a LAN-free environment, refer to: 

http://www.tivoli.com/support/storage_mgr/san/lanfree.html 

TDP or Backup Client 

TSM API 

Storage Agent 

2 

Read Data 

1 

Meta data 

LAN 

3 

3 Write Data 

SAN 

TSM server 

Client Disk 

Tape Library Disk 

Server Storage Hierarchy 

Figure 4-3 TSM LAN-free data transfer overview 

4.2 Configuring non-shared LTO with TSM 

In this section we will discuss configuring the LTO libraries with Tivoli Storage 

Manager in the AIX, Solaris and HP-UX environments. We used a FC connected 

LTO library with AIX and Solaris and direct SCSI attach for HP-UX, since at the 

time of writing, a LTO FC connection to HP-UX with the 2108-R03 SAN Data 

Gateway was not supported. 


The environment in our lab for the TSM installation with a FC connected LTO 

was: 

► 

► 

► 

► 

► 

IBM RS/6000 F50 server with IBM AIX Release 4.3.3 and IBM PCI FC 

adapter (FC 6227) 

Sun E250 SPARC server with Solaris Release 2.7 and QLogic QLA 2200 PCI 

FC adapter 

IBM 2109 FC switch 

IBM 2108-R03 SAN Data Gateway 

3583 Ultrium Scalable Tape Library with two LVD drives 

For a direct SCSI connected LTO, our environment was: 

► HP9000 L1000 server with HP-UX Release 1.00 and SCSI Ultra-2 LVD PCI 

FC adapter A5150A 

► 

3583 Ultrium Scalable Tape Library with two LVD drives 

4.2.1 Installing Tivoli Storage Manager 

We installed Tivoli Storage Manager Version 4.2 on each of our AIX, Solaris and 

HP-UX servers following the instructions in Tivoli Storage Manager for AIX Quick 

Start Guide, GC35-0402, Tivoli Storage Manager for SUN Solaris Quick Start 

Guide, GC35-0406 and Tivoli Storage Manager for HP-UX Quick Start Guide, 

GC35-0412. The default configuration sets up a disk storage pool for storing 

backup data. Note that you need TSM Version 4.1.2 or later to support the 

Ultrium 3580, 3581, 3583 and 3584 models. 

Before you can begin using tape devices as backup destinations, you must do 

the following: 

► 

► 

► 

Attach the devices to your system, which includes physically attaching the 

devices, configuring the device drivers, and determining the device names. 

Define the library, the drives in the library, a device class using the library, and 

a storage pool associated with the device class. 

Include the storage pool in your storage hierarchy. 

To perform these tasks you can use either the administrative client command line 

or Web interface. Our examples use the command line client. The next three 

sections show how to define the LTO devices in each of the specific AIX, Solaris 

and HP-UX standalone (non-shared) environments. 

Refer to the following Web site for Tivoli product requirements and supported 

devices: 

http://www.tivoli.com/support/storage_mgr/requirements.htm 


4.2.2 Configure non-shared tape library and drives for AIX 

You should have already installed the your tape drives and library (including the 

Atape device driver) according to Chapter 2, “Basic LTO setup for UNIX systems” 

on page 29 and Chapter 3, “SAN setup” on page 93. Now determine the names 

of your devices (see Example 4-1). 

Example 4-1 Installed and available drives and media changer on AIX 


rmt0 Available 3A-08-01-0,0 IBM 3580 Ultrium Tape Drive (FCP) 

rmt1 Available 3A-08-01-1,0 IBM 3580 Ultrium Tape Drive (FCP) 

smc0 Available 3A-08-01-6,0 IBM 3583 Library Medium Changer (FCP) 

For performance reasons, we suggest that you use a variable block size length 

and have hardware compression turned on. 

To display or change this setting. Use SMIT -> Devices -> Tape Drive -> 

Change / Show Characteristics of a Tape Drive. 

Example 4-2 reflects a tape drive configured with Block Size = 0, which means 

variable length blocks, and Hardware Compression on Tape = yes. Note that 

TSM can override the device settings for hardware compression by use of the 

FORMAT parameter in the DEFINE DEVICE class command (see 4.2.5, “Define 

device class and storage pool” on page 165). 

Example 4-2 Change / Show Characteristics of a tape drive 





Tape Drive 

rmt0 



fcp 

Description 

IBM 3580 Ultrium Tape 

Drive (FCP) 

Status 

Available 

Location 

3A-08-01-0,0 


fscsi0 

Connection address 2 

SCSI ID 

0x31000 

Logical Unit ID 

0x2000000000000 


0x2001006045161ff5 

New Logical Name [] 

Block Size (0=Variable Length) [0] +# 

Use Hardware Compression on Tape yes + 


Activate volume information logging no + 

Maximum size of log file (in # of entries) [500] +# 

Backward Space/Forward Space Record Mode SCSI + 

Use Immediate Bit in Rewind Commands no + 

Trailer Label Processing no + 

Define the library 

Define the library to Tivoli Storage Manager using the TSM administrative 

command line: 

DEFINE LIBRARY library_name LIBTYPE=SCSI DEVICE=dev_name 

For library_name, enter a user-specified name of the library to be defined. For 

dev_name, enter the full device name, like /dev/smc0. 

tsm: SICILY> DEFINE LIBRARY lto_3583 LIBTYPE=SCSI DEVICE=/dev/smc0 

Define the tape drives 

Next, define each of the library’s drives with: 

DEFINE DRIVE library_name drive_name DEVICE=dev_name 

ELEMENT=element_number CLEANFREQUENCY=Gigabytes/ASNEEDED/NONE 

ONLINE=yes 

For library_name, enter the name of your already defined library (lto_3583 in our 

example). Use the operating system device name of the tape drive being defined 

for device_name. Enter the full device name for dev_name, like /dev/rmt0. The 

element_address is a number that indicates the physical location of a drive within 

an automated library. TSM needs the element address to connect the physical 

location of the drive to the drive’s SCSI address. Get the SCSI element_number 

of your drives from 3584 UltraScalable Tape Library Planning and Operator 

Guide, GA32-0408 or 3583 Ultrium Scalable Tape Library Setup and Operator 

Guide, GA32-0411. 

For 3583 and 3584 element addresses, respectively, you can also get this 

information at: 

http://www.tivoli.com/support/storage_mgr/devices/atab104.htm 


Figure 4-4 shows you the SCSI element address for drives of the 3584 and 

Figure 4-5 shows you the SCSI element address for drives of the 3583. 


Figure 4-4 SCSI element number of tape drives in the 3584 

Drive 

Column3 

105h(261) 

DriveSledF 

104h(260) 

DriveSledE 

103h(259) 

DriveSledD 

102h(258) 

DriveSledC 

101h(257) 

DriveSledB 

100h(256) 

DriveSledA 

Figure 4-5 SCSI element number of tape drives in the 3583 

Or, use tapeutil, open one device and use option 14 Element Inventory. Search 

for Drive/ Address and make sure it shows the correct SCSI bus address 

matching your operating system device driver. The drive address is the element 

address of a tape drive. Example 4-3 shows you two drives with element 

addresses of 256 and 257. 


Example 4-3 Element information returned by tapeutil 

Drive Address 256 

Drive State .................... Normal 

ASC/ASCQ ....................... 0000 

Media Present .................. No 

Robot Access Allowed ........... Yes 

Source Element Address Valid ... No 

Media Inverted ................. No 

Same Bus as Medium Changer ..... Yes 

SCSI Bus Address ............... 0 

Logical Unit Number Valid ...... No 

Volume Tag ..................... 

Hit to continue... 



ASC/ASCQ ....................... 0000 








Volume Tag ..................... 

Keep in mind that your drives are not necessarily configured by the operating 

system in the same order as they are installed physically in the library, as 

described in Section 3.5, “Operating system device names” on page 142. 

Therefore, check carefully that the element address maps with your drive 

configuration. 

For cleanerfrequency we recommend specifying NONE for the 3584. This means 

the library has to take care of cleaning, and cleaning must be enabled on the 

library (automatic cleaning). In this case TSM will not store any information about 

cleaning or the cleaner cartridge, but you can use the StorWatch Specialist for 

this purpose. If you want TSM to be in charge of cleaning the drive, then you 

should use option ASNEEDED. But then you have to disable the autocleaning 

function (host cleaning) on the library using the StorWatch Specialist. Also if you 

use a 3584 and this 3584 is partioned, then every partion needs its own cleaner 

cartridge in the logical partion. In general, TSM-managed cleaning is intended for 

drives and libraries which do not have an automatic cleaning function. 


Because the 3583 does not have autoclean capability, therefore TSM has to take 

care of cleaning and you should specify cleanfrequency ASNEEDED. We 

configured our drives in the 3583 with: 

tsm: SICILY> DEFINE DRIVE lto_3583 drive0 DEVICE=/dev/rmt0 ELEMENT=256 

CLEANFREQUENCY=ASNEEDED ONLINE=yes 

tsm: SICILY> DEFINE DRIVE lto_3583 drive0 DEVICE=/dev/rmt1 ELEMENT=257 


Attention: For LTO drives, cleaning the drive on a periodic basis is not 

recommended (that is, after a set number of gigabytes has been processed by 

the drive). 

Tip: Before defining the drives and libraries, make sure that your library is 

online and your drives are available. Also there should not be any cartridge 

loaded in the drive being defined. 

Continue the implementation with “Define device class and storage pool” on 

page 165. 

4.2.3 Configure non-shared tape library and drives for Solaris 

You should have already installed your tape drives and library according to 

Chapter 2, “Basic LTO setup for UNIX systems” on page 29 and Chapter 3, “SAN 

setup” on page 93. Now determine the names of your devices (Example 4-4). 

Example 4-4 Installed and available drives and media changer on SUN Solaris 


total 52 

drwxrwxr-x 2 root sys 512 Aug 30 10:35 . 

drwxrwxr-x 13 root sys 9216 Sep 3 15:07 .. 



















Refer to Table 2-1 on page 30 regarding the special device file names to be used 

with the UNIX operating system. The recommendation for TSM is to use xst or 

xstc which is No ‘BSD-compatibility’ and ‘rewind on close’. Therefore, we can use 

either /dev/rmt/xst or /dev/rmt/xstc (compression on). 

For performance reasons we suggest that you use a variable length block and 

hardware compression from the LTO drives. The block size length is configured 

in the IBMtape.conf file, as described in 2.3.1, “IBMtape driver installation” on 

page 41. For variable block size, set the block_size=0, as shown in Example 4-5. 

Example 4-5 IBMtape.conf excerpt 

name="IBMtape" 

class="scsi" 


block_size=0 

buffering=1 

immediate=0 

trailer=0 

sili=0; 

To enable hardware compression you can specify the /dev/rmt/xstc device file 

when defining the tape drive. Note that TSM can also override the device settings 

for hardware compression by use of the FORMAT parameter in the DEFINE 

DEVICE class command (see 4.2.5, “Define device class and storage pool” on 

page 165). 


Define the library to TSM as follows: 


For library_name, enter the name of the library to be defined. For dev_name, 

enter the full device name, like /dev/rmt/2smc. 

TSM:SOL-E>DEFINE LIBRARY lto_3583 LIBTYPE=SCSI DEVICE=/dev/rmt/2smc 


Next you can define each of the drives with: 



ONLINE=yes 




for device_name. Enter the full device name for dev_name, like /dev/rmt/0stc. The 






Guide, GA32-0411. 





Figure 4-4 on page 156 shows you the SCSI Element address for drives of the 

3584 and Figure 4-5 on page 156 shows you the SCSI Element address for 

drives of the 3583. 

Or, use tapeutil, open the media changer device and use option 17 Inventory 

(Example 4-3). Search for Drive Address. The drive address is the element 

addresses of a tape drive. 




ASC/ASCQ ....................... 0000 








Volume Tag ..................... 




ASC/ASCQ ....................... 0000 









Volume Tag ..................... 

Keep in mind that your drives are not necessarily configured by the operating 

system in the same order as they are installed physically in the library, as 

described in Section 3.5, “Operating system device names” on page 142. 

Therefore, check carefully that the element address maps with your drive 



the library has to take care of cleaning and cleaning must be enabled on the 









Because the 3583 does not have the capability of autoclean, therefore TSM has 

to take care of cleaning, and you have to use cleanfrequency ASNEEDED. We 


TSM:SOL-E> DEFINE DRIVE lto_3583 drive0 DEVICE=/dev/rmt/0stc 

ELEMENT=256 CLEANFREQUENCY=ASNEEDED ONLINE=yes 

TSM:SOL-E> DEFINE DRIVE lto_3583 drive1 DEVICE=/dev/rmt/1stc 

ELEMENT=257 CLEANFREQUENCY=ASNEEDED ONLINE=yes 



the drive). 




Continue the implementation with “Define device class and storage pool” on 

page 165. 


4.2.4 Configure non-shared tape library and drives for HP-UX 

You should have already installed your tape drives and library according to 


setup” on page 93. Now determine the names of your devices (see 

Example 4-7). 

Example 4-7 Installed and available drives and media changer on HP-UX 










crw-rw-rw- 1 bin bin 245 0x056000 Aug 16 11:14 2chng 

crw-rw-rw- 1 bin bin 245 0xfffffc Aug 16 11:13 atdd.cfg 

crw-rw-rw- 1 bin bin 245 0xfffffd Aug 16 11:13 atdd.dbg 

crw-r--r-- 1 root root 245 0xfffffe Aug 15 11:24 atdd_config 









For performance reasons we suggest that you use hardware compression. 

Recording format ultirumc, which is to configure during the define device class, 

use hardware compression. More information on how to configure the device 

class is in 4.2.5, “Define device class and storage pool” on page 165. 


Define the library to TSM as follows: 


For library_name, enter the name of the library to be defined. For dev_name, 

enter the full device name, like /dev/smc0. 

TSM:EASTER> DEFINE LIBRARY lto_3583 LIBTYPE=SCSI DEVICE=/dev/rmt/2chng 



Next you can define each of the drives with: 



ONLINE=yes 



for device_name. Enter the full device name for dev_name, like /dev/0m. The 






Guide, GA32-0411. 





Figure 4-4 on page 156 shows you the SCSI Element address for drives of the 

3584 and Figure 4-5 on page 156 shows you the SCSI Element address for 

drives of the 3583. 

Or, use tapeutil, open one device and use option 14 Element Inventory. Search 

for Drive Address. The drive address is the element addresses of a tape drive. 

Example 4-3 shows you two drives with the element address of 256 and 257. 




ASC/ASCQ ....................... 0000 








Volume Tag ..................... 





ASC/ASCQ ....................... 0000 








Volume Tag ..................... 

Keep in mind that your drives are not necessarily in the same order as they are 

installed physically in the library, as described in Section 3.5, “Operating system 

device names” on page 142. Therefore, check carefully that the element address 

maps with your drive configuration. 


the library has to take care of cleaning and cleaning must be enabled on the 









Because the 3583 does not have autoclean capability, therefore TSM has to take 

care of cleaning and you should specify cleanfrequency ASNEEDED. We 


TSM:EASTER> DEFINE DRIVE lto_3583 drive0 DEVICE=/dev/rmt/0m ELEMENT=256 


TSM:EASTER> DEFINE DRIVE lto_3583 drive1 DEVICE=/dev/rmt/1m ELEMENT=257 





the drive). 




Continue the implementation with the following section — Define device class 

and storage pool. 

4.2.5 Define device class and storage pool 

Configuring the device classes and storage pools is the same for the AIX, Solaris 

and HP-UX environments. 

Define device class 

Now, configure the device class: 

DEFINE DEVCLASS devclass_name LIBRARY=library_name DEVTYPE=LTO 

FORMAT=Drive/Ultrium/UltriumC MOUNTLIMIT=mount_limit|DRIVES 

The library_name will match the name of the library you previously defined. 

The FORMAT parameter has the following options valid: 

► drive 

► 

► 

The server selects the highest format that can be supported by the drive on 

which a volume is mounted. This value will use the settings on the tape drive 

characteristics (device driver) for hardware compression. This is the default. 

ultrium 

This specifies that TSM writes data that uses the ULTRIUM recording format. 

This format results in a cartridge capacity of 100GB when using Ultrium 

100GB data cartridge. This value overrides the settings on the tape drive 

characteristics for hardware compression. 

ultriumc 

This specifies that TSM writes data that uses the ULTRIUM recording format 

with compression. This format results in a cartridge capacity of approximately 

200GB when using the Ultrium 200GB data cartridge. This value overrides 

the settings on the tape drive characteristics for hardware compression. 


Compression in general improves backup performance. Therefore, we 

recommend that you use compression either with FORMAT=ultriumc, or with 

recording FORMAT= drive, if you have set hardware compression on the device 

driver. 

► 

► 

To use hardware compression in the AIX environment, use settings in the 

tape characteristics (hardware compression=on). 

To use compression in the SUN environment, use /dev/rmt/xstc. 

If you use recording FORMAT=ultriumc, then your LAN-free clients also use 

compression regardless of whether hardware compression is enabled. 

Therefore, we suggest the easiest way is to use FORMAT=ultriumc. 

Note that TSM also provides optional client compression. If this is enabled, then 

clients compress their data before sending it to the storage device. This is 

particularly useful where the network connection between the client and server is 

slow and traffic needs to be minimized. With client compression on, using tape 

drive compression has little effect and is not recommended. You should evaluate 

your environment and requirements (probably by performing some appropriate 

testing) to determine if client compression is of benefit. If so, then enable it, and 

use the uncompressed (FORMAT=ULTRIUM) flag when defining the device 

class. If not, then tape drive compression should be used. 

The option MOUNTLIMIT specifies the maximum number of sequential access 

volumes that can simultaneously be mounted for the device class. This 

parameter is optional. The default is drives, which means that the maximum is 

set to the number of drives installed and available in the library. However, if you 

are going to share the library (for example, for LAN-free backup), Tivoli 

recommends not to use the default value, drives, but to specify instead the exact 

number of tape drives in the library. 

To define the 3583 in our UNIX environment, we called the device class 

3580_devclass, pointing to our previously defined library LTO_3583. We 

specified ULTRIUMC for the FORMAT parameter. We have two drives in the 

library, so we set the MOUNTLIMIT to 2. 

tsm: SICILY> DEFINE DEVCLASS 3580_devclass DEVTYPE=LTO 

LIBRARY=lto_3583 FORMAT=ULTRIUMC MOUNTLIMIT=drives 

Example 4-9 reflects our device class definition. 

Example 4-9 query devclass lto_devclass f=d 

Device Class Name: 3580_DEVCLASS 

Device Access Strategy: Sequential 

Storage Pool Count: 1 

Device Type: LTO 


Format: ULTRIUMC 

Est/Max Capacity (MB): 

Mount Limit: DRIVES 

Mount Wait (min): 60 

Mount Retention (min): 60 

Label Prefix: ADSM 

Library: LTO_3583 

Directory: 

Server Name: 

Retry Period: 

Retry Interval: 

Shared: 

Last Update by (administrator): TONY 

Last Update Date/Time: 08/28/01 15:40:56 

Define storage pool 

Define the storage pool with: 

DEFINE STGPOOL stgpool_name devclass_name MAXSCRATCH=100 

This storage pool will be assigned (written) to the device class we just defined, so 

we specify the name, lto_deviceclass. We used 3583tlo_stgpool as the 

stgpool_name: 

tsm: SICILY> DEFINE STGPOOL 3583lto_stgpool 3580_devclass 

MAXSCRATCH=100 

Example 4-10 shows you the query of the storage pool which we defined above. 

Example 4-10 q stgpool 3583lto_stgpool f=d 

Storage Pool Name: 3583LTO_STGPOOL 

Storage Pool Type: Primary 

Device Class Name: 3580_DEVCLASS 

Estimated Capacity (MB): 2,384,175.0 

Pct Util: 0.2 

Pct Migr: 8.0 

Pct Logical: 100.0 

High Mig Pct: 90 

Low Mig Pct: 70 

Migration Delay: 0 

Migration Continue: Yes 

Migration Processes: 

Next Storage Pool: 

Reclaim Storage Pool: 

Maximum Size Threshold: No Limit 

Access: Read/Write 

Description: 3583 LTO Storage Pool 


Overflow Location: 

Cache Migrated Files?: 

Collocate?: No 

Reclamation Threshold: 60 

Maximum Scratch Volumes Allowed: 100 

Delay Period for Volume Reuse: 0 Day(s) 

Migration in Progress?: No 

Amount Migrated (MB): 0.00 

Elapsed Migration Time (seconds): 0 

Reclamation in Progress?: No 

Volume Being Migrated/Reclaimed: 

Last Update by (administrator): TONY 


Now you can use this storage pool in your management class copygroups so that 

the device will be used for backups or archives. 

4.2.6 Inserting data and cleaner cartridges 

Once we have set up our definitions, we need to identify the cartridges which 

TSM will use, either for storing data, or for cleaning the drives. You can find 

information on how to buy LTO cartridges from IBM, both data and cleaning, from 

the Web site: 

http://www.storage.ibm.com/media/ 

LTO cartridges are also available from several other licensed manufacturers 

which vary from country to country. Check your local sources for details. 

Labeling data cartridges 

TSM requires each tape used to have a physical label written on it for 

identification. You can label volumes individually or process multiple volumes 

with variations on the label command. Here, we want to perform initial labeling of 

all the new tapes in our library. Since our 3583 model has a barcode reader and 

our tapes already have an external barcode label on them, this command will 

cause the matching label to be written onto the tape. We put the new tapes into 

the library, then used the following command: 

tsm: SICILY> LABEL LIBVOLUME LTO_3583 SEARCH=yes LABELSOURCE=barcode 

CHECKIN=scratch 

The parameter SEARCH=Yes means that TSM searches inside the library for 

any unlabeled volume with a readable barcode. It then mounts each cartridge in 

turn and writes the label onto the tape. 


After the above command, you can check on the status of the inserted volumes 

with the q libvol command. Example 4-11 shows all the labeled cartridges with 

scratch status. The element address is also displayed so that we can tell which 

physical location contains each volume. 

Example 4-11 TSM query libvol 

Library Name Volume Name Status Owner Last Use Home Element 

------------ ----------- ---------- ---------- --------- ------------ 

LTO_3583 ABA920 Scratch 4,114 










Inserting data cartridges using the I/O-Station 

For inserting additional data cartridges to the library, use the command: 

LABEL LIBVOLUME lib_name SEARCH=bulk LABELSOURCE=barcode 


If you have only a few cartridges to insert, use the I/O - station and use the option 

SEARCH=bulk. You will be prompted to insert all the volumes to be labeled into 

the I/O port and to indicate when this has been done by issuing the command (at 

an administrative command prompt): 

REPLY request_name 

The server will then load and label the volumes it finds. 

Example 4-12 shows you an output of the activity log when inserting one 

cartridge into the 3583. We used this command: 

tsm: SICILY> LABEL LIBVOLUME LTO_3583 SEARCH=bulk LABELSOURCE=barcode 


Followed by this command to continue the operation: 

tsm: SICILY> REPLY 011 

Example 4-12 TSM query actlog for insert new cartridges 

09/01/01 18:49:11 ANR2017I Administrator TONY issued command: LABEL 

LIBVOLUME LTO_3583 SEARCH=BULK LABELSOURCE=BARCODE 

CHECKIN=SCRATCH 


09/01/01 18:49:11 ANR0984I Process 23 for LABEL LIBVOLUME started in the 

BACKGROUND at 18:49:11. 

09/01/01 18:49:11 ANR8799I LABEL LIBVOLUME: Operation for library LTO_3583 

started as process 23. 

09/01/01 18:49:11 ANR0609I LABEL LIBVOLUME started as process 23. 

09/01/01 18:49:12 ANR8373I 011: Fill the bulk entry/exit port of library 

LTO_3583 with all LTO volumes to be processed within 60 

minute(s); issue 'REPLY' along with the request ID when 

ready. 

09/01/01 18:49:18 ANR8336I Verifying label of LTO volume ABA922 in drive 

3580_DEVICE1 (/dev/rmt1). 

09/01/01 18:49:44 ANR2017I Administrator TONY issued command: REPLY 011 

09/01/01 18:49:44 ANR8499I Command accepted. 

09/01/01 18:49:48 ANR2017I Administrator TONY issued command: QUERY ACTLOG 

09/01/01 18:51:00 ANR8810I Volume ABA926 has been labeled in library 

LTO_3583. 

09/01/01 18:51:42 ANR8427I CHECKIN LIBVOLUME for volume ABA926 in library 

LTO_3583 completed successfully. 

09/01/01 18:52:40 ANR8810I Volume ABA925 has been labeled in library 


09/01/01 18:53:22 ANR8427I CHECKIN LIBVOLUME for volume ABA925 in library 


Inserting unlabeled data cartridges using the I/O-Station 

If your library has a barcode reader (such as the 3583 Ultrium Scalable Tape 

Library or 3584 Ultrium UltraScalable Tape Library), you will almost certainly be 

using this reader to label the cartridges. However, if you have an LTO model (for 

example, the 3581) which is not equipped with a barcode reader you need to be 

able to manually specify a label. If the media is labeled with a barcode anyway, 

we strongly recommend that you use this string to label the volume as this will 

make it much easier for your operators to process the volumes. 

