Version 1 Release 3.0

Document Number GC26-7442-00

Note
Before using this information and the product it supports, read the information in Notices.

Ninth Edition (September 2001)

This edition applies to the IBM ESS Subsystem Device Driver 1.3.0.x and to all subsequent releases and modifications until otherwise indicated in new editions.

This edition also includes information that specifically applies to:

• AIX 4.2.1, AIX 4.3.2, AIX 4.3.3, AIX 5.1.0
• Windows NT 4.0 Service Pack 3 or higher
• Windows 2000 Service Pack 2 or higher
• HP-UX 11.00
• Solaris 2.6, Solaris 7, Solaris 8

Order publications through your IBM representative or the IBM branch office serving your locality.

© Copyright International Business Machines Corporation 1999, 2001. All rights reserved.
U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.

Contents

Figures

Tables

About this book

Chapter 1. Introducing the Subsystem Device Driver

Chapter 2. Installing and configuring SDD on an AIX host system

Chapter 3. Using SDD on an AIX host system

Chapter 4. Installing and configuring SDD on a Windows NT host system

Chapter 5. Installing and configuring SDD on a Windows 2000 host system

Chapter 6. Installing and configuring SDD on an HP host system

Chapter 7. Installing and configuring SDD on a Sun host system

Chapter 8. Using the datapath commands

Statement of Limited Warranty

Notices

Glossary

Index

Figures

1. Multipath connections between a host system and the disk storage in an ESS
2. Where SDD fits in the protocol stack
3. Output from the Display Data Path Device Configuration SMIT panel
4. Where SDD fits in the protocol stack
5. Where the SDD fits in the protocol stack
6. Where SDD fits in the protocol stack
7. IBMdpo Driver 32-bit
8. IBMdpo Driver 64-bit
9. Where SDD fits in the protocol stack

Tables

1. Required number of successful I/O operations before SDD placing a path in the Open state
2. Support for 32-bit and 64-bit applications
3. AIX PTF required fixes
4. SDD package file names
5. Major files included in the SDD installation package
6. Software support for HACMP/6000 in concurrent mode
7. Software support for HAMCP/6000 in non-concurrent mode
8. Software support for HACMP/6000 in concurrent mode on AIX 5.1.0 (32-bit kernel only)
9. Software support for HACMP/6000 in non-concurrent mode on AIX 5.1.0 (32-bit kernel only)
10. HACMP/6000 and supported SDD features
11. SDD-specific SMIT panels and how to proceed
12. SDD installation scenarios
13. HP patches necessary for proper operation of SDD
14. SDD components installed
15. System files updated
16. SDD commands and their descriptions
17. SDD installation scenarios
18. SDD package file names
19. Solaris patches necessary for proper operation of SDD
20. System files updated
21. Subsystem Device Driver components installed
22. SDD commands and their descriptions
23. Commands

About this book

This books provides step-by-step procedures for you to install, configure, and use the IBM^(R) TotalStorage^(TM) Enterprise Storage Server^(TM) Subsystem Device Driver on IBM AIX^(R), HP, Sun, Microsoft^(R) Windows NT^(R), and Microsoft^(R) Windows 2000 host systems.

Who should use this book

This book is intended for storage administrators, system programmers, and performance and capacity analysts.

Summary of changes

This book contains both information previously presented in IBM TotalStorage Enterprise Storage Server Subsystem Device Driver Installation and User's Guide Version 1 Release 2.1 (June 2001) and major technical changes to that information. Technical changes are indicated by revision bars (|) in the left margin of the book. The following sections summarize those changes.

Note:

For the last-minute changes that are not included in this book, see the README file on the SDD compact disc or visit the SDD website at:

http://www.ibm.com/storage/support/techsup/swtechsup.nsf/support/sddupdates

New information

This edition includes the following new information:

What's new in Chapter 2, Installing and configuring SDD on an AIX host system:

• The reference to the SDD 1.3.0.x support for 32-bit and 64-bit applications on AIX 4.3.3 and AIX 5.1.0 in appropriate sections.
• Reference to the SDD 1.3.0.x support for AIX 5.1.0 in appropriate sections.
• Addition of the ibmSdd_510.rte and ibmSdd_510nchacmp.rte filesets in appropriate sections.
• Addition of the addpaths file definition in Table 5.
• Reference to the addpaths command in appropriate sections.
• Procedure for Adding paths to SDD devices of a volume group.
• Reference to the SDD 1.3.0.x support for non-disruptive installation in Upgrading SDD for AIX 4.2.1, AIX 4.3.2 and AIX 4.3.3.

What's new in Chapter 3, Using SDD on an AIX host system

• Reference to the SDD support for 32-bit and 64-bit applications on AIX 4.3.3 and AIX 5.1.0 in appropriate sections.
• Reference to the SDD support for AIX 5.1.0 in appropriate sections.
• Addition of the ibmSdd_510.rte and ibmSdd_510nchacmp.rte filesets in appropriate sections.
• Reference to addpaths command in appropriate sections.
• A new section about Accessing the Add Paths to Available Data Path Devices SMIT panel.

What's new in Chapter 4, Installing and configuring SDD on a Windows NT host system

• Procedure for Configuring a Windows NT cluster with SDD.

What's new in Chapter 5, Installing and configuring SDD on a Windows 2000 host system

• Procedure for Upgrading the Subsystem Device Driver.
• Reference to the SDD 1.3.0.x Support for Windows 2000 clustering.
• Procedure for Preparing to Configure a Windows 2000 cluster with SDD.
• Procedure for Configuring a Windows 2000 cluster with SDD.

Modified information

This edition includes the following modified information:

What's modified in Chapter 2, Installing and configuring SDD on an AIX host system

• The SDD version release levels for AIX are updated as follows:
  • SDD 1.2.2.0 to SDD 1.3.0.x

What's modified in Chapter 3, Using SDD on an AIX host system

• The SDD version release levels for AIX are updated as follows:
  • SDD 1.2.2.0 to SDD 1.3.0.x

Publications

This section describes the IBM TotalStorage ESS publication library, IBM related products publications, and other related products publications. It also gives ordering information for these publications.

The IBM TotalStorage ESS publication library

See the following publications for more information about the ESS:

• IBM TotalStorage Enterprise Storage Server Introduction and Planning Guide, GC26-7294

  This guide introduces the product and lists the features you can order. It also provides guidelines on planning for the installation and configuration of the ESS.

• IBM TotalStorage Enterprise Storage Server User's Guide, SC26-7295

  This guide provides instructions for setting up and operating the ESS and for analyzing problems.

• IBM TotalStorage Enterprise Storage Server Host System Attachment Guide, SC26-7296

  This guide provides guidelines for attaching the ESS to your host system and for migrating from SCSI to fibre-channel attachment.

• IBM TotalStorage Enterprise Storage Server SCSI Command Reference, SC26-7297

  This book describes the functions of the ESS. It provides reference information for UNIX^(R), Application System/400^(R) (AS/400^(R)), and iSeries 400 hosts, such as channel commands, sense bytes, and error recovery procedures.

• IBM Enterprise Storage Server System/390 Command Reference, SC26-7298

  This book describes the functions of the ESS and provides reference information for S/390^(R) and zSeries hosts, such as channel commands, sense bytes, and error recovery procedures.

• IBM Storage Solutions Safety Notices, GC26-7229

  This book provides translations of the danger notices and caution notices that IBM uses in ESS publications.

• IBM TotalStorage Enterprise Storage Server Web Interface User's Guide, SC26-7346

  This guide provides instructions for using the IBM TotalStorage^(TM) Enterprise Storage Server Web interfaces, ESS Specialist and ESS Web Copy Services.

• IBM TotalStorage Enterprise Storage Server Configuration Planner, SC26-7353

  This guide provides work sheets for planning the logical configuration of the ESS. This book is not available in hardcopy. This guide is available on the following Web site:

  www.storage.ibm.com/hardsoft/products/ess/refinfo.htm

• IBM TotalStorage Enterprise Storage Server Quick Configuration Guide, SC26-7354

  This booklet provides flow charts for using the StorWatch Enterprise Storage Server Specialist. The flow charts provide a high-level view of the tasks the IBM service support representative performs during initial logical configuration. You can also use the flow charts for tasks that you might perform when you are modifying the logical configuration.

  The hardcopy of this booklet is a 9-inch × 4-inch fanfold.

• IBM TotalStorage Enterprise Storage Server Copy Services Command-Line Interface Reference, SC26-7434

  This reference describes the commands you can use from the ESS Web Copy Services Command-Line Interface (CLI). The CLI application provides a set of commands that allow you to write customized scripts for a host system that initiates pre-defined tasks in an ESS Web Copy Services server application. You can use the CLI commands to indirectly control ESS peer-to-peer Remote Copy and FlashCopy configuration tasks within an ESS Web Copy Services server group.

  This book is not available in hardcopy. It is available in PDF format on the following Web site:

  www.storage.ibm.com/hardsoft/products/ess/refinfo.htm

• IBM TotalStorage Enterprise Storage Server Service Guide, SY27-7605

  This service guide is for the service support representatives who are trained to install and repair the IBM 2105 Models E10, E20, F10, and F20. This service guide is broken into three volumes that contain the following chapters:

  • Volume 1: Information, Start, and Isolation
  • Volume 2: Repair, Install, and Configure
  • Volume 3: Locations, Service Terminal, Codes, and Power

IBM related products publications

The following related publications are also available:

• DFSMS/MVS^(R) Version 1 Release 5 Software Support for the IBM Enterprise Storage Server, SC26-7318

  This book describes the changes you need to make to your MVS operating systems to support the ESS on the S/390 and zSeries hosts. These changes are also included in subsequent releases of MVS, OS/390 and z/OS.

• DFSMS Advanced Copy Services, SC35-0355

  This publication helps you to understand and use IBM Advanced Copy Services functions on an S/390 or zSeries. It describes two dynamic-copy functions and several point-in-time copy functions. These functions provide backup and recovery of data if a disaster occurs to your data center. The dynamic-copy functions are Peer-to-Peer Remote Copy and Extended Remote Copy. Collectively, these functions are known as remote copy. FlashCopy^(TM) and Concurrent Copy are the point-in-time copy functions.

• IBM OS/390 Hardware Configuration Definition User's Guide, SC28-1848

  This publication provides detailed information about the IODF. It also provides details about configuring parallel access volumes (PAVs).s. OS/390 uses the IODF.

• OS/390 MVS System Messages Volume 1 (ABA - ASA), SG24-5465

  This publication lists OS/390 zSeries MVS system messages ABA to ASA.

• IBM Enterprise Storage Server, GC28-1784

  This book, from the IBM International Technical Support Organization, introduces the ESS and provides an understanding of its benefits. It also describes in detail the architecture, hardware, and functions of the ESS.

• Implementing the IBM Enterprise Storage Server, SG24-5420

  This book can help you install, tailor, and configure the ESS in your environment.

• IBM Enterprise Storage Server Performance Monitoring and Tuning Guide, SG24-5656.

  This book provides guidance on the best way to configure, monitor, and manage your ESS to ensure optimum performance.

• Enterprise Storage Solutions Handbook, SG24-5250

  This book helps you understand what comprises enterprise storage management. The concepts include the key technologies that you need to know, and the IBM subsystems, software, and solutions that are available today. It also provides guidelines for implementing various enterprise storage administration tasks, so that you can establish your own enterprise storage management environment.

• IBM StorWatch Enterprise Storage Server Expert Hands-On Usage Guide, SG24-6102

  This guide helps you to install, tailor, and configure ESS Expert, and shows you how to use Expert.

• ESS Solutions for Open Systems Storage Compaq AlphaServer, HP, and Sun, SG24-6119

  This book helps you to install, tailor, and configure the ESS when you attach Compaq AlphaServer (running Tru64 UNIX), HP, and Sun hosts. This book does not cover Compaq AlphaServer running the Open VMS operating system.

  The book focuses on the settings required to give optimal performance and on device driver levels. This book is for the experienced UNIX professional who has a broad understanding of storage concepts.

• Implementing Fibre Channel Attachment on the ESS, SG24-6113

  This book helps you to install, tailor, and configure fibre-channel attachment of open-systems hosts to the ESS. It gives you a broad understanding of the procedures involved and describes the prerequisites and requirements. It also shows you how to implement fibre-channel attachment. This book also describes the steps required to migrate to direct fibre-channel attachment from native SCSI adapters and from fibre-channel attachment through the SAN Data Gateway (SDG).

• Fibre Transport Services (FTS) Direct Attach, Physical and Configuration Planning Guide, GA22-7234

  This publication provides information about fibre-optic and ESCON-trunking systems.

• DFSMS/MVS Version 1 Remote Copy Guide and Reference, SC35-0169

  This publication provides guidelines for using remote copy functions with S/390 and zSeries hosts.

• Implementing ESS Copy Services on S/390, SG24-5680

  This publication tells you how to install, customize, and configure Copy Services on an ESS that is attached to an S/390 or zSeries host system. Copy Services functions include Peer-to-Peer Remote Copy, Extended Remote Copy, FlashCopy^(TM) and, Concurrent Copy. This publication describes the functions, prerequisites, and corequisites and describes how to implement each of the functions into your environment.

• Implementing ESS Copy Services on UNIX and Windows NT/2000, SG24-5757

  This publication tells you how to install, customize, and configure ESS Copy Services on UNIX or Windows NT host systems. Copy Services functions include Peer-to-Peer Remote Copy, FlashCopy, Extended Remote Copy, and Concurrent Copy. Extended Remote Copy and Concurrent Copy are not available for UNIX and Windows NT host systems; they are only available on the S/390 or zSeries. This publication describes the functions and shows you how to implement each of the functions into your environment. It also shows you how to implement these solutions in an HACMP cluster.

• Enterprise Systems Architecture/390 ESCON I/O Interface, SA22-7202

  This publication provides a description of the physical and logical ESA/390 I/O interface and the protocols which govern information transfer over that interface. It is intended for designers of programs and equipment associated with the ESCON I/O interface and for service personnel maintaining that equipment although anyone concerned with the functional details of the ESCON I/O interface will find it useful.

• Fibre Channel Connection (FICON) I/O Interface, Physical Layer, SA24-7172

  This publication contains information to the Fiber Channel I/O Interface. This book is also available in PDF format by accessing Resource Link on the Internet at:

  www.ibm.com/servers/resourcelink.

• z/Architecture Principles of Operation, SA22-7832

  This publication contains, for reference purposes, a detailed definition of the z/Architecture. It is written as a reference for use primarily by assembler language programmers and describes each function at the level of detail needed to prepare an assembler language program that relies on that function; although anyone concerned with the functional details of z/Architecture will find it useful.

• IBM TotalStorage Enterprise Storage Server Subsystem Device Driver Installation and User's Guide, SC26-7442-00

  This book describes how to use the IBM Subsystem Device Driver on open-systems hosts to enhance performance and availability on the ESS. The Subsystem Device Driver creates redundant paths for shared logical unit numbers. The Subsystem Device Driver permits applications to run without interruption when path errors occur. It balances the workload across paths, and it transparently integrates with applications. For information about the Subsystem Device Driver, see the following Web site:

  www.ibm.com/storage/support/techsup/swtechsup.nsf/support/sddupdates/

Other related products publications

The following related publications are also available:

• NUMA-Q^(R) ESS Integration Release Notes for NUMA Systems, Part Number 1003-80094

  This publication contains information about special procedures and limitations involved in running ESS with Copy Services on an IBM xSeries 430 and IBM NUMA-Q^(R) host system.

  It also contains information on how to:

  • Configure the ESS
  • Configure the IBM NUMA-Q and xSeries 430 host system
  • Manage the ESS from the IBM NUMA-Q and xSeries 430 host system with DYNIX/ptx tools

  It is not available through IBM ordering systems. Contact your IBM NUMA-Q sales representative to obtain this document.

• Veritas Volume Manager Installation Guide

  This publication tells you how to install VxVM. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Veritas Volume Manager Storage Administrator's Guide

  This publication tells you how to configure the disk groups. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Veritas Volume Storage Administrator Guide

  This publication tells you how to administer the disk volumes. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Veritas Volume Manager Hardware Notes

  This publication tells you how to implement dynamic multipathing. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Veritas Cluster Server Installation Guide

  This publication tells you how to install the Veritas Cluster Server. See also the companion document, Veritas Cluster Server Release Notes. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Veritas Cluster Server Release Notes

  This publication tells you how to install the Veritas Cluster Server. See also the companion document, Veritas Cluster Server Installation Guide. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Veritas Cluster Server User's Guide

  This publication tells you how to configure the Veritas Cluster Server. See also the companion document, Quick Start Guide: An Example with NFS. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Quick Start Guide: An Example with NFS

  This publication tells you how to configure the Veritas Cluster Server. See also the companion document, Veritas Cluster Server User's Guide. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

• Fibre-channel Subsystem Installation Guide

  This publication tells you how to attach the xSeries 430 and NUMA-Q host system with fibre-channel adapters. It is not available through IBM ordering systems. Contact your sales representative to obtain this document.

Ordering publications

All of the publications that are listed in The IBM TotalStorage ESS publication library are available on a compact disc that comes with the ESS, unless otherwise noted. You can also order a hardcopy of each publication. For publications on compact disc, order IBM TotalStorage Enterprise Storage Server Customer Documents, SK2T-8770.

The customer documents are also available on the following ESS Web site:

www.storage.ibm.com/hardsoft/products/ess/refinfo.htm

Web sites

For general information about IBM storage products, see the following Web site:

www.storage.ibm.com/

For information about the IBM Enterprise Storage Server (ESS), see the following Web site:

www.storage.ibm.com/hardsoft/products/ess/ess.htm

To view and print the ESS publications, see the following Web site:

ssddom02.storage.ibm.com/disk/ess/documentation.html

To get current information about the host system models, operating systems, and adapters that the ESS supports, see the following Web site:

www.storage.ibm.com/hardsoft/products/ess/supserver.htm

For information about the IBM Subsystem Device Driver, see the following Web site:

ssddom01.storage.ibm.com/techsup/swtechsup.nsf/support/sddupdates/

To attach a SAN or host system that uses an industry-standard, fibre-channel arbitrated loop (FC-AL) through the IBM 2108 SAN Data Gateway Model G07, see the following Web site:

www.storage.ibm.com/hardsoft/products/sangateway/sangateway.htm

For information about the latest updates to Copy Services components including XRC, PPRC, Concurrent Copy, and FlashCopy for S/390 and zSeries, see the following Web site:

www.storage.ibm.com/software/sms/sdm/sdmtech.htm

For information about the IBM ESS Copy Services Command-Line Interface, (CLI), see the following Web site:

• ssddom01.storage.ibm.com/techsup/swtechsup.nsf/support/sddcliupdates

How to send your comments

Your feedback is important to help us provide the highest quality information. If you have any comments about this book or any other ESS documentation, you can submit them in one of the following ways:

• e-mail

  Submit your comments electronically to:

  starpubs@us.ibm.com

  Be sure to include the name and order number of the book and, if applicable, the specific location of the text you are commenting on, such as a page number or table number.

• Mail or fax

  Fill out the Readers' Comments form (RCF) at the back of this book and return it by mail or fax (1-800-426-6209) or give it to an IBM representative. If the RCF has been removed, you may address your comments to:

  International Business Machines Corporation
  RCF Processing Department
  G26/050
  5600 Cottle Road
  San Jose, CA 95193-0001
  U.S.A.
  

  
  

Chapter 1. Introducing the Subsystem Device Driver

This chapter introduces the IBM TotalStorage Enterprise Storage Server (ESS) Subsystem Device Driver (SDD) and provides an overview of SDD functions.

Subsystem Device Driver

The Subsystem Device Driver is a pseudo device driver designed to support the multipath configuration environments in the IBM ESS. It resides in a host system with the native disk device driver and provides the following functions:

• Enhanced data availability
• Dynamic I/O load-balancing across multiple paths
• Automatic path failover protection
• Concurrent download of licensed internal code.
• Path-selection policies for the host system

Enhanced data availability

Figure 1 shows that a SDD-residing host system is attached through SCSI or fibre-channel adapters to an ESS that has internal component redundancy and multipath configuration. SDD uses this multipath configuration to enhance data availability. That is, when there is a path failure, SDD reroutes I/O operations from the failing path to an alternate operational path. This capability prevents a single failing bus adapter on the host system, SCSI or fibre-channel cable, or host-interface adapter on the ESS from disrupting data access.

Figure 1. Multipath connections between a host system and the disk storage in an ESS

Dynamic I/O load-balancing

By distributing the I/O workload over multiple active paths, SDD provides dynamic load-balancing and eliminates data flow bottlenecks. In the event of failure in one data path, SDD automatically switches the affected I/O operations to another active data path, ensuring path failover protection.

Path-failover protection system

The SDD failover protection system is designed to minimize any disruptions in I/O operations and recover I/O operations from a failing data path. SDD provides path-failover protection through the following process:

• Detecting a path failure
• Notifying the host system of the path failure
• Selecting and using an alternate data path.

SDD dynamically selects an alternate I/O path when it detects a software or hardware problem.

Concurrent download of licensed internal code

SDD is capable of concurrent download of licensed internal (microcode) code. That is, it allows you to download and install the licensed internal code while applications continue running. During the download and installation process, the host adapters inside the ESS might not respond to host I/O requests for approximately 30 seconds. SDD makes this process transparent to the host system through its path-selection and retry algorithms.

Path-selection algorithms

SDD uses similar path-selection algorithms on all the host systems. There are two modes of operation:

single-path mode

The host system has only one path that is configured to an ESS logical unit number (LUN). SDD in single-path mode has the following characteristics:

• When an I/O error occurs, SDD retries the I/O operation a sufficient number of times to bypass the interval when the ESS host adapters are not available. This I/O error might be caused by the concurrent download of licensed internal code process. See Concurrent download of licensed internal code for more information.
• SDD never puts this single path into the Dead state.

multiple-path mode

The host system has multiple paths that are configured to an ESS LUN. SDD in multiple-path mode has the following characteristics:

• If an I/O error occurs on a path, SDD does not attempt to use the path again until 2000 successful I/O operations on an operational path. This process is known as bypassing a path. SDD bypasses a failing path twice (until the I/O error count reaches three) and then puts the path in the Dead state.
• If I/O errors occur consecutively on a path and the I/O error count reaches three, SDD immediately puts the failing path into the Dead state. It uses the same bypass algorithm twice on the failed path while it is in the Dead state.
• SDD puts the failed path from the Dead state back into the Open state after a certain number of successful I/O operations on an operational path. As listed in Table 1, the required number of successful I/O operations varies from one host system to another.
  

  Table 1. Required number of successful I/O operations before SDD placing a path in the Open state
  

  Operating System	Number of I/O operations
  AIX	50 000
  Windows NT	50 000
  Windows 2000	50 000
  HP	200 000
  Solaris	200 000

  If the first I/O operation fails after the path is put back into the Open state, SDD puts the path into the Dead state immediately and permanently. You must manually bring the path online by using the datapath command.

• If an I/O error occurs on the last operational path to a device, SDD attempts to reuse (or fail back to) a previously-failed path. SDD never puts the last operational path to an ESS LUN into the Dead state even if I/O errors have occurred on the path.
• If an I/O error occurs on all the paths to an ESS LUN, SDD returns the I/O error back to the application.

Chapter 2. Installing and configuring SDD on an AIX host system

This chapter provides step-by-step procedures for you to install, configure, upgrade, and remove the Subsystem Device Driver on an AIX host system that is attached to an ESS.

Understanding how SDD works on an AIX host system

As Figure 2 shows, SDD resides above the AIX disk driver in the protocol stack and acts as a pseudo device driver. I/O operations sent to SDD are passed to the AIX disk driver after path selection. When an active path experiences a failure (such as a cable or controller failure), SDD dynamically switches to another path.

Figure 2. Where SDD fits in the protocol stack

Each SDD device represents a unique physical device on the storage subsystem. There can be up to 32 hdisk devices that represent up to 32 different paths to the same physical device.

SDD devices behave almost like hdisk devices. Most operations on an hdisk device can be performed on the SDD device, including commands such as open, close, dd, or fsck.

Support for 32-bit and 64-bit applications on AIX 4.3.2, AIX 4.3.3, and AIX 5.1.0

Table 2 summarizes SDD 1.3.0.x support for 32-bit and 64-bit applications on AIX 4.3.2, AIX 4.3.3, and AIX 5.1.0.

Table 2. Support for 32-bit and 64-bit applications

SDD Installation Filesets	Application Mode	SDD Interface	AIX Kernel Mode	SDD Support
ibmSdd_432.rte	32-bit, 64-bit	LVM, raw device	32-bit	Yes
ibmSdd_433.rte	32-bit, 64-bit	LVM, raw device	32-bit	Yes
ibmSdd_510.rte	32-bit, 64-bit	LVM, raw device	32-bit, 64-bit	Yes
ibmSdd_510nchacmp.rte	32-bit, 64-bit	LVM, raw device	32-bit, 64-bit	Yes

Switching between 32-bit and 64-bit modes on AIX 5.1.x host systems

ln -sf /usr/lib/boot/unix_64 /unix
ln -sf /usr/lib/boot/unix_64 /usr/lib/boot/unix
bosboot -ak /usr/lib/boot/unix_64
shutdown -Fr

ln -sf /usr/lib/boot/unix_mp /unix
ln -sf /usr/lib/boot/unix_mp /usr/lib/boot/unix
bosboot -ak /usr/lib/boot/unix_mp
shutdown -Fr

The kernel mode of your AIX host system is switched to 32-bit mode after the system reboot completes. As a result, SDD automatically switches to 32-bit mode.

Hardware and software requirements

You must install the following hardware and software components to ensure that SDD installs and operates successfully.

Hardware

• ESS
• Host system
• SCSI adapters and cables
• Fibre adapters and cables

Software

• ibm2105.rte ESS package
• AIX operating system
• SCSI and fibre-channel device drivers

Host system requirements

To successfully install SDD 1.3.0.x, you must have AIX 4.2.1, AIX 4.3.2, AIX 4.3.3 or AIX 5.1.0 installed on your host system along with the fixes in Table 3. SDD 1.3.0.x does not support AIX 5.1.B.

Table 3. AIX PTF required fixes

AIX level	PTF number	Component name	Component level
4.2.1	IX62304
	U451711	perfagent.tools	2.2.1.4
	U453402	bos.rte.libc	4.2.1.9
	U453481	bos.adt.prof	4.2.1.11
	U458416	bos.mp	4.2.1.15
	U458478	bos.rte.tty	4.2.1.14
	U458496	bos.up	4.2.1.15
	U458505	bos.net.tcp.client	4.2.1.19
	U462492	bos.rte.lvm	4.2.1.16
4.3.2	U461953	bos.rte.lvm	4.3.2.4

Attention:

You must check for the latest information on APARs, maintenance level fixes, and microcode updates at the following downloadable website:

service.software.ibm.com/support/rs6000

ESS requirements

To successfully install SDD, ensure that your ESS meets the following requirements:

• The ibm2105.rte ESS package must be installed on your AIX host system.
• The ESS devices must be configured as either an:
  • IBM 2105xxx (SCSI-channel attached device)
  • IBM FC 2105xxx (fibre-channel attached device)

  Note:

  xxx is the ESS model number.

SCSI requirements

To use the SDD SCSI support, ensure your host system meets the following requirements:

• The bos.adt package must be installed. The host system can be a uniprocessor or a multiprocessor system, such as SMP.
• The maximum number of SCSI adapters that is supported is 32.
• A SCSI cable is required to connect each SCSI host adapter to an ESS port.
• The SDD I/O load-balancing and failover features require a minimum of two SCSI adapters.

  Note:

  SDD also supports one SCSI adapter on the host system. With single-path access, concurrent download of licensed internal code is supported with SCSI devices.

For information about the SCSI adapters that can attach to your AIX host system, go to the following website:

www.storage.ibm.com/hardsoft/products/ess/supserver.htm

Fibre requirements

To use the SDD fibre support, ensure your host system meets the following requirements:

• The AIX host system is an IBM RS/6000^(R) with AIX 4.3.3 or AIX 5.1.0.
• The AIX host system has the fibre-channel device drivers installed along with APARs IY10201, IY10994, IY11245, IY13736, IY17902, and IY18070.
• The bos.adt package must be installed. The host system can be a uniprocessor or a multiprocessor system, such as SMP.
• A fiber-optic cable is required to connect each fibre-channel adapter to an ESS port.
• Attention: If more than one adapter is attached to a peripheral component interconnect (PCI) bus, both adapter devices will be configured. Sometimes, though, one adapter saturates the entire PCI bus and causes command timeouts.

  The Emulex LP70000E adapter should be attached to its own PCI bus and the adapter should not be shared with other PCI adapters.

  Notes:

  1. The RS/6000 Models S70, S7A, and S80 support the attachment of a maximum of four Emulex LP7000E adapters.

  2. The RS/6000 Models F50 and H50 support the attachment of a maximum of three Emulex LP7000E adapters.
• The SDD I/O load-balancing and failover features require a minimum of two fibre adapters.

For information about the fibre-channel adapters that can be used on your AIX host system go to the following website:

www.storage.ibm.com/hardsoft/products/ess/supserver.htm

Non-supported environments

SDD does not support the following environments:

• A host server with a single-path fibre-channel connection to an ESS.

  Note:

  A host server with a single fibre-channel adapter that connects through a switch to multiple ESS ports is considered a multipath fibre-channel connection; and, thus is a supported environment.

• A host server with SCSI channel connections and a single-path fibre-channel connection to an ESS.
• A host server with both a SCSI channel and fibre-channel connection to a shared LUN.

Configuring the ESS

Before you install SDD, configure your ESS for single-port or multiple-port access for each LUN. SDD requires a minimum of two independent paths that share the same logical unit to use the load-balancing and failover features.

For more information about configuring your IBM Enterprise Storage Server, see IBM TotalStorage Enterprise Storage Server Introduction and Planning Guide.

Note:

Ensure the ibm2105.rte installation package is installed.

Installing fibre-channel device drivers and configuring fibre-channel devices

AIX fibre-channel device drivers are developed by IBM for the Emulex LP7000E adapter.

This section contains the procedures for installing fibre-channel device drivers and configuring fibre-channel devices. These procedures include:

1. Installing the AIX fibre-channel device drivers
2. Installing the Emulex adapter firmware (sf320A9)
3. Configuring fibre-channel attached devices

This section also contains procedures for:

• Removing fibre-channel attached devices
• Uninstalling fibre-channel device drivers

Requirement: For the fibre-channel support, the AIX host system must be an IBM RS/6000 system with AIX 4.3.3 or AIX 5.1.0. The AIX host system should have the fibre-channel device driver installed along with APARS IY10201, IY10994, IY11245, IY13736, IY17902, and IY18070.

Installing the AIX fibre-channel device drivers

Perform the following steps to install the AIX fibre-channel device drivers:

1. Install the fibre-channel device drivers from the AIX 4.3.3 compact disc. The fibre-channel device drivers include the following filesets:

   devices.pci.df1000f7

   Adapter device driver for RS/6000 with feature code 6227

   devices.fcp.disk

   FCP disk driver

   devices.common.IBM.fc

   FCP and SCSI protocol driver

2. Check to see if APARS IY10201, IY10994, IY11245, IY13736, IY17902, and IY18070 are installed by issuing the instfix -i | grep IY10201, instfix -i | grep IY10994, instfix -i | grep IY11245, instfix -i | grep IY13736, instfix -i | grep IY17902, and instfix -i | grep IY18070 commands. If the APARS are listed, that means that they are installed. If they are installed, go to Configuring fibre-channel attached devices. Otherwise, go to step 3.
3. Install APARS IY10201, IY10994, IY11245, IY13736, IY17902, and IY18070.

Configuring fibre-channel attached devices

The newly installed devices must be configured before they can be used. There are two ways to configure these devices. You can:

• Use the cfgmgr command.
• Use the shutdown -rF command to restart the system.

Determining the Emulex adapter firmware level (sf322A0)

You are required to install new adapter firmware only if the current adapter firmware is not at the sf322A0 level. Perform the following steps to download the Emulex adapter firmware:

1. Determine the firmware level that is currently installed. Issue the

   lscfg -vl fcsN command. The adapter's vital product data is displayed.

2. Look at the ZB field. The ZB field should look something like this:

   +--------------------------------------------------------------------------------+
   |(ZB).............S2F3.22A0                                                      |
   +--------------------------------------------------------------------------------+

   To determine the firmware level, ignore the second character in the ZB field. In the example, the firmware level is sf322A0.

3. If the adapter firmware level is at the sf322A0 level, there is no need to upgrade; otherwise, the firmware level must be upgraded. To upgrade the firmware level, go to Upgrading the Emulex adapter firmware level.

Upgrading the Emulex adapter firmware level

Upgrading the firmware level consists of downloading the firmware (microcode) from your AIX host system to the adapter. Before this can be done, however, the fibre-channel attached devices must be configured. After the devices are configured, you are ready to download the firmware from the AIX host system to the FCP adapter. Perform the following steps to download the firmware:

1. Verify that the correct level of firmware is installed on your AIX host system. Locate the file called df1000f7.131.320.320.320.503. It should be in the /etc/microcode directory. This file was copied into the /etc/microcode directory during the installation of the fibre-channel device drivers.
2. From the AIX command prompt, type diag and press Enter.
3. Select the Task Selection option.
4. Select the Download Microcode option.
5. Select all the fibre-channel adapters to which you want to download firmware. Press F7. The Download panel is displayed with one of the selected adapters highlighted. Press Enter to continue.
6. Type the filename for the firmware that is contained in the /etc/microcode directory and press Enter; or use the Tab key to toggle to Latest.
7. Follow the instructions that are displayed to download the firmware, one adapter at a time.
8. After the download is complete, issue the lscfg -v -l fcsN command to verify the firmware level on each fibre-channel adapter.

Removing fibre-channel attached devices

To remove all fibre-channel attached devices, you must issue the rmdev -dl fcsN -R command for each installed FCP adapter, where N is the FCP adapter number. For example, if you have two installed FCP adapters (adapter 0 and adapter 1), you must issue both the commands: rmdev -dl fcs0 -R and the rmdev -dl fcs1 -R

Uninstalling fibre-channel device drivers

There are two methods for uninstalling all of your fibre-channel device drivers. You can:

• Use the smitty deinstall command.
• Manually uninstall the drivers using the installp command.

Perform the following steps to use the smitty deinstall command:

1. Type smitty deinstall at the AIX command prompt and press Enter. The Remove Installed Software panel is displayed.
2. Press F4. All of the software that is installed is displayed.
3. Select the file name of the fibre-channel device driver you want to uninstall. Press Enter. The selected file name is displayed in the Software Name Field of the Remove Installed Software panel.
4. Use the Tab key to toggle to No in the PREVIEW Only? field. Press Enter. The uninstallation process begins.

Perform the following steps to use the installp command from the AIX command line:

1. Type installp -ug devices.pci.df1000f7 and press Enter.
2. Type installp -ug devices.common.IBM.fc and press Enter.
3. Type installp -ug devices.fcp.disk and press Enter.

Installing the Subsystem Device Driver

You must have root access and AIX system administrator knowledge to install SDD. See the IBM Subsystem Device Driver/Data Path Optimizer on an ESS:Installation Procedures/Potential Gotchas publication for additional information about SDD installation procedures. This publication is especially help if you have SP systems. You can find this publication at the following website:

http://SSDDOM01.storage.ibm.com/techsup/swtechsup.nsf/support/sdddocs

To install SDD, use the installation package that is appropriate for your environment. Table 4 lists and describes the SDD installation package file names (filesets).

Table 4. SDD package file names

Package file names	Description
ibmSdd_421.rte	AIX 4.2.1
ibmSdd_432.rte	AIX 4.3.2 or AIX 4.3.3 (also use when running HACMP with AIX 4.3.3 in concurrent mode)
ibmSdd_433.rte	AIX 4.3.3 (only use when running HACMP with AIX 4.3.3 in non-concurrent mode)
ibmSdd_510.rte	AIX 5.1.0 (also use when running HACMP with AIX 5.1.0 in concurrent mode)
ibmSdd_510nchacmp.rte	AIX 5.1.0 (also use when running HACMP with AIX 5.1.0 in non-concurrent mode)

Notes:

1. SDD 1.3.0.x does not support AIX 5.1.B.

2. SDD 1.3.0.x installed from either the ibmSdd_432.rte or ibmSdd_433.rte fileset is a 32-bit device driver. This version supports 32-bit and 64-bit mode applications on AIX 4.3.2 and AIX 4.3.3 host systems. A 64-bit mode application can access a SDD device directly or through the logical volume manager (LVM).

3. SDD 1.3.0.x installed from the ibmSdd_433.rte fileset is supported on AIX 4.3.3 and is for HACMP/6000 environments only; It supports non-concurrent and concurrent modes. However, in order to make the best use of the manner in which the device reserves are made, IBM recommends that you:
   • Use the ibmSdd_432.rte fileset for SDD 1.3.0.x when running HACMP with AIX 4.3.3 in concurrent mode.
   • Use the ibmSdd_433.rte fileset for SDD 1.3.0.x when running HACMP with AIX 4.3.3 in non-concurrent mode.
   Table 4 lists and describes the installation package file names (filesets) for the SDD 1.3.0.x.

4. The 1.3.0.x version of SDD installed from either ibmSdd_510.rte or ibmSdd_510nchacmp.rte filesets is supported on AIX 5.1.0; It contains both 32-bit and 64-bit drivers. Based on the kernel mode currently running on the system, the AIX loader will load the correct mode of the SDD into the kernel.

5. SDD 1.3.0.x contained in the ibmSdd_510nchacmp.rte fileset supports HACMP/6000 in both concurrent and non-concurrent modes. IBM recommends that you:
   • Install SDD 1.3.0.x from the bmSdd_510.rte fileset if you run HACMP with AIX 5.1.0 in concurrent code only
   • Install SDD 1.3.0.x from the ibmSdd_510nchacmp.rte fileset if you run HACMP with AIX 5.1.0 in non-concurrent mode

6. SDD does not support a system restart from a SDD pseudo device.

7. SDD does not support placing system paging devices (e.g. /dev/hd6) on a SDD pseudo device.

8. SDD 1.3.0.x installed from the ibmSdd_421.rte, ibmSdd_432.rte and ibmSdd_510.rte filesets do not support any application that depends on a reserve/release device on AIX 4.2.1, AIX 4.3.2, AIX 4.3.3, and AIX 5.10.

9. The published AIX limitation on one system is 10,000 devices. The combined number of hdisk and vpath devices should not exceed the number of supported devices by AIX. In a multipath environment, since each path to a disk creates an hdisk, the total number of disks being configured can be reduced by the number of paths to each disk.

The installation package installs a number of major files on your AIX system.Table 5 lists the major files that are part of the SDD installation package.

Table 5. Major files included in the SDD installation package

Filename	Description
defdpo	Define method of the SDD pseudo parent data path optimizer (dpo)
cfgdpo	Configure method of the SDD pseudo parent dpo
define_vp	Define method of the SDD vpath devices
addpaths	The command that dynamically adds more paths to Subsystem Device Driver devices while they are in Available state. Note: This command is not supported with Subsystem Device Driver for AIX 4.2.1; It is not available if you have the ibmSdd_421.rte fileset installed. This feature only supports Subsystem Device Driver for AIX 4.3.2 and higher.
cfgvpath	Configure method of SDD vpath devices
cfallvpath	Fast-path configure method to configure the SDD pseudo parent dpo and all vpath devices
vpathdd	SDD
hd2vp	The SDD script that converts an ESS hdisk device volume group to a Subsystem Device Drive vpath device volume group
vp2hd	The SDD script that converts a SDD vpath device volume group to an ESS hdisk device volume group
datapath	The SDD driver console command tool
lsvpcfg	The SDD driver query configuration status command
mkvg4vp	The command that creates a SDD volume group
extendvg4vp	The command that extends SDD devices to a SDD volume group
dpovgfix	The command that fixes a SDD volume group that has mixed vpath and hdisk physical volumes
savevg4vp	The command that backs-up all files belonging to a specified volume group with SDD devices.
restvg4vp	The command that restores all files belonging to a specified volume group with SDD devices.

The following procedures assume that SDD will be used to access all of your single and multipath devices.

