59Y7292 - IBM DS Storage Manager v10 Installation and Host Support Guide

IBM System Storage DS Storage Manager Version 10

Installation and Host Support Guide

��

NoteBefore using this information and the product it supports, read the information in “Notices” on page M-1.

Seventh Edition, May 2010

© Copyright IBM Corporation 2009, 2010.US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contractwith IBM Corp.

Contents

Figures . . . . . . . . . . . . . . vii

Tables . . . . . . . . . . . . . . . ix

About this document . . . . . . . . . xiOverview . . . . . . . . . . . . . . . xiWho should read this document . . . . . . . xiNotices . . . . . . . . . . . . . . . . xiTerms to know . . . . . . . . . . . . . xiStorage Manager online help and diagnostics . . . xiiReceiving product updates and support notifications xiiGetting information, help, and service . . . . . xiii

Before you call . . . . . . . . . . . . xiiiUsing the documentation . . . . . . . . xiiiFinding Storage Manager software, controllerfirmware, and README files . . . . . . . xiiiIBM System Storage Productivity Center . . . xivEssential Web sites for support information . . xivSoftware service and support . . . . . . . xvHardware service and support . . . . . . . xv

Chapter 1. Preparing for installation 1-1Introduction . . . . . . . . . . . . . . 1-1

The DS Storage Manager software . . . . . 1-1DS Storage Manager Software Components . . 1-1Supported controller firmware . . . . . . . 1-2

Types of installation configurations . . . . . . 1-2Network configuration . . . . . . . . . 1-2

Reviewing a sample network . . . . . . 1-3The storage management station . . . . . 1-4

Direct and SAN-attached configurations . . . 1-5Creating a direct-attached configuration . . 1-5Creating a SAN-attached configuration . . . 1-5

Setting up controller addresses for softwareinstallation . . . . . . . . . . . . . . 1-6

Setting up IP addresses for storage controllers 1-6Setting up the DHCP/BOOTP server andnetwork . . . . . . . . . . . . . 1-6Steps for assigning static TCP/IP addresses tothe DS3000, DS4000, or DS5000 . . . . . 1-7

Chapter 2. Introducing the software 2-1Enterprise Management Window . . . . . . . 2-1Subsystem Management Window . . . . . . . 2-4

Chapter 3. Installing Storage Managerand Support Monitor . . . . . . . . 3-1Pre-installation requirements . . . . . . . . 3-1Installation requirements. . . . . . . . . . 3-3Installing DS Storage Manager and SupportMonitor packages automatically using theinstallation wizard . . . . . . . . . . . . 3-3

Installing Support Monitor using a consolewindow . . . . . . . . . . . . . . 3-5

Installing Storage Manager packages manually . . 3-6Software installation sequence . . . . . . . 3-6

Manual installation requirements . . . . . 3-6Uninstalling DS Storage Manager and SupportMonitor . . . . . . . . . . . . . . 3-7

Uninstalling DS Storage Manager andSupport Monitor on a Windows operatingsystem . . . . . . . . . . . . . . 3-7Uninstalling DS Storage Manager andSupport Monitor on a Unix-type operatingsystem . . . . . . . . . . . . . . 3-8

Completing the Storage Manager installation . . . 3-8Performing an automatic discovery of storagesubsystems . . . . . . . . . . . . . 3-8Performing a manual discovery of storagesubsystems . . . . . . . . . . . . . 3-9Storage subsystem password protection . . . 3-10Naming storage subsystems . . . . . . . 3-10Setting up alert notifications . . . . . . . 3-10Managing iSCSI settings . . . . . . . . 3-11

Changing target authentication . . . . . 3-12Entering mutual authentication permissions 3-13Changing target identification . . . . . 3-13Changing target discovery . . . . . . . 3-13Configuring iSCSI host ports . . . . . . 3-13Viewing or ending an iSCSI session. . . . 3-13Viewing iSCSI statistics . . . . . . . . 3-13iSNS best practices . . . . . . . . . 3-13Using DHCP . . . . . . . . . . . 3-13Using supported hardware initiators . . . 3-14Using IPv6 . . . . . . . . . . . . 3-14Network settings . . . . . . . . . . 3-14Maximum Transmission Unit settings . . . 3-15Microsoft iSCSI Software Initiatorconsiderations . . . . . . . . . . . 3-15

Downloading controller firmware, NVSRAM,ESM firmware . . . . . . . . . . . . 3-15

Determining firmware levels . . . . . . 3-16Downloading controller or NVSRAMfirmware . . . . . . . . . . . . 3-17Downloading ESM firmware . . . . . . 3-18

Downloading drive firmware . . . . . . . 3-19Downloading Storage Manager drivefirmware . . . . . . . . . . . . 3-19

DS Storage Manager premium features . . . 3-20Enabling premium features . . . . . . 3-21Saving the storage subsystem profile . . . 3-23

Chapter 4. Configuring storage . . . . 4-1Storage partitioning overview . . . . . . . . 4-2Using the Task Assistant . . . . . . . . . . 4-3Configuring hot-spare devices . . . . . . . . 4-4Creating arrays and logical drives . . . . . . 4-4

Creating an array . . . . . . . . . . . 4-5Redundant array of independent disks (RAID) 4-5

© Copyright IBM Corp. 2009, 2010 iii

Creating a logical drive . . . . . . . . . 4-8Defining the default host type . . . . . . . . 4-9Defining a host group . . . . . . . . . . 4-10

Steps for defining a host group . . . . . . 4-10Defining heterogeneous hosts . . . . . . . 4-11

Steps for defining the host and host port . . . . 4-12Mapping LUNs to a storage partition . . . . . 4-12

Mapping LUNs to a new partition . . . . . 4-12Adding LUNs to an existing partition . . . . 4-13

Configuring the IBM Systems Storage DS5100 andDS5300 for IBM i . . . . . . . . . . . . 4-13Optional premium features . . . . . . . . 4-15

Creating a FlashCopy logical drive . . . . . 4-16Using VolumeCopy . . . . . . . . . . 4-16Using the Remote Mirror option . . . . . . 4-16Drive security with Full Disk Encryption . . . 4-16

Other features . . . . . . . . . . . . . 4-18Controller cache memory . . . . . . . . 4-18Persistent Reservations . . . . . . . . . 4-19Media scan . . . . . . . . . . . . . 4-19

Errors reported by a media scan . . . . . 4-20Media scan settings . . . . . . . . . 4-21Media scan duration . . . . . . . . . 4-22

Chapter 5. Configuring hosts . . . . . 5-1Booting a host operating system using SAN boot 5-1Using multipath drivers to monitor I/O activity . . 5-3

Steps for installing the multipath driver . . . 5-6Windows MPIO or MPIO/DSM . . . . . 5-6Storport miniport HBA device driver . . . 5-6SCSIport miniport HBA device driver . . . 5-7Veritas DMP DSM driver . . . . . . . 5-7

AIX multipath drivers . . . . . . . . . 5-7Linux MPP driver . . . . . . . . . . . 5-8Veritas DMP driver . . . . . . . . . . 5-9HP-UX PV-links . . . . . . . . . . . 5-10

Using PV-links: Method 1 . . . . . . . 5-10Using PV-links: Method 2 . . . . . . . 5-12HP-UX native multipathing . . . . . . 5-15

Solaris failover drivers . . . . . . . . . 5-15Installing the MPxIO driver . . . . . . 5-15Installing the RDAC driver on Solaris . . . 5-22Installing the DMP driver . . . . . . . 5-24

Identifying devices . . . . . . . . . . . 5-27Using the SMdevices utility . . . . . . . 5-27

Using SMdevices on Windows operatingsystems . . . . . . . . . . . . . 5-27Using SMdevices on UNIX-based operatingsystems . . . . . . . . . . . . . 5-27

Identifying devices on AIX hosts. . . . . . 5-28Performing initial device discovery . . . . 5-28Initial discovery with MPIO . . . . . . 5-30

Configuring devices . . . . . . . . . . . 5-30Using the hot_add utility . . . . . . . . 5-30Using the SMrepassist utility . . . . . . . 5-30Stopping and restarting the host-agent software 5-31

Windows 2000 . . . . . . . . . . . 5-31Windows Server 2003 and 2008 . . . . . 5-31

Setting the queue depth for hdisk devices . . 5-31Calculating maximum queue depth . . . . 5-31Changing the queue depth for Windows 5-32

Changing the queue depth for AIX . . . . 5-32Steps for disabling cache mirroring . . . . . 5-33Using dynamic capacity expansion anddynamic volume expansion . . . . . . . 5-33

Performing a dynamic capacity expansionoperation . . . . . . . . . . . . 5-33Performing a dynamic volume expansionoperation . . . . . . . . . . . . 5-33

Veritas Storage Foundation with SUSE LinuxEnterprise Server . . . . . . . . . . . 5-34Veritas Storage Foundation 5.0 with Red HatEnterprise Linux . . . . . . . . . . . 5-34Checking LUN size . . . . . . . . . . 5-35Redistributing logical drives . . . . . . . 5-35

Redistributing logical drives on AIX . . . 5-35Redistributing logical drives on HP-UX . . 5-36Redistributing logical drives on Solaris . . 5-37

Resolving disk array errors on AIX . . . . . 5-37Replacing hot swap HBAs . . . . . . . . 5-39

Known issues and restrictions for AIX . . . 5-39Preparing for the HBA hot swap for AIX 5-40Replacing the hot swap HBA for AIX . . . 5-42Replacing IBM host bus adapters on a Linuxoperating system . . . . . . . . . . 5-44Replacing a PCI Hotplug HBA . . . . . 5-46Mapping the new WWPN to the DS3000,DS4000, or DS5000 Storage Subsystem forAIX and Linux. . . . . . . . . . . 5-48Completing the HBA hot swap procedure 5-48

Chapter 6. Working with full diskencryption . . . . . . . . . . . . . 6-1FDE disk drives . . . . . . . . . . . . 6-1

Securing data against a breach . . . . . . . 6-1Creating a security key . . . . . . . . 6-2Changing a security key . . . . . . . . 6-4Security key identifier . . . . . . . . 6-4Unlocking secure drives . . . . . . . 6-10

Secure erase . . . . . . . . . . . . 6-10FDE security authorizations . . . . . . . . 6-11FDE key terms. . . . . . . . . . . . . 6-12Configuring DS5000 disk encryption with FDEdrives . . . . . . . . . . . . . . . 6-13

Installing FDE drives . . . . . . . . . 6-13Enabling the DS5000 full disk encryptionfeature . . . . . . . . . . . . . . 6-14Securing a RAID array . . . . . . . . . 6-16Unlocking disk drives . . . . . . . . . 6-21Migrating disk drives . . . . . . . . . 6-23Erasing disk drives . . . . . . . . . . 6-27Global hot-spare disk drives . . . . . . . 6-30Log files . . . . . . . . . . . . . . 6-31

Frequently asked questions . . . . . . . . 6-31Securing arrays . . . . . . . . . . . 6-31Secure erase . . . . . . . . . . . . 6-32Security keys and pass phrases . . . . . . 6-32Premium features . . . . . . . . . . . 6-33Global hot-spare drives . . . . . . . . . 6-33Boot support . . . . . . . . . . . . 6-33Locked and unlocked states . . . . . . . 6-33Backup and recovery . . . . . . . . . 6-33

iv IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Other . . . . . . . . . . . . . . . 6-34

Chapter 7. Configuring and usingSupport Monitor . . . . . . . . . . 7-1Overview of the Support Monitor interface . . . 7-1Scheduling collection of the support bundle . . . 7-3Sending the support bundle to IBM Support . . . 7-3Collecting the support bundle manually . . . . 7-5Using the Support Monitor log window . . . . 7-6Solving Support Monitor problems . . . . . . 7-8

Appendix A. Using the IBM SystemStorage DS3000, DS4000, and DS5000Controller Firmware Upgrade Tool . . A-1Tool overview . . . . . . . . . . . . . A-1Checking the device health conditions . . . . . A-1

Using the upgrade tool . . . . . . . . . A-2Adding a storage subsystem . . . . . . . . A-2Downloading the firmware. . . . . . . . . A-2Viewing the IBM System Storage DS3000, DS4000,and DS5000 Controller Firmware Upgrade Tool logfile . . . . . . . . . . . . . . . . . A-3

Appendix B. Host bus adaptersettings . . . . . . . . . . . . . . B-1Setting host bus adapters . . . . . . . . . B-1

Accessing HBA settings through Fast!UTIL . . B-1Accessing host bus adapter settings . . . . . B-1Advanced Adapter Settings. . . . . . . . B-2

QLogic host bus adapter settings . . . . . . . B-3JNI and QLogic host bus adapter settings . . . . B-8

JNI HBA card settings . . . . . . . . . B-8Configuration settings forFCE-1473/FCE-6460/FCX2-6562/FCC2-6562 . B-8Configuration settings forFCE-1063/FCE2-1063/FCE-6410/FCE2-6410 . B-9Configuration settings for FCI-1063. . . . B-10Configuration settings for FC64-1063 . . . B-11

QLogic HBA card settings . . . . . . . . B-12Connecting HBAs in an FC switch environment B-13

Appendix C. Using a DS3000, DS4000,or DS5000 with a VMware ESX Serverconfiguration . . . . . . . . . . . C-1Sample configuration. . . . . . . . . . . C-1Software requirements . . . . . . . . . . C-1

Management station . . . . . . . . . . C-2Host (VMware ESX Server). . . . . . . . C-2

Hardware requirements . . . . . . . . . . C-2VMware ESX Server restrictions . . . . . . . C-3Other VMware ESX Server host information . . . C-4Configuring storage subsystems for VMware ESXServer . . . . . . . . . . . . . . . . C-4

Cross connect configuration for VMwareconnections . . . . . . . . . . . . . C-4Notes on mapping LUNs to a storage partition C-5Steps for verifying the storage configuration forVMware . . . . . . . . . . . . . . C-5

Appendix D. Using DS StorageManager with high-availability clusterservices. . . . . . . . . . . . . . D-1General information . . . . . . . . . . . D-1Using cluster services on AIX systems . . . . . D-1

High Availability Cluster Multi-Processing . . D-1Software requirements . . . . . . . . D-2Configuration limitations . . . . . . . D-2Other HACMP usage notes . . . . . . D-3

Parallel System Support Programs and GeneralParallel File System . . . . . . . . . . D-3

Software requirements . . . . . . . . D-3Configuration limitations . . . . . . . D-3Other PSSP and GPFS usage notes . . . . D-3

GPFS, PSSP, and HACMP cluster configurationdiagrams . . . . . . . . . . . . . . D-3

Using cluster services on HP-UX systems . . . . D-9Using cluster services on Solaris systems. . . . D-10

General Solaris requirements . . . . . . . D-10System dependencies . . . . . . . . . D-10

RDAC IDs . . . . . . . . . . . . D-10Single points of failure . . . . . . . . D-10

Appendix E. Viewing and setting AIXObject Data Manager (ODM) attributes. E-1Attribute definitions . . . . . . . . . . . E-1Using the lsattr command to view ODM attributes E-5

Appendix F. DS Diagnostic DataCapture (DDC) . . . . . . . . . . . F-1DDC information . . . . . . . . . . . . F-1

DDC function implementation . . . . . . . F-1How Diagnostic Data Capture works . . . F-1Recovery steps . . . . . . . . . . . F-1

DDC MEL events . . . . . . . . . . . . F-3

Appendix G. The Script Editor . . . . G-1Using the Script Editor . . . . . . . . . . G-2Adding comments to a script . . . . . . . . G-2

Appendix H. Tuning storagesubsystems . . . . . . . . . . . . H-1Load balancing. . . . . . . . . . . . . H-1Balancing the Fibre Channel I/O load . . . . . H-2Optimizing the I/O transfer rate . . . . . . . H-3Optimizing the Fibre Channel I/O request rate H-3

Determining the Fibre Channel I/O accesspattern and I/O size . . . . . . . . . . H-3Enabling write-caching . . . . . . . . . H-3Optimizing the cache-hit percentage . . . . H-3Choosing appropriate RAID levels . . . . . H-4Choosing an optimal logical-drive modificationpriority setting . . . . . . . . . . . . H-4Choosing an optimal segment size . . . . . H-4Defragmenting files to minimize disk access H-5

Appendix I. Critical event problemsolving . . . . . . . . . . . . . . I-1

Contents v

Appendix J. Additional SystemStorage DS documentation. . . . . . J-1DS Storage Manager Version 10 library . . . . . J-1DS5100 and DS5300 Storage Subsystem library . . J-2DS5020 Storage Subsystem library. . . . . . . J-2DS4800 Storage Subsystem library. . . . . . . J-3DS4700 Storage Subsystem library. . . . . . . J-4DS4500 Storage Subsystem library. . . . . . . J-4DS4400 Storage Subsystem library. . . . . . . J-5DS4300 Storage Subsystem library. . . . . . . J-5DS4200 Express Storage Subsystem library . . . . J-6DS4100 Storage Subsystem library. . . . . . . J-6DS3500 Storage Subsystem library. . . . . . . J-7DS3400 Storage Subsystem library. . . . . . . J-7DS3300 Storage Subsystem library. . . . . . . J-7DS3200 Storage Subsystem library. . . . . . . J-8DS5000 Storage Expansion Enclosure documents . . J-8DS5020 Storage Expansion Enclosure documents . . J-9DS3950 Storage Expansion Enclosure documents J-10DS4000 Storage Expansion Enclosure documents J-10Other DS and DS-related documents . . . . . J-11

Appendix K. Accessibility . . . . . . K-1

Appendix L. FDE best practices . . . L-1Physical asset protection . . . . . . . . . . L-1Data backup. . . . . . . . . . . . . . L-1FDE drive security key and the security key file L-2DS subsystem controller shell remote login . . . L-3Working with FDE drives . . . . . . . . . L-3Replacing controllers . . . . . . . . . . . L-3Storage industry standards and practices . . . . L-4

Notices . . . . . . . . . . . . . . M-1Trademarks . . . . . . . . . . . . . . M-2Important notes . . . . . . . . . . . . M-2Particulate contamination . . . . . . . . . M-3

Glossary . . . . . . . . . . . . . N-1

Index . . . . . . . . . . . . . . . X-1

vi IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Figures

1-1. Sample network using network managedand host-agent managed storage subsystems 1-3

2-1. Parts of the Enterprise ManagementWindow. . . . . . . . . . . . . 2-2

2-2. Parts of the Subsystem ManagementWindow. . . . . . . . . . . . . 2-6

3-1. Manage iSCSI settings. . . . . . . . 3-124-1. Assigning a port identifier for IBM i 4-144-2. Selecting IBM i as the host type 4-155-1. Host HBA to storage subsystem controller

multipath sample configuration for allmultipath drivers except AIX fcp_array andSolaris RDAC . . . . . . . . . . . 5-5

5-2. Host HBA to storage subsystem controllermultipath sample configuration for AIXfcp_array and Solaris RDAC multipathdrivers . . . . . . . . . . . . . 5-5

6-1. Security-enabled FDE drives: With thecorrect authorizations in place, the readingand writing of data occurs in Unlockedstate . . . . . . . . . . . . . . 6-3

6-2. A security-enabled FDE drive is removedfrom the storage subsystem: Without correctauthorizations, a stolen FDE disk cannot beunlocked, and the data remains encrypted . 6-4

6-3. Changing the security key. . . . . . . 6-56-4. Changing the security key - Complete 6-66-5. Drive properties - Secure FDE drive 6-76-6. Select file - LockKeyID . . . . . . . . 6-86-7. Drive properties - Unsecured FDE drive 6-9

6-8. Secure erase process . . . . . . . . 6-117-1. Console area . . . . . . . . . . . 7-2B-1. One-to-one zoning scheme . . . . . . B-14B-2. One-to-two zoning scheme . . . . . . B-14C-1. Sample VMware ESX Server configuration C-1C-2. Cross connect configuration for VMware

connections . . . . . . . . . . . C-5D-1. Cluster configuration with single DS3000,

DS4000, or DS5000 storage subsystem—oneto four partitions . . . . . . . . . D-4

D-2. Cluster configuration with three DS3000,DS4000, or DS5000 storagesubsystems—one partition per DS3000,DS4000, or DS5000 . . . . . . . . . D-5

D-3. Cluster configuration with four DS3000,DS4000, or DS5000 storagesubsystems—one partition per DS3000,DS4000, or DS5000 . . . . . . . . . D-6

D-4. RVSD cluster configuration with twoDS3000, DS4000, or DS5000 storagesubsystems—two partitions per DS3000,DS4000, or DS5000 . . . . . . . . . D-7

D-5. HACMP/GPFS cluster configuration withone DS3000, DS4000, or DS5000 storagesubsystem—one partition . . . . . . . D-8

D-6. HACMP/GPFS cluster configuration withtwo DS3000, DS4000, or DS5000 storagesubsystems—two partitions per DS3000,DS4000, or DS5000 . . . . . . . . . D-9

G-1. The Script Editor window . . . . . . G-1

© Copyright IBM Corp. 2009, 2010 vii

viii IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Tables

2-1. Parts of the Enterprise ManagementWindow. . . . . . . . . . . . . 2-2

2-2. Data shown in the Table View . . . . . 2-32-3. Parts of the Subsystem Management

Window. . . . . . . . . . . . . 2-62-4. Nodes in the Logical pane. . . . . . . 2-82-5. Controller status icons . . . . . . . . 2-92-6. Drive enclosure type icons . . . . . . 2-92-7. Types of nodes in the Topology view 2-102-8. Node information in the Defined Mappings

pane . . . . . . . . . . . . . 2-112-9. Node information by type of node 2-113-1. Storage Monitor-compatible subsystems and

controller firmware . . . . . . . . . 3-23-2. Installation sequence of DS Storage Manager

software packages . . . . . . . . . 3-63-3. Storage Manager package install commands 3-73-4. Storage Manager package installation verify

commands . . . . . . . . . . . . 3-74-1. Parts of the Enterprise Management

Window. . . . . . . . . . . . . 4-24-2. RAID level configurations . . . . . . . 4-64-3. Array Security Properties . . . . . . 4-174-4. Errors discovered during a media scan 4-215-1. Multipath driver by operating system 5-45-2. Number of paths each multipath driver

supports by operating system . . . . . 5-65-3. Sample SMdevices command output

(method 2) . . . . . . . . . . . 5-125-4. Sample record of logical drive preferred

and alternate paths. . . . . . . . . 5-136-1. Security authorizations . . . . . . . 6-116-2. Full disk encryption key terms . . . . . 6-126-3. DS5000 supported FDE drives . . . . . 6-137-1. Support Monitor icon meanings . . . . . 7-27-2. Support Monitor messages and descriptions 7-67-3. Problem index. . . . . . . . . . . 7-8B-1. Qlogic model QLA234x, QLA24xx,

QLE2462, QLE2460, QLE2560, QLE2562 . . B-3B-2. QLogic model QL220x (for BIOS V1.81) host

bus adapter settings by operating system . B-6B-3. Configuration settings for

FCE-1473/FCE-6460/FCX2-6562/FCC2-6562. B-8B-4. Configuration settings for

FCE-1063/FCE2-1063/FCE-6410/FCE2-6410 . B-9B-5. Configuration settings for FCI-1063 B-11B-6. Configuration settings for FC64-1063 B-12B-7. Configuration settings for QL2342 B-13E-1. Attributes for dar devices . . . . . . . E-1E-2. Attributes for dac devices . . . . . . . E-2E-3. Attributes for hdisk devices . . . . . . E-3E-4. Example 1: Displaying the attribute settings

for a dar . . . . . . . . . . . . E-5

E-5. Example 2: Displaying the attribute settingsfor a dac . . . . . . . . . . . . E-5

E-6. Example 3: Displaying the attribute settingsfor an hdisk . . . . . . . . . . . E-6

F-1. Recovery Step 2 . . . . . . . . . . F-2F-2. Recovery Step 4 . . . . . . . . . . F-2F-3. Recovery Step 5 . . . . . . . . . . F-2H-1. Performance Monitor tuning options in the

Subsystem Management window . . . . H-1H-2. Load balancing policies supported by

operating systems . . . . . . . . . H-1I-1. Critical events . . . . . . . . . . . I-1J-1. DS Storage Manager Version 10 titles by

user tasks . . . . . . . . . . . . J-1J-2. DS5100 and DS5300 Storage Subsystem

document titles by user tasks . . . . . . J-2J-3. DS5020 Storage Subsystem document titles

by user tasks . . . . . . . . . . . J-2J-4. DS4800 Storage Subsystem document titles





by user tasks . . . . . . . . . . . J-5J-9. DS4200 Express Storage Subsystem

document titles by user tasks . . . . . . J-6J-10. DS4100 Storage Subsystem document titles





by user tasks . . . . . . . . . . . J-8J-15. DS5000 Storage Expansion Enclosure

document titles by user tasks . . . . . . J-8J-16. DS5020 Storage Expansion Enclosure

document titles by user tasks . . . . . . J-9J-17. DS3950 Storage Expansion Enclosure

document titles by user tasks . . . . . J-10J-18. DS4000 Storage Expansion Enclosure

document titles by user tasks . . . . . J-10J-19. DS4000 and DS4000–related document titles

by user tasks . . . . . . . . . . . J-11K-1. DS3000 and DS4000 Storage Manager

alternate keyboard operations . . . . . K-1M-1. Limits for particulates and gases M-3

© Copyright IBM Corp. 2009, 2010 ix

x IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

About this document

Throughout this document, Storage Manager refers to all host software release levels.

This document provides information about how to plan, install, configure, and work with IBM® SystemStorage® DS Storage Manager.

Important: Check the Storage Manager README files for any updates to the list of supported operatingsystems.

See “Finding Storage Manager software, controller firmware, and README files” on page xiii to find outhow to access the most recent Storage Manager README files on the Web.