To insert an unlabeled cartridge, specify the volume name you want to use for the 

cartridge and put the cartridge in the I/O station. 

LABEL LIBVOLUME lib_name volume_name CHECKIN=scratch 

You will be prompted to insert the cartridge and to indicate when it is available by 

issuing the command (at an administrative command prompt): 

REPLY request_name 

You can insert only one cartridge at a time which will then be labeled by the 

server. 

Example 4-13 shows you an output of the activity log for the inserting process of 

one unlabeled cartridge in a 3583. We used this command: 


tsm: SICILY> LABEL LIBVOLUME LTO_3583 unlab02 CHECKIN=SCRATCH 

Followed by this command to continue the operation: 

tsm: SICILY> REPLY 013 

Example 4-13 TSM query volume output for insert unlabeled cartridge 

09/03/01 15:12:04 ANR2017I Administrator TONY issued command: LABEL 

LIBVOLUME LTO_3583 unlab02 CHECKIN=SCRATCH OVERWRITE=YES 

09/03/01 15:12:04 ANR0984I Process 28 for LABEL LIBVOLUME started in the 


09/03/01 15:12:04 ANR8799I LABEL LIBVOLUME: Operation for library LTO_3583 


09/03/01 15:12:04 ANR0609I LABEL LIBVOLUME started as process 28. 

09/03/01 15:12:04 ANR0405I Session 1368 ended for administrator TONY 

(WebBrowser). 

09/03/01 15:12:06 ANR8323I 013: Insert LTO volume UNLAB02 R/W into 

entry/exit port of library LTO_3583 within 60 minute(s); 

issue 'REPLY' along with the request ID when ready. 

09/03/01 15:12:25 ANR2017I Administrator SERVER_CONSOLE issued command: 

REPLY 013 


09/03/01 15:15:32 ANR8810I Volume UNLAB02 has been labeled in library 


09/03/01 15:16:15 ANR8427I CHECKIN LIBVOLUME for volume UNLAB02 in library 


09/03/01 15:16:16 ANR8800I LABEL LIBVOLUME for volume UNLAB02 in library 


09/03/01 15:16:16 ANR0985I Process 28 for LABEL LIBVOLUME running in the 

BACKGROUND completed with completion state SUCCESS at 

15:16:16. 

Inserting cleaner cartridges 

If you have chosen to have TSM manage library cleaning, you need to insert 

some cleaning cartridges. Use: 

tsm: SICILY> CHECKIN LIBVOLUME LTO_3583 STATUS=cleaner 

CHECKLABLE=barcode SEARCH=bulk CLEANINGS=50 

The parameter SEARCH=bulk means that TSM will search the library entry/Exit 

ports for usable volumes to checkin. You should set the CLEANINGS parameter 

to the number of uses specified for your cartridge. Insert one or more cleaner 

cartridges into the I/O station and reply to the request issued by the server as 

shown in Example 4-14. 

Example 4-14 Insert cleaner cartridge 

09/01/01 20:02:56 ANR2017I Administrator TONY issued command: CHECKIN 


LIBVOLUME LTO_3583 STATUS=CLEANER CHECKLABEL=BARCODE 

MOUNTWAIT=60 SEARCH=BULK CLEANINGS=50 

09/01/01 20:02:56 ANR0984I Process 4 for CHECKIN LIBVOLUME started in the 


09/01/01 20:02:56 ANR8422I CHECKIN LIBVOLUME: Operation for library 

LTO_3583 


09/01/01 20:02:56 ANR0609I CHECKIN LIBVOLUME started as process 4. 

09/01/01 20:03:06 ANR8468I LTO volume ABA922 dismounted from drive 

3580_DEVICE1 (/dev/rmt1) in library LTO_3583. 

09/01/01 20:03:07 ANR8373I 001: Fill the bulk entry/exit port of library 

LTO_3583 with all LTO volumes to be processed within 60 

minute(s); issue 'REPLY' along with the request ID when 

ready. 


09/01/01 20:03:21 ANR2017I Administrator TONY issued command: REPLY 001 




09/01/01 20:03:46 ANR8430I Volume CLNI17 has been checked into library 


09/01/01 20:03:46 ANR1434W No files have been identified for automatically 

storing device configuration information. 

09/01/01 20:03:46 ANR8431I CHECKIN LIBVOLUME process completed for library 

LTO_3583; 1 volume(s) found. 

09/01/01 20:03:46 ANR0985I Process 4 for CHECKIN LIBVOLUME running in the 

BACKGROUND completed with completion state SUCCESS at 

20:03:46. 

4.2.7 Performance hints for TSM and LTO 

Because TSM writes a record of each file backed up to its database for tracking, 

it will perform better with large file workloads than with small file workloads. This 

is because of the proportion of time spent doing database writes and updates as 

a fraction of the total backup time. Therefore, to get the best performance with 

TSM and LTO drives, be sure to raise the parameter TXNGroupmax on the 

server to its maximum value, 256. This parameter controls how many files are 

transferred as a group between the client and server. On the client, also set the 

parameter TXNBytelimit to its maximum of 2097152. This parameter specifies 

the number of kilobytes the client program can buffer together in a transaction 

before it sends data to the server. 

To set TXNGroupmax, edit the dsmserv.opt on the TSM server as shown in 

Example 4-15. If you are using the Storage Agent for LAN-free backup, 

(discussed in “AUDIT LIBRARY” on page 180), you should also set this 

parameter to the same value in the options file dsmsta.opt. 


Example 4-15 TXNGroupmax entry in dsmserv.opt and dsmsta.opt 

*============================================================================== 

* 

* TXNGROUPMAX 

* 

* Specifies the maximum number of files transferred as a group between 

* the client and storage agent. 

* 

* Syntax 

* +------------------+----------------------------------------------+ 

* | TXNGroupmax | value | 

* +------------------+----------------------------------------------+ 

* 

* Parameters 

* value Specifies the maximum number of files that are 

* transferred as a group between the client and 

* storage agent. The minimum value is 4 and the maximum 

* value is 256. The default value is 40. 

* 

* Examples 

TXNGroupmax 256 

* 

To set the client parameter TXNBytelimit, edit dsm.sys (for UNIX clients) or 

dsm.opt (for all other clients). The entry should look like: 

Example 4-16 TXNBytelimit entry in dsm.opt/dsm.sys 

TXNByte 2097152 

In general, small file workloads will backup faster if they are staged initially to a 

disk storage pool which then migrates to the LTO pool. More suggestions on 

storage pool configuration are in Getting Started with Tivoli Storage Manager: 

Implementation Guide, SG24-5416. We also recommended the use of 

compression on the LTO drives as previously discussed in this chapter. 

4.3 TSM library sharing 

Several TSM Servers can share the same library using the library sharing 

feature, as shown in Figure 4-2 on page 151. Tape library sharing is supported 

among any combination of Windows NT, Windows 2000, AIX, Solaris and HP-UX 

servers. Some operating system platforms were supported at the TSM Version 

3.7 level, however others were introduced later. We recommend using the latest 

code level, which was Version 4.2 at the time of writing. 


You can download the latest code fixes from: 

http://www.tivoli.com/support/storage_mgr/adsercli.htm 

All servers which will share the library need to have a FC/SAN connection to the 

tape drives in the library. One server is the library controller and is designated the 

library manager server. The other servers, known as library client, make 

requests for library servers from the library manager, no longer to the library itself 

as in non-shared library configurations. For instance, if a library client wants to 

write data to a tape, then the client has to ask the library manager to mount the 

tape volume. After the tape is mounted by the library manager, then the library 

client can write the data directly over the physical SAN path to the tape drive. 

In our lab environment, we had three SAN-attached servers, SICILY, BRAZIL and 

SOL-E, which were accessing the 3583 Ultrium Scalable Tape Library via SAN 

Data Gateway and FC switch. This set up is shown in Figure 4-6. We will show 

how to set up SICILY as the library manager and BRAZIL and SOL-E as library 

clients. 

LAN 

Sicily Brazil Sol-e 

AIX 

AIX 

Solaris 

TSM Library TSM Library TSM Library 

Manager 

Client 

Client 

2109 Fibre Channel Switch 

2108 SAN Data Gateway 

3583 

Figure 4-6 Our lab library sharing environment 


4.3.1 Configuring the Library Manager to share libraries 

First, install and configure your tape library and drives on the system which will 

be the Library Manager, as described in Section 4.2, “Configuring non-shared 

LTO with TSM” on page 152. 

Because library sharing uses server-to-server communication for sending and 

receiving request, the first step is to set up server-to-server communications 

between the library manager and the library clients. 

Set the following parameters on the Library Manager: servername, 

serverpassword, serverhladdress, serverlladdress and set crossdefine to on. 

SET SERVERNAME server_name 

SET SERVERPASSWORD server_password 

SET CROSSDEFINE on 

SET SERVERHLADDRESS server_ip_address 

SET SERVERLLADDRESS server_ip_portaddress 

We configured the Library Manager server SICILY as shown in Example 4-17. 

Example 4-17 Configure server SICILY 

tsm: SICILY>set servername sicily 

ANR2094I Server name set to SICILY. 

tsm: SICILY>set serverpassword sicily 

ANR2131I Server password set. 

tsm: SICILY>set crossdefine on 

ANR2135I Crossdefine set to ON. 

tsm: SICILY>set serverhladdress 9.1.150.40 

ANR2132I Server hladdress set to 9.1.150.40. 

tsm: SICILY>set serverlladdress 1500 

ANR2133I Server lladdress set to 1500. 

Now, configure the library so that it can be shared with the Library Clients. Do 

this with the shared=yes parameter on the UPDATE LIBRARY command. We used: 

UPDATE LIBRARY lto_3583 SHARED=yes 


4.3.2 Configuring the Library Client 

Now you need to configure each Library Client. Set each of the following 

parameters: servername, serverhladdress, serverlladdress and set crossdefine 

to on. 

SET SERVERNAME server_name 

SET SERVERPASSWORD server_password 

SET CROSSDEFINE on 

SET SERVERHLADDRESS server_ip_address 

SET SERVERLLADDRESS server_ip_portaddress 

We configured the Library Client BRAZIL as shown in Example 4-18. 

Example 4-18 Configure server BRAZIL 

tsm: BRAZIL>set servername brazil 

ANR2094I Server name set to BRAZIL. 

tsm: BRAZIL>set serverpassword brazil 

ANR2131I Server password set. 

tsm: BRAZIL>set crossdefine on 

ANR2135I Crossdefine set to ON. 

tsm: BRAZIL>set serverhladdress 9.1.150.57 

ANR2132I Server hladdress set to 9.1.150.57. 

tsm: BRAZIL>set serverlladdress 1500 

ANR2133I Server lladdress set to 1500. 

The definitions for the SOL-E Library Client are similar. 

Now, you must create a server definition for the Library Manager on each of the 

Library Clients, so that they will know which system is managing the tape 

libraries. Use the DEFINE SERVER command. 

DEFINE SERVER library_manager_servername SERVERPASSWORD=library_ 

manager_password HLADDRESS=library_manager_ip_address 

LLADDRESS=library_manager_port_number COMM=tcpip 

Here we are defining the server SICILY on the Library Client, BRAZIL: 


tsm: BRAZIL> DEFINE SERVER sicily SERVERPASSWORD=sicily 

HLADDRESS=9.1.150.40 LLADDRESS=1500 COMM=tcpip 

Again, define the server SICILY on the other Library client, SOL-E. 

4.3.3 Define library and drives on Library Client 

You should have already performed basic tape drive installation on each of the 

Library Client systems as described in Chapter 2, “Basic LTO setup for UNIX 

systems” on page 29 and Chapter 3, “SAN setup” on page 93. The installation of 

the SCSI Medium Changer (SMC) on the Library Clients is not needed as only 

the Library Manager interfaces directly with the SMC device driver. 

On each Library Client machine, define the shared library and the shared drives. 

First define the library with: 

DEFINE LIBRARY library_name LIBTYPE=shared 

PRIMARYLIBMANAGER=lib_manager_name 

Note the use of the LIBTYPE=shared parameter. You must use the same 

library_name as was defined on the library server. We defined the library on both 

BRAZIL and SICILY: 

DEFINE LIBRARY lto_3583 LIBTYPE=shared PRIMARYLIBMANAGER=sicily 

Example 4-19 shows the defined shared library on the Library Client BRAZIL. 

Example 4-19 Shared library displayed with q library f=d 

Library Name: LTO_3583 

Library Type: SHARED 

Device: 

Private Category: 

Scratch Category: 

External Manager: 

Shared: No 

Primary Library Manager: SICILY 

Last Update by (administrator): ADMIN 


Now, define each of the tape drives with: 

DEFINE DRIVE library_name drive_name DEVICE=dev_name ONLINE=YES 

For library_name, enter the name of your already defined library. Specify the 

name of your to be defined tape drive for device_name. Enter the full device 

name for dev_name, like /dev/rmt0. 


Configure drives on an AIX Library Client 

Get the device name from lsdev -Cc tape (see Example 4-21). 

Example 4-20 lsdev -Cc tape on brazil 





Define each tape drive with this command. Use the same drive names as you 

defined on the library manager. We defined: 

DEFINE DRIVE lto_3583 drive0 DEVICE=/dev/rmt0 ONLINE=yes 

DEFINE DRIVE lto_3583 drive1 DEVICE=/dev/rmt1 ONLINE=yes 

You can check the result with Q DRIVE as shown in Example 4-21. 

Example 4-21 q drive on the BRAZIL server 

Library Name Drive Name Device Type Device ON LINE 

------------ ------------ ----------- ---------------- ------------------- 

LTO_3583 3580_DEVICE0 LTO /dev/rmt0 Yes 

LTO_3583 3580_DEVICE1 LTO /dev/rmt1 Yes 

Now you can define the device class and the storagepool using the library 

definitions as described in 4.2.5, “Define device class and storage pool” on 

page 165. 

Configure drives on a SUN Solaris Library Client 

Get the device name from ls -la /dev/rmt (see Example 4-22). 

Example 4-22 ls -la /dev/rmt output from sol-e 


total 52 




















Define each tape drive with this command. Use the same drive names as you 

defined on the library manager. We defined: 

DEFINE DRIVE lto_3583 3580_drive0 DEVICE=/dev/rmt/0stc ONLINE=yes 

DEFINE DRIVE lto_3583 3580_drive1 DEVICE=/dev/rmt/1stc ONLINE=yes 

You can check the result with Q DRIVE as shown in Example 4-23. 

Example 4-23 q drive on the sol-e server 

Library Name Drive Name Device Type Device ON LINE 

------------ ------------ ----------- ---------------- ----------------- 

LTO_3583 3580_DEVICE0 LTO /dev/rmt/0stc Yes 

LTO_3583 3580_DEVICE1 LTO /dev/rmt/1stc Yes 

Now you can define the device class and the storagepool using the library 

definitions as described in 4.2.5, “Define device class and storage pool” on 

page 165. 

4.3.4 Administering shared libraries 

When using a shared LTO library, it is important to remember that the Library 

Clients cannot directly access the library robotics (medium changer) themselves. 

Any requests for tape movement must be requested through the Library 

Manager. While library clients do not directly access the library robotics, once a 

volume is mounted in a drive, the Library Clients can read or write data directly to 

the drive via the SAN data path. This means that certain commands give different 

output, or are not applicable in some circumstances. 

QUERY LIBVOLUME 

If you use the QUERY LIBVOLUME on the Library Manager the output is different to 

what it was before library sharing. The output now also displays which TSM 

server is the “owner” of each volume as shown in Example 4-24. 

Example 4-24 Q LIBVOL on a Library Manager 

Library Name Volume Name Status Owner Last Use Home Element 

------------ ----------- ---------- ---------- --------- ------------ 

LTO_3583 ABA920 Private SICILY Data 4,114 

LTO_3583 ABA922 Private SICILY Data 4,104 




LTO_3583 ABA925 Private SICILY 4,130 

LTO_3583 ABA926 Private SICILY 4,098 

LTO_3583 ABA927 Private BRAZIL Data 4,102 




The same command on a Library Client will not report the library contents: 

Example 4-25 Q LIBVOL on a Library Client 

tsm: BRAZIL> QUERY LIBVOLUME 

ANR2017I Administrator TONY issued command: QUERY LIBV 

ANR2034E QUERY LIBVOLUME: No match found using this criteria. 

QUERY VOLUME 

Querying the volumes (QUERY VOLUME) on any attached server will report only 

those volumes that belong to that server. 

Example 4-26 QUERY VOLUME on a shared library 

tsm: SICILY>q volume 

Volume Name Storage Device Estimated Pct Volume 

Pool Name Class Name Capacity Util Status 

(MB) 

------------------------ ----------- ---------- --------- ----- -------- 

ABA920 3583LTO_ST- 3580_DEVC- 95,367.0 2.4 Filling 

GPOOL LASS 

ABA922 3583LTO_ST- 3580_DEVC- 190,734.0 0.8 Filling 

GPOOL LASS 

AUDIT LIBRARY 

Performing an AUDIT LIBRARY command from a Library Client does not actually 

perform a physical audit on the shared library; instead, the library client’s 

database is checked and synchronized against the library manager’s database. 

No tape mounts are performed. 

Performing this command on the Library Manager mounts or inspects tapes in 

the library, the same as it does in non-shared configurations. 


CHECKIN and CHECKOUT LIBV 

Checking in or out of a volume from a shared library must be done from the 

Library Manager. If you try to execute the CHECKOUT LIBVOLUME command from a 

Library Client, the operation will fail with an error. 

LABEL LIBVOLUME 

When the LTO library is shared, tape labeling must be performed on the Library 

Manager. If you try to execute the LABEL LIBVOLUME command from a Library 

Client, the operation will fail with an error. 

You can share the scratch volumes between all the Library Clients. This because 

the Library Manager serves as a central repository for information about all the 

volumes contained in the library. Since all media handling requests are handled 

by the Library Manager, it can keep track of which tape belongs to which Library 

Client, and which tapes are unallocated and available for use by any Library 

Client. There is no special configuration required to utilize this feature; simply 

make sure there is an adequate supply of labeled scratch tapes in the library and 

the Library Manager will allocate them as needed to the requesting Library 

Clients. Once a particular volume is assigned to a Library Client, the Library 

Manager flags that volume as belonging to that client, and prevents any other 

client from accessing it. 

Note: The library manager is itself a fully functional TSM server - that is, it 

typically supports its own backup client workload and can access and use the 

volumes just like a library client. However it also has the particular function of 

overall management of the devices and tape inventory. 

4.4 LAN-free data transfer 

LAN-free client data transfer is based on the library sharing technology. It 

requires the clients which will send their data to have SAN access to the storage 

devices (LTO tape in our example). The TSM server also requires SAN tape 

library access. The function is provided by separating the actual application data 

(the client data written to or read from the storage pools) from the meta or control 

data (information contained in the TSM database). The metadata is transmitted 

over the LAN from the TSM client to the TSM server while the application data is 

moved over the SAN from the TSM client to the storage device. Refer to 

Figure 4-7 where we show our lab setup. Again we used SICILY as the TSM 

server and configured two LAN-free clients, BRAZIL and SOL-E. Our 

SAN-attached tape library was the 3583 Ultrium Scalable Tape Library, attached 

to the SAN via SAN Data Gateway and FC Switch. 


A core component of LAN-free technology is the Storage Agent which is a piece 

of code installed on each LAN-free client. Essentially, the Storage Agent can be 

viewed as a lightweight TSM server that has no database or storage hierarchy of 

its own. It behaves like a Library Client which contacts the TSM server for volume 

access, and reads/writes data directly to SAN-attached devices. 

Sol-e/ Brazil 

TDP or Backup Client 

TSM API 

Storage Agent 

2 

Read Data 

1 

Meta data 

LAN 

3 

3 Write Data 

SAN 

Sicily 

TSM server 

Client Disk 

Tape Library Disk 

Server Storage Hierarchy 

Figure 4-7 LAN-free backup lab example 

4.4.1 LAN-free client setup 

We installed and configured our LAN-free clients according to the detailed 

instructions in the following publications: 

► 

► 

► 

AIX - Managed System for SAN Storage Agent User’s Guide, GC36-0001 

Sun Managed System for SAN Storage Agent User’s Guide, GC36-0002 


We used Version 4.2 of the Storage Agent and the V4.2 backup/archive clients 

and API on AIX and Solaris. 

Table 4-1 is a summary of the tasks necessary to configure LAN-free 

backup-archive client data transfer. The detailed instructions are contained in the 

User’s Guides. 


Before beginning, set up your TSM server (SICILY in our example) as a Library 

Manager as described in Section 4.3, “TSM library sharing” on page 173): 

Table 4-1 Configuring LAN-free backups 

Step On Server On Client 

1) .................................................... Install tape drives according 

Chapter 2, “Basic LTO setup for 

UNIX systems” on page 29 and 

Chapter 3, “SAN setup” on 

page 93. 

2) ................................................... Install client and TSM API code. 

3) .................................................... Install and setup the Storage 

Agent. 

4) .................................................... Modify dsm.sys of 

backup-archive client with: 

ENABLELANFREE YES 

5) Define each LAN-free client 

storage agent as if it was a server. 

Use DEFINE SERVER command. 

6) Define a new policy domain for 

LAN-free clients and register (or 

update) those clients to that 

domain. 

7) Associate the copy groups for the 

default management class in the 

LAN-free client’s policy domain 

with the storage pool defined for 

the SAN-attached shared library. 

8) .................................................... Obtain Tape Device information 

(see 4.4.3 Mapped drive 

configuration for AIX LAN-free 

client and 4.4.4 Mapped drive 

configuration for Solaris LAN-free 

client). 

9) Use DEFINE DRIVEMAPPING 

command to relate the client 

device names to the server 

defined devices (see 4.4.3 

Mapped drive configuration for 

AIX LAN-free client and 4.4.4 

Mapped drive configuration for 

Solaris LAN-free client). 


Step On Server On Client 

10) .................................................... Test if data transfer goes over the 

SAN - see Section 4.4.5, “Test if 

data transfer goes over the SAN” 

on page 186 

4.4.2 Mapped drive considerations 

Consider the following guidelines for mapping your SAN drives: 

► 

► 

► 

Map all drives. Problems can occur if you do not define drive mappings on the 

server for each drive in a library. For example, during backup operations, all 

drives that have been mapped could be used by storage agents backing up 

data. Backup operations will fail if you try to use other drives in a library that 

has not been mapped. When drive mappings are defined for each drive in a 

library, backup operations wait until the next drive is available for the transfer 

of data. 

Limit the number of drives available to each LAN-free clients. Use the 

MAXNUMMP parameter on the REGISTER NODE or UPDATE NODE 

command to limit the number of drives that are available for the storage agent 

to use on behalf of the client. 

Review tape device names. For the same tape device, the drive name as 

known to the server will probably not match the device name as known to the 

storage agent, as described in Section 3.5, “Operating system device names” 

on page 142. 

Define the drivemappings on the Library Manager. Use the DEFINE DRIVEMAPPING 

command: 

DEFINE DRIVEMAPPING storageagent_name library_name drive_name 

DEVICE=dev_name_on_storageagent 

Where: 

► storageagent_name is the name of the storage agent 

► library_name is the name of the tape library 

► device_name is the name of the tape drive on the TSM server 

► dev_name_on_storageagent is the name of the tape drive on the storage 

agent (/dev/rmt0) 


4.4.3 Mapped drive configuration for AIX LAN-free client 

We configured BRAZIL as a Storage Agent and gave it the name rio_lanfree. We 

defined both drives in the library. We found out the device names which are 

visible to BRAZIL with the lsdev -Cctape command. 

Example 4-27 lsdev -Cc tape on BRAZIL 





Then on SICILY, we map BRAZIL’s view of the devices to the real drives with 

these commands: 

DEFINE DRIVEMAPPING rio_lanfree lto_3583 3580_device0 DEVICE=/dev/rmt0 

DEFINE DRIVEMAPPING rio_lanfree lto_3583 3580_device1 DEVICE=/dev/rmt1 

4.4.4 Mapped drive configuration for Solaris LAN-free client 

We configured SOL-E as a Storage Agent and gave it the name sol_lanfree. We 

defined both drives in the library. We found out the device names which are 

visible to SOL-E with the ls -la /dev/rmt command. 

Example 4-28 ls -la /dev/rmt output from sol-e 


total 52 




















Then on SICILY, we map SOL-E’s view of the devices to the real drives with these 

commands: 

DEFINE DRIVEMAPPING sol_lanfree lto_3583 3580_device0 

DEVICE=/dev/rmt/0stc 

DEFINE DRIVEMAPPING sol_lanfree lto_3583 3580_device1 

DEVICE=/dev/rmt/1stc 

4.4.5 Test if data transfer goes over the SAN 

To verify that you have configured your system correctly for LAN-free data 

movement, run a backup operation from the client. First check to see that the 

backup completes successfully. 

To check whether the data was sent over the SAN (that is, LAN-free), use the 

QUERY ACTLOG commands on the TSM server to which the LAN-free client is 

connected: 

QUERY ACTLOG SEARCH=storage_agent_name 

Note: If the query finds entries in the activity log that relates to the storage 

agent, the client is using LAN-free data transfer. 

Example 4-29 shows you the activity log during a LAN-free backup from the 

client rio_lanfree. 