To install SDD, use the System Management Interface Tool (SMIT). The SMIT facility has two interfaces, nongraphical (type smitty to invoke the nongraphical user interface) or graphical (type smit to invoke the graphical user interface). SDD is released as an installation image. The fileset name is ibmSdd_nnn.rte, where nnn represents the AIX version level (4.2.1, 4.3.2, 4.3.3 or 5.1.0). For example, the fileset name for the AIX 4.3.2 level is ibmSdd_432.rte.

Note:

The ibmSdd_432.rte installation package can be installed on an AIX 4.3.2 or AIX 4.3.3 system

Throughout this SMIT procedure, /dev/cd0 is used for the compact disc drive address. The drive address might be different in your environment. Perform the following SMIT steps to install the SDD package on your system.

1. Log in as the root user.
2. Load the compact disc into the CD-ROM drive.
3. From your desktop window, type smitty install_update and press Enter to go directly to the installation panels. The Install and Update Software menu is displayed.
4. Highlight Install and Update from LATEST Available Software and press Enter.
5. Press F4 to display the INPUT Device/Directory for Software panel.
6. Select the compact disc drive that you are using for the installation; for example, /dev/cd0, and press Enter.
7. Press Enter again. The Install and Update from LATEST Available Software panel is displayed.
8. Highlight Software to Install and press F4. The SOFTWARE to Install panel is displayed.
9. Select the installation package that is appropriate for your environment. Table 4 lists and describes the SDD installation package file names (filesets).
10. Press Enter. The Install and Update from LATEST Available Software panel is displayed with the name of the software you selected to install.
11. Check the default option settings to ensure that they are what you need.
12. Press Enter to install. SMIT responds with the following message:

    +--------------------------------------------------------------------------------+
    |     ARE YOU SURE??                                                             |
    |     Continuing may delete information you may want to keep.                    |
    |     This is your last chance to stop before continuing.                        |
    +--------------------------------------------------------------------------------+

13. Press Enter to continue. The installation process can take several minutes to complete.
14. When the installation is complete, press F10 to exit from SMIT. Remove the compact disc.

Verifying the installation

You can verify that SDD has been successfully installed by issuing the lslpp -l ibmSdd_421.rte, lslpp -l ibmSdd_432.rte, lslpp -l ibmSdd_433.rte, lslpp -l ibmSdd_510.rte or lslpp -l ibmSdd_510nchacmp.rte command.

If you have successfully installed the ibmSdd_432.rte package, the output from the lslpp -l ibmSdd_432.rte command looks like this:

+--------------------------------------------------------------------------------+
|Fileset                      Level  	 State        Description                  |
|------------------------------------------------------------------------------  |
|Path: /usr/lib/objrepos                                                         |
|  ibmSdd_432.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V432 V433 for concurrent |
|                                                   HACMP                        |
|                                                                                |
|Path: /etc/objrepos                                                             |
|  ibmSdd_432.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V432 V433 for concurrent |
|                                                   HACMP                        |
|                                                                                |
+--------------------------------------------------------------------------------+

If you have successfully installed the ibmSdd_433.rte package, the output from the lslpp -l ibmSdd_433.rte command looks like this:

+--------------------------------------------------------------------------------+
|Fileset                      Level  	 State        Description                  |
|--------------------------------------------------------------------------------|
|Path: /usr/lib/objrepos                                                         |
|  ibmSdd_433.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V433  for non-concurrent |
|                                                   HACMP                        |
|                                                                                |
|Path: /etc/objrepos                                                             |
|  ibmSdd_433.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V433  for non-concurrent |
|                                                   HACMP                        |
|                                                                                |
+--------------------------------------------------------------------------------+

If you have successfully installed the ibmSdd_510.rte package, the output from the lslpp -l ibmSdd_510.rte command looks like this:

+--------------------------------------------------------------------------------+
|Fileset                      Level  	 State        Description                  |
|--------------------------------------------------------------------------------|
|Path: /usr/lib/objrepos                                                         |
|  ibmSdd_510.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V510 for concurrent HACMP|
|                                                                                |
|Path: /etc/objrepos                                                             |
|  ibmSdd_510.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V510 for concurrent HACMP|
|                                                                                |
+--------------------------------------------------------------------------------+

If you have successfully installed the ibmSdd_510nchacmp.rte package, the output from the lslpp -l ibmSdd_510nchacmp.rte command looks like this:

+--------------------------------------------------------------------------------+
|Fileset                      Level    State        Description                  |
|--------------------------------------------------------------------------------|
|Path: /usr/lib/objrepos                                                         |
|  ibmSdd_510nchacmp.rte      1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V510 for non-concurrent  |
|                                                   HACMP                        |
|                                                                                |
|Path: /etc/objrepos                                                             |
|  ibmSdd_510nchacmp.rte      1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
|                                                   AIX V510 for non-concurrent  |
|                                                   HACMP                        |
|                                                                                |
+--------------------------------------------------------------------------------+

Configuring the Subsystem Device Driver

The following section describes the steps needed to prepare for and to configure the Subsystem Device Driver.

Preparing to configure the Subsystem Device Driver

Before you configure SDD, ensure that:

• The ESS is operational.
• The ibmSdd_nnn.rte software is installed on the AIX host system
• The ESS hdisks are configured correctly on the AIX host system.

Configure the ESS devices before you configure the SDD. If you configure multiple paths to an ESS device, make sure that all paths (hdisks) are in Available state. Otherwise, some SDD devices will lose multiple-path capability.

Perform the following steps:

1. Use the lsdev -Cc disk | grep 2105 command to check the ESS device configuration.
2. If you have already created some ESS volume groups, vary off (deactivate) all active volume groups with ESS subsystem disks by using the varyoffvg (LVM) command.

   Attention: Before you vary off a volume group, unmount all file systems in that volume group. If some ESS devices (hdisks) are used as physical volumes of an active volume group, and there are file systems of that volume group being mounted, then you must unmount all file systems, and vary off (deactivate) all active volume groups with ESS SDD disks.

Configuring the Subsystem Device Driver

Perform the following steps to configure SDD using SMIT:

1. Type smitty device from your desktop window. The Devices menu is displayed.
2. Highlight Data Path Device and press Enter. The Data Path Device panel is displayed.
3. Highlight Define and Configure All Data Path Devices and press Enter. The configuration process begins.
4. Check the SDD configuration status. See Displaying the ESS vpath device configuration.
5. Enter the varyonvg command to vary on all deactivated ESS volume groups.
6. If you want to convert the ESS hdisk volume group to SDD vpath devices, you must run the hd2vp utility. (See hd2vp and vp2hd for information about this utility.)
7. Mount the file systems for all volume groups that were previously unmounted.

Note:

The following error might occur if you run the cfgmgr command with all vpath paths (hdisks) in the Open state:

+--------------------------------------------------------------------------------+
|	0514-061 Cannot find a child device                                            |
+--------------------------------------------------------------------------------+

Ignore this error if it is returned by the cfgmgr command when all vpath paths (hdisks) are open. You can use the datapath query device command to verify the status of all vpath paths.

Verifying the SDD configuration

To check the SDD configuration, you can use either the SMIT Display Device Configuration panel or the lsvpcfg console command.

Perform the following steps to verify the SDD configuration on an AIX host system:

1. Type smitty device from your desktop window. The Devices menu is displayed.
2. Select Data Path Device and press Enter. The Data Path Device panel is displayed.
3. Select Display Data Path Device Configuration and press Enter to display the condition (Defined or Available) of all SDD pseudo devices and the paths to each device.

If any device is listed as Defined, the configuration was not successful. Check the configuration procedure again. See Configuring the Subsystem Device Driver for information about the procedure.

Perform the following steps to verify that multiple paths are configured for each adapter connected to an ESS port:

1. Type smitty device from your desktop window. The Devices menu is displayed.
2. Highlight Data Path Device and press Enter. The Data Path Device panel is displayed.
3. Highlight Display Data Path Device Adapter Status and press Enter. All attached paths for each adapter are displayed.

If you want to use the command-line interface to verify the configuration, type lsvpcfg.

You should see an output similar to this:

+--------------------------------------------------------------------------------+
|vpath0 (Avail pv vpathvg) 018FA067 = hdisk1 (Avail )                            |
|vpath1 (Avail ) 019FA067 = hdisk2 (Avail )                                      |
|vpath2 (Avail ) 01AFA067 = hdisk3 (Avail )                                      |
|vpath3 (Avail ) 01BFA067 = hdisk4 (Avail ) hdisk27 (Avail )                     |
|vpath4 (Avail ) 01CFA067 = hdisk5 (Avail ) hdisk28 (Avail )                     |
|vpath5 (Avail ) 01DFA067 = hdisk6 (Avail ) hdisk29 (Avail )                     |
|vpath6 (Avail ) 01EFA067 = hdisk7 (Avail ) hdisk30 (Avail )                     |
|vpath7 (Avail ) 01FFA067 = hdisk8 (Avail ) hdisk31 (Avail )                     |
|vpath8 (Avail ) 020FA067 = hdisk9 (Avail ) hdisk32 (Avail )                     |
|vpath9 (Avail pv vpathvg) 02BFA067 = hdisk20 (Avail ) hdisk44 (Avail )          |
|vpath10 (Avail pv vpathvg) 02CFA067 = hdisk21 (Avail ) hdisk45 (Avail )         |
|vpath11 (Avail pv vpathvg) 02DFA067 = hdisk22 (Avail ) hdisk46 (Avail )         |
|vpath12 (Avail pv vpathvg) 02EFA067 = hdisk23 (Avail ) hdisk47 (Avail )         |
|vpath13 (Avail pv vpathvg) 02FFA067 = hdisk24 (Avail ) hdisk48 (Avail )         |
+--------------------------------------------------------------------------------+

The output shows:

• The name of each pseudo device (for example, vpath13)
• The Defined or Available condition of a pseudo device
• Whether or not the pseudo device is defined to AIX as a physical volume (the pv flag)
• The name of the volume group the device belongs to (for example, vpathvg)
• The unit serial number of the ESS LUN (for example, 02FFA067)
• The names of the AIX disk devices making up the pseudo device and their configuration and physical volume status

Changing the path-selection policy

SDD supports path-selection policies that increase the performance of a multipath-configured ESS and make path failures transparent to applications. The following path-selection policies are supported:

load balancing (lb)

The path to use for an I/O operation is chosen by estimating the load on the adapter to which each path is attached. The load is a function of the number of I/O operations currently in process. If multiple paths have the same load, a path is chosen at random from those paths.

round robin (rr)

The path to use for each I/O operation is chosen at random from those paths not used for the last I/O operation. If a device has only two paths, SDD alternates between the two.

failover only (fo)

All I/O operations for the device are sent to the same (preferred) path until the path fails because of I/O errors. Then an alternate path is chosen for subsequent I/O operation.

The path-selection policy is set at the SDD device level. The default path-selection policy for a SDD device is load balancing. You can change the policy for a SDD device with the chdev command.

Before changing the path-selection policy, determine the active attributes for the SDD device. Type the lsattr -El vpathN command. Press Enter, where N represents the vpath-number, N=[0,1,2,...]. The output should look similar to this:

+--------------------------------------------------------------------------------+
|pvid         0004379001b90b3f0000000000000000 Data Path Optimizer Parent False  |
|policy       df                               Scheduling Policy          True   |
|active_hdisk hdisk1/30C12028                  Active hdisk               False  |
|active_hdisk hdisk5/30C12028                                                    |
+--------------------------------------------------------------------------------+

The path-selection policy is the only attribute of a SDD device that can be changed. The valid policies are rr, lb, fo, and df. Here are the explanations for these policies:

rr

round robin

fo

failover only

lb

load balancing

df

(load balancing) default policy

Attention: By changing a SDD device's attribute, the chdev command unconfigures and then reconfigures the device. You must ensure the device is not in use if you are going to change its attribute. Otherwise, the command fails.

Use the following command to change the SDD path-selection policy:

chdev  -l  vpathN   -a  policy=[rr/fo/lb/df]

Adding paths to SDD devices of a volume group

You can add more paths to SDD devices that belong to a volume group after you have initially configured SDD. This section shows you how to add paths to SDD devices from AIX 4.2.1 and AIX 4.3.2 or higher host systems.

Adding paths from AIX 4.3.2 or higher host systems

Adding paths from AIX 4.2.1 host systems

To activate additional paths to a SDD device, the related SDD devices must be unconfigured and then reconfigured. The SDD conversion scripts should be run to enable the necessary SDD associations and links between the SDD vpath (pseudo) devices and the ESS hdisk devices.

Note:

Ensure that logical volume sharing is enabled at the ESS for all applicable devices. Logical volume sharing is enabled using the ESS Specialist. See IBM TotalStorage Enterprise Storage Server Web Interface User's Guide for information about enabling volume sharing.

Perform the following steps to activate additional paths to SDD devices of a volume group from your AIX 4.2.1 host system:

1. Identify the volume groups containing the SDD devices to which you want to add additional paths. Type the following command:

   lspv
   

2. Check if all the physical volumes belonging to that SDD volume group are SDD devices (vpathNs). If they are not, you need to fix the problem.

   Attention: You must fix this problem with the volume group before proceeding to step 3. Otherwise, the volume group loses path failover capability. To fix the problem, type the following command:

   dpovgfix vg-name
   

   Vg-name represents the volume group.
3. Identify the associated file systems for the selected volume group. Type the following command:

   lsvgfs vg-name
   

4. Identify the associated mounted file systems for the selected volume group. Type the following command:

   mount
   

5. Unmount the file systems of the selected volume group listed in step 3. Type the following command:

   umount mounted-filesystem
   

6. Run the vp2hd volume group conversion script to convert the volume group from SDD devices to ESS hdisk devices. Type the following command to run the script:

   vp2hd vg-name
   

   When the conversion script completes, the volume group is in the Active condition (varied on).
7. Vary off the selected volume group in preparation for SDD reconfiguration. Type the following command:

   varyoffvg  vg-name
   

8. Run the AIX configuration manager cfgmgr to recognize all new hdisk devices. You can do this in one of two ways:
   • Run the cfgmgr command n times, where n represents the number of paths for the SDD. (See Note on page *** for an explanation of why cfgmgr should be run n times.)
   • Run the cfgmgr -l [scsiN/fcsN] command for each relevant SCSI or FCP adapter.

   Note:

   Ensure that all logical drives on the ESS are identified as hdisks before continuing.

9. Unconfigure affected SDD devices to the Defined condition by using the rmdev -l vpathN command; where N represents the vpath-number you want to set to the Defined condition N=[0,1,2,...]. This command allows you to unconfigure only SDD devices for which you are adding paths.

   Note:

   Use the rmdev -l dpo -R command if you need to unconfigure all Subsystem Device Driver devices. SDD volume groups must be inactive before unconfiguring. This command attempts to unconfigure all SDD devices recursively.

10. Reconfigure SDD devices by using either the System Management Interface Tool (SMIT) or the command-line interface.

    If you are using SMIT, perform the following steps:

    1. Type Smitty device from your desktop window. The Devices menu is displayed.
    2. Highlight Data Path Devices and press Enter. The Data Path Devices menu is displayed.
    3. Highlight Define and Configure All Data Path Devices and press Enter. SMIT executes a script to define and configure all SDD devices that are in the Defined condition.
    If you are using the command-line interface, type the mkdev -l vpathN command for each SDD device or type the cfallvpath command to configure all SDD devices.
11. Verify your datapath configuration using either SMIT or the command-line interface.

    If you are using SMIT, perform the following steps:

    1. Type Smitty device from your desktop window. The Devices menu is displayed.
    2. Highlight Data Path Devices and press Enter. The Data Path Devices menu is displayed.
    3. Highlight Display Data Path Device Configuration and press Enter.
    If you are using the command-line interface, type the lsvpcfg command to display the SDD configuration status.

    SDD devices should show two or more hdisks associated with each SDD device when failover protection is required.

12. Vary on the volume groups selected in step 3. Type the following command:

    varyonvg  vg-name
    

13. Run the hd2vp script to convert the volume group from ESS hdisk devices back to SDD vpath devices. Type the following command:

    hd2vp vg-name
    

14. Mount all file systems for the volume groups that were previously unmounted.

Unconfiguring SDD devices

Before you unconfigure SDD devices, all the file systems belonging to the SDD volume groups must be unmounted. Then, run the vp2hd conversion script to convert the volume group from SDD devices (vpathN) to ESS subsystem devices (hdisks).

Note:

If you are running HACMP with ibmSdd_433.rte fileset installed on your host system, there are special requirements regarding unconfiguring and removing SDD 1.3.0.x. vpath devices. See Special requirements.

Using the System Management Interface Tool (SMIT), you can unconfigure the SDD devices in two ways. Either you can unconfigure without deleting the device information from the Object Database Management (ODM) database, or you can delete device information from the ODM database. If you unconfigure without deleting the device information, the device remains in the Defined condition. Using either SMIT or the mkdev -l vpathN command, you can return the device to the Available condition.

If you delete the device information from the ODM database, that device is removed from the system. To return it, follow the procedure described in "Configuring the Subsystem Device Driver".

Perform the following steps to unconfigure SDD devices:

1. Type smitty device from your desktop window. The Devices menu is displayed.
2. Highlight Devices and press Enter. The Devices menu is displayed.
3. Highlight Data Path Device and press Enter. The Data Path Device panel is displayed.
4. Highlight Remove a Data Path Device and press Enter. A list of all SDD devices and their conditions (either Defined or Available) is displayed.
5. Select the device that you want to unconfigure. Select whether or not you want to delete the device information from the ODM database.
6. Press Enter. The device is unconfigured to the condition that you selected.
7. To unconfigure more SDD devices you have to repeat steps 4-6 for each SDD device.

Notes:

1. The fast-path command to unconfigure all SDD devices from the Available to the Defined condition is: rmdev -l dpo -R

2. The fast-path command to remove all Subsystem Device Driver devices from your system is: rmdev -dl dpo -R

Removing the Subsystem Device Driver

Before you remove the SDD package from your AIX host system, all the SDD devices must be removed from your host system. The fast-path rmdev -dl dpo -R command removes all the SDD devices from your system. After all SDD devices are removed, perform the following steps to remove SDD.

1. Type smitty deinstall from your desktop window to go directly to the Remove Installed Software panel.
2. Type ibmSdd_421.rte, ibmSdd_432.rte, ibmSdd_433.rte, ibmSdd_510.rte, or ibmSdd_510nchacmp.rte in the SOFTWARE name field and press Enter.
3. Press the Tab key in the PREVIEW Only? field to toggle between Yes and No. Select No to remove the software package from your AIX host system.

   Note:

   If you select Yes, the process stops at this point and previews what you are removing. The results of your pre-check are displayed without removing the software. If the condition for any SDD device is either Available or Defined, the process fails.

4. Select No for the remaining fields on this panel.
5. Press Enter. SMIT responds with the following message:

   +--------------------------------------------------------------------------------+
   |     ARE YOU SURE??                                                             |
   |     Continuing may delete information you may want to keep.                    |
   |     This is your last chance to stop before continuing.                        |
   +--------------------------------------------------------------------------------+

6. Press Enter to begin the removal process. This might take a few minutes.
7. When the process is complete, the SDD software package is removed from your system.

lslpp -l ibmSdd_421.rte
lslpp -l ibmSdd.rte.421
lslpp -l ibmSdd_432.rte
lslpp -l ibmSdd.rte.432
lslpp -l ibmSdd_433.rte
lslpp -l ibmSdd.rte.433
 

Attention: If a SDD volume group's physical volumes are mixed with hdisk devices and vpath devices, you must run the dpovgfix utility to fix this problem. Otherwise, SDD will not function properly. Use the dpovgfix vg_name command to fix this problem.

Using concurrent download of licensed internal code

Concurrent download of licensed internal code is the capability to download and install licensed internal code on an ESS while applications continue to run. This capability is supported for single-path (SCSI only) and multiple-path (SCSI or FCP) access to an ESS.

Attention: During the download of licensed internal code, the AIX error log might overflow and excessive system paging space could be consumed. When the system paging space drops too low it could cause your AIX system to hang. To avoid this problem, you can perform the following steps prior to doing the download:

1. Save the existing error report by typing the following command from the AIX command-line interface:

   > errpt > file.save
   

2. Delete the error log from the error log buffer by typing the following command:

   > errclear 0
   

3. Enlarge the system paging space by using the SMIT tool.
4. Stop the AIX error log daemon by typing the following command:

   /usr/lib/errstop
   

Once you have completed steps 1- 4, you can perform the download of the ESS licensed internal code. After the download completes, type /usr/lib/errdemon from the command-line interface to restart the AIX error log daemon.

Understanding the SDD support for High Availability Cluster Multi-Processing (HACMP/6000)

You can run the Subsystem Device Driver in concurrent and non-concurrent multihost environments in which more than one host is attached to the same LUNs on an ESS. RS/6000 servers running HACMP/6000 in concurrent or non-concurrent mode are supported. Different SDD releases support different kinds of environments. (See Table 6, Table 8, Table 7 and Table 9.)

HACMP/6000 provides a reliable way for clustered IBM RS/6000 servers which share disk resources to recover from server and disk failures. In a HACMP/6000 environment, each RS/6000 server in a cluster is a node. Each node has access to shared disk resources that are accessed by other nodes. When there is a failure, HACMP/6000 transfers ownership of shared disks and other resources based on how you define the relationship among nodes in a cluster. This process is known as node failover or node failback. HACMP supports two modes of operation:

non-concurrent

Only one node in a cluster is actively accessing shared disk resources while other nodes are standby.

concurrent

Multiple nodes in a cluster are actively accessing shared disk resources.

SDD supports RS/6000 servers connected to shared disks with SCSI adapters and drives as well as FCP adapters and drives. The kind of attachment support depends on the version of SDD that you have installed. Table 6 and Table 8 summarizes the software requirements to support HACMP/6000:

Table 6. Software support for HACMP/6000 in concurrent mode

Subsystem Device Driver Version and Release Level	HACMP 4.3.1 + APARs	HACMP 4.4 + APARs
SDD 1.1.4.0 (SCSI only)	IY07392 IY03438 IY11560 IY08933 IY11564 IY12021 IY12056 F model requires IY11110	IY11563 IY11565 IY12022 IY12057 F model requires IY11480
SDD 1.2.0.0 (SCSI/FCP)	IY07392 IY13474 IY03438 IY08933 IY11560 IY11564 IY12021 IY12056 F model requires IY11110	IY13432 IY11563 IY11565 IY12022 IY12057 F model requires IY11480
SDD 1.2.2.x (SCSI/FCP)	IY07392 IY13474 IY03438 IY08933 IY11560 IY11564 IY12021 IY12056 F model requires IY11110	IY13432 IY11563 IY11565 IY12022 IY12057 F model requires IY11480
SDD 1.3.0.x (SCSI/FCP)	IY07392 IY13474 IY03438 IY08933 IY11560 IY11564 IY12021 IY12056 F model requires IY11110	IY13432 IY11563 IY11565 IY12022 IY12057 F model requires IY11480

Table 7. Software support for HAMCP/6000 in non-concurrent mode

Subsystem Device Driver Version and Release Level	HACMP 4.3.1 + APARs	HACMP 4.4 + APARs
SDD 1.2.2.x (SCSI/FCP)	IY07392 IY13474 IY03438 IY08933 IY11560 IY11564 IY12021 IY12056 IY14682 F model requires IY11110	IY13432 IY11563 IY11565 IY12022 IY12057 IY14683 F model requires IY11480
ibmSdd_433.rte fileset for SDD 1.3.0.x (SCSI/FCP)	IY07392 IY13474 IY03438 IY08933 IY11560 IY11564 IY12021 IY12056 IY14682 F model requires IY11110	IY13432 IY11563 IY11565 IY12022 IY12057 IY14683 F model requires IY11480

Table 8. Software support for HACMP/6000 in concurrent mode on AIX 5.1.0 (32-bit kernel only)

Subsystem Device Driver Version and Release Level	HACMP 4.4 + APARs
ibmSdd_510.rte fileset for SDD 1.3.0.x (SCSI/FCP)	IY11563 IY11565 IY12022 IY12057 IY13432 IY14683 IY17684 IY19089 IY19156 F model requires IY11480

Table 9. Software support for HACMP/6000 in non-concurrent mode on AIX 5.1.0 (32-bit kernel only)

Subsystem Device Driver Version and Release Level	HACMP 4.4 + APARs
ibmSdd_510nchacmp.rte fileset for SDD 1.3.0.x (SCSI/FCP)	IY11563 IY11565 IY12022 IY12057 IY13432 IY14683 IY17684 IY19089 IY19156 F model requires IY11480

Note:

For the most up-to-date list of required APARs go to the following website: www.storage.ibm.com/hardsoft/products/ess/supserver.htm

Even though SDD supports HACMP/6000, certain combinations of features are not supported. Table 10 lists those combinations:

Table 10. HACMP/6000 and supported SDD features

Feature	RS/6000 node running HACMP
ESS concurrent code load	Yes
Subsystem Device Driver load balancing	Yes
SCSI	Yes
FCP (fibre)	Yes
Single-path fibre	No
SCSI and fibre-channel connections to the same LUN from one host (mixed environment)	No

What's new in SDD for HACMP/6000

The ibmSdd_433.rte and ibmSdd_510nchacmp.rte filesets for SDD 1.3.0.x have different features compared with ibmSdd_432.rte and ibmSdd_510.rte filesets for SDD 1.3.0.x. The ibmSdd_433.rte and ibmSdd_510nchacmp.rte filesets implement the SCSI-3 Persistent Reserve command set, in order to support HACMP in non-concurrent mode with single-point failure protection. The ibmSdd_433.rte and ibmSdd_510nchacmp.rte filesets require the ESS G3 level microcode on the ESS to support the SCSI-3 Persistent Reserve command set. If the ESS G3 level microcode is not installed, the ibmSdd_433.rte and ibmSdd_510nchacmp.rte filesets will switch the multi-path configuration to a single-path configuration. There is no single-point failure protection for single-path configurations.

The ibmSdd_433.rte and ibmSdd_510nchacmp.rte filesets have a new attribute under its pseudo parent (dpo), that reflects whether the ESS supports the Persistent Reserve Command set or not. The attribute name is persistent_resv. If SDD detects that G3 level microcode is installed, the persistent_resv attribute is created in the CuAt ODM and its value is set to yes; otherwise this attribute only exists in the PdAt ODM and its value is set to no (default). You can use the following command to check the persistent_resv attribute, after the SDD device configuration is complete:

odmget  -q  "name = dpo"  CuAt

If your attached ESS has the G3 microcode, the output should look similar to this:

			name  =  "dpo"
			attribute =  "persistent_resv"
			value  =  "yes"
			generic  =  "D"
			rep  =  "sl"
			nls_index  =  0

In order to implement the Persistent Reserve command set, each host server needs a unique 8-byte reservation key. There are 2 ways to get a unique reservation key. In HACMP/6000 environments, HACMP/6000 generates a unique key for each node in the ODM database. When SDD cannot find that key in the ODM database, it generates a unique reservation key by using the middle 8 bytes of the output from the uname -m command.

To check the Persistent Reserve Key of a node, provided by HACMP, issue the command:

odmget  -q  "name = ioaccess" CuAt

The output should look similar to this:

      name = "ioaccess"
      attribute = "perservekey"
      value = "01043792"
      type = "R"
      generic = ""
      rep = "s"
      nls_index = 0

Special requirements

There is a special requirement regarding unconfiguring and removing the ibmSdd_433.rte and ibmSdd_510nchacmp.rte filesets for SDD 1.3.0.x vpath devices. You must unconfigure and remove the vpath devices before you unconfigure and remove the vpath devices' underlying ESS hdisks. Otherwise if the ESS hdisks are unconfigured and removed first, the persistent reserve will not be released, even though the vpath devices have been successfully unconfigured and removed.

SDD does not automatically create the pvid attribute in the ODM database for each vpath device. The AIX disk driver automatically creates the pvid attribute in the ODM database, if a pvid exists on the physical device; however, SDD does not. Therefore, the first time you import a new SDD volume group to a new cluster node, you must import the volume group using hdisks as physical volumes. Next, run the hd2vp conversion script (see SDD utility programs) to convert the volume group's physical volumes from ESS hdisks to vpath devices. This conversion step not only create pvid attributes for all vpath devices which belong to that imported volume group, it also deletes the pvid attributes for these vpath devices' underlying hdisks. Later on you can import and vary on the volume group directly from the vpath devices. These special requirements apply to both concurrent and non-concurrent volume groups.

Under certain conditions, the state of a physical device's pvid on a system is not always as expected. So it is necessary to determine the state of a pvid as displayed by the lspv command, in order to select the appropriate import volume group action.

There are four scenarios:

Scenario 1. lspv displays pvid's for both hdisks and vpath:

	>lspv
	hdisk1		003dfc10a11904fa	None
	hdisk2		003dfc10a11904fa	None
	vpath0		003dfc10a11904fa	None

Scenario 2. lspv displays pvid's for hdisks only:

	>lspv
	hdisk1		003dfc10a11904fa	None
	hdisk2		003dfc10a11904fa	None
	vpath0		none					None

For both Scenario 1 and Scenario 2, the volume group should be imported using the hdisk names and then converted using the hd2vp command:

	>importvg -y vg_name -V major# hdisk1
	>hd2vp vg_name

Scenario 3. lspv displays the pvid for vpath only:

	>lspv
	hdisk1		none					None
	hdisk2		none					None
	vpath0		003dfc10a11904fa	None

For Scenario 3, the volume group should be imported using the vpath name:

	>importvg -y vg_name -V major# vpath0

Scenario 4. lspv does not display the pvid on the hdisks or the vpath:

	>lspv
	hdisk1		none			None
	hdisk2		none			None
	vpath0		none			None

For Scenario 4, the pvid will need to be placed in the odm for the vpath devices and then the volume group can be imported using the vpath name:

	>chdev -l vpath0 -a pv=yes
	>importvg -y vg_name -V major# vpath0

Note:

See Importing a volume group with SDD for a detailed procedure for importing a volume group with the SDD devices.

How to recover paths that are lost during HACMP/6000 node failover

Normally, when there is a node failure, HACMP/6000 transfers ownership of shared disks and other resources, through a process known as node failover. Certain situations, such as, a loose or disconnected SCSI or fibre-adapter card, can cause your vpath devices to lose one or more underlying paths during node failover. Perform the following steps to recover these paths:

• Check to ensure that all the underlying paths (hdisks) are in the Available state
• Run the addpaths command to add the lost paths back to the SDD devices.

If your vpath devices have lost one or more underlying paths that belong to an active volume group, you can use either the Add Paths to Available Data Path Devices SMIT panel or run the addpaths command from the AIX command line to recover the lost paths. Go to Adding paths to SDD devices of a volume group for more information about the addpaths command.

Note:

Simply running the cfgmgr command while the vpath devices are in the Available state will not recover the lost paths; that is why you need to run the addpaths command to recover the lost paths.

SDD does not support the addpaths command for AIX 4.2.1; it is not available if you have the ibmSdd_421.rte fileset installed (this feature only supports SDD for AIX 4.3.2 and higher.) If you have the ibmSdd_421.rte fileset installed, and if your vpath devices have lost one or more underlying paths and they belong to an active volume group, perform the following steps to recovery the lost paths:

Note:

• When there is a node failure, HACMP/6000 transfers ownership of shared disks and other resources, through a process known as node failover. To recover these paths, you need to first check to ensure that all the underlying paths (hdisks) are in the Available state. Next, you need to unconfigure and reconfigure your SDD vpath devices.
• Simply running the cfgmgr command while vpath devices are in the Available state will not recover the lost paths; that is why you need to unconfigure and reconfigure the vpath devices.

Notes:

1. HACMP/6000 running in concurrent mode is supported with the ibmSdd_432.rte fileset for SDD 1.1.4. (SCSI only)

2. HACMP/6000 running in concurrent mode is supported with the ibmSdd_432.rte fileset for SDD 1.2.0.x or later (SCSI and fibre) and ibmSdd_510.rte fileset for SDD 1.3.0.x or later (SCSI and fibre)

3. The ibmSdd_433.rte fileset for SDD 1.2.2.x (or later) and the ibmSdd_510nchacmp.rte fileset for SDD 1.3.0.x are for HACMP/6000 environments only; these versions support non-concurrent and concurrent modes. However, in order to make the best use of the manner in which the device reserves are made, IBM recommends that you:
   • Use either ibmSdd_432.rte fileset for SDD 1.2.2.x or 1.3.0.x, or the ibmSdd_510.rte fileset for SDD 1.3.0.x when running HACMP in concurrent mode.
   • Use either ibmSdd_433.rte fileset for SDD 1.2.2.x or 1.3.0.x, or the ibmSdd_510nchacmp.rte fileset for SDD 1.3.0.x when running HACMP in non-concurrent mode.

4. HACMP/6000 is not supported on all models of the ESS.

5. For information about supported ESS models and required ESS microcode levels, go to the following website: www.ibm.com/storage/support/techsup/swtechsup.nsf/support/sddupdates

Chapter 3. Using SDD on an AIX host system

This chapter provides instructions for using the Subsystem Device Driver. It shows you how to configure SDD to provide I/O load-balancing and path failover protection.

Providing load-balancing and failover protection

SDD provides load-balancing and failover protection for AIX applications and for the LVM when ESS vpath devices are used. These devices must have a minimum of two paths to a physical logical unit number (LUN) for failover protection to exist.

Displaying the ESS vpath device configuration

To provide failover protection, an ESS vpath device must include a minimum of two paths. Both the SDD vpath device and the ESS hdisk devices must all be in the Available condition. In the following example, vpath0, vpath1, and vpath2 all have a single path and, therefore, will not provide failover protection because there is no alternate path to the ESS LUN. The other SDD vpath devices have two paths and, therefore, can provide failover protection.

To display which ESS vpath devices are available to provide failover protection, use either the Display Data Path Device Configuration SMIT panel, or run the lsvpcfg command. Perform the following steps to use SMIT:

1. Type smitty device from your desktop window. The Devices panel is displayed.
2. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
3. Select Display Data Path Device Configuration and press Enter.

You will see an output similar to the following:

Figure 3. Output from the Display Data Path Device Configuration SMIT panel

+--------------------------------------------------------------------------------+
|vpath0 (Avail pv vpathvg) 018FA067 = hdisk1 (Avail )                            |
|vpath1 (Avail ) 019FA067= hdisk2 (Avail )                                       |
|vpath2 (Avail ) 01AFA067 = hdisk3 (Avail )                                      |
|vpath3 (Avail ) 01BFA067 = hdisk4 (Avail ) hdisk27 (Avail )                     |
|vpath4 (Avail ) 01CFA067 = hdisk5 (Avail ) hdisk28 (Avail )                     |
|vpath5 (Avail ) 01DFA067 = hdisk6 (Avail ) hdisk29 (Avail )                     |
|vpath6 (Avail ) 01EFA067 = hdisk7 (Avail ) hdisk30 (Avail )                     |
|vpath7 (Avail ) 01FFA067 = hdisk8 (Avail ) hdisk31 (Avail )                     |
|vpath8 (Avail ) 020FA067 = hdisk9 (Avail ) hdisk32 (Avail )                     |
|vpath9 (Avail pv vpathvg) 02BFA067 = hdisk20 (Avail ) hdisk44 (Avail )          |
|vpath10 (Avail pv vpathvg) 02CFA067 = hdisk21 (Avail ) hdisk45 (Avail )         |
|vpath11 (Avail pv vpathvg) 02DFA067 = hdisk22 (Avail ) hdisk46 (Avail )         |
|vpath12 (Avail pv vpathvg) 02EFA067 = hdisk23 (Avail ) hdisk47 (Avail )         |
|vpath13 (Avail pv vpathvg) 02FFA067 = hdisk24 (Avail ) hdisk48 (Avail )         |
+--------------------------------------------------------------------------------+

The following information is displayed:

• The name of each SDD vpath device, such as vpath1.
• The configuration condition of the SDD vpath device. It is either Defined or Available. There is no failover protection if only one path is in the Available condition. At least two paths to each SDD vpath device must be in the Available condition to have failover protection.

  Example of vpath devices with or without failover protection: vpath0, vpath1, or vpath2 has a single path and therefore does not have failover protection. The other ESS vpath devices each have two paths and thus can provide failover protection. The requirement for failover protection is that the ESS vpath device, and at least two hdisk devices making it up, must be in the Available condition.

  Attention: The configuration condition also indicates whether or not the SDD vpath device is defined to AIX as a physical volume (pv flag). If pv is displayed for both SDD vpath devices and ESS hdisk devices, you might not have failover protection. Run the dpovgfix command to fix this problem.

• The name of the volume group to which the device belongs, such as vpathvg
• The unit serial number of the ESS LUN, such as 019FA067
• The names of the AIX disk devices that comprise the SDD vpath devices, their configuration conditions, and the physical volume status.

You can also use the datapath command to display information about a SDD vpath device. This command displays the number of paths to the device. For example, the datapath query device 10 command might produce this output:

+--------------------------------------------------------------------------------+
|DEV#:  10  DEVICE NAME: vpath10   TYPE: 2105B09   SERIAL: 02CFA067              |
|==================================================================              |
|Path#      Adapter/Hard Disk   State     Mode    Select        Errors           |
|    0          scsi6/hdisk21    OPEN   NORMAL        44             0           |
|    1          scsi5/hdisk45    OPEN   NORMAL        43             0           |
+--------------------------------------------------------------------------------+

The sample output shows that device vpath10 has two paths and both are operational.

Configuring a volume group for failover protection

You can create a volume group with SDD vpath devices using the Logical Volume Groups SMIT panel. Choose the SDD vpath devices that have failover protection for the volume group.

It is possible to create a volume group that has only a single path (see Figure 3) and then add paths later by reconfiguring the ESS. (See Adding paths to SDD devices of a volume group for information about adding paths to a SDD device.) However, a SDD volume group does not have failover protection if any of its physical volumes only has a single path.

Perform the following steps to create a new volume group with SDD vpaths:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This procedure uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Volume Group panel is displayed.
4. Select Volume Group and press Enter. The Add Volume Group with Data Path Devices panel is displayed.
5. Select Add Volume Group with Data Path Devices and press Enter.

   Note:

   Press F4 while highlighting the PHYSICAL VOLUME names field to list all the available SDD vpaths.

If you use a script file to create a volume group with SDD vpath devices, you must modify your script file and replace the mkvg command with the mkvg4vp command.

All the functions that apply to a regular volume group also apply to a SDD volume group. Use SMIT to create a logical volume group (mirrored, striped, or compressed) or a file system (mirrored, striped, or compressed) on a SDD volume group.

Once you create the volume group, AIX creates the SDD vpath device as a physical volume (pv). In Figure 3, vpath9 through vpath13 are included in a volume group and they become physical volumes. To list all the physical volumes known to AIX, use the lspv command. Any ESS vpath devices that were created into physical volumes are included in the output similar to the following:

+--------------------------------------------------------------------------------+
|hdisk0         0001926922c706b2    rootvg                                       |
|hdisk1         none                None                                         |
|...                                                                             |
|hdisk10        none                None                                         |
|hdisk11        00000000e7f5c88a    None                                         |
|...                                                                             |
|hdisk48        none                None                                         |
|hdisk49        00000000e7f5c88a    None                                         |
|vpath0         00019269aa5bc858    None                                         |
|vpath1         none                None                                         |
|vpath2         none                None                                         |
|vpath3         none                None                                         |
|vpath4         none                None                                         |
|vpath5         none                None                                         |
|vpath6         none                None                                         |
|vpath7         none                None                                         |
|vpath8         none                None                                         |
|vpath9         00019269aa5bbadd    vpathvg                                      |
|vpath10        00019269aa5bc4dc    vpathvg                                      |
|vpath11        00019269aa5bc670    vpathvg                                      |
|vpath12        000192697f9fd2d3    vpathvg                                      |
|vpath13        000192697f9fde04    vpathvg                                      |
+--------------------------------------------------------------------------------+

To display the devices that comprise a volume group, enter the lsvg -p vg-name command. For example, the lsvg -p vpathvg command might produce the following output:

+--------------------------------------------------------------------------------+
|PV_NAME           PV STATE    TOTAL PPs   FREE PPs    FREE DISTRIBUTION         |
|vpath9            active      29          4           00..00..00..00..04        |
|vpath10           active      29          4           00..00..00..00..04        |
|vpath11           active      29          4           00..00..00..00..04        |
|vpath12           active      29          4           00..00..00..00..04        |
|vpath13           active      29          28          06..05..05..06..06        |
+--------------------------------------------------------------------------------+

The example output indicates that the vpathvg volume group uses physical volumes vpath9 through vpath13.