OverviewUse this document to perform the following tasks:v Determine the hardware and software that you will require to install the storage management software.v Integrate the necessary hardware components into your network.v Install the DS Storage Manager software.v Upgrade controller firmware, if necessary.v Identify storage management features that are unique to your installation.

Who should read this documentThis document is intended for system and storage administrators who are responsible for installingstorage administration software. Readers should have knowledge of Redundant Array of IndependentDisks (RAID), Small Computer System Interface (SCSI), Fibre Channel, and SATA technology. Theyshould also have working knowledge of the applicable operating systems that are used with themanagement software.

Notices

This document contains the following notices, which are designed to highlight key information:v Note: These notices provide important tips, guidance, or advice.v Important: These notices provide information that might help you avoid inconvenient or problem

situations.v Attention: These notices indicate possible damage to programs, devices, or data. An attention notice is

placed just before the instruction or situation in which damage could occur.

Terms to know

For information on terminology, see the Help section of the Storage Manager Enterprise ManagementWindow, the Subsystem Management Window, or the “Glossary” on page N-1.

It is important to understand the distinction between the following two terms when you read thisdocument.

Management stationA management station is a system that is used to manage the storage subsystem. You can attachit to the storage subsystem in either of the following ways:

© Copyright IBM Corp. 2009, 2010 xi

v Through a TCP/IP Ethernet connection to the controllers in the storage subsystemv Through a TCP/IP connection to the host-agent software that is installed on a host computer,

which in turn is either directly attached to the storage subsystem through the fibre-channel I/Opath or through a TCP/IP Ethernet connection to the controllers

Host computerA host computer is a system that is directly attached to the storage subsystem through afibre-channel I/O path. This system is used to perform the following tasks:v Serve data (typically in the form of files) from the storage subsystemv Function as a connection point to the storage subsystem for a remote-management station

Note:

1. The terms host and host computer are used interchangeably throughout this document.2. A host computer can also function as a management station.

Storage Manager online help and diagnostics

You can access the help systems from the Enterprise Management and Subsystem ManagementWindows® in DS Storage Manager by clicking Help on the toolbar or pressing F1.

Enterprise Management Help windowUse this online help system to learn more about working with the entire management domain.

Subsystem Management Help windowUse this online help system to learn more about managing individual storage subsystems.

After you install IBM DS Storage Manager, consider installing the HBA management and diagnosticapplication if available. The QLogic SANsurfer and Emulex HBAnyware applications are diagnosticprograms that you can use to verify the status of the I/O connections before you use the storagesubsystem.

Note: If your storage subsystem is connected to a Fibre Channel host bus adapter (HBA) in the hostserver in a SAN environment, consider purchasing the IBM Tivoli® Storage Manager software applicationfor SAN management and troubleshooting.

Receiving product updates and support notifications

Be sure to download the latest versions of the following packages at the time of initial installation andwhen product updates become available:v DS Storage Manager host softwarev Storage subsystem controller firmwarev Drive expansion enclosure ESM firmwarev Drive firmware

Important: Keep your system up-to-date with the latest firmware and other product updates bysubscribing to receive support notifications.

For more information about how to register for support notifications, see the following IBM Support Webpage and click on My notifications:

http://www.ibm.com/systems/support

You can also check the Stay Informed section of the IBM Disk Support Web site, at the following address:

xii IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

http://www.ibm.com/systems/support

http://www.ibm.com/systems/support/storage/disk

Getting information, help, and service

If you need help, service, or technical assistance or just want more information about IBM products, youwill find a wide variety of sources available from IBM to assist you. This section contains informationabout where to go for additional information about IBM and IBM products, what to do if you experiencea problem with your system, and whom to call for service, if it is necessary.

Before you call

Before you call, take these steps to try to solve the problem yourself:v Check all cables to make sure that they are connected.v Check the power switches to make sure that the system is turned on.v Use the troubleshooting information in your system documentation, and use the diagnostic tools that

come with your system.v Check for technical information, hints, tips, and new device drivers at the IBM System Storage Disk

Support Web site pages that are listed in this section.v Use an IBM discussion forum on the IBM Web site to ask questions.

You can solve many problems without outside assistance by following the troubleshooting proceduresthat IBM provides in the DS Storage Manager online help or in the documents that are provided withyour system and software. The information that comes with your system also describes the diagnostictests that you can perform. Most subsystems, operating systems, and programs come with informationthat contains troubleshooting procedures and explanations of error messages and error codes. If yoususpect a software problem, see the information for the operating system or program.

Using the documentation

Information about your IBM system and preinstalled software, if any, is available in the documents thatcome with your system; this includes printed books, online documents, README files, and help files. Seethe troubleshooting information in your system documentation for instructions for using the diagnosticprograms. The troubleshooting information or the diagnostic programs might tell you that you needadditional or updated device drivers or other software.

Finding Storage Manager software, controller firmware, and READMEfiles

DS Storage Manager software and controller firmware versions are available on the product CD and canalso be downloaded from the Web.

Important: Before you install DS Storage Manager software, consult the README. Updated READMEfiles contain the latest device driver versions, firmware levels, limitations, and other information notfound in this document.

Storage Manager README files are found on the Web, at the following address:

http://www.ibm.com/systems/support/storage/disk1. On the Support for IBM System Storage and TotalStorage products page, from the Product family

drop-down menu, select Disk systems. From the Product drop-down menu, select your product (forexample, DS5100 Midrange Disk System). Click Go.

About this document xiii



2. In the Support & downloads box, again click Download. The Software and device drivers pageopens.

3. In the Storage Manager section of the table, locate your operating system and version level (forexample, IBM DS5000 Storage Manager v10.xx.xx.xx for AIX - IBM System Storage), and click onthe version link in the right-hand column. The DS5000 Storage Manager download page opens.

4. On the download page, in the table under File details, click on the *.txt file link, and the READMEwill open in your Web browser.

IBM System Storage Productivity Center

The IBM System Storage Productivity Center (SSPC) is an integrated hardware and software solution thatprovides a single point of entry for managing IBM System Storage DS3000 systems, DS4000® systems,DS5000 systems, DS8000® systems, IBM System Storage SAN Volume Controller clusters, and othercomponents of your data storage infrastructure. Therefore, you can use the IBM System StorageProductivity Center to manage multiple IBM System Storage product configurations from a singlemanagement interface.

To learn how to incorporate the DS Storage Manager with the IBM System Storage Productivity Center,see the IBM System Storage Productivity Center Information Center at the following Web site:

publib.boulder.ibm.com/infocenter/tivihelp/v4r1/index.jsp

Note: Support for the DS Storage Manager element of the IBM System Storage Productivity Centerbegins June 2008.

Essential Web sites for support information

The most up-to-date information about your IBM storage subsystems and DS Storage Manager, includingdocumentation and the most recent software, firmware, and NVSRAM downloads, can be found at thefollowing Web sites:

IBM System Storage Disk Storage SystemsFind links to software and firmware downloads, READMEs, and support pages for all IBMSystem Storage disk storage systems:


IBM System Storage Interoperation Center (SSIC)Find technical support information for your specific storage subsystem/host configuration,including the latest firmware versions for your system, by using this interactive Web-basedutility:

http://www.ibm.com/systems/support/storage/config/ssic

IBM DS3000, DS4000, DS5000, and BladeCenter® Boot Disk System Premium Feature ActivationActivate a premium feature by using this Web-based utility:

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

IBM System Storage Productivity CenterFind the latest documentation supporting the IBM System Storage Productivity Center, a newsystem that is designed to provide a central management console for IBM System StorageDS3000, DS4000, DS5000, DS8000, and SAN Volume Controller:

publib.boulder.ibm.com/infocenter/tivihelp/v4r1/index.jsp

IBM System Storage SupportFind the latest support information for host operating systems, HBAs, clustering, storage areanetworks (SANs), DS Storage Manager software and controller firmware:

xiv IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

http://publib.boulder.ibm.com/infocenter/tivihelp/v4r1/index.jsp



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

http://publib.boulder.ibm.com/infocenter/tivihelp/v4r1/index.jsp

www.ibm.com/systems/support/storage

Storage Area Network (SAN) SupportFind information about using SAN switches, including links to SAN user guides and otherdocuments:

www.ibm.com/systems/support/storage/san

Support for IBM System p® AIX 5L™ and Linux® serversFind the latest support information for System p AIX®, Linux, BladeCenter, and i5/OS® servers:

www.ibm.com/systems/support/supportsite.wss/brandmain?brandind=5000025

Support for IBM System x® serversFind the latest support information for System x Intel®- and AMD-based servers:

http://www.ibm.com/systems/support/

eServer™ System p and AIX Information CenterFind everything you need to know about using AIX with System p and POWER® servers:

publib.boulder.ibm.com/infocenter/pseries/index.jsp?

IBM System Storage productsFind information about all IBM System Storage products:

www.ibm.com/systems/storage

IBM Publications CenterFind IBM publications:

www.ibm.com/shop/publications/order/

Software service and support

Through IBM Support Line, for a fee you can get telephone assistance with usage, configuration, andsoftware problems. For information about which products are supported by Support Line in your countryor region, go to the following Web site:

www.ibm.com/services/sl/products

For more information about the IBM Support Line and other IBM services, go to the following Web sites:v www.ibm.com/servicesv www.ibm.com/planetwide

Hardware service and support

You can receive hardware service through IBM Integrated Technology Services or through your IBMreseller, if your reseller is authorized by IBM to provide warranty service. Go to the following Web sitefor support telephone numbers:

www.ibm.com/planetwide

In the U.S. and Canada, hardware service and support is available 24 hours a day, 7 days a week. In theU.K., these services are available Monday through Friday, from 9 a.m. to 6 p.m.

About this document xv

http://www.ibm.com/systems/support/storage

http://www.ibm.com/systems/support/storage/san/

http://www.ibm.com/systems/support/supportsite.wss/brandmain?brandind=5000025

http://www.ibm.com/systems/support/

http://publib.boulder.ibm.com/infocenter/pseries/index.jsp?

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

http://www.ibm.com/shop/publications/order/

http://www.ibm.com/services/sl/products

http://www.ibm.com/services

http://www.ibm.com/planetwide

http://www.ibm.com/planetwide

xvi IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Chapter 1. Preparing for installation

The following information is integral to helping you prepare for the successful installation of the storagemanagement software.v “The DS Storage Manager software”v “Supported controller firmware” on page 1-2v “Types of installation configurations” on page 1-2v “Setting up controller addresses for software installation” on page 1-6

IntroductionIBM System Storage DS® Storage Manager consists of a set of client and host tools that enable you tomanage the IBM DS3000, DS4000, and DS5000 Storage Subsystems from a storage management station.

DS3000, DS4000, and DS5000 Storage Manager is supported on the following operating systems:v AIXv Windows 2003 and Windows 2008v Linux (RHEL and SLES)v HP-UXv Solaris

The DS3000, DS4000, and DS5000 products are also supported when attached to NetWare, VMware ESXServer, and System p Virtual IO Server (VIOS) hosts, as well as on i5/OS as a guest client on VIOS.

Information about i5/OS support can be found at the following Web site:

www.ibm.com/systems/i/os/

For additional information, please refer to the System Storage Interoperation Center found at thefollowing Web site:

http://www.ibm.com/systems/support/storage/config/ssic.

The DS Storage Manager software

The DS Storage Manager software is used to configure, manage and troubleshoot the IBM System Storagesubsystem. It is used primarily to configure RAID arrays and logical drives, assign logical drives to hosts,replace and rebuild failed disk drives, expand the size of the arrays and logical drives, and convert fromone RAID level to another. It allows troubleshooting and management tasks, such as checking the statusof the storage subsystem components, updating the firmware of the RAID controllers, and managing thestorage subsystem. Finally, it offers advanced functions such as FlashCopy®, Volume Copy, and EnhancedRemote Mirroring.

For the latest firmware versions that are supported by each storage subsystem model, see the DSREADME file for your operating system.

DS Storage Manager Software Components

DS Storage Manager contains the following client software components. The following components mightvary depending on the operating system:

© Copyright IBM Corp. 2009, 2010 1-1

http://www-03.ibm.com/systems/i/os/


SMruntime softwareDS Storage Manager Java™ compiler

SMesm softwareDS Storage Manager ESM firmware delivery package

SMclient softwareDS Storage Manager client package

SMagent softwareDS Storage Manager agent package

SMutil softwareDS Storage Manager utility package

Storage Manager Profiler Support MonitorThe IBM DS Storage Manager Profiler Support Monitor tool is a component of IBM DS StorageManager version 10.60.x5.17 and later. In addition to the DS Storage Manager Profiler SupportMonitor code, the Apache Tomcat web server and MySQL database software packages areinstalled as part of the tool. For more information about the Support Monitor tool, see Chapter 7,“Configuring and using Support Monitor,” on page 7-1.

Supported controller firmware

All controller firmware versions are available free-of-charge.

To ensure the highest level of compatibility and error-free operation, ensure that the controller firmwarefor your storage subsystem is the latest firmware version for the storage subsystem model.

Important: DS4000 and DS5000 storage subsystems support Storage Manager version 10.50.xx.xx,controller firmware 5.41.xx.xx and later. Controller firmware versions earlier than 5.41.xx.xx are no longersupported or managed on these subsystems.

For detailed information explaining how to download the most current firmware version level, see“Downloading controller firmware, NVSRAM, ESM firmware” on page 3-15.

Types of installation configurations

A storage management station can be either of the following configurations:v A remote system, connected to an Ethernet network, that is used to manage one or more storage

subsystemsv A host that is connected to the storage subsystem with a Fibre Channel, iSCSI, or SAS input/output

(I/O) path that is also used to manage the attached storage subsystems

Network configurationBefore you begin installing the storage management software, ensure that the network components areset up and operating properly and that you have all the host and controller information necessary for thecorrect operation of the software.

Note: When connecting the storage subsystem to an Ethernet switch, set the switch port settings toautonegotiate.

1-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Reviewing a sample networkFigure 1-1 shows an example of a network that contains both a network managed storage subsystem(Network A) and a host-agent-managed storage subsystem (Network B).

Network managed storage subsystem: Network A is a network managed storage subsystem. Both themanagement station and the storage subsystem are connected to the network. Network A contains thefollowing components:v A DHCP/BOOTP serverv A network management station (NMS) for Simple Network Management Protocol (SNMP) trapsv A host that is connected to a storage subsystem through a fibre-channel I/O pathv A management station that is connected by an Ethernet cable to the storage subsystem controllers

Note: If the controller static TCP/IP addresses or default TCP/IP addresses are used, you do not need toset up the DHCP/BOOTP server.

Host-agent-managed storage subsystem: Network B is a host-agent-managed storage subsystem. Thestorage subsystem is not connected to the network. Network B contains the following components:v A host that is connected to a storage subsystem through a supported I/O pathv A management station that is connected by an Ethernet cable to the host computer

Figure 1-1. Sample network using network managed and host-agent managed storage subsystems

Chapter 1. Preparing for installation 1-3

The storage management stationThe storage management station is the system that is responsible for managing all, or a portion of, a storagenetwork. It communicates with the network management agents that reside in the managed nodes usinga network management protocol, such as Simple Network Management Protocol (SNMP).

Storage management commands are sent to the storage subsystem controllers, where the controllerfirmware validates and runs the commands, and then returns status and configuration information to theclient software.

Network-managed systems: The following steps provide an overview of the tasks involved in setting upthe network installation of a network managed (out-of-band) system:

Important: A maximum of eight storage management stations can concurrently monitor an out-of-bandmanaged storage array. This limit does not apply to systems that manage the storage array through thein-band management method.1. Install all hardware components (host computers, storage subsystems, and cables) that you want to

connect to the network. Refer to the installation guides for the specific hardware components.2. Establish a naming convention for the storage subsystems that will be connected to the network.3. Record the storage subsystem names and management types.

Note: Throughout the remaining steps, you will need to record such information as the hardwareEthernet and IP addresses.

4. Determine the hardware Ethernet address for each controller in storage subsystems connected to thenetwork.

5. For a network managed system only: If you are using a default controller IP address, go to step 7.Otherwise, obtain the TCP/IP address and host name for each of the controllers in the storagesubsystems on the network from the network administrator.

6. Set up the DHCP/BOOTP server to provide network configuration information for a specificcontroller. If you are using a controller static IP addresses, skip this step.

7. Verify that the TCP/IP software is installed.8. Set up the host or domain name server (DNS) table.9. Power on the devices that are connected to the network.

Host-agent-managed systems: The following steps provide an overview of the tasks involved in settingup a network installation of a host-managed (in-band) system:1. Install all hardware components (host computers, storage subsystems, and cables) that you want to

connect to the network. Refer to the installation guides for the specific hardware components.2. Establish a naming convention for the storage subsystems that will be connected to the network.3. Record the storage subsystem names and management types.

Note: Throughout the remaining steps, you will need to record such information as the hardwareEthernet and IP addresses.

4. Obtain the IP address and host name of the host computer on which the host-agent software will runfrom the network administrator.

Note: SMagent is part of the Storage Manager software package and is required on the host that isconnected to the storage subsystem through the fibre channel.

5. Verify that the TCP/IP software is installed.6. Power on the devices that are connected to the network.


Note: Even though you can install the storage management software on a host, the host still uses theTransmission Control Protocol/Internet Protocol (TCP/IP) to communicate with the host-agent. Thehost-agent communicates with the controller over the Fibre Channel connection through the access volume.

Direct and SAN-attached configurationsDS Storage Manager supports IBM Storage Subsystems in direct-attached configurations or in a SANenvironment through switches.

Creating a direct-attached configuration

Important: Storage subsystems with iSCSI ports do not support direct-attached connections from the hostsystems to the storage subsystem iSCSI ports.

Before you begin, verify that:v You can connect one or two servers to the DS3000, DS4000, and DS5000 Storage Subsystems.v No external hubs are being used.v Two-server DS4400 or DS4500 configurations require four host-side minihubs, with exactly one

connection from each HBA to a minihub.

Note:

1. Only the DS4400 and DS4500 Storage Subsystems have minihubs.2. See the Installation and User's Guide for your storage subsystem for more information.

Complete the following steps to set up a direct-attached configuration:1. Connect the HBAs to each controller (or minihub) port of the DS3000, DS4000, or DS5000 Storage

Subsystem.2. Use the DS Storage Manager automatic discovery feature to make sure that the subsystem is

discovered.

Creating a SAN-attached configurationA SAN-attached configuration can consist of Fibre Channel, SAS, or iSCSI connections.

If you use Fibre Channel HBAs in your SAN-attached configuration, the HBA and the storage subsystemhost port connections should be isolated in fabric zones to minimize the possible interactions between theports in a SAN fabric environment. Multiple storage subsystems can be configured to the same set ofHBAs through a Fibre Channel switch. For more information about Fibre Channel zoning schemes, see“Connecting HBAs in an FC switch environment” on page B-13.

Attention: A single-HBA configuration can lead to loss of access data in the event of a path failure. Ifyou have a single HBA in a SAN-attached configuration, both controllers in the DS3000, DS4000, orDS5000 subsystem must be connected to the HBA through a switch, and both controllers must be withinthe same SAN zone as the HBA.

Complete the following steps to set up a SAN-attached configuration:1. Connect the HBAs to the switch or switches.2. Connect the DS3000, DS4000, or DS5000 Storage Subsystems to the switch or switches.3. Set the required zoning or VLANs on the Fibre Channel switches or Ethernet switches, if applicable.4. Use the DS Storage Manager automatic discovery feature to make sure that the subsystem is

discovered.


Setting up controller addresses for software installationHow you plan to manage the storage subsystems determines where you must install the softwarecomponents. Before you can install software components, IP addresses must be assigned for the storagecontrollers.

Note: The controllers should be connected to a LAN port, set for rate auto-negotiate. The controllers donot function properly when connected to a switch port that is set for a hard rate.

Setting up IP addresses for storage controllersComplete the following procedures after you install SMruntime and SMclient, as described in theinstallation section for your host operating system.

You must set up a DHCP or BOOTP server and network with the following components:v A DHCP or BOOTP serverv A network management station (NMS) for Simple Network Management Protocol (SNMP) trapsv A host that is connected to a storage subsystem through a fibre-channel I/O pathv A management station that is connected by an Ethernet cable to the storage subsystem controllers

Note: You can avoid DHCP/BOOTP server and network tasks by assigning static IP addresses to thecontroller. If you do not wish to assign static TCP/IP addresses with the DS Storage Manager, using theDS3000, DS4000, or DS5000 default TCP/IP addresses, see the following IBM support Web site at:

www.ibm.com/systems/support/storage

Note: To manage storage subsystems through a firewall, configure the firewall to open port 2463 to TCPdata.

Setting up the DHCP/BOOTP server and networkComplete the following steps to set up the DHCP/BOOTP server and network:1. Get the MAC address from each controller blade. (See the “Identifying Ethernet MAC addresses”

procedure.)2. Complete whichever of the following steps is appropriate for your server:

v On a DHCP server, create a DHCP record for each of the MAC addresses. Set the lease duration tothe longest time possible.

v On a BOOTP server, edit the bootptab file to add in the entries that associate the MAC address tabwith the TCP/IP address.

3. Connect the DS3000, DS4000, or DS5000 Storage Subsystem Ethernet ports to the network.4. Boot the DS3000, DS4000, or DS5000 Storage Subsystem.

When you are finished, complete the steps in the “Steps for assigning static TCP/IP addresses to theDS3000, DS4000, or DS5000” on page 1-7.

Identifying Ethernet MAC addresses: To manage your storage subsystem using the direct-managementmethod, you need to identify the hardware Ethernet medium access control (MAC) address for eachcontroller.

Every storage subsystem has a label with the hardware Ethernet MAC address number. The number willhave the format xx.xx.xx.xx.xx.xx, where x represents a letter or a number. For example, an Ethernet MACaddress might be 00.a0.b8.20.00.d8.

Instructions and label locations for particular storage subsystems are listed in the following sections.


http://www.ibm.com/systems/support/storage

Identifying the Ethernet MAC addresses on a DS4800, DS5100, or DS5300 Storage Subsystem: The machinetype, model number, and serial number are located on top of each RAID controller unit. The MACaddresses are located near the Ethernet ports on each RAID controller.

Note: You can access the controllers from the back of a DS4800, DS5100, or DS5300 chassis.

Identifying the Ethernet MAC addresses on a DS3000, DS3500, DS3950, DS4200, DS4700, or DS5020 StorageSubsystem: The MAC addresses on these storage subsystems are located near the Ethernet ports on eachRAID controller.

Note: You can access the controllers from the back of the storage subsystem chassis.

Identifying the Ethernet MAC addresses on FAStT500, DS4100, DS4400, and DS4500 Storage Subsystems: Toidentify the hardware Ethernet MAC address for FAStT500, DS4100, DS4400 and DS4500 StorageSubsystems, perform the following steps:1. Remove the front bezel from the storage subsystem, and carefully pull the bottom of the bezel out to

release the pins. Then slide the bezel down.2. On the front of each controller, look for a label with the hardware Ethernet MAC address. The

number will be in the form xx.xx.xx.xx.xx.xx (for example, 00.a0.b8.20.00.d8).3. Record each Ethernet address.4. To replace the bezel, slide the top edge under the lip on the chassis. Then push the bezel bottom until

the pins snap into the mounting holes.

Identifying the Ethernet MAC addresses on FAStT200 and DS4300: To identify the hardware Ethernet MACaddress for machine types 3542 (FAStT200) and 1722 (DS4300), perform the following steps:1. Locate the Ethernet MAC address at the back of the unit, under the controller Fibre Channel host

ports. The number will be in the form xx.xx.xx.xx.xx.xx (for example, 00.a0.b8.20.00.d8).2. Record each Ethernet address.

Steps for assigning static TCP/IP addresses to the DS3000, DS4000, or DS5000Complete the following steps to assign static TCP/IP addresses to the DS3000, DS4000, or DS5000 StorageSubsystem controllers, using default TCP/IP addresses that are assigned to the DS3000, DS4000, orDS5000 Storage Subsystem controllers during manufacturing:1. Make a direct management connection to the DS3000, DS4000, or DS5000 Storage Subsystem, using

the default TCP/IP addresses for the controllers. To find the default TCP/IP addresses for yourstorage subsystem, see the Installation and User's Guide that came with the hardware.

2. Start SMclient. The Enterprise Management Window opens.3. In the Enterprise Management Window, click on the name of the default storage subsystem. The

Subsystem Management Window opens.4. In the Subsystem Management Window, right-click the controller icon and select Change → Network

Configuration in the pull-down menu. The Change Network Configuration window opens.5. In the Change Network Configuration window, click on the Controller A and Controller B tabs and

type the new TCP/IP addresses in their appropriate fields. Click OK.6. Close the Subsystem Management Window, wait five minutes, then delete the default DS3000,

DS4000, or DS5000 Storage Subsystem entry in the Enterprise Management Window.7. Add a new storage subsystem entry in the Enterprise Management Window, using the new TCP/IP

address.



Chapter 2. Introducing the software

This chapter describes the basic layout of the Storage Manager software. The storage managementsoftware has two windows that provide management functionality and a graphical representation of yourdata storage array: the Enterprise Management Window (EMW) and the Subsystem ManagementWindow (SMW).

Note: The IBM System Storage DS Storage Manager software is also referred to as the storagemanagement software.

In general, you will use the following process when using the storage management software. You use theEMW to add the storage arrays that you want to manage and then monitor the storage subsystems.Through the EMW, you will also receive alert notifications of critical errors affecting the storagesubsystems. If you are notified in the EMW that a storage subsystem has a non-Optimal status, you canstart the SMW for the affected storage subsystem to show detailed information about the storagesubsystem condition.