Example 4-29 queryactlog search=rio_lanfree 

09/05/01 18:28:00 ANR0408I Session 68 started for server RIO_LANFREE 

(AIX-RS/6000) (Tcp/Ip) for library sharing. 

09/05/01 18:29:17 ANR0409I Session 68 ended for server RIO_LANFREE 

(AIX-RS/6000). 


(AIX-RS/6000) (Tcp/Ip) for event logging. 

09/05/01 18:29:50 ANR0400I (Session: 69, Origin: RIO_LANFREE) Session 111 

started for node RIO_LANFREE (AIX) (ShMem). 


started for node RIO_LANFREE (AIX) (ShMem). 

09/05/01 18:29:50 ANR8337I (Session: 69, Origin: RIO_LANFREE) LTO volume 

UNLAB02 mounted in drive 3580_DEVICE1 (/dev/rmt1). 





09/05/01 18:33:32 ANR0403I Session 67 ended for node RIO_LANFREE (AIX). 

09/05/01 18:33:35 ANE4952I (Session: 66, Node: RIO_LANFREE) Total number of 

objects inspected: 5 



objects backed up: 4 


objects updated: 0 


objects rebound: 0 


objects deleted: 0 


objects expired: 0 


objects failed: 0 


bytes transferred: 1.64 GB 

09/05/01 18:33:35 ANE4963I (Session: 66, Node: RIO_LANFREE) Data transfer 

time: 

228.31 sec 

09/05/01 18:33:35 ANE4966I (Session: 66, Node: RIO_LANFREE) Network data 

transfer rate: 7,570.18 KB/sec 

09/05/01 18:33:35 ANE4967I (Session: 66, Node: RIO_LANFREE) Aggregate data 

transfer rate: 4,987.07 KB/sec 

09/05/01 18:33:35 ANE4968I (Session: 66, Node: RIO_LANFREE) Objects 

compressed by: 0% 

09/05/01 18:33:35 ANE4964I (Session: 66, Node: RIO_LANFREE) Elapsed 

processing time: 00:05:46 

09/05/01 18:33:39 ANR8336I (Session: 69, Origin: RIO_LANFREE) Verifying 

label of LTO volume UNLAB02 in drive 3580_DEVICE1 

(/dev/rmt1). 



09/05/01 18:34:14 ANR8468I (Session: 69, Origin: RIO_LANFREE) LTO volume 

UNLAB02 dismounted from drive 3580_DEVICE1 (/dev/rmt1) in 

library LTO_3583. 




ended for node RIO_LANFREE (AIX). 

09/05/01 18:35:23 ANR0403I Session 66 ended for node RIO_LANFREE (AIX). 

09/05/01 18:40:14 ANR2017I (Session: 69, Origin: RIO_LANFREE) Administrator 

SERVER_CONSOLE issued command: HALT 



09/05/01 18:40:15 ANR0991I (Session: 69, Origin: RIO_LANFREE) Storage agent 

shutdown complete. 

09/05/01 18:40:15 ANR0479W Session 69 for server RIO_LANFREE (AIX-RS/6000) 

terminated - connection with server severed. 

09/06/01 08:45:39 ANR2017I Administrator ADMIN issued command: QUERY ACTLOG 

search=rio_lanfree 

09/06/01 08:46:49 ANR2017I Administrator ADMIN issued command: QUERY ACTLOG 

BEGINDATE=09/06/2001 BEGINTIME=NOW-24:00 ENDDATE=09/06/- 

2001 ENDTIME=NOW SEARCH=rio_lanfree ORIGINATOR=ALL 


Check also the performance panel from your SAN switch and verify that the data 

transfer goes over the correct FC ports. On Figure 4-2 you see the performance 

panel from the FC switch in our lab. The TSM server SICILY HBA is connected 

on port 14, the TSM LAN-free client BRAZIL (rio_lanfree) is connected on port 12 

and the SDG for the 3583 is connected to port 0. As you can see, data is being 

transferred on both port 0 and port 12, indicating flow between the LAN-free 

client and the tape drive. But there is no activity on port 14, which is the TSM 

server SICILY. Therefore, it is not involved in writing the backup data to the tape 

drive. 

Table 4-2 Performance panel during a LAN-free backup 


4.5 TSM and 3584 redundant library control paths 

The 3584 Ultrium UltraScalable Tape Library allows multiple control paths for 

each logical library to be enabled. This provides improved redundancy to the 

3584. If there was only a single control path available, a failure in this path would 

result in loss of communication with the library. Therefore, multiple control paths 

eliminate this single point of failure. The issue is that software applications are 

not able to automatically handle these handle multiple control paths to one 

logical library. 

If you enable multiple control paths on your 3584, you will see several SCSI 

media changer devices on AIX, like /dev/smc0, /dev/smc1, depending on how 

many control paths have been enabled. 

Figure 4-8 shows you a 3584 with two drives. Both drives have the control path 

enabled. 

Server 

SCSI 

3584 Lxx 

Control path 

enabled 

FC 

Aapter 

FC 

Control path 

enabled 


Figure 4-8 3584 with multiple control paths enabled 

The lsdev -Cc tape shows you two tape drives and two smc devices. See 

Example 4-30. 

Example 4-30 3584 connect over two HBA 







In TSM you should configure only one control path to the library. If you try to 

configure the second control path, then you will get a second library. Therefore, 

do not configure the second control path in TSM. 

If or when the drive with the enabled control path fails, you can maintain access 

to the library by updating it to point to another control path. Use the UPDATE 

LIBRARY command: 

UPDATE LIBRARY library_name DEVICE=dev_name 

We configured our library initially using the first control path: 

DEFINE LIBRARY lto_3584 LIBTYPE=SCSI DEVICE=/dev/smc0 

If we had a failure on the first drive occur, we could switch to the second path of 

the SCSI medium changer with: 

UPDATE library lto_3584 DEVICE=/dev/smc1 

This moves the control path device from the existing device to /dev/smc1. 


5 

Chapter 5. 

Implementing VERITAS 

NetBackup DataCenter with 

LTO 

In this chapter we give you detailed information on the implementation of the 

VERITAS NetBackup DataCenter product with LTO drives and libraries. This 

includes: 

► 

► 

► 

Product overview 

Installation and library definition (Solaris environment) 

Implementation and use (Solaris environment) 



► 

► 

Install the VERITAS NetBackup DataCenter product in a UNIX environment 

with LTO 

Implement and use the LTO library with VERITAS NetBackup DataCenter 


5.1 VERITAS NetBackup DataCenter overview 

VERITAS NetBackup DataCenter provides complete platform-independent data 

protection for the UNIX, Windows NT/2000, and NetWare enterprise 

environments. Global enterprises can access a single, easy-to-use central 

control site to assure continuity of all aspects of backup and recovery. With a 

flexible four-tier architecture and database-aware agents and options, VERITAS 

NetBackup DataCenter is fully scalable. 

VERITAS NetBackup DataCenter has a four-tier architecture that which allows it 

to scale to manage very large and complex environments. 

Master Server 

Media Server 

Client Agents 

The NetBackup DataCenter Master Server acts as the 

“brains” for actions such as scheduling and tracking 

client backups. It can have one or more tape 

devices/libraries attached for backing up data from 

multiple clients. The Master Server manages the catalog 

of backups. 

If organizations have data in several locations or have 

data-intensive applications such as data warehouses, 

they can implement Media Servers that locally back up 

large applications while backing up other clients over the 

network. A Media Server can share a tape library with 

the Master Server or another Media Server, or work with 

its own tape devices. If a Media Server fails, the attached 

client’s backups can be routed to another Media Server. 

Agents represent the third tier of the NetBackup 

DataCenter architecture and these back up servers and 

workstations. Normally, this tier represents the largest 

number of individual machines but not necessarily the 

most data. Both the Media Servers and clients can be 

centrally managed from the Master Server. 

Global Data Manager For organizations that centralize management of multiple 

Master Servers and/or widely distributed environments, 

a fourth tier is available: VERITAS Global Data Manager 

for NetBackup. Global Data Manager offers centralized 

management and control of all VERITAS NetBackup 

DataCenter storage domains in the enterprise. It also 

facilitates consistent policy management along with the 

ability to monitor storage domains anywhere in the 

enterprise. 


A storage domain consists of one Master Server and one or more Media Servers. 

In campus situations and where operations may be dispersed in multiple 

geographic locations, there may be more than one storage domain. One instance 

of Global Data Manager that controls multiple storage domains is called an 

enterprise domain. 

Shared Storage Option, or SSO is an optional feature available for VERITAS 

NetBackup DataCenter. It is a heterogeneous Storage Area Network (SAN) 

solution that allows individual tape drives (standalone or in a robotic tape library) 

to be shared dynamically between multiple VERITAS NetBackup DataCenter 

servers. The drives are connected to each host via enabling hardware, such as 

switches, hubs, or multiplexors. SSO allows better utilization of hardware 

resources by providing the ability for drives to be shared, rather than tying 

individual units to specific servers. 

More information is available at the VERITAS NetBackup DataCenter Web site, 

including supported hardware configurations, and all product manuals: 

http://www.veritas.com 

For brevity, we will now refer to the product as NetBackup DataCenter. 

5.2 Installing NetBackup DataCenter on Solaris 

In this section we show how to install NetBackup DataCenter for Solaris. 

5.2.1 Test environment 

We installed NetBackup DataCenter in this environment: 

► 

► 

► 

Sun Ultra 10 server, with Solaris operating system Version 2.8 and HVD SCSI 

adapter 

VERITAS NetBackup DataCenter Version 3.4 with patch NB_34_1 

3583 Ultrium Scalable Tape Library with two HVD drives 

This is a simple configuration where the Master Server, Media Server and client 

are all contained on the one system. 

Chapter 5. Implementing VERITAS NetBackup DataCenter with LTO 193

5.2.2 Device drivers 

NetBackup DataCenter does not use the IBM supplied LTO device drivers. If you 

have already installed and configured them, they should be deinstalled from the 

Solaris operating system using the procedure described in 2.3.5, “Removing 

IBMtape driver from the system (uninstallation)” on page 47. You will need to 

make sure that the correct drive entries are in the /kernel/drv/st.conf, if they were 

previously entered to the /kernel/drv/IBMtape.conf. 

NetBackup DataCenter uses native operating system drivers (for example, st for 

Solaris) and installs its own SCSI passththrough driver (sg) during the initial 

installation of the software. For more information on device driver configuration, 

refer to the manual VERITAS NetBackup 3.4 Media Manager Device 

Configuration Guide for UNIX, 100-01517. 

You need to add information to the file /kernel/drv/st.conf to allow the native 

device driver it to recognize the LTO library. Our modified file, including the tape 

configuration and target/LUN entries looked like this (note there are some 

redundant entries which could be removed): 

tape-config-list = 

“IBM ULTRIUM-TD1", "IBM Ultrium", "LTO_Ultrium", 

"IBM ULT3580-TD1", "IBM 3580 Ultrium", "LTO_Ultrium"; 

LTO_Ultrium = 1,0x24,0,0x45863d,2,0x00,0x01,0; 

name="st" class="scsi" target=0 lun=0; 

























After a reconfiguration reboot (reboot -- -r), the dmesg output included the 

following lines, indicating two 3580 drives detected at target 0, lun0 and target 1, 

lun0: 

Oct 8 11:42:49 mktg6 scsi: [ID 365881 kern.info] /pci@1f,0/pci@1/scsi@3/st@0,0 

(st14): 

Oct 8 11:42:49 mktg6 

Oct 8 11:42:49 mktg6 scsi: [ID 193665 kern.info] st14 at glm2: target 0 lun 0 

Oct 8 11:42:49 mktg6 genunix: [ID 936769 kern.info] st14 is 

/pci@1f,0/pci@1/scsi@3/st@0,0 

Oct 8 11:42:49 mktg6 scsi: [ID 365881 kern.info] /pci@1f,0/pci@1/scsi@3/st@1,0 

(st15): 

Oct 8 11:42:49 mktg6 

Oct 8 11:42:49 mktg6 scsi: [ID 193665 kern.info] st15 at glm2: target 1 lun 0 

5.2.3 Installation 

At this release of NetBackup DataCenter, V3.4, you need to install base code, 

and then a patch. 

Install base code 

Insert the NetBackup DataCenter CD in the CD-ROM drive and mount it. Change 

the directory to where the CD is mounted (for example, /cdrom/cdrom0) and start 

the installation script by entering the command: 

./install 

You will see the following output: 

Example 5-1 Initial NetBackup DataCenter installation menu 

VERITAS Installation Script 

Copyright 2000 VERITAS Software Corporation 

Installation Options 

1 NetBackup 

2 NetBackup Client Software 

q To quit from this script 

Choose an option [default: q]: 1 


Since this will be a Master Server, select option 1. The installation continues: 

Example 5-2 Select NetBackup DataCenter installation directory 

Processing package instance from 

 

VERITAS Media Manager 

(sparc) 3.4,REV=2000.06.23.16.09 

Copyright 1993 - 2000 VERITAS Software Corporation, All Rights Reserved 

Media Manager binaries will be installed in 

/opt/openv/volmgr 

and a link will be created from /usr/openv 

to /opt/openv, 

is this okay? (y) [y,n,?,q] y 

Selecting y will install in the default directory. 

Enter y to continue with the installation of the VERITAS Media Manager and the 

installation proceeds as shown in Example 5-3. 

Example 5-3 Install VERITAS Media Manager 










Installing VERITAS Media Manager as 

## Executing preinstall script. 

create /opt/openv 

Creating link from /usr/openv to /opt/openv. 


/opt/openv/man/share/man/man1/tpreq.1 

/opt/openv/man/share/man/man1/tpreq_bs.1 

.... omitted complete list of copied files ... 

/opt/openv/man/share/man/man1m/tldcd.1m 

/opt/openv/man/share/man/man1m/tlhcd.1m 


Copied files to /kernel/drv and to /kernel/drv/sparcv9. 


Doing add_drv of the sg driver 

Leaving existing sg configuration. 

If you wish to update the configuration you need to 

rm -f /kernel/drv/sg.conf 

and rerun /usr/openv/volmgr/bin/driver/sg.install. 

Media Manager configuration was successful. 


The Media Manager is now installed. You can see that the SCSI passthrough 

driver, sg was added. The installation automatically continues with client 

installation. Enter y when prompted to confirm loading of the Solaris client, then 

you can decide whether to load clients for any other operating system. 

Example 5-4 NetBackup DataCenter client installation 

Processing package instance from 

 

VERITAS NetBackup 

(sparc) 3.4,REV=2000.06.23.16.03 

Copyright 1993 - 2000 VERITAS Software Corporation, All Rights Reserved 

NetBackup 3.4GA will be installed. 

There is an existing link from /usr/openv to /opt/openv. 

NetBackup will be installed in 

/opt/openv/netbackup 

is this okay? (y) [y,n,?,q] y 

The Solaris clients will be loaded. 

Do you want to load any other NetBackup 

clients onto the server? (y) [y,n,?] y 

You will see a list of all the other clients available on the CD and can select 

individual clients or all of them as required. In our example, we chose to install all 

the clients. 

Example 5-5 Selecting NetBackup DataCenter clients for installation 

Choose the Platform Client types you wish to install 

by selecting the platform type one at a time 

or select ALL client platforms. 

Platform Client Options 

----------------------- 

1. DEC Alpha 

2. Auspex 

3. CRAY/J90 


4. DataGeneral 

5. HP9000 

6. INTEL (FreeBSD) 

7. Linux (RedHat) 

8. MACINTOSH (MacOSXS) 

9. NCR 

10. Pyramid 

11. IBM RS6000 

12. SCO 

13. Sequent 

14. SGI 

15. Solaris 

16. ALL client platforms 

q Quit client selection 

Enter Choice (Default is 16) [1-16,?,q] 16 

You have chosen to install: 

All client platforms 

Is this the list you wish to use? (y) [y,n,?] y 

The client installation proceeds as shown: 

Example 5-6 Installing NetBackup DataCenter clients 











Installing VERITAS NetBackup as 

## Executing preinstall script. 


/opt/openv/NB-Java.install 

/opt/openv/NB-Java.tar.Z 



opt/openv/netbackup/client/Linux/RedHat/xbp 



The NetBackup binaries are installed. 

Read the release notes carefully. 

If you ran pkgadd instead of ./install, you must run 

/usr/openv/netbackup/bin/install_bp 

to initiate NetBackup. 

Installation of was successful 

Next, the installation asks for the license keys. These are supplied with your 

purchased product. 

Example 5-7 Add license keys 

Running /usr/openv/netbackup/bin/install_bp 

A NetBackup BusinesServer or DataCenter license key is needed 

for installation to continue. 

Enter license key: xxxxxxxxxx 

Evaluation NetBackup DataCenter with Dec 8, 2001 expiration date will be 

registered. 

Is this OK? (y/n) [y] y 

xxxxxxxxxx 

Evaluation NetBackup DataCenter Base product with all the features enabled 

has been registered. 

Now the installation script asks about our configuration. In this simple example, 

we have just one server in the domain. Therefore, we answer yes to make this 

system (MKTG6) the master server and indicate that there are no other media 

servers. 

Example 5-8 Install master server 

Installing NetBackup DataCenter version: 3.4GA 

Is mktg6 the master server? (y/n) [y] 

Do you have any media (slave) servers? (y/n) [n] 

Checking for a bpcd entry in /etc/inetd.conf:Found. 

Checking for a vopied entry in /etc/inetd.conf:Found. 


Checking for a bpjava-msvc entry in /etc/inetd.conf:Found. 

Checking /etc/services for the needed NetBackup and Media Manager services. 

Found the necessary NetBackup services in /etc/services. 

To make NetBackup and Media Manager start up automatically when the system is 

restarted, the S77netbackup script found in /usr/openv/netbackup/bin/goodies 

has been placed in the /etc/rc2.d directory. 

NetBackup shutdown command found in /etc/rc0.d. 

Extracting NetBackup-Java product files ... 

Installing NetBackup-Java, and the Java runtime environment for SunOS 5.8. 

Configuring NetBackup-Java runtime environment. 

The installation of the NetBackup-Java product on mktg6 is complete. 

Finally, the installation requests some more configuration information. We select 

to have this host store the global device information and to start the NetBackup 

DataCenter processes immediately. We will not create example class and 

schedule definitions (these will be created later), and will take the default 

database indexing options. More information on these and other installation 

options is available in the product documentation and beyond the scope of this 

redbook. Finally, select q to finish the installation. 

Example 5-9 Completion of NetBackup DataCenter install 

In order for device discovery and auto-configuration to work properly in a 

DataCenter environment, particularly where peripherals are connected to 

multiple servers, one host must serve as the repository for global device 

configuration information. 

Enter which host will store global device information. 

(default: mktg6): 

To be able to install the client software the NetBackup 

processes need to be started. Do you want to start the 

NetBackup processes so client software can be installed? (y/n) [y] 

Starting the NetBackup database manager process (bpdbm). 

Do you want to create class and schedule examples that you can view or use 

when you are configuring your own classes and schedules? (y/n) [y]n 

Do you want to start the NetBackup bprd process so 

backups and restores can be initiated? (y/n) [y] 

Starting the NetBackup request daemon process (bprd). 

Client database indexing reduces the search time when restoring 

client files, but it takes about 2% more disk space. 

Do you want to index the client database files? (y/n) [y] 

The default index level is 9 levels. Use the default? (y/n) [y] 

Running index_clients process in background mode. 

Output from the process will be written to /tmp/index_clients.output. 


Do you want to add additional license keys now? (y/n) [y]n 

Use /usr/openv/netbackup/bin/admincmd/get_license_key 

to add, delete or list license keys at a later time. 

VERITAS Installation Script 

Copyright 2000 VERITAS Software Corporation 

Installation Options 

1 NetBackup 

2 NetBackup Client Software 

q To quit from this script 

Choose an option [default: q]: q 

A trace of the install can be found in /tmp/install_trace.3384 

That file can be deleted after you are sure the install was successful. 

Install patch 

You now need to install a patch, NB_34_1 for supporting the IBM LTO drives. You 

can download this patch from: 

http://support.veritas.com 

In our example, the patch was downloaded to /var/sadm/VRTS. Install the patch 

as shown in the example. 

Example 5-10 Install patch 

mktg6# cd /var/sadmVRTS 

mktg6# ls -l 

total 91952 

-rw-r--r-- 1 root other 72922 Sep 18 18:25 patchNB_34_1.README 

-rw-r--r-- 1 root other 32081920 Sep 18 18:18 patchNB_34_1.solaris.tar 

-rw-r--r-- 1 root other 14841539 Sep 18 18:31 

patchNB_34_1.solaris.tar.Z 

drwxr-xr-x 3 root other 512 Sep 18 18:21 usr 

-rwxr-xr-x 1 root other 39683 Sep 18 18:24 Vrts_patch.install 

mktg6# ./Vrts_patch.install 

There is 1 patch available in /var/sadm/VRTS: 

(* denotes installed patch) 

NB_34_1 

Enter patch name (or q): NB_34_1 

Install patch NB_34_1 Mon Oct 8 13:25:16 PDT 2001 Rev. 1.40 


Saving existing binaries before patch /var/sadm/VRTS/patchNB_34_1.solaris.tar.Z 

updates them. 

Copy /usr/openv/netbackup/bin/admincmd/bpauthsync to 

/usr/openv/patch/NB_34_1/save/bpauthsync.b4patchNB_34_1 


Copy /usr/openv/lib/client/Solaris/Solaris2.6/libvafs.so to 

/usr/openv/patch/NB_34_1/save/libvafs.so.b4patchNB_34_1 

Extract binaries /var/sadm/VRTS/patchNB_34_1.solaris.tar.Z 

x usr/openv/netbackup/bin/admincmd/bpauthsync, 831416 bytes, 1624 tape blocks 

x usr/openv/netbackup/bin/admincmd/bpdbjobs, 945992 bytes, 1848 tape blocks 

.... omitted complete list of extracted files ... 

x usr/openv/lib/client/RS6000/AIX4.2/libvafs.sl, 2106 bytes, 5 tape blocks 

x usr/openv/lib/client/Solaris/Solaris2.6/libvafs.so, 2448 bytes, 5 tape blocks 

Installation of patch NB_34_1 completed Mon Oct 8 13:25:16 PDT 2001 Rev. 

1.40. 

There is 1 patch available in /var/sadm/VRTS: 

(* denotes installed patch) 

NB_34_1 * 

Enter patch name (or q): q 

Exiting ./Vrts_patch.install 

The initial installation is now complete. 

Check daemons have started 

Check that NetBackup DataCenter has started correctly by using the bpps 

command. 

mktg6# cd /usr/openv/netbackup/bin 

mktg6# ./bpps -a 

NB Processes 

------------ 

root 3769 1 0 13:24:19 ? 0:00 /usr/openv/netbackup/bin/bprd 

root 3721 1 0 13:24:03 ? 0:00 /usr/openv/netbackup/bin/bpdbm 

MM Processes 

------------ 

root 3847 1 0 13:24:35 ? 0:00 /usr/openv/volmgr/bin/vmd 


All three daemons should be running, including bprd, the request daemon, 

bpdbm, database manager and vmd, volume manager daemon. 

Starting and stopping NetBackup DataCenter daemons 

The installation procedure should insert the startup script, s77netbackup in the 

startup directory /etc/rc2.d. If you need to start NetBackup DataCenter manually 

at any time, you can re-execute the script from the 

/usr/openv/netbackup/bin/goodies directory. Alternatively, you can issue the 

following two commands: 

/usr/openv/volmgr/bin/ltid 

/usr/openv/netbackup/bin/initbprd 

To stop NetBackup DataCenter, use the command: 

/usr/openv/netbackup/bin/goodies/bp.kill_all 

5.3 Configuration 

Configuring NetBackup DataCenter involves detecting and defining the devices, 

configuring storage units and volume groups, and defining your backup policy. 

5.3.1 Detect tape drives 

First, check that the devices have been correctly configured to the operating 

system. The ls command in the /dev/rmt directory will list all the device special 

files available for the two defined drives. The drives are at target 0, lun0 and 

target 1, lun 0, respectively. 

mktg6# cd /dev/rmt 

mktg6# ls -l 

total 96 

lrwxrwxrwx 1 root root 43 Oct 8 11:43 0 -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0: 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 0b -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:b 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0bn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:bn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 0c -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:c 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0cb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:cb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 0cbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:cbn 


lrwxrwxrwx 1 root root 45 Oct 8 11:43 0cn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:cn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 0h -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:h 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0hb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:hb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 0hbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:hbn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0hn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:hn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 0l -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:l 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0lb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:lb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 0lbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:lbn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0ln -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:ln 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 0m -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:m 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0mb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:mb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 0mbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:mbn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0mn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:mn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 0n -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:n 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 0u -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:u 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0ub -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:ub 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 0ubn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:ubn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 0un -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@0,0:un 

lrwxrwxrwx 1 root root 43 Oct 8 11:43 1 -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0: 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 1b -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:b 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1bn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:bn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 1c -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:c 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1cb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:cb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 1cbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:cbn 


lrwxrwxrwx 1 root root 45 Oct 8 11:43 1cn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:cn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 1h -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:h 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1hb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:hb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 1hbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:hbn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1hn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:hn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 1l -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:l 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1lb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:lb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 1lbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:lbn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1ln -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:ln 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 1m -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:m 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1mb -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:mb 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 1mbn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:mbn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1mn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:mn 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 1n -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:n 

lrwxrwxrwx 1 root root 44 Oct 8 11:43 1u -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:u 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1ub -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:ub 

lrwxrwxrwx 1 root root 46 Oct 8 11:43 1ubn -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:ubn 

lrwxrwxrwx 1 root root 45 Oct 8 11:43 1un -> 

../../devices/pci@1f,0/pci@1/scsi@3/st@1,0:un 

Each special file has the format [0-127][l,m,h,u,c][b][n]. The first field, [0-127] is 

the device number which is assigned by the operating system in order of 

detection. The field [l,m,h,u,c] specifies the density (low, medium, high, ultra, 

compressed). The b flag specifies the optional BSD behavior and n indicates a 

no rewind device. VERITAS requires you to use a BSD, no-rewind device. 