Importing a volume group with SDD

You can import a new volume group definition from a set of physical volumes with SDD vpath devices using the Volume Groups SMIT panel.

Note:

To use this command, you must either have root user authority or be a member of the system group.

Attention:

SDD does not automatically create the pvid attribute in the ODM database for each vpath device. The AIX disk driver automatically creates the pvid attribute in the ODM database, if a pvid exists on the physical device. Therefore, the first time you import a new SDD volume group to a new cluster node, you must import the volume group using hdisks as physical volumes. Next, run the hd2vp conversion script (see SDD utility programs) to convert the volume group's physical volumes from ESS hdisks to vpath devices. This conversion step not only creates pvid attributes for all vpath devices which belong to that imported volume group, it also deletes the pvid attributes for these vpath devices' underlying hdisks. Later on you can import and vary on the volume group directly from the vpath devices. These special requirements apply to both concurrent and non-concurrent volume groups.

Under certain conditions, the state of a pvid on a system is not always as we expected. So it is necessary to determine the state of a pvid as displayed by the lsvp command, in order to select the appropriate action.

There are four scenarios:

Scenario 1. lspv displays pvid's for both hdisks and vpath:

	>lspv
	hdisk1		003dfc10a11904fa	None
	hdisk2		003dfc10a11904fa	None
	vpath0		003dfc10a11904fa	None

Scenario 2. lspv displays pvid's for hdisks only:

	>lspv
	hdisk1		003dfc10a11904fa	None
	hdisk2		003dfc10a11904fa	None
	vpath0		none					None

For both Scenario 1 and Scenario 2, the volume group should be imported using the hdisk names and then converted using the hd2vp command:

	>importvg -y vg_name -V major# hdisk1
	>hd2vp vg_name

Scenario 3. lspv displays the pvid for vpath only:

	>lspv
	hdisk1		none					None
	hdisk2		none					None
	vpath0		003dfc10a11904fa	None

For Scenario 3, the volume group should be imported using the vpath name:

	>importvg -y vg_name -V major# vpath0

Scenario 4. lspv does not display the pvid on the hdisks or the vpath:

	>lspv
	hdisk1		none			None
	hdisk2		none			None
	vpath0		none			None

For Scenario 4, the pvid will need to be placed in the ODM for the vpath devices and then the volume group can be imported using the vpath name:

	>chdev -l vpath0 -a pv=yes
	>importvg -y vg_name -V major# vpath0

See "Special requirements" for special requirements regarding unconfiguring and removing the ibmSdd_433.rte or ibmSdd_510nchacmp.rte filesets for SDD 1.3.0.x vpath devices.

Perform the following steps to import a volume group with SDD devices:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Volume Group panel is displayed.
4. Select Volume Groups and press Enter. The Volume Groups panel is displayed.
5. Select Import a Volume Group and press Enter. The Import a Volume Group panel is displayed.
6. In the Import a Volume Group panel, perform the following tasks:
   • Type in the volume group you want to import.
   • Type in the physical volumes that you want to import over.
   • Press Enter after making all desired changes.

   You can press the F4 key for a list of choices.

Exporting a volume group with SDD

You can export a volume group definition from a set of physical volumes with SDD vpath devices using the Volume Groups SMIT panel.

The exportvg command removes the definition of the volume group specified by the Volume Group parameter from the system. Since all system knowledge of the volume group and its contents are removed, an exported volume group is no longer accessible. The exportvg command does not modify any user data in the volume group.

A volume group is a nonshared resource within the system; it should not be accessed by another system until it has been explicitly exported from its current system and imported on another. The primary use of the exportvg command, coupled with the importvg command, is to allow portable volumes to be exchanged between systems. Only a complete volume group can be exported, not individual physical volumes.

Using the exportvg command and the importvg command, you can also switch ownership of data on physical volumes shared between two systems.

Note:

To use this command, you must either have root user authority or be a member of the system group.

Perform the following steps to export a volume group with SDD devices:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Volume Group panel is displayed.
4. Select Volume Groups and press Enter. The Volume Groups panel is displayed.
5. Select Export a Volume Group and press Enter. The Export a Volume Group panel is displayed.
6. Type in the volume group to export and press Enter.

You can use the F4 key to select which volume group you want to export.

How failover protection can be lost

AIX can only create volume groups from disk (or pseudo) devices that are physical volumes. If a volume group is created using a device that is not a physical volume, AIX makes it a physical volume as part of the procedure of creating the volume group. A physical volume has a physical volume identifier (pvid) written on its sector 0 and also has a pvid attribute attached to the device attributes in the CuAt ODM. The lspv command lists all the physical volumes known to AIX. Here is a sample output from this command:

+--------------------------------------------------------------------------------+
|hdisk0         0001926922c706b2    rootvg                                       |
|hdisk1         none                None                                         |
|...                                                                             |
|hdisk10        none                None                                         |
|hdisk11        00000000e7f5c88a    None                                         |
|...                                                                             |
|hdisk48        none                None                                         |
|hdisk49        00000000e7f5c88a    None                                         |
|vpath0         00019269aa5bc858    None                                         |
|vpath1         none                None                                         |
|vpath2         none                None                                         |
|vpath3         none                None                                         |
|vpath4         none                None                                         |
|vpath5         none                None                                         |
|vpath6         none                None                                         |
|vpath7         none                None                                         |
|vpath8         none                None                                         |
|vpath9         00019269aa5bbadd    vpathvg                                      |
|vpath10        00019269aa5bc4dc    vpathvg                                      |
|vpath11        00019269aa5bc670    vpathvg                                      |
|vpath12        000192697f9fd2d3    vpathvg                                      |
|vpath13        000192697f9fde04    vpathvg                                      |
+--------------------------------------------------------------------------------+

In some cases, access to data is not lost, but failover protection might not be present. Failover protection can be lost in several ways:

1. Through the loss of a device path
2. By creating a volume group from single-path vpath (pseudo) devices
3. As a side effect of running the disk change method
4. Through running the mksysb restore command
5. By manually deleting devices and running the configuration manager (cfgmgr)

The following sections provide more information about the ways that failover protection can be lost.

Through the loss of a device path

Due to hardware errors, SDD might remove one or more paths to a vpath pseudo device. A pseudo devices loses failover protection when it only has a single path. You can use the datapath query device command to show the state of paths to a pseudo device. You cannot use any Dead path for I/O operations.

By creating a volume group from single-path vpath (pseudo) devices

A volume group created using any single-path pseudo devices does not have failover protection because there is no alternate path to the ESS LUN.

As a side effect of running the disk change method

It is possible to modify attributes for an hdisk device by running the chdev command. The chdev command invokes the hdisk configuration method to make the requested change. In addition, the hdisk configuration method sets the pvid attribute for an hdisk if it determines that the hdisk has a pvid written on sector 0 of the LUN. This causes the vpath pseudo device and one or more of its hdisks to have the same pvid attribute in the ODM. If the volume group containing the vpath pseudo device is activated, the LVM uses the first device it finds in the ODM with the desired pvid to activate the volume group.

As an example, if you issue the lsvpcfg command, the following output is displayed:

+--------------------------------------------------------------------------------+
|vpath0 (Avail pv vpathvg) 018FA067 = hdisk1 (Avail )                            |
|vpath1 (Avail ) 019FA067 = hdisk2 (Avail )                                      |
|vpath2 (Avail ) 01AFA067 = hdisk3 (Avail )                                      |
|vpath3 (Avail ) 01BFA067 = hdisk4 (Avail ) hdisk27 (Avail )                     |
|vpath4 (Avail ) 01CFA067 = hdisk5 (Avail ) hdisk28 (Avail )                     |
|vpath5 (Avail ) 01DFA067 = hdisk6 (Avail ) hdisk29 (Avail )                     |
|vpath6 (Avail ) 01EFA067 = hdisk7 (Avail ) hdisk30 (Avail )                     |
|vpath7 (Avail ) 01FFA067 = hdisk8 (Avail ) hdisk31 (Avail )                     |
|vpath8 (Avail ) 020FA067 = hdisk9 (Avail ) hdisk32 (Avail )                     |
|vpath9 (Avail pv vpathvg) 02BFA067 = hdisk20 (Avail ) hdisk44 (Avail )          |
|vpath10 (Avail pv vpathvg) 02CFA067 = hdisk21 (Avail ) hdisk45 (Avail )         |
|vpath11 (Avail pv vpathvg) 02DFA067 = hdisk22 (Avail ) hdisk46 (Avail )         |
|vpath12 (Avail pv vpathvg) 02EFA067 = hdisk23 (Avail ) hdisk47 (Avail )         |
|vpath13 (Avail pv vpathvg) 02FFA067 = hdisk24 (Avail ) hdisk48 (Avail )         |
+--------------------------------------------------------------------------------+

The following example of a chdev command could also set the pvid attribute for an hdisk:

chdev -l hdisk46 -a queue_depth=30

For this example, the output of the lsvpcfg command would look similar to this:

+--------------------------------------------------------------------------------+
|vpath0 (Avail pv vpathvg) 018FA067 = hdisk1 (Avail )                            |
|vpath1 (Avail ) 019FA067 = hdisk2 (Avail )                                      |
|vpath2 (Avail ) 01AFA067 = hdisk3 (Avail )                                      |
|vpath3 (Avail ) 01BFA067 = hdisk4 (Avail ) hdisk27 (Avail )                     |
|vpath4 (Avail ) 01CFA067 = hdisk5 (Avail ) hdisk28 (Avail )                     |
|vpath5 (Avail ) 01DFA067 = hdisk6 (Avail ) hdisk29 (Avail )                     |
|vpath6 (Avail ) 01EFA067 = hdisk7 (Avail ) hdisk30 (Avail )                     |
|vpath7 (Avail ) 01FFA067 = hdisk8 (Avail ) hdisk31 (Avail )                     |
|vpath8 (Avail ) 020FA067 = hdisk9 (Avail ) hdisk32 (Avail )                     |
|vpath9 (Avail pv vpathvg) 02BFA067 = hdisk20 (Avail ) hdisk44 (Avail )          |
|vpath10 (Avail pv vpathvg) 02CFA067 = hdisk21 (Avail ) hdisk45 (Avail )         |
|vpath11 (Avail pv vpathvg) 02DFA067 = hdisk22 (Avail ) hdisk46 (Avail pv vpathvg|
|vpath12 (Avail pv vpathvg) 02EFA067 = hdisk23 (Avail ) hdisk47 (Avail )         |
|vpath13 (Avail pv vpathvg) 02FFA067 = hdisk24 (Avail ) hdisk48 (Avail )         |
+--------------------------------------------------------------------------------+

The output of the lsvpcfg command shows that vpath11 contains hdisk22 and hdisk46. However, hdisk46 is the one with the pv attribute set. If you run the lsvg -p vpathvg command again, you might see something like this:

+--------------------------------------------------------------------------------+
|vpathvg:                                                                        |
|PV_NAME           PV STATE    TOTAL PPs   FREE PPs    FREE DISTRIBUTION         |
|vpath10           active      29          4           00..00..00..00..04        |
|hdisk46           active      29          4           00..00..00..00..04        |
|vpath12           active      29          4           00..00..00..00..04        |
|vpath13           active      29          28          06..05..05..06..06        |
+--------------------------------------------------------------------------------+

Notice that now device vpath11 has been replaced by hdisk46. That is because hdisk46 is one of the hdisk devices included in vpath11 and it has a pvid attribute in the ODM. In this example, the LVM used hdisk46 instead of vpath11 when it activated volume group vpathvg. The volume group is now in a mixed mode of operation because it partially uses vpath pseudo devices and partially uses hdisk devices. This is a problem that must be fixed because failover protection is effectively disabled for the vpath11 physical volume of the vpathvg volume group.

Note:

The way to fix this problem with the mixed volume group is to run the dpovgfix vg-name command after running the chdev command.

Through running the

If a system is restored from a mksysb restore file or tape, the vpath pseudo device pvid attribute is not set. All logical volumes made up of vpath pseudo devices use hdisk devices instead of vpath devices. You can correct the problem by using the hd2vp shell script to convert the volume group back to using vpath devices.

By manually deleting devices and running the configuration manager

rmdev -dl vpath3 rmdev -dl hdisk4 rmdev -dl hdisk27
cfgmgr

Recovering from mixed volume groups

Run the dpovgfix shell script to recover a mixed volume group. The syntax is dpovgfix vg-name. The script tries to find a pseudo device corresponding to each hdisk in the volume group and replaces the hdisk with the vpath pseudo device. In order for the shell script to be executed, all mounted file systems of this volume group have to be unmounted. After successful completion of the dpovgfix shell script, mount the file systems again.

Extending an existing SDD volume group

You can extend a volume group with SDD vpath devices using the Logical Volume Groups SMIT panel. The SDD vpath devices to be added to the volume group should be chosen from those that can provide failover protection. It is possible to add a SDD vpath device to a SDD volume group that has only a single path (vpath0 in Figure 3) and then add paths later by reconfiguring the ESS. With a single path, failover protection is not provided. (See Adding paths to SDD devices of a volume group for information about adding paths to a SDD device.)

Perform the following steps to extend a volume group with SDD devices:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Volume Group panel is displayed.
4. Select Volume Group and press Enter. The Add Volume Group with Data Path Devices panel is displayed.
5. Select Add Volume Group with Data Path Devices and press Enter.
6. Type in the volume group name and physical volume name and press Enter. You can also use the F4 key to list all the available SDD devices, and you can select the devices you want to add to the volume group.

If you use a script file to extend an existing SDD volume group, you must modify your script file and replace the extendvg command with the extendvg4vp command.

Backing-up all files belonging to a Subsystem Device Driver volume group

You can back up all files belonging to a specified volume group with Subsystem Device Driver vpath devices using the Volume Groups SMIT panel.

To backup a volume group with SDD devices, go to Accessing the Back Up a Volume Group with Data Path Devices SMIT panel.

If you use a script file to back up all files belonging to a specified SDD volume group, you must modify your script file and replace the savevg command with the savevg4vp command.

Attention: Backing-up files (running the savevg4vp command) will result in the loss of all material previously stored on the selected output medium. Data integrity of the archive may be compromised if a file is modified during system backup. Keep system activity at a minimum during the system backup procedure.

Restoring all files belonging to a Subsystem Device Driver volume group

You can restore all files belonging to a specified volume group with Subsystem Device Driver vpath devices using the Volume Groups SMIT panel.

To restore a volume group with SDD devices and go to Accessing the Remake a Volume Group with Data Path Devices SMIT panel.

If you use a script file to restore all files belonging to a specified SDD volume group, you must modify your script file and replace the restvg command with the restvg4vp command.

SDD-specific SMIT panels

SDD supports several special SMIT panels. Some SMIT panels provide SDD-specific functions, while other SMIT panels provide AIX functions (but requires SDD-specific commands). For example, the Add a Volume Group with Data Path Devices function uses the SDD mkvg4vp command, instead of the AIX mkvg command. Table 11 lists the SDD-specific SMIT panels and how you can use them.

Table 11. SDD-specific SMIT panels and how to proceed

SMIT panels	How to proceed:
Display Data Path Device Configuration	Go to: Accessing the Display Data Path Device Configuration SMIT panel
Display Data Path Device Status	Accessing the Display Data Path Device Status SMIT panel
Display Data Path Device Adapter Status	Accessing the Display Data Path Device Adapter Status SMIT panel
Define and Configure all Data Path Devices	Accessing the Define and Configure All Data Path Devices SMIT panel
Add Paths to Available Data Path Devices Note: SDD does not support the addpaths command for AIX 4.2.1; it supports this command for AIX 4.3.2 or higher.	Accessing the Add Paths to Available Data Path Devices SMIT panel
Configure a Defined Data Path Device	Accessing the Configure a Defined Data Path Device SMIT panel
Remove a Data Path Device	Accessing the Remove a Data Path Device SMIT panel
Add a Volume Group with Data Path Devices	Accessing the Add a Volume Group with Data Path Devices SMIT panel
Add a Data Path Volume to a Volume Group	Accessing the Add a Data Path Volume to a Volume Group SMIT panel
Remove a copy from a datapath Logical Volume	Accessing the Remove a copy from a datapath Logical Volume SMIT panel
Back Up a Volume Group with Data Path Devices	Accessing the Back Up a Volume Group with Data Path Devices SMIT panel
Remake a Volume Group with Data Path Devices	Accessing the Remake a Volume Group with Data Path Devices SMIT panel

Accessing the Display Data Path Device Configuration SMIT panel

Perform the following steps to access the Display Data Path Device Configuration panel:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select Devices and press Enter. The Devices panel is displayed.
3. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
4. Select Display Data Path Device Configuration and press Enter.

Accessing the Display Data Path Device Status SMIT panel

Perform the following steps to access the Display Data Path Device Status panel:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type SMIT to invoke the graphical user interface.

2. Select Devices and press Enter. The Devices panel is displayed.
3. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
4. Select Display Data Path Device Status and press Enter.

Accessing the Display Data Path Device Adapter Status SMIT panel

Perform the following steps to access the Display Data Path Device Status panel:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select Devices and press Enter. The Devices panel is displayed.
3. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
4. Select Display Data Path Device Status and press Enter.

Accessing the Define and Configure All Data Path Devices SMIT panel

To access the Define and Configure All Data Path Devices panel, perform the following steps:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select Devices and press Enter. The Devices panel is displayed.
3. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
4. Select Define and Configure All Data Path Devices and press Enter.

Accessing the Configure a Defined Data Path Device SMIT panel

Perform the following steps to access the Configure a Defined Data Path Device panel:

1. Type SMITTY from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type SMIT to invoke the graphical user interface.

2. Select Devices and press Enter. The Devices panel is displayed.
3. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
4. Select Configure a Defined Data Path Device and press Enter.

Accessing the Remove a Data Path Device SMIT panel

Perform the following steps to access the Remove a Data Path Device panel:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select Devices and press Enter. The Devices panel is displayed.
3. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
4. Select Remove a Data Path Device and press Enter.

Accessing the Add a Volume Group with Data Path Devices SMIT panel

Perform the following steps to access the Add a volume group with data path devices panel:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Volume Group panel is displayed.
4. Select Volume Groups and press Enter. The Add Volume Group with Data Path Devices panel is displayed.
5. Select Add Volume Group with Data Path Devices and press Enter.

   Note:

   Press F4 while highlighting the PHYSICAL VOLUME names field to list all the available SDD vpaths.

Accessing the Add a Data Path Volume to a Volume Group SMIT panel

Perform the following steps to access the Add a Data Path Volume to a Volume Group panel:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical) and press Enter. The System Storage Management (Physical & Logical) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Logical Volume Manager panel is displayed.
4. Select Volume Group and press Enter. The Volume Group panel is displayed.
5. Select Add a Data Path Volume to a Volume Group and press Enter.
6. Type the volume group name and physical volume name and press Enter. Alternately, you can use the F4 key to list all the available SDD vpath devices and use the F7 key to select the physical volumes you want to add.

Accessing the Remove a copy from a datapath Logical Volume SMIT panel

Perform the following steps to access the Remove a copy from a datapath Logical Volume panel:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select Logical Volume manager and press Enter. The Logical Volume manager panel is displayed.
3. Select Volume Groups and press Enter. The Volume Groups panel is displayed.
4. Select Set Characteristics of a Volume Group and press Enter. The Set Characteristics of a Volume Group panel is displayed.
5. Select Remove a Copy from a datapath Logical Volume and press Enter. The Remove a Physical Volume from a Volume Group panel is displayed.

Accessing the Back Up a Volume Group with Data Path Devices SMIT panel

Perform the following steps to access the Back Up a Volume Group with Data Path Devices panel and to backup a volume group with SDD devices:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Volume Group panel is displayed.
4. Select Volume Groups and press Enter. The Volume Groups panel is displayed.
5. Select Back Up a Volume Group with Data Path Devices and press Enter. The Back Up a Volume Group with Data Path Devices panel is displayed.
6. In the Back Up a Volume Group with Data Path Devices panel, perform the following steps:
   • Type in the Backup DEVICE or FILE name.
   • Type in the Volume Group to back up.
   • Press Enter after making all desired changes.

   Tip: You can also use the F4 key to list all the available SDD devices, and you can select the devices or files you want to backup. Attention: Backing-up files (running the savevg4vp command) will result in the loss of all material previously stored on the selected output medium. Data integrity of the archive may be compromised if a file is modified during system backup. Keep system activity at a minimum during the system backup procedure.

Accessing the Remake a Volume Group with Data Path Devices SMIT panel

Perform the following steps to access the Remake a Volume Group with Data Path Devices panel and restore a volume group with SDD devices:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
3. Select Logical Volume Manager and press Enter. The Volume Group panel is displayed.
4. Select Volume Groups and press Enter. The Volume Groups panel is displayed.
5. Select Remake a Volume Group with Data Path Devices and press Enter. The Remake a Volume Group with Data Path Devices panel is displayed.
6. Type in the Restore DEVICE or FILE name, and press Enter. You can also use the F4 key to list all the available SDD devices, and you can select the devices or files you want to restore.

SDD utility programs

For more information about this command, go to Adding paths to SDD devices of a volume group.

hd2vp and vp2hd

SDD provides two conversion scripts, hd2vp and vp2hd. The hd2vp script converts a volume group from ESS hdisks into SDD vpaths, and the vp2hd script converts a volume group from SDD vpaths into ESS hdisks. Use the vp2hd program when you want to configure your applications back to original ESS hdisks, or when you want to remove the SDD from your AIX host system.

Note:

You must convert all your applications and volume groups to the original ESS hdisk device special files before removing SDD.

The syntax for these conversion scripts is as follows:

hd2vp  vgname
vp2hd  vgname

These two conversion programs require that a volume group contain either all original ESS hdisks or all SDD vpaths. The program fails if a volume group contains both kinds of device special files (mixed volume group).

Tip: Always use SMIT to create a volume group of SDD devices. This avoids the problem of a mixed volume group.

dpovgfix

You can use the dpovgfix script tool to recover mixed volume groups.

Performing AIX system management operations on adapters and ESS hdisk devices might cause original ESS hdisks to be contained within a SDD volume group. This is known as a mixed volume group. Mixed volume groups happen when a SDD volume group is inactivated (varied off), and certain AIX commands to the hdisk put the pvid attribute of hdisk back into the ODM database. The following is an example of a command that does this:

   chdev -1 hdiskN -a queue_depth=30 

If this disk is an active hdisk of a vpath that belongs to a SDD volume group, and you run the varyonvg command to activate this SDD volume group, LVM might pick up the hdisk device rather than the vpath device. The result is that a SDD volume group partially uses SDD vpath devices, and partially uses ESS hdisk devices. The result is the volume group loses path failover capability for that physical volume. The dpovgfix script tool fixes this problem. The command syntax is:

dpovgfix vg-name

lsvpcfg

You can use the lsvpcfg script tool to display the configuration status of SDD devices. This displays the configuration status for all SDD devices. The lsvpcfg command can be issued in two ways.

1. The command can be issued without parameters. The command syntax is:

   lsvpcfg
   

   See Verifying the SDD configuration for an example of the output and what it means.

2. The command can also be issued using the vpath device name as a parameter. The command syntax is:

   lsvpcfg vpathN0 vpathN1 vpathN2
   

   You will see output similar to this:

   +--------------------------------------------------------------------------------+
   |vpath10 (Avail pv ) 13916392 = hdisk95 (Avail ) hdisk179 (Avail )               |
   |vpath20 (Avail ) 02816392 = hdisk23 (Avail ) hdisk106 (Avail )                  |
   |vpath30 (Avail ) 10516392 = hdisk33 (Avail ) hdisk116 (Avail )                  |
   +--------------------------------------------------------------------------------+

   See Verifying the SDD configuration for an explanation of the output.

mkvg4vp

You can use the mkvg4vp command to create a SDD volume group. For more information about this command, go to Configuring a volume group for failover protection.

extendvg4vp

You can use the extendvg4vp command to extend an existing SDD volume group. For more information about this command, go to Extending an existing SDD volume group.

Using ESS devices directly

After you configure the SDD, it creates SDD devices (vpath devices) for ESS LUNs. ESS LUNs are accessible through the connection between the AIX host server SCSI or FCP adapter and the ESS ports. The AIX disk driver creates the original or ESS devices (hdisks). Therefore, with SDD, an application now has two ways in which to access ESS devices.

To use the SDD load-balancing and failover features and access ESS devices, your application must use the SDD vpath devices rather than the ESS hdisk devices.

Two types of applications use ESS disk storage. One type of application accesses ESS devices through the SDD vpath device (raw device). The other type of application accesses ESS devices through the AIX logical volume manager (LVM). For this type of application, you must create a volume group with the SDD vpath devices.

If your application used ESS hdisk device special files directly before installing SDD, convert it to using the SDD vpath device special files. After installing SDD, perform the following steps:

1. Type smitty from your desktop window. The System Management Interface Tool is displayed.

   Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

2. Select Devices and press Enter. The Devices panel is displayed.
3. Select Data Path Devices and press Enter. The Data Path Devices panel is displayed.
4. Select Display Data Path Device Configuration. The system displays all SDD vpaths with their attached multiple paths (hdisks).
5. Search the list of hdisks to locate the hdisks your application is using.
6. Replace each hdisk with its corresponding SDD vpath device.

   Note:

   Depending upon your application, the manner in which you replace these files is different. If this is a new application, use the SDD vpath rather than hdisk to use the SDD load-balancing and failover features.

Note:

Alternately, you can type lsvpcfg from the command-line interface rather than using SMIT. This displays all configured SDD vpath devices and their underlying paths (hdisks).

Using ESS devices through AIX LVM

Attention:

• You must use the System Management Interface Tool (SMIT). The SMIT facility runs in two interfaces, nongraphical (type smitty to invoke the nongraphical user interface) or graphical (type SMIT to invoke the graphical user interface).
• Do not use the mkvg command directly. Otherwise, the path failover capability could be lost.

If your application accesses ESS devices through LVM, determine the volume group that it uses before you convert volume groups. Then, perform the following steps to convert the volume group from the original ESS device hdisks to the SDD vpaths:

1. Determine the file systems or logical volumes that your application accesses.
2. Type smitty from your desktop window. The System Management Interface Tool is displayed.
3. Select System Storage Management (Physical & Logical Storage) and press Enter. The System Storage Management (Physical & Logical Storage) panel is displayed.
4. Select Logical Volume Manager and press Enter. The Logical Volume Manager panel is displayed.
5. Select Logical Volume and press Enter. The Logical Volume panel is displayed.
6. Select List All Logical Volumes by Volume Group to determine the logical volumes that belong to this volume group and their logical volume mount points.
7. Press Enter. The logical volumes are listed by volume group.

   To determine the file systems, perform the following steps:

   1. Type smitty from your desktop window. The System Management Interface Tool is displayed.
   2. Select Logical Volume Manager and press Enter. The Logical Volume Manager panel is displayed.
   3. Select File Systems and press Enter. The File Systems panel is displayed.
   4. Select List All File Systems to locate all file systems that have the same mount points as the logical volumes.
   5. Press Enter. The file systems are listed.
   6. Note the file system name of that volume group and the file system mount point, if it is mounted.
   7. Unmount these file systems.
8. Enter the following to convert the volume group from the original ESS hdisks to SDD vpaths:

     hd2vp vgname
   

9. When the conversion is complete, mount all file systems that you previously unmounted.

When the conversion is complete, your application now accesses ESS physical LUNs through SDD vpath devices. This provides load balancing and failover protection for your application.

Migrating a non-SDD volume group to an ESS SDD multipath volume group in concurrent mode

Before you migrate your non-SDD volume group to a SDD volume group, make sure that you have completed the following tasks:

• The SDD for the AIX host system is installed and configured. To see if SDD is installed, issue one of the following commands:

  lslpp -l ibmSdd_421.rte
  lslpp -l ibmSdd_432.rte
  lslpp -l ibmSdd_433.rte
  lslpp -l ibmSdd_510.rte
  lslpp -l ibmSdd_510nchacmp.rte 
  

  An example of output from the lslpp command is:

  +--------------------------------------------------------------------------------+
  |Fileset                      Level    State        Description                  |
  |--------------------------------------------------------------------------------|
  |Path: /usr/lib/objrepos                                                         |
  |  ibmSdd_432.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
  |                                                   AIX V432 V433 for concurrent |
  |                                                   HACMP                        |
  |                                                                                |
  |Path: /etc/objrepos                                                             |
  |  ibmSdd_432.rte             1.3.0.x  COMMITTED    IBM Subsystem Device Driver  |
  |                                                   AIX V432 V433 for concurrent |
  |                                                   HACMP                        |
  |                                                                                |
  +--------------------------------------------------------------------------------+

• The ESS subsystem devices to which you want to migrate have multiple paths configured per LUN. To check the status of your SDD configuration, use the System Management Interface Tool (SMIT) or issue the lsvpcfg command from the command line. To use SMIT:
  • Type smitty and press Enter from your desktop window. The System Management Interface Tool panel is displayed.

    Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

  • Select Devices and press Enter. The Devices panel is displayed.
  • Select Data Path Device and press Enter. The Data Path Device panel is displayed.
  • Select Display Data Path Device Configuration and press Enter. A list is displayed of the pseudo devices and whether there are multiple paths configured for the devices.
• Make sure the SDD vpath devices you are going to migrate to do not belong to any other volume group, and that the corresponding physical device (ESS LUN) does not have a pvid written on it. Use the lsvpcfg command output to check the SDD vpath devices that you are going to use for migration. Make sure there is no pv displayed for this vpath and its paths (hdisks). If a LUN has never belonged to any volume group, there is no pvid written on it. In case there is a pvid written on the LUN and the LUN does not belong to any volume group, you need to clear the pvid from the LUN before using it to migrate a volume group. The commands to clear the pvid are:

  chdev -l  hdiskN  -a  pv=clear
  chdev -l  vpathN  -a  pv=clear
  

  Attention: Exercise care when clearing a pvid from a device with this command. Issuing this command to a device that does belong to an existing volume group can cause system failures.

You should complete the following steps to migrate a non-SDD volume group to a multipath SDD volume group in concurrent mode:

1. Add new SDD vpath devices to an existing non-SDD volume group:
   1. Type smitty and press Enter from your desktop window. The System Management Interface Tool panel is displayed.

      Tip: The SMIT facility runs in two interfaces, nongraphical and graphical. This step uses the nongraphical interface. You can type smit to invoke the graphical user interface.

   2. Select System Storage Management (Physical & Logical) and press Enter. The System Storage Management (Physical & Logical) panel is displayed.
   3. Select Logical Volume Manager and press Enter. The Logical Volume Manager panel is displayed.
   4. Select Volume Group and press Enter. The Volume Group panel is displayed.
   5. Select Add a Data Path Volume to a Volume Group and press Enter.
   6. Type the volume group name and physical volume name and press Enter. Alternately, you can use the F4 key to list all the available SDD vpath devices and use the F7 key to select the physical volumes you want to add.
2. Mirror logical volumes from the original volume to a Subsystem Device Driver ESS volume. Use the command:

   smitty mklvcopy  
   

   Use the new Subsystem Device Driver vpath devices for copying all logical volumes. Do not forget to include JFS log volumes.

   Note:

   The command smitty mklvcopy copies one logical volume at a time. A fast-path command to mirror all the logical volumes on a volume group is mirrorvg.

3. Synchronize logical volumes (LVs) or force synchronization. Use the following command to synchronize all the volumes:

   smitty syncvg 
   

   There are two options on the smitty panel:

   • Synchronize by Logical Volume
   • Synchronize by Physical Volume

   The fast way to synchronize logical volumes is to select the Synchronize by Physical Volume option.

4. Remove the mirror and delete the original LVs. Use the following command to remove the original copy of the logical volumes from all original non-Subsystem Device Driver physical volumes:

   smitty rmlvcopy 
   

5. Remove the original non-Subsystem Device Driver devices from the volume group. Use the command:

   smitty reducevg
   

   The Remove a Physical Volume panel is displayed. Remove all non-SDD devices.

Notes:

1. A non-SDD volume groups can consist of non-ESS or ESS hdisk devices.

2. There is no failover protection unless multiple paths are configured for each LUN.

Example of migrating an existing non-SDD volume group to Subsystem Device Driver vpath devices in concurrent mode

This procedure shows how to migrate an existing AIX volume group to use SDD vpath (pseudo) devices that have multipath capability. You do not take the volume group out of service. The example shown starts with a volume group, vg1, made up of one ESS device, hdisk13.

Tip: This procedure uses the System Management Interface Tool (SMIT). The SMIT facility runs in two interfaces, nongraphical (type smitty to invoke the nongraphical user interface) or graphical (type SMIT to invoke the graphical user interface).

To perform the migration, you must have vpath devices available that are greater than or equal to the size of each of the hdisks making up the volume group. In this example, we have a pseudo device, vpath12, with two paths, hdisk14 and hdisk30, that we will migrate the volume group to.

1. Add the vpath device to the volume group as an Available volume:
   1. Type smitty and press Enter from your desktop window. The System Management Interface Tool panel is displayed.
   2. Select System Storage Management (Physical & Logical) and press Enter. The System Storage Management (Physical & Logical) panel is displayed.
   3. Select Logical Volume Manager and press Enter. The Logical Volume Manager panel is displayed.
   4. Select Volume Group and press Enter. The Volume Group panel is displayed.
   5. Select Add a Data Path Volume to a Volume Group and press Enter.
   6. Type vg1 in the Volume Group Name field. Type vpath12 in the Physical Volume Name field. Press Enter.

      You can also enter the command:

      extendvg4vp -f vg1 vpath12
      

2. Mirror logical volumes from the original volume to the new SDD vpath volume:
   1. Type smitty and press Enter from your desktop window. The System Management Interface Tool panel is displayed.
   2. Select System Storage Management (Physical & Logical) and press Enter. The System Storage Management (Physical & Logical) panel is displayed.
   3. Select Logical Volume Manager and press Enter. The Logical Volume Manager panel is displayed.
   4. Select Volume Group and press Enter. The Volume Group panel is displayed.
   5. Select Mirror a Volume Group and press Enter. The Mirror a Volume Group panel is displayed.
   6. Type a volume group name. Type a physical volume name. Press Enter.

      You can also enter the command:

      mirrorvg vg1 vpath12
      

3. Synchronize the logical volumes in the volume group:
   1. Type smitty and press Enter from your desktop window. The System Management Interface Tool panel is displayed.
   2. Select System Storage Management (Physical & Logical) and press Enter. The System Storage Management (Physical & Logical) panel is displayed.
   3. Select Logical Volume Manager and press Enter. The Logical Volume Manager panel is displayed.
   4. Select Volume Group and press Enter. The Volume Group panel is displayed.
   5. Select Synchronize LVM Mirrors and press Enter. The Synchronize LVM Mirrors panel is displayed.
   6. Select Synchronize by Physical Volume.
   You can also enter the command:

   syncvg -p hdisk13 vpath12
   

4. Delete copies of all logical volumes from the original physical volume:
   1. Type smitty and press Enter from your desktop window. The System Management Interface Tool panel is displayed.
   2. Select Logical Volumes and press Enter. The Logical Volumes panel is displayed.
   3. Select Set Characteristic of a Logical Volume and press Enter. The Set Characteristic of a Logical Volume panel is displayed.
   4. Select Remove Copy from a Logical Volume and press Enter. The Remove Copy from a Logical Volume panel is displayed.
   You can also enter the command:

   rmlvcopy loglv01
   1 hdisk13 and rmlvcopy lv01 1 hdisk13
   

5. Remove the old physical volume from the volume group:
   1. Type smitty and press Enter from your desktop window. The System Management Interface Tool panel is displayed.
   2. Select Logical Volume manager and press Enter. The Logical Volume manager panel is displayed.
   3. Select Volume Groups and press Enter. The Volume Groups panel is displayed.
   4. Select Set Characteristics of a Volume Group and press Enter. The Set Characteristics of a Volume Group panel is displayed.
   5. Select Remove a Physical Volume from a Volume Group and press Enter. The Remove a Physical Volume from a Volume Group panel is displayed.
   You can also enter the command:

   reducevg vg1 hdisk13
   

Using the trace function

SDD supports AIX trace functions. The SDD trace ID is 2F8. Trace ID 2F8 traces routine entry, exit, and error paths of the algorithm. To use it, manually turn on the trace function before the program starts to run, then turn off the trace function either after the program stops, or any time you need to read the trace report. To start the trace function, type:

   trace -a -j 2F8

To stop the trace function, type:

   trcstop

To read the report, type:

   trcrpt | pg

Note:

To perform the AIX trace function, you must have the bos.sysmgt.trace fileset installed on your system.

Error log messages

SDD logs error conditions into the AIX errlog system. To check if SDD has generated an error log message, type the following command:

   errpt -a | grep VPATH

The following list shows the SDD error log messages and explains each one:

VPATH_XBUF_NOMEM

An attempt was made to open a SDD vpath file and to allocate kernel-pinned memory. The system returned a null pointer to the calling program and kernel-pinned memory was not available. The attempt to open the file failed.

VPATH_PATH_OPEN

SDD device file failed to open one of its paths (hdisks). An attempt to open a vpath device is successful if at least one attached path opens. The attempt to open a vpath device fails only when all the vpath device paths fail to open.

VPATH_DEVICE_OFFLINE

Several attempts to retry an I/O request for a vpath device on a path have failed. The path state is set to Dead and the path is taken offline. Use the datapath command to set the offline path to online. For more information, see Chapter 8, Using the datapath commands.

VPATH_DEVICE_ONLINE

SDD supports Dead path auto_failback and Dead path reclamation. A Dead path is selected to send an I/O, after it has been bypassed by 2000 I/O requests on an operational path. If the I/O is successful, the Dead path is put Online, and its state is changed back to Open; a Dead path is put Online, and its state changes to Open after it has been bypassed by 50 000 I/O requests on an operational path.

New and modified error log messages by SDD for HACMP

The following list shows the new and modified error log messages generated by SDD installed from the ibmSdd_433.rte or ibmSdd_510nchacmp.rte fileset. This SDD release is for HACMP environments only. See What's new in SDD for HACMP/6000 for more information on this release.

VPATH_DEVICE_OPEN

The SDD device file failed to open one of its paths (hdisks). An attempt to open a vpath device is successful if at least one attached path opens. The attempt to open a vpath device fails only when all the vpath device paths fail to open. In addition, this error log message is posted when the vpath device fails to register its underdying paths or fails to read the persistent reserve key for the device.

VPATH_OUT_SERVICE

There is no path available to retry a I/O request that failed for a vpath device. The I/O request is returned to the calling program and this error log is posted.

VPATH_FAIL_RELPRESERVE

An attempt was made to close a vpath device that was not opened with the RETAIN_RESERVE option on the persistent reserve. The attempt to close the vpath device was successful; however, the persistent reserve was not released. The user is notified that the persistent reserve is still in effect, and this error log is posted.

VPATH_RESV_CFLICT

An attempt was made to open a vpath device, but the reservation key of the vpath device is different from the reservation key currently in effect. The attempt to open the device fails and this error log is posted. The device could not be opened because it is currently reserved by someone else.

Chapter 4. Installing and configuring SDD on a Windows NT host system

This chapter provides instructions for installing and configuring the Subsystem Device Driver on an Windows NT host system attached to an ESS. For updated and additional information not included in this chapter, see the README file on the compact disc or visit the SDD website at: www.ibm.com/storage/support/techsup/swtechsup.nsf/support/sddupdates

Figure 4. Where SDD fits in the protocol stack

Notes:

1. If you attempt to install over an existing version of SDD or Data Path Optimizer (DPO), the installation fails. You must uninstall any previous version of SDD or DPO before installing this version of SDD.

2. SDD 1.2.1 or higher is required to support Windows NT clustering.

3. Windows NT clustering requires Windows NT 4.0 Enterprise Edition.

4. SDD 1.2.1 or higher does not support I/O load-balancing in a Windows NT clustering environment.

5. You cannot store the Windows NT operating system or a paging file on a SDD-controlled multipath device. This environment is not supported.

6. You must have Windows NT 4.0 Service Pack 3 or higher installed on your system.

7. SDD only supports 32-bit mode applications on a Windows NT host system.

Hardware and software requirements

You must have the following hardware and software components in order to successfully install SDD.

Hardware

• ESS
• Host system
• SCSI adapters and cables
• Fibre adapters and cables

Software

• Windows NT operating system
• SCSI and fibre-channel device drivers

Host system requirements

To successfully install SDD, your Windows NT host system should be an Intel-based system with Windows NT Version 4.0 Service Pack 3 or higher installed. The host system can be a uni-processor or a multi-processor system.