Important: Depending on your version of storage management software, the views, menu options, andfunctionality might differ from the information presented in this document. For information aboutavailable functionality, refer to the online help topics that are supplied with your version of the storagemanagement software.

Enterprise Management WindowParts of the Enterprise Management Window

The Enterprise Management Window (EMW) has the areas that provide options for managing yourstorage subsystem.


Table 2-1. Parts of the Enterprise Management Window

Number Description

1 "Enterprise Management" in the title bar text indicates that this is the EMW.

2 The toolbar contains icons that are shortcuts to common commands. To show thetoolbar, select View → Toolbar.

3 The EMW contains two tabs:

v Devices - Shows discovered storage subsystems and their status and also showsunidentified storage subsystems.

v Setup - Allows you to perform initial setup tasks with the storage managementsoftware.

Learn about using the EMW Devices tab

The Devices tab in the EMW presents two views of the storage subsystems that are managed by thestorage management station:v Tree view

gc53113501_em

w_1

Figure 2-1. Parts of the Enterprise Management Window


v Table view

Tree view

The Tree view provides a tree-structured view of the nodes in the storage subsystem. The Tree viewshows two types of nodes:v Discovered Storage Subsystemsv Unidentified Storage Subsystems

Both the Discovered Storage Subsystems node and the Unidentified storage subsystems node are childnodes of the storage management station node.

The Discovered Storage Subsystems node has child nodes that represent the storage subsystems that arecurrently managed by the storage management station. Each storage subsystem is labeled with itsmachine name and is always present in the Tree view. When storage subsystems and hosts with attachedstorage subsystems are added to the EMW, the storage subsystems become child nodes of the DiscoveredStorage Subsystems node.

Note: If you move the mouse over the storage management station node, a tooltip shows the controller'sIP address.

The Unidentified Storage Subsystems node shows storage subsystems that the storage managementstation cannot access because the name or IP address does not exist.

You can perform these actions on the nodes in the Tree view:v Double-click the storage management station node and the Discovered Storage Subsystems node to

expand or collapse the view of the child nodes.v Double-click a storage subsystem node to launch the Subsystem Management Window for that storage

subsystem.v Right-click the Discovered Storage Subsystems node to open a pop-up menu that contains the

applicable actions for that node.

The right-click menu for the Discovered Storage Subsystems node contains these options:v Add Storage Subsystem

v Automatic Discovery

v Refresh

These options are the same as the options in the Tools menu. For more information, refer to the Using theEnterprise Management Window online help topic.

Table view

Each managed storage subsystem is represented by a single row in the Table view. The columns in theTable view show data about the managed storage subsystem.

Table 2-2. Data shown in the Table View

Column Description

Name The name of the managed storage array. If the managed storagesubsystem is unnamed, the default name is Unnamed.

Type The type of managed storage subsystem. This type is represented by anicon.

Status An icon and a text label that report the true status of the managedstorage subsystem.

Chapter 2. Introducing the software 2-3

Table 2-2. Data shown in the Table View (continued)

Column Description

Management Connections Out-of-Band - This storage subsystem is an out-of-band storagesubsystem.

In-Band - This storage subsystem is an in-band storage subsystem thatis managed through a single host.

Out-of-Band, In-Band -This storage subsystem is a storage subsystemthat is both out-of-band and in-band.

Click Details to see more information about any of these connections.

Comment Any comments that you have entered about the specific managedstorage subsystem.

Sort the rows in the Table view in ascending order or descending order by either clicking a columnheading or by selecting one of these commands:v View → By Name

v View → By Status

v View → By Management Connection

v View → By Comment

Learn about using the EMW Setup tab

The EMW Setup tab is a gateway to tasks that you can perform when you set up a storage subsystem.Using the EMW Setup tab, you can perform these tasks:v Add a storage subsystemv Name or rename a storage subsystemv Configure an alertv Manage a storage subsystem by launching the SMWv Open the Inherit Systems Settings window

Showing managed storage subsystems in the Table view

You can change the way that managed storage subsystems appear in the Table view.v Select the storage management station node to show all of the known managed storage subsystems in

the Table view.v Select a Discovered Storage Subsystem node or Undiscovered Storage Subsystem node in the Tree view

to show any storage subsystems that are attached to that specific host in the Table view.

Note: If you have not added any storage subsystems, the Table view is empty.v Select a storage subsystem node in the Tree view to show only that storage subsystem in the Table

view.

Note: Selecting an Unidentified node in the Tree view shows an empty Table view.

Subsystem Management WindowStarting the Subsystem Management Window

To start the Subsystem Management Window (SMW) from the Enterprise Management Window (EMW),do one of the following:


v Click the Devices tab, and double-click the name of the storage subsystem that you want to manage.v Click the Devices tab, right-click the name of the storage subsystem you want to manage, and select

Manage Storage Subsystem.v Click the Devices tab, and select Tools → Manage Storage Subsystem.v Click the Setup tab, and select Manage Storage Subsystem. In the Select Storage Subsystem dialog,

select the name of the storage subsystem that you want to manage, and click OK.

The SMW is specific to an individual storage subsystem; therefore, you can manage only a single storagesubsystem within an SMW; however, you can start more than one SMW from the EMW tosimultaneously manage multiple storage subsystems.

Parts of the Subsystem Management Window

The Subsystem Management Window (SMW) provides the following options for managing your storagesubsystem.


Table 2-3. Parts of the Subsystem Management Window

Number Description

1 "Subsystem Management" indicates that this is the SMW.

2 The name of the storage subsystem that you are managing and its status.

3 The toolbar contains icons that are shortcuts to common commands. To show theToolbar, select View → Toolbar.

4 The name of the storage subsystem that you are managing and its status.

gc53113501_sm

w_2

Figure 2-2. Parts of the Subsystem Management Window


Table 2-3. Parts of the Subsystem Management Window (continued)

Number Description

5 The Subsystem Management Window has these tabs:

v Summary - Shows an overview of the configuration of the storage subsystem.

v Logical/Physical - Contains the Logical pane and the Physical pane.

v Mappings - Contains the Topology pane and the Defined Mappings pane.

v Setup - Contains links to tasks that you can perform when setting up a storagesubsystem, such as configuring the storage subsystem, setting the storagesubsystem password, and other tasks.

v Support - Contains links to tasks, such as recovering from storage subsystemfailure; gathering support information, such as the Event Log and a descriptionof a storage subsystem; and sending the information to a Customer andTechnical Support representative.

Learn about using the Summary tab

The Summary tab in the Subsystem Management Window (SMW) shows information about the storagesubsystem. Links to the Storage Subsystem Profile dialog, relevant online help topics, and the storageconcepts tutorial also appear. Additionally, the link to the Recovery Guru dialog is shown when thestorage subsystem needs attention.

In the Summary tab, you can view this information:v The status of the storage subsystemv The hardware components in the storage subsystemv The capacity of the storage subsystemv The hosts, the mappings, and the storage partitions in the storage subsystemv The arrays and logical drives in the storage subsystem

Learn about using the Logical/Physical tab

The Logical/Physical tab in the Subsystem Management Window (SMW) contains two panes: the Logicalpane and the Physical pane.

Note: You can resize either pane by dragging the splitter bar, located between the two panes, to the rightor to the left.

Logical pane

The Logical pane provides a tree-structured view of the logical nodes. Click the plus (+) sign or theminus (-) sign adjacent to a node to expand or collapse the view. You can right-click a node to open apop-up menu that contains the applicable actions for that node.


Nodes in the Logical pane

The storage subsystem, or root node, has the types of child nodes shown in the following table.

Table 2-4. Nodes in the Logical pane

Child nodes of the root node Description of the child nodes

Unconfigured Capacity This node represents the storage subsystem capacity that is not configuredinto an array.Note: Multiple Unconfigured Capacity nodes might appear if yourstorage subsystem contains mixed drive types. Each drive type has anassociated Unconfigured Capacity node shown under the TotalUnconfigured Capacity node if unassigned drives are available in thedrive tray.

Array This node has two types of child nodes:

v Logical Drive - This node represents a configured and defined logicaldrive. Multiple logical drive nodes can exist under an Array node. See“Types of Logical Drives? for a description of these arrays.

v Free Capacity - This node represents a region of capacity that you canuse to create one or more new logical drives within the storagesubsystem. Multiple Free Capacity nodes can exist under an Arraynode.

Types of logical drives

These types of logical drives appear under the Array node:v Standard logical drives.v Primary logical drives that participate in a mirror relationship in the primary role. Primary logical

drives are standard logical drives with a synchronized mirror relationship. The remote secondarylogical drive that is associated with the primary logical drive appears as a child node.

v Secondary logical drives appear directly under the Array node when the local storage subsystemcontains this logical drive.

v Mirror repository logical drives.v Snapshot repository logical drives.v Snapshot logical drives are child nodes of their associated base logical drive.v Source logical drives are standard logical drives that participate in a logical drive copy relationship.

Source logical drives are used as the copy source for a target logical drive. Source logical drives accepthost I/O requests and store application data. A source logical drive can be a standard logical drive, asnapshot logical drive, a snapshot base logical drive, or a Remote Logical Drive Mirroring primarylogical drive.

v Target logical drives are standard logical drives that participate in a logical drive copy relationship andcontain a copy of the data from the source logical drive. Target logical drives are read only and do notaccept write requests. A target logical drive can be made from a standard logical drive, the base logicaldrive of a snapshot logical drive, or a Remote Logical Drive Mirror primary logical drive. The logicaldrive copy overwrites any existing logical drive data if an existing logical drive is used as a target.

Physical pane

The Physical pane provides this information:v A view of the hardware components in a storage subsystem, including their status.v The hardware components that are associated with a selected node in the Logical pane.

You can right-click a hardware component to open a pop-up menu that contains the applicable actions forthat component.


Note: The orientation of the Physical pane is determined by the actual layout of the storage subsystem.For example, if the storage subsystem has horizontal drives, the storage management software showshorizontal drives in the Physical pane.

Controller Status

The status of each controller is indicated by an icon in the Physical pane. The following table, describesthe various controller icons.

Table 2-5. Controller status icons

Icon Status

Online, Optimal

Offline

Data Transfer Disabled

Service Mode

Slot Empty

Needs Attention (if applicable for your hardwaremodel)

Suspended (if applicable for your hardwaremodel)

Association

v The blue association dot shown adjacent to a controller in the controller enclosure indicates the currentowner of a selected logical drive in the Logical pane.

v The blue association dot adjacent to a drive indicates that the drive is associated with a selected logicaldrive in the Logical pane.

View button

The View button on each enclosure shows the status of the secondary components within the enclosure.

Drive enclosures

For each drive enclosure that is attached to the storage subsystem, a drive enclosure appears in thePhysical pane. If your storage subsystem contains mixed drive types, a drive type icon appears on the leftof the drive enclosure to indicate the type of drives in the enclosure. The following table describes thedifferent drive type icons that might appear.

Table 2-6. Drive enclosure type icons

Icon Status

This enclosure contains only Fibre Channel drives.

This drive enclosure contains only Full Disk Encryption(FDE) security-capable drives.


Table 2-6. Drive enclosure type icons (continued)

Icon Status

This drive enclosure contains only Serial Attached SCSIdrives.

This drive enclosure contains only Serial ATA drives.

Learn about using the Mappings tab

The Mappings tab in the Subsystem Management Window (SMW) contains two panes: the Topology paneand the Defined Mappings pane.

Note: You can resize either pane by dragging the splitter bar, located between the two panes, to the rightor to the left.

Topology pane

The Topology pane shows a tree-structured view of logical nodes related to storage partitions. Click theplus (+) sign or the minus (-) sign adjacent to a node to expand or collapse the view. You can right-click anode to open a pop-up menu that contains the applicable actions for that node.

Nodes in the Topology pane

The storage subsystem, or the root node, has four types of child nodes.

Table 2-7. Types of nodes in the Topology view

Child nodes of the root node Description of the child nodes

Undefined Mappings The Undefined Mappings node has one type of child node:

v Individual Undefined Mapping - Represents a logical drive withan undefined mapping. Multiple Logical Drive nodes can existunder an Undefined Mappings node.

Default Group Note: If the DS Storage Manager Storage Partitioning premium featureis disabled, all of the created logical drives are in the Default Group.

A Default Group node has two types of child nodes:

v Host Group - Defined host groups that are not participating inspecific mappings are listed. This node can have host child nodes,which can have child host port nodes.

v Host - Defined hosts that are not part of a specific host group butare part of the Default Group and are not participating in specificmappings are listed. This node can have child host port nodes.

Host Group A Host Group node has one type of child node:

v Host - Defined hosts that belong to this defined host group arelisted. This node can have child host port nodes.

Note: The host nodes that are child nodes of this host group can alsoparticipate in mappings specific to the individual host rather than thehost group.

Host A Host node has one type of child node:

v Host Ports - This node has child nodes that represent all of the hostports or single ports on a host adapter that are associated with thishost.


Storage Partition icon

The Storage Partition icon, when present in the Topology pane, indicates that a storage partition has beendefined for the Default Group, a host group, or a host. This icon also appears in the status bar whenstorage partitions have been defined.

Defined Mappings pane

The Defined Mappings pane shows the mappings that are associated with a node selected in theTopology pane.

This information appears for a selected node.

Table 2-8. Node information in the Defined Mappings pane

Column name Description

Logical Drive name The user-supplied logical drive name.

The factory-configured access logical drive also appears in this column.Note: An access logical drive mapping is not required for storagesubsystem with an in-band connection and might be removed.

Accessible by Shows the Default Group, a defined host group, or a defined host thathas been granted access to the logical drive in the mapping.

LUN The LUN assigned to the specific logical drive that the host or hosts useto access the logical drive.

Logical Drive Capacity The logical drive capacity in units of GB.

Type The type of logical drive: standard logical drive or snapshot logical drive.

You can right-click a logical drive name in the Mappings pane to open a pop-up menu. The pop-upmenu contains options to change and remove the mappings.

The information shown in the Defined Mappings pane varies according to what node you select in theTopology pane, as shown in this table.

Table 2-9. Node information by type of node

Node selected Information that appears in the Defined Mappings pane

Root (storage subsystem) node All defined mappings.

Default Group node or any child nodeof the Default Group

All mappings that are currently defined for the Default Group (if any).

Host Group node (outside of DefaultGroup)

All mappings that are currently defined for the Host Group.

Host node that is a child node of a HostGroup node

All mappings that are currently defined for the Host Group, plus anymappings specifically defined for a specific host.

HBA Host Ports node or individual hostport node outside of the Default Group

All mappings that are currently defined for the HBA host port'sassociated host.

Learn about using the SMW Setup tab

The SMW Setup tab provides links to these tasks:v Locating the storage subsystemv Renaming the storage subsystemv Setting a storage subsystem password


v Configuring the storage subsystemv Defining the hosts and host portsv Mapping logical drives to hostsv Saving configuration parameters in a filev Configuring the Ethernet management portsv Viewing and enabling the premium features

You can click a link to open the corresponding dialog.

Learn about using the Support tab

The Support tab in the SMW provides links to these tasks:v Recovering from a storage subsystem failure by using the Recovery Guruv Gathering support information, such as the Event Log and a description of the storage subsystem, to

send to a Customer and Technical Support representativev Viewing the description of all components and properties of the storage subsystemv Downloading the controller firmware, the NVSRAM, the drive firmware, the ESM firmware, and the

ESM configuration settingsv Viewing the Event Log of the storage subsystemv Viewing the online help topicsv Viewing the version and copyright information of the storage management software

Managing multiple software versions

When you open the Subsystem Management Window (SMW) to manage a storage subsystem, the versionof software that is appropriate for the version of firmware that the storage subsystem uses is opened. Forexample, you can manage two storage subsystems using this software; one storage subsystem hasfirmware version 6.14, and the other has firmware version 7.5x. When you open an SMW for a particularstorage subsystem, the correct SMW version is used. The storage subsystem with firmware version 6.14uses version 9.14 of the storage management software, and the storage subsystem with firmware version7.5x uses version 10.5x of the storage management software. You can verify the version that you arecurrently using by selecting Help → About in the SMW.


Chapter 3. Installing Storage Manager and Support Monitor

This chapter describes requirements and procedures for installing DS Storage Manager software,including the Support Monitor tool. The Support Monitor tool is a component of the IBM DS StorageManager version 10.60.x5.17 and later. The Apache Tomcat web server and MySQL database softwarepackages will also be installed, as a part of the Support Monitor tool.

The installation instructions include the following sections:v “Pre-installation requirements”v “Installation requirements” on page 3-3v “Installing DS Storage Manager and Support Monitor packages automatically using the installation

wizard” on page 3-3v “Installing Storage Manager packages manually” on page 3-6v “Completing the Storage Manager installation” on page 3-8

Attention: For cluster configurations, complete all applicable configuration procedures for each storagesubsystem before you install the storage management software on a second host or cluster server.

Pre-installation requirements

This section describes the requirements that must be met before IBM DS Storage Manager with SupportMonitor can be installed.

The supported management station operating systems for DS Storage Manager are:v AIXv Windows 2003 and Windows 2008v Linux (RHEL and SLES)v HP-UXv SUN Solaris

Support Monitor must be installed on the same management stations as the IBM DS Storage Managersoftware. The supported management station operating systems for Support Monitor are:v Microsoft® Windows® 2003 SP2, Windows 2008, Windows 2008 R2, Windows XP (Service Pack 2 or

later), and Windows Vista (x86, x64, and IA64 editions)v Red Hat Enterprise Linux 4 and 5 (x86, x86_64, and IA64 editions)v Red Hat Linux on POWERv SUSE Linux Enterprise Server 9, 10, and 11 (x86, x86_64, IA64, and Linux on POWER editions)v SUN Solaris 10 (Sparc and x86 editions)v HP-UX (PA-RISC and IA64 editions)v IBM AIX® 5.2, AIX 5.3 and AIX 6.1.

Important: If a MySQL database application or Apache Tomcat web server application is installed on themanagement station, it must be uninstalled before the Support Monitor can be installed.

Note: With Storage Manager version 10.50.xx.xx, controller firmware 5.41.xx.xx and later are supported.Controller firmware versions earlier than 5.41.xx.xx are no longer supported or managed.


The management station must also meet the following hardware, software, and configurationrequirements:v Microprocessor speed of 1.6 GHz or faster.v Minimum of 2 GB of system memory. If any other applications are installed in the management station,

additional memory might be required.v Minimum of 1.5 GB of free disk space for the tool and for the saved support bundles.v The TCP/IP stack must be enabled. If Support Monitor is installed, the management station Ethernet

port TCP/IP address must be static, and its Ethernet port must be on the same Ethernet subnet as themonitored storage subsystem Ethernet management ports. DHCP server IP address is not supported. IfSupport Monitor is not installed, the IP address does not have to be static.

v The following requirements apply only to the Support Monitor tool:– Make sure that your storage subsystem meets the subsystem model and controller firmware version

requirements listed in the following table:

Table 3-1. Storage Monitor-compatible subsystems and controller firmware

DS storage subsystem Storage Monitor-compatibility Controller firmware compatibility

DS3200 No n/a

DS3300 No n/a

DS3400 No n/a

DS3950 Yes 7.60.28.xx and later

DS4100 No n/a



DS4400 No n/a







– To use Support Monitor, one of the following Web browsers must be installed:- Internet Explorer 7.0 or later- Netscape version 6.0 or later- Mozilla version 1.0 or later- Firefox version 3.x or later

– Any installed MySQL database application on the management station must be manuallyuninstalled before you install the Support Monitor tool.

– Any installed Apache Tomcat web server software on the management station must be manuallyuninstalled before you install the IBM DS Storage Manager Profiler Support Monitor tool.

– The Support Monitor uses port 162 by default to receive event data from the server. To prevent portconflicts with other applications running on the server, make sure that no other applications useport 162.


Installation requirements

You can install the DS Storage Manager automatically using the DS Storage Manager installation wizard,or you can install each package manually.

For DS Storage Manager installations on UNIX®, your system must have graphics capability to use theinstallation wizard; if your system does not have graphics capability, you can use the shell command toinstall DS Storage Manager without graphics. You can also skip this section and install the stand-alonehost software packages using the procedures described in “Installing Storage Manager packagesmanually” on page 3-6. All of the packages are included with the installation CD.

Installing DS Storage Manager and Support Monitor packagesautomatically using the installation wizardIBM DS Storage Manager version 10.60.x5.17 and later includes software for a Web-based tool calledSupport Monitor. The DS Storage Manager and Support Monitor software are both installed when youuse the installation wizard. However, the DS Storage Monitor and Support Monitor are installed inseparate parts. The DS Storage Manager Client program and other Storage Manager software componentswill be installed first, followed by the DS Storage Manager Profiler Support Monitor tool. A separateprogress status bar is displayed for each part.

Before you install the DS Storage Manager and Support Monitor software, read the following “Importantinstallation notes about Support Monitor” and “Installing the DS Storage Manager and Support Monitor”sections.

Important installation notes about Support Monitorv The Support Monitor tool is packaged in the same SMIA installer package as the Storage Manager Host

software package. There is not a separate installer package for the Support Monitor tool.v The DS Storage Manager Client program must be installed with the Support Monitor tool. Support

Monitor will not run correctly without the DS Storage Manager client program.v The Support Monitor tool will be installed by default when either the Typical (Full Installation) or

Management installation type is selected in the Wizard Select Installation Type window. The SupportMonitor tool will not be installed when the Host installation type is selected.

v If you select the Custom installation type option in the wizard Select Installation Type window, theSupport Monitor tool will be displayed as a selected component for installation. To install the DSStorage Manager without the Support Monitor tool, clear the Support Monitor box.If you are installing the DS Storage Manager on multiple management stations that manage the sameset of storage subsystems, use the custom installation type in subsequent installations of the DS StorageManager software and clear the Support Monitor box to prevent it from being installed in more thanone management station. If the tool is installed in more than one management station, the storagesubsystem will be servicing multiple requests at 2:00 a.m. every day for support bundle collection. Thismight cause problems during support-bundle collection.

v If MySQL data or Apache Tomcat web server program is installed, the Storage Manager ProfilerSupport Monitor installation terminates and an installation error message is displayed. The StorageManager Profiler Support Monitor install log is stored in the directory C:\Program Files...\IBM_DS inWindows operating systems and in the directory /opt/IBM_DS/ in Unix-type operating systems. The filename of the log is IBMStorageManagerProfiler_install.log.

Installing the DS Storage Manager and Support Monitor

To install the DS Storage Manager software (including the Support Monitor tool) with the installationwizard, complete the following steps:

Chapter 3. Installing Storage Manager and Support Monitor 3-3

1. For a Windows-based operating system the default drive is C, and for a UNIX-based operating systemit is the root file system. Download the files from the DS Storage Manager CD, or from the DS SystemStorage Disk Support Web site, to a directory on your system.

2. Install the DS Storage Manager software with the DS Storage Manager Profiler Support Monitor tool:v If your management station has a Microsoft Windows operating system, complete the following

steps:a. Double-click the IBM DS Storage Manager package executable icon.b. Follow the instructions in the Installation wizard to install the DS Storage Manager software

with the DS Storage Manager Profiler Support Monitor tool. If you accept the defaultinstallation directory, the Storage Manager Profiler Support Monitor will be installed inC:\Program Files...\IBM_DS\ IBMStorageManagerProfiler Server.

c. When you select the installation type, you can choose one of the following options:– Typical (Full) Installation: Installs all Storage Manager software packages necessary for both

managing the storage subsystem from this host and providing I/O connectivity to the storage– Management Station: Installs the packages required to manage and monitor the storage

subsystem (SMruntime and SMclient)– Host: Installs the packages required to provide I/O connectivity to the storage subsystem

(SMruntime, SMagent, and SMutil)– Custom: Allows you to select which packages you want to install. To install the DS Storage

Manager without the Support Monitor tool, select the customer installation and clear theSupport Monitor box.

d. Configure any antivirus software to not scan the MySQL directory. In Windowsoperating-system environments, the directory is C:\Program Files...\IBM_DS\IBMStorageManagerProfiler Server/mysql.

e. Click Start > All Programs > DS Storage Manager 10 client > Storage Manager 10 client tostart the DS Storage Manager Client program. Add the storage subsystems that you want tomanage and monitor in the Enterprise Management Window (EMW) of the Storage ManagerClient program.

v If your management station has a Unix-based operating system, such as Linux, AIX, or Solaris,complete the following steps:a. Log in as root.b. If the IBM DS storage manager software package .bin file does not have executable permission,

use the chmod +x command to make it executable.c. Execute the .bin file and follow the instructions in the Installation wizard to install the software.

If you accept the default installation directory, the Storage Manager Profiler Support Monitorwill be installed in /opt/IBM_DS/IBMStorageManagerProfiler_Server.When you select the installation type, you can choose one of the following options:– Typical (Full) Installation: Installs all Storage Manager software packages necessary for both

managing the storage subsystem from this host and providing I/O connectivity to the storage– Management Station: Installs the packages required to manage and monitor the storage

subsystem (SMruntime and SMclient)– Host: Installs the packages required to provide I/O connectivity to the storage subsystem

(SMruntime, SMagent, and SMutil)– Custom: Allows you to select which packages you want to install. To install the DS Storage

Manager without the Support Monitor tool, select the customer installation and clear theSupport Monitor box.

d. Configure any antivirus software to not scan the MySQL directory. In Unix-typeoperating-system environments, the directory is /opt/IBM_DS/IBMStorageManagerProfiler_Server/mysql.


e. Type SMclient in the console window and press Enter to start the DS Storage Manager Clientprogram. Add the storage subsystems that you want to manage and monitor to the EnterpriseManagement Window (EMW) of the Storage Manager Client program.