5.3.2 Scan for tape devices 

Once we have verified that the tape drives are available, the VERITAS command 

/usr/openv/netbackup/bin/sgscan looks for installed tape devices and robots 

which are visible to it. 

mktg6# cd /usr/openv/netbackup/bin 

mktg6# ./sgscan 

/dev/sg/c0t0l0: Tape (/dev/rmt/0): "IBM ULT3580-TD1" 

/dev/sg/c0t1l0: Tape (/dev/rmt/1): "IBM ULT3580-TD1" 

/dev/sg/c0t2l0: Changer: "IBM ULT3583-TL" 

Our two drives and medium changer have been correctly identified. Now we 

begin the device configuration. 

5.3.3 Device configuration wizard 

For the rest of the configuration, we used the graphical interface for NetBackup 

DataCenter. This could be at the Solaris system where NetBackup DataCenter 

was installed, using the CDE interface, or remotely using the Java System 

Administration client. 

When you start the System Administration client, you first need to specify the 

NetBackup DataCenter host to connect to, and an operating system userid and 

password as shown in Example 5-1. Our host is called MKTG6. 


Figure 5-1 System administration login 

Because it is the first time we have run this interface, the NetBackup DataCenter 

Assistant displays as in Figure 5-2. 


Figure 5-2 NetBackup DataCenter assistant 

Close this window and the regular console displays (Figure 5-3). 


Figure 5-3 System administration main window 

Select Media and Device management and on the displayed window, select the 

line Media Manager for your host (mktg6) on the left hand side. You will notice 

there is no media available in the right hand side panels. See Figure 5-4. 

Figure 5-4 Media and device management 


Click the magic wand button (sixth icon from the left) to start the device 

configuration wizard. Click Next on the initial display. 

The Device Hosts window is shown in Figure 5-5. This will initiate automatic 

discovery and configuration of devices on the selected hosts. Click Add and enter 

your hostname (MKTG6). Click Next to start the scan. 

Figure 5-5 Select hosts to scan 

The scan shows up two attached drives as expected (Figure 5-6). 

Figure 5-6 Display detected drives 


To view the detected devices, click Next. The Backup devices will display, as 

expected, two LTO drives and one medium changer are shown, as in Figure 5-7. 

The NetBackup DataCenter software has correctly identified the device types. 

Figure 5-7 Display discovered devices configuration 

The State should show as “Configured” to indicate the device has been correctly 

identified. Since the device configuration is correct, click Next. NetBackup 

DataCenter then reinitializes its daemons and internally configures the devices 

(Figure 5-8). 


Figure 5-8 Updating device configuration 

Finally, make the configured devices available to NetBackup DataCenter by 

defining storage units for them. A storage unit can later be used by a backup 

policy as a destination device for backed up data. 

Make sure the robot line is checked in Figure 5-9. Click Next. This will create a 

default configuration for the storage unit which will view in the next section. This 

completes the device configuration wizard. 

Figure 5-9 Configure storage units 


5.3.4 Importing media 

Now, return to the administration window (Figure 5-10), and expand the entries 

for Media Manager mktg6, Device Hosts, mktg6, Robot0 - TLD, and Tape Drives. 

You will now see the two newly configured drives listed. 

Figure 5-10 Show defined drives 

Right click on the Robot0 line and select Inventory Robot. From the display 

shown in Figure 5-11, select the device host (MKTG6) and robot TLD(0) and 

make sure the option Show contents of robot is active. Click Start. 


Figure 5-11 Setup robot inventory 

This process scans each cell in the library to determine which slots contain 

tapes. The barcode reader of the 3583 Ultrium Scalable Tape Library is used to 

read the labels. 

Monitor the Results part of the screen for progress as shown in Example 5-11. 

Example 5-11 Robot inventory results 

"Show Contents of Robot" 

-------------------------------------------------------------------- 

10/08/2001 14:06:37 

Robot: TLD(0) on mktg6 

Operation: Inventory 

-------------------------------------------------------------------- 

Robot Contents 

SlotTape Barcode 

====================== 

1 No 

2 No 

3 Yes 159AATL1 



6 Yes 448ACOL1 


7 No 

8 No 

9 No 

10 No 

11 No 


13 No 

14 No 

15 No 

16 No 

17 No 


19 No 

20 No 

21 No 

22 No 

23 No 

24 No 

You should check the output to see that the cartridges detected match what you 

believe is physically present in the library. You should also make sure that all 

cartridges found are available for use by NetBackup DataCenter. If there are 

cartridges in the library which contain data from or belong to another application, 

you should remove them and re-run the inventory operation to prevent them from 

being recognized and possibly over-written by NetBackup DataCenter. 

Now, select Perform volume configuration update. If you have loaded 

additional tapes in the I/O slot to be added into the library, you can also select the 

option Empty inport prior to update. This will load those tapes into the robot 

and add them to the inventory. Alternatively, if you have ejected tapes from the 

robot, remove them if this option is selected to avoid them being moved back into 

the robot. 

The volume configuration update operation will load the discovered inventory into 

the NetBackup DataCenter database (all the tapes in the robot). Click Start — 

the inventory will be loaded and the output displayed in the Results panel. Click 

Close when you have finished viewing the results. 

Example 5-12 Volume configuration update results 

"Perform volume configuration update" 

-------------------------------------------------------------------- 

10/08/2001 14:12:25 

Robot: TLD(0) on mktg6 

Operation: Inventory and Update 

Volume Database Host: mktg6 

-------------------------------------------------------------------- 


Generating list of recommended changes ... 

Proposed Change(s) to Update the Volume Configuration 

===================================================== 

Logically add new media 9AATL1 (barcode 159AATL1) to robot slot 3. 



Logically add new media 8ACOL1 (barcode 448ACOL1) to robot slot 6. 



Updating volume configuration ... 

Processing new media added to the robotic library by logically 

adding media with new media IDs as follows... 

Media IDSlot 

============ 

9AATL1 3 

6AATL1 4 

4AATL1 5 

8ACOL1 6 

8AATL1 12 

7AATL1 18 

Volume configuration successfully updated. 

-------------------------------------------------------------------- 

In the main panel we now see the configured volumes listed for the MKTG6 

Media Manager Host as in Figure 5-12. 

Figure 5-12 Media Manager with configured volumes 


Here you can see the default configuration which has been set up by the device 

configuration wizard. A default volume group called 00_000_TLD was defined, 

along with a pool of NetBackup DataCenter. The inventoried media volumes 

were imported into the default pool. You can define additional pools and assign 

these or new media to them later if required. The Media Type, HCART has been 

automatically assigned as this is what is used for LTO cartridges. 

5.3.5 Storage units 

Exit from the GUI (File -> Exit). This returns you to the to the Global Data 

Manager Main Window as shown in Figure 5-3 on page 209. Click on Storage 

Unit Management to show the details of the automatic setup performed by the 

device configuration wizard. Select mktg6, (first line), and expand the lines mktg6 

Media Server and All Storage Units. This will display all the created storage units. 

A storage unit is a group of one or more devices of a specific type and density 

attached to a server. NetBackup DataCenter uses storage units to store backups. 

In our configuration (Figure 5-13) we have one storage unit called 

mktg6-hcart-robot-tld-0. 

Figure 5-13 Storage unit configuration 

Double click the storage unit to display the details as in Figure 5-14. On this 

panel you can enable multiplexing by setting Maximum multiplexing per drive 

to a number greater than one. When multiplexing is enabled, NetBackup 

DataCenter sends concurrent multiple backups from one or more clients to a 

single drive and multiplexes the backups onto the media. 


Figure 5-14 Storage unit properties 

5.3.6 Define backup policy 

Exit from the storage unit display and return to the main window (Figure 5-3 on 

page 209). Select Backup Policy Management. Select mktg6 Master Server to 

connect. To add a new backup class, select the laser icon as shown in 

Figure 5-15. 

Figure 5-15 Add a new backup class 


Choose a name for the class which reflects the purpose of the class. In our 

sample scenario, the class will be used to backup the system MKTG6 so we call 

the class MKTG_FULL as in Figure 5-16. 

Figure 5-16 Enter backup class name 

Click Next with the Use add class wizard box checked to start the Backup Policy 

Configuration Wizard. On the first window (Figure 5-17), indicate the class type 

— because ours is a simple Solaris operating system backup, select Standard. 

Click Next to continue. 

Figure 5-17 Backup policy configuration wizard 


Now we select the client names to be included in the class. Click the Add button. 

Enter in the host names (in our case MKTG6) and select the operating system 

(Solaris 8) as in Figure 5-18. 

Figure 5-18 Select clients and OS for backup class 

Add more clients as needed and click Close when finished. The client list is now 

displayed, as in Figure 5-19. Click Next. 


Figure 5-19 Display client list 

Now, specify which directories or paths will be backed up. You can choose to 

backup all local drives or explicitly specify the directories to backup using the 

Add button as shown in Figure 5-20. Click Next. 


Figure 5-20 Configure backup directories 

Now specify the type(s) of backup which will be performed. The choices are 

shown in Figure 5-21. If you select just full backup, then every backup will be a 

complete unconditional backup of the paths selected previously. If you select full 

backup you can also choose to perform incremental backups (either cumulative 

or differential). Click Next when you have specified the type of backup to perform. 


Figure 5-21 Configure backup type 

Next you specify the rotation schedule which defines when backups will be 

performed. Customize the window (Figure 5-22) as required to indicate how often 

full and incremental backups should be performed. Here you also indicate for 

how long backups should be kept. Click Next. 


Figure 5-22 Configure schedule rotation 

Now, define the schedule timings as in Figure 5-23. The wizard uses large time 

periods for when to start schedules, however more exact timings can be defined 

later if necessary. Click Next. 


Figure 5-23 Define backup start windows 

Click Finish to complete the wizard. 

The newly defined class is displayed in Figure 5-24. 

Figure 5-24 Display backup class 


To make further customizations, double click the class (MKTG_FULL). You can 

change the attributes shown in Figure 5-25. For storage unit, the default is Any 

Available, but if you have more than one unit defined, here you could select to 

use specific storage units. You can see that the NetBackup DataCenter 

application has assigned our default volume pool NetBackup to this Class — 

again you could override this to use any additionally created volume pools. If you 

selected Enable Multiplexing when defining the storage unit (see 5.3.5, “Storage 

units” on page 217), you also need to select Allow multiple data streams if you 

want the clients in this particular class to multiplex. You can also enable client 

compression in this panel. This is mainly used when the clients are attached via 

a slow communications link. 

Figure 5-25 Backup class attributes 

Also from Figure 5-24 you can customize the schedules (for example, to define a 

more specific start time), the files or filesystems to include for backup and add or 

change clients in the class. 

You have now completed the basic setup of the NetBackup DataCenter with the 

LTO library. 


5.4 Testing the configuration 

To test our configuration, right click on the class (MKTG_FULL) and select 

Manual Backup as shown in Figure 5-26. 

. 

Figure 5-26 Initiate client backup 

Next you can specify the particular client(s) and backup type which you want to 

perform (Figure 5-27). We will perform a full backup on our client MKTG6. Click 

OK to start the backup. 


Figure 5-27 Configure client backup 

The message informs you that the backup has started. Select Activity Monitor 

from the main window (Figure 5-3 on page 209) to monitor the progress. Click on 

the job to display detailed status (Figure 5-28). You can see that NetBackup 

DataCenter selects an empty cartridge (9AATL1) from the pool to write the 

backup on and mounts it in the drive. 


Figure 5-28 Monitor backup job details 

When the job has completed the status indicates success as in Figure 5-29. The 

tape cartridge is dismounted from the drive. 


Figure 5-29 End of backup job status 

Now we know that the LTO drive and robot is configured correctly to NetBackup 

DataCenter. To restore files, select Backup, Archive and Restore from the main 

window (Figure 5-3 on page 209). Select the files you want to restore from the 

file tree as shown in Figure 5-30. 


Figure 5-30 Restoring files 

The tape volume is mounted and the files are copied back to the client. 

5.5 Miscellaneous tape topics 

Finally, we discuss two topics related to VERITAS NetBackup DataCenter — 

using compression and labeling cartridges. 

5.5.1 Compression 

The configuration wizard by default sets the drive to use hardware compression, 

and this is normally recommended. If you want to disable compression, select 

Media and Device Management from the main window (Figure 5-3 on 

page 209). Expand the entries as shown in Figure 5-31 to display the drives 

located under Tape Drives. 


Figure 5-31 Display drives 

Right click a drive and select Change to show the attributes of that drives as in 

Figure 5-32. To disable compression, you would indicate a different special file 

name in the No rewind device field, for example, /dev/rmt/0bn. You should 

always use a special file with No Rewind and BSD behavior (as described in 

5.3.1, “Detect tape drives” on page 203). This allows NetBackup DataCenter to 

control the tape positioning, rather than the device driver. 


Figure 5-32 Change drive attributes 

Note the Robot Drive Number — this reflects the physical installed position of 

the drive in the robot and has been determined by the configuration wizard. 

When Solaris does a boot reconfigure, devices are configured in LUN order. We 

recommend that you configure the library so that the drive in first robot position 

has the lowest target ID and therefore is /dev/rmt0x. You can view the 

configuration of the library and drive with the NetBackup DataCenter provided 

utilities, tldtest and robtest, which are installed in /opt/openv/volmgr/bin. This 

utilities provide similar functions to tapeutil, which is provided with the 

IBM-provided device drivers. Consult the NetBackup DataCenter documentation 

for more information on how to use these commands. 

5.5.2 Cartridge labeling 

The Media Manager uses the last six digits of the barcode label for generating 

media IDs. This can potentially cause problems with an LTO library, since LTO 

barcodes have the format XXXXXXL1. Media Manager will drop the first two 

digits of the barcode label, therefore if there are two cartridges with barcode 

labels 100000L1 and 200000L1 they would both show up with media IDs of 

0000L1. 


To avoid this issue, you can only use media where the last six digits have unique 

numbers to avoid conflicts. Alternatively, you can use a new feature available for 

NetBackup DataCenter which allows you to define your own rules for generating 

media IDs. 

For more information on how to do this, look at VERITAS TechNote 236271, 

which can be viewed at: 

http://support.veritas.com/pub/support/Products 


6 

Chapter 6. 

Implementing Legato 

NetWorker with LTO 

In this chapter we give you detailed information on the implementation of the 

Legato NetWorker product with LTO drives and libraries. This includes: 

► 

► 

► 

► 

Product overview 

Installation and library definition 




Implementation and use 




Product exploitation 

– LTO library sharing 

– LAN free backup 



► 

► 

► 

Install the Legato NetWorker product in a UNIX environment with LTO 

Implement and use the LTO library with Legato NetWorker 

Exploit Legato NetWorker capabilities with LTO 


6.1 Legato NetWorker overview 

The Legato NetWorker product is a suite of storage management software that 

provides backup, recovery, and other services to computers, with a wide variety 

of operating systems and data types. NetWorker products for different operating 

systems are interoperable, which provides you with flexibility in designing a 

storage management setup that works best with your current computing 

environment. For detailed product information, visit the Legato Web site: 

http://www.legato.com 

The NetWorker product has four major components: the NetWorker server, 

client, storage node, and the Legato License Manager. 

NetWorker server 

NetWorker client 

NetWorker storage node 

The NetWorker server software provides control and 

scheduling for NetWorker operations. On the 

NetWorker server, you can enter the enabler 

licenses for the NetWorker server and all the 

functions the NetWorker server controls, such as 

autochanger modules, additional client connections, 

and ClientPak licenses. You can also define the 

clients, devices, and media that the NetWorker 

server controls, define the schedules for backups 

and other operations, and monitor all the NetWorker 

operations. 

The NetWorker client software provides on-demand 

backup and recovery functionality, communicating 

with the NetWorker server. You install the NetWorker 

client software on all computers which need to back 

up to the NetWorker server. 

You can back up data directly to devices attached to 

either a NetWorker server or to a NetWorker storage 

node. A storage node controls storage devices such 

as tape drivers, autochangers, and silos. 

Legato License Manager The Legato License Manager provides a central 

place for managing the licenses of all your 

NetWorker clients and servers. Instead of managing 

licenses separately, you can maintain all your 

NetWorker licenses from a single computer. 

Installation of the License Manager is an option 

when installing the NetWorker software. 


6.2 Legato NetWorker installation 

Now we give you examples of the Legato NetWorker installation for UNIX 

platforms, such IBM AIX, SUN Solaris and HP-UX. For more details, refer to the 

official Legato documentation that can be found at this link: 

http://www.legato.com/support/documentation/ 

In these examples, we are documenting the installation of Legato NetWorker in a 

Storage Area Network environment for SUN Solaris environment, with Fibre 

Channel HBA on the server, and with a SCSI point to point connection for AIX 

and HP-UX environments. 

In each installation example, we are using a 3583 Ultrium Scalable Tape Library 

with two SCSI-LVD drives, connected to an IBM 2108 SAN Data Gateway for the 

SAN environment. 

6.3 Installation in an AIX environment 

For a detailed explanation of the installation process, refer to Legato NetWorker 

Installation Guide - AIX Version. This book can be downloaded in PDF format 

from the Legato Web site: 

http://web1.legato.com/cgi-bin/catalog?sf=Releases&level=10-1 

Be sure that Legato NetWorker is supported in your specific environment. To 

verify this, cross check these IBM and Legato Web sites: 




http://www.storage.ibm.com/hardsoft/tape/conntrix/isv358html 

http://www.legato.com/products/compatibility/networker.cfm 

In the following example, we install Legato NetWorker in the environment: 

► IBM RISC/6000 F50 server with IBM AIX operating system release 4.3.3 and 

integrated LVD SCSI adapter 

► Legato NetWorker 6.1 Build 186 

► 

3583 Ultrium Scalable Tape Library with LVD drives 

Now we detail the installation process for the above environment, with the 

required steps to configure the 3583 Ultrium Scalable Tape Library. 

Chapter 6. Implementing Legato NetWorker with LTO 237

6.3.1 Software installation (AIX) 

We assume that you are installing the Legato NetWorker software from a 

directory on the AIX server. These are the required steps: 

1. First expand the compressed file using the following command: 

# gunzip -s networker_aix.tar.gz 

This process replaces the original file with an uncompressed file of the same 

name, without the .gz last level qualifier. 

2. Extract the uncompressed file to the same directory with the following 

command: 

# tar -xvpf networker_aix.tar 

This operation expands this file in many other files and directories that make 

up the installation package. 

3. If not already logged in, log in as root, then change to the directory where the 

Legato NetWorker installation files are, and enter the following command at 

the shell prompt: 

# ./nsr_ize -i -s 

You are prompted with the following messages: 

NetWorker(TM) - Release 6.1.Build.186 

Copyright (c) 1990-2001, Legato Systems, Inc. All rights reserved. 

This product includes software developed by the University of 

California, Berkeley and its contributors. 

nsr_ize is about to install NetWorker server software on machine`brazil'. 

Install the NetWorker man pages [yes]? n 

4. Answer yes/no depending if you want the Legato manuals to be installed. 

Then: 

Install the aix43n NetWorker programs [yes]? 

Directory where aix43n NetWorker programs should be installed [/bin]? 

Directory to use for client, licensing and server information [/nsr]? 

5. Press Enter to all the questions. Then configure the tape devices: 

The following is a list of the tape devices available on your system: 



smc0 Available 30-58-00-6,0 IBM 3583 Library Medium Changer 

The name needed by NetWorker is usually `/dev/' plus the name of the device 

as given the first column of the above table, followed by `.1'to indicate a 

no-rewind device name. 


Enter the tape or disk device(s) that are going to be used by the NetWorker 

server. Use the no-rewind name for each tape device (i.e., use /dev/rmt0.1 

instead of /dev/rmt0). If you do not choose a device a default device will 

be created for you. 

Enter device name ([Return] if no more): /dev/rmt0.1 

Answer exactly as indicated in the example (add .1 to the device name, that 

means it is a non rewinding device). 

Important: You can only use non rewinding devices with the NetWorker 

server. If you use a rewinding device, the read/write head is repositioned at the 

beginning of the volume, and the previously backed-up data is overwritten. 

You should receive this message: 

Device rmt0.1 is of NetWorker type: 3580 

6. Repeat this operation for every LTO tape device that you want to configure to 

NetWorker. Press Enter when you have no more devices to add. Then the 

prompt asks: 

Start NetWorker daemons at end of install [yes]? 

Install the NetWorker device drivers (required for storage nodes and 

servers) [yes]? 

7. Press Enter to both questions, since you need to install the NetWorker device 

drivers to be able to use the SCSI medium changer. At the end of the 

installation, you should see the following message: 

NetWorker successfully installed on `brazil'! 

8. Check if the daemons are running with the command: 

# ps -ef |grep -E “nsr|lgto” 

The command output should look like this, depending on the packages 

installed: 

root 4484 11512 0 15:38:07 - 0:00 /usr/bin/nsrmmdbd 

root 10902 11512 0 15:38:13 - 0:00 /usr/bin/nsrmmd -n 2 

root 11152 12044 0 15:37:57 - 0:00 /bin/nsrexecd 

root 11512 1 0 15:38:01 - 0:01 /bin/nsrd 

root 12044 1 0 15:37:57 - 0:00 /bin/nsrexecd 

root 12586 11512 0 15:38:09 - 0:00 /usr/bin/nsrindexd 

root 15524 18328 0 15:42:26 pts/1 0:00 grep -E nsr|lgto 

root 17202 11512 0 15:38:11 - 0:00 /usr/bin/nsrmmd -n 1 


oot 19064 1 0 15:37:57 - 0:00 /bin/lgtolmd -p /nsr/lic -n 1 

If you do not see the processes, start NetWorker manually with the command: 

# /etc/rc.nsr start 

In Table 6-1 you see the daemons required for each NetWorker component. 

Table 6-1 Legato NetWorker daemons for AIX 

NetWorker installation packages NetWorker daemons 




nsrd, nsrexecd, nsrindexd, nsrmmdbd, 

nsrmmd 

nsrexecd 

nsrexecd, nsrmmd 

Before using the Legato NetWorker with the LTO library (if you are using a 3581 

Autoloader drive, a 3583 Ultrium Scalable Tape Library or a 3584 Ultrium 

UltraScalable Tape Library), refer to 6.7.1, “Autochanger configuration (AIX)” on 

page 250. 

6.4 Installation in a Solaris environment 


Installation Guide - Solaris Version, this book can be downloaded in PDF format 










In the following example, we install Legato NetWorker in the environment: 

► SUN Microsystems Enterprise250 SPARC server with Solaris operating 

system release 2.7 and QLogic 2200 HBA 


► 





6.4.1 Software installation (Solaris) 


directory on the Solaris server. These are the required steps: 


# gunzip networker_sol.tar.gz 


name, without the .gz last level qualifier 


command: 

# tar -xvpBf networker_sol.tar 

This operation expands this file in many other files and directories that make 

up the installation package. You should find the following directories: 

# ls -l 

total 184478 

drwxr-xr-x 5 root other 512 Jun 22 05:42 LGTOclnt 

drwxr-xr-x 4 root other 512 Jun 22 05:41 LGTOdrvr 

drwxr-xr-x 5 root other 512 Jun 22 05:42 LGTOlicm 

drwxr-xr-x 4 root other 512 Jun 22 05:42 LGTOman 

drwxr-xr-x 4 root other 512 Jun 22 05:42 LGTOnode 

drwxr-xr-x 4 root other 512 Jun 22 05:42 LGTOserv 

drwxr-xr-x 4 root other 512 Jun 22 06:30 xdsmhsm 

3. Before continuing the installation, it is a safe practice to make a copy of your 

current configuration. Use the following commands: 

# cp /etc/rpc /etc/rpc/rpc.old 

# cp /etc/syslog.conf /etc/syslog.conf.old 

4. If you were not logged in as root, log in now, then change to the directory 

where the Legato NetWorker installation files are, and enter the following 

command: 

# pkgadd -d /opt/legato 

You are prompted with the following messages: 


1 LGTOclnt NetWorker for Solaris (Backup/Recover) Client (sparc) 

6.1.Build.186 

2 LGTOdrvr NetWorker for Solaris (Backup/Recover) Device Drivers 

(sparc) 6.1.Build.186 


3 LGTOlicm NetWorker for Solaris (Backup/Recover) Licensing Manager 


4 LGTOman NetWorker for Solaris (Backup/Recover) Man (sparc) 

6.1.Build.186 

5 LGTOnode NetWorker for Solaris (Backup/Recover) Storage Node 


6 LGTOserv NetWorker for Solaris (Backup/Recover) Server (sparc) 

6.1.Build.186 


(default: all) [?,??,q]: 

5. Because you are installing the server package, we recommended that you 

select all the packages at the same time. Press Enter. 