ESS requirements

To successfully install SDD, ensure that your host system is configured to the ESS as an Intel-based PC (personal computer) server with Windows NT 4.0 or higher.

SCSI requirements

To use the SDD SCSI support, ensure your host system meets the following requirements:

• The maximum number of SCSI adapters that is supported is 32.
• A SCSI cable is required to connect each SCSI host adapter to an ESS port.
• The SDD I/O load-balancing and failover features require a minimum of two SCSI adapters.

  Note:

  SDD also supports one SCSI adapter on the host system. With single-path access, concurrent download of licensed internal code is supported with SCSI devices. However, the load balancing and failover features are not available.

• For information about the SCSI adapters that can attach to your Windows NT host system go to www.storage.ibm.com/hardsoft/products/ess/supserver.htm

Fibre requirements

To use the SDD fibre support, ensure your host system meets the following requirements:

• The maximum number of fibre-channel adapters that are supported is 256.
• A fiber-optic cable is required to connect each fibre-channel adapter to an ESS port.
• The SDD I/O load-balancing and failover features require a minimum of two fibre adapters.
• For information about the fibre-channel adapters that can attach to your Windows NT host system go to the website at: www.storage.ibm.com/hardsoft/products/ess/supserver.htm

Non-supported environments

SDD does not support the following environments:

• A host server with a single-path fibre-channel connection to an ESS.

  Note:

  A host server with a single fibre adapter that connects through a switch to multiple ESS ports is considered a multipath fibre-channel connection; and, thus is a supported environment.

• A host server with SCSI channel connections and a single-path fibre-channel connection to an ESS.
• A host server with both a SCSI channel and fibre-channel connection to a shared LUN.

Configuring the ESS

Before you install SDD, configure your ESS for single-port or multiple-port access for each LUN. SDD requires a minimum of two independent paths that share the same LUN to use the load-balancing and failover features.

For information about configuring your ESS, see IBM Enterprise Storage Server Introduction and Planning Guide, GC26-7294.

Configuring SCSI adapters

Attention: Failure to disable the BIOS of attached non-boot devices may cause your system to attempt to boot from an unexpected non-boot device.

Before you install and use SDD, you must configure your SCSI adapters. For SCSI adapters that attach boot devices, ensure that the BIOS for the adapter is enabled. For all other adapters that attach non-boot devices, ensure the BIOS for the adapter is disabled.

Note:

When the adapter shares the SCSI bus with other adapters, the BIOS must be disabled.

Configuring fibre-channel adapters

You must configure the fibre-channel adapters that are attached to your Windows NT host system before you install SDD. Follow the adapter-specific configuration instructions to configure the adapters attached to your Windows NT host systems. Make sure that your Windows NT host system has Service Pack 3 or higher.

See IBM TotalStorage Enterprise Storage Server Host System Attachment Guide for more information about installing and configuring fibre-channel adapters to your Windows NT host systems.

Installing the Subsystem Device Driver

To install all components, you must have 1 MB (MB equals approximately 1 000 000 bytes) of disk space available, and you must have Windows NT 4.0 Service Pack 3 or higher installed on your system.

You must log on as an administrator user to install SDD.

Perform the following steps to install SDD filter and application programs on your system:

1. Log on as the administrator user.
2. Insert the SDD installation compact disc into the CD-ROM drive.
3. Start the Windows NT Explorer program.
4. Select the CD-ROM drive. A list of all the installed directories on the compact disc is displayed.
5. Select the \winnt\IBMSdd directory.
6. Run the setup.exe program. This starts the install-shield.
7. Click Next. The Software License agreement is displayed.
8. Click Yes. The User Information panel is displayed.
9. Type your name and your company name.
10. Click Next. The Choose Destination Location panel is displayed.
11. Click Next. The Setup panel is displayed.
12. Select the type of setup you prefer from the following setup choices: IBM recommends that you select Typical.

    Typical

    Selects all options.

    Compact

    Selects the minimum required options only (the installation driver and the README file).

    Custom

    Select the options that you need.

13. Click Next. The Setup Complete panel is displayed.
14. Click Finish. The SDD program prompts you to start your computer again.
15. Click Yes to start your computer again. When you log on again, you see a Subsystem Device Driver Management entry in your Program menu containing the following files:
    1. SDD management
    2. Subsystem Device Driver manual
    3. README file

Note:

You can use the datapath query device command to verify the SDD installation. SDD is successfully installed if the command runs successfully.

Uninstalling the Subsystem Device Driver

Perform the following steps to uninstall SDD on a Windows NT host system:

1. Log on as the administrator user.
2. Click Start -->Settings -->Control Panel. The Control Panel window opens.
3. Open Add/Remove Programs in Control Panel. The Add/Remove Programs window opens.
4. In the Add/Remove Programs window, select SDD from the Currently installed programs selection list.
5. Click on the Add/Remove button.

   Attention: After uninstalling the previous version, you must immediately install the new SDD version to avoid any potential data loss (See Installing the Subsystem Device Driver for instructions).

Displaying the current version of the Subsystem Device Driver

You can display the current SDD version on a Windows NT host system by viewing the sddpath.sys file properties. Perform the following steps to view the properties of sddpath.sys file:

1. Click Start -->Run -->Programs -->Accessories -->Windows Explorer. Windows will open Windows Explorer.
2. In Windows Explorer, go to your_installation_directory_letter:\Winnt\system32\drivers directory.

   (your_installation_directory_letter\ is the directory letter where you have installed the sddpath.sys file)

3. Click the sddpath.sys file in your_installation_directory_letter:\Winnt\system32\drivers directory

   where your_installation_directory_letter refers to the letter of the directory in which you have installed the sddpath.sys file.

4. Right-click on the sddpath.sys file and then click Properties. The sddpath.sys properties window will open.
5. In the sddpath.sys properties window, click the Version panel. The file version and copyright information about sddpath.sys will be displayed.

Upgrading the Subsystem Device Driver

If you attempt to install over an existing version of SDD or Data Path Optimizer (DPO), the installation fails. You must uninstall any previous version of the SDD or DPO before installing a new version of SDD.

Perform the following steps to upgrade to a newer SDD version:

1. Uninstall the previous version of SDD (See Uninstalling the Subsystem Device Driver for instructions).

   Attention: After uninstalling the previous version, you must immediately install the new version of SDD to avoid any potential data loss.

2. Install the new version of SDD (See Installing the Subsystem Device Driver for instructions).

Configuring the Subsystem Device Driver

To activate SDD, you need to restart your Windows NT system after it is installed. In fact, a restart is required to activate multipath support whenever a new file system or partition is added.

Note:

You must log on as an administrator user to have access to the Windows NT disk administrator.

Adding paths to SDD devices

Attention: Ensure that SDD is installed before you add a new path to a device. Otherwise, the Windows NT server's ability to access existing data on that device could be lost.

This section contains the procedures for adding paths to SDD devices in multipath environments. These procedures include:

1. Reviewing the existing SDD configuration information
2. Installing and configuring additional paths
3. Verifying additional paths are installed correctly

Reviewing the existing SDD configuration information

Before adding any additional hardware, you should review the configuration information for the adapters and devices currently on your Windows NT server.

You should verify that the number of adapters and the number of paths to each ESS volume match the known configuration. Perform the following steps to display information about the adapters and devices:

1. Click Start --> Program --> Subsystem Device Driver --> Subsystem Device Driver Management. An MS-DOS window is displayed.
2. Type datapath query adapter and press Enter. The output should include information about all the installed adapters. In this example, one SCSI adapter has 10 active paths. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |                                                                                |
   |Active Adapters :1                                                              |
   |                                                                                |
   |Adpt#     Adapter Name   State     Mode     Select     Errors  Paths  Active    |
   |    0  Scsi Port6 Bus0  NORMAL   ACTIVE        542          0     10      10    |
   |                                                                                |
   |                                                                                |
   +--------------------------------------------------------------------------------+

3. Next, type datapath query device and press Enter. In this example, 10 devices are attached to the SCSI path. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |                                                                                |
   |Total Devices : 10                                                              |
   |                                                                                |
   |DEV#:   0  DEVICE NAME: Disk2 Part0  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part0     OPEN   NORMAL         14          0    |
   |                                                                                |
   |DEV#:   1  DEVICE NAME: Disk2 Part1  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part1     OPEN   NORMAL         94          0    |
   |                                                                                |
   |DEV#:   2  DEVICE NAME: Disk3 Part0  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part0     OPEN   NORMAL         16          0    |
   |                                                                                |
   |DEV#:   3  DEVICE NAME: Disk3 Part1  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part1     OPEN   NORMAL         94          0    |
   |                                                                                |
   |DEV#:   4  DEVICE NAME: Disk4 Part0  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part0     OPEN   NORMAL         14          0    |
   |                                                                                |
   |DEV#:   5  DEVICE NAME: Disk4 Part1  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part1     OPEN   NORMAL         94          0    |
   |                                                                                |
   |DEV#:   6  DEVICE NAME: Disk5 Part0  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part0     OPEN   NORMAL         14          0    |
   |                                                                                |
   |DEV#:   7  DEVICE NAME: Disk5 Part1  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part1     OPEN   NORMAL         94          0    |
   |                                                                                |
   |DEV#:   8  DEVICE NAME: Disk6 Part0  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part0     OPEN   NORMAL         14          0    |
   |                                                                                |
   |DEV#:   9  DEVICE NAME: Disk6 Part1  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part1     OPEN   NORMAL         94          0    |
   |                                                                                |
   +--------------------------------------------------------------------------------+

Installing and configuring additional paths

Perform the following steps to install and configure additional paths to a vpath device:

1. Install any additional hardware on the Windows NT server.
2. Install any additional hardware to the ESS.
3. Configure the new paths to the server.
4. Restart the Windows NT server. Restarting will ensure correct multi-path access to both existing and new storage, and your Windows NT server.
5. Verify that the path is added correctly. See Verifying additional paths are installed correctly

Verifying additional paths are installed correctly

After installing additional paths to SDD devices, you should verify:

• That all additional paths have been installed correctly.
• The number of adapters and the number of paths to each ESS volume match the updated configuration.
• The Windows disk numbers of all primary paths are labeled as path #0.

Perform the following steps to verify that the additional paths have been installed correctly:

1. Click Start --> Program --> Subsystem Device Driver --> Subsystem Device Driver Management. An MS-DOS window appears.
2. Type datapath query adapter and press Enter. The output should include information about any additional adapters that were installed. In this example, an additional path is installed to the previous configuration. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |Active Adapters :2                                                              |
   |                                                                                |
   |Adpt#     Adapter Name   State     Mode     Select     Errors  Paths  Active    |
   |    0  Scsi Port6 Bus0  NORMAL   ACTIVE        188          0     10      10    |
   |    1  Scsi Port7 Bus0  NORMAL   ACTIVE        204          0     10      10    |
   |                                                                                |
   |                                                                                |
   +--------------------------------------------------------------------------------+

3. Type datapath query device and press Enter. The output should include information about any additional devices that were installed. In this example, the output includes information about the new SCSI adapter that was assigned. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |                                                                                |
   |                                                                                |
   |Total Devices : 10                                                              |
   |                                                                                |
   |DEV#:   0  DEVICE NAME: Disk2 Part0  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part0     OPEN   NORMAL          5          0    |
   |    1    Scsi Port7 Bus0/Disk7 Part0     OPEN   NORMAL          9          0    |
   |                                                                                |
   |DEV#:   1  DEVICE NAME: Disk2 Part1  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part1     OPEN   NORMAL         32          0    |
   |    1    Scsi Port7 Bus0/Disk7 Part1     OPEN   NORMAL         32          0    |
   |                                                                                |
   |DEV#:   2  DEVICE NAME: Disk3 Part0  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part0     OPEN   NORMAL          7          0    |
   |    1    Scsi Port7 Bus0/Disk8 Part0     OPEN   NORMAL          9          0    |
   |                                                                                |
   |DEV#:   3  DEVICE NAME: Disk3 Part1  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part1     OPEN   NORMAL         28          0    |
   |    1    Scsi Port7 Bus0/Disk8 Part1     OPEN   NORMAL         36          0    |
   |                                                                                |
   |DEV#:   4  DEVICE NAME: Disk4 Part0  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part0     OPEN   NORMAL          8          0    |
   |    1    Scsi Port7 Bus0/Disk9 Part0     OPEN   NORMAL          6          0    |
   |                                                                                |
   |DEV#:   5  DEVICE NAME: Disk4 Part1  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part1     OPEN   NORMAL         35          0    |
   |    1    Scsi Port7 Bus0/Disk9 Part1     OPEN   NORMAL         29          0    |
   |                                                                                |
   |DEV#:   6  DEVICE NAME: Disk5 Part0  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part0     OPEN   NORMAL          6          0    |
   |    1   Scsi Port7 Bus0/Disk10 Part0     OPEN   NORMAL          8          0    |
   |                                                                                |
   |DEV#:   7  DEVICE NAME: Disk5 Part1  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part1     OPEN   NORMAL         24          0    |
   |    1   Scsi Port7 Bus0/Disk10 Part1     OPEN   NORMAL         40          0    |
   |                                                                                |
   |DEV#:   8  DEVICE NAME: Disk6 Part0  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part0     OPEN   NORMAL          8          0    |
   |    1   Scsi Port7 Bus0/Disk11 Part0     OPEN   NORMAL          6          0    |
   |                                                                                |
   |DEV#:   9  DEVICE NAME: Disk6 Part1  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part1     OPEN   NORMAL         35          0    |
   |    1   Scsi Port7 Bus0/Disk11 Part1     OPEN   NORMAL         29          0    |
   |                                                                                |
   +--------------------------------------------------------------------------------+

   Note:

   The definitive way to identify unique volumes on the ESS is by the serial number displayed. The volume appears at the SCSI level as multiple disks (more properly, Adapter/Bus/ID/LUN), but it's the same volume on the ESS. The example above shows two paths to each partition (path 0: Scsi Port6 Bus0/Disk2; And path 1: Scsi Port7 Bus0/Disk7).

   The example shows partition 0 (Part0) for each of the device. This partition stores information about windows partition on the drive. The operating system masks this partition from the user, but it still exists. In general, you'll see one more partition from the output of Datapath Query Device than what is being displayed from the Disk Administrator application.

Adding or modifying multipath storage configuration to the ESS

This section contains the procedures for adding new storage to existing configuration in multipath environments. These procedures include:

1. Reviewing the existing SDD configuration information
2. Adding new storage to existing configuration
3. Verifying new storage is installed correctly

Reviewing the existing SDD configuration information

Before adding any additional hardware, you should review the configuration information for the adapters and devices currently on your Windows NT server.

You should verify that the number of adapters and the number of paths to each ESS volume match the known configuration. Perform the following steps to display information about the adapters and devices:

1. Click Start --> Program --> Subsystem Device Driver --> Subsystem Device Driver Management. An MS-DOS window is displayed.
2. Type datapath query adapter and press Enter. The output should include information about all the installed adapters. In this example, two SCSI adapters are installed on the Windows NT host server. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |                                                                                |
   |Active Adapters :2                                                              |
   |                                                                                |
   |Adpt#     Adapter Name   State     Mode     Select     Errors  Paths  Active    |
   |    0  Scsi Port6 Bus0  NORMAL   ACTIVE        188          0     10      10    |
   |    1  Scsi Port7 Bus0  NORMAL   ACTIVE        204          0     10      10    |
   |                                                                                |
   |Previous configuration with one additional path                                 |
   |                                                                                |
   |                                                                                |
   |                                                                                |
   +--------------------------------------------------------------------------------+

3. Next, type datapath query device and press Enter. In this example, 10 devices are attached to the SCSI path. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |                                                                                |
   |Total Devices : 10                                                              |
   |                                                                                |
   |DEV#:   0  DEVICE NAME: Disk2 Part0  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part0     OPEN   NORMAL          5          0    |
   |    1    Scsi Port7 Bus0/Disk7 Part0     OPEN   NORMAL          9          0    |
   |                                                                                |
   |DEV#:   1  DEVICE NAME: Disk2 Part1  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part1     OPEN   NORMAL         32          0    |
   |    1    Scsi Port7 Bus0/Disk7 Part1     OPEN   NORMAL         32          0    |
   |                                                                                |
   |DEV#:   2  DEVICE NAME: Disk3 Part0  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part0     OPEN   NORMAL          7          0    |
   |    1    Scsi Port7 Bus0/Disk8 Part0     OPEN   NORMAL          9          0    |
   |                                                                                |
   |DEV#:   3  DEVICE NAME: Disk3 Part1  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part1     OPEN   NORMAL         28          0    |
   |    1    Scsi Port7 Bus0/Disk8 Part1     OPEN   NORMAL         36          0    |
   |                                                                                |
   |DEV#:   4  DEVICE NAME: Disk4 Part0  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part0     OPEN   NORMAL          8          0    |
   |    1    Scsi Port7 Bus0/Disk9 Part0     OPEN   NORMAL          6          0    |
   |                                                                                |
   |DEV#:   5  DEVICE NAME: Disk4 Part1  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part1     OPEN   NORMAL         35          0    |
   |    1    Scsi Port7 Bus0/Disk9 Part1     OPEN   NORMAL         29          0    |
   |                                                                                |
   |DEV#:   6  DEVICE NAME: Disk5 Part0  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part0     OPEN   NORMAL          6          0    |
   |    1   Scsi Port7 Bus0/Disk10 Part0     OPEN   NORMAL          8          0    |
   |                                                                                |
   |DEV#:   7  DEVICE NAME: Disk5 Part1  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part1     OPEN   NORMAL         24          0    |
   |    1   Scsi Port7 Bus0/Disk10 Part1     OPEN   NORMAL         40          0    |
   |                                                                                |
   |DEV#:   8  DEVICE NAME: Disk6 Part0  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part0     OPEN   NORMAL          8          0    |
   |    1   Scsi Port7 Bus0/Disk11 Part0     OPEN   NORMAL          6          0    |
   |                                                                                |
   |DEV#:   9  DEVICE NAME: Disk6 Part1  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part1     OPEN   NORMAL         35          0    |
   |    1   Scsi Port7 Bus0/Disk11 Part1     OPEN   NORMAL         29          0    |
   |                                                                                |
   +--------------------------------------------------------------------------------+

Adding new storage to existing configuration

Perform the following steps to install additional storage:

1. Install any additional hardware to the ESS.
2. Configure the new storage to the server.
3. Restart the Windows NT server. Restarting will ensure correct multi-path access to both existing and new storage, and your Windows NT server.
4. Verify that the new storage is added correctly. See Verifying new storage is installed correctly

Verifying new storage is installed correctly

After adding new storage to existing configuration, you should verify:

• That the new storage is correctly installed and configured.
• The number of adapters and the number of paths to each ESS volume match the updated configuration.
• The Windows disk numbers of all primary paths are labeled as path #0.

Perform the following steps to verify that the additional storage have been installed correctly:

1. Click Start --> Program --> Subsystem Device Driver --> Subsystem Device Driver Management. An MS-DOS window appears.
2. Type datapath query adapter and press Enter. The output should include information about all the installed adapters. In this example, two SCSI adapters are installed on the Windows NT host server. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |                                                                                |
   |Active Adapters :2                                                              |
   |                                                                                |
   |Adpt#     Adapter Name   State     Mode     Select     Errors  Paths  Active    |
   |    0  Scsi Port6 Bus0  NORMAL   ACTIVE        295          0     16      16    |
   |    1  Scsi Port7 Bus0  NORMAL   ACTIVE        329          0     16      16    |
   |                                                                                |
   |                                                                                |
   +--------------------------------------------------------------------------------+

3. Type datapath query device and press Enter. The output should include information about any additional devices that were installed. In this example, the output includes information about the new devices that were assigned. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |                                                                                |
   |Total Devices : 16                                                              |
   |                                                                                |
   |DEV#:   0  DEVICE NAME: Disk2 Part0  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part0     OPEN   NORMAL          9          0    |
   |    1   Scsi Port7 Bus0/Disk10 Part0     OPEN   NORMAL          5          0    |
   |                                                                                |
   |DEV#:   1  DEVICE NAME: Disk2 Part1  TYPE: 2105E20   SERIAL: 00A12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk2 Part1     OPEN   NORMAL         26          0    |
   |    1   Scsi Port7 Bus0/Disk10 Part1     OPEN   NORMAL         38          0    |
   |                                                                                |
   |DEV#:   2  DEVICE NAME: Disk3 Part0  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part0     OPEN   NORMAL          9          0    |
   |    1   Scsi Port7 Bus0/Disk11 Part0     OPEN   NORMAL          7          0    |
   |                                                                                |
   |DEV#:   3  DEVICE NAME: Disk3 Part1  TYPE: 2105E20   SERIAL: 00B12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk3 Part1     OPEN   NORMAL         34          0    |
   |    1   Scsi Port7 Bus0/Disk11 Part1     OPEN   NORMAL         30          0    |
   |                                                                                |
   |DEV#:   4  DEVICE NAME: Disk4 Part0  TYPE: 2105E20   SERIAL: 31512028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part0     OPEN   NORMAL          8          0    |
   |    1   Scsi Port7 Bus0/Disk12 Part0     OPEN   NORMAL          6          0    |
   |                                                                                |
   |DEV#:   5  DEVICE NAME: Disk4 Part1  TYPE: 2105E20   SERIAL: 31512028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk4 Part1     OPEN   NORMAL         35          0    |
   |    1   Scsi Port7 Bus0/Disk12 Part1     OPEN   NORMAL         28          0    |
   |                                                                                |
   |DEV#:   6  DEVICE NAME: Disk5 Part0  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part0     OPEN   NORMAL          5          0    |
   |    1   Scsi Port7 Bus0/Disk13 Part0     OPEN   NORMAL          9          0    |
   |                                                                                |
   |DEV#:   7  DEVICE NAME: Disk5 Part1  TYPE: 2105E20   SERIAL: 00D12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk5 Part1     OPEN   NORMAL         28          0    |
   |    1   Scsi Port7 Bus0/Disk13 Part1     OPEN   NORMAL         36          0    |
   |                                                                                |
   |DEV#:   8  DEVICE NAME: Disk6 Part0  TYPE: 2105E20   SERIAL: 40812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part0     OPEN   NORMAL          5          0    |
   |    1   Scsi Port7 Bus0/Disk14 Part0     OPEN   NORMAL          9          0    |
   |                                                                                |
   |DEV#:   9  DEVICE NAME: Disk6 Part1  TYPE: 2105E20   SERIAL: 40812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk6 Part1     OPEN   NORMAL         25          0    |
   |    1   Scsi Port7 Bus0/Disk14 Part1     OPEN   NORMAL         38          0    |
   |                                                                                |
   |DEV#:  10  DEVICE NAME: Disk7 Part0  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk7 Part0     OPEN   NORMAL          7          0    |
   |    1   Scsi Port7 Bus0/Disk15 Part0     OPEN   NORMAL          7          0    |
   |                                                                                |
   |DEV#:  11  DEVICE NAME: Disk7 Part1  TYPE: 2105E20   SERIAL: 50812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk7 Part1     OPEN   NORMAL         34          0    |
   |    1   Scsi Port7 Bus0/Disk15 Part1     OPEN   NORMAL         30          0    |
   |                                                                                |
   |DEV#:  12  DEVICE NAME: Disk8 Part0  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk8 Part0     OPEN   NORMAL          7          0    |
   |    1   Scsi Port7 Bus0/Disk16 Part0     OPEN   NORMAL          7          0    |
   |                                                                                |
   |DEV#:  13  DEVICE NAME: Disk8 Part1  TYPE: 2105E20   SERIAL: 60012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk8 Part1     OPEN   NORMAL         29          0    |
   |    1   Scsi Port7 Bus0/Disk16 Part1     OPEN   NORMAL         35          0    |
   |                                                                                |
   |DEV#:  14  DEVICE NAME: Disk9 Part0  TYPE: 2105E20   SERIAL: 00812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk9 Part0     OPEN   NORMAL          6          0    |
   |    1   Scsi Port7 Bus0/Disk17 Part0     OPEN   NORMAL          8          0    |
   |                                                                                |
   |DEV#:  15  DEVICE NAME: Disk9 Part1  TYPE: 2105E20   SERIAL: 00812028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port6 Bus0/Disk9 Part1     OPEN   NORMAL         28          0    |
   |    1   Scsi Port7 Bus0/Disk17 Part1     OPEN   NORMAL         36          0    |
   |                                                                                |
   |                                                                                |
   +--------------------------------------------------------------------------------+

   Note:

   The definitive way to identify unique volumes on the ESS is by the serial number displayed. The volume appears at the SCSI level as multiple disks (more properly, Adapter/Bus/ID/LUN), but it's the same volume on the ESS. The example above shows two paths to each partition (path 0: Scsi Port6 Bus0/Disk2; And path 1: Scsi Port7 Bus0/Disk10).

   The example shows partition 0 (Part0) for each of the device. This partition stores information about windows partition on the drive. The operating system masks this partition from the user, but it still exists. In general, you'll see one more partition from the output of Datapath Query Device than what is being displayed from the Disk Administrator application.

Chapter 5. Installing and configuring SDD on a Windows 2000 host system

This chapter provides instructions to install and set up the Subsystem Device Driver on a Windows 2000 host system attached to an ESS. For updated and additional information not included in this chapter, see the README file on the compact disc or visit the SDD website at: www.ibm.com/storage/support/techsup/swtechsup.nsf/support/sddupdates

Figure 5. Where the SDD fits in the protocol stack

Notes:

1. You cannot store the Windows 2000 operating system or a paging file on a SDD-controlled multi-path device. This environment is not supported.

2. You cannot run SDD in a non-concurrent environment in which more than one host is attached to the same logical unit number (LUN) on a Enterprise Storage Server; for example, in a multi-host environment. However, concurrent multi-host environments are supported.

3. SDD supports 32-bit mode applications on a Windows 2000 host system.

4. SDD 1.3.0.0 or higher is required to support Windows 2000 clustering.

5. SDD 1.3.0.0 or higher does not support I/O load-balancing in a Windows 2000 clustering environment.

Hardware and software requirements

You must have the following hardware and software components in order to install SDD:

Hardware

• ESS
• Host system
• SCSI adapters and cables
• Fibre adapters and cables

Software

• Windows 2000 operating system with Service Pack 2 or higher
• SCSI and fibre-channel device drivers

Host system requirements

To successfully install SDD, your Windows 2000 host system should be an Intel-based system. Your host system should have Windows 2000 Service Pack 2 installed. The host system can be a uni-processor or a multi-processor system.

To install all components, you must have 1 MB (MB equals approximately 1 000 000 bytes) of disk space available.

ESS requirements

To successfully install SDD, make sure that you configure the ESS devices as IBM 2105xxx (where xxx is the ESS model number) on your Windows 2000 host system.

SCSI requirements

To use the SDD SCSI support, ensure your host system meets the following requirements:

• The maximum number of SCSI adapters that is supported is 32.
• A SCSI cable is required to connect each SCSI host adapter to an ESS port.
• The SDD I/O load-balancing and failover features require a minimum of two SCSI adapters.

  Note:

  SDD also supports one SCSI adapter on the host system. With single-path access, concurrent download of licensed internal code is supported with SCSI devices. However, the load-balancing and failover features are not available.

• For information about the SCSI adapters that can attach to your Windows 2000 host system go to www.storage.ibm.com/hardsoft/products/ess/supserver.htm

Fibre requirements

To use the SDD fibre support, ensure your host system meets the following requirements:

• The maximum number of fibre-channel adapters that are supported is 32.
• A fiber-optic cable is required to connect each fibre-channel adapter to an ESS port.
• The SDD I/O load-balancing and failover features require a minimum of two fibre-channel adapters.
• For information about the fibre-channel adapters that can attach to your Windows 2000 host system go to the Web site at: www.storage.ibm.com/hardsoft/products/ess/supserver.htm

Non-supported environments

SDD does not support the following environments:

• A host server with a single-path fibre-channel connection to an ESS is not supported.

  Note:

  A host server with a single fibre adapter that connects through a switch to multiple ESS ports is considered a multipath fibre-channel connection; and, thus is a supported environment.

• A host server with SCSI channel connections and a single-path fibre-channel connection to an ESS is not supported.
• A host server with both a SCSI channel and fibre-channel connection to a shared LUN is not supported.

Configuring the ESS

Before you install SDD, configure your ESS for single-port or multiple-port access for each LUN. SDD requires a minimum of two independent paths that share the same logical unit to use the load balancing and failover features.

For information about configuring your ESS, see the IBM Enterprise Storage Server Introduction and Planning Guide.

Note:

During heavy usage, the Windows 2000 operating system might slow down while trying to recover from error conditions.

Configuring SCSI adapters

Before you install and use SDD, you must configure your SCSI adapters. For SCSI adapters that attach boot devices, ensure that the BIOS for the adapter is enabled. For all other adapters that attach non-boot devices, ensure the BIOS for the adapter is disabled.

Note:

When the adapter shares the SCSI bus with other adapters, the BIOS must be disabled.

Configuring fibre-channel adapters

You must configure the fibre-channel adapters that are attached to your Windows 2000 host system before you install SDD. Follow the adapter-specific configuration instructions to configure the adapters attached to your Windows 2000 host systems. Make sure that your Windows 2000 host system has Service Pack 2 or higher.

See IBM TotalStorage Enterprise Storage Server Host System Attachment Guide for more information about installing and configuring fibre-channel adapters to your Windows 2000 host systems.

Installing SDD on a Windows 2000 host system

The following section describes how to install SDD. Make sure that all hardware and software requirements are met before you install the Subsystem Device Driver. See Hardware and software requirements for more information.

Note:

You must log on as an administrator user to install SDD.

Perform the following steps to install the SDD filter and application programs on your system:

1. Log on as the administrator user.
2. Insert the SDD installation CD-ROM into the selected drive. The SDD panel is displayed.
3. Start the Windows 2000 Explorer program.
4. Select the CD-ROM drive. A list of all the installed directories on the compact disc is displayed.
5. Select the \win2k\IBMSdd directory.
6. Run the setup.exe program. The Installshield starts.
7. Click Next. The Software Licensing Agreement panel is displayed.
8. Click Yes. The User Information panel is displayed.
9. Type your name and your company name.
10. Click Next. The Choose Destination Location panel is displayed.
11. Click Next. The Setup panel is displayed.
12. Select the type of setup you prefer from the following setup choices described below. IBM recommends that you select Typical.

    Typical

    Selects all options.

    Compact

    Selects the minimum required options only (the installation driver and README file).

    Custom

    Select the options that you need.

13. Click Next. The Setup Complete panel is displayed.
14. Click Finish. The SDD program prompts you to start your computer again.

15. Click Yes to start your computer again. When you log on again, you see a Subsystem Device Driver entry in your Program menu containing the following files:
    1. Subsystem Device Driver management
    2. Subsystem Device Driver manual
    3. README file

Note:

You can verify that SDD has been successfully installed by issuing the datapath query device command. If the command executes, SDD is installed.

Uninstalling the Subsystem Device Driver

Perform the following steps to uninstall SDD on a Windows 2000 host system:

1. Log on as the administrator user.
2. Click Start -->Settings -->Control Panel. The Control Panel opens.
3. Open Add/Remove Programs in Control Panel. The Add/Remove Programs window opens.
4. In the Add/Remove Programs window, select the Subsystem Device Driver from the Currently installed programs selection list.
5. Click on the Change/Remove button.

   Attention: After uninstalling the previous version, you must immediately install the new version of SDD to avoid any potential data loss (See Installing SDD on a Windows 2000 host system for instructions).

Displaying the current version of the Subsystem Device Driver

You can display the current version of SDD on a Windows 2000 host system by viewing the sddpath.sys file properties. Perform the following steps to view the properties of sddpath.sys file:

1. Click Start -->Run -->Programs -->Accessories -->Windows Explorer to open Windows Explorer.
2. In Windows Explorer, go to your_installation_directory_drive_letter:\Winnt\system32\drivers directory

   where your_installation_directory_drive_letter is the letter of the directory in which you have installed the sddpath.sys file.

3. Click the sddpath.sys file in your_installation_directory_drive_leltter:\Winnt\system32\drivers directory
4. Right-click on the sddpath.sys file and then click Properties. The sddpath.sys properties window opens.
5. In the sddpath.sys properties window, click the Version panel. The file version and copyright information about sddpath.sys displays.

Configuring the Subsystem Device Driver

To activate SDD, you need to restart your Windows 2000 system after it is installed. In fact, a restart is required to activate multipath support whenever a new file system or partition is added.

Note:

You must log on as an administrator user to have access to the Windows 2000 Computer Management.

Adding paths to SDD devices

Attention: Ensure that SDD is installed before you add additional paths to a device. Otherwise, the Windows 2000 server's ability to access existing data on that device could be lost.

Before adding any additional hardware, you should review the configuration information for the adapters and devices currently on your Windows 2000 server. Perform the following steps to display information about the adapters and devices:

1. Click Start --> Program --> Subsystem Device Driver --> Subsystem Device Driver Management. An MS-DOS window is displayed.
2. Type datapath query adapter and press Enter. The output should include information about all the installed adapters. In this example, one SCSI adapter is installed on the Windows 2000 host server. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |Active Adapters :1                                                              |
   |                                                                                |
   |Adpt#     Adapter Name   State     Mode     Select     Errors  Paths  Active    |
   |    0  Scsi Port1 Bus0  NORMAL   ACTIVE       4057          0      8       8    |
   +--------------------------------------------------------------------------------+

3. Next, type datapath query device and press Enter. In this example, 8 devices are attached to the SCSI path. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |Total Devices : 8                                                               |
   |                                                                                |
   |DEV#:   0  DEVICE NAME: Disk7 Part7  TYPE: 2105E20   SERIAL: 01312028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk7 Part0     OPEN   NORMAL       1045          0    |
   |                                                                                |
   |DEV#:   1  DEVICE NAME: Disk6 Part6  TYPE: 2105E20   SERIAL: 01212028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk6 Part0     OPEN   NORMAL        391          0    |
   |                                                                                |
   |DEV#:   2  DEVICE NAME: Disk5 Part5  TYPE: 2105E20   SERIAL: 01112028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk5 Part0     OPEN   NORMAL       1121          0    |
   |                                                                                |
   |DEV#:   3  DEVICE NAME: Disk4 Part4  TYPE: 2105E20   SERIAL: 01012028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk4 Part0     OPEN   NORMAL        332          0    |
   |                                                                                |
   |DEV#:   4  DEVICE NAME: Disk3 Part3  TYPE: 2105E20   SERIAL: 00F12028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk3 Part0     OPEN   NORMAL        375          0    |
   |                                                                                |
   |DEV#:   5  DEVICE NAME: Disk2 Part2  TYPE: 2105E20   SERIAL: 31412028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk2 Part0     OPEN   NORMAL        258          0    |
   |                                                                                |
   |DEV#:   6  DEVICE NAME: Disk1 Part1  TYPE: 2105E20   SERIAL: 31312028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk1 Part0     OPEN   NORMAL        267          0    |
   |                                                                                |
   |DEV#:   7  DEVICE NAME: Disk0 Part0  TYPE: 2105E20   SERIAL: 31212028           |
   |=====================================================================           |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk0 Part0     OPEN   NORMAL        268          0    |
   |                                                                                |
   +--------------------------------------------------------------------------------+

Perform the following steps to activate additional paths to a vpath device:

1. Install any additional hardware on the Windows 2000 server or the ESS.
2. Restart the Windows 2000 server.
3. Verify that the path is added correctly. See Verifying additional paths are installed correctly.

Verifying additional paths are installed correctly

After installing additional paths to SDD devices, you should verify that the additional paths have been installed correctly.

Perform the following steps to verify that the additional paths have been installed correctly:

1. Click Start --> Program --> Subsystem Device Driver --> Subsystem Device Driver Management. An MS-DOS window appears.
2. Type datapath query adapter and press Enter. The output should include information about any additional adapters that were installed. In this example, an additional SCSI adapter has been installed. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |Active Adapters :2                                                              |
   |                                                                                |
   |Adpt#     Adapter Name   State     Mode     Select     Errors  Paths  Active    |
   |    0  Scsi Port1 Bus0  NORMAL   ACTIVE       1325          0      8       8    |
   |    1  Scsi Port2 Bus0  NORMAL   ACTIVE       1312          0      8       8    |
   +--------------------------------------------------------------------------------+

3. Type datapath query device and press Enter. The output should include information about any additional devices that were installed. In this example, the output includes information about the new SCSI adapter and the new device numbers that were assigned. The following output is displayed:

   +--------------------------------------------------------------------------------+
   |Total Devices : 8                                                               |
   |                                                                                |
   |DEV#:   0  DEVICE NAME: Disk7 Part7  TYPE: 2105E20   SERIAL: 01312028           |
   |=========================================================================       |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk7 Part0     OPEN   NORMAL        190          0    |
   |    1   Scsi Port2 Bus0/Disk15 Part0     OPEN   NORMAL        179          0    |
   |                                                                                |
   |DEV#:   1  DEVICE NAME: Disk6 Part6  TYPE: 2105E20   SERIAL: 01212028           |
   |=========================================================================       |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk6 Part0     OPEN   NORMAL        179          0    |
   |    1   Scsi Port2 Bus0/Disk14 Part0     OPEN   NORMAL        184          0    |
   |                                                                                |
   |DEV#:   2  DEVICE NAME: Disk5 Part5  TYPE: 2105E20   SERIAL: 01112028           |
   |=========================================================================       |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk5 Part0     OPEN   NORMAL        194          0    |
   |    1   Scsi Port2 Bus0/Disk13 Part0     OPEN   NORMAL        179          0    |
   |                                                                                |
   |DEV#:   3  DEVICE NAME: Disk4 Part4  TYPE: 2105E20   SERIAL: 01012028           |
   |=========================================================================       |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk4 Part0     OPEN   NORMAL        187          0    |
   |    1   Scsi Port2 Bus0/Disk12 Part0     OPEN   NORMAL        173          0    |
   |                                                                                |
   |DEV#:   4  DEVICE NAME: Disk3 Part3  TYPE: 2105E20   SERIAL: 00F12028           |
   |=========================================================================       |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk3 Part0     OPEN   NORMAL        215          0    |
   |    1   Scsi Port2 Bus0/Disk11 Part0     OPEN   NORMAL        216          0    |
   |                                                                                |
   |DEV#:   5  DEVICE NAME: Disk2 Part2  TYPE: 2105E20   SERIAL: 31412028           |
   |=========================================================================       |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk2 Part0     OPEN   NORMAL        115          0    |
   |    1   Scsi Port2 Bus0/Disk10 Part0     OPEN   NORMAL        130          0    |
   |                                                                                |
   |DEV#:   6  DEVICE NAME: Disk1 Part1  TYPE: 2105E20   SERIAL: 31312028           |
   |=======================================================================         |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk1 Part0     OPEN   NORMAL        122          0    |
   |    1    Scsi Port2 Bus0/Disk9 Part0     OPEN   NORMAL        123          0    |
   |                                                                                |
   |DEV#:   7  DEVICE NAME: Disk0 Part0  TYPE: 2105E20   SERIAL: 31212028           |
   |=========================================================================       |
   |Path#              Adapter/Hard Disk    State     Mode     Select     Errors    |
   |    0    Scsi Port1 Bus0/Disk0 Part0     OPEN   NORMAL        123          0    |
   |    1    Scsi Port2 Bus0/Disk8 Part0     OPEN   NORMAL        128          0    |
   +--------------------------------------------------------------------------------+

Chapter 6. Installing and configuring SDD on an HP host system

This chapter provides instructions to install and set up the Subsystem Device Driver on an HP host system attached to an ESS. For updated and additional information not included in this manual, please see the README file on the compact disc or go to the SDD website at: www.ibm.com/storage/support/techsup/swtechsup.nsf/support/sddupdates

Understanding how SDD works on an HP host system

SDD resides above the HP SCSI disk driver (sdisk) in the protocol stack (see Figure 6).

Figure 6. Where SDD fits in the protocol stack

SDD devices behave exactly like sdisk devices. Any operation on an sdisk device can be performed on the SDD device, including commands such as mount, open, close, umount, dd, newfs, or fsck. For example, with SDD you use the mount /dev/dsk/vpath0 /mnt1 command instead of the HP-UX mount /dev/dsk/clt2d0 /mnt1 command.