The only time that you need to configure the Storage Manager Profiler Support Monitor tool iswhen you want to change the support bundle collection time for the monitored storage subsystems.The Storage Manager Profiler Support Monitor tool automatically collects the support bundles fromthe storage subsystems that were added to the Enterprise Management Window of the StorageManager Client program daily at 2:00 a.m.

Note: During the installation, the question Automatically Start Monitor? is displayed. This refersto the Event Monitor service. The Event Monitor must be enabled for both the automatic ESMsynchronization and the automatic support bundle collection of critical events. To enable the EventMonitor, select Automatically Start Monitor.

Installing Support Monitor using a console window

For management stations without a graphics adapter installed, the DS Storage Manager software packagecan be installed silently with the option -i silent or option -i console.

The -i silent option will cause the DS Storage Manager Software Installer package to be installed usingthe default installer settings. The -i console option will prompt the user for installed options beforestarting the software installation, just like the Installation wizard. However, the prompts will be displayedin console window text instead of graphical windows.

Snippets of the DS Storage Manager installation with the -i silent and -i console options are shown inthe following example.[usr@RHManaStation ~]# ./SMIA-LINUX-10.60.A5.17.bin -i consolePreparing to install...Extracting the JRE from the installer archive...Unpacking the JRE...Extracting the installation resources from the installer archive...Configuring the installer for this system’s environment...

Launching installer...

Preparing CONSOLE Mode Installation...

===============================================================================Choose Locale...----------------

1- Deutsch->2- English

3- Español4- Français5- Italiano6- Português (Brasil)

CHOOSE LOCALE BY NUMBER:2... ... ...

...[usr@RHManaStation ~]# ./SMIA-LINUX-10.60.A5.17.bin -i silentPreparing to install...Extracting the JRE from the installer archive...Unpacking the JRE...Extracting the installation resources from the installer archive...Configuring the installer for this system’s environment...

Launching installer...


Preparing SILENT Mode Installation...

===============================================================================IBM System Storage DS Storage Manager 10(created with InstallAnywhere by Macrovision)-------------------------------------------------------------------------------------

===============================================================================Installing...-------------

[==================|==================|==================|==================][------------------|------------------|------------------|------------------]

... ... ...

Installing Storage Manager packages manuallyFor Unix-type operating systems such as AIX, Linux, Sun Solaris, and HP-UX, individual DS StorageManger software packages are provided. See Table 3-2 for the installation sequence of the individualsoftware packages.

Use the procedure in this section to manually install the DS Storage Manager software on a storagemanagement station. Make sure you install the packages in the correct order.

Important: There is no individual software package for the DS Storage Manager Support Monitor tool. Ifyou want to install the Support Monitor tool, you must use the DS Storage Manager software installerpackage.

Note: There is no manual installation option for Windows operating systems. For all installations of DSStorage Manager on Windows, the individual software packages are included in a single host softwareinstaller.

Software installation sequence

Install the DS Storage Manager software packages in the sequences shown in Table 3-2.

Note: These packages are available for UNIX systems that may be running without a graphical userinterface.

Table 3-2. Installation sequence of DS Storage Manager software packages

Step Package

1 SMruntime

2 SMesm

3 SMclient¹

4 SMagent

5 SMutil

¹SMclient is dependent on SMruntime, which is a Java compiler for SMclient. SMruntime must beinstalled first.

Manual installation requirements

Before installing the software, ensure that the DS Storage Manager files are available in a directory on thesystem.


Manual installation steps

For your installation, modify the following commands, as needed. No restart is required during theinstallation process. The verification process returns a table that describes the software installation,including the install package file name, version number, action, and action status.1. Install the <SMpackage> by typing the command appropriate for your operating system.

Note: The manual install commands listed in the following table are only for UNIX-based operatingsystems.

Table 3-3. Storage Manager package install commands

Operating system Package name Install command

AIX SMruntime.AIX-10.xx.xx.xx.bff #installp -a -d /path_name/SMruntime.AIX-10.xx.xx.xx.bff SMruntime.aix.rte

HP-UX SMruntime_10.xx.xx.xx.depot #swinstall -s /cdrom/HP-UX/SMruntime_10.xx.xx.xx.depot

Solaris SMruntime-SOL-10.xx.xx.xx.pkg #pkgadd -d path/filename.pkg

Linux on POWER SMruntime-LINUX-10.xx.xx.xx-x.i586.rpm

#rpm -ihv SMruntime-LINUX-10.xx.xx.xx-x.i586.rpm

Windows SMIA-WS-xx.xx.xx.xx.exe

2. Verify that the installation was successful by typing the command appropriate for your operatingsystem.

Table 3-4. Storage Manager package installation verify commands

Operating system Verify command

AIX # lslpp -ah <SMpackage>.aix.rte

HP-UX # swverify -v <SMpackage>

Solaris # pkginfo -l <SMpackage>

Linux on POWER # rpm -qa|grep <SMpackage>

If the verification process returns an error, contact your IBM service representative.

Uninstalling DS Storage Manager and Support MonitorUse the applicable procedure in this section to uninstall Support Monitor, or both DS Storage Managerand Support Monitor, on either a Windows or Unix-type operating system.

Uninstalling DS Storage Manager and Support Monitor on a Windows operatingsystemUse the information in this section to uninstall the DS Storage Manager and Support Monitor software.

To uninstall the software on Windows operating systems, complete the following steps:1. Open the Control Panel window.2. If you have Windows 2003 or Windows XP, double-click Add/Remove Programs. If you have

Windows 2008, double-click Program and Features. The new window opens.3. Select IBM DS Storage Manager Host Software version 10.xx.x5.yy, where xx and yy are the

applicable version numbers of your software.4. Click the Change/Remove button and follow the instructions in the Uninstall IBM System Storage DS

Storage Manager 10 wizard to uninstall just the Support Monitor tool or both the Support Monitortool and the DS Storage Manager software. The process of uninstalling the software might leave filesthat were created by the DS Storage Manager and the Support Monitor after the installation was


complete. These files might include trace files, repository files, and other administrative files. Deletethese files manually to completely remove DS Storage Manager and Support Monitor.

Note: You can also uninstall the DS Storage Manager Profiler Support Monitor by running theuninstall.exe execution file in the C:\Program Files ...\IBM_DS\IBMStorageManagerProfiler Serverdirectory.

Uninstalling DS Storage Manager and Support Monitor on a Unix-type operatingsystem

To uninstall the software on Unix-type operating systems, complete the following steps:1. Open the /opt/IBM_DS/Uninstall IBM System Storage DS Storage Manager 10 directory that contains

the uninstaller binary.2. Run the uninstall script Uninstall_IBM_System_Storage_DS_Storage_Manager_10 in the console

window to uninstall just the Support Monitor or both the Support Monitor and DS Storage Managersoftware. The process of uninstalling the software might leave files that were not part of the originalinstallation. These files might include trace files, repository files, and other administrative files. Deletethese files manually to completely remove the DS Storage Manager and Support Monitor.

Note: You can also uninstall the DS Storage Manager Profiler Support Monitor by running the uninstallexecution file in the /opt/IBM_DS/IBM_DS\IBMStorageManagerProfiler_Server directory.

Completing the Storage Manager installationThis section contains procedures for using the Enterprise Management and Subsystem Managementfeatures of DS Storage Manager to complete the storage management installation tasks for all hostoperating systems.

To complete a Storage Manager installation, the following tasks must be performed:v Initial automatic discovery of storage subsystemsv Initial manual discovery of storage subsystemsv Naming the storage subsystemsv Setting up alert notificationsv Downloading controller firmware and NVSRAMv DS Storage Manager premium featuresv Saving a storage subsystem profile

Each of these steps is discussed in detail in the sections that follow.

The Enterprise Management Window opens when you start the DS Storage Manager. You can use theEnterprise Management Window to do the following:v Add and discover the storage subsystemsv View all storage subsystems in your management domainv Perform batch storage subsystem management tasks using the Script Editor

Performing an automatic discovery of storage subsystemsComplete the following steps to perform an initial automatic discovery of storage subsystems:1. Click Start → Programs.2. Click IBM DS Storage Manager Client. The client software starts and displays the Enterprise

Management Window and the Confirm Initial Automatic Discovery window.3. Click Yes to begin an initial automatic discovery of hosts and storage subsystems attached to the local

subnetwork.


After the initial automatic discovery is complete, the Enterprise Management Window displays allhosts and storage subsystems attached to the local subnetwork.

Note: The Enterprise Management Window can take up to a minute to refresh after an initialautomatic discovery.

4. Verify that each host and storage subsystem displays in the Enterprise Management Window.If a host or storage subsystem is not displayed, perform the following tasks:v Check the hardware and hardware connections for possible problems. Refer to your machine type's

Storage Subsystems Installation, User's, and Maintenance Guide for specific procedures.v Refer to the Enterprise Management online help for additional information about discovering

storage subsystems.v If you are using the network management method (commonly known as out-of-band management),

verify that all hosts and storage subsystems are connected to the same subnet network. If you areusing the host-agent method (commonly known as in-band management), ensure that the FibreChannel connection between the host and storage subsystems is made.

v Make sure that all of the preparation steps for setting up the storage subsystem for a networkmanaged system are completed. Use the Add Device option to add the IP addresses of the storagesubsystem. Add both IP addresses of the controller; otherwise, you will get a "partially-manageddevice" error message when you try to manage the storage subsystem.

Note: To use the auto-discovery method, the storage subsystem and this host must be on the samesubnet. Otherwise use the manual method to add a subsystem.

v If you are using the host-agent management method, perform the following steps:a. Make sure that the SMagent is installed in the host.b. Verify that you have a Fibre Channel connection from the storage subsystems to the host that

has the SMagent installed.c. Verify that all of the preparation steps are complete, then perform the following steps:

1) Run the hot_add utility.2) Restart the SMagent.3) Right-click on the host, and click Tools → Rescan in the Enterprise Management Window.

Note: In certain situations, a storage subsystem might be duplicated in the device tree after anautomatic discovery. You can remove a duplicate storage management icon from the device tree byusing the Remove Device option in the Enterprise Management Window.

5. Verify that the status of each storage subsystem is Optimal. If a device shows a status of Unresponsive,right-click the device and select Remove Device to delete it from the management domain. Then usethe Add Device option to add it to the management domain again. Refer to the EnterpriseManagement Window online help for instructions on removing and adding devices.

Performing a manual discovery of storage subsystemsYou can manually add more hosts or storage subsystems. You can use this option to selectively manage agroup of storage subsystems from an SMclient. You can also use this option to add additional devices tobe managed that were not discovered during the SMclient initial discovery. For more information aboutthis option, see the Enterprise Management Window online help.

Important:

v When you add new storage subsystems to the existing storage subsystems in a SAN that are managedthrough the host-agent software, you must stop and restart the host-agent service. When the host-agentservice restarts, the new storage subsystem is detected. Then, go to the Enterprise ManagementWindow and click Tools → Rescan to add the new storage subsystems to the management domain.


v When you add new storage subsystems to existing storage subsystems that are managed using thedirect-management method, be sure to specify the IP addresses for both controllers.

Storage subsystem password protectionFor added security, you can configure a password for each storage subsystem that you manage byclicking Storage Subsystem → Change Password. After you have set the password for each storagesubsystem, you are prompted for that password the first time you attempt a destructive operation in theSubsystem Management window. You are asked for the password only once during a single managementsession.

Important: There is no way to reset the password once it is set. Ensure that the password information iskept in a safe and accessible place. Contact IBM technical support for help if you forget the password tothe storage subsystem.

Naming storage subsystemsAs you set up your network, decide on a naming convention for the storage subsystems. (The namingconvention IBM recommends is to use the device type, followed by the serial number; an example mightbe 1815 1312345 XXXX xxx xxxx.) When you install the storage management software and start it for thefirst time, all storage subsystems in the management domain display as <unnamed>. Use the SubsystemManagement Window to rename the individual storage subsystems.

Consider the following factors when you name storage subsystems:v There is a 30-character limit. All leading and trailing spaces are deleted from the name.v Use a unique, meaningful naming scheme that is easy to understand and remember.v Avoid arbitrary names or names that might quickly lose their meaning.v The software adds the prefix Storage Subsystem when displaying storage subsystem names. For

example, if you name a storage subsystem Engineering, it is displayed as:Storage Subsystem Engineering

To name your storage subsystem, perform the following steps:1. In the Enterprise Management Window, right-click the storage subsystem and select Rename. The

Rename Storage Subsystem window opens.

Note: If any of your hosts are running path failover drivers, update the storage subsystem name inyour path failover driver's configuration file before rebooting the host machine to ensureuninterrupted access to the storage subsystem.

2. Type the name of the storage subsystem. Then click OK. To continue, click Yes on the warning screen.3. Repeat this procedure for each unnamed storage subsystem. For more information, see the topic on

renaming storage subsystems in the Subsystem Management Window online help.4. Proceed to “Setting up alert notifications.”

Setting up alert notificationsAfter you add devices to the management domain, you can set up alert notifications to report criticalevents on the storage subsystems. The following alert-notification options are available:v Notification to a designated network management station (NMS) using Simple Network Management

Protocol (SNMP) trapsv Notification to designated e-mail addressesv Notification to designated alphanumeric pagers (requires separately supplied software to convert

e-mail messages)


Note: You can only monitor storage subsystems within the management domain. If you do not install theEvent Monitor service, the Enterprise Management Window must remain open. If you close the window,you will not receive any alert notifications from the managed storage subsystems. Refer to the EnterpriseManagement Window online help for additional information.

Alert notification with SNMP traps

To set up alert notification to an NMS using SNMP traps, perform the following steps:1. Insert the IBM DS Storage Manager CD into the CD-ROM drive on an NMS. You need to set

up the designated management station only once.2. Copy the SMxx.x.MIB file from the SMxxMIB directory to the NMS.3. Follow the steps required by your NMS to compile the management information base (MIB)

file. (For details, contact your network administrator or see the documentation specific to yourparticular storage management product.)

Alert notification without SNMP traps

To set up alert notification without using SNMP traps, select Storage subsystem → Edit →Configure alerts from the Enterprise Management window.

Managing iSCSI settingsClick the Setup tab in the Subsystem Management window. A window similar to the one in Figure 3-1 onpage 3-12 opens.

Note: The link to Configure iSCSI Host Ports on the Subsystem Management window is available onlyfor storage subsystems that support an iSCSI host attachment. Currently, the following storagesubsystems support iSCSI host attachment:v DS3300v DS3500v DS3950v DS5020


The following iSCSI options are available from the Storage Subsystem menu:v Change Target Authenticationv Enter Mutual Authentication Permissionsv Change Target Identificationv Change Target Discoveryv Configure iSCSI Host Portsv View/End iSCSI Sessionsv View iSCSI Statistics

Changing target authenticationSelect Change Target Authentication to specify the target challenge handshake authentication protocol(CHAP) secret that the initiator must use during the security negotiation phase of the iSCSI login. Bydefault, None is selected. To change the selection, click CHAP, and then enter the CHAP secret. You canalso select the option to generate a random secret. This enables 1-way CHAP.

Figure 3-1. Manage iSCSI settings


Entering mutual authentication permissionsBefore you select Enter Mutual Authentication Permissions, you must define a host port for the initiatorand enable Target Authentication. After the host port is listed, select the host from the list and click ChapSecret to specify the secret that is passed to the initiator from the target to authenticate it. This enablesMutual CHAP (2-way).

Changing target identificationSelect Change Target Identification to specify a target alias that is to be used during device discovery.You must provide a unique name for the target that consists of fewer than 30 characters.

Note: You will connect to the target by using the fully qualified IQN that is listed above the alias.

Changing target discoverySelect Change Target Discovery to perform device discovery by using the iSCSI simple naming service(iSNS). After you select this option, select the Use iSNS Server check box. You can also select whetherthe iSNS server is discovered using a DHCP server on your network, and you can manually specify anInternet Protocol version 4 (IPv4) or IPv6 address. When you click the Advanced tab, you can assign adifferent TCP/IP port for your iSNS server for additional security.

Note: To provide the required port login information for correct device discovery, all iSCSI ports must beable to communicate with the same iSNS server.

Configuring iSCSI host portsSelect Configure iSCSI Host Ports to configure all of the TCP/IP settings. You can choose to enable ordisable IPv4 and IPv6 on all of the ports. You can also statically assign IP addresses or let them bediscovered using DHCP. Under Advanced IPv4 Settings, you can assign VLAN Tags (802.1Q) or set theEthernet Priority (802.1P). Under Advanced Host Port Settings, you can specify a unique iSCSI TCP/IPport for that target port. You can also enable Jumbo Frames from this option. The supported frame sizesare 1500 and 9000.

Viewing or ending an iSCSI sessionSelect View/End iSCSI Sessions to view all of the connected iSCSI sessions to the target. From this page,you can also close an existing session by forcing a target ASYNC logout of the initiator session.

Viewing iSCSI statisticsSelect View iSCSI Statistics to view a list of all iSCSI session data, for example, the number of headerdigest errors, number of data digest errors, and successful protocol data unit counts. You can also set abaseline count after a corrective action to determine whether the problem is solved.

iSNS best practicesThere are many considerations for using an iSNS server correctly. Make sure that you correctly assignyour iSNS server address that is provided during the DHCP lease discovery of your initiator or target.This enables ease of discovery when you use initiator-based solutions. If you are unable to do this, andmust manually assign the iSNS server to your software or hardware initiators, you should make sure thatall of the storage subsystem iSCSI ports and iSCSI initiators are in the same network segment (or makesure that the routing between the separate network segments is correct). If you do not do this, you willbe unable to discover all ports during the iSCSI discovery process, and you might not be able to correctlyperform a controller or path failover.

Using DHCPDo not use DHCP for the target portals. If you use DHCP, you should assign DHCP reservations so thatleases are maintained consistently across restarts of the storage subsystem. If static IP reservations are notprovided, the initiator ports can lose communication to the controller and might not be able to reconnectto the device.


Using supported hardware initiatorsAs of the date of this document, only the following hardware initiators are supported:v IBM iSCSI Server TX Adapterv IBM iSCSI Server SX Adapterv QLogic iSCSI Single-Port PCIe HBA for IBM System xv QLogic iSCSI Dual-Port PCIe HBA for IBM System x

All of the hardware initiators that are supported use the same base firmware code and the SANsurfermanagement application. Before you install and configure these adapters, make sure that you haveinstalled the latest management application and the latest firmware code. After you confirm this,configure each adapter one at a time. To make sure that failovers are performed correctly, connect eachadapter by using one of the following two basic configurations:v If you have a simple configuration in which all adapters and target ports are in the same network

segment, each adapter should be able to log in to any target port.v If you have a complex configuration, each adapter is allowed a single path to each controller device.

To log in correctly to all available target ports from the hardware initiator, complete the following steps.

Note: Failure to perform the steps in the following procedure might result in path failover inconsistenciesand incorrect operation of the storage subsystem.1. Start the SANsurfer management utility.2. Connect to the system that is running the qlremote agent.3. Select the adapter that you want to configure.4. Select Port 0 or Port 1 for the adapter.5. Click Target Settings.6. Click the plus sign (+) in the far right of the window.7. Type either the IPv4 or IPv6 address of the target port to which you want to connect.8. Click OK

9. Select Config Parameters.10. Scroll until you see ISID.11. For connection 0, the last character that is listed should be 0. For connection 1, it should be 1, for

connection 2, it should be 2, and so on.12. Repeat steps 6 through 11 for each connection to the target that you want to create.13. After all of the sessions are connected, select Save Target Settings. If you are using the QLogic iSCSI

Single-Port or Dual-Port PCIe HBA for IBM System x to support IPv6, you should allow the host busadapter firmware to assign the local link address.

Using IPv6The storage subsystem iSCSI ports support the Internet Protocol version 6 (IPv6) TCP/IP. Note that onlythe final four octets can be configured if you are manually assigning the local link address. The leadingfour octets are fe80:0:0:0. The full IPv6 address is required when you are attempting to connect to thetarget from an initiator. If you do not provide the full IPv6 address, the initiator might fail to beconnected.

Network settingsUsing a DS storage subsystem that supports iSCSI host attachment in a complex network topologyintroduces many challenges. If possible, try to isolate the iSCSI traffic to a dedicated network. If this isnot possible, follow this suggestion. If you are using a hardware-based initiator, the Keep Alive timeoutshould be set to 120 seconds. To set the Keep Alive timeout, complete the following steps:1. Start the SANsurfer Management Utility and connect to the server.2. Select the adapter and the adapter port that is to be configured.


3. Select the port options and firmware.

The default connection timeout is 60 seconds. This setting is correct for simple network topologies.However, in a more complex configuration, if a network convergence occurs and you are not using FastSpanning Tree and separate spanning tree domains, you might encounter I/O timeouts. If you are using aLinux iSCSI software initiator, modify the ConnFailTimeout parameter to account for the spanning treeissue. The ConnFailTimeout value should be set to 120 seconds.

Maximum Transmission Unit settingsAll devices on a link that are expected to communicate with each other (i.e., on the same VLAN, etc.)should be configured with the same Maximum Transmission Unit (MTU) size. The MTU size is either aconfiguration item or hard-coded in the device, and is not negotiated between endpoints during login orconnection establishment. If a device receives a packet that is larger than the MTU size then it drops thepacket. If a router receives a packet whose size doesn't exceed the MTU size of the link on which it wasreceived, but exceeds the MTU size of the forwarding link, then the router either fragments the packet(IPv4) or else returns a “packet too large” ICMP error message. Make sure that all of the components ona network link are using the same MTU size value.

For Storage systems that support iSCSI, the default MTU setting is 1500 Bytes. There is an option to select9000 Bytes for jumbo frames. All components (host, switch, routers, and targets) must have JumboFrames (large MTU) enabled in order for end-to-end Jumbo frames to work effectively. If Jumbo frames isnot enabled on all components, one or more of the following can occur:v Dropped frames.v No connection/dropped connections due to error messages relating to the packet being too large.v Fragmentation of jumbo frames.

Microsoft iSCSI Software Initiator considerationsThe native multipath I/O (MPIO) that is provided with the Microsoft iSCSI Software Initiator (version2.03 or later) is not supported. You must use the DSM that is provided with the Storage Managersoftware to make sure that failover and I/O access are correct. If the native MPIO from the MicrosoftiSCSI Software Initiator is used, it causes unwanted effects.

Downloading controller firmware, NVSRAM, ESM firmwareThis section provides instructions for downloading storage subsystem controller firmware, NVSRAM,storage expansion enclosure ESM firmware, and drive firmware. Normally, the storage subsystemfirmware download sequence is as follows:1. Controller firmware2. Controller NVSRAM3. ESM firmware4. Drive firmware

Review the readme file that is provided with updated controller firmware, NVSRAM, ESM firmware, anddrive firmware for any necessary changes to the firmware download sequence.

Note: Perform a Collect All Support Data before performing the controller and NVSRAM procedure.Please refer to Appendix I, “Critical event problem solving,” on page I-1 for details.

Important: The following procedures assume you are using the latest available controller firmwareversion. Access the latest versions of storage subsystem controller firmware, NVSRAM, and expansionenclosure ESM firmware at the following IBM Web site:


For the most recent Storage Manager README files for your operating system, see “Finding StorageManager software, controller firmware, and README files” on page xiii.



Before upgrading a DS4800, DS4700, or a DS4200 Storage Subsystem to controller firmware version07.1x.xx.xx, first see the procedures in Appendix A, “Using the IBM System Storage DS3000, DS4000, andDS5000 Controller Firmware Upgrade Tool,” on page A-1.

Important:

1. IBM supports firmware download with I/O, sometimes referred to as concurrent firmware download,with some DS3000, DS4000, and DS5000 Storage Subsystems. Before proceeding with concurrentfirmware download, check the README file packaged with the firmware code or your particularoperating system's DS Storage Manager host software for any restrictions to this support.

2. Suspend all I/O activity while downloading firmware and NVSRAM to a DS3000, DS4000, or DS5000Storage Subsystem with a single controller or you will not have redundant controller connectionsbetween the host server and the DS3000, DS4000, or DS5000 Storage Subsystem.

Important: Always check the DS3000, DS4000, or DS5000 Storage Subsystem controller firmwareREADME file for any controller firmware Dependencies and Prerequisites before applying the firmwareupdates to the DS3000, DS4000, or DS5000 Storage Subsystem. Updating any components of the DS3000,DS4000, or DS5000 Storage Subsystem firmware without complying with the Dependencies andPrerequisites may cause down time (to fix the problems or recover).

If your controller's existing firmware is 06.1x.xx.xx or later, you will have the option to select theNVSRAM for download at the same time that you upgrade/download the new controller firmware.Additionally, you will have the option to download the firmware and NVSRAM immediately, but activateit later, when it may be more appropriate. See the online help for more information.

Note: The option to activate firmware at a later time is not supported on the DS4400.

Determining firmware levelsBefore you download any firmware upgrades, be sure you know your current firmware version. Thereare two different methods to determine DS3000, DS4000, or DS5000 Storage Subsystem, expansion unit,drive, and ESM firmware versions. Each method uses the DS Storage Manager client that manages theDS3000, DS4000, or DS5000 Storage Subsystem with the attached expansion unit.

Method one: Go to the Subsystem Management window and select Storage Subsystem → View Profile.When the Storage Subsystem Profile window opens, select the All tab and scroll through Profile ForStorage Subsystem to locate the following information.

Note: The Profile For Storage Subsystem page contains all the profile information for the entiresubsystem; therefore, it may be necessary to scroll through a large amount of information to locate thefirmware version numbers.

DS3000, DS4000, and DS5000 Storage SubsystemsFirmware information types are:v NVSRAM versionv Appware version (Appware is a reference to controller firmware.)v Bootware version (Bootware is a reference to controller firmware.)

See the following example of profile information.