NetWorker for Solaris (Backup/Recover) Client(sparc) 6.1.Build.186 

Legato Systems, Inc. NetWorker(TM) - Release 6.1.Build.186 

Copyright (c) 1990-2001, Legato Systems, Inc. All rights reserved. 

This product includes software developed by the University of California, 

Berkeley and its contributors. 

To set up a NetWorker storage node, licensing manager or client, you need 

to supply a directory for the nsrexecd state file. 

Below is a list of some of the filesystems, with their free space, which 

you might consider: 

/opt : 690707 

/usr : 448510 

/ : 345166 

/var : 52042 

Directory to use for client, licensing and server information [/opt/nsr]? 

6. Press Enter again to accept the default, or select another directory for 

installation. 

The nsrexecd program restricts access to a select set of NetWorker servers. 

Please enter the names of each computer running a NetWorker server that 

will back up this computer, one name at a time. If a computer has more 

than one network interface, please enter each interface's name (one at a 

time). 

Enter the first NetWorker server's name [no more]: sol-e.almaden.ibm.com 

Enter the next NetWorker server's name [no more]: 

7. Enter the server name/names that you want to use with NetWorker. Press 

Enter again when no more servers have to be added. Then at the following 

prompt: 


Start NetWorker daemons at end of install [yes]? no 

Answer no, because after this single package installation, the process 

continues with the other packages, and if you start the daemons now then you 

have to stop again the daemons later during the procedure. Then the 

installation begins: 






## Checking for setuid/setgid programs 




Installing NetWorker for Solaris (Backup/Recover) Client as 



Completing Installation 

Starting NetWorker daemons 

NetWorker successfully installed on `sol-e'! 


8. After the last message, if the installation has been successful, the procedure 

continues with the following package, with the same questions. Check the 

following messages during the installation process: 





9. The next (and last) package to install, is the server package. You are 

prompted with the following question: 

Enter the tape or disk device(s) that are going to be used by the NetWorker 

server. Use the no-rewind, BSD-semantics name for each tape device (i.e., 

use /dev/rmt/0mbn instead of /dev/rmt/0mb). If you do not choose a device 

a default device will be created for you. 

Enter device name ([Return] if no more): /dev/rmt/1stcbn 

10.Use the device special file name indicating a no rewind device, as in the 

example. For performance reasons, we suggest that you use a special device 

file name that has compression turned on. For more information on the device 

special file name to use, see Table 2-1 on page 30. 



server. If you use a rewinding device, the read/write head is repositioned at the 

beginning of the volume, and the previously backed-up data is overwritten. 

11.Then choose the device type (LTO Ultrium): 

Select the device type for "/dev/rmt/1stbn". 

The possible types are: 

3480 - 1/2 inch cartridge tape. 

3570 - IBM 3570. 

3590 - IBM 3590. 

4890 - StorageTek TwinPeaks cartridge tape. 

4mm 

- 4mm DDS-1 tape. 

4mm_4GB - 4mm DDS-2 tape. 

4mm_8GB - 4mm DDS-2 tape with compression. 



8mm 

- 8mm tape. 

8mm_5GB - 8mm double density tape. 

8mm_20GB - 8mm 20GB tape (Mammoth). 

8mm_AIT - 8mm 20GB AIT tape. 

8mm_AIT-2 - 8mm 50GB AIT2 tape. 

8mm_Mammoth-2 - 8mm 60GB Mammoth2 tape. 

8mm_Mammoth-3 - 8mm 100GB Mammoth3 tape. 

9490 - StorageTek TimberLine cartridge tape. 

9840 - StorageTek 9840 cartridge tape. 

9940 - StorageTek 9940 cartridge tape. 

dlt 

- Digital Linear Tape. 

dlt1 

- Benchmark DLT1. 

dlt7000 - 35/70GB Digital Linear Tape. 

dlt8000 - 40/80GB Digital Linear Tape. 

sdlt 

- 100/200GB SuperDLT. 

dtf 

- Sony Digital Tape Format. 

dst 

- AMPEX dst. 

file 

- file type device. 

himt 

- 1/2 inch magnetic tape. 

LTO_Ultrium - LTO Ultrium tape. 

optical - Optical Disk. 

qic 

- 1/4 inch data cartridge. 

qic_13GB - 13 GB 1/4 inch data cartridge. 

sd3 

- StorageTek RedWood SD-3. 

slr 

- Tandberg SLR family of QIC drives. 

tz85 

- DEC Digital Linear Tape. 

tz86 

- DEC Digital Linear Tape. 

tz87 

- DEC 10/20GB Digital Linear Tape. 

tz88 


tz89 



tzs20 

tkz90 

vhs 

VXA 

- DEC 8mm AIT tape. 

- DEC 3590 compatible tape. 

- VHS Format Tape. 

- VXA Tape drives. 

Select device type for /dev/rmt/1stbn [4mm]: LTO_Ultrium 

12.Repeat the same operation for every tape device that you want to be 

managed by Legato NetWorker. When finished, press Enter. The installation 

process continues and when complete, you should see the following 

message: 


13.Now you are ready to start the NetWorker daemons with the following 

command: 

# /etc/init.d/networker start 

Check if the daemons are running with the command: 

# ps -ef |grep nsr 

The command output should look like this: 

root 236 235 0 Aug 22 ? 0:01 /usr/sbin/nsrexecd 

root 235 1 0 Aug 22 ? 0:01 /usr/sbin/nsrexecd 

root 240 1 0 Aug 22 ? 0:01 /usr/sbin/lgtolmd -p /nsr/lic -n 1 

root 243 1 0 Aug 22 ? 0:27 /usr/sbin/nsrd 

root 324 243 0 Aug 22 ? 0:00 /usr/sbin/nsrindexd 

root 313 243 0 Aug 22 ? 0:00 /usr/sbin/nsrmmdbd 

root 337 243 0 Aug 22 ? 0:00 /usr/sbin/nsrmmd -n 1 

root 353 243 0 Aug 22 ? 0:00 /usr/sbin/nsrmmd -n 2 

Depending on the packages installed, in Table 6-2 you see the daemons 

required for each NetWorker component. 

Table 6-2 Legato NetWorker daemons for Solaris 

NetWorker installation packages 

NetWorker daemons 





nsrmmd 

nsrexecd 





UltraScalable Tape Library), check 6.8.1, “Autochanger configuration (SUN)” on 

page 253. 

6.5 Installation in a HP-UX environment 


Installation Guide - HP-UX Version. This book can be downloaded in PDF format 










In the following example, we are installing Legato NetWorker in the environment: 

► HP9000 L-series server with HP-UX operating system release 11 and LVD 

SCSI adapter 


► 




6.5.1 Software installation (HP-UX) 


directory on the HP-UX server. These are the required steps: 


# gunzip networker_hp11_32.tar.gz 


name, without the .gz last level qualifier 


command: 


# tar -xvf networker_hp11_32.tar 

The output is the installation package file named nw_hp11_32.pkg. 

3. Start the package installation process using the HP-UX GUI installation 

interface: 

# swinstall 

The Software Selection window appears, followed by the Specify Source 

window. Make sure that Source Depot Type is set to Local Directory, and 

enter the package file name and path in the Source Host Name field, as 

described in Figure 6-1. 

Figure 6-1 Specify source window of the software installation GUI interface 

4. Press OK and you will see the NetWorker package displayed as in Figure 6-2. 

Figure 6-2 Select Legato NetWorker package 


5. Double click the line that displays Legato NetWorker, and the window in 

Figure 6-3 appears: 

Figure 6-3 Subproducts selection window of the software install dialog 

6. Select the Actions menu and mark the packages that you want to install. To 

install the server software you need to select at least: 

– NWr-Client 

– NWr-Node 

– NWr-Server 

7. Select Install (analysis) from the Actions menu, and then if there were no 

errors, click OK to begin the packages installation. 

8. If the installation process was successful you should see the panel in 

Figure 6-4. 

Figure 6-4 Install window after the successful package installation 


9. To start the NetWorker daemons after the successful installation of the 

product, use the following command: 

# /sbin/init.d/networker start 

And, check if the daemons are running with the command: 

# ps -ef |grep -E “nsr|lgto” 

The command output should look like this: 

root 2304 2301 0 10:37:51 ? 0:00 /opt/networker/bin/nsrexecd 

root 2316 2308 0 10:37:56 ? 0:00 /opt/networker/bin/nsrindexd 

root 2317 2308 0 10:37:58 ? 0:00 /opt/networker/bin/nsrmmd -n 1 

root 2301 1 0 10:37:50 ? 0:00 /opt/networker/bin/nsrexecd 

root 2305 1 0 10:37:51 ? 0:00 /opt/networker/bin/lgto.... 1 

root 2308 1 0 10:37:53 ? 0:00 /opt/networker/bin/nsrd 

root 2322 2070 1 10:38:26 pts/ta 0:00 grep -E nsr|lgto 

root 2315 2308 0 10:37:54 ? 0:00 /opt/networker/bin/nsrmmdbd 

Depending on the packages installed, in Table 6-3 you see the daemons 

required for each NetWorker component. 

Table 6-3 Legato NetWorker daemons for HP-UX 

NetWorker installation packages 

NetWorker Daemons 





nsrmmd 

nsrexecd 




UltraScalable Tape Library), check 6.9.1, “Autochanger configuration (HP-UX)” 

on page 257. 

6.6 Legato NetWorker implementation and use 

In the following section, we describe how to correctly implement the required 

tasks to get the LTO library and drives operational in your data center 

environment after the Legato NetWorker product installation. We then show how 

to verify proper library operation with NetWorker. 


6.7 Implementation in AIX environment 

Before using the Legato NetWorker product, you need to configure the LTO 

medium changer to the software. Here are the steps to implement the required 

definitions to use the LTO library. 

6.7.1 Autochanger configuration (AIX) 

1. Use the Legato inquire command to check which tape devices are 

connected and visible to the system. 

This gives you the following output: 

# /etc/LGTOuscsi/inquire -l 

-l flag found: searching all LUNs, which may take over 10 minutes per 

adapter for some fibre channel adapters. Please be patient. 

scsidev@5.0.0:IBM ULT3583-TL1 |Autochanger (Jukebox) /dev/smc4 

scsidev@5.0.0:IBM ULT3580-TD1 |Tape, /dev/rmt2.1 

scsidev@5.1.0:IBM ULT3580-TD1 |Tape, /dev/rmt3.1 

Here you see the devices recognized as SCSI attached to this system and the 

information on their SCSI identification. For example, in scsidev@5.1.0, the 

last two numbers indicate the target ID and the LUN number (tgt 1, LUN 1). 

Remember this information, because it will be used later in the autochanger 


Note: Refer to “Legato NetWorker autochanger maintenance commands” 

on page 273 for a list of Legato NetWorker autochanger commands, such 

as inquire. 

2. Before configuring the 3583 medium changer to be managed by Legato 

NetWorker as an autochanger device (if you had previously installed the 

IBM-provided Atape device driver), you have to remove the medium changer 

device from the AIX configuration. You will have to do this if the inquire 

command detected a device /dev/smcn. To remove the medium changer, use 

the command: 

# rmdev -l smcn -d 

Where, n is the number associated with the medium changer special device 

name. 


Alternatively, you can use the smit interface to remove the device from the 

configuration, as detailed in 2.2.6, “Deleting LTO devices” on page 38. After 

the command execution you should no longer see the medium changer in the 

list of defined tape drives (use the following command). 

# lsdev -Cctape 

rmt2 Available 1A-08-00-0,0 IBM 3580 Ultrium Tape Drive 

rmt3 Available 1A-08-00-1,0 IBM 3580 Ultrium Tape Drive 

3. Now you can use the jbconfig command to define an autochanger like the 

IBM LTO medium changer for NetWorker automatic storage management. 

When you run jbconfig, you are prompted for the type of autochanger to 

install: 

# jbconfig 

1) Install a SmartMedia Jukebox. 

2) Install an Autodetected SCSI Jukebox. 

3) Install an Autodetected NDMP SCSI Jukebox. 

4) Install an SJI Jukebox. 

5) Install an STL Silo. 

What kind of Jukebox are you installing? [1] 2 

For an IBM LTO library, select option [2]. It could take some minutes to auto 

detect the autochanger, because the jbconfig program scans all the SCSI 

ports connected to the server. You should see the messages: 

These are the SCSI Jukeboxes currently attached to your system: 

1) scsidev@5.6.0: Standard SCSI Jukebox, Vendor , Product 

 

Which one do you want to install? 1 

Select the entry corresponding to the 3583 library, (in our example, [1]) and 

press Enter. Then: 

Installing an 'Standard SCSI Jukebox' jukebox. 

Name you would like to assign to the jukebox device? ibmlib1 

A brief description of the jukebox device? [scsidev@5.6.0: Standard SCSI 

Jukebox] 3583 Ultrium library 

Pathname of the control port for the jukebox device? [scsidev@5.6.0] 

Enter a name and a description for the library, then the pathname as returned 

by the previous inquire command (if the default name in brackets is correct, 

just press enter). 

Then you see the following prompt: 


Do you want automated device cleaning support enabled? (yes/no) no 

Answer no, if you are configuring the autochanger for a 3584 Ultrium 

UltraScalable Tape Library. Because this model supports automatic cleaning 

driven by the library, IBM recommends that you use this, rather than enabling 

the cleaning to be done by the application software. Therefore, select no for 

this option. If you have another LTO product like the 3580 drive, the 3581 

Autoloader or the 3583 Ultrium Scalable Tape Library, you can choose either 

to enable the Legato NetWorker automatic device cleaning (enter yes) or to 

perform your own manual cleaning (enter no), as these LTO models do not 

support automatic library cleaning. At the next prompt: 

Will media drive 1 be shared by multiple device paths? (yes/no) no 

Answer no, if you don’t want to activate Autochanger Sharing. Refer to 6.12.2, 

“Sharing autochangers between NetWorker hosts” on page 267 if you want to 

use this feature. Next you see the messages: 

Enter hostname that owns media drive 1: ? [brazil.almaden.ibm.com] 

Enter pathname of media drive 1: ? [/dev/rmt2.1] 

using '/dev/rmt2.1' as device name 

Use the device special file name indicating a no rewind device, as you did 

earlier in the Legato NetWorker installation. For more information on the 

device special file name to use, see Table 2-1 on page 30. If the device was 

not been configured yet, the following prompt appears: 

This media device has not been configured yet. Please select a media 

device type for /dev/rmt2.1. 

1) 3480 23) dst 

2) 3570 24) dst (NT) 

3) 3590 25) dtf 

4) 4890 26) file 

5) 4mm 27) himt 

6) 4mm 12GB 28) logical 

7) 4mm 20GB 29) LTO Ultrium 

8) 4mm 4GB 30) optical 

9) 4mm 8GB 31) qic 

10) 8mm 32) SD3 

11) 8mm 20GB 33) sdlt 

12) 8mm 5GB 34) SLR 

13) 8mm AIT 35) tkz90 

14) 8mm AIT-2 36) tz85 

15) 8mm Mammoth-2 37) tz86 

16) 9490 38) tz87 

17) 9840 39) tz88 


18) 9940 40) tz89 

19) dlt 41) tzs20 

20) dlt1 42) vhs 

21) dlt7000 43) VXA 

22) dlt8000 

Choice? 29 

Select the LTO Ultrium device type. These questions are related to the media 

drive 1, that is, the first tape drive connected to the library. The same 

questions apply to every other tape drive installed in the tape library. Then: 

Your jukebox does not report that it has a bar code reader, but it does 

report that it can handle volume tags. Some jukeboxes that have barcode 

readers report this way. If your jukebox has a barcode reader, do you want 

bar code reader support enabled? (yes/no) yes 

Do you want volume labels to match bar code labels? (yes/no) yes 

These two last questions apply only if you have a 3583 Ultrium Scalable Tape 

Library or 3584 Ultrium UltraScalable Tape Library, both of which have a 

barcode reader. Answer yes, if you have one of these models. The last 

message received should be: 

Jukebox has been added successfully 

4. After the successful completion of the autochanger configuration, you should 

be able to operate the IBM LTO medium changer from the Legato NetWorker 

administration interface. 

6.8 Implementation in a Solaris environment 




6.8.1 Autochanger configuration (SUN) 

1. Use the Legato inquire command to check the tape devices connected to the 

system. 


/etc/LGTOuscsi/inquire -l 

-l flag found: searching all LUNs, which may take over 10 minutes per 

adapter for some fibre channel adapters. Please be patient. 


scsidev@2.1.1:IBM ULT3583-TL 2.50|Autochanger (Jukebox) 

scsidev@2.1.2:IBM ULT3580-TD1 16E0|Tape 





Remember this information, it will be used later in the autochanger 




as inquire. 

You can also see the firmware microcode level (2.50 for the medium changer 

and 16E0 for the drives). 


NetWorker as an autochanger device, if you had previously installed the IBM 

provided Atape device driver, you have to remove the medium changer device 

from the Solaris configuration. You will have to do this if the inquire command 

detected an Autochanger device. 

To remove it, edit the IBMtape.conf file located in the /kernel/drv directory. 

Comment out (or delete) the entry with the SCSI target ID and LUN number of 

the medium changer device (in this example, it is target=1 lun=1): 

#name="IBMtape" 

# class="scsi" 

# target=1 lun=1 

# block_size=0 

# buffering=1 

# immediate=0 

# trailer=0 

# sili=0; 

Use the # symbol to comment out the entry. Save the changes. Then, to 

create an entry in the Solaris standard devices configuration file st.conf, which 

is located in the same directory (/kernel/drv): 

name="st" class="scsi" 

target=1 lun=1; 

Add this entry to the bottom of the file. The target and LUN numbers are the 

addresses of the LTO medium changer, as reported by the inquire command. 


Then reboot the server using the following command: 

# reboot -- -r 

3. After the server has successfully restarted, you can configure the 

autochanger to Legato NetWorker, using the jbconfig command. When you 

run jbconfig, you are prompted for the type of autochanger to install: 

# jbconfig 







For an IBM LTO library, select option [2]. Then jbconfig searches for attached 

SCSI jukeboxes. You will be prompted: 



 


Select the entry corresponding to the Ultrium library. At the next prompt: 


Name you would like to assign to the jukebox device? ibmlib 


Jukebox] 3583 Ultrium library 

Enter a user-assigned name for the autochanger and a brief description. 

Then: 


Enter the pathname as reported by the previous inquire command (in the 

example, scsidev@2.1.1). Then: 



UltraScalable Tape Library, because this library supports automatic cleaning 






to enable the Legato NetWorker automatic device cleaning (enter yes) or to 






use this feature. Then you see the following messages: 

Enter hostname that owns media drive 1: ? [sol-e.almaden.ibm.com] 

Enter pathname of media drive 1: ? [/dev/rmt/0stcbn] 

using '/dev/rmt/0stcbn' as device name 

Use the device special file name indicating a no rewind device, as you did 

earlier in the Legato NetWorker installation. For performance reasons, it is 

suggested to use a special device file name that has compression enabled. 

For more information on the device special file name to use, see Table 2-1 on 

page 30. If the device has not been configured yet, the following prompt 

appears: 


device type for /dev/rmt/0stbn 

1) 3480 23) dst 

2) 3570 24) dst (NT) 

3) 3590 25) dtf 

4) 4890 26) file 

5) 4mm 27) himt 




9) 4mm 8GB 31) qic 

10) 8mm 32) SD3 

11) 8mm 20GB 33) sdlt 

12) 8mm 5GB 34) SLR 

13) 8mm AIT 35) tkz90 

14) 8mm AIT-2 36) tz85 


16) 9490 38) tz87 

17) 9840 39) tz88 

18) 9940 40) tz89 

19) dlt 41) tzs20 

20) dlt1 42) vhs 

21) dlt7000 43) VXA 

22) dlt8000 


Choice? 29 

















6.9 Implementation in HP-UX environment 




6.9.1 Autochanger configuration (HP-UX) 

1. Use the Legato inquire command to check the tape devices connected to the 

system: 


# /etc/LGTOuscsi/inquire 

scsidev@2.1.3:IBM ULT3583-TL 2.50|Autochanger (Jukebox) 







Remember this information, it will be used later in the autochanger 




as inquire. 

You can also see the firmware microcode level (2.50 for the medium changer 

and 16E0 for the drives). 


NetWorker as an autochanger device, if you had previously installed the IBM 

provided ATDD device driver, you have to remove it from the HP-UX kernel 

configuration. To remove it, use the procedure detailed in 2.4.5, “Deleting the 

LTO medium changer” on page 56. 

3. Use the jbconfig command to define the IBM LTO medium changer for 

NetWorker automatic storage management. When you run jbconfig, you are 

prompted for the type of autochanger to install: 

# jbconfig 







For an IBM LTO library, select option [2]. Then jbconfig searches for attached 

SCSI jukeboxes. When you are prompted: 


1) scsidev@3.2.0: Standard SCSI Jukebox, Vendor , Product 


 


Select the appropriate changer (ULT3583-TL is the name of the IBM LTO 

medium changer) and press Enter. 


Name you would like to assign to the jukebox device? ibmlib 



Jukebox] IBM LTO Ultrium 3583 

Assign a name and a brief description to the changer, then at the next prompt: 


Just press Enter, because the default control port for the changer is assigned. 

Then the next selections are: 



UltraScalable Tape Library. Because this model supports automatic cleaning 





to enable the Legato NetWorker automatic device cleaning (enter yes), or to 






use this feature. Then you see the following messages: 

Enter hostname that owns media drive 1: ? [easter.almaden.ibm.com] 

Enter pathname of media drive 1: ? [/dev/rmt/0mnb] 

using '/dev/rmt/0mnb' as device name 

Use the device special file name indicating a no rewind device. For more 

information on the device special file name to use, see Table 2-1 on page 30. 



server. If you use a rewinding device, the read/write head is repositioned at 

the beginning of the volume, and the previously backed-up data is 

overwritten. 

For performance reasons, we suggest that you use a device that has 

compression turned on. Refer to 2.4.4, “Configuring tape devices” on page 55 

to see how to check the device configuration. If the device has not been 

configured yet, the following prompt appears: 


device type for /dev/rmt/0mnb 

1) 3480 23) dst 

2) 3570 24) dst (NT) 

3) 3590 25) dtf 

4) 4890 26) file 

5) 4mm 27) himt 




9) 4mm 8GB 31) qic 

10) 8mm 32) SD3 

11) 8mm 20GB 33) sdlt 

12) 8mm 5GB 34) SLR 

13) 8mm AIT 35) tkz90 

14) 8mm AIT-2 36) tz85 


16) 9490 38) tz87 

17) 9840 39) tz88 

18) 9940 40) tz89 

19) dlt 41) tzs20 

20) dlt1 42) vhs 

21) dlt7000 43) VXA 

22) dlt8000 

Choice? 29 




Should the drive be configured as a NDMP device? (yes/no) no 















6.10 Installation verification 

Start the Legato NetWorker administration Graphical User Interface (GUI) using 

the following command: 

# nwadmin 

And, check if the LTO drives are present in the list of configured devices. The 

administration main panel should look similar to the one shown in Figure 6-5. 


Figure 6-5 Legato NetWorker administration GUI 

You can see the LTO Ultrium devices listed in the Devices window. 

Use the Media pull down menu to select the Jukeboxes window and check if the 

LTO library is present. The windows should look like those in Figure 6-6. 


Figure 6-6 Jukeboxes window for the LTO tape library 

To test the LTO library and drives, use the NetWorker administration dialog to 

mount and dismount tape volumes in the LTO drives, do an inventory of the 

library and label volumes to check the correct functionality of the medium 

changer. We show how to label volumes and inventory the library in 6.12.4, 

“Inventorying and labeling volumes with barcode” on page 268. 

6.11 Adding a new LTO device to Legato NetWorker 

If your UNIX environment already has Legato NetWorker installed, and you are 

implementing IBM LTO library and drives for the first time, you may need to add 

one or more tape devices to your NetWorker configuration. 


Open the NetWorker administration program (nwadmin). From the Media menu, 

select Devices to open the devices window. Click Create, and in the Name 

attribute, enter the device pathname. Refer to Table 2-1 on page 30 to find the 

right name for your platform. 

Remember to use a no rewind tape device for Legato NetWorker as this is 

required by the software. 

From the Media Type attribute, select LTO Ultrium, as shown in Figure 6-7. 

Figure 6-7 Configuring a new tape device 

The Target Sessions parameter, should be set to 1 for tape devices. If a device 

is already receiving the number of backup sessions determined by the target 

sessions value, the NetWorker server uses the next least utilized device for your 

backups. So for example, if you have two tape drives, you can run two backup 

sessions in parallel only if you specify one as the Target Sessions for each tape 

device. 