SDD acts as a pass-through agent. I/O operations sent to SDD are passed to an sdisk driver after path selection. When an active path experiences a failure (such as a cable or controller failure), SDD dynamically switches to another path. The device driver dynamically balances the load, based on the workload of the adapter.

SDD also supports one SCSI adapter on the host system. With single-path access, concurrent download of licensed internal code is supported. However, the load balancing and failover features are not available.

Notes:

1. SDD does not support a system boot from a SDD pseudo device.

2. SDD does not support placing a system paging file on a SDD pseudo device.

Support for 32-bit and 64-bit applications on HP-UX 11.0

SDD supports 32-bit and 64-bit applications on HP-UX 11.0.

Attention: HP patches (as appropriate for a 32-bit or 64-bit application) must be installed on your host system to ensure that SDD operates successfully. See Table 13.

Hardware and software requirements

You must meet the following minimum hardware and software requirements for installing SDD on your HP host system:

• A PA-RISC system running HP-UX 11.00
• A multi-port storage subsystem, such as ESS
• At least one SCSI host adapter (two are required for load balancing and failover)
• You need a SCSI cable to connect each SCSI host adapter to a storage system controller port.
• Subsystem LUNs which have been created and confirmed for multi-port access
• For information on the fibre-channel adapters that can be used on your HP host system go to www.storage.ibm.com/hardsoft/products/ess/supserver.htm
• You need a fiber-optic cable to connect each fibre-channel adapter to a ESS port.

To install SDD and use the input-output load balancing and failover features, you need a minimum of two SCSI or fibre-channel adapters.

Notes:

1. A host server with a single fibre adapter that connects through a switch to multiple ESS ports is considered a multipath fibre-channel connection.

2. A host server with a single-path fibre connection to an ESS is not supported.

3. A host server with SCSI channel connections and a single-path fibre connection to an ESS is not supported.

4. A host server with both a SCSI channel and fibre channel connection to a shared LUN is not supported.

Configuring the ESS

Before you install SDD, configure your ESS for single-port or multiple-port access for each LUN. The Subsystem Device Driver requires that you provide a minimum of two independent paths that share the same logical unit to use the load balancing and failover features.

For information about configuring your ESS, see IBM Enterprise Storage Server Introduction and Planning Guide, GC26-7294.

Planning for installation

Before you install SDD on your HP host, you need to understand what kind of software runs on your host. The way you install SDD depends on the kind of software you have running. There are two types of special device files that are supported:

• Block device files
• Character device files

There are three possible scenarios for installing SDD. The scenario you choose depends on the kind of software you have installed:

Scenario 1

Your system has no software applications (other than UNIX) or DBMSs that talk directly to the HP-UX disk device layer

Scenario 2

Your system already has a software application or DBMS, such as Oracle, that talks directly with the HP-UX disk device layer

Scenario 3

Your system already has SDD and you want to upgrade the software

The following table further describes the various installation scenarios and how you should proceed.

Table 12. SDD installation scenarios

Installation Scenario	Description	How To Proceed
Scenario 1	SDD not installed No software application or DBMS that talks directly to sdisk interface	Go to: Installing the Subsystem Device Driver Standard UNIX applications
Scenario 2	SDD not installed Existing application package or DBMS that talks directly to sdisk interface	Go to: Installing the Subsystem Device Driver Using applications with SDD
Scenario 3	SDD installed	Go to Upgrading the Subsystem Device Driver

For SDD to operate properly on HP-UX 11.0, ensure that the following patches in Table 13 are installed on your host system:

Table 13. HP patches necessary for proper operation of SDD

Application mode:	Install HP Patch:	Patch Description:
32-bit	PHKL_20674	Fix VxFS unmount hang & NMF, sync panics
32-bit	PHKL_20915	Trap-related panics/hangs
32-bit	PHKL_21834	Fibre channel Mass Storage Driver Patch
32-bit	PHKL_22759	SCSI IO Subsystem Cumulative patch
32-bit	PHKL_23001	Signal, threads, spinlock, scheduler, IDS, q3p
32-bit	PHKL_23406	Probe, sysproc, shmem, thread cumulative patch
32-bit or 64-bit	PHKL_21392	VxFS performance, hang, icache, DPFs
32-bit or 64-bit	PHKL_21624	Boot, JFS, PA8600, 3Gdata, NFS, IDS, PM, VM, async
32-bit or 64-bit	PHKL_21989	SCSI IO Subsystem Cumulative patch
64-bit	PHKL_21381	Fibre Channel Mass Storage driver

Installing the Subsystem Device Driver

You need to complete the following procedure if you are installing SDD for the first time on your HP host system:

1. Make sure the SDD compact disc is available.
2. Insert the compact disc into your CD-ROM drive.
3. Mount the CD-ROM drive using the mount command. Here is an example of the mount command:

   mount /dev/dsk/c0t2d0 /cdrom
   

   or

   mount /dev/dsk/c0t2d0 /your_installation_directory
   

   Note:

   /cdrom or /your_installation_directory is the name of the directory you want to mount the CD-ROM drive.

4. Run sam

   > sam
   

5. Select Software Management.
6. Select Install Software to Local Host.
7. At this point, the SD Install - Software Selection panel is displayed. Almost immediately afterwards, a Specify Source menu is displayed:
   • For Source Depot Type, select the local CD-ROM.
   • For Source Depot Path, choose the directory and the IBMdpo.depot file

     For 32-bit mode applications, use:

     /cdrom/hp32bit/IBMdpo.depot

     or

     /your_installation_directory/hp32bit/IBMdpo.depot

     For 64-bit mode applications, use:

     /cdrom/hp64bit/IBMdpo.depot

     or

     /your_installation_directory/hp32bit/IBMdpo.depot

   • Click OK.
8. You will see an output similar to the one in Figure 7 or Figure 8.

   Figure 7. IBMdpo Driver 32-bit
   

   +--------------------------------------------------------------------------------+
   |Name           Revision          Information                        Size(Kb)    |
   |IBMdpo_tag ->  B.11.00.01        IBMdpo Driver 32-bit               nnnn        |
   +--------------------------------------------------------------------------------+

   Figure 8. IBMdpo Driver 64-bit
   

   +--------------------------------------------------------------------------------+
   |Name           Revision          Information                        Size(Kb)    |
   |IBMdpo_tag ->  B.11.00.01        IBMdpo Driver 64-bit               nnnn        |
   +--------------------------------------------------------------------------------+

   1. Choose the IBMdpo_tag product.
   2. Select Actions from the Bar menu, then select Mark for Install.
   3. Select Actions from the Bar menu, then select Install (analysis). You will see an Install Analysis panel, and on it you will see the status of Ready.
   4. Select OK to proceed. A Confirmation panel is displayed which states that the installation will begin.
   5. Type Yes and press Enter. The analysis phase starts.
   6. After the analysis phase has finished, another Confirmation panel is displayed informing you that the system will be restarted after installation is complete. Type Yes and press Enter. The installation of IBMdpo will now proceed.
   7. Next, an Install panel is displayed which informs you about the progress of the IBMdpo software installation. This is what the panel looks like:

      +--------------------------------------------------------------------------------+
      |Press 'Product Summary' and/or 'Logfile' for more target information.           |
      |Target          : XXXXX                                                         |
      |Status          : Building kernel                                               |
      |Percent Complete     : 17%                                                      |
      |Kbytes Installed   :  276 of 1393                                               |
      |Time Left (minutes) : 1                                                         |
      |Product Summary       Logfile                                                   |
      |Done                             Help                                           |
      +--------------------------------------------------------------------------------+

      The Done option is not available when the installation is in progress. It becomes available after the installation process completes.
9. Click Done. A Note window is displayed informing you that the local system will restart with the newly installed software.
10. Select OK to proceed. The following message is displayed on the machine console before it restarts:

    +--------------------------------------------------------------------------------+
    |* A reboot of this system is being invoked. Please wait.                        |
    |                                                                                |
    |*** FINAL System shutdown message (XXXXX) ***                                   |
    |System going down IMMEDIATELY                                                   |
    +--------------------------------------------------------------------------------+

Note:

You can use the datapath query device command You can use the datapath query device command to verify the SDD installation. SDD is successfully installed if the command executes successfully.

Post-installation

After SDD is installed, the device driver resides above the HP SCSI disk driver (sdisk) in the protocol stack. In other words, SDD now talks to the HP-UX device layer. The SDD software installation procedure installs a number of SDD components and updates some system files. Those components and files are listed in the following tables:

Table 14. SDD components installed

File	Location	Description
libvpath.a	/usr/conf/lib	SDD device driver
vpath	/usr/conf/master.d	SDD configuration file
Executables	/opt/IBMdpo/bin	Configuration and status tools
README.sd	/opt/IBMdpo	README file
defvpath	/sbin	SDD configuration file used during startup

Table 15. System files updated

File	Location	Description
system	/stand/build	Forces the loading of the SDD device driver
lvmrc	/etc	Causes defvpath to run at boot time

Table 16. SDD commands and their descriptions

Command	Description
cfgvpath	Configures vpath devices
defvpath	Second part of cfgvpath configuration during boot time
showvpath	Lists the configuration mapping between SDD devices and underlying disks
datapath	SDD driver console command tool

If you are not using a DBMS or an application package that talks directly to the sdisk interface, then the installation procedure is nearly complete. However, you still need to customize HP-UX so that standard UNIX applications can use SDD. Go to section Standard UNIX applications. If you have a DBMS or an application package installed that talks directly to the sdisk interface, such as Oracle, go to Using applications with SDD and read the information specific to the application you will be using.

Note:

During the installation process, the following files were copied from the IBMdpo_depot to the system:

# Kernel-related files

• /usr/conf/lib/libvpath.a
• /usr/conf/master.d/vpath

# SDD driver related files

• /opt/IBMdpo
• /opt/IBMdpo/bin
• /opt/IBMdpo/README.sd
• /opt/IBMdpo/bin/cfgvpath
• /opt/IBMdpo/bin/datapath
• /opt/IBMdpo/bin/defvpath
• /opt/IBMdpo/bin/libvpath.a
• /opt/IBMdpo/bin/pathtest
• /opt/IBMdpo/bin/showvpath
• /opt/IBMdpo/bin/vpath
• /sbin/defvpath

In addition, the /stand/vmunix kernel was created with the device driver. The /stand/system directory was modified in order to add the device driver entry into the file. After these files were created, the /opt/IBMdpo/bin/cfgvpath program was initiated in order to create vpaths in the /dev/dsk and /dev/rdsk directories for all IBM disks which are available on the system. This information is stored in the /opt/IBMdpo file for use after rebooting the machine.

Note:

SDD devices are found in /dev/rdsk and /dev/dsk. The device is named according to the SDD number. A device with a number of 0 would be /dev/rdsk/vpath0.

Upgrading the Subsystem Device Driver

Upgrading the SDD consists of uninstalling and reinstalling the IBMdpo package. If you are upgrading SDD, go to Uninstalling the Subsystem Device Driver and then go to Installing the Subsystem Device Driver.

Using applications with SDD

If your system already has a software application or an DBMS installed that communicates directly with the HP-UX disk device drivers, you need to insert the new SDD device layer between the software application and the HP-UX disk device layer. You also need to customize the software application in order to have it communicate with the SDD devices instead of the HP-UX devices.

In addition, many software applications and DBMSs need to control certain device attributes such as ownership and permissions. Therefore, you must ensure that the new SDD devices that these software applications or DBMSs access in the future have the same attributes as the HP-UX sdisk devices that they replace. You need to customize the application or DBMS to accomplish this.

This section contains the procedures for customizing the following software applications and DBMS for use with SDD:

• Standard UNIX applications
• Network File System file systems
• Oracle.

Standard UNIX applications

If you have not already done so, install SDD using the procedure in Installing the Subsystem Device Driver. When this is done, SDD resides above the HP SCSI disk driver (sdisk) in the protocol stack. In other words, SDD now talks to the HP-UX device layer. To use standard UNIX applications with SDD, you must make some changes to your logical volumes. You must either convert your existing logical volumes or create new ones.

Standard UNIX applications such as newfs, fsck, mkfs, and mount, that normally take a disk device or raw disk device as a parameter, also accept the SDD device as a parameter. Similarly, entries in files such as vfstab and dfstab (in the format of cntndnsn) can be replaced by entries for the corresponding SDD devices' vpathNs. Make sure that the devices that are replaced are replaced with the corresponding SDD device. Running the showvpath command lists all SDD devices and their underlying disks.

In order to use the SDD driver for an existing logical volume, it is necessary to remove the existing logical volume and volume group and recreate it using the SDD device.

Attention: Do not use the SDD for critical file systems needed at boot time, such as /(root), /stand, /usr, /tmp or /var. Doing so may render your system unusable if SDD is ever uninstalled (for example, as part of an upgrade).

Converting existing logical volumes

The task of converting an existing logical volume to use SDD can be broken down into the following subtasks:

1. Determining the size of the logical volume
2. Removing the existing logical volume
3. Removing the existing volume group
4. Recreating the logical volume.

Note:

You must have super-user privileges to perform these subtasks.

As an example, suppose you have a logical volume called lvol1 under a volume group vgibm, which is currently using the disk directly, (for example, through path /dev path /dev/dsk/c3t4d0). You would like to convert logical volume lvol1 to use SDD. In order to recreate the logical volume, you first need to determine the size of the logical volume.

Determining the size of the logical volume

Use the lvdisplay command to determine this:

# lvdisplay | grep LV Size

A message is displayed:

+--------------------------------------------------------------------------------+
|LV Size (Mbytes) 100                                                            |
+--------------------------------------------------------------------------------+

In this case, the logical volume size is 100 megabytes. Next, remove the logical volume from the system.

Removing the existing logical volume

Before the logical volume is removed, it must be unmounted. Here is an example of using the umount command to unmount logical volume lvol1:

# umount /dev/vgibm/lvol1

Next, remove the logical volume. You can use the following command to remove logical volume lvol1:

# lvremove /dev/vgibm/lvol1

A message is displayed:

+--------------------------------------------------------------------------------+
|The logical volume "/dev/vgibm/lvol1" is not empty;                             |
|do you really want to delete the logical volume (y/n)                           |
+--------------------------------------------------------------------------------+

Type Y and press Enter. A message is displayed that is similar to the following:

+--------------------------------------------------------------------------------+
|Logical volume "/dev/vgibm/lvol1" has been successfully removed.                |
|Volume Group configuration for /dev/vgibm has been saved in                     |
|/etc/lvmconf/vgibm.conf                                                         |
+--------------------------------------------------------------------------------+

When prompted to delete the logical volume, type y.

Next, remove the volume group.

Removing the existing volume group

You can use the following command to remove the volume group vgibm:

# vgremove /dev/vgibm

You see a message similar to this:

+--------------------------------------------------------------------------------+
|Volume group "/dev/vgibm" has been successfully removed.                        |
+--------------------------------------------------------------------------------+

Now recreate the logical volume.

Recreating the logical volume

Recreating the logical volume consists of a number of smaller steps:

1. Recreating the physical volume
2. Recreating the volume group
3. Recreating the logical volume
4. Setting the proper timeout value for the logical volume manager.

Recreating the physical volume

Use the following command to recreate the physical volume:

# pvcreate /dev/rdsk/vpath0

You see a message similar to this:

+--------------------------------------------------------------------------------+
|Physical volume "/dev/rdsk/vpath0" has been successfully created.               |
+--------------------------------------------------------------------------------+

This assumes that the SDD device associated with the underlying disk is vpath0. Verify this with the showvpath command:

# /opt/IBMdpo/bin/showvpath

A message is displayed:

+--------------------------------------------------------------------------------+
|vpath0:                                                                         |
|	/dev/dsk/c3t4d0                                                                |
+--------------------------------------------------------------------------------+

Next, recreate the volume group.

Recreating the volume group

Use the following command to recreate the volume group:

# vgcreate /dev/vgibm /dev/dsk/vpath0

You see a message that says:

+--------------------------------------------------------------------------------+
|Increased the number of physical extents per physical volume to 2187.           |
|Volume group "/dev/vgibm" has been successfully created.                        |
|Volume Group configuration for /dev/vgibm has been saved in                     |
|/etc/lvmconf/vgibm.conf                                                         |
+--------------------------------------------------------------------------------+

Now recreate the logical volume.

Recreating the logical volume

Attention: The recreated logical volume should be the same size as the original volume; otherwise, the recreated volume cannot store the data that was on the original.

Use the following command to recreate the logical volume:

# lvcreate -L 100 -n lvol1 vgibm

You see a message that says:

+--------------------------------------------------------------------------------+
|Logical volume "/dev/vgibm/lvol1" has been successfully created with            |
|character device "/dev/vgibm/rlvol1".                                           |
|Logical volume "/dev/vgibm/lvol1" has been successfully extended.               |
|Volume Group configuration for /dev/vgibm has been saved in                     |
|/etc/lvmconf/vgibm.conf                                                         |
+--------------------------------------------------------------------------------+

Note that the -L 100 parameter comes from the size of the original logical volume, determined by using the lvdisplay command. In this example, the original logical volume was 100 MB in size.

Setting the correct timeout value for the logical volume manager

Attention: The timeout values for the logical volume manager must be set correctly for SDD to operate properly. This is particularly true if you are going to be using concurrent microcode download.

If you are going to be using concurrent microcode download with single-path SCSI, perform the following steps to set the correct timeout value for the logical volume manager:

1. Ensure the timeout value for a SDD logical volume is set to default. Type lvdisplay /dev/vgibm/lvol1 and press Enter. If the timeout value is not default, type lvchange -t 0 /dev/vgibm/lvol1 and press Enter to change it. (vgibm is the name of the logical volume group previously configured to use SDD; in your environment the name may be different.)
2. Change the timeout value for a SDD physical volume to 240. Type pvchange -t 240 /dev/dsk/vpathn and press Enter. (n refers to the vpath number.) If you are not sure about the vpath number, type /opt/IBMdpo/bin/showvpath and press Enter to obtain this information.

If you are going to be using concurrent microcode download with multi-path SCSI, perform the following steps to set the proper timeout value for the logical volume manager:

1. Ensure the timeout value for a SDD logical volume is set to default. Type lvdisplay /dev/vgibm/lvoly and press Enter. If the time-out value is not default, type lvchange -t 0 /dev/vgibm/lvoly and press Enter to change it. (vgibm is the name of logical volume group previously configured to use SDD; in your environment the name may be different, y=[0,1,2,...].)
2. Change the timeout value for a SDD physical volume to 240. Type pvchange -t 240 /dev/dsk/vpathn and press Enter. (n refers to the vpath number.) If you are not sure about the vpath number, type /opt/IBMdpo/bin/showvpath and press Enter to obtain this information.

Note:

The recreated logical volume must be mounted before it can be accessed.

Attention: In some cases it may be necessary to use standard HP recovery procedures to fix a volume group that has become damaged or corrupted. For information on using recovery procedures, such as, vgscan, vgextend, vpchange, or vgreduce, see the HP-UX Reference Volume 2 at the Web site: docs.hp.com.

Creating new logical volumes

The task of creating a new logical volume to use SDD can be broken down into the following subtasks:

Note:

You must have super-user privileges to perform the following subtasks.

1. Determining the major number of the logical volume device
2. Creating a device node for the logical volume device
3. Creating a physical volume
4. Creating a volume group
5. Creating a logical volume
6. Creating a file system on the volume group
7. Mounting the logical volume.

In order to create a new logical volume that uses SDD, you first need to determine the major number of the logical volume device.

Determining the major number of the logical volume device

Use the lsdev command to determine this:

# lsdev | grep lv

A message is displayed:

+--------------------------------------------------------------------------------+
|64          64         lv              lvm                                      |
+--------------------------------------------------------------------------------+

The first number is the major number of the character device, which is what you want to use. Next, create a device node for the logical volume device.

Creating a device node for the logical volume device

Creating a device node actually consists of:

1. Creating a directory in /dev for the volume group
2. Changing to the /dev directory
3. Creating a device node for the logical volume device.

Creating a directory in /dev for the volume group

Use the following command to create a directory in /dev for the volume group:

# mkdir /dev/vgibm

In this example, vgibm is the name of the directory.

Next, change to the directory that you just created

Changing to the /dev directory

Use the following command to change to the /dev directory:

# cd /dev/vgibm

Next, create a device node for the logical volume device.

Creating a device node for the logical volume device

If you do not have any other logical volume devices, you can use a minor number of 0x010000. In this example, assume that you have no other logical volume devices. Use the following command to create a device node:

# mknod group c 64 0x010000

Now create a physical volume.

Creating a physical volume

Use the following command to create a physical volume:

# pvcreate /dev/rdsk/vpath0

Now create the volume group

Creating a volume group

Use the following command to create a volume group:

# vgcreate /dev/vgibm /dev/dsk/vpath0

Now create the logical volume.

Creating a logical volume

Use the following command to create logical volume lvol1 :

# lvcreate -L 100 -n lvol1 vgibm

The -L 100 makes a 100 MB volume group; you can make it larger if you want to. Now you are ready to create a file system on the volume group.

Creating a file system on the volume group

Use the following command to create a file system on the volume group:

# newfs -F hfs /dev/vgibm/rlvol1

Finally, mount the logical volume (assuming that you have a mount point called /mnt).

Mounting the logical volume

Use the following command to mount the logical volume lvol1:

# mount /dev/vgibm/lvol1 /mnt

Attention: In some cases it may be necessary to use standard HP recovery procedures to fix a volume group that has become damaged or corrupted. For information on using recovery procedures, such as, vgscan, vgextend, vpchange, or vgreduce, see the HP-UX Reference Volume 2 at the website: docs.hp.com.

Network File System file server

The procedures in this section show how to install SDD for use with an exported file system (Network File System file server).

Setting up Network File System for the first time

Follow the instructions in this section if you are installing exported file systems on SDD devices for the first time:

1. If you have not already done so, install SDD using the procedure in Installing the Subsystem Device Driver.
2. Determine which SDD (vpathN) volumes you will use as file system devices.
3. Create file systems on the selected SDD devices using the appropriate utilities for the type of file system you will use. If you are using the standard HP-UX LJFS file system, use the following command:

   # newfs /dev/rdsk/vpathN
   

   In this example, N is the SDD device instance of the selected volume. Create mount points for the new file systems.

4. Install the file systems into the directory /etc/fstab. Be sure to set the mount at boot field to yes.
5. Install the file system mount points into /etc/exports for export.
6. Reboot.

Installing SDD on a system that already has Network File System file server

Follow the instructions in this section if you have Network File System file server already configured for exported file systems that reside on a multi-port subsystem, and if you want to use SDD partitions instead of sdisk partitions to access them.

1. List the mount points for all currently exported file systems by looking in the /etc/exports directory.
2. Match the mount points found in step 1 with sdisk device link names (files named /dev/(r)dsk/cntndn) by looking in the /etc/fstab directory.
3. Match the sdisk device link names found in step 2 with SDD device link names (files named /dev/(r)dsk/vpathN) by running the showvpath command.
4. Make a backup copy of the current /etc/fstab file.
5. Edit the /etc/fstab file, replacing each instance of an sdisk device link named /dev/(r)dsk/cntndn with the corresponding SDD device link.
6. Reboot. Verify that each exported file system passes the boot time fsck pass, that each mounts properly, and that each is exported and available to NFS clients.

If there is a problem with any exported file system after completing step 6, restore the original /etc/fstab file and reboot to restore Network File System service. Then review your steps and try again.

Oracle

Notes:

1. Procedures listed below require you to have Oracle documentation on hand.

2. You must have super-user privileges to perform these procedures.

3. These procedures were tested with Oracle 8.0.5 Enterprise server, with the 8.0.5.1 patch set from Oracle.

Installing an Oracle database for the first time

You can set up your Oracle database in one of two ways. You can set it up to use a file system or raw partitions. The procedure for installing your database differs depending on the choice you make.

If using a file system

1. If you have not already done so, install SDD using the procedure in Installing the Subsystem Device Driver.
2. Create and mount file systems on one or more SDD partitions (Oracle recommends three mount points on different physical devices).
3. Follow the Oracle Installation Guide for instructions on installing to a file system. (During the Oracle installation, you will be asked to name three mount points. Supply the mount points for the file systems you created on the SDD partitions).

If using raw partitions

Notes:

1. Make sure that the ownership and permissions of the SDD devices are the same as the ownership and permissions of the raw devices they are replacing.

2. Make sure that all the databases are closed before making changes.

In the following procedure you will be replacing the raw devices with the SDD devices.

1. If you have not already done so, install SDD using the procedure in Installing the Subsystem Device Driver.
2. Create the Oracle Software Owner user in the server's local /etc/passwd file. You must also complete the following related activities:
   1. Complete the rest of the Oracle pre-installation tasks described in the Oracle8 Installation Guide.
   2. Plan the installation of Oracle8 on a file system residing on a SDD partition.
   3. Set up the Oracle user's ORACLE_BASE and ORACLE_ HOME environment variables to be directories of this file system.
   4. Create two more SDD-resident file systems on two other SDD volumes. Each of the resulting three mount points should have a subdirectory named oradata, to be used as a control file and redo log location for the Installer's Default Database (a sample database) as described in the Oracle8 Installation Guide. Oracle recommends using raw partitions for redo logs. To use SDD raw partitions as redo logs, create symbolic links from the three redo log locations to SDD raw device links (files named /dev/rdsk/vpathNs, where N is the SDD instance number, and s is the partition ID) that point to the slice.
3. Determine which SDD (vpathN) volumes you will use as Oracle8 database devices.
4. Partition the selected volumes using the HP-UX format utility. If SDD raw partitions are to be used by Oracle8 as database devices, be sure to leave disk cylinder 0 of the associated volume unused. This protects UNIX disk labels from corruption by Oracle8, as described in the Oracle8 Installation Guide in the information on raw devices.
5. Ensure that the Oracle Software Owner has read and write privileges to the selected SDD raw partition device files under the /devices directory.
6. Set up symbolic links from the oradata directory (under the first of the three mount points) that link the database files system<db>.dbf, tempdb.dbf, rbsdb.dbf, toolsd.bdbf, and usersdb.dbf to SDD raw device links (files named /dev/rdsk/vpathNs) pointing to partitions of the appropriate size, where " db" is the name of the database that you are creating. (The default is test.)
7. Install the Oracle8 Server following the instructions in the Oracle8 Installation Guide. Be sure to be logged in as the Oracle Software Owner when you run the orainst /m command. Select the Install New Product - Create Database Objects option. Select Raw Devices for storage type. Specify the raw device links set up in steps 2 and 6 for the redo logs and database files of the default database.
8. To set up other Oracle8 databases you must set up control files, redo logs, and database files following the guidelines in the Oracle8 Administrator's Reference. Make sure any raw devices and file systems you set up reside on SDD volumes.
9. Launch the sqlplus utility.
10. Use the create database SQL command, specifying the control, log, and system data files that you have set up.
11. Use the create tablespace SQL command to set up each of the temp, rbs, tools, and users database files that you created.
12. Use the create rollback segment SQL command to create the three redo log files that you set. For the syntax of these three create commands, see the Oracle8 Server SQL Language Reference Manual.

Installing SDD on a system that already has Oracle in place

Your installation procedure for a new SDD install will differ depending on whether you are using a file system or raw partitions for your Oracle database.

If using a file system

Follow this procedure if you are installing SDD for the first time on a system with an Oracle database that uses a file system:

1. Record the raw disk partitions being used (they are in the cntndnsn format) or the partitions where the Oracle file systems reside. You can get this information from /etc/vfstab if you know where the Oracle files are. Your database administrator can tell you where the Oracle files are, or you can check for directories with the name oradata.
2. Complete the basic installation steps in Installing the Subsystem Device Driver.
3. Change to the directory where you installed the SDD utilities. Run the showvpath command.
4. Check the display to see whether you find a cntndn directory that is the same as the one where the Oracle files are.
5. Use the SDD partition identifiers instead of the original HP-UX identifiers when mounting the file systems.

   If you would originally have used:

   mount /dev/dsk/c1t3d2 /oracle/mp1
   

   You now use:

   mount /dev/dsk/vpath2 /oracle/mp1
   

   (assuming that you had found vpath2 to be the SDD identifier)

Follow the instructions in the Oracle Installation Guide for setting ownership and permissions.

If using raw partitions

Follow this procedure if you have Oracle8 already installed and want to reconfigure it to use SDD partitions instead of sdisk partitions (for example, partitions accessed through /dev/rdsk/cntndn files).

All Oracle8 control, log, and data files are accessed either directly from mounted file systems, or using links from the oradata subdirectory of each Oracle mount point set up on the server. Therefore, the process of converting an Oracle installation from sdisk to SDD has two parts:

• Changing the Oracle mount points' physical devices in /etc/fstab from sdisk device partition links to the SDD device partition links that access the same physical partitions.
• Recreating links to raw sdisk device links to point to raw SDD device links that access the same physical partitions.

Converting an Oracle installation from sdisk to SDD

Following are the conversion steps:

1. Back up your Oracle8 database files, control files, and redo logs.
2. Obtain the sdisk device names for the Oracle8 mounted file systems by looking up the Oracle8 mount points in /etc/fstab and extracting the corresponding sdisk device link name. (for example, /dev/rdsk/c1t4d0)
3. Launch the sqlplus utility.
4. Type the command:

   select * from sys.dba_data_files;
   

   Determine the underlying device that each data file resides on, either by looking up mounted file systems in /etc/fstab, or by extracting raw device link names directly from the select command output.

5. Fill in the following table, which is for planning purposes:
   

   Oracle Device Link	File Attributes			SDD Device Link
   Owner		Group	Permissions
   /dev/rdsk/c1tld0	oracle	dba	644	/dev/rdsk/vpath4

6. Fill in column 2 by running ls -l on each device link listed in column 1 and extracting the link source device file name.
7. Fill in the File Attributes columns by running ls -l on each Actual Device Node from column 2.
8. Install SDD following the instructions in the Installing the Subsystem Device Driver.
9. Fill in the Subsystem Device Driver Device Links column by matching each cntndnsn device link listed in the Oracle Device Link column with its associated vpathN device link name by running the command:

   /opt/IBMdpo/bin/showvpath
   

10. Fill in the Subsystem Device Driver Device Nodes column by running ls -l on each SDD Device Link and tracing back to the link source file.
11. Change the attributes of each node listed in the Subsystem Device Driver Device Nodes column to match the attributes listed to the left of it in the File Attributes column using the UNIX chown, chgrp, and chmod commands.
12. Make a copy of the existing /etc/fstab file. Edit the /etc/fstab file, changing each Oracle device link to its corresponding SDD device link.
13. For each link found in an oradata directory, recreate the link using the appropriate SDD device link as the source file instead of the associated sdisk device link listed in the Oracle Device Link column.
14. Reboot the server.
15. Verify that all file system and database consistency checks complete successfully.

Uninstalling the Subsystem Device Driver

Note:

You must uninstall the current level of SDD must be uninstalled before upgrading to a newer level.

Complete the following procedure to uninstall SDD:

1. Reboot or unmount all SDD file systems.
2. If you are using SDD with a database, such as Oracle, edit the appropriate database configuration files (database partition) to remove all the SDD devices.
3. Run sam

   > sam
   

4. Select Software Management.
5. Choose Remove Software.
6. Choose Remove Local Host Software.
7. Choose the IBMdpo_tag selection.
   1. Select Actions from the Bar menu, then select Mark for Remove.
   2. Select Actions from the Bar menu, then select Remove (analysis). You will see a Remove Analysis panel, and on it you will see the status of Ready.
   3. Select OK to proceed. A Confirmation panel is displayed which states that the uninstall will begin.
   4. Type Yes. The analysis phase starts.
   5. After the analysis phase has finished, another Confirmation panel is displayed informing you that the system will be rebooted after the uninstall is complete. Type Yes and press Enter. The uninstall of IBMdpo will now proceed.
   6. Next, an Uninstall panel is displayed which informs you about the progress of the IBMdpo software uninstall. This is what the panel looks like:

      +--------------------------------------------------------------------------------+
      |Target        : XXXXX                                                           |
      |Status        : Executing unconfigure                                           |
      |Percent Complete    : 17%                                                       |
      |Kbytes Removed      : 340 of 2000                                               |
      |Time Left (minutes) : 5                                                         |
      |Removing Software   : IBMdpo_tag,...........                                    |
      +--------------------------------------------------------------------------------+

      The Done option is not available when the installation is in progress. It becomes available after the installation process completes.
8. Click Done. A Note panel is displayed informing you that the local system will reboot with the newly installed software.
9. Select OK to proceed. The following message is displayed on the machine console before it reboots:

   +--------------------------------------------------------------------------------+
   |* A reboot of this system is being invoked. Please wait.                        |
   |                                                                                |
   |*** FINAL System shutdown message (XXXXX) ***                                   |
   |System going down IMMEDIATELY                                                   |
   +--------------------------------------------------------------------------------+

Note:

When Subsystem Device Driver has been successfully uninstalled, the first part of the procedure for upgrading the Subsystem Device Driver is complete. To complete an upgrade, you need to reinstall Subsystem Device Driver. See the installation procedure in Installing the Subsystem Device Driver.

The uninstall of SDD involved the following actions:

• The /sbin/defvpath file was removed.
• The /usr/conf/master.d/vpath, and /usr/conf/lib/libvpath.a files were removed.
• The files in the /opt/IBMdpo directory were removed.
• The /opt/IBMdpo directory was removed.
• The SDD driver entry was removed from the /stand/system file.
• The SDD driver was removed from the /stand/vmunix kernel.

Changing a SDD hardware configuration

When adding or removing multi-port SCSI devices from your system, you must reconfigure SDD to recognize the new devices. Perform the following steps to reconfigure SDD:

1. Reboot the system:

   shutdown -r 0
   

2. Run cfgvpath to reconfigure vpath:

   /opt/IBMdpo/bin/cfgvpath -c
   

3. Reboot the system:

   shutdown -r 0
   

Chapter 7. Installing and configuring SDD on a Sun host system

This chapter provides instructions to install and set up the Subsystem Device Driver on an host system attached to an ESS. For updated and additional information not included in this manual, see the README file on the compact disc or visit the Subsystem Device Driver Web site www.ibm.com/storage/support/techsup/swtechsup.nsf/support/sddupdates

Understanding how SDD works on a Sun host

SDD resides above the Sun SCSI disk driver (sd) in the protocol stack. There can be a maximum of eight sd devices underneath each SDD device in the protocol stack. Each sd device represents a different path to the physical device. There can be up to eight sd devices that represent up to eight different paths to the physical device.

Figure 9. Where SDD fits in the protocol stack

SDD devices behave exactly like sd devices. Any operation on an sd device can be performed on the SDD device, including commands such as mount, open, close, umount, dd, newfs, or fsck. For example, with SDD you enter mount /dev/dsk/vpath0c /mnt1 instead of the Solaris mount /dev/dsk/c1t2d0s2 /mnt1 command.

SDD acts as a pass-through agent. I/Os sent to the device driver are passed to an sdisk driver after path selection. When an active path experiences a failure (such as a cable or controller failure), the device driver dynamically switches to another path. The device driver dynamically balances the load, based on the workload of the adapter.

The SDD also supports one SCSI adapter on the host system. With single-path access, concurrent download of licensed internal code is supported. However, the load balancing and failover features are not available.

Notes:

1. SDD only supports 32-bit applications on Solaris 2.6.

2. SDD supports 32-bit and 64-bit applications on Solaris 7.

3. SDD supports 32-bit and 64-bit applications on Solaris 8.

4. SDD does not support a system boot from a SDD pseudo device.

5. SDD does not support placing a system paging file on a SDD pseudo device.

Hardware and software requirements

You must meet the following minimum hardware and software requirements to install the SDD on your host system:

• A Sparc system running Solaris 2.6, Solaris 7, or Solaris 8
• A multi-port storage subsystem; for example, multi-active redundant RAID controllers (ESS)
• One or more pairs of SCSI or fibre-channel host adapters
• Subsystem LUNs which have been created and confirmed for multi-port access. There should be up to eight sdisk instances for each LUN. (One for each path on the server)
• A SCSI cable to connect each SCSI host adapter to a storage system controller port per cable
• A fiber-optic cable to connect each fibre-channel adapter to a ESS port
• For information on the SCSI or fibre-channel adapters that can be used on your Sun host system go to the website www.storage.ibm.com/hardsoft/products/ess/supserver.htm

To install SDD and use the input-output load balancing and failover features, you need a minimum of two SCSI or fibre-channel adapters.

Notes:

1. A host server with a single fibre adapter that connects through a switch to multiple ESS ports is considered a multipath fibre-channel connection.

2. A host server with a single-path fibre connection to an ESS is not supported.

3. A host server with SCSI channel connections and a single-path fibre connection to an ESS is not supported.

4. A host server with both a SCSI channel and fibre-channel connection to a shared LUN is not supported.

Configuring the ESS

Before you install SDD, configure your ESS for single-port or multiple-port access for each LUN. SDD requires a minimum of two independent paths that share the same logical unit to use the load balancing and failover features.

For information about configuring your ESS, see IBM Enterprise Storage Server Introduction and Planning Guide, GC26-7294.

Planning for installation

Before you install SDD on your Sun host, you need to understand what kind of software is running on it. The way you install SDD depends on the kind of software you are running. Basically, there are three types of software that talk directly to raw or block disk device interfaces such as sd and SDD:

• UNIX file systems, where there is no logical volume manager present
• Logical volume managers (LVMs), such as Sun's Solstice Disk Suite. LVMs allow the system manager to logically integrate, for example, several different physical volumes to create the image of a single large volume.
• Major application packages, such as certain database managers (DBMSs).

There are three possible scenarios for installing SDD. The scenario you choose depends on the kind of software you have installed:

Scenario 1

Your system has no volume manager, DBMS, or software applications (other than UNIX) that talk directly to the Solaris disk device layer.

Scenario 2

Your system already has a volume manager, software application, or DBMS, such as Oracle, that talks directly with the Solaris disk device drivers.

Scenario 3

Your system already has SDD and you want to upgrade the software.

Table 17 further describes the various installation scenarios and how you should proceed.

Table 17. SDD installation scenarios

Installation Scenario	Description	How To Proceed
Scenario 1	Subsystem Device Driver not installed No volume managers No software application or DBMS installed that talks directly to sd interface	Go to: Installing the Subsystem Device Driver Standard UNIX applications
Scenario 2	Subsystem Device Driver not installed Existing volume manager, software application, or DBMS installed that talks directly to sd interface	Go to: Installing the Subsystem Device Driver Using applications with SDD
Scenario 3	Subsystem Device Driver installed	Go to: Upgrading the Subsystem Device Driver

Table 18 lists the install package file names that come with SDD.

Table 18. SDD package file names

Package file names	Description
sun32bit/IBMdpo	Solaris 2.6
sun64bit/IBMdpo	Solaris 7
sun64bit/IBMdpo	Solaris 8

For SDD to operate properly, ensure that the Solaris patches in Table 19 are installed on your operating system.

Table 19. Solaris patches necessary for proper operation of SDD

	Solaris 2.6	Solaris 7
glm	105580-15	106925-04
isp	105600-19	106924-06
sd & ssd	105356-16	107458-10

Attention: Analyze and study your operating and application environment to ensure there are no conflicts with these patches prior to their installation.

Go to the following website for the latest information on Solaris patches sunsolve.Sun.COM

Installing the Subsystem Device Driver

You need to complete the following procedure if you are installing SDD for the first time on your Sun host.