Controller in Enclosure 0, Slot AStatus: OnlineCurrent configurationFirmware version: 07.10.23.00.Appware version: 07.10.23.00.Bootware version: 07.10.23.00.NVSRAM version: N1814D47R1010V05


DrivesFirmware version

See the following example of SATA drive data.

Product ID: ST3750640NS 43W9715 42D0003IBMPackage version: EP58Firmware version: 3.AEP

ATA TranslatorProduct ID: BR-2401-3.0Vendor: SLIFirmware Version: LP1158

ESM

ESM card firmware version

See the following example of ESM data.

ESM card status: OptimalFirmware version: 9898Configuration settings version: FD 00.52 03/08/2007

Method two: Select the appropriate procedure from the following options and complete it to obtain thespecified firmware version.

To obtain the controller firmware version:Right-click the Controller icon in the Physical View pane of the Subsystem Management windowand select Properties. The Controller Enclosure properties window opens and displays theproperties for that controller.

You must perform this step for each individual controller.

To obtain the drive firmware version:Right-click the Drive icon in the Physical View pane of the Subsystem Management window andselect Properties. The Drive Properties window opens and displays the properties for that drive.

You must perform this step for each individual drive.

To obtain the ESM firmware version:

1. In the Physical View pane of the Subsystem Management window, click the Drive EnclosureComponent icon (which is the icon furthest to the right). The Drive Enclosure ComponentInformation window opens.

2. Click the ESM icon in the left pane. The ESM information displays in the right pane of theDrive Enclosure Component Information window.

3. Locate the firmware version of each ESM in the drive enclosure.

Downloading controller or NVSRAM firmware

Note: It is highly recommend that you perform a Collect All Support Data before upgrading controllerfirmware and NVSRAM. Refer to Appendix I, “Critical event problem solving,” on page I-1 for datacollection procedures.

This section provides instructions for downloading DS3000, DS4000, or DS5000 storage server controllerfirmware and NVSRAM. Normally, the DS3000, DS4000, and DS5000 Storage Subsystem firmwaredownload sequence starts with controller firmware, followed by the NVSRAM and then the ESMfirmware, and concludes with the drive firmware.


Important:

v Users upgrading from 06.xx to 07.xx must use the Controller Firmware Upgrade Tool. See Appendix A,“Using the IBM System Storage DS3000, DS4000, and DS5000 Controller Firmware Upgrade Tool,” onpage A-1.

v Users already at the 07.xx firmware level are not required to use the Controller Firmware Upgrade Toolto upgrade to another 07.xx level. There are diagnostic capabilities in the Upgrade Tool that are verybeneficial.

To download firmware version 06.1x.xx.xx or later, and NVSRAM, perform the following steps:1. From the Enterprise Management Window, select a storage subsystem.2. Click Tools → Manage Device. The Subsystem Management Window opens.3. Click Advanced → Maintenance → Download → Controller firmware.... The download firmware

window opens.

Note: DS3000, DS4000, or DS5000 Storage Subsystems with controller firmware versions 06.1x.xx.xx,and higher, support the downloading of the NVSRAM file together with the firmware file; therefore,the following window will display only if your exiting controller firmware is version 06.1x.xx.xx orhigher. This download feature is not supported in DS3000, DS4000, or DS5000 Storage Subsystemswith controller firmware 05.4x.xx.xx or earlier. If your exiting controller firmware is version05.4x.xx.xx or lower, only a window for downloading firmware will display.

4. Click Browse next to the Selected firmware file: field to identify and select the file with the newfirmware.

5. Select the Download NVSRAM file with firmware option and click Browse next to the SelectedNVSRAM file: field to identify and select the file and identify and select the correct NVSRAMfilename. For most configurations, unless your configuration has unique conditions, upgrade theNVSRAM at the same time as the controller firmware. If you choose to transfer and activateimmediately, do not select Transfer files but don't activate them (activate later); otherwise, click thebox to select Transfer files but don't activate them (activate later). To activate the firmware at a latertime, in the Subsystem Management window, click Advanced → Maintenance → Activate ControllerFirmware.

Downloading ESM firmwareThis section provides instructions for downloading DS3000, DS4000, or DS5000 Storage ExpansionEnclosure ESM firmware. Normally, the DS3000, DS4000, or DS5000 Storage Subsystem firmwaredownload sequence starts with controller firmware, followed by the NVSRAM and then the ESMfirmware, and concludes with the drive firmware.

Steps for downloading the ESM firmware: Perform the following steps to download the ESM firmware:1. In the DS3000, DS4000, or DS5000 Storage Subsystem Management window, select Advanced →

Maintenance → Download → ESM firmware.... A Download Environmental Card Firmware windowopens.

2. Click Select All to direct the download to all enclosures. You can also select one enclosure orcombinations of enclosures by pressing the Ctrl key while selecting the individual enclosures.

3. Click Browse to identify and select the filename of the ESM firmware file and click Start to begin theESM firmware download.

4. A Confirm Download window opens. Type Yes and click OK to start the download process.5. Click Cancel to close the window when the ESM firmware download to all selected enclosures is

complete.

Note: Suspend all I/O activity while ESM firmware downloads if you select multiple enclosures fordownloading ESM firmware. If you select only one enclosure for download at a time, you candownload ESM firmware while the server conducts I/O activity.


Automatic ESM firmware synchronization: When you install a new ESM into an existing storageexpansion enclosure in a DS3000, DS4000, or DS5000 Storage Subsystem that supports automatic ESMfirmware synchronization, the firmware in the new ESM is automatically synchronized with the firmwarein the existing ESM. This resolves any ESM firmware mismatch conditions automatically.

To enable automatic ESM firmware synchronization, ensure that your system meets the followingrequirements:v The Storage Manager Event Monitor must be installed and running.v The Storage Subsystem must be defined in the Enterprise Management window of the Storage

Manager client (SMclient).

Note: Storage Manager currently supports automatic ESM firmware synchronization with EXP710 andEXP810 storage expansion enclosures only. Contact IBM for information about support for other types ofstorage expansion enclosures in the future. To correct ESM firmware mismatch conditions in storageexpansion enclosures without automatic ESM firmware synchronization support, you must download thecorrect ESM firmware file by using the ESM firmware download menu function in the SMclientSubsystem Management window.

Downloading drive firmware

Important: The following procedures assume you are using the most current available controllerfirmware version. If you are using an earlier firmware version, see “Finding Storage Manager software,controller firmware, and README files” on page xiii to obtain the appropriate firmware versiondocumentation.

This section provides instructions for downloading DS3859, DS4000, or DS5000 drive firmware. See theonline help for additional information.

Important:

1. IBM supports firmware download with I/O, sometimes referred to as concurrent firmware download.This feature is not supported for drive firmware.

2. Before starting the drive firmware download process, do the following:v Complete a full backup of all data residing on the drives that you select for firmware upgrade.v Unmount the file systems on all logical drives accessing the drives that you select for firmware

upgrade.v Stop all I/O activity before downloading drive firmware to a DS3000, DS4000, or DS5000 Storage

Subsystem.

Downloading Storage Manager drive firmwareTo download drive firmware for DS Storage Manager, perform the following steps:1. From the Enterprise Management Window, select a storage subsystem.2. On the Enterprise Management Window's menu bar, click Tools → Manage Device. The Subsystem

Management Window opens.3. On the Subsystem Management Window's menu bar, click Advanced → Maintenance → Download →

Drive Firmware/Mode pages .... The Download Drive Firmware wizard window opens to theIntroduction page. Read the instructions displayed and click Next.

Note: Storage Manager offers you the option to download and update up to four different firmwarefile types simultaneously.

4. Click Browse to locate the server directory that contains the firmware that you plan to download.5. Select the firmware file that you plan to download and click OK. The file appears listed in the

Selected Packages window pane.


6. Select the firmware file for any additional drive types that you intend to download and click OK.Additional files appear listed in the Selected Packages window pane. A maximum total of four drivetypes are possible.

7. Click Browse to repeat step 6 until you have selected each firmware file that you plan to download.8. When you have finished specifying the firmware packages for download, select Next.9. The Select Drive window opens, containing two tabs, a Compatible Drives tab and an Incompatible

Drives tab. The Compatible Drives tab contains a list of the drives compatible to the firmwarepackage types that you selected. From that list, select the drives to which you plan to download thedrive firmware that you selected in steps 6 and 7.

Note: The firmware that you propose to download should be listed on the Compatible Drives tab.If your particular drives' product ID matches the firmware type, however, and it is not listed ascompatible on the tab, contact your IBM technical support representative for additional instructions.

10. Select the Compatible Drives tab.Press and hold the Ctrl key while using your mouse to select multiple drives individually, or pressand hold the shift key while using your mouse to select multiple drives listed in series. Thecompatible firmware that you selected in steps 5 on page 3-19 and 6 will download to the drivesthat you select.

11. Click Finish to initiate download of the drive firmware to each compatible drive that you selected instep 9.

12. The Download Drive Firmware warning opens and prompts: "Do you want to continue?" Type yesand click OK to start the drive firmware download.

13. The Download Progress window opens. Do not intervene until the download process completes.14. Every drive scheduled for firmware download will be designated as in progress until successful or

failed.

Note: Complete the following two steps if you receive a failure.a. Click the Save as button to save the error log.b. On the Subsystem Management Window's menu bar, click Advanced →Trouble Shooting → Open

Event Log and complete the following tasks necessary to save the storage subsystem event logbefore contacting your IBM Service Representative and proceeding to step 16.1) Click the Select all button.2) Click Save the Storage Subsystem Event Log.

15. When the Close button appears active, the drive firmware download process is complete.16. Click Close to exit the Download Progress window.17. Use either of the following procedures to determine or verify what level of drive firmware resides on

a particular drive:v Right-click on that drive in the Logical/Physical View in the Subsystem Management Window and

click Properties. The associated drive firmware version will be listed in the drive properties table.v Right-click on Storage Subsystem → View Profile in the Logical/Physical View of the Subsystem

Management Window.

DS Storage Manager premium features

DS Storage Manager supports the following premium features, which are separately available forpurchase from IBM or an IBM Business Partner:

Copy ServicesThe following copy services are available with Storage Manager:v FlashCopyv VolumeCopy


v Enhanced Remote Mirror Option

For more detailed information about the copy services features, see the IBM System Storage DSStorage Manager Copy Services User's Guide.

Storage PartitioningStorage Partitioning is standard on all DS3000, DS4000, and DS5000 Storage Subsystems that aresupported by DS3000, DS4000, and DS5000 controller firmware versions. For more informationabout Storage Partitioning, see the “Storage partitioning overview” on page 4-2.

FC/SATA Intermix premium featureThe IBM System Storage DS3000, DS4000, and DS5000 Fibre Channel and Serial ATA Intermixpremium feature supports the concurrent attachment of Fibre Channel and SATA storageexpansion enclosures to a single DS3000, DS4000, or DS5000 controller configuration.

This Intermix premium feature enables you to create and manage distinct arrays that are builtfrom either Fibre Channel disks or SATA disks, and allocate logical drives to the appropriateapplications using a single DS3000, DS4000, or DS5000 Storage Subsystem.

For important information about using the Intermix premium feature, including configuration,firmware versions required for specific Intermix configurations, and setup requirements, see theIBM System Storage DS Storage Manager Fibre Channel and Serial ATA Intermix Premium FeatureInstallation Overview.

See your IBM representative or reseller for information regarding future DS3000, DS4000, andDS5000 Storage Subsystem support for the FC/SATA Intermix premium feature.

Enabling premium featuresYou must perform the following tasks to enable a premium feature on your storage subsystem:v “Obtaining the feature enable identifier”v “Generating the feature key file” on page 3-22v “Enabling the premium feature” on page 3-22

Note: The procedure for enabling a premium feature depends on your version of DS Storage Manager.

To obtain the storage subsystem premium feature identifier string, ensure that your controller unit andstorage expansion enclosures are connected, powered on, and managed using the SMclient.

Obtaining the feature enable identifier: Each storage subsystem has its own unique feature enableidentifier. This identifier ensures that a particular feature key file is applicable only to that storagesubsystem.

Note: Before you obtain the feature enable identifier, complete the following prerequisites:1. Make sure that you have available the feature activation code from the premium feature Web

activation card, as well as the model, machine type, and serial number of the storage subsystem.2. Make sure that the controller unit and disk drive expansion units are connected, powered on, and

configured.

Complete the following steps to obtain the Feature Enable Identifier:1. Click Start → Programs → Storage Manager xx Client. The Enterprise Management Window opens.2. In the Enterprise Management Window, double-click the storage subsystem for which you want to

enable the premium feature. The Subsystem Management window opens for the selected storagesubsystem.

3. Complete one of the following actions, depending on your version of DS Storage Manager:v If you are using DS Storage Manager version 9.x or earlier, click Storage Subsystem → Premium

Features → List. The List Premium Features window opens and displays the feature enableidentifier.


v If you are using DS Storage Manager version 10 or later, click Storage Subsystem → PremiumFeatures.... The Premium Features and Feature Pack Information window opens. The FeatureEnable Identifier is displayed at the bottom of the new window.

4. Record the feature enable identifier.

Note: To prevent a mistake when recording the feature enable identifier, copy the 32-characteridentifier and paste it in the premium feature key request field.

5. Click Close to close the window.6. Continue to next section, “Generating the feature key file.”

Note: To check the status of an existing premium feature in DS Storage Manager version 9.x or earlier,select Storage Subsystem → Premium Features → List from the pull-down menu.

Generating the feature key file: You can generate the feature key file by using the Premium FeatureActivation tool that is located at the following Web site:

https://www-912.ibm.com/PremiumFeatures1. Complete the steps in the Web site. Feature key file is e-mailed to you.2. On your hard drive, create a new directory that you can find easily. (For example, name the directory

FlashCopy feature key.)3. Save the premium feature key file in the new directory.

Enabling the premium feature:Enabling the premium feature in DS Storage Manager 9.x or earlier

To enable a premium feature in DS Storage Manager version 9.x or earlier, complete the following steps:1. In the Subsystem Management window, click Premium Features → Enable.2. Browse to the appropriate key file in the directory you saved it to in the previous task, “Generating

the feature key file.”3. Click OK.4. Verify that the premium feature is enabled:

a. In the Subsystem Management window, click Storage Subsystem → Premium Features → List. TheList Premium Features window opens. The window shows the following information:v The premium features that are enabled on the storage subsystemv The feature enable identifier

b. Click Close to close the window.

Enabling the premium feature in DS Storage Manager 10 or later

To enable a premium feature in DS Storage Manager version 10 or later, complete the following steps:1. In the Subsystem Management window, click Storage Subsystem → Premium Features.... The

Premium Features and Feature Pack Information window opens.2. To enable a premium feature from the list, click the Enable... button. A window will open that allows

you to select the premium feature key file to enable the premium feature. Follow the on-screeninstructions.

3. Verify that the premium feature is enabled by inspecting the displayed list of premium features in thePremium Features and Feature Pack Information window.

4. Click Close to close the Premium Features and Feature Pack Information window.


https://www-912.ibm.com/PremiumFeatures

Note: To enable and verify a feature pack, click the Change button in the Premium Feature andFeature Pack Information window. A window opens that allows you to select the key file to enablethe premium feature pack. To verify that the feature pack is enabled, inspect the content of theFeature Pack installed on storage subsystem field.

Important: Enabling a premium feature pack requires that the controllers be restarted. If the storagesubsystem for which the premium feature pack will be enabled is in running, make sure you scheduledowntime to restart the controllers.

Disabling premium features: In normal system operations, you do not need to disable the premiumfeatures. However, if you want to disable a premium feature, make sure that you have the key file or theIBM DS3000, DS4000, or DS5000 premium feature entitlement card with the premium feature activationcode for generating the key file. You need this key file to re-enable the premium feature at a later time.

Note:

1. For DS3000 storage subsystems with controller firmware version 7.35 or earlier, you cannot use the DSStorage Manager interface to disable a premium feature. Instead, you must use the Storage Managercommand line (SMcli) scripts to disable the premium feature.

2. If you want to enable the premium feature in the future, you must reapply the Feature Key file forthat feature.

3. You can disable the Remote Mirror Option without deactivating the feature. If the feature is disabledbut activated, you can perform all mirroring operations on existing remote mirrors. However, whenthe feature is disabled you cannot create any new remote mirrors. For more information aboutactivating the Remote Mirror Option, see the IBM System Storage DS Storage Manager Copy ServicesUser's Guide or see “Using the Activate Remote Mirroring Wizard” in the DS Storage Manager onlinehelp.

4. If a premium feature becomes disabled, you can access the Web site and repeat this process.

Disabling the premium feature in DS Storage Manager 9.x or earlier

To enable a premium feature in DS Storage Manager version 9.x or earlier, complete the following steps:1. In the Subsystem Management window, click Storage Subsystem → Premium Features → Disable. The

Disable Premium Feature window opens, which shows all of the premium features that are enabled.2. Select one item in the list, and then click OK. A confirmation message states that a premium feature

should not be disabled.3. Click Yes. The Working dialog displays while the feature is being disabled. When the feature has been

disabled, the Working dialog closes.

Disabling the premium feature in DS Storage Manager 10 or later

To enable a premium feature in DS Storage Manager version 10 or later, complete the following steps:1. In the Subsystem Management window, click Storage Subsystem → Premium Features.... The

Premium Features and Feature Pack Information window opens.2. Select the premium feature you want to disable and click the Disable button.

For additional assistance, contact your local IBM service provider.

Saving the storage subsystem profile

Important: You should save a storage subsystem profile whenever you modify the arrays and logicaldrives in your storage subsystem. This saved profile contains detailed controller information, includinglogical and physical disk configuration information that can help you recover the configuration in theevent of a catastrophic failure. Do not save a profile for a storage subsystem on that same storagesubsystem.


To save a storage subsystem profile, select Storage Subsystem → View Profile in the Storage SubsystemManagement window and click the Save As button when the Storage Subsystem Profile window opens.To save the full profile, select all of the radio buttons. In addition, you can also select Advanced →Troubleshooting → Collect All Support Data to collect all the various types of inventory, status,diagnostic and performance data from this storage subsystem and save them to a single compressed file.


Chapter 4. Configuring storage

Upon the successful installation of the DS Storage Manager software, your next step is to configure thestorage subsystem. The following information will explain the tasks necessary for configuration:v “Storage partitioning overview” on page 4-2v “Using the Task Assistant” on page 4-3v “Creating arrays and logical drives” on page 4-4v “Defining the default host type” on page 4-9v “Defining a host group” on page 4-10v “Steps for defining the host and host port” on page 4-12v “Mapping LUNs to a storage partition” on page 4-12v “Optional premium features” on page 4-15

Tip: Storage Manager version 10.50 now has an improved Enterprise Management Window, containing aSetup tab and a Devices tab. By default, the Setup tab opens first.


Table 4-1. Parts of the Enterprise Management Window

Number Description

1 "Enterprise Management" in the title bar text indicates that this is the EMW.

2 The toolbar contains icons that are shortcuts to common commands. To show thetoolbar, select View → Toolbar.

3 The EMW contains two tabs:

v Devices - Shows discovered storage arrays and their status and also showsunidentified storage arrays.

v Setup - Allows you to perform initial setup tasks with the storage managementsoftware.

Storage partitioning overviewWhen you begin to create your storage partitions using the procedures in this section, be aware of thefollowing information:

gc53113501_em

w_1


v The Storage Manager task assistant provides a Storage Partitioning wizard that you can use to defineyour host and host ports, and map LUNs to the storage partitions. If your subsystem is runningcontroller firmware 05.xx.xx.xx, you cannot use the wizard. Both types of procedures are documentedin this section.

v These procedures assume that you have already created a physical connection between the host andthe storage subsystem controllers, and that you have also connected and zoned the SAN switch (ifany). If you have not completed these connections, please note that Storage Manager will not be able tolist the WWPNs of the HBAs during these procedures. In this case you will need to type the WWPNsinto the appropriate fields during the steps for defining a host and host ports.

v Create the host group at the storage subsystem level. Do not create host groups at the default grouplevel.Exception: If you are running a DS4100 or a DS4300 configuration without partitioning enabled, youcan use the default host group.

v In a cluster partition, perform logical drive mappings on the host group level so that all the hosts cansee the same storage. In a normal partition, perform logical drive mappings on the host level.

v To set up and assign IBM i LUNs on the DS5300 and DS5100 storage subsystems with the wizard, see“Configuring the IBM Systems Storage DS5100 and DS5300 for IBM i” on page 4-13 for informationspecific to IBM i configuration.

Using the Task AssistantThe DS Storage Manager Task Assistant provides a convenient, central location from which you canchoose to perform the most common tasks in the Enterprise Management window and in the SubsystemManagement window. You can use the Task Assistant to complete many of the procedures described inthis section.

Important: If you have controller firmware version 7.50 or later, the Storage Manager task descriptionsmight differ slightly from the tasks in the following lists.

In the Enterprise Management window, the Task Assistant consists of shortcuts to these tasks:v Creating arrays and logical drivesv Defining host groups (partitions)v Defining hostsv Mapping LUNs to a host or partition

In the Subsystem Management window, the Task Assistant consists of shortcuts to these tasks:v Configuring storage subsystemsv Defining hostsv Creating a new storage partitionv Mapping additional logical drivesv Saving configurations

If there is a problem with the storage subsystem, a shortcut to the Recovery Guru appears, where youcan learn more about the problem and find solutions to correct the problem.

To open the Task Assistant, choose View → Task Assistant from either the Enterprise Managementwindow or the Subsystem Management window, or click the Task Assistant button in the toolbar. TheTask Assistant window opens.

Important: If you have controller firmware version 7.50 or later, the Storage Manager procedure foraccessing the Task Assistant functionality is slightly different. There is no button and no separate windowfor Task Assistant. Click the Setup tab in the Subsystem Management window to access the TaskAssistant menu on the Initial Setup Tasks screen.

Chapter 4. Configuring storage 4-3

Note: The Task Assistant is automatically invoked every time you open the Subsystem Managementwindow unless you check the Don't show the task assistant at startup again check box at the bottom ofthe window.

Configuring hot-spare devicesYou can assign available physical drives in the storage subsystem as hot-spare drives to keep dataavailable. A hot spare is a drive that contains no data and that acts as a standby in case a drive fails in aRaid 1, Raid 10, Raid 3, Raid 5 or Raid 6 array. If the logical drive in an array fails, the controllersautomatically use a hot-spare drive to replace the failed physical drive while the storage subsystem isoperating. The controller uses redundancy data to automatically reconstruct the data from the failedphysical drive to the replacement (hot-spare) drive. This is called reconstruction. The hot-spare drive addsanother level of redundancy to the storage subsystem. If a physical drive fails in the storage subsystem,the hot-spare drive is automatically substituted without requiring a physical swap. There are two ways toassign hot-spare drives:v Automatically assign drives - If you select this option, hot spare drives are automatically created for

the best hot-spare coverage using the drives that are available. This option is always available.v Manually assign individual drives - If you select this option, hot-spare drives are created out of those

drives that were previously selected in the Physical View. This option is not available if you have notselected any drives in the Physical View.

If you choose to manually assign the hot-spare drives, select a drive with a capacity equal to or largerthan the total capacity of the drive you want to cover with the hot-spare. For example, if you have an 18GB drive with configured capacity of 8 GB, you could use a 9 GB or larger drive as a hot-spare.Generally, you should not assign a drive as a hot-spare unless its capacity is equal to or greater than thecapacity of the largest drive on the storage subsystem. For maximum data protection, you should useonly the largest capacity drives for hot-spare drives in mixed capacity hard drive configurations. There isalso an option to manually unassign individual drives.

If a drive fails in the array, the hot-spare can be substituted automatically for the failed drive withoutrequiring your intervention. If a hot-spare is available when a drive fails, the controller uses redundancydata to reconstruct the data onto the hot-spare. After the failed drive is physically replaced, you can useeither of the following options to restore the data:v When you have replaced the failed drive, the data from the hot-spare is copied back to the replacement

drive. This action is called copyback.v You can assign the hot-spare as a permanent member of the array. Performing the copyback function is

not required for this option. If you do not have a hot-spare, you can still replace a failed drive whilethe array is operating.

If you do not have a hot spare, you can still replace a failed drive while the array is operating. If thedrive is part of a RAID Level 1, RAID Level 3, RAID Level 5, RAID Level 6, or RAID Level 10 volumegroup, the controller uses redundancy data to automatically reconstruct the data onto the replacementdrive.

Manually unassign drives—If you select this option, the hot-spare drives that you selected in the PhysicalView are unassigned. This option is not available if you have not selected any drives in the PhysicalView.

Note: SATA drives and Fibre Channel drives cannot act as hot-spares for each other.

Creating arrays and logical drivesAn array is a set of Fibre Channel or SATA hard drives that are logically grouped together to form aRedundant Array of Independent Disks (RAID).


A logical drive is a logical structure, which is the basic structure that you create to store data on thestorage subsystem. The operating system recognizes a logical drive as a single drive. Choose a RAIDlevel to meet application needs for data availability and maximize Fibre Channel I/O performance.

Note: For cluster configurations, if you add or delete logical drives, you must make them known to bothnodes A and B.

Creating an arrayIn the DS Storage Manager Subsystem Management window, perform the following steps to create anarray from unconfigured capacity nodes.1. Use either of the following two methods to create a new array:

v Select Total Unconfigured Capacity node, and click Array → Create, or;v Select and right-click Total Unconfigured Capacity node, and click Create Array.The Introduction (Create Array) window opens.

2. Click Next to continue.3. The Array Name & Drive Selection (Create Array) window opens.

v Array name—Enter a name for the new array. This name can be a maximum of 30 characters.v Drive selection choices—Select one of the following two options.

– Automatic (Recommended)—Choose from a list of automatically generated drive and capacityoptions. (This option is preselected by default.)