For the Cleaning Required field, answer no if you are configuring the 

autochanger for a 3584 Ultrium UltraScalable Tape Library. Because this model 

supports automatic cleaning driven by the library, IBM recommends that you use 

this feature, rather than enabling the cleaning to be done by the application 

software. Therefore, select no for this option. If you have another LTO product like 

the 3580 drive, the 3581 Autoloader or the 3583 Ultrium Scalable Tape Library, 

you can choose either to enable the Legato NetWorker automatic device 

cleaning (enter yes), or to perform your own manual cleaning (enter no), as these 

LTO models do not support automatic library cleaning. 

The Automatic Media Management feature can be activated from this panel 

only if you are using a manual tape drive. This feature gives the NetWorker 

server automatic control over media loaded in the storage device. If you have an 

LTO library or autoloader, this feature can be enabled only from the Jukeboxes 

resources dialog. We discuss this feature more in 6.12.3, “Auto media 

management” on page 268. 

Important: Do not activate the Automatic Media Management feature 

unless you are using your LTO tape library exclusively with NetWorker. If this is 

activated, NetWorker assumes that all the volumes loaded in the autochanger 

will be handled exclusively as NetWorker media. This could lead to unwanted 

data overwrites if there are other non-NetWorker volumes in the autochanger. 

6.12 Legato NetWorker exploitation 

In this section we give you further information on Legato NetWorker capabilities 

related to tape processing. These are: 

► 

► 

► 

Dynamic drive sharing (DDS) 

Auto media management 

Using the barcode reader with NetWorker 

These are all functions that you can enable and configure to exploit the Legato 

NetWorker product with the IBM LTO tape library and drives. 

6.12.1 Dynamic drive sharing (DDS) 

Dynamic drive sharing (DDS) is a feature that provides NetWorker software with 

the ability to recognize shared drives. Shared drives are drives which are 

available to more than one NetWorker server. DDS enables NetWorker software 

to be aware of all the servers with access to these drives. If a drive is required for 

an operation, it will skip any shared drives that are in use and route the backups 

or recoveries to other available shared drives. 


Note: DDS is only supported in a SAN Fibre Channel environment, not in a 

direct-connect, twin-tailed SCSI environment. 

Enabling DDS on your system provides you with the following benefits: 

► 

► 

Reduces storage costs — A single tape drive can be shared among several 

storage nodes. In fact, since the NetWorker software uses the same open 

tape format for UNIX, Windows NT, NetWare and Linux, you can actually 

share the same tape between different platforms (assuming that respective 

save sets belong to the same pool). 

Reduces LAN traffic —You can reconfigure clients as SAN storage nodes 

which can send save sets over the SAN to shared drives. 

In the example in Figure 6-8, we are using one NetWorker server, two NetWorker 

storage nodes (SN1 and SN2), and a 3583 Ultrium Scalable Tape Library with 

two drives. 


Storage Node 

SN1 

SAN 

Storage Node 

SN2 

IBM 3583 

SDG Router 

SCSI 

SCSI 

I/Face 

Drive 1 

Drive 2 

shared autochanger 

Figure 6-8 Legato NetWorker Dynamic Drive Sharing (DDS) example 


In this example, SN1 sends a backup to drive_1 and subsequently, SN2 starts a 

backup to drive_1. If the backup from SN2 can use the same tape in drive_1, it 

will. If not, the backup started by SN2 will be sent to the next available drive, 

which is drive_2. 

6.12.2 Sharing autochangers between NetWorker hosts 

Legato NetWorker allows the control and usage of individual tape devices within 

a medium changer on different NetWorker hosts (a NetWorker server or storage 

node). This feature is known as autochanger sharing. 

Autochanger sharing allows one NetWorker host to control the autochanger’s 

robotic arm, while other NetWorker hosts (as well as the host controlling the 

robotic arm) can each control and use specific tape devices. A specific device 

can be controlled only by a single NetWorker host, therefore the drive sharing is 

static. Figure 6-9 shows the autochangers sharing drives between different 

NetWorker hosts. 


Storage Node 

SN1 

Medium 

Changer 

IBM 3583 

Drive 6 

Drive 5 

Drive 4 

Drive 3 

Drive 2 

Drive 1 

shared autochanger 

Storage Node 

SN2 

SCSI medium changer control 

SCSI tape device control 

Storage Node 

SN3 

Figure 6-9 How Legato NetWorker autochanger sharing works 

You can set the LTO medium changer as shared during the initial autochanger 

setup using the jbconfig command. This command must be run from the 

NetWorker host that controls the autochanger. 


When running jbconfig, you will be prompted to enter the NetWorker hostname 

controlling the first autochanger device. The default hostname value is the local 

NetWorker host on which jbconfig is running. However, you can change the 

default hostname value to another NetWorker host. Then you are prompted to 

enter the controlling host for each device in the autochanger. This would 

correspond to the host which is cabled to each drive. 

6.12.3 Auto media management 

Auto media management further automates the backup procedures by freeing 

administrators from the task of labeling volumes before backup. When auto 

media management is enabled in the Jukebox resource, Legato NetWorker 

assumes that all the volumes loaded in the medium changer may be handled 

exclusively as NetWorker media. Volumes that appear unlabeled are considered 

blank and are automatically labeled, mounted, and overwritten with new data. 

Important: Be careful about sharing NetWorker volumes between different 

storage devices if you have auto media management enabled. You could 

potentially relabel and overwrite valuable data. 

To enable auto media management, you have to enable the Auto Media 

Management attribute in the Device resource for 3580 manual devices and in the 

Jukebox resource for 3581 Autoloader, 3583 Ultrium Scalable Tape Library, and 

3584 Ultrium UltraScalable Tape Library. 

6.12.4 Inventorying and labeling volumes with barcode 

Labeling is the process of identifying the volumes to NetWorker, so it can locate 

the right cartridge for a backup and restore operation. 

Inventorying is the process Legato NetWorker uses to register the locations of 

the volumes in the autochanger slots when it assigns a label. Inventorying 

happens automatically when new volumes are loaded and labeled by NetWorker. 

Inventorying with bar code labels is fast and efficient, because you do not have to 

mount the volumes into the device. Instead, the medium changer scans the 

external bar code labels with a laser light while the volumes remain in their slots. 

Inventorying with bar code labels greatly reduces the time it takes to locate a 

volume or determine the contents of an LTO tape library. Bar code labeling is 

therefore recommended where the hardware supports it, for example, with the 

3583 Ultrium Scalable Tape Library or 3584 Ultrium UltraScalable Tape Library. 


When you use the label function of the Legato NetWorker administration 

interface with an autochanger with barcode reader, you don’t have to specify the 

label number, because the medium changer barcode reader matches the 

external label with the tape volume internal label. As shown in Figure 6-10, the 

Jukebox Labeling window has already set the Barcode label value in the 

Starting with field. 

Figure 6-10 Jukebox labeling 

The NetWorker server uses bar code labels only to inventory volumes. It is 

important to update and maintain an accurate account of the medium changer 

contents by using the NetWorker inventory command. Otherwise, locating a 

volume for backup or recovery (whether or not it uses bar code labels) is 

impossible, because the NetWorker server does not know which slot contains the 

needed volume. 

Sometimes, a manual inventory is required. For example, if you need to open the 

3583 or 3584 tape library door and move the cartridges, or extract them for 

maintenance or other reasons, remember to start an inventory of the library from 

the Legato NetWorker administration interface after you put the library online. 

Use the option Media -> Inventory, then click OK in the Jukebox Inventory 

window, as shown in Figure 6-11. 


Figure 6-11 Inventorying the library 


A 

Appendix A. 

Useful commands 

In this appendix we give a list of commonly commands which are particularly 

relevant when installing and configuring tape drives and libraries. It includes: 

► 

► 

► 

► 

► 

Table of commonly used AIX commands 

Table of commonly used Solaris commands 

Tivoli Storage Manager tape drive and library commonly used commands 

Legato NetWorker autochanger maintenance commands 

Using tapeutil to verify correct device attachment 

AIX commands 

Here is a useful list of common AIX commands: 

Table A-1 AIX commands 

Command 

Description 

lscfg -v 

lscfg -vl rmtx 

lscfg -vl fcsx 

lsattr -E -l rmtx 

General device configuration 

Tape drive configuration 

FC HBA configuration 

Device parameters 


Command 

errpt -a 

lslpp -l Atape.driver 

lsdev -Cc tape 

Description 

Error log. Call hardware support if errors 

are found for TAPE_ERR* or SCSI* error 

labels. 

Driver version 

Configured tape drives 

SUN Solaris commands 

Here is a useful list of common Solaris commands: 

Table A-2 Commonly used SUN Solaris commands 

Command 

Description 

boot -r 

catman -w 

exec ksh 

set -o emacs 

dmesg |more 

exportfs 

format 

init 0 

init 6 

iostat -dnx 2 

kill -9 [processnumber] 

mount [IPaddress]:/[mount-point] 

ps -ef 

reboot -- -r 

uname -a 

eject 

Boot sequence with device scan 

Format manuals 

Use emacs command-line edit in Korn 

shell 

Show the boot log file 

Show free NFS volumes 

Show all hard disks 

OS shutdown 

OS shutdown and reboot 

I/O statistics 

Kill specified process number 

Show mounts from host with specified IP 

address 

List of all processes 

Init 6 and rescan devices 

Show OS level 

Ejects an unmounted CD ROM 


Tivoli Storage Manager tape and tape library commands 

Table A-3 Commonly used TSM tape commands 

Command 

QUERY LIBRARY 

QUERY DRIVE 

QUERY DEVCLASS 

QUERY LIBVOLUME 

QUERY VOLUME 

QUERY VOLUME ACCESS=READONLY 

CHECKIN LIBVOLUME 

CHECKOUT LIBVOLUME 

LABEL LIBVOLUME 

AUDIT LIBRARY 

REPLY nnn 

Description 

Information about the library 

Information about tape drives 

Information about device classes 

Find volumes (scratch and private) 

Find private volumes 

Find volumes in read only status 

Check in volumes to the library 

Check out volumes to the library 

Write TSM labels on cartridges 

Check consistency of library inventory 

Reply to console request 

Legato NetWorker autochanger maintenance commands 

NetWorker device driver software provides the following maintenance commands 

for diagnosing problems on tape devices and medium changers. 

Table A-4 Legato NetWorker autochangers maintenance commands 

Command 

Description 

lusbinfo 

lusdebug 

lrescan 

lreset 

changers 

inquire 

ldunld 

msense 

Prints out SCSI information 

Sets the library debugging level 

Rescans for devices 

Resets the SCSI bus 

Lists the SCSI autochangers attached to 

the system 

Lists the devices available 

Loads or unloads a tape device 

Retrieves mode sense data 

Appendix A. Useful commands 273

Command 

pmode 

tur 

writebuf 

sjirjc 

Description 

Prints mode sense data 

Tests whether the unit is ready 

Writes a device buffer 

Tests the jukebox 

Verifying proper device attachment with tapeutil 

Before you start to use your devices for production work with your applications, or 

if you encounter difficulties with your devices, you may want to verify that the 

hardware, connections, and device drivers are working together properly. Before 

you can do this, you should have already installed your LTO according to 


setup” on page 93. 

AIX tape and medium changer device attachment test 

The following procedure tests the attachment of a medium changer device and a 

tape drive to an AIX system. The procedure assumes that the device is attached 

at /dev/rmt0 and that there is no cartridge in the drive. You also need an 

unassigned cartridge available for loading in at least one of the slots. 

1. Open an AIX window. 

2. Enter tapeutil. The tapeutil menu will be displayed (as in Figure 2-4 on 

page 58). 

3. Enter 1 (Open a Device). 

4. Enter /dev/rmt0 when prompted for the device name. 

5. Enter 1 (Read/Write). 

6. Enter 3 (Device Info). 

7. Enter 5 (Inquiry). Specify 0 when prompted for an inquiry page. This 

concludes a very basic test of the device, SCSI connection, and the device 

driver. You may stop the test here or continue to perform a more complete 

test. 

8. Enter 14 (Element Inventory). 

9. From the output of the previous step, select a writable, scratch cartridge and 

determine its element ID. Also, select the element ID of the first tape drive. 


10.Enter 16 (Move Medium), then supply the address of the cartridge, followed 

by the address of the tape drive. Verify that the cartridge moved. 

11.Enter 6 (Test Unit Ready) until no error occurs. 

12.Enter 30 (Read and Write Tests). 

13.Enter 1 (Read/Write). Press the Enter key three times to accept the defaults 

and run the test. 

14.Enter 31 (Unload Tape). Wait until the Medium is unloaded. 

15.Enter 16 (Move Medium), then supply the address of the tape drive, followed 

by the address of the cartridge slot where it was before. Verify that the 

cartridge moved. 

16.Enter 2 (Close a Device). 

17.Enter Q (Quit Program). 

Solaris and HP tape and medium changer device attachment test 

The following procedure tests the attachment of a medium changer device and a 

tape device to a Sun or HP system. The procedure assumes that your media 

changer is configured at: 

► 

► 

/dev/rmt/0smc - for Sun System 

/dev/rmt/0chng - for HP System 

And, the tape drive is attached at: 

► /dev/rmt/1st - for Sun System 

► /dev/rmt/1m - for HP System 

And, it assumes that there is no cartridge in the drive. 

You also need an unassigned cartridge available for loading in at least one of the 

slots. 

1. Open a terminal session. 

2. Enter: 

– /opt/IBMtapeutil/tapeutil - for Sun System 

– tapeutil - for HP System 

The tapeutil menu will be displayed (as in Figure 2-4 on page 58). 

3. Enter 1 (Open a Device), when prompted for the device name: 

– Enter /dev/rmt/0smc - for Sun System 

– /dev/rmt/0chng - for HP System 


4. Enter 3 (Device Info). 

5. Enter 14 (Element Inventory). 

6. From the output of the previous step, select a writable, scratch cartridge and 

determine its element ID. Also, select the element ID the first tape drive. 

7. Enter 16 (Move Medium), then supply the address of the cartridge, followed 

by the address of the tape drive. Verify that the cartridge moved. 

8. Enter 6 (Test Unit Ready) until no error occurs. 

9. Enter 30 (Read and Write Tests). 

10.Enter 1 (Read/Write). Press the Enter key three times to accept the defaults 

and run the test. 

11.Enter 31 (Unload Tape). Wait until the Medium is unloaded. 

12.Enter 16 (Move Medium), then supply the address of the tape drive, followed 

by the address of the cartridge slot where it was before. Verify that the 

cartridge moved. 

13.Enter 2 (Close a Device). 

14.Enter Q (Quit Program). 

Using tapeutil element inventory 

When you select option 14. Element Inventory in tapeutil, you will receive a list of 

each element in the library and its status. First the robot itself is listed, followed 

by each of the slots in the Import/Export station. 

Reading element status... 

Robot Address 1 

Robot State .................... Normal 

ASC/ASCQ ....................... 0000 




Volume Tag ..................... 

Import/Export Station Address 16 

Import/Export State ............ Normal 

ASC/ASCQ ....................... 0000 


Import Enabled ................. Yes 

Export Enabled ................. Yes 





Volume Tag ..................... 

Scroll down to continue the listing of the rest of the Import/Export slots. Next, the 

drives are listed: 



ASC/ASCQ ....................... 0000 








Volume Tag ..................... 



ASC/ASCQ ....................... 0000 








Volume Tag...................... 

Finally, the individual storage slots are listed: 

Slot Address 4096 

Slot State ..................... Normal 

ASC/ASCQ ....................... 0000 

Media Present .................. Yes 


Source Element Address ......... 4096 


Volume Tag ..................... ABA920L1 

Slot Address 4097 

Slot State ..................... Normal 

ASC/ASCQ ....................... 0000 






Volume Tag ..................... 

You can see that slot 4096 is occupied, since Media Present is set to Yes and the 

Volume Tag contains ABA920L1, indicating this tape cartridge is located in this 

slot. Slot 4097 is empty, since Media Present is set to No. Therefore, when 

selecting option 16 Move Medium, assuming we knew the tape in slot 4096 was 

not already used by an application, we could specify source address of 4096 and 

destination address 256. This would load that tape into the tape drive. To unload 

the media (after selecting 31 Unload Tape) we would move the medium from 256 

back to 4096. 


B 

Appendix B. 

Additional information 

This appendix provides additional information that is useful when implementing 

IBM LTO tape drives and libraries. 

The topics are: 

► Simple Network Management Protocol (SNMP) alerts 

► Configuring SNMP for 3583 Ultrium Scalable Tape Library 

► Configuring SNMP for 3584 Ultrium UltraScalable Tape Library 

► Guidelines for booting SAN Data Gateway and FC hosts 

► Some performance considerations for tape devices 


Simple Network Management Protocol (SNMP) alerts 

Occasionally, the LTO library may encounter a situation that you want to know 

more about, such as an open door that causes the library to stop. The 3583 and 

3584 libraries provide a standard TCP/IP protocol called Simple Network 

Management Protocol (SNMP) to send alerts about conditions (such as, an 

opened door) over a TCP/IP LAN network to an SNMP manager. These alerts 

are called SNMP traps. The SNMP connection requires an Ethernet port, 

therefore your library must have the RMU feature (for 3583) or Ethernet port (for 

3584) installed. 

SNMP is itself a sub-protocol of TCP/IP and is based on a hierarchical structure 

represented by strings of “object identifiers” or OIDs. Much of the data required 

by network managers is mapped to instances of the OIDs. Standard values are 

included in higher levels of the structure, providing a standard for identifying 

those data. Brand-specific extensions are provided by the manufacturers. These 

standard and proprietary OID collections are called Management Information 

Bases (MIBs). The specification for SNMP, contained in RFC 1157 is available at: 

ftp://ftp.isi.edu/in-notes/rfc1157.txt 

Using the information supplied in each SNMP trap, the SNMP manager (such as 

Tivoli NetView) can alert operations staff of possible problems or required 

operator interventions. The SNMP traps are transmitted over the Ethernet to the 

SNMP monitor station. 

The 3583 Ultrium Scalable Tape Library has the SNMP monitoring function 

included in the RMU module. 

The RMU supports SNMP version 2.0 and can act as an SNMP server, 

responding to SNMP requests and generating SNMP traps. An internal library 

serial interface enables the RMU to acquire TapeAlert 3.0 compatibility 

information from the drives and to send this information to an SNMP server. In 

the event of a power loss, the RMU will detect the loss and generate an SNMP 

trap for notification. 

The 3584 Ultrium UltraScalable Tape Library has the SNMP support included in 

the L frame. Up to five different SNMP managers can be configured to receive 

trap events. 

For both SNMP-capable LTO tape libraries, you need an Ethernet LAN 

connection to the SNMP manager to trap SNMP events. We described how to 

configure the LAN connection for these libraries in 2.9.1, “IBM 3583 StorWatch 

Specialist” on page 76 and 2.9.2, “IBM 3584 StorWatch Specialist” on page 81. 


We will describe how to send traps to a Tivoli NetView Version 6.01 on a 

Windows 2000 server. For more information on NetView, see the redbook Tivoli 

NetView 6.01 and Friends, SG24-6019. 

Configuring SNMP for the 3583 tape library 

To configure and activate SNMP event monitoring for the 3583 Ultrium Scalable 

Tape Library, use the LTO Ultrium StorWatch Specialist panels. You cannot use 

the 3583 Operator Panel to configure this function. 

From the main specialist panel (shown in Figure 2-7 on page 77), on the black 

column at the left hand side of the panel, select SNMP MIB. The panel shown in 

Figure B-1 appears: 

Figure B-1 3583 Storwatch Specialist SNMP MIB download panel 

As described in the text of the panel, the MIB file is a specific configuration file for 

the SNMP environment with the information for the 3583 tape library. Download 

the file to the SNMP management server. If you are running Tivoli NetView on 

this server, you should store the file in the directory \usr\ov\snmp_mibs. You will 

then load this file to your SNMP manager, using the specific application’s 

process. 

Appendix B. Additional information 281

Go to the 3583 specialist Configuration panel. In the SNMP Configuration 

section at the low left side of the panel, enter the TCP/IP address and other 

related parameters for the SNMP server, and make sure the Alerts Enabled 

pulldown is set to ON as shown in Figure B-2. 

Figure B-2 3583 Storwatch Specialist: configuring SNMP 

Click on the Submit button to apply the changes. A confirmation panel appears. 

Press Enter to confirm the updates. 

After you have configured the RMU to send SNMP alerts to the SNMP server, 

configure your SNMP monitoring product (like Tivoli NetView) to use the MIB 

supplied file and to monitor the events. An example of the Tivoli NetView 6.0.1 

panel is shown in Figure B-3. We can see our tape drive BAVARIA discovered as 

a node. 


Figure B-3 Tivoli NetView SNMP monitoring 

Configuring SNMP for the 3584 tape library 

Access the 3584 Operator Panel Dialog to configure and activate SNMP event 

monitoring on the 3584 Ultrium UltraScalable Tape Library. You cannot use the 

StorWatch Specialist to configure this function. 

From the library’s Activity panel, press MENU. Then press Settings -> Network 

Settings -> SNMP. The panel in Figure B-4 appears. 


Panel 0175 

Ethernet 

______________________________ 

Current Settings Frame 1: 

MAC Address: 18:36:F3:98:4F:9A 

IP Addresses: 10.1.1.1 

Subnet Mask: 255.255.255.0 

Gateway: 10.1.1.254 

[Change Settings] 

______________________________ 


Figure B-4 3584 Operator Panel: SNMP settings 

To enable or disable SNMP, press UP or DOWN to highlight Enable/Disable 

SNMP and press ENTER. The Enable/Disable SNMP panel displays with the 

current state of SNMP (Figure B-5). Press UP or DOWN to specify ENABLED or 

DISABLED for SNMP. The SNMP panel redisplays the new setting. To accept the 

new setting and return to the previous panel, press CANCEL. 

Panel 0166 

Enable/Disable SNMP 

______________________________ 

SNMP is Enabled 

Press UP or DOWN to toggle the 

SNMP setting, then press ENTER 

to activate the change. 

______________________________ 

CANCEL UP DOWN ENTER 

Figure B-5 3584 Operator Panel: enable/disable SNMP 


To view or change the destination IP address of a monitoring station to which 

SNMP alerts will be sent, press UP or DOWN to highlight Destination IP 

Addresses and press ENTER. Then press UP or DOWN to specify the 

destination IP address that you want to view or change, then press ENTER. The 

Set Destination IP Address panel displays with the current value of the 

destination IP address that you specified (Figure B-6). Press UP or DOWN to 

change to the value that you want, then press ENTER. The panel redisplays with 

the new destination IP address. To accept the new setting and return to the 

previous panel, press CANCEL. 

Panel 0172 

Set Destination IP Address 

______________________________ 

IP Address 1: 10.25.36.1 

Use [UP] and [DOWN] to change 

highlighted value. 

______________________________ 


Figure B-6 3584 Operator Panel: set destination IP address 

To view or change the remote port that is associated with the IP address, press 

UP or DOWN to highlight Remote Port and press ENTER. The Remote Port 

panel displays the number of the port that is associated with one of the five 

destination IP addresses. To change the value of the remote port, press UP or 

DOWN to specify the associated number and press ENTER. The Set Remote 

Port panel displays with the current value of the remote port (Figure B-7). Press 

UP or DOWN to increment or decrement the value. When the desired number 

displays, press ENTER. The library displays the message Changing Remote Port 

X to YYY. When the update is finished, Remote Port X is now YYY displays. Press 

CANCEL to return to the SNMP panel. 


Panel 0169 

Set Remote Port 

______________________________ 

Remote Port: [162] 

Use [UP] and [DOWN] to change 

highlighted value. 

______________________________ 


Figure B-7 3584 Operator Panel: set remote port 

To view or change the community name, press UP or DOWN to highlight 

Community Name and press ENTER. The Community Name panel displays 

with the current password (Figure B-8). To change the password, press UP or 

DOWN to specify the character that you want and press ENTER. Repeat this 

operation for the following characters. When you are done, the library displays 

the message Updating Community Name and the update of the password 

begins. When the update is finished, Community Name change displays. Press 

ENTER to display the new password. 


Panel 0167 

Community Name 

______________________________ 

Community Name: 

Public 

Press UP/DOWN to select the 

currect charachter. ENTER 

select the next charachter. 

ENTER on a blank commits 

changes. 

______________________________ 


Figure B-8 3584 Operator Panel: set community name 

To send a test trap, press UP or DOWN to highlight Send a Test Trap and press 

ENTER. The library displays the message “A test trap with the text.This is a test 

SNMP trap.has been sent to all defined target hosts. Press ENTER to continue”. 

Press ENTER. The library sends an SNMP trap to all SNMP IP addresses at the 

remote port that you specified. The trap contains the machine type, model 

number, and serial number of the library, as well as other fields. 

After you have configured the 3584 to send SNMP alerts to the SNMP server, 

configure your SNMP monitoring product (like Tivoli NetView) to monitor the 

events. An example of the Tivoli NetView 6.0.1 panel is shown in Figure B-3. 