1. Make sure the SDD compact disc is available.
2. Insert the compact disc into your CD-ROM drive.
3. Change to the install directory:

   # cd /cdrom/cdrom0/sun32bit or
   # cd /cdrom/cdrom0/sun64bit 
   

4. Run pkgadd, and point the -d option of pkgadd to the directory containing IBMdpo. For example,:

   pkgadd -d /cdrom/cdrom0/sun32bit IBMdpo or
   pkgadd -d /cdrom/cdrom0/sun64bit IBMdpo
   

5. You should messages similar to this:

   +--------------------------------------------------------------------------------+
   |Processing package instance <IBMdpo> from <var/spool/pkg>                       |
   |                                                                                |
   |                                                                                |
   |IBM DPO driver                                                                  |
   |(sparc) 1                                                                       |
   |## 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 <IBMdpo> [y,n,?]               |
   +--------------------------------------------------------------------------------+

6. Type Y and press Enter to proceed.
7. You should see messages similar to this:

   +--------------------------------------------------------------------------------+
   |Installing IBM DPO driver as <IBMdpo>                                           |
   |                                                                                |
   |## Installing part 1 of 1.                                                      |
   |/etc/defvpath                                                                   |
   |/etc/rc2.d/S00vpath-config                                                      |
   |/etc/rcS.d/S20vpath-config                                                      |
   |/kernel/drv/vpathdd                                                             |
   |/kernel/drv/vpathdd.conf                                                        |
   |/opt/IBMdpo/cfgvpath                                                            |
   |/opt/IBMdpo/datapath                                                            |
   |/opt/IBMdpo/devlink.vpath.tab                                                   |
   |/opt/IBMdpo/etc.system                                                          |
   |/opt/IBMdpo/pathtest                                                            |
   |/opt/IBMdpo/showvpath                                                           |
   |/usr/sbin/vpathmkdev                                                            |
   |[ verifying class <none>  ]                                                     |
   |## Executing postinstall script.                                                |
   |                                                                                |
   |DPO: Configuring 24 devices (3 disks * 8 slices)                                |
   |                                                                                |
   |Installation of <IBMdpo> was successful.                                        |
   |                                                                                |
   |The following packages are available:                                           |
   |1 IBMcli ibm2105cli                                                             |
   |         (sparc) 1.1.0.0                                                        |
   |2 IBMdpo IBM DPO driver Version: May-10-2000 16:51                              |
   |         (sparc) 1                                                              |
   |Select package(s) you wish to process (or 'all' to process                      |
   |all packages). (default: all) [?,??,q]:                                         |
   +--------------------------------------------------------------------------------+

   Type q and press Enter to proceed.
8. You should see messages similar to this:

   +--------------------------------------------------------------------------------+
   |*** IMPORTANT NOTICE ***                                                        |
   |This machine must now be rebooted in order to ensure                            |
   |sane operation. Execute                                                         |
   |       shutdown -y -i6 -g0                                                      |
   |and wait for the "Console Login:" prompt.                                       |
   |                                                                                |
   |DPO is now installed. Proceed to Post-Installation.                             |
   +--------------------------------------------------------------------------------+

Note:

You can verify that SDD has been successfully installed by issuing the datapath query device command. If the command executes, SDD is installed.

Post-installation

After the installation is complete, manually unmount the compact disc. Run the umount /cdrom command from the root directory. Go to the CD-ROM drive and press the Eject button.

After SDD is installed, your system must be rebooted to ensure proper operation. Type the command:

# shutdown -i6 -g0 -y

Note:

SDD devices are found in the /dev/rdsk and /dev/dsk directories. The device is named according to the SDD instance number. A device with an instance number of 0 would be: /dev/rdsk/vpath0a where a denotes the slice. Therefore, /dev/rdsk/vpath0c would be instance zero (0) and slice 2.

After SDD is installed, the device driver resides above the Sun SCSI disk driver (sd) in the protocol stack. In other words, SDD now talks to the Solaris device layer. The SDD software installation procedure installs a number of SDD components and updates some system files. Those components and files are listed in the following tables:

Table 20. System files updated

File	Location	Description
/etc/system	/etc	Forces the loading of SDD
/etc/devlink.tab	/etc	Tells the system how to name SDD devices in /dev

Table 21. Subsystem Device Driver components installed

File	Location	Description
vpathdd	/kernel/drv	Device driver
vpathdd.conf	/kernel/drv	SDD config file
Executables	/opt/IBMdpo/bin	Configuration and status tools
S20vpath-config	/etc/rcS.d	Boot initialization script*

Table 22. SDD commands and their descriptions

Command	Description
cfgvpath	Configures vpath devices
showvpath	Lists all SDD devices and their underlying disks
vpathmkdev	Create SDD devices for /dev/dsk entries
datapath	SDD driver console command tool

Note:

* This script must come before other LVM initialization scripts, such as Veritas initialization scripts.

If you are not using a volume manager, software application, or DBMS that talks directly to the sd interface, then the installation procedure is nearly complete. If you have a volume manager, software application, or DBMS installed that talks directly to the sd interface, such as Oracle, go to Using applications with SDD and read the information specific to the application you will be using.

Upgrading the Subsystem Device Driver

Upgrading SDD consists of uninstalling and reinstalling the IBMdpo package. If you are upgrading SDD, go to Uninstalling the Subsystem Device Driver and then go to Installing the Subsystem Device Driver.

Using applications with SDD

If your system already has a volume manager, software application, or DBMS installed that communicates directly with the Solaris disk device drivers, you need to insert the new SDD device layer between the program and the Solaris disk device layer. You also need to customize the volume manager, software application, or DBMS in order to have it communicate with the SDD devices instead of the Solaris devices.

In addition, many software applications and DBMSs need to control certain device attributes such as ownership and permissions. Therefore, you must ensure that the new SDD devices that these software applications or DBMSs access in the future have the same attributes as the Solaris sd devices that they replace. You need to customize the software application or DBMS to accomplish this.

This section describes how to use the following applications with SDD:

• Standard UNIX applications
• Network File System file systems
• Oracle
• Veritas Volume Manager.

Standard UNIX applications

If you have not already done so, install SDD using the procedure in section Installing the Subsystem Device Driver. When this is done, the device driver resides above the Solaris SCSI disk driver (sd) in the protocol stack. In other words, SDD now talks to the Solaris device layer.

Standard UNIX applications, such as newfs, fsck, mkfs, and mount, that normally take a disk device or raw disk device as a parameter, also accept the SDD device as a parameter. Similarly entries in files such as vfstab and dfstab (in the format of cntndnsn) can be replaced by entries for the corresponding SDD devices' vpathNs. Make sure that the devices that are replaced are replaced with the corresponding SDD device. Running the showvpath command lists all SDD devices and their underlying disks.

Note:

SDD does not support being used for the root (/), /var, /usr, /opt, /tmp and swap partitions.

Network File System file server

The procedures in this section show how to install SDD for use with an Exported File System (Network File System file server).

Setting up Network File System for the first time

Follow the instructions in this section if you are installing exported file systems on SDD devices for the first time:

1. If you have not already done so, install SDD using the procedure in Installing the Subsystem Device Driver.
2. Determine which SDD (vpathN) volumes you will use as file system devices.
3. Partition the selected volumes using the Solaris format utility.
4. Create file systems on the selected SDD devices using the appropriate utilities for the type of file system you will use. If you are using the standard Solaris UFS file system, use the following command:

   # newfs /dev/rdsk/vpathNs
   

   In this example, N is the SDD device instance of the selected volume. Create mount points for the new file systems.

5. Install the file systems into the /etc/fstab directory. Be sure to set the mount at boot field to yes.
6. Install the file system mount points into the directory /etc/exports for export.
7. Reboot.

Installing SDD on a system that already has Network File System file server

Follow the instructions in this section if you have Network File System file server already configured for exported file systems that reside on a multiport subsystem, and if you want to use SDD partitions instead of sd partitions to access them.

1. List the mount points for all currently exported file systems by looking in the /etc/exports directory.
2. Match the mount points found in step 1 with sdisk device link names (files named /dev/(r)dsk/cntndn) by looking in the /etc/fstab directory.
3. Match the sd device link names found in step 2 with SDD device link names (files named /dev/(r)dsk/vpathN) by running the showvpath command.
4. Make a backup copy of the current /etc/fstab file.
5. Edit the /etc/fstab file, replacing each instance of an sd device link named /dev/(r)dsk/cntndn with the corresponding Subsystem Device Driver device link.
6. Reboot. Verify that each exported file system passes the boot time fsck pass, that each mounts properly, and that each is exported and available to NFS clients.

If there is a problem with any exported file system after completing step 6, restore the original /etc/fstab file and reboot to restore Network File System service. Then review your steps and try again.

Oracle

Notes:

1. Procedures listed below require you to have Oracle documentation on hand.

2. You must have super-user privileges to perform these procedures.

3. These procedures were tested with Oracle 8.0.5 Enterprise server, with the 8.0.5.1 patch set from Oracle.

Installing a Oracle database for the first time

You can set up your Oracle database in one of two ways. You can set it up to use a file system or raw partitions. The procedure for installing your database differs depending on the choice you make.

If using a file system

1. If you have not already done so, install SDD using the procedure in Installing the Subsystem Device Driver.
2. Create and mount file systems on one or more SDD partitions. (Oracle recommends three mount points on different physical devices.)
3. Follow the Oracle Installation Guide for instructions on installing to a file system. (During the Oracle installation, you will be asked to name three mount points. Supply the mount points for the file systems you created on the SDD partitions.)

If using raw partitions

Notes:

1. Make sure all the databases are closed before going further.

2. Make sure that the ownership and permissions of the SDD devices are the same as the ownership and permissions of the raw devices they are replacing.

3. Do not use disk cylinder 0 (sector 0), which is the disk label. Using it corrupts the disk. For example, slice 2 on Sun is the whole disk. If you use this device without repartitioning it to start at sector 1, the disk label is corrupted.

1. If you have not already done so, install SDD using the procedure outlined in Installing the Subsystem Device Driver.
2. Create the Oracle Software Owner user in the server's local /etc/passwd file. You must also complete the following related activities:
   1. Complete the rest of the Oracle pre-installation tasks described in the Oracle8 Installation Guide. Plan to install Oracle8 on a file system residing on a SDD partition.
   2. Set up the Oracle user's ORACLE_BASE and ORACLE_ HOME environment variables to be directories of this file system.
   3. Create two more SDD-resident file systems on two other SDD volumes. Each of the resulting three mount points should have a subdirectory named oradata, to be used as a control file and redo log location for the Installer's Default Database (a sample database) as described in the Installation Guide. Oracle recommends using raw partitions for redo logs. To use SDD raw partitions as redo logs, create symbolic links from the three redo log locations to SDD raw device links (files named /dev/rdsk/vpathNs, where N is the SDD instance number, and s is the partition ID) that point to the slice.
3. Determine which SDD (vpathN') volumes you will use as Oracle8 database devices
4. Partition the selected volumes using the Solaris format utility. If SDD raw partitions are to be used by Oracle8 as database devices, be sure to leave sector 0/disk cylinder 0 of the associated volume unused. This protects UNIX disk labels from corruption by Oracle8.
5. Ensure the Oracle Software Owner has read and write privileges to the selected SDD raw partition device files under the /devices/pseudo directory.
6. Set up symbolic links in the oradata directory under the first of the three mount points created in step 2 to link the database files to SDD raw device links (files named /dev/rdsk/vpathNs) pointing to partitions of the appropriate size.
7. Install the Oracle8 Server following the instructions in the Oracle Installation Guide. Be sure to be logged in as the Oracle Software Owner when you run the orainst /m command. Select the Install New Product - Create Database Objects option. Select Raw Devices for storage type. Specify the raw device links set up in step 2 for the redo logs. Specify the raw device links set up in step 3 for the database files of the default database.
8. To set up other Oracle8 databases you must set up control files, redo logs, and database files following the guidelines in the Oracle8 Administrator's Reference. Make sure any raw devices and file systems you set up reside on SDD volumes.
9. Launch the sqlplus utility.
10. Use the create database SQL command, specifying the control, log, and system data files that you have set up.
11. Use the create tablespace SQL command to set up each of the temp, rbs, tools, and users database files that you created.
12. Use the create rollback segment SQL command to create the three redo log files that you set. For the syntax of these three create commands, see the Oracle8 Server SQL Language Reference Manual.

Installing SDD on a system that already has Oracle in place

Your installation procedure for a new SDD install will differ depending on whether you are using a file system or raw partitions for your Oracle database.

If using a file system

Follow this procedure if you are installing SDD for the first time on a system with a Oracle database that uses a file system:

1. Record the raw disk partitions being used (they are in the cntndnsn format) or the partitions where the Oracle file systems reside. You can get this information from /etc/vfstab if you know where the Oracle files are. Your database administrator can tell you where the Oracle files are, or you can check for directories with the name oradata.
2. Complete the basic installation steps in Installing the Subsystem Device Driver.
3. Change to the directory where you installed the SDD utilities. Enter the showvpath command.
4. Check the display to see whether you find a cntndn directory that is the same as the one where the Oracle files are. For example, if the Oracle files are on c1t8d0s4, look for c1t8d0s2. If you find it, you will know that /dev/dsk/vpath0c is the same as /dev/dsk/clt8d2s2. (SDD partition identifiers end in abcdefg rather than s0, s1, s2, etc.) Write this down. The output from the showvpath command looks similar to this:

   +--------------------------------------------------------------------------------+
   |vpath0c                                                                         |
   |    c1t8d0s2   /devices/pci@1f,0/pci@1/scsi@2/sd@1,0:c,raw                      |
   |    c2t8d0s2   /devices/pci@1f,0/pci@1/scsi@2,1/sd@1,0:c,raw                    |
   |                                                                                |
   +--------------------------------------------------------------------------------+

5. Use the SDD partition identifiers instead of the original Solaris identifiers when mounting the file systems.

   If you would originally have used:

   mount /dev/dsk/c1t3d2s4 /oracle/mp1
   

   You now use:

   mount /dev/dsk/vpath2e /oracle/mp1
   

   (assuming you had found vpath2c to be the SDD identifier)

Follow the instructions in Oracle Installation Guide for setting ownership and permissions.

If using raw partitions

Follow this procedure if you have Oracle8 already installed and want to reconfigure it to use SDD partitions instead of sd partitions (for example, partitions accessed through /dev/rdsk/cntndn files).

If the Oracle8 installation is accessing Veritas logical volumes, go to Veritas Volume Manager for information about installing SDD with that application.

All Oracle8 control, log, and data files are accessed either directly from mounted file systems, or through links from the oradata subdirectory of each Oracle mount point set up on the server. Therefore, the process of converting an Oracle installation from sd to SDD has two parts:

1. Changing the Oracle mount points' physical devices in /etc/fstab from sd device partition links to the SDD device partition links that access the same physical partitions.
2. Recreating any links to raw sd device links to point to raw SDD device links that access the same physical partitions.

Converting an Oracle installation from sd to SDD partitions

Perform the following steps to convert an Oracle installation from sd to SDD partitions:

1. Back up your Oracle8 database files, control files, and redo logs.
2. Obtain the sd device names for the Oracle8 mounted file systems by looking up the Oracle8 mount points in /etc/vfstab and extracting the corresponding sd device link name (for example, /dev/rdsk/c1t4d0s4).
3. Launch the sqlplus utility.
4. Type the command:

   select * from sys.dba_data_files;
   

   The output lists the locations of all data files in use by Oracle. Determine the underlying device that each data file resides on, either by looking up mounted file systems in /etc/vfstab or by extracting raw device link names directly from the select command output.

5. Run the ls -l command on each device link found in step 4 and extract the link source device file name. For example, if you type command:

   # ls -l /dev/rdsk/c1t1d0s4
   

   You might see output that is similar to this:

   +--------------------------------------------------------------------------------+
   |/dev/rdsk/c1t1d0s4 /devices/pci@1f,0/pci@1/scsi@2/sd@1,0:e                      |
   +--------------------------------------------------------------------------------+

6. Write down the file ownership and permissions by running the ls -lL command on either the files in /dev/ or /devices (it yields the same result). For example, if you type the command:

   # ls -lL /dev/rdsk/c1t1d0s4
   

   You might see output that is similar to this:

   +--------------------------------------------------------------------------------+
   |crw-r--r-- oracle dba 32,252 Nov 16 11:49 /dev/rdsk/c1t1d0s4                    |
   +--------------------------------------------------------------------------------+

7. Complete the basic installation steps in Installing the Subsystem Device Driver.
8. Match each cntndns device with its associated vpathNs device link name by running the showvpath command. Remember that vpathNs partition names use the letters [a-h] in the s position to indicate slices [0-7] in the corresponding cntndnsn slice names.
9. Run the ls -l command on each SDD device link.
10. Write down the SDD device nodes for each SDD device link by tracing back to the link source file.
11. Change the attributes of each SDD device to match the attributes of the corresponding disk device using the chgrp and chmod commands.
12. Make a copy of the existing /etc/vfstab file for recovery purposes. Edit the /etc/vfstab file, changing each Oracle device link to its corresponding SDD device link.
13. For each link found in an oradata directory, recreate the link using the appropriate SDD device link as the source file instead of the associated sd device link. As you perform this step, generate a reversing shell script that can restore all the original links in case of error.
14. Reboot the server.
15. Verify that all file system and database consistency checks complete successfully.

Veritas Volume Manager

For these procedures, you should have a copy of the Veritas Volume Manager System Administrator's Guide, and Veritas Volume Manager Command Line Interface for Solaris. These publications can be found at the following website:

www.sun.com/products-n-solutions/hardware/docs/Software/Storage_Software/VERITAS_Volume_Manager/index.html

These procedures were tested using Veritas 3.0.1. The Sun patches 105223 and 105357 must be installed with Veritas (this is a Veritas requirement).

Notes:

1. You must have super-user privileges to perform these procedures.

2. SDD does not support being used for the root (/), /var, /usr, /opt, /tmp and swap partitions.

Installing Veritas Volume Manager for the first time

Follow the instructions in this section if you are installing Veritas on the multiport subsystem's server for the first time. Installing Veritas for the first time on a SDD system consists of:

1. Installing SDD using the procedure in Installing the Subsystem Device Driver, if you have not already done so.
2. Adding a Solaris hard disk device to the Veritas root disk group (rootdg).
3. Adding a SDD device to Veritas.
4. Creating a new disk group from a SDD device.
5. Creating a new volume from a SDD device.

Adding a Solaris hard disk device to the Veritas root disk group (rootdg)

During the installation, Veritas requires that at least one disk device be added to the Veritas root disk group (rootdg). This device must be a standard Solaris hard disk device, and not a SDD device. It is important that the last disk in the rootdg be a regular disk and not a SDD device. Therefore, it is recommended that you use a different disk group for your SDD disks.

SDD disks may only be added to a Veritas disk group as a whole, for example, any previous partitioning is ignored. The c partition (the whole disk) is used, so the SDD device name for the disk in the /dev/dsk and /dev/rdsk directories would be vpath0c, for example. Veritas always looks in these directories by default, so only the device name is needed, for example, vpath0c, when issuing Veritas commands.

Partitioning of the given disk once it has been added to a Veritas disk group is achieved by dividing the Veritas disk into Veritas subdisks.

Adding a SDD device to Veritas

The following is an example of a command that adds a SDD device to Veritas:

vxdisk -f init vpath0c

After running this command, the Veritas graphical user interface tool (VMSA) can be used to create a new disk group and, a new volume from a SDD device.

Attention: VMSA and the command-line interface are the only supported methods of creating new disks or volumes with Veritas.

Creating a new disk group from a SDD device

The following command creates a new disk group from the SDD physical device. In this example, the new disk group is called ibmdg and the disk is vpath0c.

vxdg init ibmdg vpath0c

You can add a SDD device to an existing disk group using the vxdgadd command.

Creating a new volume from a SDD device

This command gets the maximum size of the disk vpath0c in blocks:

/usr/sbin/vxassist -g ibmdg -p maxsize [vpath0c]

Write down the output of the last command and use it in the next command, which creates a volume called ibmv within the disk group called ibmdg.

The command to create a volume is:

/usr/sbin/vxassist -g ibmdg make ibmv 17846272 layout=nostripe

You can change the size of the volume and use less that the maximum number of blocks.

Installing SDD on a system that already has Veritas Volume Manager in place

This section describes the Veritas command-line instructions needed to reconfigure a Veritas volume for use as a SDD disk device. This reconfiguration consists of:

• Adding SDD devices to the disk group that corresponds to the existing sd disks.
• Setting the size of a SDD device to that of the original disk.
• Setting the size of the original device to zero.

At the conclusion, you will have a disk group that contains twice the number of devices as the original disk group. The new SDD devices in the disk group will be the same size as the original sd disks. The Solaris operating system will use the SDD devices, and not the original sd disk.

Note:

Versions of Veritas that support multi-pathing (dpm) must be disabled. See Veritas Volume Manager Release Notes for instructions on doing this. Some versions of Veritas do not support the disabling of multi-pathing (dpm). In that case, you must first upgrade to a version of Veritas that supports this before proceeding. See the Veritas Volume Manager documentation for further details.

The following procedure assumes that you have:

1. Configured Veritas volumes to use Solaris disk device drivers for accessing the multiport subsystem drives.
2. Created SDD devices that refer to the same multiport subsystem drive.

These instructions allow you to replace all sd references to the original hard disks that occur in the Veritas volume's configuration with references to the SDD devices. The example provided shows the general method for replacing the sd device with the corresponding SDD device in an existing Veritas volume.

Note:

At least one device in the rootdg disk group must be a non-SDD disk; do not attempt to change all the disks in rootdg to SDD devices.

The example uses the following identifiers:

ibmv

the Veritas volume

ibmv-01

the plex associated with the ibmv volume

disk01-01

Veritas VM disk containing the original Sun hard disk device

vpath0c

the SDD device that refers to the same hard disk that disk01-01 does

c1t1d0s2

the sd disk associated with vpath0c, and disk01-01

disk02

Veritas VM disk containing the vpath0c device

rootdg

the name of the Veritas disk group to which ibmv belongs

A simplifying assumption is that the original volume, ibmv, contains exactly one subdisk. However, the method outlined here should be easy to adapt to other cases.

Before proceeding, record the multiport subsystem device links (/dev/(r)dsk/cntndnsn) being used as Veritas volume device files. Next, determine the corresponding SDD device link (/dev/(r)dsk/vpathNs) using the showvpath command. Record this information.

Reconfiguring a Veritas Volume to use a SDD disk device

1. If you have not already done so, install SDD using the procedure in Installing the Subsystem Device Driver.
2. Display information about the disk used in the volume ibmv.

   vxdisk list c1t1d0
   

   The resulting display includes information about the disk, including its public and private offset and length:

   +--------------------------------------------------------------------------------+
   |public:  slice=4 offset=0 len=17846310                                          |
   |private:  slice=3 offset=1 len=2189                                             |
   |                                                                                |
   +--------------------------------------------------------------------------------+

   From this information, calculate the parameters privlen (length of the private region) and puboffset (offset of the public region). In this case, privlen=2189, and puboffset=2190 because puboffset is one block more than the length of privlen.

3. Initialize the SDD device for use by Veritas as a simple disk, using the privlen and puboffset values from step 2.

   vxdisk -f init vpath0c puboffset=2190 privlen=2189
   

4. Add the SDD device to the disk group:

   vxdg -g rootdg adddisk disk02=vpath0c
   

5. Make sure that the file systems that are part of this volume are not mounted and then stop the volume

   umount /ibmvfs
   vxvol -g rootdg stop ibmv
   

6. Get the volume length (in sectors). This information is used in later steps. For this example, a volume length of 17846310 is assumed.

   vxprint ibmv
   

7. Disassociate the plex but do not delete it.

   vxplex -g rootdg dis ibmv-01
   vxvol -g rootdg set len=0 vol01
   

   Attention: The plex should remain to serve as backup should backing out of the SDD installation be necessary.

8. Create a subdisk from the SDD VM disk:

   vxmake -g rootdg sd disk02-01 disk02,0,17846310 
   

   (Use len from step 6)
9. Create a new plex called ibmv-02 containing the disk02-01 subdisk

   vxmake -g rootdg plex ibmv-02 sd=disk02-01
   

10. Attach the plex to the volume

    vxplex -g rootdg att ibmv ibmv-02
    vxvol set len=17846310 ibmv
    

    (Use length from step 6)
11. Make the volume active:

    vxvol -g rootdg init active ibmv
    

    Notes:

    1. When a disk is initialized for use by Veritas, it is repartitioned as a sliced disk containing a private region at slice 3 and a public region at slice 4. The length and offsets of these regions can be displayed using the vxdisk list cntndn command.

    2. When using an sd device as a SDD device, you must initialize the SDD disk as a simple disk. This simple disk uses only a single slice (slice 2). The private region starts at block 1, after the disk's VTOC region, which is situated at block 0. Note that the length of the private region varies with the type of disk used, with the public region following the private region.

At this stage you can delete the original disk, after verifying that everything is working correctly.

Solstice DiskSuite

For these procedures, you need access to the Solaris answerbook facility. These procedures were tested using Solstice DiskSuite 4.2, with the patch, 106627-04 (DiskSuite patch), installed. You should have a copy of the DiskSuite Administration Guide available to complete these procedures.

Notes:

1. You must have super-user privileges to perform these procedures.

2. SDD vpath does not support being used for the root (/), /var, /usr, /opt, /tmp and swap partitions.

Installing Solstice DiskSuite for the first time

Perform the following steps if you are installing Solstice DiskSuite on the multiport subsystem's server for the first time. The installation of Solstice DiskSuite for the first time on a SDD system consists of:

1. Installing SDD using the procedure in Installing the Subsystem Device Driver, if you have not already done so.
2. Configuring the Sparc server to recognize all devices over all paths using the boot -r command.
3. Installing the Solstice DiskSuite packages and the answerbook. Do not reboot yet.

   Note:

   Do not install the DiskSuite Tool (metatool)

4. Determine which vpath devices you will use to create Disk Suite metadevices. Partition these devices by selecting them in the Solaris format utility. The devices appear as vpathNs, where N is the vpath driver instance number). Use the partition submenu, just as you would for an sd device link of the form, cntndn. If you want to know which cntndn links correspond to a particular vpath device, type the showvpath command and press Enter. Reserve at least three partitions of three cylinders each for use as DiskSuite Replica database locations.

   Note:

   You do not need to partition any sd (cntndn) devices.

5. Set up the replica databases on a partitions of its own. This partition needs to be at least three partitions of three cylinders, and do not use a partition that includes Sector 0 for this database replica partition. Follow the instructions for setting up replica databases on vpathN's partitions, where N is the vpath device instance number and s is the letter denoting the three cylinder partition, or slice, of the device that you wish to use as a replica. Remember that partitions [a-h] of a vpath device correspond to slices [0-7] of the underlying multiport subsystem device.
6. Follow the instructions in the DiskSuite Administration Guide to build the types of metadevices you need, using the metainit command and the /dev/(r)dsk/vpathNs device link names, wherever the instructions specify /dev/(r)dsk/cntndnsn device link names.
7. Insert the setup of all vpathNs devices used by DiskSuite into the /etc/opt/SUNWmd/md.tab file

Installing SDD on a system that already has Solaris DiskSuite in place

Perform the following steps if Solaris DiskSuite is already installed:

1. Back up all data.
2. Back up the current Solstice configuration by making a copy of the /etc/opt/SUNWmd/md.tab file, and recording the output of the metastat and metadb -i commands. Make sure all sd device links in use by DiskSuite are entered in md.tab, and that they all come up properly after a reboot.
3. Installing SDD using the procedure in Installing the Subsystem Device Driver, if you have not already done so. After the installation completes, type the shutdown -i6 -y -g0 command and press Enter. This verifies the vpath installation.

   Note:

   Do not do a reconfiguration reboot

4. Using a plain sheet of paper, make a two-column list matching up the /dev/(r)dsk/cntndnsn device links found in step 2 with the corresponding /dev/(r)dsk/vpathNs device links using the showvpath command.
5. Delete each replica database currently configured with an /dev/(r)dsk/cntndnsn device, by using the metadb -d -f <device> command, and replace it with the corresponding /dev/(r)dsk/vpathNs device found in step 2, by using the metadb -a <device> command.
6. Create a new md.tab file, inserting the corresponding vpathNs device link name in place of each cntndnsn device link name. Do not do this for boot device partitions (vpath does not currently support these). When you are confident that the new file is correct, install it in the /etc/opt/SUNWmd directory.
7. Reboot the server or proceed to the next step, if you wish to avoid rebooting your system.
8. Stop all applications using DiskSuite, including file systems.
9. Enter the following commands for each existing metadevice:

   metaclear <device>
   metainit -a
   

10. Restart your applications.

Note:

To back out vpath in case of any problems following step 7, reverse the procedures in step 6, reinstall the original md.tab into /etc/opt/SUNWmd, and run the command pkgrm IBMdpo, and reboot.

Setting up UFS logging on a new system

For these procedures, you need access to the Solaris answerbook facility.

Notes:

1. You must have super-user privileges to perform these procedures.

Perform the following steps if you are installing a new UFS logging file system on vpath devices:

1. Installing SDD using the procedure in Installing the Subsystem Device Driver, if you have not already done so.
2. Determine which vpath (vpathNs) volumes you will use as file system devices. Partition the selected vpath volumes using the Solaris format utility. Be sure to create partitions for UFS logging devices as well as for UFS master devices.
3. Create file systems on the selected vpath UFS master device partitions using the newfs command.
4. Install Solstice DiskSuite if you have not already done so.
5. Create the metatrans device using metainit. For example, assume /dev/dsk/vpath0d is your UFS master device used in step 3, /dev/dsk/vpath0e is its corresponding log device, and d0 is the trans device you want to create for UFS logging. Type metainit d0 -t vpath0d vpath0e and press Enter.
6. Create mount points for each UFS logging file system you have created using steps 3 and 5.
7. Install the file systems into the /etc/vfstab directory, specifying /dev/md/(r)dsk/d <metadevice number> for the raw and block devices. Be sure to set the mount at boot field to yes.
8. Reboot.

Installing vpath on a System that already has UFS Logging in Place

Perform the following steps if you have UFS logging file systems already residing on a multiport subsystem and you wish to use vpath partitions instead of sd partitions to access them.

1. Make a list of the DiskSuite metatrans devices for all existing UFS logging file systems by looking in the /etc/vfstab directory. Make sure that all configured metatrans devices are correctly set up in the /etc/opt/SUNWmd/md.tab file. If the devices are not set up now, set them up before continuing. Save a copy of md tab.
2. Match the device names found in step 1 with sd device link names (files named /dev/(r)dsk/cntndnsn) through the metastat command.
3. Install SDD using the procedure in Installing the Subsystem Device Driver, if you have not already done so.
4. Match the sd device link names found in step 2 with vpath device link names (files named /dev/(r)dsk/vpathNs) by executing the /opt/IBMdpo/bin/showvpath command.
5. Unmount all current UFS logging file systems known to reside on the multiport subsystem through the umount command.
6. Type metaclear -a and press Enter.
7. Create new metatrans devices from the vpathNs partitions found in step 4 corresponding to the sd device links found in step 2. Remember that vpath partitions [a-h] correspond to sd slices [0-7]. Use the metainit d <metadevice number> -t <"vpathNs" - master device> <"vpathNs" - logging device> command . Be sure to use the same metadevice numbering as was originally used with the sd partitions. Edit the /etc/opt/SUNWmd/md.tab file to change each metatrans device entry to use vpathNs devices.
8. Reboot.

Note:

If there is a problem with a metatrans device after steps 7 and 8, restore the original /etc/opt/SUNWmd/md.tab file and reboot. Review your steps and try again.

Create:
metadb -a -c 3 -f vpath0f # add database replicas
metainit d0 1 1 vpath0e # add metadevice
 
Info
metastat
metadb -i
 
Delete:
metaclear d0 # delete metadevice
metadb -d -f vpat

Uninstalling the Subsystem Device Driver

Note:

You must uninstall the current level of SDD before upgrading to a newer level.

Attention: Do not reboot between the uninstall and the reinstall of SDD.

Upgrading SDD consists of uninstalling and reinstalling the IBMdpo package. Perform the following steps to uninstall SDD:

1. Reboot or umount all SDD file systems.
2. If you are using SDD with a database, such as Oracle, edit the appropriate database configuration files (database partition) to remove all the SDD devices.
3. If you are using a database, restart the database.
4. Type # pkgrm IBMdpo and press Enter.

   Attention: A number of different installed packages is displayed. Make sure you specify the correct package to uninstall.

   A message similar to the following is displayed:

   +--------------------------------------------------------------------------------+
   |The following packages are available:                                           |
   |1 IBMcli ibm2105cli                                                             |
   |         (sparc) 1.1.0.0                                                        |
   |2 IBMdpo IBM DPO driver Version: May-10-2000 16:51                              |
   |         (sparc) 1                                                              |
   |                                                                                |
   +--------------------------------------------------------------------------------+

5. Type Y and press Enter. A message similar to the following is displayed:

   +--------------------------------------------------------------------------------+
   |## Removing installed package instance <IBMdpo>                                 |
   |                                                                                |
   |This package contains scripts that 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            |
   |                                                                                |
   +--------------------------------------------------------------------------------+

6. Type Y and press Enter. A message similar to the following is displayed:

   +--------------------------------------------------------------------------------+
   |## Verifying package dependencies.                                              |
   |## Processing package information.                                              |
   |## Executing preremove script.                                                  |
   |Device busy                                                                     |
   |Cannot unload module: vpathdd                                                   |
   |Will be unloaded upon reboot.                                                   |
   |## Removing pathnames in class <none>                                           |
   |/usr/sbin/vpathmkdev                                                            |
   |/opt/IBMdpo                                                                     |
   |/kernel/drv/vpathdd.conf                                                        |
   |/kernel/drv/vpathdd                                                             |
   |/etc/rcS.d/S20vpath-config                                                      |
   |/etc/rc2.d/S00vpath-config                                                      |
   |/etc/defvpath                                                                   |
   |## Updating system information.                                                 |
   |                                                                                |
   |Removal of <IBMdpo> was successful.                                             |
   |                                                                                |
   +--------------------------------------------------------------------------------+

   Attention: Do not reboot at this time.

   Note:

   When SDD has been successfully uninstalled, the first part of the procedure for upgrading the SDD is complete. To complete the upgrade, you now need to reinstall SDD. See Installing the Subsystem Device Driver for detailed procedures.

Changing a SDD hardware configuration

When adding or removing multiport SCSI devices from your system, you must reconfigure SDD to recognize the new devices. Perform the following steps to reconfigure SDD:

1. Shut down the system. Type shutdown -i0 -g0 -y and press Enter.
2. Do a configuration reboot. From the OK prompt, type boot -r and press Enter. This uses the current SDD entries during reboot, not the new entries. The reboot forces the new disks to be recognized.
3. After the reboot, run the SDD configuration utility to make the changes to the directory /opt/IBMdpo/bin. Type cfgvpath -c and press Enter.
4. Shut down the system. Type shutdown -i6 -g0 -y and press Enter.
5. After the reboot, change to the /opt/IBMdpo/bin directory.

   cd /opt/IBMdpo/bin
   

6. Type drvconfig and press Enter. This reconfigures all the drives.
7. Type vpathmkdev and press Enter. This creates all the vpath devices.

Chapter 8. Using the datapath commands

SDD provides commands that you can use to display the status of adapters that are used to access managed devices, or to display the status of devices that the device driver manages. You can also set individual path conditions either to online or offline, or you can set all paths that are connected to an adapter or bus either to online or offline. This chapter includes descriptions of these commands. Table 23 provides an alphabetical list of these commands, a brief description, and where to go in this chapter for more information.

Table 23. Commands

Command	Description	Page
datapath query adapter	Displays information about adapters	datapath query adapter command
datapath query adaptstats	Displays performance information for all SCSI and FCS adapters that are attached to SDD devices	datapath query adaptstats command
datapath query device	Displays information about devices	datapath query device command
datapath query devstats	Displays performance information for a single SDD device or all SDD devices	datapath query devstats command
datapath set adapter	Sets all device paths that are attached to an adapter to online or offline	datapath set adapter command
datapath set device	Sets the path of a device to online or offline	datapath set device command

datapath query adapter command

The datapath query adapter command displays information about a single adapter or all adapters.

Syntax

>>-datapath query adapter-adapter number-----------------------><
 
 

Parameters

adapter number

The adapter number for which you want information displayed. If you do not enter an adapter number, information about all adapters is displayed.

Examples

If you enter the following command, datapath query adapter, the following output is displayed:

+--------------------------------------------------------------------------------+
|Active Adapters :4                                                              |
|                                                                                |
|Adpt#     Adapter Name   State     Mode     Select     Errors  Paths  Active    |
|    0            scsi3  NORMAL   ACTIVE  129062051          0     64       0    |
|    1            scsi2  NORMAL   ACTIVE   88765386        303     64       0    |
|    2           fscsi2  NORMAL   ACTIVE  407075697       5427   1024       0    |
|    3           fscsi0  NORMAL   ACTIVE  341204788      63835    256       0    |
+--------------------------------------------------------------------------------+

The terms used in the output are defined as follows:

Adpt #

The number of the adapter.

Adapter Name

The name of the adapter.

State

The condition of the named adapter. It can be either:

Normal

Adapter is in use.

Degraded

One or more paths are not functioning.

Failed

The adapter is no longer being used by SDD .

Mode

The mode of the named adapter, which is either Active or Offline.

Select

The number of times this adapter was selected for input or output.

Errors

The number of errors on all paths that are attached to this adapter.

Paths

The number of paths that are attached to this adapter.

Note:

In the Windows NT host system, this is the number of physical and logical devices that are attached to this adapter.

Active

The number of functional paths that are attached to this adapter. The number of functional paths is equal to the number of paths minus any that are identified as failed or offline.

datapath query adaptstats command

The datapath query adaptstats command displays performance information for all SCSI and FCS adapters that are attached to SDD devices. If you do not enter a device number, information about all devices is displayed.

Syntax

>>-datapath query adaptstats-adapter number--------------------><
 
 

Parameters

adapter number

The adapter number for which you want information displayed. If you do not enter an adapter number, information about all adapters is displayed.

Examples

If you enter the following command, datapath query adaptstats 0, the following output is displayed:

Adapter #:  0
=============
                Total Read  Total Write  Active Read  Active Write   Maximum
I/O:                  1442     41295166            0             2        75
SECTOR:             156209    750217654            0            32      2098
 
/*-------------------------------------------------------------------------*/                                                       
 

The terms used in the output are defined as follows:

Total Read

• I/O: total number of completed Read requests
• SECTOR: total number of sectors that have been read

Total Write

• I/O: total number of completed Write requests
• SECTOR: total number of sectors that have been written

Active Read

• I/O: total number of Read requests in process
• SECTOR: total number of sectors to read in process

Active Write

• I/O: total number of Write requests in process
• SECTOR: total number of sectors to write in process

Maximum

• I/O: the maximum number of queued I/O requests
• SECTOR: the maximum number of queued sectors to Read/Write

datapath query device command

The datapath query device command displays information about a single device or all devices. If you do not enter a device number, information about all devices is displayed.

Syntax

>>-datapath query device-device number-------------------------><
 
 

Parameters

device number

The device number refers to the device index number, rather than the SDD device number.

Examples

If you enter the following command, datapath query device 35, the output is displayed as follows:

DEV#:  35  DEVICE NAME: vpath0  TYPE: 2105E20   SERIAL: 60012028
================================================================
Path#      Adapter/Hard Disk   State      Mode    Select        Errors
 0             scsi6/hdisk58    OPEN    NORMAL   7861147             0
 1             scsi5/hdisk36    OPEN    NORMAL   7762671             0                                                       
 

Note:

Usually, the device number and the device index number are the same. However, if the devices are configured out of order, the two numbers are not always consistent. To find the corresponding index number for a specific device, you should always run the datapath query device command first.

The terms used in the output are defined as follows:

Dev#

The number of this device.

Name

The name of this device.

Type

The device product ID from inquiry data.

Serial

The logical unit number (LUN) for this device.

Path

The path number.

Adapter

The name of the adapter that the path is attached to.

Hard Disk

The name of the logical device that the path is bound to.

State

The condition of the named device:

Open

Path is in use.

Close

Path is not being used.

Dead

Path is no longer being used. It was either removed by SDD due to errors or manually removed using the datapath set device M path N offline or datapath set adapter N offline command.

Invalid

Path verification failed. The path was not opened.

Mode

The mode of the named device. It is either Normal or Offline.

Select

The number of times this path was selected for input or output.

Errors

The number of errors on a path that is attached to this device.

datapath query devstats command

The datapath query devstats command displays performance information for a single SDD device or all SDD devices. If you do not enter a device number, information about all devices is displayed.

Syntax

>>-datapath query devstats-device number-----------------------><
 
 

Parameters

device number

The device number refers to the device index number, rather than the SDD device number.