– Manual (Advanced)—Choose specific drives to obtain capacity for the new array.v Click Next to continue.

4. The RAID Level and Capacity (Create Array) window opens. Specify the RAID level (redundancyprotection).

5. Select the number of drives (overall capacity) for the new array.6. Click Finish.7. The Array Created window opens. If, at this point, you want continue the process to create a logical

drive, click Yes; if you want to wait to create a logical drive at another time, click No.

Redundant array of independent disks (RAID)Redundant array of independent disks (RAID) is available on all operating systems and relies on a seriesof configurations, called levels, to determine how user and redundancy data is written and retrieved fromthe drives. The DS3000, DS4000, or DS5000 controller firmware supports six RAID level configurations:v RAID-0v RAID-1v RAID-3v RAID-5v RAID-6v RAID-10

Each level provides different performance and protection features.

RAID-1, RAID-3, RAID-5, and RAID-6 write redundancy data to the drive media for fault tolerance. Theredundancy data might be a copy of the data (mirrored) or an error-correcting code that is derived fromthe data. If a drive fails, the redundant data is stored on a different drive from the data that it protects.The redundant data is used to reconstruct the drive information on a hot-spare replacement drive.RAID-1 uses mirroring for redundancy. RAID-3, RAID-5, and RAID-6 use redundancy information,sometimes called parity, that is constructed from the data bytes and striped along with the data on eachdisk.


Note: RAID-0 does not provide data redundancy.

Table 4-2. RAID level configurations

RAID level Short description Detailed description

RAID-0 Non-redundant, striping mode RAID-0 offers simplicity, but does notprovide data redundancy. A RAID-0array spreads data across all drives inthe array. This normally provides thebest performance but there is not anyprotection against single drive failure.If one drive in the array fails, alllogical drives contained in the arrayfail. This RAID level is notrecommended for highdata-availability needs. RAID-0 isbetter for noncritical data.

RAID-1 or RAID-10 Striping/Mirroring mode v A minimum of two drives isrequired for RAID-1: one for theuser data and one for the mirroreddata. The DS3000, DS4000, orDS5000 Storage Subsystemimplementation of RAID-1 isbasically a combination of RAID-1and RAID-10, depending on thenumber of drives selected. If onlytwo drives are selected, RAID-1 isimplemented. If you select four ormore drives (in multiples of two),RAID 10 is automaticallyconfigured across the volumegroup: two drives for user data,and two drives for the mirroreddata.

v RAID-1 provides high performanceand the best data availability. On aRAID-1 logical drive, data iswritten to two duplicate diskssimultaneously. On a RAID-10logical drive, data is striped acrossmirrored pairs.

v RAID-1 uses disk mirroring tomake an exact copy of data fromone drive to another drive. If onedrive fails in a RAID-1 array, themirrored drive takes over.

v RAID-1 and RAID-10 are costly interms of capacity. One-half of thedrives are used for redundant data.


Table 4-2. RAID level configurations (continued)

RAID level Short description Detailed description

RAID-3 High-bandwidth mode v RAID-3 requires one dedicateddisk in the logical drive to holdredundancy information (parity).User data is striped across theremaining drives.

v RAID-3 is a good choice forapplications such as multimedia ormedical imaging that write andread large amounts of sequentialdata. In these applications, the I/Osize is large, and all drives operatein parallel to service a singlerequest, delivering high I/Otransfer rates.

RAID-5 High I/O mode v RAID-5 stripes both user data andredundancy information (parity)across all of the drives in thelogical drive.

v RAID-5 uses the equivalent of onedrive's capacity for redundancyinformation.

v RAID-5 is a good choice inmulti-user environments such asdatabase or file-system storage,where the I/O size is small andthere is a high proportion of readactivity. When the I/O size is smalland the segment size isappropriately chosen, a single readrequest is retrieved from a singleindividual drive. The other drivesare available to concurrentlyservice other I/O read requestsand deliver fast read I/O requestrates.

RAID-6 Block-level striping with dualdistributed parity.

RAID-6 is an evolution of RAID-5and is designed for tolerating twosimultaneous HDD failures bystoring two sets of distributedparities.

v RAID Level 6 uses the equivalentof the capacity of two drives (in avolume group) for redundancydata.

v RAID Level 6 protects against thesimultaneous failure of two drivesby storing two sets of distributedparities.

Note: One array uses a single RAID level and all redundancy data for that array is stored within thearray.


The capacity of the array is the aggregate capacity of the member drives, minus the capacity that isreserved for redundancy data. The amount of capacity that is needed for redundancy depends on theRAID level that is used.

To perform a redundancy check, go to Advanced → Recovery → Check array redundancy. Theredundancy check performs one of the following actions:v Scans the blocks in a RAID-3, RAID-5, or RAID-6 logical drive and checks the redundancy information

for each blockv Compares data blocks on RAID-1 mirrored drives

Important: A warning box opens when you select the Check array redundancy option that cautions youto only use the option when instructed to do so by the Recovery Guru. It also informs you that if youneed to check redundancy for any reason other than recovery, you can enable redundancy checkingthrough Media Scan.

Creating a logical driveIn the DS Storage Manager Subsystem Management window, perform the following steps to create alogical drive:1. In the Logical/Physical View tab of the Introduction (Create Logical Drive) window, you will see

how much Free Capacity is available for all existing arrays. Select the Free Capacity of an array forwhich you want to create a new logical drive; then, right-click and select Create Logical Drive.

2. Click Next.3. In the Specify Capacity/Name (Create Logical Drive) window, specify the following parameters for

the logical drive you are creating.

New logical drive capacityThe capacity can either be the entire unconfigured capacity in an array or a portion of thearray's capacity.

Units Choose GB, MB, or TB, depending upon the capacity available.

Name Choose a name that is unique in the storage subsystem, up to a maximum of 30 characters.4. Under Advanced logical drive parameters, select from one of the following options:

v Use recommended settingsv Customize settings (I/O characteristics and controller ownership)You can create the logical drive using the DS3000, DS4000, or DS5000 Storage Subsystemrecommended settings, or you can customize your I/O characteristics, controller ownership, andlogical-drive-to-LUN mapping settings. To use the recommended (default) settings, select Userecommended settings, and click Next. Proceed to step 6. If you want to customize your settings,select Customize settings, and click Next. Proceed to step 5.

5. In the Advanced logical drive parameters window, specify the appropriate I/O characteristics(characteristics type, segment size, and cache read-ahead multiplier).The I/O characteristics settings can be set automatically or they can be manually specified, based onone of the following logical drive usages—file system, database, or multimedia. Click Next. TheSpecify Logical Drive-to-LUN Mapping (Create Logical Drive) window opens.

6. In the Specify Logical Drive-to-LUN Mapping (Create Logical Drive) window, specify the logicaldrive-to-LUN mapping.The logical drive-to-LUN mapping preference can be one of the following two settings:

Default mappingThe Automatic setting specifies that a LUN is automatically assigned to the logical driveusing the next available LUN within the default host group. This setting grants logical driveaccess to host groups or host computers that have no specific logical drive-to-LUN mappings(those that were designated by the default host group node in the Topology view). If the


Storage Partition feature is not enabled, you must specify the Automatic setting. In addition,you can also change the host type to match the host operating system.

Map later using the Mappings ViewThis setting specifies that you are not going to assign a LUN to the logical drive duringcreation. This setting enables you to define a specific logical drive-to-LUN mapping andcreate storage partitions using the Mappings Defined option. When you enable storagepartitioning, specify this setting.

7. Click Finish to create the logical drive. The Creation Successful (Create Logical Drive) windowopens and states that the logical drive was successfully created.

8. In the Creation Successful (Create Logical Drive) window, click Yes and proceed to step 9, if youwant to create another logical drive; otherwise, click No. When the Completed (Create Logical Drive)window opens, click OK, and continue with step 10.

9. In the Allocate Capacity (Create Logical Drive) window, choose to create the new logical drive fromfree capacity on the same array, free capacity on a different array, or from unconfigured capacity(create a new array). Then continue with step 1 on page 4-8.

10. The Completed (Create Logical Drive) window opens. Click OK.11. Register the logical drive with the operating system.

After you create logical drives with automatic logical drive-to-LUN mappings, follow the appropriateinstructions in the Installation and Support Guide for your operating system to allow it to discoverthe new logical drive.

Defining the default host typeBefore you use the logical drives in a host computer, you must specify the correct host type. The hosttype determines how the storage subsystem controllers will work with each particular operating systemon the hosts to which it is connected. If all of the host computers connected to the same storagesubsystem are running the same operating system, and you do not want to define partitioning, you candefine a default host type.

To verify the current default host type, perform the following steps:1. In the Subsystem Management window, click Storage subsystem → View profile. A Storage

Subsystem Profile window opens.2. Click the Mappings folder tab and scroll down to NVSRAM Host Type Index Definitions. The host

type name of the index that has the word DEFAULT next to it is the default host type.3. Click Close.

The host type setting that you specify when you configure Storage Manager determines how the storagesubsystem controllers work with the operating systems on the connected hosts. All Fibre Channel HBAports that are defined with the same host type are handled the same way by the DS3000, DS4000, andDS5000 controllers. This determination is based on the specifications that are defined by the host type.Some of the specifications that differ according to the host type setting include the following options:

Auto Drive TransferEnables or disables the Auto-Logical Drive Transfer feature (ADT).

Enable Alternate Controller Reset PropagationDetermines whether the controller will propagate a Host Bus Reset/Target Reset/Logical UnitReset to the other controller in a dual controller subsystem to support Microsoft® ClusteringServices.

Allow Reservation on Unowned LUNsDetermines the controller response to Reservation/Release commands that are received for LUNsthat are not owned by the controller.


Sector 0 Read Handling for Unowned Volumes—Enable Sector 0 Reads for Unowned VolumesApplies only to host types with the Auto-Logical Drive Transfer feature enabled. For non-ADThosts, this option will have no effect.

Maximum Sectors Read from Unowned VolumesSpecifies the maximum allowable sectors (starting from sector 0) that can be read by a controllerthat does not own the addressed volume. The value of these bits specifies the maximum numberof additional sectors that can be read in addition to sector 0.

Reporting of Deferred ErrorsDetermines how the DS3000, DS4000, and DS5000 controller's deferred errors are reported to thehost.

Do Not Report Vendor Unique Unit Attention as Check ConditionDetermines whether the controller will report a vendor-unique Unit Attention condition as aCheck Condition status.

World Wide Name In Standard InquiryEnables or disables Extended Standard Inquiry.

Ignore UTM LUN OwnershipDetermines how inquiry for the Universal Access LUN (UTM LUN) is reported. The UTM LUNis used by the DS3000, DS4000, or DS5000 Storage Manager host software to communicate to theDS3000, DS4000, or DS5000 Storage Subsystem in DS3000, DS4000, or DS5000 Storage Subsystemin-band management configurations.

Report LUN Preferred Path in Standard Inquiry DataReports the LUN preferred path in bits 4 and 5 of the Standard Inquiry Data byte 6.

In most DS3000, DS4000, and DS5000 configurations, the NVSRAM settings for each supported host typefor a particular operating system environment are sufficient for connecting a host to the DS3000, DS4000,and DS5000 Storage Subsystems. You should not need to change any of the host type settings forNVSRAM.

If you think you need to change the NVSRAM settings, please contact your IBM support representativebefore proceeding.

To define a default host type, perform the following steps:1. Click Storage subsystem → Change → Default host-type. The Default Host-type window opens.2. From the pull-down list, select the host type.3. Click OK.

Note: In the Veritas Storage Foundation Linux environment the default host type must be set to 13.

Defining a host groupA host group is an entity in the Storage Partitioning topology that defines a logical collection of hostcomputers that require shared access to one or more logical drives. You can grant individual hosts in adefined host group access to storage partitions, independently of the host group. You can make logicaldrive-to-LUN mappings to the host group or to an individual host in a host group.

Steps for defining a host groupBefore you begin: Note that you must create the host group at the storage subsystem level. Do not createhost groups at the default group level.

Exception: You can use the default host group, if you are running a DS3000, DS4000, or DS5000configuration without partitioning enabled.


Complete the following steps to define a host group:1. Click the Mappings View tab on the Subsystem Management window.2. In the Topology section of the Mappings window, highlight the name of the storage subsystem, and

click Mappings → Define → Host Group.

Note: Make sure that the storage subsystem is highlighted in the left panel of the SubsystemManagement window. Do not highlight Undefined Mappings.

3. Type a name for the new host group. Click Add, and then click Close.4. Highlight the new host group and click Mappings → Define → Host.5. Type the name of the host to which the storage subsystem is attached. Click Add, and then click

Close.6. Highlight the host that you just added, then right-click and select Define Host Port.7. Select the host port identifier (WWPN) for the first HBA (for example, 10:00:00:00:c9:24:0c:3f). If you

do not see the identifier that you are looking for, see the note at the end of this procedure.

Note: If you are configuring storage for IBM i, the port will be on the first adapter. IBM i requirestwo adapters to make a valid configuration.

8. Change the host type and click Add.

Important: Failure to change the host type from the default might cause undesired results. See theREADME file for DS Storage Manager for a list of host types you can use for each host operatingsystem.

9. If you are configuring an additional HBA to this partition, choose the host port for the next HBA andclick Add, and then click Close.

Important: If you do not see the host port identifier that you want in the host port identifier drop downmenu, you can enter it manually. Otherwise, verify that the switch is properly zoned and cabled.

Defining heterogeneous hostsThe heterogeneous hosts feature enables hosts that are running different operating systems to access asingle storage subsystem. DS Storage Manager supports up to 512 storage partitions on some subsystems,which enables a multiple host-type subsystem to share storage capacity, consolidate storage, and reducestorage management costs.

Note:

1. On DS4800 and DS5000 Storage Subsystems, DS Storage Manager allows a maximum of 512 storagepartitions.

2. On the DS4700 and DS4200 Storage Subsystems, DS Storage Manager allows a maximum of 128storage partitions.

Host computers can run on completely different operating systems or variants of the same operatingsystem. When you define a host type in the Define New Host Port window, the heterogeneous hostsfeature enables the controllers in the storage subsystem to tailor their behavior (such as LUN reportingand error conditions) to the needs of the operating system or variant of the host that is sending theinformation.

Before you begin setting up the configuration for you heterogeneous host, see the IBM System Storage DSStorage Manager Concepts Guide.

Note:

1. During host-port definition, you must set each host type to the appropriate operating system so thatthe firmware on each controller can respond correctly to the host.


2. You must enable storage partitioning.

Attention: Partitioning is a premium feature. You must use the partition key save at installation or go tothe IBM Web page for feature codes to reactivate and obtain a new feature key.

Steps for defining the host and host portComplete the following steps to define the host and host ports by using the Define a host and host portswizard:1. In the Topology section of the Mappings view of the Subsystem Management window, right-click the

new host group and select Define Host. The Introduction (Define Host) window opens.2. Click Next. The Specify Host Name/HBA Attribute (Define Host) window opens.3. Type the host name in the Specify Host Name/HBA Attribute (Define Host) window. In the left panel,

select the correct WWPN of the HBA host port. Click Add.

Note: If you there is not yet a physical connection between the host and the DS3000, DS4000, orDS5000 controllers, the WWPNs will not display. In this case, you must type the correct WWPN intothe field.

4. You must now provide an alias name for the host port. Click Edit, then type an alias name (forexample, Port1).

5. On configurations with two or more HBAs, repeat step3 and step4 for each host port that you need todefine, then proceed to step 6.

6. Click Next. The Specify Host Type window opens.7. Select the correct host type from the drop down menu and click Next. If you are configuring storage

for IBM i, make sure you select IBM i from the Host type (operating system) list.

Note: In advanced setups, a LUN 0 might be assigned to a host group or host definition that will notallow a IBM i as the host type. To fix this problem, remove the LUN 0 setting, change the operatingsystem to IBM i, and then add the LUN that was previously removed.The Review window opens.Failure to change the host type from the default to a specific host operating system will causeundesired results.

8. Review the information for accuracy, and make any necessary changes. Then click Next.9. After Storage Manager finishes defining the host and host ports, a dialog window opens. If you need

to define another host, select Define another host. To finish, click Exit. The wizard closes.

Mapping LUNs to a storage partitionThis section explains how to map LUNs to a storage partition using the following procedures:v “Mapping LUNs to a new partition”v “Adding LUNs to an existing partition” on page 4-13

Mapping LUNs to a new partitionThe following procedure enables you to map LUNs to a newly created partition:1. Select the Mappings view of the Subsystem Management window.2. In the Topology section, right-click the host on which you want to map LUNs, and select Define

Storage Partitioning. The Define Storage Partitioning window opens.3. In the Define Storage Partitioning window, select Host, then click Next.4. Select the logical drive by name, on the right side of the window.5. Accept the default LUN ID, or change it, then click Add.6. Repeat Step 5 for each LUN that you want to map to the partition.


Note: You can also use the Storage Partitioning wizard feature of the Storage Manager task assistant tomap LUNs to a new storage partition.

Adding LUNs to an existing partitionComplete the following steps to map new LUNs to an existing partition. Repeat these steps, as necessary,for each LUN you want to add to the partition.1. Select the Mappings view of the Subsystem Management window.2. In the Topology section, right-click the host or host group on which you want to map LUNs, and

select Define Additional Mappings. The Define Additional Mapping window opens.3. In the Define Additional Mapping window, select the following options, and then click Add:

v Host group or hostv Logical unit number (LUN)(0-255)v Logical drive

Configuring the IBM Systems Storage DS5100 and DS5300 for IBM i

Use the information in the following sections, in combination with the “Creating arrays and logicaldrives” on page 4-4 and “Defining a host group” on page 4-10 sections, to set up and assign IBM i LUNson the DS5300 and DS5100 storage subsystems with the DS Storage Manager software.

Assigning a port identifier for IBM i

When you use DS Storage Manager to enter a port identifier that IBM i will use, the port will be on thefirst adapter. IBM i requires two adapters to make a valid configuration. The following illustration showsthe set up screen where you assign the port identifier.


Defining IBM i as the host type

When you use DS Storage Manager to define a host type, select IBM i from the Host type (operatingsystem) list.

Important: In advanced setups, a LUN 0 might be assigned to a host group or host definition that willnot allow a IBM i as the host type. To fix this problem, remove the LUN 0 setting, change the operatingsystem to IBM i, and then add the LUN that was previously removed.The following illustration shows the set up screen where you define the IBM i as the host type.

Figure 4-1. Assigning a port identifier for IBM i


Optional premium featuresOptional premium features also available are FlashCopy, VolumeCopy, Remote Mirror, and Full DiskEncryption. For more information about these features, contact your IBM reseller or IBM marketingrepresentative.

Note: For more extensive information about these optional premium features, see the IBM System StorageDS Storage Manager Copy Services User's Guide.

Figure 4-2. Selecting IBM i as the host type


Creating a FlashCopy logical driveA FlashCopy logical drive is a logical point-in-time image of a logical drive, called a base logical drive. AFlashCopy logical drive has the following features:v It is created quickly and requires less disk space than an actual logical drive.v It can be assigned a host address, so that you can perform backups by using the FlashCopy logical

drive while the base logical drive is online and accessible.v You can use the FlashCopy logical drive to perform application testing or both scenario development

and analysis. This does not effect the actual production environment.v The maximum number of FlashCopy logical drives allowed is one half of the total logical drives

supported by your controller model.

For additional information about the FlashCopy feature and how to manage Flash Copy logical drives,refer to the Storage Manager Subsystem Management online help.

Important: The FlashCopy drive cannot be added or mapped to the same server that has the baselogical drive of the FlashCopy logical drive in a Windows 2000, Windows Server 2003, or NetWareenvironment. You must map the FlashCopy logical drive to another server.

Perform the following steps to create a FlashCopy logical drive:1. To ensure that you have the accurate point-in-time image of the base logical drive, stop applications

and flush cache I/O to the base logical drive.2. Open the Subsystem Management Window. From the Logical View, right-click the base logical drive.3. Select Create FlashCopy Logical Drive. The Create FlashCopy Logical Drive Wizard starts.4. Follow the on-screen instructions.5. Refer to the Subsystem Management online help for instructions on how to add the FlashCopy logical

drive to the host.

Using VolumeCopy

The VolumeCopy feature is a firmware-based mechanism for replicating logical drive data within astorage array. This feature is designed as a system management tool for tasks such as relocating data toother drives for hardware upgrades or performance management, data backup, or restoring snapshotlogical drive data. Users submit VolumeCopy requests by specifying two compatible drives. One drive isdesignated as the source and the other as the target. The VolumeCopy request is persistent so that anyrelevant result of the copy process can be communicated to the user. For more information about thisfeature, contact your IBM reseller or marketing representative.

Using the Remote Mirror option

The Remote Mirror option is a premium feature. It is used for online, real-time replication of databetween storage subsystems over a remote distance. In the event of a disaster or unrecoverable error atone storage subsystem, the Remote Mirror option enables you to promote a second storage subsystem totake over responsibility for normal input/output (I/O) operations. For more information about thisfeature, see the IBM Remote Support Manager for Storage - Planning, Installation and User's Guide, or contactyour IBM reseller or marketing representative.

Drive security with Full Disk Encryption

Full Disk Encryption (FDE) is a premium feature that prevents unauthorized access to the data on a drivethat is physically removed from the storage array. Controllers in the storage array have a security key.Secure drives provide access to data only through a controller that has the correct security key. FDE is apremium feature of the storage management software and must be enabled either by you or your storagevendor.


The FDE premium feature requires security capable drives. A security capable drive encrypts data duringwrites and decrypts data during reads. Each security capable drive has a unique drive encryption key.

When you create a secure array from security capable drives, the drives in that array become securityenabled. When a security capable drive has been security enabled, the drive requires the correct securitykey from a controller to read or write the data. All of the drives and controllers in a storage array sharethe same security key. The shared security key provides read and write access to the drives, while thedrive encryption key on each drive is used to encrypt the data. A security capable drive works like anyother drive until it is security enabled.

Whenever the power is turned off and turned on again, all of the security-enabled drives change to asecurity locked state. In this state, the data is inaccessible until the correct security key is provided by acontroller.

You can view the FDE status of any drive in the storage array from the Drive Properties dialog. Thestatus information reports whether the drive is:v Security Capablev Secure—Security enabled or disabledv Read/Write Accessible—Security locked or unlocked

You can view the FDE status of any array in the storage array from the Array Properties dialog. Thestatus information reports whether the storage array is:

Table 4-3. Array Security Properties

Security Capable—yes Security Capable—no

Secure—yes The array is composed of all FDEdrives and is in a Secure state.

Not applicable. Only FDE drives canbe in a Secure state.

Secure—no The array is composed of all FDEdrives and is in a Non-Secure state.

The array is not entirely composed ofFDE drives.

When the FDE premium feature has been enabled, the Drive Security menu appears in the Storage Arraymenu. The Drive Security menu has these options:v Create Security Keyv Change Security Keyv Save Security Keyv Unlock Drives

Note: If you have not created a security key for the storage array, only the Create Security Key option isactive. If you have created a security key for the storage array, the Create Security Key option is inactivewith a check mark to the left. The Change Security Key option and the Save Security Key option are nowactive.

The Unlock Drives option is active if there are any security locked drives in the storage array.

When the FDE premium feature has been enabled, the Secure Drives option appears in the VolumeGroup menu. The Secure Drives option is active if these conditions are true:v The selected storage array is not security enabled but is comprised entirely of security capable drives.v The storage array contains no snapshot base volumes or snapshot repository volumes.v The volume group is in an Optimal state.v A security key is set up for the storage array.

The Secure Drives option is inactive if the conditions are not true.


The Secure Drives option is inactive with a check mark to the left if the array is already security enabled.

You can erase security-enabled drives so that you can reuse the drives in another array or in anotherstorage array. When you erase security-enabled drives, you make sure that the data cannot be read. Whenall of the drives that you have selected in the Physical pane are security enabled, and none of the selecteddrives is part of an array, the Secure Erase option appears in the Drive menu.

For more information about FDE drives and working with Full Disk Encryption, see Chapter 6, “Workingwith full disk encryption,” on page 6-1. See also Appendix L, “FDE best practices,” on page L-1 forinformation about maintaining security on storage systems equipped with FDE disks.

Other features

Controller cache memory

Write caching enables the controller cache memory to store write operations from the host computer,which improves system performance; however, a controller can fail with user data in its cache that hasnot been transferred to the logical drive. Also, the cache memory can fail while it contains unwrittendata. Write-cache mirroring protects the system from either of these possibilities. Write-cache mirroringenables cached data to be mirrored across two redundant controllers with the same cache size. The datathat is written to the cache memory of one controller is also written to the cache memory of the othercontroller. That is, if one controller fails, the other controller completes all outstanding write operations.

Note: You can enable the write-cache mirroring parameter for each logical drive but when write-cachemirroring is enabled, half of the total cache size in each controller is reserved for mirroring the cache datafrom the other controller.

To prevent data loss or damage, the controller writes cache data to the logical drive periodically. Whenthe cache holds a specified start percentage of unwritten data, the controller writes the cache data to thelogical drive. When the cache is flushed down to a specified stop percentage, the flush is stopped. Forexample, the default start and stop settings for a logical drive are 80% and 20% of the total cache size,respectively. With these settings, the controller starts flushing the cache data when the cache reaches 80%full and stops flushing cache data when the cache is flushed down to 20% full. For maximum data safety,you can choose low start and stop percentages, for example, a start setting of 25% and a stop setting of0%; however, these low start and stop settings increase the chance that data that is needed for a hostcomputer read will not be in the cache, decreasing the cache-hit percentage and, therefore, the I/Orequest rate. It also increases the number of disk writes necessary to maintain the cache level, increasingsystem overhead and further decreasing performance. If a power outage occurs, data in the cache that isnot written to the logical drive is lost, even if it is mirrored to the cache memory of both controllers;therefore, there are batteries in the controller enclosure that protect the cache against power outages.