Guidelines for booting SAN Data Gateway and FC hosts 

Device access problems sometimes occur when it is required to shutdown the 

IBM SAN Data Gateway. To avoid losing access to your FC devices when a 

shutdown is required, the SAN Data Gateway and its attached devices must be 

booted in a specific order. Rebooting is required after updating firmware and 

when adding or removing SCSI devices. The following procedure describes the 

recommended boot order, and is also applicable to the SAN Data Gateway 

Router. 


Attention: You must stop all I/O between the host and SCSI devices before 

rebooting the SAN Data Gateway. 

1. SCSI devices 

Turn on the SCSI devices first. All SCSI devices attached to the SAN Data 

Gateway must be turned on before turning on or rebooting the SAN Data 

Gateway. 

2. SAN Data Gateway 

The SAN Data Gateway scans the SCSI buses when it is booting. If you add 

or remove SCSI devices after the Gateway has booted, it will not 

automatically detect the changes. You can reboot the SDG using the 

StorWatch SAN Data Gateway Specialist client, from a telnet session or a 

serial terminal session. You can also use the scsiRescan command to scan 

the SCSI channels. 

3. Fibre Channel host 

Before turning on or restarting hosts that are connected via Fibre Channel to 

the SAN Data Gateway, you must wait until the SAN Data Gateway has 

finished booting as indicated by the Ready light on the front panel blinking at a 

rate of about once per second. 

► 

► 

► 

Although some operating systems may provide software methods that allow 

SCSI devices to be dynamically added or removed after the host has booted, 

we highly recommend that you reboot the host to ensure reliable operation. 

If you update SAN Data Gateway firmware, you must reboot the Gateway to 

use the new firmware. We highly recommended that you also reboot the host 

to ensure compatibility with any new firmware features or functions. 

If you update SCSI device firmware, the SAN Gateway Explorer application 

won’t display the new firmware version until the SAN Data Gateway has 

issued a SCSI inquiry, which occurs when it rescans the SCSI buses. The 

same is true for the StorWatch SAN Data Gateway Specialist client 

application and the service terminal. 

Performance considerations 

Performance analysis must be approached by determining which component of 

the data path impacts performance. Typically, a performance problem can be 

isolated by looking at one aspect of the data path at a time. The data path mainly 

consists of: 

► 

Client file system 


► 

► 

► 

► 

► 

► 

► 

Client server hardware 

Network 

Application server hardware 

Application server file system 

Application itself (for example, database) 

Connection to the tape drives 

Size of files which you backing up 

If you think you have a performance problem, then you have to determine which 

link of your chain is causing the problem, that is, where does the bottleneck 

occur? So therefore look at each link separately. The UNIX command dd can be 

used to get a baseline performance level. 

Attention: The numbers presented here are not intended to reflect any real 

performance which might be achieved. The tests were run without particular 

attention to optimizing performance and are intended to be used for 

comparative results only. 

We tested the tape performance on two hosts, BRAZIL and SOL-E. We created a 

file of 256 MB called rbs01.dbf. First, we tested the file system performance with: 

time dd if=filename of=/dev/null bs=128k 

Then we tested the tape performance with: 

time dd if=filename of=/dev/rmt0 bs=128k 

Example B-1 shows the results obtained on BRAZIL. There is good file system 

performance (over 74 MB/s) and the tape performance is also in the expected 

range (nearly 11 MB/s). 

Example: B-1 Performance on brazil 

# ls -l 

-rw-r--r-- 1 root sys 268439552 Sep 05 18:21 rbs01.dbf 

# time dd if=/software/lto/rbs01.dbf of=/dev/null bs=128k 

2048+1 records in. 

2048+1 records out. 

real 

user 

sys 

# 

0m3.43s 

0m0.02s 

0m3.42s 


# time dd if=/software/lto/rbs01.dbf of=/dev/rmt0 bs=128k 

2048+1 records in. 

2048+1 records out. 

real 

user 

sys 

0m23.68s 

0m0.02s 

0m3.95s 

Example B-2 shows you the result on SOL-E. As you can see the file system test 

takes longer than on BRAZIL, which can be attributed to less CPU, disk or I/O 

capacity. Therefore, the tape performance is also relatively slower. 

Example: B-2 performance on sol-e 

# time dd if=/usr/sepp/rbs01.dbf of=/dev/null bs=256k 

1024+1 records in 

1024+1 records out 

real 0m17.43s 

user 0m0.02s 

sys 0m3.80s 

# time dd if=/usr/sepp/rbs01.dbf of=/dev/rmt/0st bs=256k 



real 1m19.85s 

user 0m0.02s 

sys 0m3.77s 

# time dd if=/usr/sepp/rbs01.dbf of=/dev/rmt/0stc bs=256k 



real 

user 

sys 

0m33.39s 

0m0.05s 

0m3.53s 

In addition you can see in Example B-2 how the performance is different 

depending on whether hardware compression is used (dev/rmt/0st) or not 

(dev/rmt/0stc). 


C 

Appendix C. 

IBM LTO Tape Library 

sharing and partitioning 

In this appendix we describe the sharing and partitioning of the 3583 Ultrium 

Scalable Tape Library and 3584 Ultrium UltraScalable Tape Library. 

The topics are: 

► Tape library sharing and partitioning definitions 

► Sharing the 3583 Ultrium Scalable Tape Library 

► Sharing the 3584 Ultrium UltraScalable Tape Library 

► Partitioning the 3584 Ultrium UltraScalable Tape Library using the StorWatch 

Specialist 

► 

Partitioning the 3584 Ultrium UltraScalable Tape Library using the Operator 

Panel 


IBM LTO tape libraries attached to multiple servers 

The purpose of this appendix is to address frequently asked questions about 

attaching multiple servers to IBM Ultrium Tape Libraries. 

The following content is based on the document IBM LTO Tape Library Sharing 

White Paper, which is available from your IBM representative. 

Tape Library sharing and partitioning definitions 

Here are some of the most commonly used terms related to sharing and 

partitioning tape libraries: 

Library Sharing 

Tape Drive Pooling 

Multipathing 

Library Partitioning 

Multiple servers attached to a tape library, sharing both 

the library and the robotics. The tape drives within the 

library may or may not be shared (pooled) among the 

attached servers. Tape library sharing is a prerequisite for 

tape drive pooling. 

The sharing of one or more tape drives between multiple 

servers. 

The capability of a tape library to provide multiple paths to 

the SCSI Medium Changer. Multipathing capability is a 

prerequisite for partitioning. The 3584 Ultrium 

UltraScalable Tape Library and 3575 Magstar MP tape 

libraries are multipath libraries. The 3583 Ultrium Scalable 

Tape Librarytape library is a single path library. 

The capability to define the library's drives and slots as 

multiple logical libraries. The server attached to each 

partition has no knowledge of any drives or slots outside 

the partition. The 3584 Ultrium UltraScalable Tape Library 

and Ultrium 3575 Tape Libraries are capable of 

partitioning. Partitioning provides one approach to tape 

library sharing. 

Sharing the 3583 Ultrium Scalable Tape Library 

Figure C-1 shows the single path design of the Ultrium 3583 tape library, and 

illustrates one possible configuration for attaching the library to an IBM pSeries 

or RS/6000. 


SCSI Interface Module 

(SCSI Medium Changer) 

3583 


Tape 

Drives 

IBM p Series (RS/6000) 

Figure C-1 3583 Ultrium Scalable Tape Library single path design 

One SCSI bus from the pSeries is attached to the SCSI Medium Changer and 

two tape drives. The three other SCSI slots on this pSeries are attached to the 

remaining tape drives. 

The pSeries supports different speeds of SCSI adapters. Higher bus speeds 

allow additional tape drives to be daisy-chained on a SCSI bus without 

performance degradation. 

As shown in Figure C-1, no other servers could be attached to the SCSI Medium 

Changer unless a second initiator was added to the pServer's SCSI bus that is 

attached to the SCSI Medium Changer. This configuration option is discussed 

later in “Sharing the 3583 tape library using multiple initiators” on page 295. 

The address for the SCSI Medium Changer is LUN1 of the tape drive in the 3584 

Ultrium library. In contrast, the SCSI Medium Changer is a SCSI target ID in the 

3583 Ultrium library. 

Because every iSeries server (AS/400) SCSI bus has to have a library control 

path, and the 3583 has only one library control path, only one iSeries server 

(AS/400) SCSI bus can be attached to a 3583. LTO support for the iSeries server 

(AS/400) is limited to one drive per SCSI bus. This means that the a 3583 library 

attached to an iSeries server can only contain one tape drive. 

The 3583 Fibre Channel support (with integrated SDG), provides for the iSeries 

multi-initiator, single target, multiple LUN support. This means you can have 

multiple iSeries hosts attached to the 3583. Each of these hosts can have 

multiple drives, because the drives are attached to the SAN Data Gateway 

Module, which is seen as a single target by the hosts. All six drives could be 

configured for a single iSeries host. 

Appendix C. IBM LTO Tape Library sharing and partitioning 293

network 

In contrast, AIX, NT, and Solaris do not require a separate control path for every 

SCSI adapter. As a result, they can take advantage of all six tape drives in a 

Ultrium 3583. 

Three approaches could potentially provide tape library sharing with an Ultrium 

3583. 

Sharing the 3583 tape library using SAN and TSM 

The first approach involves attaching the 3583 via a SAN. Figure C-2 shows a 

SAN attached configuration which would enable tape library sharing. As shown, 

the SAN allows multiple servers to be attached to multiple tape drives. This 

provides the necessary physical attachment to support tape drive pooling. This 

configuration supports both library sharing and tape drive pooling. The library 

sharing and tape drive pooling capability are provided by Tivoli Storage Manager 

(TSM) Version 3.7 and higher. 

TSM LM 

Server 


T 

TSM 

Server 

TSM 

Server 

TSM 

Server 

SAN 

T 

T 

T 

3583 

Figure C-2 3583 Ultrium Scalable Tape Library sharing with SAN and TSM 

Using this approach, when a TSM server needs a tape, it communicates with the 

TSM Library Manager server via the LAN. The TSM Library Manager (LM) server 

handles all tape robotic requirements (mount, dismount) through the SCSI 

Medium Changer. The other TSM servers have no requirement to communicate 

with the SCSI Medium Changer, therefore the lack of multipathing capability is 

not an issue. Once the TSM LM Server has the tape mounted, the requesting 

TSM server performs its required tape I/O. TSM is providing the library sharing 

and tape drive pooling capability. TSM provides this capability for AIX, NT, 

Windows 2000, HP-UX and Solaris platforms. We discussed TSM library sharing 

in depth in 4.3, “TSM library sharing” on page 173. 


The servers do not have to be running the same operating system, but the 

application software needs to be the same. For instance, you could have TSM for 

AIX, TSM for NT, and TSM for Solaris all sharing a pool of tape drives. 

This TSM capability is only supported in a SAN environment. TSM library 

sharing supports a wide range of libraries today. The list of supported libraries 

can be found at: 

http://www.tivoli.com/support/storage_mgr/san/libsharing.html 

TSM support for library sharing with 358x drives/libraries is planned to be 

available in the near future. 

Sharing the 3583 tape library without a SAN 

Tape drive pooling requires a SAN configuration to provide the physical 

any-to-any attachment. But library sharing itself does not require a SAN. 

Figure C-3 shows this approach. 


T 

Server #1 

network 

Server #2 

Server #3 

T 

T 

Server #4 

T 

3583 

Figure C-3 3583 Ultrium Scalable Tape Library sharing without a SAN 

Here, no tape pooling is possible, because there is not an any-to-any connection 

between the servers and the tape drives. But the servers could be sharing the 

library/robotics assuming they were all communicating with server #1. ISV 

software might also provide this function (see 6.12.2, “Sharing autochangers 

between NetWorker hosts” on page 267 for Legato NetWorker sharing 

capabilities). 

Sharing the 3583 tape library using multiple initiators 

The third approach to having multiple RS/6000 or pSeries servers sharing a 3583 

is to use a multiple initiator configuration. This is shown is Figure C-4. 


terminator 

IBM pSeries 

(RS/6000) 

HACMP 

y-cable 

IBM pSeries 

(RS/6000) 

HACMP 

Standby 

terminator 

3583 

Figure C-4 3583 Ultrium Scalable Tape Library sharing with multiple initiators 

In this high availability (HA) configuration there are multiple initiators on one 

SCSI bus. Here both servers could see the SCSI target ID of the SCSI Media 

Changer. But in this configuration, only one server at a time can have any activity 

on the SCSI bus. It requires IBM’s HACMP (High Availability Cluster 

Multi-Processing) software. 

The above configuration will work (using SCSI Y-cables) if the 3583 is HVD. If the 

3583 is LVD, you need to use standard VHDCI to HD68 cables. (No Y-cables). 

The reason for this is when you use a Y-cable, you need to disable the built-in 

terminator in the SCSI adapter. This is easily accomplished with the pSeries or 

RS/6000 feature code 6204 and 6207 HVD SCSI adapters by removing the 

terminating resistors on the adapter card. Problems have been observed when 

disabling the built-in terminators on the LVD 6205 SCSI adapter card, therefore 

establishing the need to use standard SCSI cables rather than Y-cables. 

Sharing the 3584 Ultrium UltraScalable Tape Library 

Figure C-5 shows the multipath design of the 3584 tape library. Every drive can 

have a path defined to the SCSI Medium Changer (SMC). 


2 

FRAME CONTROLLER 


FRAME CONTROLLER 


AS/400s 

w BRMS 

Logical 

Library 1 

DRIVE 

1 

to Host 1 

DRIVE 

1 

to Host 1 

DRIVE 

AS/400 

DRIVE 

2 

Logical 

Library 2 

DRIVE 

3 

DRIVE 

to Host 2 

NT 

Logical 

Library 

1 

DRIVE 

3 

DRIVE 

4 

to Host 2 

4 

to Host 3 

Logical 

Library 3 

DRIVE 

5 

DRIVE 

6 

to Host 3 

RS/6000 

DRIVE 

5 

DRIVE 

6 

Multiple hosts 'owning' separate 

logical library inventories & drives 

Up to 12 

Hosts per 

Frame 

Multiple hosts sharing one 

logical library inventory 

Up to 72 

Hosts Per 

Library 

Figure C-5 3584 Ultrium UltraScalable Tape Library multipath design 

The library on the left has been partitioned into three logical libraries. In the AIX 

and NT partitions, only the first drive has a library control path defined. Most 

applications on these platforms are not enabled for multiple control path support. 

The iSeries (AS/400) is unique in that every SCSI bus has to have a library 

control path defined. The iSeries attached to the left hand library has two SCSI 

buses and, therefore, two library control paths defined. 

These three servers are sharing the 3584 Ultrium library. 

The library on the right has not been partitioned, and has only one logical library. 

Every iSeries SCSI bus has a library control path defined. The capability for 

iSeries attached to LTO tape drives (made generally available on November 17, 

2000) supports one drive per SCSI bus. In the near future support is planned for 

two LTO tape drives per iSeries SCSI bus. 

The iSeries SCSI buses are either 20MB/s SCSI2 Fast and Wide or 40MB/s Ultra 

SCSI. Given the high throughput capability of the LTO Ultrium tape drive, it does 

not make sense to put more than two LTO tape drives on an iSeries SCSI bus. 

The iSeries servers attached to the right hand library are also sharing the library 

even though the library has not been partitioned. In the case, BRMS is providing 

the library serialization. The iSeries Servers are sharing the library, but no tape 

drive pooling is being done. BRMS supports tape library sharing and a shared 

scratch pool, but not tape drive pooling. 


Configuring the 3584 with multiple logical libraries 

You can create multiple logical libraries that allow a single library to be used by 

multiple servers by partitioning the physical library’s storage slots and tape drives 

into two or more logical libraries. Each logical library consists of: 

► 

► 

► 

► 

Tape drives 

Storage slots 

Input/output (I/O) slots 

Cartridge accessor 

Each logical library has its own control path (a logical path into the library through 

which a server sends standard SCSI Medium Changer commands to control the 

logical library). A logical library cannot share another logical library’s tape drives 

and storage slots. However, it does share the I/O slots and the cartridge 

accessor on a first-come, first-served basis. 

In addition, a logical library cannot include mixed media. That is, you must 

configure LTO and DLT elements if present (that is, drives, storage slots, I/O 

slots, and grippers) into separate logical libraries. 

You can partition multiple logical libraries by using one of two methods: 

► Opening the door of the library and manually labeling the elements (storage 

slot columns and drives) that you want to include in each logical library (see 

“Configuring the 3584 tape library using labels” on page 298). This method 

enables you to view your partition at a glance whenever you open the front 

doors. 

► 

Identifying the number of elements that you want to include in each logical 

library, then using the 3584 StorWatch Specialist configuration wizard or the 

operator panel menus to create the logical libraries (see “Configuring the 

3584 tape library using menus” on page 304). This method makes it 

unnecessary for you to manually label the elements, but you cannot view your 

partition by opening the front doors. 

Configuring the 3584 tape library using labels 

First, plan the number and location of storage slot columns and tape drives that 

you want in each logical library. 

Open the front door of one or more frames, and on the drive side of the library 

attach a logical library bar code label to each tape drive and storage slot column 

that you defined. 


Attention: Ensure that the drives within each logical library are in contiguous 

locations. 

Close the front door of the library. 

Then to configure your library by using logical library bar code labels, and use 

one of the following methods. 

Using the StorWatch Specialist 

To use the 3584 StorWatch Specialist Web interface, perform the following: 

1. Enter the library’s Ethernet IP address on the URL line of the browser and 

press Enter. The introduction window displays (see Figure 2-12 on page 83). 

2. Click Settings, then Library Configuration. The Logical Libraries window 

displays (Figure C-6). 

Figure C-6 3854 StorWatch Specialist Logical Libraries entry panel 


3. Click the drop-down box and select the logical library that you want to 

configure. The window displays that library’s current configuration. 

4. Click the Configuration Wizard link and follow the instructions on the 

windows until the Select Configuration Method window displays. The 

window in Figure C-7 opens: 

Figure C-7 3854 StorWatch Specialist automated configuration wizard 

5. Click the Automated configuration button, then click Next. The library scans 

the bar code labels and creates the logical library. 

6. When the library has finished scanning the barcode labels, the window in 

Figure C-8 appears. 


Figure C-8 3854 StorWatch Specialist configuration review 

7. If the logical library’s configuration is correct, press Next and then Finish. The 

library is successfully configured as shown in Figure C-9. 


Figure C-9 3854 StorWatch Specialist configuration final window 

Using the Operator Panel 

1. From the library’s Activity panel, press MENU. The Main Menu displays 

(Figure C-10). 


Panel 0002 

Main Menu 

______________________________ 

Library Status 

Manual Operations 

Settings 

Usage Statistics 

Vital Product Data 

Service 

______________________________ 

BACK DOWN ENTER 

Figure C-10 3584 Operator Panel Main Menu 

2. Press UP or DOWN to highlight Settings, then press ENTER. The Settings 

menu displays. 

3. Press UP or DOWN to highlight Configuration and press ENTER. The 

Configuration menu displays. 

4. Press UP or DOWN to highlight Configure Library and press ENTER. The 

library displays the message: If you continue with configuration the library 

will go offline. Press ENTER to continue. 

5. Press ENTER twice. The library displays the message Searching for installed 

devices and may take approximately two minutes to discover the physical 

configuration. The Physical Configuration panel displays, with the library’s 

existing physical configuration: the panel shows the total quantity of drives, 

storage slots, and I/O slots in the library’s physical configuration. 

6. Press ENTER. The library displays the message Do you want to commit the 

new physical configuration? 

7. Press YES to accept the new physical configuration and to set up any logical 

library configurations. The library displays the message Searching for logical 

library labels, as it reads the labels that you previously attached. 

8. It then displays the Configuration Summary panel for Logical Library 1 

(Figure C-11). The panel contains the range of SCSI element addresses for 

the cartridge storage slots and the drives. 


Panel 0103 

Configuration Summary 

______________________________ 

Key:[F=Frame, C=Column, R=Row] 

Logical Library 1 

Storage Slots: 0064 

Elem Addr Range: 1025 - 1088 

Location Start: [F1,C01,R02] 

Location End: [F1,C03,R21] 

Drives: 001 

Elem Addr Range: 0257 - 0257 

Location Start: [F1,R01] 

Location End: 

[F1,R01] 

Control Path Drives: 

[F01,R01] 

______________________________ 

ENTER 

Figure C-11 3854 Operator Panel Configuration Summary 

9. Press ENTER to display the Configuration Summary panel for each logical 

library. After displaying the panel of the last logical library, the library displays 

the message Do you want to commit the new logical configuration? 

10.|Press YES to accept the new configuration (the library may take several 

minutes to process). When finished, it displays the message The 

configuration process is complete. 

Configuring the 3584 tape library using menus 

To configure your library by using menus, use one of the following methods. 

Using the StorWatch Specialist 

1. Enter the library’s Ethernet IP address on the URL line of the browser and 

press Enter. The introduction window displays (see Figure 2-12 on page 83). 

2. Select Settings, then Library Configuration to view the current logical 

library configuration. 

3. Click the Configuration Wizard link and follow the instructions on the 

windows until the Choose Configuration Method panel displays (Figure C-7 

on page 300). 

4. Click the Advanced configuration button and follow the instructions on the 

windows to specify one or more logical libraries and their attached physical 

devices (such as cartridge storage slots and drives). 


Note: If you have both DLT and LTO media in your physical library, you will 

have more than one configuration range and more than one selection in 

the windows that follow (a configuration range includes all available 

contiguous drives and a single media type). For example, you may first 

configure a range of DLT elements, then a range of LTO elements. 

5. Assuming you are configuring both LTO and DLT logical libraries, select the 

number of logical libraries for LTO (Figure C-12). 

Figure C-12 3854 StorWatch Specialist: using menus for LTO range 

6. Click Next. In the following panel (Figure C-13) you must allocate the number 

of drives and slots for each logical library defined for the LTO range. 


Figure C-13 3854 StorWatch Specialist: configure drives and slots for LTO range 

7. Repeat the same operation at number [6] and [7] for the DLT logical library’s 

range. 

8. When you are finished configuring the DLT range, the window in Figure C-14 

appears. 


Figure C-14 3854 StorWatch Specialist: reviewing the configuration 

9. If the logical libraries configuration is correct, press Next and then Finish. 

The library is successfully configured as shown in Figure C-9 on page 302. 

Using the Operator Panel 

1. From the library’s Activity window, press MENU. The Main Menu displays 

(Figure C-10 on page 303). 

2. Press UP or DOWN to highlight Settings, then press ENTER. The Settings 

menu displays. 

3. Press UP or DOWN to highlight Configuration and press ENTER. The 

Configuration menu displays. 

4. Press UP or DOWN to highlight Advanced configuration and press ENTER 

(Figure C-15). The library displays the message: If you continue with 

configuration the library will go offline. Press ENTER to continue. 


Panel 0101 

Configuration 

______________________________ 

Display Configuration 

Configure Library 

Advanced Configuration 

______________________________ 

BACK DOWN ENTER 

Figure C-15 3584 Operator Panel: configuration 

5. Press ENTER twice. The library displays the message Searching for installed 

devices and may take approximately two minutes to discover the physical 

configuration. The Physical Configuration panel displays, with the library’s 

existing physical configuration: the panel shows the total quantity of drives, 

storage slots, and I/O slots in the library’s physical configuration. 

6. Press ENTER. The library displays the message Do you want to commit the 

new physical configuration? 

7. Press YES to accept the new physical configuration and to set up any logical 

library configurations. The Set Logical Libraries panel displays with the type 

of media used by the logical library (Figure C-16). 


Panel 0105 

Set Logical Libraries 

______________________________ 

Media Type : LTO 

Frames 1 - 1 

Number of Logical Libraries 1 

Select 1 - 12 

______________________________ 

CANCEL UP ENTER 

Figure C-16 3584 Operator Panel: set logical libraries 

8. Specify the quantity of logical libraries that you want by pressing UP or 

DOWN to increment or decrement the value. 

9. When the desired quantity of libraries displays, press ENTER. The Set 

Storage Slots panel displays (Figure C-17). 

Panel 0106 

Set Storage Slots 

______________________________ 


Storage Slots 139 

Select 1 - 140 

Cartridges in the following 

storage slots will now be 

part of logical library 2 

Location Start: 

Location End: 

[F1,C07,R43] 

[F1,C07,R42] 

______________________________ 


Figure C-17 3584 Operator Panel: set storage slots 

10.specify the quantity of storage slots that you want in the logical library by 

pressing UP or DOWN to increment or decrement the value. When the 


desired quantity of storage slots displays, press ENTER. The Set Drives 

panel displays (Figure C-18). 

Panel 0107 

Set Drives 

______________________________ 


Nunber of Drives 1 

Select 1 - 12 

______________________________ 


Figure C-18 3584 Operator Panel: set drives 

11.Specify the quantity of drives that you want in the logical library by pressing 

UP or DOWN to increment or decrement the value. When the desired quantity 

of drives displays, press ENTER. 

12.It then displays the Configuration Summary panel for Logical Library 1 

(Figure C-11 on page 304). The panel contains the range of SCSI element 

addresses for the cartridge storage slots and the drives. 

13.Press ENTER to display the Configuration Summary panel for each logical 

library. After displaying the panel of the last logical library, the library displays 

the message Do you want to commit the new logical configuration? 