Examples

If you enter the following command, datapath query devstats 0, the following output is displayed:

Device #:   0
=============
                Total Read  Total Write  Active Read  Active Write   Maximum
I/O:                   387     24502563            0             0        62
SECTOR:               9738    448308668            0             0      2098
 
Transfer Size:      <= 512        <= 4k       <= 16K        <= 64K     > 64K
                   4355850      1024164     19121140          1665       130
 
/*-------------------------------------------------------------------------*/                                                       
 

The terms used in the output are defined as follows:

Total Read

• I/O: total number of completed Read requests
• SECTOR: total number of sectors that have been read

Total Write

• I/O: total number of completed Write requests
• SECTOR: total number of sectors that have been written

Active Read

• I/O: total number of Read requests in process
• SECTOR: total number of sectors to read in process

Active Write

• I/O: total number of Write requests in process
• SECTOR: total number of sectors to write in process

Maximum

• I/O: the maximum number of queued I/O requests
• SECTOR: the maximum number of queued sectors to Read/Write

Transfer size

• <= 512: the number of I/O requests received, whose transfer size is 512 bytes or less
• <= 4k: the number of I/O requests received, whose transfer size is 4 KB or less (where KB equals 1024 bytes)
• <= 16K: the number of I/O requests received, whose transfer size is 16 KB or less (where KB equals 1024 bytes)
• <= 64K: the number of I/O requests received, whose transfer size is 64 KB or less (where KB equals 1024 bytes)
• > 64K: the number of I/O requests received, whose transfer size is greater than 64 KB (where KB equals 1024 bytes)

datapath set adapter command

The datapath set adapter command sets all device paths attached to an adapter either to online or offline.

Notes:

1. This command will not remove the last path to a device.

2. The datapath set adapter offline command fails if there is any device having the last path attached to this adapter.

3. This command can be issued even when the device are closed.

4. If all paths are attached to a single fibre-channel adapter, that connects to multiple ESS ports through a switch, the set adapter 0 offline command fails; all the paths are not set offline.

Syntax

>>-datapath set adapter-adapter number-+- online--+------------><
                                       '- offline-'
 
 

Parameters

adapter number

The adapter number that you want to change.

online

Sets the adapter online.

offline

Sets the adapter offline.

Examples

If you enter the following command, datapath set adapter 0 offline, adapter 0 changes to Offline mode and its state changes to failed; while all paths attached to adapter 0 change to Offline mode and their states change to Dead, if they were in the Open state.

datapath set device command

The datapath set device command sets the path of a device either to online or offline.

Notes:

1. You cannot remove the last path to a device from service. This prevents a data access failure from occurring.

2. This command can be issued even when the device is closed.

Syntax

>>-datapath set device-device number path number-+- online--+--><
                                                 '- offline-'
 
 

Parameters

device number

The device index number that you want to change.

path number

The path number that you want to change.

online

Sets the path online.

offline

Removes the path from service.

Examples

If you enter the following command, datapath set device 0 path 0 offline, path 0 for device 0 changes to Offline mode.

Statement of Limited Warranty

Part 1 - General Terms

International Business Machines Corporation

Armonk, New York, 10504

This Statement of Limited Warranty includes Part 1 - General Terms and Part 2 - Country or region-unique Terms. The terms of Part 2 may replace or modify those of Part 1. The warranties provided by IBM in this Statement of Limited Warranty apply only to Machines you purchase for your use, and not for resale, from IBM or your reseller. The term "Machine" means an IBM machine, its features, conversions, upgrades, elements, or accessories, or any combination of them. The term "Machine" does not include any software programs, whether pre-loaded with the Machine, installed subsequently or otherwise. Unless IBM specifies otherwise, the following warranties apply only in the country or region where you acquire the Machine. Nothing in this Statement of Warranty affects any statutory rights of consumers that cannot be waived or limited by contract. If you have any questions, contact IBM or your reseller.

Unless IBM specifies otherwise, the following warranties apply only in the country or region where you acquire the Machine. If you have any questions, contact IBM or your reseller.

Machine: IBM 2105 (Models E10, E20, F10, and F20) Enterprise Storage Server (ESS)

Warranty Period: Three Years *

*Contact your place of purchase for warranty service information. Some IBM Machines are eligible for On-site warranty service depending on the country or region where service is performed.

The IBM Warranty for Machines

IBM warrants that each Machine 1) is free from defects in materials and workmanship and 2) conforms to IBM's Official Published Specifications ("Specifications"). The warranty period for a Machine is a specified, fixed period commencing on its Date of Installation. The date on your sales receipt is the Date of Installation, unless IBM or your reseller informs you otherwise.

During the warranty period IBM or your reseller, if approved by IBM to provide warranty service, will provide repair and exchange service for the Machine, without charge, under the type of service designated for the Machine and will manage and install engineering changes that apply to the Machine.

If a Machine does not function as warranted during the warranty period, and IBM or your reseller are unable to either 1) make it do so or 2) replace it with one that is at least functionally equivalent, you may return it to your place of purchase and your money will be refunded. The replacement may not be new, but will be in good working order.

Extent of Warranty

The warranty does not cover the repair or exchange of a Machine resulting from misuse, accident, modification, unsuitable physical or operating environment, improper maintenance by you, or failure caused by a product for which IBM is not responsible. The warranty is voided by removal or alteration of Machine or parts identification labels.

THESE WARRANTIES ARE YOUR EXCLUSIVE WARRANTIES AND REPLACE ALL OTHER WARRANTIES OR CONDITIONS, EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OR CONDITIONS OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THESE WARRANTIES GIVE YOU SPECIFIC LEGAL RIGHTS AND YOU MAY ALSO HAVE OTHER RIGHTS WHICH VARY FROM JURISDICTION TO JURISDICTION. SOME JURISDICTIONS DO NOT ALLOW THE EXCLUSION OR LIMITATION OF EXPRESS OR IMPLIED WARRANTIES, SO THE ABOVE EXCLUSION OR LIMITATION MAY NOT APPLY TO YOU. IN THAT EVENT, SUCH WARRANTIES ARE LIMITED IN DURATION TO THE WARRANTY PERIOD. NO WARRANTIES APPLY AFTER THAT PERIOD.

Items Not Covered by Warranty

IBM does not warrant uninterrupted or error-free operation of a Machine.

Unless specified otherwise, IBM provides non-IBM machines WITHOUT WARRANTIES OF ANY KIND.

Any technical or other support provided for a Machine under warranty, such as assistance via telephone with "how-to" questions and those regarding Machine setup and installation, will be provided WITHOUT WARRANTIES OF ANY KIND.

Warranty Service

To obtain warranty service for the Machine, contact your reseller or IBM. In the United States, call IBM at 1-800-IBM-SERV (426-7378). In Canada, call IBM at 1-800-465-6666 . You may be required to present proof of purchase.

IBM or your reseller provides certain types of repair and exchange service, either at your location or at a service center, to keep Machines in, or restore them to, conformance with their Specifications. IBM or your reseller will inform you of the available types of service for a Machine based on its country or region of installation. IBM may repair the failing Machine or exchange it at its discretion.

When warranty service involves the exchange of a Machine or part, the item IBM or your reseller replaces becomes its property and the replacement becomes yours. You represent that all removed items are genuine and unaltered. The replacement may not be new, but will be in good working order and at least functionally equivalent to the item replaced. The replacement assumes the warranty service status of the replaced item.

Any feature, conversion, or upgrade IBM or your reseller services must be installed on a Machine which is 1) for certain Machines, the designated, serial-numbered Machine and 2) at an engineering-change level compatible with the feature, conversion, or upgrade. Many features, conversions, or upgrades involve the removal of parts and their return to IBM. A part that replaces a removed part will assume the warranty service status of the removed part.

Before IBM or your reseller exchanges a Machine or part, you agree to remove all features, parts, options, alterations, and attachments not under warranty service.

You also agree to

1. ensure that the Machine is free of any legal obligations or restrictions that prevent its exchange;.
2. obtain authorization from the owner to have IBM or your reseller service a Machine that you do not own; and
3. where applicable, before service is provided
   1. follow the problem determination, problem analysis, and service request procedures that IBM or your reseller provides,
   2. secure all programs, data, and funds contained in a Machine,
   3. provide IBM or your reseller with sufficient, free, and safe access to your facilities to permit them to fulfill their obligations, and
   4. inform IBM or your reseller of changes in a Machine's location.

IBM is responsible for loss of, or damage to, your Machine while it is 1) in IBM's possession or 2) in transit in those cases where IBM is responsible for the transportation charges.

Neither IBM nor your reseller is responsible for any of your confidential, proprietary or personal information contained in a Machine which you return to IBM or your reseller for any reason. You should remove all such information from the Machine prior to its return.

Production Status

Each IBM Machine is manufactured from new parts, or new and used parts. In some cases, the Machine may not be new and may have been previously installed. Regardless of the Machine's production status, IBM's appropriate warranty terms apply.

Limitation of Liability

Circumstances may arise where, because of a default on IBM's part or other liability, you are entitled to recover damages from IBM. In each such instance, regardless of the basis on which you are entitled to claim damages from IBM (including fundamental breach, negligence, misrepresentation, or other contract or tort claim), IBM is liable for no more than

1. damages for bodily injury (including death) and damage to real property and tangible personal property; and
2. the amount of any other actual direct damages, up to the greater of U.S. $100,000 (or equivalent in local currency) or the charges (if recurring, 12 months' charges apply) for the Machine that is the subject of the claim.

   This limit also applies to IBM's suppliers and your reseller. It is the maximum for which IBM, its suppliers, and your reseller are collectively responsible.

UNDER NO CIRCUMSTANCES IS IBM LIABLE FOR ANY OF THE FOLLOWING: 1) THIRD-PARTY CLAIMS AGAINST YOU FOR DAMAGES (OTHER THAN THOSE UNDER THE FIRST ITEM LISTED ABOVE); 2) LOSS OF, OR DAMAGE TO, YOUR RECORDS OR DATA; OR 3) SPECIAL, INCIDENTAL, OR INDIRECT DAMAGES OR FOR ANY ECONOMIC CONSEQUENTIAL DAMAGES (INCLUDING LOST PROFITS OR SAVINGS), EVEN IF IBM, ITS SUPPLIERS OR YOUR RESELLER IS INFORMED OF THEIR POSSIBILITY. SOME JURISDICTIONS DO NOT ALLOW THE EXCLUSION OR LIMITATION OF INCIDENTAL OR CONSEQUENTIAL DAMAGES, SO THE ABOVE LIMITATION OR EXCLUSION MAY NOT APPLY TO YOU.

Part 2 - Country or region-unique Terms

ASIA PACIFIC

AUSTRALIA: The IBM Warranty for Machines: The following paragraph is added to this Section: The warranties specified in this Section are in addition to any rights you may have under the Trade Practices Act 1974 or other legislation and are only limited to the extent permitted by the applicable legislation.

Extent of Warranty: The following replaces the first and second sentences of this Section: The warranty does not cover the repair or exchange of a Machine resulting from misuse, accident, modification, unsuitable physical or operating environment, operation in other than the Specified Operating Environment, improper maintenance by you, or failure caused by a product for which IBM is not responsible.

Limitation of Liability: The following is added to this Section: Where IBM is in breach of a condition or warranty implied by the Trade Practices Act 1974, IBM's liability is limited to the repair or replacement of the goods or the supply of equivalent goods. Where that condition or warranty relates to right to sell, quiet possession or clear title, or the goods are of a kind ordinarily acquired for personal, domestic or household use or consumption, then none of the limitations in this paragraph apply.

PEOPLE'S REPUBLIC OF CHINA: Governing Law: The following is added to this Statement: The laws of the State of New York govern this Statement.

INDIA: Limitation of Liability: The following replaces items 1 and 2 of this Section: 1. liability for bodily injury (including death) or damage to real property and tangible personal property will be limited to that caused by IBM's negligence; 2. as to any other actual damage arising in any situation involving nonperformance by IBM pursuant to, or in any way related to the subject of this Statement of Limited Warranty, IBM's liability will be limited to the charge paid by you for the individual Machine that is the subject of the claim.

NEW ZEALAND: The IBM Warranty for Machines: The following paragraph is added to this Section: The warranties specified in this Section are in addition to any rights you may have under the Consumer Guarantees Act 1993 or other legislation which cannot be excluded or limited. The Consumer Guarantees Act 1993 will not apply in respect of any goods which IBM provides, if you require the goods for the purposes of a business as defined in that Act.

Limitation of Liability: The following is added to this Section: Where Machines are not acquired for the purposes of a business as defined in the Consumer Guarantees Act 1993, the limitations in this Section are subject to the limitations in that Act.

EUROPE, MIDDLE EAST, AFRICA (EMEA)

The following terms apply to all EMEA countries or regions.

The terms of this Statement of Limited Warranty apply to Machines purchased from an IBM reseller. If you purchased this Machine from IBM, the terms and conditions of the applicable IBM agreement prevail over this warranty statement.

Warranty Service

If you purchased an IBM Machine in Austria, Belgium, Denmark, Estonia, Finland, France, Germany, Greece, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland or United Kingdom, you may obtain warranty service for that Machine in any of those countries or regions from either (1) an IBM reseller approved to perform warranty service or (2) from IBM.

If you purchased an IBM Personal Computer Machine in Albania, Armenia, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Georgia, Hungary, Kazakhstan, Kirghizia, Federal Republic of Yugoslavia, Former Yugoslav Republic of Macedonia (FYROM), Moldova, Poland, Romania, Russia, Slovak Republic, Slovenia, or Ukraine, you may obtain warranty service for that Machine in any of those countries or regions from either (1) an IBM reseller approved to perform warranty service or (2) from IBM.

The applicable laws, Country or region-unique terms and competent court for this Statement are those of the country or region in which the warranty service is being provided. However, the laws of Austria govern this Statement if the warranty service is provided in Albania, Armenia, Belarus, Bosnia and Herzegovina, Bulgaria, Croatia, Czech Republic, Federal Republic of Yugoslavia, Georgia, Hungary, Kazakhstan, Kirghizia, Former Yugoslav Republic of Macedonia (FYROM), Moldova, Poland, Romania, Russia, Slovak Republic, Slovenia, and Ukraine.

The following terms apply to the country or region specified:

EGYPT: Limitation of Liability: The following replaces item 2 in this Section: 2. as to any other actual direct damages, IBM's liability will be limited to the total amount you paid for the Machine that is the subject of the claim.

Applicability of suppliers and resellers (unchanged).

FRANCE: Limitation of Liability: The following replaces the second sentence of the first paragraph of this Section:

In such instances, regardless of the basis on which you are entitled to claim damages from IBM, IBM is liable for no more than: (items 1 and 2 unchanged).

GERMANY: The IBM Warranty for Machines: The following replaces the first sentence of the first paragraph of this Section:

The warranty for an IBM Machine covers the functionality of the Machine for its normal use and the Machine's conformity to its Specifications.

The following paragraphs are added to this Section:

The minimum warranty period for Machines is six months.

In case IBM or your reseller are unable to repair an IBM Machine, you can alternatively ask for a partial refund as far as justified by the reduced value of the unrepaired Machine or ask for a cancellation of the respective agreement for such Machine and get your money refunded.

Extent of Warranty: The second paragraph does not apply.

Warranty Service: The following is added to this Section: During the warranty period, transportation for delivery of the failing Machine to IBM will be at IBM's expense.

Production Status: The following paragraph replaces this Section: Each Machine is newly manufactured. It may incorporate in addition to new parts, reused parts as well.

Limitation of Liability: The following is added to this Section:

The limitations and exclusions specified in the Statement of Limited Warranty will not apply to damages caused by IBM with fraud or gross negligence and for express warranty.

In item 2, replace "U.S. $100,000" with "1,000,000 DM."

The following sentence is added to the end of the first paragraph of item 2:

IBM's liability under this item is limited to the violation of essential contractual terms in cases of ordinary negligence.

IRELAND: Extent of Warranty: The following is added to this Section:

Except as expressly provided in these terms and conditions, all statutory conditions, including all warranties implied, but without prejudice to the generality of the foregoing all warranties implied by the Sale of Goods Act 1893 or the Sale of Goods and Supply of Services Act 1980 are hereby excluded.

Limitation of Liability: The following replaces items one and two of the first paragraph of this Section:

1. death or personal injury or physical damage to your real property solely caused by IBM's negligence; and 2. the amount of any other actual direct damages, up to the greater of Irish Pounds 75,000 or 125 percent of the charges (if recurring, the 12 months' charges apply) for the Machine that is the subject of the claim or which otherwise gives rise to the claim.

Applicability of suppliers and resellers (unchanged).

The following paragraph is added at the end of this Section:

IBM's entire liability and your sole remedy, whether in contract or in tort, in respect of any default shall be limited to damages.

ITALY: Limitation of Liability: The following replaces the second sentence in the first paragraph:

In each such instance unless otherwise provided by mandatory law, IBM is liable for no more than: (item 1 unchanged) 2) as to any other actual damage arising in all situations involving nonperformance by IBM pursuant to, or in any way related to the subject matter of this Statement of Warranty, IBM's liability, will be limited to the total amount you paid for the Machine that is the subject of the claim.

Applicability of suppliers and resellers (unchanged).

The following replaces the second paragraph of this Section:

Unless otherwise provided by mandatory law, IBM and your reseller are not liable for any of the following: (items 1 and 2 unchanged) 3) indirect damages, even if IBM or your reseller is informed of their possibility.

SOUTH AFRICA, NAMIBIA, BOTSWANA, LESOTHO AND SWAZILAND: Limitation of Liability: The following is added to this Section:

IBM's entire liability to you for actual damages arising in all situations involving nonperformance by IBM in respect of the subject matter of this Statement of Warranty will be limited to the charge paid by you for the individual Machine that is the subject of your claim from IBM.

TURKIYE: Production Status: The following replaces this Section:

IBM fulfills customer orders for IBM Machines as newly manufactured in accordance with IBM's production standards.

UNITED KINGDOM: Limitation of Liability: The following replaces items 1 and 2 of the first paragraph of this Section:

1. death or personal injury or physical damage to your real property solely caused by IBM's negligence;

2. the amount of any other actual direct damages or loss, up to the greater of Pounds Sterling 150,000 or 125 percent of the charges (if recurring, the 12 months' charges apply) for the Machine that is the subject of the claim or which otherwise gives rise to the claim;

The following item is added to this paragraph:

3. breach of IBM's obligations implied by Section 12 of the Sale of Goods Act 1979 or Section 2 of the Supply of Goods and Services Act 1982.

Applicability of suppliers and resellers (unchanged).

The following is added to the end of this Section:

IBM's entire liability and your sole remedy, whether in contract or in tort, in respect of any default will be limited to damages.

Notices

This information was developed for products and services offered in the U.S.A.

IBM may not offer the products, services, or features discussed in this document in other countries. Consult your local IBM representative for information on the products and services currently available in your area. Any reference to an IBM product, program, or service is not intended to state or imply that only that IBM product, program, or service may be used. Any functionally equivalent product, program, or service that does not infringe any IBM intellectual property right may be used instead. However, it is the user's responsibility to evaluate and verify the operation of any non-IBM product, program, or service.

IBM may have patents or pending patent applications covering subject matter described 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:
IBM Director of Licensing
IBM Corporation
North Castle Drive
Armonk, NY 10504-1785
U.S.A

The following paragraph does not apply to the United Kingdom or any other country where such provisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THIS PUBLICATIONS "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express or implied warranties in certain transactions, therefore, this statement may not apply to you.

This information could include technical inaccuracies or typographical errors. Changes are periodically made to the information herein; these changes will be incorporated in new editions of the publications. IBM may make improvements and/or changes in the product(s) and/or program(s) described in this publication at any time without notice.

IBM may use or distribute any of the information you supply in any way it believes appropriate without incurring any obligation to you.

Information concerning non-IBM products was obtained from the suppliers of those products, their published announcements or other publicly available sources. IBM has not tested those products and cannot confirm the accuracy of performance, compatibility or any other claims related to non-IBM products. Questions on the capabilities of non-IBM products should be addressed to the suppliers of those products.

Trademarks

The following terms are trademarks of the International Business Machines Corporation in the United States, other countries, or both:

AIX

AS/400

DFSMS/MVS

ES/9000

ESCON

FICON

FlashCopy

HACMP/6000

IBM

Enterprise Storage Server

IBM TotalStorage

eServer

MVS/ESA

Netfinity

NetVista

NUMA-Q

Operating System/400

OS/390

OS/400

RS/6000

S/390

Seascape

SNAPSHOT

SP

StorWatch

System/360

System/370

System/390

TotalStorage

Versatile Storage Server

VM/ESA

VSE/ESA

Microsoft and Windows NT are trademarks of Microsoft Corporation in the United States, other countries, or both.

Java and all Java-based trademarks are trademarks of Sun Microsystems, Inc. in the United States, other countries, or both.

UNIX is a registered trademark of The Open Group in the United States and other countries.

Other company, product, and service names may be trademarks or service marks of others.

Electronic emission notices

This section contains the electronic emission notices or statements for the United States and other countries.

Federal Communications Commission (FCC) statement

This equipment has been tested and complies with the limits for a Class A digital device, pursuant to Part 15 of the FCC Rules. These limits are designed to provide reasonable protection against harmful interference when the equipment is operated in a commercial environment. This equipment generates, uses, and can radiate radio frequency energy and, if not installed and used in accordance with the instruction manual, might cause harmful interference to radio communications. Operation of this equipment in a residential area is likely to cause harmful interference, in which case the user will be required to correct the interference at his own expense.

Properly shielded and grounded cables and connectors must be used to meet FCC emission limits. IBM is not responsible for any radio or television interference caused by using other than recommended cables and connectors, or by unauthorized changes or modifications to this equipment. Unauthorized changes or modifications could void the users authority to operate the equipment.

This device complies with Part 15 of the FCC Rules. Operation is subject to the following two conditions: (1) this device might not cause harmful interference, and (2) this device must accept any interference received, including interference that might cause undesired operation.

Industry Canada compliance statement

This Class A digital apparatus complies with Canadian ICES-003.

Cet appareil numérique de la classe A est conform à la norme NMB-003 du Canada.

European community compliance statement

This product is in conformity with the protection requirements of EC Council Directive 89/336/EEC on the approximation of the laws of the Member States relating to electromagnetic compatibility. IBM cannot accept responsibility for any failure to satisfy the protection requirements resulting from a nonrecommended modification of the product, including the fitting of non-IBM option cards.

Germany only

Zulassungsbescheinigung laut Gesetz ueber die elektromagnetische Vertraeglichkeit von Geraeten (EMVG) vom 30. August 1995.

Dieses Geraet ist berechtigt, in Uebereinstimmung mit dem deutschen EMVG das EG-Konformitaetszeichen - CE - zu fuehren.

Der Aussteller der Konformitaetserklaeung ist die IBM Deutschland.

Informationen in Hinsicht EMVG Paragraph 3 Abs. (2) 2:

  Das Geraet erfuellt die Schutzanforderungen nach EN 50082-1 und
  EN 55022 Klasse A.

EN 55022 Klasse A Geraete beduerfen folgender Hinweise:

Nach dem EMVG:

"Geraete duerfen an Orten, fuer die sie nicht ausreichend entstoert
sind, nur mit besonderer Genehmigung des Bundesministeriums
fuer Post und Telekommunikation oder des Bundesamtes fuer Post und
Telekommunikation
betrieben werden. Die Genehmigung wird erteilt, wenn keine
elektromagnetischen Stoerungen zu erwarten sind." (Auszug aus dem
EMVG, Paragraph 3, Abs.4)
 
Dieses Genehmigungsverfahren ist nach Paragraph 9 EMVG in Verbindung
mit der entsprechenden Kostenverordnung (Amtsblatt 14/93)
kostenpflichtig.

Nach der EN 55022:

"Dies ist eine Einrichtung der Klasse A. Diese Einrichtung kann im
Wohnbereich Funkstoerungen verursachen; in diesem Fall kann vom
Betreiber verlangt werden, angemessene Massnahmen durchzufuehren
und dafuer aufzukommen."

Anmerkung:

Um die Einhaltung des EMVG sicherzustellen, sind die Geraete wie in den
Handbuechern angegeben zu installieren und zu betreiben.

Japanese Voluntary Control Council for Interference (VCCI) class A statement

Korean government Ministry of Communication (MOC) statement

Please note that this device has been approved for business purpose with regard to electromagnetic interference. If you find this is not suitable for your use, you may exchange it for a nonbusiness purpose one.

Taiwan class A compliance statement

IBM agreement for licensed internal code

You accept the terms of this Agreement ¹ by your initial use of a machine that contains IBM Licensed Internal Code (called "Code"). These terms apply to Code used by certain machines IBM or your reseller specifies (called "Specific Machines"). International Business Machines Corporation or one of its subsidiaries ("IBM") owns copyrights in Code or has the right to license Code. IBM or a third party owns all copies of Code, including all copies made from them.

If you are the rightful possessor of a Specific Machine, IBM grants you a license to use the Code (or any replacement IBM provides) on, or in conjunction with, only the Specific Machine for which the Code is provided. IBM licenses the Code to only one rightful possessor at a time.

Under each license, IBM authorizes you to do only the following:

1. execute the Code to enable the Specific Machine to function according to its Official Published Specifications (called "Specifications");
2. make a backup or archival copy of the Code (unless IBM makes one available for your use), provided you reproduce the copyright notice and any other legend of ownership on the copy. You may use the copy only to replace the original, when necessary; and
3. execute and display the Code as necessary to maintain the Specific Machine.

You agree to acquire any replacement for, or additional copy of, Code directly from IBM in accordance with IBM's standard policies and practices. You also agree to use that Code under these terms.

You may transfer possession of the Code to another party only with the transfer of the Specific Machine. If you do so, you must 1) destroy all your copies of the Code that were not provided by IBM, 2) either give the other party all your IBM-provided copies of the Code or destroy them, and 3) notify the other party of these terms. IBM licenses the other party when it accepts these terms. These terms apply to all Code you acquire from any source.

Your license terminates when you no longer rightfully possess the Specific Machine.

Actions you must not take

You agree to use the Code only as authorized above. You must not do, for example, any of the following:

1. Otherwise copy, display, transfer, adapt, modify, or distribute the Code (electronically or otherwise), except as IBM may authorize in the Specific Machine's Specifications or in writing to you;
2. Reverse assemble, reverse compile, or otherwise translate the Code unless expressly permitted by applicable law without the possibility of contractual waiver;
3. Sublicense or assign the license for the Code; or
4. Lease the Code or any copy of it.

Glossary

This glossary includes terms for the Enterprise Storage Server (ESS) and other Seascape solution products.

This glossary includes selected terms and definitions from:

• The American National Standard Dictionary for Information Systems, ANSI X3.172-1990, copyright 1990 by the American National Standards Institute (ANSI), 11 West 42nd Street, New York, New York 10036. Definitions derived from this book have the symbol (A) after the definition.
• The Information Technology Vocabulary developed by Subcommittee 1, Joint Technical Committee 1, of the International Organization for Standardization and the International Electrotechnical Commission (SIO/IEC JTC1/SC1). Definitions derived from this book have the symbol (I) after the definition. Definitions taken from draft international standards, committee drafts, and working papers being developed by ISO/IEC JTC1/SCI have the symbol (T) after the definition, indicating that final agreement has not been reached among the participating National Bodies of SCI.

This glossary uses the following cross-reference form:

See

This refers the reader to one of three kinds of related information:

• A related term
• A term that is the expanded form of an abbreviation or acronym
• A synonym or more preferred term

A

access

(1) To obtain the use of a computer resource.

(2) In computer security, a specific type of interaction between a subject and an object that results in flow of information from one to the other.

Access-any mode

One of the two access modes that can be set for the ESS during initial configuration. It enables all fibre-channel-attached host systems with no defined access profile to access all logical volumes on the ESS. With a profile defined in ESS Specialist for a particular host, that host has access only to volumes that are assigned to the WWPN for that host. See pseudo-host and worldwide port name (WWPN).

active Copy Services server

The Copy Services server that manages the Copy Services domain. Either the primary or the backup Copy Services server can be the active Copy Services server. The backup Copy Services server is available to become the active Copy Services server if the primary Copy Services server fails.

alert

A message or log that a storage facility generates as the result of error event collection and analysis. An alert indicates that a service action is required.

allegiance

In Enterprise Systems Architecture/390, a relationship that is created between a device and one or more channel paths during the processing of certain conditions. See implicit allegiance, contingent allegiance, and reserved allegiance.

allocated storage

On an ESS, the space allocated to volumes, but not yet assigned. See assigned storage.

American National Standards Institute (ANSI)

An organization of producers, consumers, and general interest groups that establishes the procedures by which accredited organizations create and maintain voluntary industry standards in the United States. (A)

Anonymous host

In ESS Specialist, the label on a pseudo-host icon representing a host connection that uses the fibre-channel protocol (FCP) and that is not completely defined on the ESS. See pseudo-host and Access-any mode.

ANSI

See American National Standards Institute.

APAR

See authorized program analysis report.

arbitrated loop

For fibre-channel connections, a topology that enables the interconnection of a set of nodes. See point-to-point connection and switched fabric.

array

An ordered collection, or group, of physical devices (disk drive modules) that are used to define logical volumes or devices. More specifically, regarding the ESS, an array is a group of disks designated by the user to be managed by the RAID-5 technique. See redundant array of inexpensive disks (RAID).

ASCII

American Standard Code for Information Interchange. An ANSI standard (X3.4-1977) for assignment of 7-bit numeric codes (plus 1 bit for parity; some organizations, including IBM, have also used that bit to expand the basic code set) to represent alphabetic and numeric characters and common symbols.

assigned storage

On an ESS, the space allocated to a volume and assigned to a port.

authorized program analysis report (APAR)

A report of a problem caused by a suspected defect in a current, unaltered release of a program.

availability

The degree to which a system or resource is capable of performing its normal function. See data availability.

B

backup Copy Services server

One of two Copy Services servers, the other being the primary Copy Services server, in a Copy Services domain. The backup Copy Services server is available to become the active Copy Services server if the primary Copy Services server fails. A Copy Services server is software running in one of the two clusters of an ESS, managing data-copy operations for that Copy Services server group. See primary Copy Services server.

bay

Physical space on an ESS used for installing SCSI, ESCON, and fibre channel host adapter cards. The ESS has four bays, two in each cluster. See service boundary.

bit

(1) binary digit.

(2) The storage medium required to store a single binary digit.

(3) Either of the digits 0 or 1 when used in the binary numeration system. (T)

block

A group of consecutive bytes used as the basic storage unit in fixed-block architecture (FBA). All blocks on the storage device are the same size (fixed size). See fixed-block architecture (FBA) and data record.

byte

(1) A group of eight adjacent binary digits that represent one EBCDIC character.

(2) The storage medium required to store eight bits. See bit.

C

cache

A buffer storage that contains frequently accessed instructions and data, thereby reducing access time.

cache fast write

A form of the fast-write operation in which the subsystem writes the data directly to cache where it is available for later destaging.

cache hit

An event that occurs when a read operation is sent to the cluster, and the requested data is found in cache. The opposite of cache miss.

cache memory

Memory, typically volatile memory, that a subsystem uses to improve access times to instructions or data. The cache memory is typically smaller and faster than the primary memory or storage medium. In addition to residing in cache memory, the same data also resides on the storage devices in the storage facility.

cache miss

An event that occurs when a read operation is sent to the cluster, but the data is not found in cache. The opposite of cache hit.

call home

A communication link established between the ESS and service provider. The ESS can use this link to place a call to IBM or to another service provider when it requires service. With access to the machine, service personnel can perform service tasks, such as viewing error logs and problem logs or initiating trace and dump retrievals. See heartbeat and remote technical assistance information network.

cascading

(1) Connecting network controllers to each other in a succession of levels, to concentrate many more lines than a single level permits.

(2) In high-availability cluster multiprocessing (HACMP), cascading pertains to a cluster configuration in which the cluster node with the highest priority for a particular resource acquires the resource if the primary node fails. The cluster node relinquishes the resource to the primary node upon reintegration of the primary node into the cluster.

catcher

A server that service personnel use to collect and retain status data sent to it by an ESS.

CCR

See channel-command retry.

CCW

See channel command word.

CD-ROM

See compact disc, read-only memory.

CEC

See computer-electronic complex.

channel

In Enterprise Systems Architecture/390, the part of a channel subsystem that manages a single I/O interface between a channel subsystem and a set of control units.

channel command retry (CCR)

In Enterprise Systems Architecture/390, the protocol used between a channel and a control unit that enables the control unit to request that the channel reissue the current command.

channel command word (CCW)

In Enterprise Systems Architecture/390, a data structure that specifies an I/O operation to the channel subsystem.

channel path

In Enterprise Systems Architecture/390, the interconnection between a channel and its associated control units.

channel subsystem

In Enterprise Systems Architecture/390, the part of a host computer that manages I/O communication between the program and any attached control units.

channel-subsystem image

In Enterprise Systems Architecture/390, the logical functions that a system requires to perform the function of a channel subsystem. With ESCON multiple image facility (EMIF), one channel subsystem image exists in the channel subsystem for each logical partition (LPAR). Each image appears to be an independent channel subsystem program, but all images share a common set of hardware facilities.

CKD

See count key data.

CLI

See command-line interface.

cluster

(1) A partition in the ESS capable of performing all ESS functions. With two clusters in the ESS, any operational cluster can take over the processing of a failing cluster.

(2) On an AIX platform, a group of nodes within a complex.

cluster processor complex (CPC)

The unit within a cluster that provides the management function for the storage server. It consists of cluster processors, cluster memory, and related logic.

command-line interface (CLI)

An interface provided by an operating system that defines a set of commands and enables a user (or a script-like language) to issue these commands by entering text in response to the command prompt on the operating system's console (e.g., DOS commands, UNIX shell commands). IBM provides certain commands that can be installed with certain operating systems and that can be used to communicate with a Copy Services server. This set of commands is referred to as the Copy Services command line interface, or CLI for short.

compact disc, read-only memory (CD-ROM)

High-capacity read-only memory in the form of an optically read compact disc.

compression

(1) The process of eliminating gaps, empty fields, redundancies, and unnecessary data to shorten the length of records or blocks.

(2) Any encoding that reduces the number of bits used to represent a given message or record.

computer-electronic complex (CEC)

The set of hardware facilities associated with a host computer.

Concurrent Copy

A facility on a storage server that enables a program to make a backup of a data set while the logical volume remains available for subsequent processing. The data in the backup copy is frozen at the point in time that the server responds to the request.

concurrent installation of licensed internal code

Process of installing licensed internal code on an ESS while applications continue to run.

concurrent maintenance

Service that is performed on a unit while it is operational.

concurrent media maintenance

Service performed on a disk drive module (DDM) without losing access to the data.

configure

To define the logical and physical configuration of the input/output (I/O) subsystem through the user interface provided for this function on the storage facility.

consistent copy

A copy of a data entity (a logical volume, for example) that contains the contents of the entire data entity at a single instant in time.

console

A user interface to a server, such as can be provided by a personal computer.

contingent allegiance

In Enterprise Systems Architecture/390, a relationship that is created in a control unit between a device and a channel when unit-check status is accepted by the channel. The allegiance causes the control unit to guarantee access; the control unit does not present the busy status to the device. This enables the channel to retrieve sense data that is associated with the unit-check status on the channel path associated with the allegiance.

control unit (CU)

(1) A device that coordinates and controls the operation of one or more input/output devices, and synchronizes the operation of such devices with the operation of the system as a whole.

(2) In Enterprise Systems Architecture/390, a storage server with ESCON, FICON, or OEMI interfaces. The control unit adapts a native device interface to an I/O interface supported by an ESA/390 host system. On an ESS, the control unit would be the parts of the storage server that support the attachment of emulated CKD devices over ESCON, FICON, or OEMI interfaces. See cluster.

control-unit image

In Enterprise Systems Architecture/390, a logical subsystem that is accessed through an ESCON or FICON I/O interface. One or more control-unit images exist in each control unit. Each image appears to be an independent control unit, but each image shares a common set of hardware facilities. The ESS can emulate 3990-3, 3990-3 TPF, 3990-6, or 2105 control units.

control-unit initiated reconfiguration (CUIR)

Software mechanism used by the ESS to request that an operating system verify that one or more subsystem resources can be taken off-line for service. The ESS can use this process to automatically vary channel paths offline and online to facilitate bay service or concurrent code installation. Depending on the operating system, support for this process may be model-dependent, may depend on the Subsystem Device Driver, or may not exist.

Coordinated Universal Time (UTC)

The international standard of time that is kept by atomic clocks around the world.

Copy Services client

Software that runs on each ESS cluster in the Copy Services server group and that performs the following functions:

• Communicates configuration, status, and connectivity information to the Copy Services server.
• Performs data-copy functions on behalf of the Copy Services server.

Copy Services server group

A collection of user-designated ESS clusters participating in Copy Services functions managed by a designated active Copy Services server. A Copy Services server group is also called a Copy Services domain.

count field

The first field of a count key data (CKD) record. This eight-byte field contains a four-byte track address (CCHH). It defines the cylinder and head that are associated with the track, and a one-byte record number (R) that identifies the record on the track. It defines a one-byte key length that specifies the length of the record's key field (0 means no key field). It defines a two-byte data length that specifies the length of the record's data field (0 means no data field). Only the end-of-file record has a data length of zero.

count key data (CKD)

In Enterprise Systems Architecture/390, a data-record format employing self-defining record formats in which each record is represented by up to three fields--a count area identifying the record and specifying its format, an optional key area that can be used to identify the data area contents; and an optional data area that typically would contain the user data for the record. For CKD records on the ESS, the logical volume size is defined in terms of he device emulation mode (3390 or 3380 track format). See data record.

CPC

See cluster processor complex.

CRC

See cyclic redundancy check.

CU

See control unit.

CUIR

See control-unit initiated reconfiguration .

customer console

See console.

CUT

See Universaile Tempes du Coordinaire.

cyclic redundancy check (CRC)

A redundancy check in which the check key is generated by a cyclic algorithm. (T)

cylinder

A unit of storage on a CKD device. A cylinder has a fixed number of tracks.

D

DA

See device adapter and SSA adapter.

daisy chain

See serial connection.

DASD

See direct access storage device.

DASD fast write (DFW)

Caching of active write data by a storage server by journaling the data in nonvolatile storage, avoiding exposure to data loss.

data availability

The degree to which data is available when needed, typically measured as a percentage of time that the system would be capable of responding to any data request (e.g., 99.999% available).

data compression

A technique or algorithm used to encode data such that the encoded result can be stored in less space than the original data. The original data can be recovered from the encoded result through a reverse technique or reverse algorithm. See compression.

Data Facility Storage Management Subsystem

An operating environment that helps automate and centralize the management of storage. To manage storage, DFSMS provides the storage administrator with control over data class, storage class, management class, storage group, and automatic class selection routine definitions.

data field

The optional third field of a count key data (CKD) record. The count field specifies the length of the data field. The data field contains data that the program writes.

data record

The basic unit of S/390 and ZSeries storage on an ESS, combining a count field, a key field (optional), and a data field (optional), also known as a count-key-data (CKD) record. Data records are stored on a track. The records are sequentially numbered starting with 0. The first record R0 is typically called the track descriptor record and contains data normally used by the operating system to manage the track. The number of records is limited by the size of the track and the architectural limit of 256 records. The count field is always 8 bytes long and contains the lengths of the key and data fields, the key field has a length of 0 to 255 bytes, and the data field has a length of 0 to 65,535 or the maximum that will fit on the track. Typically, customer data appears in the data field. The use of the key field is dependent on the software managing the storage. See count-key-data (CKD) and fixed-block architecture (FBA).

data sharing

The ability of homogeneous or divergent host systems to concurrently utilize data that they store on one or more storage devices. The storage facility enables configured storage to be accessible to any, or all, attached host systems. To use this capability, the host program must be designed to support data that it is sharing.

DDM

See disk drive module (DDM).

DDM group

See disk drive module group.

dedicated storage

Storage within a storage facility that is configured such that a single host system has exclusive access to the storage.

demote

To remove a logical data unit from cache memory. A subsystem demotes a data unit in order to make room for other logical data units in the cache. It might also demote a data unit because the logical data unit is not valid. A subsystem must destage logical data units with active write units before they can be demoted.

destaging

Movement of data from an online or higher priority to an offline or low priority device..

device

In Enterprise Systems Architecture/390, a disk drive.

device adapter (DA)

A physical component of the ESS that provides communication between the clusters and the storage devices. The ESS has eight device adapters that it deploys in pairs, one from each cluster. DA pairing enables the ESS to access any disk drive from either of two paths, providing fault tolerance and enhanced availability.

device address

In Enterprise Systems Architecture/390, the field of an ESCON or FICON device-level frame that selects a specific device on a control-unit image.

device interface card

A physical subunit of a storage cluster that provides the communication with the attached DDMs.

device number

In Enterprise Systems Architecture/390, a four-hexadecimal-character identifier, for example 13A0, that the systems administrator associates with a device to facilitate communication between the program and the host operator. The device number is associated with a subchannel.

device sparing

A subsystem function that automatically copies data from a failing DDM to a spare DDM. The subsystem maintains data access during the process.

direct access storage device (DASD)

(1) A mass storage medium on which a computer stores data.