Note: The controller battery backup CRU change interval is three years from the date that the backupbattery CRU was installed for all models of the following DS4000 Storage Subsystems only: FAStT200,FAStT500, DS4100, DS4300, and DS4400. There is no replacement interval for the cache battery backupCRU in other DS4000 Storage Subsystems.

The storage management software features a battery-age clock that you can set when you replace abattery. This clock keeps track of the age of the battery (in days) so that you know when it is time toreplace the battery.

Note:

1. For the FAStT200, DS4100, and DS4300 or DS4300 Turbo disk systems, the battery CRU is locatedinside each controller CRU.


2. For the DS4800, DS5100, and DS5300, the batteries CRU are located in the Interconnect-batteries CRU.Write caching is disabled when batteries are low or discharged. If you enable a parameter calledwrite-caching without batteries on a logical drive, write caching continues even when the batteries in thecontroller enclosure are removed.

Attention: For maximum data integrity, do not enable the write-caching without batteries parameter,because data in the cache is lost during a power outage if the controller enclosure does not have workingbatteries. Instead, contact IBM service to get a battery replacement as soon as possible to minimize thetime that the subsystem is operating with write-caching disabled.

Persistent Reservations

Attention: The Persistent Reservations option should be used only with guidance from an IBMtechnical-support representative.

The Persistent Reservations option enables you to view and clear volume reservations and associatedregistrations. Persistent reservations are configured and managed through the cluster server software, andprevent other hosts from accessing particular volumes.

Unlike other types of reservations, a persistent reservation is used to perform the following functions:v Reserve access across multiple host ports—Provide various levels of access controlv Query the storage array about registered ports and reservationsv Provide for persistence of reservations in the event of a storage system power loss

The storage management software allows you to manage persistent reservations in the SubsystemManagement window. The Persistent Reservation option enables you to perform the following tasks:v View registration and reservation information for all volumes in the storage arrayv Save detailed information about volume reservations and registrationsv Clear all registrations and reservations for a single volume or for all volumes in the storage array

For detailed procedures, see the Subsystem Management Window online help. You can also managepersistent reservations through the script engine and the command line interface. For more information,see the Enterprise Management Window online help.

Media scan

A media scan is a background process that runs on all logical drives in the storage subsystem for which itis enabled, providing error detection on the drive media. Media scan checks the physical disks for defectsby reading the raw data from the disk and, if there are errors, writing it back. The advantage of enablingmedia scan is that the process can find media errors before they disrupt normal logical-drive read andwrite functions. The media scan process scans all logical-drive data to verify that it is accessible.

Note: The background media scan operation does not scan hot-spare or unused optimal hard drives(those that are not part of a defined logical drive) in a DS3000, DS4000, or DS5000 Storage Subsystemconfiguration. To perform a media scan on hot-spare or unused optimal hard drives, you must convertthem to logical drives at certain scheduled intervals and then revert them back to their hot-spare orunused states after you scan them.

There are two ways in which media scan can run:v Background media scan is enabled with logical drive redundancy data checks not enabled.

When redundancy checking is not enabled, the DS3000, DS4000, or DS5000 Storage Subsystem scans allblocks in the logical drives, including the redundancy blocks, but it does not check for the accuracy ofthe redundancy data.


This is the default setting when using Storage Manager to create logical drives and it is recommendedthat you not change this setting.

v Background media scan is enabled with logical drive redundancy data checks enabled.For RAID-3, RAID-5, or RAID-6 logical drives, a redundancy data check scans the data blocks,calculates the redundancy data, and compares it to the read redundancy information for each block. Itthen repairs any redundancy errors, if required. For a RAID-1 logical drive, a redundancy data checkcompares data blocks on mirrored drives and corrects any data inconsistencies.This setting is not recommended due to the effect redundancy checking has on the server performance.

When enabled, the media scan runs on all logical drives in the storage subsystem that meet the followingconditions:v The logical drive is in an optimal statusv There are no modification operations in progressv The Media Scan parameter is enabled

Note: The media scan must be enabled for the entire storage subsystem and enabled on each logicaldrive within the storage subsystem to protect the logical drive from failure due to media errors.

Media scan only reads data stripes, unless there is a problem. When a block in the stripe cannot be read,the read comment is retried a certain number times. If the read continues to fail, the controller calculateswhat that block should be and issues a write-with-verify command on the stripe. As the disk attempts tocomplete the write command, if the block cannot be written, the drive reallocates sectors until the datacan be written. Then the drive reports a successful write and Media Scan checks it with another read.There should not be any additional problems with the stripe. If there are additional problems, the processrepeats until there is a successful write, or until the drive is failed due to many consecutive write failuresand a hotspare takes over. Repairs are only made on successful writes and the drives are responsible forthe repairs. The controller only issues writewithverify commands. Therefore, data stripes can be readrepeatedly and report bad sectors but the controller calculates the missing information with RAID.

In a DS3000, DS4000, or DS5000 dual controller storage subsystem, there are two controllers handling I/O(Controllers A and B). Each logical drive that you create has a preferred controller which normallyhandles I/O for it. If a controller fails, the I/O for logical drives owned by the failed controller fails overto the other controller. Media scan I/O is not impacted by a controller failure and scanning continues onall applicable logical drives when there is only one remaining active controller.

If a drive is failed during the media scan process due to errors, normal reconstruction tasks are initiatedin the controllers operating system and Media Scan attempts to rebuild the array using a hotspare drive.While this reconstruction process occurs, no more media scan processing is done on that particular array.

Note: Because additional I/O reads are generated for media scanning, there might be a performanceimpact depending on the following factors:v The amount of configured storage capacity in the DS3000, DS4000, or DS5000 Storage Subsystem.

The greater the amount of configured storage capacity in the DS3000, DS4000, or DS5000 StorageSubsystem, the higher the performance impact is.

v The configured scan duration for the media scan operations.The longer the scan, the lower the performance impact.

v The status of the redundancy check option (enabled or disabled).If redundancy check is enabled, the performance impact is higher due to the need to read the data andrecalculated.

Errors reported by a media scanThe media scan process runs continuously in the background when it is enabled. Every time a scan cycle(that is, a media scan of all logical drives in a storage subsystem) completes, it restarts immediately. The


media scan process discovers any errors and reports them to the storage subsystem major event log(MEL). “Errors reported by a media scan” on page 4-20 lists the errors that are discovered during amedia scan.

Table 4-4. Errors discovered during a media scan

Error Description and result

Unrecovered media error The drive could not read the data on its first attempt, or on any subsequentattempts.

For logical drives or arrays with redundancy protection (RAID-1, RAID-3 andRAID-5), data is reconstructed, rewritten to the drive, and verified. The error isreported to the event log.

For logical drives or arrays without redundancy protection (RAID-0 anddegraded RAID-1, RAID-3, RAID-5, and RAID-6 logical drives), the error is notcorrected but is reported to the event log.

Recovered media error The drive could not read the requested data on its first attempt but succeeded ona subsequent attempt.

The data is rewritten to the drive and verified. The error is reported to the eventlog.

Note: Media scan makes three attempts to read the bad blocks.

Redundancy mismatches Redundancy errors are found.

The first 10 redundancy mismatches that are found on a logical drive arereported to the event log.

Note: This error could occur only when the optional redundancy checkbox isenabled, when the media scan feature is enabled, and the logical drive or array isnot RAID-0.

Unfixable error The data could not be read and parity or redundancy information could not beused to regenerate it. For example, redundancy information cannot be used toreconstruct data on a degraded logical drive.

The error is reported to the event log.

Media scan settingsTo maximize the protection and minimize the I/O performance impact, the DS3000, DS4000, or DS5000Storage Subsystem is shipped with the following default media scan settings:v The media scan option is enabled for all logical drives in the storage subsystem. Therefore, every time

a logical drive is created, it is created with the media scan option enabled. If you want to disablemedia scanning, you must disable it manually for each logical drive.

v The media scan duration is set to 30 days. This is the time in which the DS3000, DS4000, and DS5000controllers must complete the media scan of a logical drive. The controller uses the media scanduration, with the information about which logical drives must be scanned, to determine a constantrate at which to perform the media scan activities. The media scan duration is maintained regardless ofhost I/O activity.Thirty days is the maximum duration setting. You must manually change this value if you want toscan the media more frequently. This setting is applied to all logical drives in the storage subsystem.For example, you cannot set the media scan duration for one logical drive at two days and the otherslogical drives at 30 days.

v By default, the redundancy check option is not enabled on controller firmware versions earlier than7.60.39.00. For controller firmware versions earlier than 7.60.39.00, you must manually set this optionfor each of the logical drives on which you want to have redundancy data checked.


For controller firmware version 7.60.39.00 and later, the redundancy check option is enabled as adefault setting for any newly created logical drives. If you want an existing logical drive that wascreated prior to installing version 7.60.39.00 or later to have the redundancy check option enabled, youmust enable the option manually.Without redundancy check enabled, the controller reads the data stripe to see that all the data can beread. If it reads all the data, it discards the data and moves to the next stripe. When it cannot read ablock of data, it reconstructs the data from the remaining blocks and the parity block and issues awrite with verify to the block that could not be read. If the block has no data errors, media scan takesthe updated information and verifies that the block was fixed. If the block cannot be rewritten, thedrive allocates another block to take the data. When the data is successfully written, the controllerverifies that the block is fixed and moves to the next stripe.

Note: With redundancy check, media scan goes through the same process as without redundancycheck, but, in addition, the parity block is recalculated and verified. If the parity has data errors, theparity is rewritten. The recalculation and comparison of the parity data requires additional I/O whichcan affect performance.

Important: Changes to the media settings will not go into effect until the current media scan cyclecompletes.

To change the media scan settings for the entire storage subsystem, perform the following steps:1. Select the storage subsystem entry in the Logical/Physical view of the Subsystem Management

window.2. Click Storage Subsystem → Change → Media Scan Settings.

To change the media scan settings for a given logical drive, perform the following steps:1. Select the logical drive entry in the Logical/Physical view of the Subsystem Management window.2. Click Logical Drive → Change → Media Scan Settings.

Media scan durationWhen media scan is enabled, a duration window is specified (in days) which indicates how long thestorage subsystem will give the media scan process to check all applicable logical drives. The durationwindow can be shortened or increased to meet the customer requirements. The shorter the duration, themore often a drive is scanned and consequently, the more robust the situation will be. However, the moreoften a drive is scanned, the higher the performance impact.

Whenever the storage subsystem has some idle time, it starts or continues media scanning operations. Ifapplication generated disk I/O work is received, it gets priority. Therefore, the media scan process canslow down, speed up, or in some cases be suspended as the work demands change. If a storagesubsystem receives a great deal of application-generated disk I/O, it is possible for the Media Scan to fallbehind in its scanning. As the storage subsystem gets closer to the end of the duration window duringwhich it should finish the media scan, the background application starts to increase in priority (i.e. moretime is dedicated to the media scan process). This increase in priority only increases to a certain pointbecause the DS3000, DS4000, and DS5000 Storage Subsystem priority is process application-generateddisk I/O. In this case, it is possible that the media scan duration will be longer than the media scanduration settings.

Note: If you change the media scan duration setting, the changes will not take effect until the currentmedia scan cycle completes or the controller is reset.


Chapter 5. Configuring hosts

Upon completion of the tasks required for configuring the storage subsystems, your next step is to enableall hosts to see the storage subsystems.v “Booting a host operating system using SAN boot”v “Using multipath drivers to monitor I/O activity” on page 5-3v “Identifying devices” on page 5-27v “Configuring devices” on page 5-30

Booting a host operating system using SAN boot

SAN boot, also referred to as remote boot, is the ability to alternately boot the host operating system froma Storage Area Network device. In this case, the device is a LUN from a DS3000, DS4000, or DS5000Storage Subsystem. Some advantages to putting the boot device on the SAN include the following:v Server consolidation—As customers move to thin diskless servers that take up less space, each server

can now boot from an image of the operating system on the SAN.v Simplifies recovery from server failures—Operating system reinstallation is not required.v Rapid disaster recovery—The local SAN can be replicated at a remote recovery site.

The following are requirements and recommendations:v SAN configuration, zoning of boot devices, multipath configurationsv Active path to boot LUN

During the installation process, prior to installing and enabling a multipath driver, only one path to theboot LUN should be enabled.

v HBA BIOSSelectable boot, or boot bios, must be enabled.

Complete the following steps:1. SAN configuration

v ZoningSAN zoning is a method of arranging Fibre Channel devices into logical groups over the physicalconfiguration of the fabric. Each device in a SAN may be placed into multiple zones.

v Single path (active) to boot LUNDuring the installation, it is important that you remove all paths except one to your boot LUN. Thisis defined as the primary path between a host HBA and the controller owning the selected bootLUN. This can be done by doing a port disable on the switch for the other physical paths.

2. Configuring the storage arrayCreate LUNs and map to the host. (You will need to know the HBA WWNN, which you can get fromthe adapter label prior to installation in the server.)

3. Configuring the HBAs for boot from SANv QLogic

Configuring a QLogic Fibre Channel HBA device is a prerequisite of the SAN installation. Theconfiguration steps are provided here.a. Verify that the QLogic Fibre Channel HBA devices configured for the host have been set up to

have their boot BIOS enabled. This will allow for booting the installation LUN after the initialinstallation.


b. During power-on of the host, select Ctrl+Q to enter the QLogic boot BIOS.c. Select the HBA to be used for booting. This would be the HBA directly connected to, or zoned

to, the LUNs preferred controller.d. Configure the DS3000, DS4000, or DS5000 device from which the system will be booted. If the

utility cannot see the correct device or LUNs, check the SAN and DS3000, DS4000, or DS5000configurations before continuing. For more information, refer to the QLogic BIOS setupdocumentation. Before proceeding with the final steps for configuring the QLogic Fibre ChannelHBAs, you must perform setup of the HBAs on the storage array. Follow the steps here exactlyas they are presented.1) The HBA must be logged into the storage subsystem; and, even though no LUN will be

available yet, you can use the BIOS to discover the storage subsystem.2) Once the storage subsystem has discovered the HBA WWPNs, you must configure them as

the HBAs to the boot LUN, using the host mapping procedures.3) The BIOS can now be used to discover the newly configured LUN.

e. Save and exit.f. Reboot the server.

v EmulexConfiguring an Emulex Fibre Channel HBA device is a prerequisite of the SAN installation. Theconfiguration steps are provided here.

Note: It is assumed that the boot BIOS has been installed on the Emulex card. This is usually doneas part of the firmware installation with an Emulex utility such as HBAnywhere or LP6DUTIL.Before continuing, verify that the boot BIOS is installed.a. During power-on of the host, select Alt+E to enter the Emulex boot BIOS.b. Select the HBA to be used for booting. This would be the HBA directly connected to, or zoned

to, the LUNs preferred controller.c. Configure the adapter parameters, so the boot BIOS is enabled.d. Configure the BIOS, so the boot LUN (defined as part of the DS3000, DS4000, or DS5000 setup)

is available and selected as the preferred boot device. For more information, refer to the EmulexBIOS setup documentation. Before proceeding with the final steps for configuring the EmulexFibre Channel HBA, you must perform setup of the HBA on the storage array. Follow the stepshere exactly as they are presented.1) The HBA must be logged into the storage subsystem, and even though no LUN is available

yet, you can use the BIOS to discover the storage subsystem.2) Once the storage subsystem has discovered the HBA WWPNs, you must configure them as

the HBAs to the boot LUN, using the host mapping procedures.3) The BIOS can now be used to discover the newly configured LUN.

e. Press x to save.f. Reboot the server.

4. Starting the installation by booting from the installation mediav Select SAN LUN

During the installation, your operating system media will ask on which drive (or LUN) you wish toperform the installation. Select the drive that corresponds to your DS3000, DS4000, or DS5000device.

v Have drivers availableIn some cases such as with Windows, there is a point during the installation where it will ask youfor third party device drivers. This is where you will be able to select the HBA driver that you haveavailable on another media such as a floppy disk.

v Disk partitioning


In general, you should choose the default partitioning. Make sure the LUN you choose is largeenough for the operating system. For Linux, and most other operating systems, 20 GB is enough forthe boot device. For swap partitions, make sure the size is at least the size of your server's physicalmemory.

5. Completing the installationv You must complete the following steps to finish the SAN boot procedure:

a. Reboot again, and open the boot options menu. This time, the boot device that you set up in theprevious steps is ready to be used. Select the option to boot from a hard drive/SAN, and selectthe HBA that is associated with the SAN disk device on which the installation was completed.

b. The installation boot device should now be listed in the bootable devices that are discovered onthe selected HBA. Select the appropriate device, and boot.

c. Set the installed boot device as the default boot device for the system.

Note: This step is not required. However, it is recommended to enable unattended reboots afterthis procedure is complete.

If all of the preceding steps were completed accurately, the system is now booted in single-pathmode from the SAN. Complete the following steps to verify the installation:a. Check the mounted devices, and verify that the root is mounted in the appropriate location.b. Verify that the swap device and other configured partitions are correctly mounted.This completes the single-path SAN boot procedure for systems.

v Multipath driverFor additional information, see “Using multipath drivers to monitor I/O activity.”

v Enable all pathsAdditional paths between the storage array and server can now be added. If the server is going tobe used to manage the storage array, Storage Manager can now be installed on the server.To complete the Linux installation, perform the following steps:a. Verify that the persistent binding for /var/mpp/devicemapping is up-to-date. The

/var/mpp/devicemapping file tells RDAC which storage array to configure first. If additionalstorage arrays will be added to the server, the storage array with the boot/root volume shouldalways be first in the device mapping file. To update this file, run the following command:# mppUpdate

b. After the # mppUpdate command is run, cat the /var/mpp/devicemapping file, using the followingcommand:# cat /var/mpp/devicemapping 0:<DS4x00 SAN Boot Device>

The storage array for the boot/root volume should be at entry 0. If the boot/root volume is notat entry 0, edit the file to reorder the storage array entries so the array for the boot/root volumeis at entry 0. Run the # mppUpdate command, again. The installation is now complete.

Using multipath drivers to monitor I/O activity

Host systems that are attached to the DS3000, DS4000, or DS5000 storage for I/O activity require amultipath driver (sometimes referred to as an RDAC or failover driver) for Fibre Channel pathredundancy. The multipath driver monitors I/O paths. If a component failure occurs in one of the FibreChannel paths, the multipath driver reroutes all I/O to a different path. Your multipath driver willdepend on the operating system you have installed. See Table 5-1 on page 5-4.

In the Microsoft Windows environment another multipath driver, referred to as Windows RDAC, waspreviously provided with Storage Manager host software version 9.xx.__.__ and earlier. Support forWindows RDAC was terminated with the release of Storage Manager host software version 10.xx.__.__

Chapter 5. Configuring hosts 5-3

and later in conjunction with controller firmware version 7.xx.xx.xx and later. In addition, support forAIX fcp_array is being phased out. AIX fcp_array users should migrate to the AIX MPIO multipathdriver at the earliest time window.

An IBM Fibre Channel host bus adapter (HBA) provides the interface between a host server and aDS3000, DS4000, or DS5000 Storage Subsystem. IBM DS3000, DS4000, and DS5000 Fibre Channel HBAsare high-performance, direct memory access, bus-master, host adapters that are designed for high-endsystems. These HBAs support all Fibre Channel peripheral devices that support private-loop,direct-attach, and fabric-loop attachment. The IBM Host Adapter device driver enables your operatingsystem to communicate with the Fibre Channel HBA.

Table 5-1. Multipath driver by operating system

Operating system Multipath driver

AIX fcp_array (also called RDAC), MPIO, or SDDPCM

HP-UX LVM , native multipathing, or IBM SDD

Linux MPP (also called RDAC) or Veritas DMP

NetWare Novell MPE

Solaris RDAC, MPxIO or Veritas DMP

SVC SDD

VMWare NMP

Windows MPIO DSM or Veritas DMP DSM

Before you begin: With the exception of Windows MPIO, multipath driver files are not included on theDS Storage Manager installation CD. Check the Storage Manager README file for the minimum file setversions required for your operating system. To learn how to find the README files on the Web, see“Finding Storage Manager software, controller firmware, and README files” on page xiii. To install themultipath driver, follow the instructions in “Steps for installing the multipath driver” on page 5-6.

Multipathing refers to a host's ability to recognize multiple paths to the storage device. This is done byutilizing multiple HBA ports or devices within the host server connected to SAN fabric switches, whichare also connected to the multiple ports on the storage devices. For the storage products referred to asDS3000, DS4000, or DS5000, these devices have two controllers within the subsystem that manage andcontrol the disk drives. These controllers behave in an active/passive fashion. Ownership and control of aparticular LUN is done by one controller. The other controller is in a passive mode until a failure occurs,at which time the LUN ownership is transferred to that controller. Each controller may have more thanone fabric port for connectivity to the SAN fabric.

Figure 5-1 on page 5-5 shows a sample multipath configuration for all supported operating systemsexcept AIX fcp_array and Solaris RDAC multipath configurations. Figure 5-2 on page 5-5 shows a samplemultipath configuration for the AIX fcp_array, Microsoft Windows RDAC (no longer supported), andSolaris RDAC multipath configurations.


Most multipath drivers can support multiple paths. Table 5-2 on page 5-6 shows the number of pathseach driver can support. Note that the AIX fcp_array and Solaris RDAC can only support two paths, oneto each controller.

FC Switch FC Switch

Server 1 Server 2

HBA 1 HBA 2 HBA 1 HBA 2

Storage subsystem

Controller A Controller B

Figure 5-1. Host HBA to storage subsystem controller multipath sample configuration for all multipath drivers exceptAIX fcp_array and Solaris RDAC

FC Switch FC Switch

Server 1 Server 2


Storage subsystem


Figure 5-2. Host HBA to storage subsystem controller multipath sample configuration for AIX fcp_array and SolarisRDAC multipath drivers


Table 5-2. Number of paths each multipath driver supports by operating system

Driver Number of paths Default

AIX MPIO unlimited Not Applicable

AIX RDAC 2 Not Applicable

HP-UX native 65536 Not Applicable

HP-UX PVlinks 8192 Not Applicable

Linux MPP unlimited 4

Linux Veritas DMP unlimited Not Applicable

Solaris MPxIO unlimited Not Applicable

Solaris RDAC 2 Not Applicable

Solaris Veritas DMP unlimited Not Applicable

SVC 32 Not Applicable

VMWare unlimited - 8 or less recommended Not Applicable

Windows MPIO DSM 32 paths per LUN, 16 per controller 4

Windows Veritas DMP DSM unlimited Not Applicable

Steps for installing the multipath driverYou must install a multipath driver on all hosts attached to your storage subsystem, whether or not thesehosts will have multiple paths to the storage subsystem. This section describes how to check the currentmultipath driver program driver version level, update the multipath device driver, and verify that themultipath update is complete.

Windows MPIO or MPIO/DSMThis multipath driver is included in the DS Storage Manager host software package for Windows. MPIOis a DDK kit from Microsoft for developing code that manages multipath devices. The DDK kit contains acore set of binary drivers, which are installed with the IBM DS3000, DS4000, or DS5000 Device SpecificModule (DSM) and which are designed to provide a transparent system architecture relying on MicrosoftPlug and Play. These binary drivers provide LUN multipath functionality while maintaining compatibilitywith existing Microsoft Windows device driver stacks. The MPIO driver performs the following tasks:v Detects and claims the physical disk devices presented by the Storage Subsystems based on

Vendor/Product ID strings and manage the logical paths to the physical devicesv Presents a single instance of each LUN to the rest of the Windows operating systemv Provides an optional interface via WMI for use by user-mode applicationsv Relies on the vendor's (IBM) customized Device-Specific Module (DSM) for the information on the

behavior of storage subsystem devices on the following:– I/O routing information– Conditions requiring a request to be retried, failed, failed over or fail-back; for example,

Vendor-Unique errors– Handles miscellaneous functions such as Release/Reservation commands

Multiple Device-Specific Modules (DSMs) for different disk storage subsystems can be installed in thesame host server.

Storport miniport HBA device driverFor Windows operating systems, Storage Manager provides the MPIO DSM device driver that is based onthe Microsoft Storport Miniport device driver model.


The Storport miniport device driver model was introduced in the Microsoft Windows Server 2003 releaseas a replacement for the SCSIport miniport device driver model. It is the only supported device drivermodel for Windows Server 2003 and Windows Server 2008 editions, supporting the AMD64 and EM64Tservers. It does not support the buschange=0 parameter to bypass the Microsoft Windows operatingsystem Plug and Play driver; instead, it works with the Plug and Play driver to detect the removal andinsertion of devices at the fibre channel host bus adapter port.

Currently, only the DS4100, DS4200, DS4300 standard/base or turbo models, DS4400, DS4500, DS4700 andDS4800 Storage Subsystems support this Storport-based device driver. The DS4100, DS4300 standard/baseor turbo models, DS4400 and DS4500 subsystem models must have controller firmware version 6.12.27.xxor higher installed.

See the Storage Manager README file for Microsoft Windows operating systems for any other additionalrequirements, such as controller firmware versions or updates.

SCSIport miniport HBA device driverFor the Windows 2000 operating system environment, only the device drivers based on the SCSIportminiport device driver (not the Storport model) are currently supported.

In previous SCSIport device driver releases, the buschange=0 parameter allows the RDAC multipathdriver to control and monitor device insertion and removal from the HBA port by preventing theMicrosoft Plug and Play device drive from managing the HBA port. This new SCSIport device driverversion that is used with MPIO does not support the buschange=0 parameter.