14.|Press YES to accept the new configuration (the library may take several 

minutes to process). When finished, it displays the message The 

configuration process is complete. 

Summary 

Figure C-19 summarizes the Ultrium Tape Library/Robotic Sharing 

considerations. 


Tape Library / Robotic Sharing 

Single Path (3583) 

Hom ogeneous Appl 

- TS M (A IX , N T, S un ) 

or 

-ISV 

Single Logical Library 

Hom ogeneous Appl 

- T S M (A IX , N T, S U N ) 

or 

- B R M S (O S /4 00 ) 

or 

- IS V 

M ultiPath (3584, 3575) 

M u ltip le L o g ic a l L ib ra rie s 

Heterogeneous Appl 

- TSM (Any Platform ) 

and/or 

- B R M S (O S/400) 

and/or 

- ISV 

Figure C-19 Summary of the IBM LTO Tape Libraries sharing capabilities 

Tape libraries can be either single path or multipath. The 3583 is a single path 

tape library. Single path tape libraries require external software to allow library 

sharing. Tivoli Storage Manager (TSM) is an example application that would 

provide this functionality. TSM versions 3.7 and above provides both tape library 

sharing and tape drive pooling. Because tape drive pooling requires a SAN, TSM 

tape library sharing only works in a SAN environment. TSM plans to support the 

358x family of tape libraries in the near future. This homogeneous application 

can support heterogeneous platforms. 

Multipath libraries can be configured as a single logical library or as multiple 

logical libraries. When configured as a single logical library, the multipath library 

can do everything a single path library can plus provide IBM iSeries (AS/400) 

library sharing. 

When configured with multiple logical libraries, a multipath library can all do all of 

the above plus support the sharing of heterogeneous applications. For example, 

a 3584 configured as multiple logical partitions could support TSM, BRMS, 

Legato NetWorker, and VERITAS NetBackup DataCenter all sharing the 3584 

library and robotics. 



Related publications 

The publications listed in this section are considered particularly suitable for a 

more detailed discussion of the topics covered in this redbook. 

IBM Redbooks 

Other resources 

For information on ordering these publications, see “How to get IBM Redbooks” 

on page 315. 

► 

► 

► 

► 

► 

► 

► 

► 

► 

Designing an IBM Storage Area Network, SG24-5758 

Implementing an Open IBM SAN, SG24-6116 

IBM SAN Survival Guide, SG24-6143 


The IBM LTO Ultrium Tape Libraries Guide, SG24-5946 

IBM SAN Survival Guide Featuring the IBM 2109, SG24-6127 

Tivoli NetView 6.01 and Friends, SG24-6019 

Tivoli Storage Management Concepts, SG24-4877 

Getting Started with Tivoli Storage Manager: Implementation Guide, 

SG24-5416 

These publications are also relevant as further information sources: 

► 

► 

► 

► 

► 

► 

► 

3584 UltraScalable Tape Library Planning and Operator Guide, GA32-0408 

3583 Ultrium Scalable Tape Library Setup and Operator Guide, GA32-0411 

3581 Ultrium Tape Autoloader Setup, Operator and Service Guide, 

GA32-0412 

3580 Ultrium Tape Drive Setup, Operator and Service Guide, GA32-0415 

IBM Ultrium Device Drivers Installation and User’s Guide, GA32-0430 

IBM Storage Area Network Data Gateway Module Setup, Operator and 

Service Guide, GA32-0436 

TSM for AIX Managed System for SAN Storage Agent User’s Guide, 

GC36-0001 


► 

► 

► 

► 

► 

► 

TSM for Sun Solaris Managed System for SAN Storage Agent User's Guide, 

GC36-0002 

IBM SAN Fibre Channel Switch, 2109 Model S08 User's Guide, SC26-7349 

IBM SAN Fibre Channel Switch 2109 Model S16 User's Guide, SC26-7351 

IBM Storage Area Network Data Gateway Installation and User's Guide, 

SC26-7304 

IBM SAN Data Gateway Router 2108 Model R03 Installation and User's 

Guide, SC26-7355 

Tivoli Storage Manager for AIX Quick Start Version 4 Release 2, GC35-0402 

► Tivoli Storage Manager for Sun Solaris Quick Start Version 4 Release 2, 

GC35-0406 

► Tivoli Storage Manager for HP-UX Quick Start version 4 release 2, 

GC35-0412 

Referenced Web sites 

These Web sites are also relevant as further information sources: 

► LTO Organization homepage 

http://www.lto.org 

► IBM Ultrium products homepage 

http://www.storage.ibm.com/hardsoft/tape/lto/index.html 

► IBM Storage Area Networks homepage 

http://www.storage.ibm.com/san 

► QLogic homepage 

http://www.qlogic.com 

► Emulex homepage 

http://www.emulex.com 

► AIX documentation 

http://techsupport.services.ibm.com/server/library 

► Solaris documentation 

http://docs.sun.com/ 

► HP-UX documentation 

http://docs.hp.com/ 

► Tivoli Storage Manager homepage 


► 

► 

http://www.tivoli.com/tsm 

VERITAS homepage 

http://www.veritas.com/ 

Legato homepage 

http://www.legato.com/ 

How to get IBM Redbooks 

Search for additional Redbooks or Redpieces, view, download, or order hardcopy 

from the Redbooks Web site: 


Also download additional materials (code samples or diskette/CD-ROM images) 

from this Redbooks site. 

Redpieces are Redbooks in progress; not all Redbooks become Redpieces and 

sometimes just a few chapters will be published this way. The intent is to get the 

information out much quicker than the formal publishing process allows. 

IBM Redbooks collections 

Redbooks are also available on CD-ROMs. Click the CD-ROMs button on the 

Redbooks Web site for information about all the CD-ROMs offered, as well as 

updates and formats. 

Related publications 315


Special notices 

References in this publication to IBM products, programs or services do not imply 

that IBM intends to make these available in all countries in which IBM operates. 

Any reference to an IBM product, program, or service is not intended to state or 

imply that only IBM's product, program, or service may be used. Any functionally 

equivalent program that does not infringe any of IBM's intellectual property rights 

may be used instead of the IBM product, program or service. 

Information in this book was developed in conjunction with use of the equipment 

specified, and is limited in application to those specific hardware and software 

products and levels. 

IBM may have patents or pending patent applications covering subject matter in 

this document. The furnishing of this document does not give you any license to 

these patents. You can send license inquiries, in writing, to the IBM Director of 

Licensing, IBM Corporation, North Castle Drive, Armonk, NY 10504-1785. 

Licensees of this program who wish to have information about it for the purpose 

of enabling: (i) the exchange of information between independently created 

programs and other programs (including this one) and (ii) the mutual use of the 

information which has been exchanged, should contact IBM Corporation, Dept. 

600A, Mail Drop 1329, Somers, NY 10589 USA. 

Such information may be available, subject to appropriate terms and conditions, 

including in some cases, payment of a fee. 

The information contained in this document has not been submitted to any formal 

IBM test and is distributed AS IS. The use of this information or the 

implementation of any of these techniques is a customer responsibility and 

depends on the customer's ability to evaluate and integrate them into the 

customer's operational environment. While each item may have been reviewed 

by IBM for accuracy in a specific situation, there is no guarantee that the same or 

similar results will be obtained elsewhere. Customers attempting to adapt these 

techniques to their own environments do so at their own risk. 

Any pointers in this publication to external Web sites are provided for 

convenience only and do not in any manner serve as an endorsement of these 

Web sites. 


The following terms are trademarks of other companies: 

Tivoli, Manage. Anything. Anywhere.,The Power To Manage., Anything. 

Anywhere.,TME, NetView, Cross-Site, Tivoli Ready, Tivoli Certified, Planet Tivoli, 

and Tivoli Enterprise are trademarks or registered trademarks of Tivoli Systems 

Inc., an IBM company, in the United States, other countries, or both. In 

Denmark, Tivoli is a trademark licensed from Kjøbenhavns Sommer - Tivoli A/S. 

C-bus is a trademark of Corollary, Inc. in the United States and/or other 

countries. 

Java and all Java-based trademarks and logos are trademarks or registered 

trademarks of Sun Microsystems, Inc. in the United States and/or other 

countries. 

Microsoft, Windows, Windows NT, and the Windows logo are trademarks of 

Microsoft Corporation in the United States and/or other countries. 

PC Direct is a trademark of Ziff Communications Company in the United States 

and/or other countries and is used by IBM Corporation under license. 

ActionMedia, LANDesk, MMX, Pentium and ProShare are trademarks of Intel 

Corporation in the United States and/or other countries. 

UNIX is a registered trademark in the United States and other countries licensed 

exclusively through The Open Group. 

SET, SET Secure Electronic Transaction, and the SET Logo are trademarks 

owned by SET Secure Electronic Transaction LLC. 

Other company, product, and service names may be trademarks or service 

marks of others. 


Abbreviations and acronyms 

AIX 

BRMS 

BSD 

CDE 

DDS 

DHCP 

DLT 

ESRM 

FC 

FC-AL 

FTP 

GA 

GBIC 

HA 

HACMP 

HBA 

HD68 

HP 

HVD 

I/O 

Advanced Interactive 

Executive 

Backup, Recovery and Media 

Services 

Berkeley Software Distribution 

Common Desktop 

Environment 

Dynamic Drive Sharing 

Dynamic Host Configuration 

Protocol 

Digital Linear Tape 

Enterprise Storage Resource 

Management 

Fibre Channel 

Fibre Channel Arbitrated Loop 

File Transfer Protocol 

Generally Available 

Gigabit Interface Converter 

High Availability 

High Availability Cluster 

Multi-Processing 

Host Bus Adapter 

High Density 68-pin 

Hewlett-Packard 

High voltage Differential 

Input/Output 


LTO-DC 

LUN 

LVD 

MAC 

MB/s 

MIB 

OID 

PTF 

RFC 

RISC 

RMU 

ROM 

SAM 

SAN 

SCSI 

SDG 

SMC 

SMIT 

SNMP 

Linear Tape-Open 

LTO Data Compression 

Logical Unit Number 

Low Voltage Differential 

Media Access Control 

Megabytes per second 

Management Information 

Base 

Object Identifier 

Program Temporary Fix 

Request for Comment 

Reduced Instruction Set 

Computer 

Remote Management Unit 

Read Only Memory 

System Administration 

Manager 

Storage Area Network 

Small Computer Systems 

Interface 

SAN Data Gateway 

SCSI Medium Changer 

System Management 

Interface Tool 

Simple Network Management 

Protocol 


ISV 

ITSO 

LAN 

LM 

LPP 

International Business 

Machines Corporation 

Independent Software Vendor 

International Technical 

Support Organization 

Local Area Network 

Library Manager 

Licensed Program Product 

SNS 

SSO 

SSO 

SUN 

TB 

TCP/IP 

Simple Name Server 

Shared Storage Option 

Shared Storage Option 

Stanford University 

Networking 

terabytes 

Transmission Control 

Protocol/Internet Protocol 


TSM 

VHDCI 

VPN 

VPS 

WWN 

WWNN 

WWPN 

Tivoli Storage Manager 

Very High Density Cable 

Interconnect 

Virtual Private Network 

Virtual Private SAN 


World Wide Node Name 

World Wide Port Name 


Index 

Numerics 

3584 base library 13 

3584 tape library using labels 298 

3584 tape library using menus 304 

3584 with multiple logical libraries 298 

A 

adding a new LTO device 263 

AIX 

device attachment test 274 

FC adapter filesets 61 

FC device filesets 61 

get HBA WWN 105 

install SDG StorWatch Specialist 129 

LTO with TSM 154 

setting device names 124 

AIX commands 271 

271 

cfgmgr 99, 111, 124, 127 

dd 289 

lsattr 123 

lscfg 105, 143 

lsdev 99, 123, 154 

rmdev 111 

alias 104, 107 

Atape device driver 31 

Atape driver installation 33 

Atdd driver installation 49 

auto media management 268 

autochanger configuration (AIX) 250 

autochanger configuration (HP-UX) 257 

autochanger configuration (SUN) 253 

autochanger maintenance commands 273 

B 

bootcode 75 

booting FC hosts 287 

booting SAN Data Gateway 287 

BRMS 297 

C 

cartridge labeling 233 

cartridges 

labeling 168 

cfgmgr 99, 111, 124, 127 

channel zoning 125, 141 

check daemons have started 202 

cleaning tape drives 171 

common commands 271 

compression 31, 94, 154, 159, 162, 231, 243, 256, 

260 

configure non-shared tape library 154 

configuring non-shared LTO 152 

configuring SNMP 281 

configuring tape and medium changer devices 35, 

43 

configuring tape devices 55 

configuring the Library Client 176 

configuring the Library Manager 175 

connecting tape drives to a SDG 97 

D 

daisy-chaining 24, 94, 98, 293 

data flow 25 

define and apply zone configuration 109 

define backup policy 218 

define device class 165 

define storage pool 167 

deleting LTO devices 47 

deleting LTO medium changer 56 

detect tape drives 203 

determine how many SCSI drives 23 

device attachment with tapeutil 274 

device configuration wizard 206 

device driver 25, 194 

disableCC 127 

DLT 15 

DLT expansion frame 3584-D42 15 

dmesg 106, 114, 122 

drives 

blocksize 154, 159 

data transfer rate 94 

device special names 30 

element numbers 143, 163 

multiple paths 98 


no rewind device 31 

serial numbers 143 

setting device names with AIX 124 

tape pooling 292 

upgrade firmware 90 

drvconfig 115 

dynamic drive sharing (DDS) 265 

E 

element numbers 143, 163 

emacs 272 

Emulex HBA 

bootcode 75 

configuration 73 

deinstall driver 73 

enabling software zoning 103 

firmware 75 

install device driver 69 

persistent binding 121 

Emulex HBA driver configuration 73 

Emulex HBA driver installation 69 

Emulex HBA driver removal 73 

Emulex HBA firmware and bootcode 75 

Emulex HBA on Solaris 121 

ethAddrSet 127 

expansion frame 14 

F 

FC-AL 94 

FCode 68 

fcShowDevs 128, 139 

fcShowNames 106, 128 

firmware 75 

drive 90 

IBM SAN Data Gateway 131 

library 89 

RMU 91 

upgrade 88 

G 

GBIC 12 

H 

hardware zoning 102 

HBA 24 

HD68 27, 28 

Hewlett-Packard 4 

HP-UX 



HP-UX commands 

swinstall 247 

HVD 10, 12, 26 

HVD model 9 

I 

IBM 2109 Fibre Channel Switch 

create alias 107 

define zone 109 

StorWatch Specialist 95, 103 

zoning 103 

IBM 3580 Ultrium Tape Drive 8 

IBM 3581 Ultrium Tape Autoloader 9 

IBM 3583 Ultrium Scalable Tape Library 10, 88 

configure SNMP 281 

I/O station 11 

library sharing 292 

RMU 11, 76, 280 

SAN Data Gateway Integrated Module 11, 23, 

293 

StorWatch Specialist 11, 13, 76 


IBM 3584 Ultrium UltraScalable Tape Library 12 

configure SNMP 283 

Ethernet attachment 13, 81 

expansion frame 14 

I/O station 13 


logical libraries 298 

multiple control paths 189 


IBM SAN Data Gateway 8, 21, 125 

data transfer rate 97 

fcShowNames 106 

get WWN 106 

install firmware 131 

install StorWatch Specialist 129 

rebooting 287 

setup 126 


IBM SAN Data Gateway commands 

disableCC 127 

ethAddrSet 127 

fcShowDevs 128, 139 

fcShowNames 128 


initializeBox 127 

reboot 128, 288 

scsiRescan 128, 288 

setHost 127 

userAdd 127 

IBM SAN Data Gateway Router 

rebooting 287 

implementation in a Solaris environment 253 

implementation in AIX environment 250 

implementation in HP-UX environment 257 

importing media 213 

initializeBox 127 

inserting data and cleaner cartridges 168 

inserting unlabeled data cartridges 170 

install base code 195 

install patch 201 

install StorWatch SDG Specialist 129 

installation in a HP-UX environment 246 

installation in a Solaris environment 240 

installation in an AIX environment 237 

installation using the SMIT interface 34 

installation verification 261 

installing library and device drivers 30 

installing NetBackup DataCenter 193 

installing Tivoli Storage Manage 153 

Integrated SAN Data Gateway Module 11, 23, 293 

inventorying volumes with barcode 268 

iSeries 297 

BRMS 297 

LTO library sharing 293 

ISV software 18 

K 

Korn shell 272 

L 

labeling data cartridges 168 

labeling volumes with barcode 268 

LAN-free data transfer 152, 181 

Legato NetWorker 236 

auto media management 268 

autochanger sharing 267 

cleaning tape drives 252, 255, 259 

client 236 

compression 243, 256, 260 

configure jukebox 251, 255, 258 

daemons 240, 245, 249 

drive sharing 265 

dynamic drive sharing 265 

installation 237 

inventorying libraries 268 

labeling cartridges 269 

License Manager 236 

server 236 

storage node 236 

Legato NetWorker commands 273 

inquire 250, 253, 257 

inventory 269 

jbconfig 255, 258, 267 

nsr_ize 238 

nwadmin 261 

Legato NetWorker exploitation 265 

Legato NetWorker implementation 249 

Legato NetWorker use 249 

library partitioning 292 


Legato NetWorker 265 

tape drive pooling 294 

TSM 173, 294 

Linear Tape Open initiative 4 

logical libraries 292 

logical library 298 

lpfc.conf 73, 121 

lputil 75 

lsattr 123 

lscfg 105, 143 

lsdev 99, 123, 154 

LTO 4, 29 

3580 see IBM 3580 Ultrium Tape Drive 

3581 see IBM 3581 Ultrium Tape Autoloader 

3583 see IBM 3583 Ultrium Scalable Tape Library 

3584 see IBM 3584 Ultrium UltraScalable Tape 

Library 

buying media 168 

duplicate drive definitions 99 


operating system device names 142 

LTO driver installation for AIX 31 

LTO driver installation for Solaris 40 

LTO expansion frame 3584-D32 14 

LTO medium changer configuration 50 

LTO product announcements 5 

LTO Ultrium 4 

LTO-DC 13 

LUN masking 142 

LVD 9, 12, 26 

Index 323

LVD model 9 

M 

mapped drive considerations 184 

media suppliers 168 

medium changer device attachment test 274 

multipathing 98, 292 

multiple paths to tape drives 98 

multiplexing 217, 226 

P 

performance considerations 288 

performance hints for TSM and LTO 172 

persistent binding 99, 112 

Emulex HBA 121 

QLogic HBA 118 

Q 

qla2200.conf 67, 118 

QLogic HBA 


deinstall driver 68 

FCode 68 

persistent binding 118 

QLogic HBA driver installation 64 

QLogic HBA driver removal 68 

QLogic HBA FCode 68 

QLogic HBA on Solaris 118 

R 

reboot 128 

Redbooks Web site 315 

Contact us 18 

redundant library control paths 189 

Remote Management Unit (RMU) 11 

removing Atape driver 39 

rmdev 111 

RMU 76, 280 


RS/6000 Fibre Channel HBA installation 61 

RS/6000 HBA microcode installation 62 

S 

SAN 193 

channel zoning 141 

data transfer rate 94, 95 

performance statistics 96 

zoning 99, 100 

scan for tape devices 206 

SCSI cable connectors 27 

SCSI direct attach 20 

SCSI Medium Changer 98 

SCSI target IDs 121 

scsiRescan 128, 288 

Seagate 4 

serial numbers 143 

setHost 127 

sharing autochangers 267 

Simple Name Server 102 

SNMP 11, 280 

SNMP alerts 280 

software installation (AIX) 238 

software installation (HP-UX) 246 

software installation (Solaris) 241 

software zoning 102 

Solaris 


get HBA WWN 106 

HBA configuration 67, 73 



Solaris commands 272 

dd 289 

dmesg 45, 106, 114, 122, 195 

drvconfig 115 

lputil 75 

pkgadd 241 

special device file name 30 

starting NetBackup DataCenter daemons 203 

stopping NetBackup DataCenter daemons 203 

storage units 217 

StorWatch 13 


3583 76 

3584 82 

IBM SAN Data Gateway 288 

StorWatch Tape Library Specialist 75 

T 

tape drive pooling 292, 294 

tape libraries attached to multiple servers 292 

tape library commands 273 

tape library partitioning 291 

tape library sharing 291 

tape library using multiple initiators 295 


tape library using SAN and TSM 294 

tape library without a SAN 295 

tape topics 231 

tapeutil 

device attachment tests 274 

element inventory 276 

upgrade drive firmware 90 

upgrade library firmware 89 

tapeutil element inventory 276 

test environment 193 

test the configuration 227 

Tivoli Storage Manager see TSM 

TSM 

cleaning tape drives 157, 161, 164 

client 150 

define device class 165 

define storage pool 167 

device classes 165 

dsm.opt 173 

dsm.sys 173 

dsmserv.opt 172 

insert cleaner cartridge 171 

insert data cartridges 169 


labeling cartridges 168 

library client 151, 174 

library manager 151 

library manager server 174 

library sharing 150, 173, 294 

performance hints 172 

server 149 

server-to-server communications 151 

storage pools 167 

tape library 150 

TXNBytelimit 173 

TXNGroupmax 172 

unlabeled cartridges 170 

Version 4.2 149 

TSM commands 273 

CHECKIN LIBVOLUME 171 

DEFINE DEVCLASS 165 

DEFINE DRIVE 155, 158, 159, 161, 163, 164 

DEFINE LIBRARY 155, 159, 162 

DEFINE STGPOOL 167 

LABEL LIBVOLUME 168, 169, 170 

QUERY LIBVOLUME 169 

REPLY 169 

SET CROSSDEFINE 175, 176 

SET SERVERHLADDRESS 175, 176 

SET SERVERLLADDRESS 175, 176 

SET SERVERNAME 175, 176 

SET SERVERPASSWORD 175, 176 

UPDATE LIBRARY 175 

TSM commonly used terms 149 

U 

ultrium 4, 165 

ultriumc 165 


use StorWatch SDG Specialist 129 

useful commands 271 

userAdd 127 

using the StorWatch Specialist 299 

V 

verifying the Atape driver installation 36 

verifying the Atdd driver installation 54 

verifying the IBMtape driver installation 46 

VERITAS NetBackup DataCenter 192 

add client names 220 

backup class 218 

backup policy 218 

Backup Policy Configuration Wizard 219 

backup types 222 

client agents 192 


daemons 202 

Device Configuration Wizard 210 

Global Data Manager 192 


inventory robot 213 

Java administration client 206 

manual backup 227 

master server 192 

media server 192 

multiplexing 217, 226 

pool 217 

rotation schedule 223 

SCSI passththrough driver 194 

Shared Storage Option (SSO) 193 

storage units 212, 217, 226 

volume configuration 215 

Volume Group 217 

VERITAS NetBackup DataCenter commands 

bp.kill_all 203 

bpps 202 

initbprd 203 

Index 325

ltid 203 

robtest 233 

s77betbackup 203 

sgscan 206 

tldtest 233 

VHDCI 11, 27, 28 

VHDCI cable connectors 27 

Virtual Private Network 99 

Virtual Private SAN 125, 142 

W 

Windows 


WWN 

of HBA on AIX 105 

of HBA on Solaris 106 

of SDG 106 

WWNN 102, 119, 121 

WWPN 102, 121 

Z 

zoning 99, 100 

alias 104, 107 

defining 109 

SNS 102 

software 102 

zoning example 111 


Using IBM LTO Ultrium with Open Systems 

(0.5” spine) 

0.475”0.875” 

250 459 pages

Back cover 

® 

Using IBM LTO Ultrium 

with Open Systems 

How to setup Ultrium 

tape drives in 

multiple open 

environments 

Integration with 

popular backup 

applications 

Direct SCSI and SAN 

attachment 

This IBM Redbook is a follow-on to The IBM Ultrium Tape 

Libraries Guide, SG24-5946, and will help you plan, install 

and configure IBM Ultrium LTO tape drives and libraries in 

UNIX environments. It focuses on the setup and customization 

of these drives and libraries in both direct-attached SCSI and 

SAN configurations for the AIX, Solaris and HP-UX operating 

systems. 

The first part of the book focuses on understanding how to 

attach and configure the drives and libraries and covers basic 

installation and administration. The second part documents 

how to use these products with popular data backup 

applications — Tivoli Storage Manager, VERITAS NetBackup 

DataCenter and Legato NetWorker. 

This redbook will help IBM personnel, Business Partners and 

customers to better understand and implement the IBM 

Ultrium LTO product line in UNIX environments. 

INTERNATIONAL 

TECHNICAL 

SUPPORT 

ORGANIZATION 

BUILDING TECHNICAL 

INFORMATION BASED ON 

PRACTICAL EXPERIENCE 

IBM Redbooks are developed by 

the IBM International Technical 

Support Organization. Experts 

from IBM, Customers and 

Partners from around the world 

create timely technical 

information based on realistic 

scenarios. Specific 

recommendations are provided 

to help you implement IT 

solutions more effectively in 

your environment. 

For more information: 


SG24-6502-00 ISBN 0738423599

Using IBM LTO Ultrium with Open Systems - RS/6000 Home

Create successful ePaper yourself

Delete template?

Save as template?