(2) A disk device.

disk drive

Standard term for a disk-based nonvolatile storage medium. The ESS uses hard disk drives as the primary nonvolatile storage media to store host data.

disk drive module (DDM)

A field replaceable unit that consists of a single disk drive and its associated packaging.

disk drive module group

In the ESS, a group of eight disk drive modules (DDMs) contained in an 8-pack and installed as a unit.

DNS

See domain name system (DNS).

domain

(1) That part of a computer network in which the data processing resources are under common control.

(2) In TCP/IP, the naming system used in hierarchical networks.

(3) A Copy Services server group, in other words, the set of clusters designated by the user to be managed by a particular Copy Services server.

domain name system (DNS)

In TCP/IP, the server program that supplies name-to-address translation by mapping domain names to internet addresses. The address of a DNS server is the internet address of the server that hosts the DNS software for the network.

drawer

A unit that contains multiple DDMs and provides power, cooling, and related interconnection logic to make the DDMs accessible to attached host systems.

drive

(1) A peripheral device, especially one that has addressed storage media. See disk drive module (DDM).

(2) The mechanism used to seek, read, and write information on a storage medium.

duplex

A communication mode in which data can be sent and received at the same time.

dynamic sparing

The ability of a storage server to move data from a failing disk drive module (DDM) to a spare DDM while maintaining storage functions.

E

E10

A previous model of the ESS.

E20

A previous model of the ESS.

EBCDIC

See extended binary-coded decimal interchange code.

EC

See engineering change.

ECKD

See extended count key data.

electrostatic discharge (ESD)

An undesirable discharge of static electricity that can damage equipment and degrade electrical circuitry.

emergency power off (EPO)

A means of turning off power during an emergency, usually a switch.

EMIF

See ESCON multiple image facility.

enclosure

A unit that houses the components of a storage subsystem, such as a control unit, disk drives, and power source.

end of file

A coded character recorded on a data medium to indicate the end of the medium. On a CKD direct access storage device, the subsystem indicates the end of a file by including a record with a data length of zero.

engineering change (EC)

An update to a machine, part, or program.

Enterprise Systems Architecture/390^(R) (ESA/390) and z/Architecture

IBM architectures for mainframe computers and peripherals. Processor systems that follow the ESA/390 architecture include the ES/9000^(R) family, while the zSeries server uses the z/Architecture.

Enterprise Systems Connection (ESCON)

(1) An ESA/390 and zSeries computer peripheral interface. The I/O interface uses ESA/390 logical protocols over a serial interface that configures attached units to a communication fabric.

(2) A set of IBM products and services that provide a dynamically connected environment within an enterprise.

EPO

See emergency power off.

ERP

See error recovery procedure.

error recovery procedure (ERP)

Procedures designed to help isolate and, where possible, to recover from errors in equipment. The procedures are often used in conjunction with programs that record information on machine malfunctions.

ESA/390

See Enterprise Systems Architecture/390.

ESCD

See ESCON director.

ESCON

See Enterprise System Connection (ESCON).

ESCON channel

An S/390 or zSeries channel that supports ESCON protocols.

ESCON director (ESCD)

An I/O interface switch that provides for the interconnection of multiple ESCON interfaces in a distributed-star topology.

ESCON host systems

S/390 or zSeries hosts that attach to the ESS with an ESCON adapter. Such host systems run on MVS, VM, VSE, or TPF operating systems.

ESCON multiple image facility (EMIF)

In Enterprise Systems Architecture/390, a function that enables LPARs to share an ESCON channel path by providing each LPAR with its own channel-subsystem image.

EsconNet

In ESS Specialist, the label on a pseudo-host icon representing a host connection that uses the ESCON protocol and that is not completely defined on the ESS. See pseudo-host and Access-any mode.

ESD

See electrostatic discharge.

eserver

See IBM .

ESS

See IBM TotalStorage Enterprise Storage Server (ESS).

ESS Expert

See IBM StorWatch Enterprise Storage Server Expert.

ESS Specialist

See IBM TotalStorage Enterprise Storage Server Specialist.

ESS Copy Services

See IBM TotalStorage Enterprise Storage Server Copy Services.

ESSNet

See IBM TotalStorage Enterprise Storage Server Network (ESSNet).

Expert

See IBM StorWatch Enterprise Storage Server Expert.

extended binary-coded decimal interchange code (EBCDIC)

A coding scheme developed by IBM used to represent various alphabetic, numeric, and special symbols with a coded character set of 256 8-bit codes.

extended count key data (ECKD)

An extension of the CKD architecture.

Extended Remote Copy (XRC)

A function of a storage server that assists a control program to maintain a consistent copy of a logical volume on another storage facility. All modifications of the primary logical volume by any attached host are presented in order to a single host. The host then makes these modifications on the secondary logical volume.

extent

A continuous space on disk that is occupied by or reserved for a particular data set, data space, or file. The unit of increment is a track. See multiple allegiance and parallel access volumes (PAV).

F

F10

A model of the ESS featuring a single-phase power supply. It has fewer expansion capabilities than the Model F20.

F20

A model of the ESS featuring a three-phase power supply. It has more expansion capabilities than the Model F10, including the ability to support a separate expansion rack.

fabric

In fibre-channel technology, a routing structure, such as a switch, receives addressed information and routes to the appropriate destination. A fabric can consist of more than one switch. When multiple fibre-channel switches are interconnected, they are said to be cascaded.

failback

Cluster recovery from failover following repair. See failover.

failover

On the ESS, the process of transferring all control of a storage facility to a single cluster when the other cluster in the storage facility fails.

fast write

A write operation at cache speed that does not require immediate transfer of data to a DDM. The subsystem writes the data directly to cache, to nonvolatile storage, or to both. The data is then available for destaging. A fast-write operation reduces the time an application must wait for the I/O operation to complete.

FBA

See fixed-block architecture.

FC-AL

See Fibre Channel-Arbitrated Loop.

FCP

See fibre-channel protocol.

FCS

See fibre-channel standard.

feature code

A code that identifies a particular orderable option and used service personnel to process hardware and software orders. Individual optional features are each identified by a unique feature code.

fibre-channel (FC)

Fibre-channel is an architecture that supports full-duplex communication over a serial interface that configures attached units to a communication fabric.

The ESS supports data transmission over fibre-optic cable through its fibre-channel adapters.

Fibre Channel-Arbitrated Loop (FC-AL)

An implementation of the fibre-channel technology that uses a ring topology for communication. In this topology, two or more fibre-channel end points are interconnected through a looped interface. The ESS supports this topology.

fibre-channel protocol (FCP)

For fibre-channel communication the protocol has five layers. The layers define how fibre-channel ports interact through their physical links to communicate with other ports.

fibre-channel standard (FCS)

An ANSI standard for a computer peripheral interface. The I/O interface defines a protocol for communication over a serial interface that configures attached units to a communication fabric. The protocol has two layers. The IP layer defines basic interconnection protocols. The upper layer supports one or more logical protocols. Refer to ANSI X3.230-199x.

FICON

Acronym derived from FIbre-channel CONnection, a fibre-channel communications protocol designed for IBM mainframe computers and peripherals.

FiconNet

In ESS Specialist, the label on a pseudo-host icon representing a host connection that uses the FICON protocol and that is not completely defined on the ESS. See pseudo-host and Access-any mode.

field replaceable unit (FRU)

An assembly that is replaced in its entirety when any one of its components fails. In some cases, a field replaceable unit may contain other field replaceable units.

FIFO

See first-in-first-out.

firewall

A protection against unauthorized connection to a computer or a data storage system. The protection is usually in the form of software on a gateway server that grants access to users who meet authorization criteria.

first-in-first-out (FIFO)

A queuing technique in which the next item to be retrieved is the item that has been in the queue for the longest time. (A)

fixed-block architecture (FBA)

An architecture for logical devices that specifies the format of and access mechanisms for the logical data units on the device. The logical data unit is a block. All blocks on the device are the same size (fixed size). The subsystem can access them independently.

fixed-block devices

An architecture for logical devices that specifies the format of the logical data units on the device. The logical data unit is a block. All blocks on the device are the same size (fixed size); the subsystem can access them independently. This is the required format of the logical data units for host systems that attach with a Small Computer System Interface (SCSI) or fibre-channel interface. See Small Computer System Interface (SCSI).

FlashCopy

An optional feature for the ESS that can make an instant copy of data, that is, a point-in-time copy of a volume.

FRU

See field replaceable unit.

full duplex

See duplex.

G

GB

See gigabyte.

gigabyte (GB)

A gigabyte of storage is 10⁹ bytes. A gigabyte of memory is 2³⁰ bytes.

group

See disk drive module group or Copy Services server group.

H

HA

See host adapter.

HACMP

Software that provides host clustering, so that a failure of one host is recovered by moving jobs to other hosts within the cluster; named for high-availability cluster multiprocessing

hard disk drive (HDD)

(1) A storage medium within a storage server used to maintain information that the storage server requires.

(2) A mass storage medium for computers that is typically available as a fixed disk (such as the disks used in system units of personal computers or in drives that are external to a personal computer) or a removable cartridge.

HDA

See head and disk assembly.

HDD

See hard disk drive.

hdisk

An AIX term for storage space.

head and disk assembly (HDA)

The portion of an HDD associated with the medium and the read/write head.

heartbeat

A status report sent at regular intervals from the ESS. The service provider uses this report to monitor the health of the call home process. See call home and heartbeat call home record, remote technical assistance information network.

heartbeat call home record

Machine operating and service information sent to a service machine. These records might include such information as feature code information and product logical configuration information.

home address

A nine-byte field at the beginning of a track that contains information that identifies the physical track and its association with a cylinder.

host adapter (HA)

A physical subunit of a storage server that provides the ability to attach to one or more host I/O interfaces. The Enterprise Storage Server has four HA bays, two in each cluster. Each bay supports up to four host adapters.

host processor

A processor that controls all or part of a user application network. In a network, the processing unit in which the data communication access method resides. See host system.

host system

(1) A computer system that is connected to the ESS. The ESS supports both mainframe (S/390 or zSeries) hosts as well as open-systems hosts. S/390 or zSeries hosts are connected to the ESS through ESCON or FICON interfaces. Open-systems hosts are connected to the ESS by SCSI or fibre-channel interfaces.

(2) The data processing system to which a network is connected and with which the system can communicate.

(3) The controlling or highest level system in a data communication configuration.

hot plug

Pertaining to the ability to add or remove a hardware facility or resource to a unit while power is on.

I

IBM

The brand name for a series of server products that are optimized for e-commerce. The products include the iSeries, pSeries, xSeries, and zSeries.

IBM product engineering (PE)

The third-level of IBM service support. Product engineering is composed of IBM engineers who have experience in supporting a product or who are knowledgeable about the product.

IBM StorWatch Enterprise Storage Server Expert (ESS Expert)

The software that gathers performance data from the ESS and presents it through a Web browser.

IBM TotalStorage Enterprise Storage Server (ESS)

A member of the Seascape^(R) product family of storage servers and attached storage devices (disk drive modules). The ESS provides for high-performance, fault-tolerant storage and management of enterprise data, providing access through multiple concurrent operating systems and communication protocols. High performance is provided by four symmetric multiprocessors, integrated caching, RAID support for the disk drive modules, and disk access through a high-speed serial storage architecture (SSA) interface.

IBM TotalStorage Enterprise Storage Server Specialist (ESS Specialist)

The Web browser-based configuration management interface to the ESS.

IBM TotalStorage Enterprise Storage Server Copy Services (ESS Copy Services)

The Web browser-based interface for managing the data-copy functions of FlashCopy and PPRC.

IBM TotalStorage Enterprise Storage Server Network (ESSNet)

A private network providing Web browser access to the ESS. IBM installs the ESSNet on an IBM workstation supplied with the first ESS delivery when they install the ESS. ESSNet I is a version of the ESSNet that is installed on an IBM workstation running a Microsoft Windows operation system. ESSNet II is a version of the ESSNet that is installed on an IBM workstation running a Linux operation system.

ID

See identifier.

identifier (ID)

A unique name or address that identifies things such as programs, devices, or systems.

IML

See initial microprogram load.

implicit allegiance

In Enterprise Systems Architecture/390, a relationship that a control unit creates between a device and a channel path when the device accepts a read or write operation. The control unit guarantees access to the channel program over the set of channel paths that it associates with the allegiance.

initial microprogram load (IML)

To load and initiate microcode or firmware that controls a hardware entity such as a processor or a storage server.

initial program load (IPL)

To load and initiate the software, typically an operating system, that controls a host computer.

initiator

A SCSI device that communicates with and controls one or more targets. An initiator is typically an I/O adapter on a host computer. A SCSI initiator is analogous to an S/390 channel. A SCSI logical unit is analogous to an S/390 device. See target.

i-node

The internal structure in an AIX operating system that describes the individual files in the operating system. It contains the code, type, location, and owner of a file.

input/output (I/O)

Pertaining to (a) input, output, or both or (b) a device, process, or channel involved in data input, data output, or both.

Internet Protocol (IP)

In the Internet suite of protocols, a protocol without connections that routes data through a network or interconnecting networks and acts as an intermediary between the higher protocol layers and the physical network.

invalidate

To remove a logical data unit from cache memory, because it cannot support continued access to the logical data unit on the device. This removal may be the result of a failure within the storage server or a storage device that is associated with the device.

I/O

See input/output.

I/O device

An addressable read and write unit, such as a disk drive device, magnetic tape device, or printer.

I/O interface

An interface that enables a host to perform read and write operations with its associated peripheral devices.

IP

See Internet Protocol.

IPL

See initial program load.

iSeries

An IBM server Series product that emphasizes integration. See AS/400 and iSeries.

J

Java virtual machine (JVM)

A software implementation of a central processing unit (CPU) that runs compiled Java code (applets and applications).

JVM

See Java virtual machine.

K

KB

See kilobyte.

key field

The second (optional) field of a CKD record. The key length is specified in the count field. The key length determines the field length. The program writes the data in the key field and use the key field to identify or locate a given record. The subsystem does not use the key field.

kilobyte (KB)

(1) For processor storage, real, and virtual storage, and channel volume, 2¹⁰ or 1024 bytes.

(2) For disk storage capacity and communications volume, 1000 bytes.

KPOH

See thousands of power-on hours.

L

LAN

See local area network.

last-in first-out (LIFO)

A queuing technique in which the next item to be retrieved is the item most recently placed in the queue. (A)

LBA

See logical block address.

LCU

See logical control unit.

least recently used (LRU)

(1) The algorithm used to identify and make available the cache space that contains the least-recently used data.

(2) A policy for a caching algorithm that chooses to remove from cache the item that has the longest elapsed time since its last access.

LED

See light-emitting diode.

LIC

See licensed internal code.

licensed internal code (LIC)

Microcode that IBM does not sell as part of a machine, but licenses to the customer. LIC is implemented in a part of storage that is not addressable by user programs. Some IBM products use it to implement functions as an alternate to hard-wired circuitry.

LIFO

See last-in first-out.

light-emitting diode (LED)

A semiconductor chip that gives off visible or infrared light when activated.

link address

On an ESCON or FICON interface, the portion of a source or destination address in a frame that ESCON or FICON uses to route a frame through an ESCON or FICON director. ESCON or FICON associates the link address with a specific switch port that is on the ESCON or FICON director. Equivalently, it associates the link address with the channel-subsystem or control unit link-level functions that are attached to the switch port.

link-level facility

The ESCON or FICON hardware and logical functions of a control unit or channel subsystem that allow communication over an ESCON or FICON write interface and an ESCON or FICON read interface.

local area network (LAN)

A computer network located on a user's premises within a limited geographic area.

local e-mail

An e-mail configuration option for storage servers that are connected to a host-system network that does not have a domain name system (DNS) server.

loop

The physical connection between a pair of device adapters in the ESS. See device adapter (DA).

logical address

On an ESCON or FICON interface, the portion of a source or destination address in a frame used to select a specific channel-subsystem or control-unit image.

logical block address (LBA)

The address assigned by the ESS to a sector of a disk.

logical control unit (LCU)

See control-unit image.

logical data unit

A unit of storage that is accessible on a given device.

logical device

The facilities of a storage server associated with the processing of I/O operations directed to a single host-accessible emulated I/O device. The associated storage is referred to as a logical volume. The logical device is mapped to one or more host-addressable units, such as a device on an S/390 I/O interface or a logical unit on a SCSI I/O interface, such that the host initiating I/O operations to the I/O-addressable unit interacts with the storage on the associated logical device.

logical partition (LPAR)

A set of functions that create the programming environment that is defined by the ESA/390 architecture. ESA/390 architecture uses this term when more than one LPAR is established on a processor. An LPAR is conceptually similar to a virtual machine environment except that the LPAR is a function of the processor. Also the LPAR does not depend on an operating system to create the virtual machine environment.

logical path

For Copy Services, a relationship between a source logical subsystem and target logical subsystem that is created over a physical path through the interconnection fabric used for Copy Services functions.

logical subsystem (LSS)

A construct within a storage server that consists of a group of up to 256 logical devices. A storage server can have up to 16 CKD logical subsystems (4096 CKD logical devices) and also up to 16 fixed-block (FB) logical subsystems (4096 FB logical devices). The logical subsystem facilitates configuration of the storage server and may have other implications relative to the operation of certain functions. There is a one-to-one mapping between a CKD logical subsystem and an S/390 control-unit image.

For S/390 or zSeries hosts, a logical subsystem represents a logical control unit (LCU). Each control-unit image is associated with only one logical subsystem. See control-unit image.

logical unit

The open-systems term for a logical disk drive.

logical unit number (LUN)

A SCSI term for a unique number used on a SCSI bus to enable it to differentiate between up to eight separate devices, each of which is a logical unit.

logical volume

The storage medium associated with a logical disk drive. A logical volume typically resides on one or more storage devices. The ESS administrator defines this unit of storage. The logical volume, when residing on a RAID-5 array, is spread over 6 +P or 7 +P drives, where P is parity. A logical volume can also reside on a non-RAID storage device. See count key data and fixed block address.

logical volume manager (LVM)

A set of system commands, library routines, and other tools that allow the user to establish and control logical volume storage. The LVM maps data between the logical view of storage space and the physical disk drive module (DDM).

longitudinal redundancy check (LRC)

Also called a longitudinal parity check, a method of error-checking during data transfer involving checking parity on a row of binary digits that are members of a set forming a matrix.

longwave laser adapter

A connector used between host and ESS to support longwave fibre-channel communication.

LPAR

See logical partition.

LRC

See longitudinal redundancy check.

LRU

See least recently used.

LSS

See logical subsystem.

LUN

See logical unit number.

LVM

See logical volume manager.

M

machine level control (MLC)

A database that contains the EC level and configuration of products in the field.

machine reported product data (MRPD)

Product data gathered by a machine and sent to a destination such as an IBM support server or RETAIN. These records might include such information as feature code information and product logical configuration information.

mainframe

A computer, usually in a computer center, with extensive capabilities and resources to which other computers may be connected so that they can share facilities. (T)

maintenance analysis procedure (MAP)

A hardware maintenance document that gives an IBM service representative a step-by-step procedure for tracing a symptom to the cause of a failure.

management information base (MIB)

(1) A schema for defining a tree structure that identifies and defines certain objects that can be passed between units using an SNMP protocol. The objects passed typically contain certain information about the product such as the physical or logical characteristics of the product.

(2) Shorthand for referring to the MIB-based record of a network device. Information about a managed device is defined and stored in the management information base (MIB) of the device. Each ESS has a MIB. SNMP-based network management software uses the record to identify the device.See simple network management protocol.

MAP

See maintenance analysis procedure.

MB

See megabyte.

MCA

See Micro Channel architecture.

mean time between failures (MTBF)

(1) A projection of the time that an individual unit remains functional. The time is based on averaging the performance, or projected performance, of a population of statistically independent units. The units operate under a set of conditions or assumptions.

(2) For a stated period in the life of a functional unit, the mean value of the lengths of time between consecutive failures under stated conditions. (I) (A)

medium

For a storage facility, the disk surface on which data is stored.

megabyte (MB)

(1) For processor storage, real and virtual storage, and channel volume, 2²⁰ or 1 048 576 bytes.

(2) For disk storage capacity and communications volume, 1 000 000 bytes.

MES

See miscellaneous equipment specification.

MIB

See management information base.

Micro Channel architecture (MCA)

The rules that define how subsystems and adapters use the Micro Channel bus in a computer. The architecture defines the services that each subsystem can or must provide.

MIH

See missing-interrupt handler.

mirrored pair

Two units that contain the same data. The system refers to them as one entity.

mirroring

In host systems, the process of writing the same data to two disk units within the same auxiliary storage pool at the same time.

miscellaneous equipment specification (MES)

IBM field-installed change to a machine.

MLC

See machine level control.

missing-interrupt handler (MIH)

An MVS and MVS/XA facility that tracks I/O interrupts. MIH informs the operator and creates a record whenever an expected interrupt fails to occur before a specified elapsed time is exceeded.

mobile service terminal (MoST)

The mobile terminal used by service personnel.

MoST

See mobile service terminal.

MRPD

See machine reported product data.

MTBF

See mean time between failures.

multiple allegiance

ESS hardware function independent of software support that enables concurrent access to the same logical volume on the ESS from multiple system images, as long as the system images are accessing different extents. See extent and parallel access volumes.

multiple virtual storage (MVS)

Implies MVS/390, MVS/XA, MVS/ESA, and the MVS element of the OS/390 operating system.

MVS

See multiple virtual storage.

N

node

The unit that is connected in a fibre-channel network. An ESS is a node in a fibre-channel network.

non-RAID

A disk drive set up independently of other disk drives and not set up as part of a disk drive module group to store data using the redundant array of disks (RAID) data-striping methodology.

non-removable medium

A recording medium that cannot be added to or removed from a storage device.

non-retentive data

Data that the control program can easily recreate in the event it is lost. The control program may cache non-retentive write data in volatile memory.

nonvolatile storage (NVS)

(1) Typically refers to nonvolatile memory on a processor rather than to a nonvolatile disk storage device. On a storage facility, nonvolatile storage is used to store active write data to avoid data loss in the event of a power loss.

(2) A storage device whose contents are not lost when power is cut off.

NVS

See nonvolatile storage.

O

octet

In Internet Protocol (IP) addressing, one of the four parts of a 32-bit integer presented in dotted decimal notation. dotted decimal notation consists of four 8-bit numbers written in base 10. For example, 9.113.76.250 is an IP address containing the octets 9, 113, 76, and 250.

OEMI

See original equipment manufacturer's information.

open system

A system whose characteristics comply with standards made available throughout the industry and that therefore can be connected to other systems complying with the same standards. Applied to the ESS, such systems are those hosts that connect to the ESS through SCSI or SCSI-FCP adapters.

organizationally unique identifier (OUI)

An IEEE-standards number that identifies an organization with a 24-bit globally unique assigned number referenced by various standards. OUI is used in the family of 802 LAN standards, such as Ethernet and Token Ring.

original equipment manufacturer's information (OEMI)

A reference to an IBM guideline for a computer peripheral interface. The interface uses ESA/390 logical protocols over an I/O interface that configures attached units in a multidrop bus topology.

OUI

See organizationally unique identifier.

P

panel

The formatted display of information that appears on a display screen.

parallel access volume (PAV)

An advanced function of the ESS that enables OS/390 systems to issue concurrent I/O requests against a CKD logical volume by associating multiple devices of a single control-unit image with a single logical device. Up to 8 device addresses can be assigned to a parallel access volume. PAV enables two or more concurrent writes to the same logical volume, as long as the writes are not to the same extents. See extent and multiple allegiance.

parity

A data checking scheme used in a computer system to ensure the integrity of the data. The RAID implementation uses parity to recreate data if a disk drive fails.

path group

The ESA/390 term for a set of channel paths that are defined to a control unit as being associated with a single logical partition (LPAR). The channel paths are in a group state and are online to the host. See logical partition (LPAR).

path group identifier

The ESA/390 term for the identifier that uniquely identifies a given logical partition (LPAR). The path group identifier is used in communication between the LPAR program and a device. The identifier associates the path group with one or more channel paths, thereby defining these paths to the control unit as being associated with the same LPAR.

PAV

See parallel access volume.

PCI

See peripheral component interconnect.

PE

See IBM product engineering.

Peer-to-Peer Remote Copy (PPRC)

A function of a storage server that maintains a consistent copy of a logical volume on the same storage server or on another storage server. All modifications that any attached host performs on the primary logical volume are also performed on the secondary logical volume.

peripheral component interconnect (PCI)

An architecture for a system bus and associated protocols that supports attachments of adapter cards to a system backplane.

physical path

A single path through the I/O interconnection fabric that attaches two units. For Copy Services, this is the path from a host adapter on one ESS (through cabling and switches) to a host adapter on another ESS.

point-to-point connection

For fibre-channel connections, a topology that enables the direct interconnection of ports. See arbitrated loop and switched fabric.

POST

See power-on self test.

power-on self test (POST)

A diagnostic test run by servers or computers when they are turned on.

PPRC

See Peer-to-Peer Remote Copy.

predictable write

A write operation that can cache without knowledge of the existing format on the medium. All writes on FBA DASD devices are predictable. On CKD DASD devices, a write is predictable if it does a format write for the first data record on the track.

primary Copy Services server

One of two Copy Services servers, the other being the backup Copy Services server, in a Copy Services domain. The primary Copy Services server is the active Copy Services server until it fails; it is then replaced by the backup Copy Services server. A Copy Services server is software running in one of the two clusters of an ESS and performs data-copy operations within that group. See active Copy Services server. See backup Copy Services server.

product engineering

See IBM product engineering.

program

On a computer, a generic term for software that controls the operation of the computer. Typically, the program is a logical assemblage of software modules that perform multiple related tasks.

program-controlled interruption

An interruption that occurs when an I/O channel fetches a channel command word with the program-controlled interruption flag on.

program temporary fix (PTF)

A temporary solution or bypass of a problem diagnosed by IBM in a current unaltered release of a program

promote

To add a logical data unit to cache memory.

protected volume

An AS/400 term for a disk storage device that is protected from data loss by RAID techniques. An AS/400 does not mirror a volume configured as a protected volume, while it does mirror all volumes configured as unprotected volumes. The ESS, however, can be configured to indicate that an AS/400 volume is protected or unprotected, and give it RAID protection in either case. This allows AS/400 data to have RAID protection on the ESS, while enabling the AS/400 to perform mirroring on all its data, providing redundancy for recovering from host adapter failures, interconnection failures, or device failures.

pSeries

An IBM server Series product that emphasizes performance. See RS/6000 and pSeries.

pseudo-host

A host connection that is not explicitly defined to the ESS and that has access to at least one volume configured on the ESS. Such a host connection using the FICON protocol is represented by a FiconNet pseudo-host icon, the ESCON protocol by an EsconNet pseudo-host icon, and the FCP protocol by an Anonymous pseudo-host icon. "Anonymous host" is a commonly used synonym for "pseudo-host". The ESS adds a pseudo-host icon only when the ESS access mode is set to Access any. See Access-any mode.

PTF

See program temporary fix.

R

REQ/ACK

Short for request for acknowledgement and acknowledgement between data transmitter and receptor to verify connection.

R0

See track-descriptor record.

rack

See enclosure.

RAID

See redundant array of inexpensive disks and array. RAID also is expanded to redundant array of independent disks.

RAID 5

A type of RAID that optimizes cost-effective performance through data striping while providing fault tolerance for up to two failed disk drives by distributing parity across all of the drives in the array plus one parity disk drive. The ESS automatically reserves spare disk drives when it assigns arrays to a device adapter pair (DA pair). See device adapter (DA).

random access

A mode of accessing data on a medium in a manner that requires the storage device to access nonconsecutive storage locations on the medium.

redundant array of inexpensive disks (RAID)

A methodology of grouping disk drives for managing disk storage to insulate data from a failing disk drive.

remote technical assistance information network (RETAIN)

The initial service tracking system for IBM service support for capturing heartbeat and call-home records. See support catcher and support catcher telephone number.

reserved allegiance

In Enterprise Systems Architecture/390, a relationship that is created in a control unit between a device and a channel path when a Sense Reserve command is completed by the device. The allegiance causes the control unit to guarantee access (busy status is not presented) to the device. Access is over the set of channel paths that are associated with the allegiance; access is for one or more channel programs, until the allegiance ends.

RETAIN

See remote technical assistance information network.

S

S/390 and zSeries

IBM enterprise servers based on Enterprise Systems Architecture/390 (ESA/390) and z/Architecture, respectively. "S/390" is a shortened form of the original name "System 390". See zSeries.

S/390 and zSeries storage

Storage arrays and logical volumes that are defined in the ESS as connected to S/390 and zSeries servers. This term is synonymous with count-key-data (CKD) storage.

SAID

See system adapter identification number.

SAM

See sequential access method.

SAN

See storage area network.

SBCON

See Single-Byte Command Code Sets Connection.

screen

The physical surface of a display device upon which information is shown to users.

SCSI

See Small Computer System Interface (SCSI).

SCSI device

A disk drive connected to the host through a SCSI I/O interface. A SCSI device is either an initiator or a target. See Small Computer System Interface (SCSI).

SCSI host systems

Host systems attached to the ESS with a SCSI interface. Such host systems run on UNIX, OS/400 and iSeries, Windows NT, Windows 2000, or Novell NetWare operating systems.

SCSI ID

A unique identifier assigned to a SCSI device that is used in protocols on the SCSI interface to identify or select the device. The number of data bits on the SCSI bus determines the number of available SCSI IDs. A wide interface has 16 bits, with 16 possible IDs.

Seascape architecture

A storage system architecture developed by IBM for open-systems servers and S/390 and zSeries host systems. It provides storage solutions that integrate software, storage management, and technology for disk, tape, and optical storage.

serial connection

A method of device interconnection for determining interrupt priority by connecting the interrupt sources serially.

self-timed interface (STI)

An interface that has one or more conductors that transmit information serially between two interconnected units without requiring any clock signals to recover the data. The interface performs clock recovery independently on each serial data stream and uses information in the data stream to determine character boundaries and inter-conductor synchronization.

sequential access

A mode of accessing data on a medium in a manner that requires the storage device to access consecutive storage locations on the medium.

sequential access method (SAM)

An access method for storing, deleting, or retrieving data in a continuous sequence, based on the logical order of the records in the file.

serial storage architecture (SSA)

An IBM standard for a computer peripheral interface. The interface uses a SCSI logical protocol over a serial interface that configures attached targets and initiators in a ring topology.

server

(1) A type of host that provides certain services to other hosts that are referred to as clients.

(2) A functional unit that provides services to one or more clients over a network.

service boundary

Identifies a group of components that are unavailable when one of the components of the group is being serviced. Service boundaries are provided on the ESS, for example, in each host bay and each cluster.

service information message (SIM)

A message sent by a storage server to service personnel through an S/390 operating system.

service personnel

Individuals or a company authorized to service the ESS. This term also refers to a service provider, a service representative, or an IBM service support representative (SSR). An IBM SSR installs the ESS.

service processor

A dedicated processing unit used to service a storage facility.

service support representative (SSR)

Individuals or a company authorized to service the ESS. This term also refers to a service provider, a service representative, or an IBM service support representative (SSR). An IBM SSR installs the ESS.

shared storage

Storage within an ESS that is configured so that multiple homogeneous or divergent hosts can concurrently access the storage. The storage has a uniform appearance to all hosts. The host programs that access the storage must have a common model for the information on a storage device. The programs must be designed to handle the effects of concurrent access.

shortwave laser adapter

A connector used between host and ESS to support shortwave fibre-channel communication.

SIM

See service-information message.

simple network management protocol (SNMP)

A network management protocol in the Internet suite of protocols that is used to monitor routers and attached network devices. SNMP is an application layer protocol in the Open Systems Interconnection reference model. See management information base (MIB).

simplex volume

A volume that is not part of a FlashCopy, XRC, or PPRC volume pair.

Single-Byte Command Code Sets Connection (SBCON)

The ANSI standard for the ESCON or FICON I/O interface.

Small Computer System Interface (SCSI)

(1) An ANSI standard for a logical interface to computer peripherals and for a computer peripheral interface. The interface uses a SCSI logical protocol over an I/O interface that configures attached initiators and targets in a multidrop bus topology.

(2) A standard hardware interface that enables a variety of peripheral devices to communicate with one another.

SMIT

See System Management Interface Tool.

SMP

See symmetric multi-processor.

SNMP

See simple network management protocol.

software transparency

Criteria applied to a processing environment that states that changes do not require modifications to the host software in order to continue to provide an existing function.

spare

A disk drive on the ESS that can replace a failed disk drive. A spare can be predesignated to allow automatic dynamic sparing. Any data preexisting on a disk drive that is invoked as a spare is destroyed by the dynamic sparing copy process.

spatial reuse

A feature of serial storage architecture that enables a device adapter loop to support many simultaneous read/write operations. See serial storage architecture.

Specialist

See IBM TotalStorage Enterprise Storage Server Specialist.

SSA

See serial storage architecture.

SSA adapter

A physical adapter based on serial storage architecture. The device adapters used to connect disk drive modules to an ESS cluster are SSA adapters. See serial storage architecture.

SSID

See subsystem identifier.

SSR

See service support representative.

stacked status

In Enterprise Systems Architecture/390, the condition when the control unit is holding status for the channel, and the channel responded with the stack-status control the last time the control unit attempted to present the status.

stage operation

The operation of reading data from the physical disk drive into the cache.

staging

To move data from an offline or low-priority device back to an online or higher priority device, usually on demand of the system or on request of the user.

STI

See self-timed interface.

storage area network

A network that connects a company's heterogeneous storage resources.

storage complex

Multiple storage facilities.

storage device

A physical unit that provides a mechanism to store data on a given medium such that it can be subsequently retrieved. See disk drive module.

storage facility

(1) A physical unit that consists of a storage server integrated with one or more storage devices to provide storage capability to a host computer.

(2) A storage server and its attached storage devices.

storage server

A physical unit that manages attached storage devices and provides an interface between them and a host computer by providing the function of one or more logical subsystems. The storage server can provide functions that are not provided by the storage device. The storage server has one or more clusters.

striping

A technique that distributes data in bit, byte, multibyte, record, or block increments across multiple disk drives.

subchannel

A logical function of a channel subsystem associated with the management of a single device.

subsystem identifier (SSID)

A number that uniquely identifies a logical subsystem within a computer installation.

support catcher

A server to which a machine sends a trace or a dump package.

support catcher telephone number

The telephone number that connects the support catcher server to the ESS to receive a trace or dump package. See support catcher. See remote technical assistance information network (RETAIN).

switched fabric

One of three a fibre-channel connection topologies supported by the ESS. See arbitrated loop and point-to-point.

symmetric multi-processor (SMP)

An implementation of a multi-processor computer consisting of several identical processors configured in a way that any subset of the set of processors is capable of continuing the operation of the computer. The ESS contains four processors set up in SMP mode.

synchronous write

A write operation whose completion is indicated after the data has been stored on a storage device.

System/390

See S/390.

system adapter identification number (SAID)

System Management Interface Tool (SMIT)

An interface tool of the AIX operating system for installing, maintaining, configuring, and diagnosing tasks.

System Modification Program (SMP)

A program used to install software and software changes on MVS systems.

T

target

A SCSI device that acts as a slave to an initiator and consists of a set of one or more logical units, each with an assigned logical unit number (LUN). The logical units on the target are typically I/O devices. A SCSI target is analogous to an S/390 control unit. A SCSI initiator is analogous to an S/390 channel. A SCSI logical unit is analogous to an S/390 device. SEe Small Computer System Interface (SCSI).

TB

See terabyte.

TCP/IP

See Transmission Control Protocol/Internet Protocol.

terabyte (TB)

(1) Nominally, 1 000 000 000 000 bytes, which is accurate when speaking of bandwidth and disk storage capacity.

(2) For ESS cache memory, processor storage, real and virtual storage, a terabyte refers to 2⁴⁰ or 1 099 511 627 776 bytes.

thousands of power-on hours (KPOH)

A unit of time used to measure the mean time between failures (MTBF).

time sharing option (TSO)

An operating system option that provides interactive time sharing from remote terminals.

TPF

See transaction processing facility.

track

A unit of storage on a CKD device that can be formatted to contain a number of data records. See home address, track-descriptor record, and data record.

track-descriptor record (R0)

A special record on a track that follows the home address. The control program uses it to maintain certain information about the track. The record has a count field with a key length of zero, a data length of 8, and a record number of 0. This record is sometimes referred to as R0.

transaction processing facility (TPF)

A high-availability, high-performance IBM operating system, designed to support real-time, transaction-driven applications. The specialized architecture of TPF is intended to optimize system efficiency, reliability, and responsiveness for data communication and database processing. TPF provides real-time inquiry and updates to a large, centralized database, where message length is relatively short in both directions, and response time is generally less than three seconds. Formerly known as the Airline Control Program/Transaction Processing Facility (ACP/TPF).

Transmission Control Protocol/Internet Protocol (TCP/IP)

(1) The Transmission Control Protocol and the Internet Protocol, which together provide reliable end-to-end connections between applications over interconnected networks of different types.

(2) The suite of transport and application protocols that run over the Internet Protocol.

transparency

See software transparency.

TSO

See time sharing option.

U

UFS

UNIX filing system.

ultra-SCSI

An enhanced Small Computer System Interface.

unit address

The ESA/390 term for the address associated with a device on a given control unit. On ESCON or FICON interfaces, the unit address is the same as the device address. On OEMI interfaces, the unit address specifies a control unit and device pair on the interface.

unprotected volume

An AS/400 term that indicates that the AS/400 host recognizes the volume as an unprotected device, even though the storage resides on a RAID array and is therefore fault tolerant by definition. The data in an unprotected volume can be mirrored. Also referred to as an unprotected device.

UTC

See Coordinated Universal Time (UTC).

utility device

The ESA/390 term for the device used with the Extended Remote Copy facility to access information that describes the modifications performed on the primary copy.

V

virtual machine (VM)

A virtual data processing machine that appears to be for the exclusive use of a particular user, but whose functions are accomplished by sharing the resources of a real data processing system.

vital product data (VPD)

Information that uniquely defines the system, hardware, software, and microcode elements of a processing system.

VM

See virtual machine.

volume

In Enterprise Systems Architecture/390, the information recorded on a single unit of recording medium. Indirectly, it can refer to the unit of recording medium itself. On a nonremovable-medium storage device, the term can also indirectly refer to the storage device associated with the volume. When multiple volumes are stored on a single storage medium transparently to the program, the volumes can be reffered to as logical volumes.

VPD

See vital product data.

W

Web Copy Services

See IBM TotalStorage Enterprise Storage Server Copy Services.

worldwide node name (WWNN)

A unique 64-bit identifier for a host containing a fibre-channel port. See worldwide port name (WWPN).

worldwide port name (WWPN)

A unique 64-bit identifier associated with a fibre-channel adapter port. It is assigned in an implementation- and protocol-independent manner.

write hit

A write operation in which the requested data is in the cache.

write penalty

The performance impact of a classical RAID 5 write operation.

WWPN

See worldwide port name.

X

XRC

See Extended Remote Copy.

xSeries

An IBM Series product that emphasizes architecture.

Z

zSeries

An IBM Series product that emphasizes near-zero downtime. See S/390 and zSeries.

zSeries storage

See S/390 and zSeries storage.

Numerics

2105

The machine number for the IBM Enterprise Storage Server (ESS). See IBM Enterprise Storage Server (ESS). 2105-100 is an ESS expansion rack.

3390

The machine number of an IBM disk storage system. The IBM Enterprise Storage Server (ESS), when interfaced to IBM S/390 or zSeries hosts, is set up to appear as one or more 3390 devices, with a choice of 3390-2, 3390-3, 3990-6, or 3390-9 track formats.

3990

The machine number of an IBM control unit.

7133

The machine number of an IBM disk storage system. The Model D40 and 020 drawers of the 7133 can be installed in the 2105-100 expansion rack of the IBM Enterprise Storage Server (ESS).

8-pack

See disk drive module group.