Attention: Not all DS4000 and DS5000 controller firmware versions support this functionality. OnlyDS4000 and DS5000 controller firmware versions 06.12.27.xx (and later) for DS4300 standard/base orturbo models, and DS4500 subsystems or versions 6.16.8x.xx (and later) for DS4200, DS4700 and DS4800subsystems support this new SCSIport miniport device driver.

Before installing the device driver, check the README file that is included in the device driver packagefile, as well as the README file included with the DS Storage Manager host software for Windows, tosee which device drivers and controller firmware versions are supported for DS3000, DS4000, or DS5000Storage Subsystems. See “Finding Storage Manager software, controller firmware, and README files” onpage xiii to find out how to access the most recent Storage Manager README files on the Web. Followthe README device driver installation instructions associated with your operating system.

Note: Read the device driver README for any required modifications to the default HBA BIOS and hostoperating system registry settings to ensure optimal performance. If you make any changes to the HBABIOS settings, the machine must be rebooted for the changes to be enabled.

For more information, see the Installation and User's Guide for your particular Fibre Channel HBAmodel.

Veritas DMP DSM driverSee the Symantec Storage Foundation for Windows documentation for instructions about installing theVeritas DMP DSM driver at http://www.symantec.com/business/support/.

AIX multipath driversAn AIX host system requires either the AIX Redundant Disk Array Controller (RDAC) or the MPIO failoverdriver for fibre channel path redundancy. In supported Veritas environments, RDAC is the supportedfailover driver.

The failover driver monitors I/O paths. If a component failure occurs in one of the Fibre Channel paths,the failover driver reroutes all I/O to another path.


http://www.symantec.com/business/support/

Note: AIX supports both Redundant Disk Array Controller (RDAC) and Multiple Path IO. Thesemultipath drivers are part of the native AIX operating system. See the AIX installation guide for detailsabout the installation of these drivers.

Linux MPP driverThis section describes how to install the MPP (RDAC) driver for a Linux configuration.

Important: Before you install MPP, make sure that the partitions and LUNs are configured and assignedand that the correct HBA driver is installed.

Complete the following steps to install MPP:1. Download the MPP driver package from the IBM DS3000, DS4000, or DS5000 System Storage Disk

Support Web site.2. Create a directory on the host and download the MPP driver package to that directory.3. Uncompress the file by typing the following command:

# tar -zxvf rdac-LINUX-package_version-source.tar.gz

where package_version is the SLES or RHEL package version number. Result: A directory calledlinuxrdac-version# or linuxrdac is created.

4. Open the README that is included in the linuxrdac-version# directory.5. In the README, find the instructions for building and installing the driver and complete all of the

steps.

Note: Make sure you reboot the server before you proceed to the next step.6. Type the following command to list the installed modules:

# lsmod

7. Verify that module entries are included in the lsmod list, as follows:

Module entries for SLES or RHEL:

v mppVhbav mppUpperv lpfc (or qla2xxx for BladeCenter configurations)v lpfcdfc (if ioctl module is installed)

Note: If you do not see the mpp_Vhba module, the likely cause is that the server was rebooted beforethe LUNs were assigned, so the mpp_Vhba module was not installed. If this is the case, assign theLUNs now, reboot the server, and repeat this step.

8. Type the following command to verify the driver version:

# mppUtil -V

Result: The Linux multipath driver version displays.9. Type the following command to verify that devices are configured with the RDAC driver:

# ls -1R /proc/mpp

Result: An output similar to the following example displays:


# ls -1R /proc/mpp/proc/mpp:total 0dr-xr-xr-x 4 root root 0 Oct 24 02:56 DS4100-sys1crwxrwxrwx 1 root root 254, 0 Oct 24 02:56 mppVBusNode

/proc/mpp/ DS4100-sys1:total 0dr-xr-xr-x 3 root root 0 Oct 24 02:56 controllerAdr-xr-xr-x 3 root root 0 Oct 24 02:56 controllerB-rw-r--r-- 1 root root 0 Oct 24 02:56 virtualLun0-rw-r--r-- 1 root root 0 Oct 24 02:56 virtualLun1-rw-r--r-- 1 root root 0 Oct 24 02:56 virtualLun2-rw-r--r-- 1 root root 0 Oct 24 02:56 virtualLun3-rw-r--r-- 1 root root 0 Oct 24 02:56 virtualLun4-rw-r--r-- 1 root root 0 Oct 24 02:56 virtualLun5

/proc/mpp/ DS4100-sys1/controllerA:total 0dr-xr-xr-x 2 root root 0 Oct 24 02:56 lpfc_h6c0t2

/proc/mpp/ DS4100-sys1/controllerA/lpfc_h6c0t2:total 0-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN0-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN1-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN2-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN3-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN4-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN5

/proc/mpp/ DS4100-sys1/controllerB:total 0dr-xr-xr-x 2 root root 0 Oct 24 02:56 lpfc_h5c0t0

/proc/mpp/ DS4100-sys1/controllerB/lpfc_h5c0t0:total 0-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN0-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN1-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN2-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN3-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN4-rw-r--r-- 1 root root 0 Oct 24 02:56 LUN5

Note: After you install the RDAC driver, the following commands and man pages are available:v mppUtilv mppBusRescanv mppUpdatev RDAC

Veritas DMP driverSee the Symantec Storage Foundation for Windows documentation for instructions about installing theVeritas DMP driver at http://www.symantec.com/business/support/.

Note: The Array Support Library (ASL) that supports DMP on the DS3000, DS4000, or DS5000 mightneed to be loaded. The ASL might be a separate file available from Symantec or it might be integratedwith Volume Manager, depending on the version of Storage Foundation.



HP-UX PV-links

If an HP-UX system is attached with two host bus adapters to the DS3000, DS4000, or DS5000 StorageSubsystem, you can establish redundant access to storage by using physical volume links (PV-links), afeature of the HP-UX operating system. PV-links achieve access redundancy by using devices with bothprimary and secondary paths to the same device.

Important:

v There are two methods for establishing redundant access to storage using PV-links:– If you have DS3000, DS4000, or DS5000 controller firmware version 07.xx.xx.xx, 06.xx.xx.xx, or

05.xx.xx.xx installed, use method 1.– If you have DS4000 or DS5000 controller firmware version 04.xx.xx.xx installed, use method 2.

v For both methods, you must have SMutil installed on the host.

Using PV-links: Method 1If you have DS4000 or DS5000 controller firmware version 06.1.xx.xx or higher, or 05.xx.xx.xx installed,use the following procedure to enable multipath I/O by using PV-links:1. Run the hot_add command from HP-UX at the shell prompt. This command updates any new devices

that are created or added. A dump is generated. When the hot_add command runs, each new logicaldrive that is created in the Subsystem Management window represents a disk device to the operatingsystem.

#hot_add

2. Run the SMdevices command. The system provides a dump similar to the example in the table thatfollows. Notice that every logical drive and logical drive access unit has been assigned a logical unitnumber (LUN). Each logical drive has two paths. Each DS3000, DS4000, and DS5000 controller hasone logical drive access. For example, a subsystem that contains two DS3000, DS4000, or DS5000controllers has two logical drive accesses.

#SMdevices


/dev/rdsk/c166t0d0 [Storage Subsystem DS4000, Logical Drive Accounting, LUN 0,Logical Drive WWN <600a0b80000f56d00000001e3eaead2b>,Preferred Path (Controller-B): In Use]/dev/rdsk/c166t0d1 [Storage Subsystem DS4000, Logical Drive HR, LUN 1,Logical Drive WWN <600a0b80000f5d6c000000273eaeae30>,Alternate Path (Controller-B): Not In Use]/dev/rdsk/c166t0d2 [Storage Subsystem DS4000, Logical Drive Finance,LUN 2, Logical Drive WWN <600a0b80000f5d6c000000253eaeadf8>,Alternate Path (Controller-B): Not In Use]/dev/rdsk/c166t0d3 [Storage Subsystem DS4000, Logical Drive Purchasing,LUN 3, Logical Drive WWN <600a0b80000f5d6c000000243eaeadbe>,Alternate Path (Controller-B): Not In Use]/dev/rdsk/c166t0d4 [Storage Subsystem DS4000, Logical Drive Development,LUN 4, Logical Drive WWN <600a0b80000f56d00000001d3eaeacef>,Preferred Path (Controller-B): In Use]/dev/rdsk/c166t3d7 [Storage Subsystem DS4000, Logical Drive Access, LUN 31,Logical Drive WWN <600a0b80000f56d00000001b00000000>]

/dev/rdsk/c172t0d0 [Storage Subsystem DS4000, Logical Drive Accounting, LUN 0,Logical Drive WWN <600a0b80000f56d00000001e3eaead2b>,Alternate Path (Controller-A): Not In Use]/dev/rdsk/c172t0d1 [Storage Subsystem DS4000, logical Drive HR, LUN 1,Logical Drive WWN <600a0b80000f5d6c000000273eaeae30>,Preferred Path (Controller-A): In Use]/dev/rdsk/c172t0d2 [Storage Subsystem DS4000, Logical Drive Finance, LUN 2,Logical Drive WWN <600a0b80000f5d6c000000253eaeadf8>,Preferred Path (Controller-A): In Use]/dev/rdsk/c172t0d3 [Storage Subsystem DS4000, Logical Drive Purchasing, LUN 3,Logical Drive WWN <600a0b80000f5d6c000000243eaeadbe>,Preferred Path (Controller-A): In Use]/dev/rdsk/c172t0d4 [Storage Subsystem DS4000, Logical Drive Development, LUN 4,Logical Drive WWN <600a0b80000f56d00000001d3eaeacef>,Alternate Path (Controller-A): Not In Use]/dev/rdsk/c172t3d7 [Storage Subsystem DS4000, Logical Drive Access, LUN 31,Logical Drive WWN <600a0b80000f5d6c0000002200000000>]

Note: If you do not see the logical drives and logical drive accesses after running the hot_add andSMdevices commands, restart the HP-UX host by running the reboot command.

#reboot

3. Determine the preferred and alternate path for each logical drive by examining the output from theSMdevices command, as shown in the previous example. Notice that each device is listed twice; oneinstance is the preferred path and one instance is the alternate path.

Preferred pathIn the sample output that is shown below, the preferred path is /dev/rdsk/c166t0d0:

/dev/rdsk/c166t0d0 [Storage Subsystem DS4000, Logical DriveAccounting, LUN 0, Logical Drive WWN <600a0b80000f56d00000001e3eaead2b>,Preferred Path (Controller-B): In Use]

Alternate pathIn the sample output that is shown below, the alternate path is /dev/rdsk/c172t0d0:

/dev/rdsk/c172t0d0 [Storage Subsystem DS4000, Logical DriveAccounting, LUN 0, Logical Drive WWN <600a0b80000f56d00000001e3eaead2b>,Alternate Path (Controller-A): NotIn Use]


Using PV-links: Method 2If you have DS4000 or DS5000 controller firmware version 4.xx.xx.xx installed, use the followingprocedures to enable multipath I/O by using PV-links:v Determine the preferred and alternate pathsv Create the logical drives and logical drive groups

Determining preferred and alternate paths: Complete the following steps to determine the preferredand alternate paths:1. Run the hot_add command from HP-UX at the shell prompt. This command updates any new devices

that are created or added. A dump is generated. When the hot_add command runs, each new logicaldrive that is created in the Subsystem Management window represents a disk device to the operatingsystem.

#hot_add

2. Run the SMdevices command. The system provides a dump similar to the example in Table 5-3.Notice that every logical drive and logical drive access unit has been assigned a logical unit number(LUN). Each logical drive has two paths. Each DS3000, DS4000, or DS5000 controller have one logicaldrive access. For example, a subsystem that contains two DS3000, DS4000, or DS5000 controllers havetwo logical drive accesses.

#SMdevices

Table 5-3. Sample SMdevices command output (method 2)

/dev/rdsk/c166t0d0 [Storage Subsystem DS4000, Logical Drive Accounting, LUN 0,Logical Drive WWN <600a0b80000f56d00000001e3eaead2b>]

/dev/rdsk/c166t0d1 [Storage Subsystem DS4000, Logical Drive HR, LUN 1,Logical Drive WWN <600a0b80000f5d6c000000273eaeae30>]/dev/rdsk/c166t0d2 [Storage Subsystem DS4000, Logical Drive Finance, LUN 2,Logical Drive WWN <600a0b80000f5d6c000000253eaeadf8>]/dev/rdsk/c166t0d3 [Storage Subsystem DS4000, Logical Drive Purchasing, LUN 3,Logical Drive WWN <600a0b80000f5d6c000000243eaeadbe>]/dev/rdsk/c166t0d4 [Storage Subsystem DS4000, Logical Drive Development, LUN 4,Logical Drive WWN <600a0b80000f56d00000001d3eaeacef>]/dev/rdsk/c166t3d7 [Storage Subsystem DS4000,Logical Drive Access, LUN 31,Logical Drive WWN <600a0b80000f56d00000001b00000000>]

/dev/rdsk/c172t0d0 [Storage Subsystem DS4000, Logical Drive Accounting, LUN 0,Logical Drive WWN <600a0b80000f56d00000001e3eaead2b>]/dev/rdsk/c172t0d1 [Storage Subsystem DS4000, logical Drive HR, LUN 1,Logical Drive WWN <600a0b80000f5d6c000000273eaeae30>]/dev/rdsk/c172t0d2 [Storage Subsystem DS4000, Logical Drive Finance, LUN 2,Logical Drive WWN <600a0b80000f5d6c000000253eaeadf8>]/dev/rdsk/c172t0d3 [Storage Subsystem DS4000, Logical Drive Purchasing, LUN 3,Logical Drive WWN <600a0b80000f5d6c000000243eaeadbe>]/dev/rdsk/c172t0d4 [Storage Subsystem DS4000, Logical Drive Development, LUN 4,Logical Drive WWN <600a0b80000f56d00000001d3eaeacef>]/dev/rdsk/c172t3d7 [Storage Subsystem DS4000, Logical Drive Access, LUN 31,Logical Drive WWN <600a0b80000f5d6c0000002200000000>]

Note: If you do not see the logical drives and logical drive accesses after running the hot_add andSMdevices commands, restart the HP-UX host by running the reboot command.

#reboot

3. Determine the preferred and alternate path for each logical drive by examining the output from theSMdevices command, as shown in the example in Table 5-3.


Notice that each device is listed twice; one instance is the preferred path and one instance is thealternate path. Also, notice that each device has a worldwide name (WWN). Part of the WWN of eachlogical drive is unique for each controller in the DS3000, DS4000, or DS5000 Storage Subsystem. If youexamine the WWNs for the logical drive access in Table 5-3 on page 5-12, you notice that they differin only five digits, f56d0 and f5d6c.The devices in Table 5-3 on page 5-12 are viewed through the controllers c166 and c172. To determinethe preferred path of a specific logical drive seen by the operating system perform the following steps:a. Find the WWN for each logical drive access. In this case, Logical Drive Access 1 is associated with

c166 and has the WWN of f56d0.

/dev/rdsk/c166t3d7 [Storage Subsystem DS4000, Logical Drive Access, LUN 31,Logical Drive WWN <600a0b80000f56d00000001b00000000>]

Logical Drive Access 2 is associated with c172 and has the WWN of f5d6c:

/dev/rdsk/c172t3d7 [Storage Subsystem DS4000, Logical Drive Access, LUN 31,Logical Drive WWN <600a0b80000f5d6c0000002200000000>]

b. Identify the preferred device path name for the attached storage device by matching the logicaldrive WWN to a logical drive access WWN. In this case, the WWN for LUN 0 is associated withcontroller c166 and c172. Therefore, the preferred path for LUN 0 is /dev/rdsk/c166t0d0, which iscontroller c166:

/dev/rdsk/c166t0d0 [Storage Subsystem DS4000, Logical DriveAccounting, LUN 0, Logical Drive g<600a0b80000f56d00000001e3eaead2b>]

The alternate path is /dev/rdsk/c172t0d0, which is controller c172:

/dev/rdsk/c172t0d0 [Storage Subsystem DS4000, Logical DriveAccounting, LUN 0, Logical Drive WWN <600a0b80000f56d00000001e3eaead2b>]

c. To keep a record for future reference, enter this path information for LUN 0 into a matrix (similarto the one in Table 5-4).

Table 5-4. Sample record of logical drive preferred and alternate paths

LUN Logical drive name Preferred path Alternate path

0 Accounting /dev/rdsk/c166t0d0 /dev/rdsk/c172t0d0

1 HR /dev/rdsk/c172t0d1 /dev/rdsk/c166t0d1

2 Finance dev/rdsk/c172t0d2 /dev/rdsk/c166t0d2

3 Purchasing /dev/rdsk/c172t0d3 /dev/rdsk/c166t0d3

4 Development /dev/rdsk/c166t0d4 /dev/rdsk/c172t0d4

d. Repeat Step 3.a through Step 3.c, for each logical drive that is seen by the operating system.

Creating volumes and volume groups: After you have determined the preferred and alternate paths,and have recorded them in a matrix for future reference, perform the following steps to create volumesand volume groups.

Important: Do not use SAM for DS3000, DS4000, or DS5000 storage configuration; if you do, you mightget unexpected results.

Note: The steps in this procedure refer to LUN 0 in Table 5-4.1. Create a physical volume and define the primary paths for the attached storage devices. The primary

path will be the preferred path. Type the following command at the shell prompt:


#pvcreate /dev/rdsk/c166t0d0

The system confirms the creation of the new physical volume.2. Create volume groups.

Note: For more information on how to create volume groups, refer to HP-UX documentation or toman pages.a. Make a directory for volume group by typing the following commands. This directory must reside

in the /dev directory.

#cd /dev#mkdir /vg1

b. Create the group special file in the /dev directory for the volume group by typing the followingcommand:

#mknod /dev/vg1/group c 64 0x010000

c. Create a volume group and define physical volume names (primary link) for the attached storagedevice by typing the following command:

#vgcreate /dev/vg1/ /dev/dsk/c166t0d0

d. Define the secondary path name (alternate path) for the attached-storage device by typing thefollowing command:

#vgextend vg1 /dev/dsk/c172t0d0

Note: You can also use the vgextend command to add additional storage devices to an existingvolume group. Add the primary path first, then add the alternate path, as shown in the followingexample:1) Add the primary path for LUN1.


2) Add the secondary path for LUN1.


3. Create logical volumes. For more information, refer to HP-UX documentation.4. Create file systems for the logical volumes.5. Repeat step1 on page 5-13 through step 4 to create additional volume groups. For more information,

refer to HP-UX documentation.6. Verify the primary (preferred) and secondary (alternate) paths for each device by typing the following

command:

#vgdisplay -v vgname

where vgname is the volume group name.


HP-UX native multipathing

Multipathing is native to HP-UX 11i v3. It is built in to the mass storage subsystem, and it is available toapplications without any special configuration.

For more information about native multipathing in HP-UX 11iv3, consult the documentation found at thisURL:

http://docs.hp.com/en/oshpux11iv3.html

Solaris failover drivers

A failover driver monitors I/O paths. If a component failure occurs in one of the Fibre Channel paths, thefailover driver reroutes all I/O to another path.

Solaris host systems require one of the following failover drivers:v Solaris Multiplexed I/O (MPxIO)v RDACv Veritas VolumeManager with Dynamic Multipathing (DMP)

Note:

1. RDAC is not supported on Solaris 10. You must use either Solaris MPxIO or the Veritas DMP failoverdriver.

2. With Solaris 10, MPxIO capability is built in. If you want to use MPxIO with previous versions ofSolaris, you must install SUN StorEdge SAN Foundation Suite.

This section includes the following procedures:v “Installing the MPxIO driver”v “Installing the RDAC driver on Solaris” on page 5-22v “Installing the DMP driver” on page 5-24

Installing the MPxIO driverMultiplexed I/O (MPxIO) is a Sun Solaris multipath driver architecture. This failover driver enablesstorage arrays to be accessed through multiple host controller interfaces from a single instance of thestorage array. MPxIO helps protect against storage subsystem outages because of controller failures. Ifone controller fails, MPxIO automatically switches to an alternate controller.

MPxIO is fully integrated within the Solaris 10 operating system. For Solaris 8 and 9 operating systems,MPxIO is available as part of the Sun StorEdge SAN Foundation Suite, and must be installed separately.

For the latest supported version of Sun StorEdge SAN Foundation Suite, the latest Solaris kernel patches,and the most recent updates to information about using MPxIO, check the DS Storage Manager READMEfile for Solaris. (See “Finding Storage Manager software, controller firmware, and README files” on pagexiii for steps to find the README file on the Web.)

This section contains the following topics:v “Device name change considerations for MPxIO” on page 5-16v “Acquiring the latest MPxIO driver version” on page 5-16v “Steps for enabling the MPxIO failover driver” on page 5-16v “Disabling the MPxIO multipath driver” on page 5-22

Note: For more information, please refer to the following Sun documents, which you can find at the SunWeb site:


http://docs.hp.com/en/oshpux11iv3.html

http://docs.sun.comv Sun StorEdge SAN Foundation Software Installation Guide

v Sun StorEdge SAN Foundation Software Configuration Guide

v Sun Solaris Fibre Channel and Storage Multipathing Administration Guide

Device name change considerations for MPxIO:

In the /dev and /devices trees, devices are named differently from their original names when MPxIO isenabled. For example:

Device name with MPxIO disabled/dev/dsk/c1t1d0s0

MPxIO-enabled device name/dev/rdsk/c0t600A0B800011121800006B31452CC6A0d0s2

You must configure applications that directly consume the device to use the new names whenever theMPxIO configuration is enabled or disabled.

In addition, the /etc/vfstab file and the dump configuration also contain references to device names.When you use the stmsboot command to enable or disable MPxIO, as described in the next sections,/etc/vfstab and the dump configuration are automatically updated with the new device names.

Acquiring the latest MPxIO driver version: The method of acquiring MPxIO depends upon whichversion of Solaris you have installed, as described in the following list:

Solaris 10MPxIO is fully integrated within the Solaris 10 operating system, and does not need to beinstalled separately. MPxIO with Solaris 10 is updated using regular Solaris 10 patches, which areavailable at the following Sun Technical Support Web site:

http://sunsolve.sun.com

Note: It is recommended that you install the regular kernel jumbo patch, because there aredependencies between the various patches that make up the driver stack.

Solaris 8 and 9Because MPxIO is not included with Solaris 8 and 9, you must download the required SAN suite(Sun StorEdge SAN Foundation Suite) from the Sun Technical Support Web site:


On this page, click SAN 4.4 release Software/Firmware Upgrades & Documentation.

Note: Install the software using the provided install_it.ksh script.

Steps for enabling the MPxIO failover driver: This section describes how to enable MPxIO by usingthe stmsboot command. In addition to enabling MPxIO, this command also updates the device names inthe /etc/vfstab file and the dump configuration files during the next reboot.

Note: In Solaris 10, the stmsboot command is used to enable or disable MPxIO on all devices.

Before you begin:

1. Install the Solaris operating system, and the latest patches.2. Ensure that the Solaris host type was selected when the host was defined.

Steps for enabling MPxIO on Solaris 8 and 9:


http://docs.sun.com



1. Install the latest version of Sun StorEdge SAN Foundation Suite and required patches, using the SunStorEdge install_it script. For more information, see the Sun StorEdge SAN Foundation Suite x.xxInstallation Guide (where x.xx is the version of StorEdge software).

2. Edit the /kernel/drv/scsi_vhci.conf configuration file to ensure that the VID/PID is not specified inthis file. Also, ensure that the following entries exist in the file:mpxio-disable=”no”;load-balance=”none”;auto-failback=”enable”;

Exception: In a cluster environment where logical drives (LUNs) are shared between multiple Sunservers, you might need to set the auto-failback parameter to disable to prevent the followingphenomenon, which can occur when one of the servers has a failed path to one of the shared LUNs:If a host in a cluster server configuration loses a physical path to a DS3000, DS4000, or DS5000Storage Subsystem controller, LUNs that are mapped to the cluster group can periodically failoverand then failback between cluster nodes until the failed path is restored. This behavior is the result ofthe automatic logical drive failback feature of the multipath driver. The cluster node with a failedpath to a DS3000, DS4000, or DS5000 controller issues a failover command for all LUNs that weremapped to the cluster group to the controller that it can access. After a programmed interval, thenodes that did not have a failed path will issue a failback command for the LUNs because they canaccess the LUNs on both controllers, resulting in the cluster node with the failed path not being ableto access certain LUNs. This cluster node will then issue a failover command for all LUNs, repeatingthe LUN failover-failback cycle.Note: See the System Storage Interoperation Center at the following Web page for supported clusterservices:www.ibm.com/systems/support/storage/config/ssic

3. If you made any changes to the /kernel/drv/scsi_vhci.conf file in the previous step, save the fileand reboot the server using the following command:

# shutdown –g0 –y –i6

4. If needed, update the Fibre Channel HBA firmware.5. Create the DS3000, DS4000, or DS5000 logical drives and map them to the Fibre Channel HBA ports

in the Sun servers.

Steps for enabling MPxIO on Solaris 10: Before you begin: Keep in mind the following considerations forstmsboot -e [enable] -d [disable] and -u [update]:v When you run the stmsboot command, it is recommended that you accept the default to Reboot the

system now.v The stmsboot command saves copies of the original /kernel/drv/fp.conf and /etc/vfstab files before

modifying them, so you can use the saved files to recover from any unexpected problems.v Ensure that the eeprom boot device is set to boot from the current boot device.

Complete the following steps to enable MPxIO on all Fibre Channel devices:1. Run the stmsboot -e command, and select the default [y] to reboot the system:

# stmsboot -e

WARNING: This operation will require a reboot.Do you want to continue ? [y/n] (default: y) yThe changes will come into effect after rebooting the system.Reboot the system now ? [y/n] (default: y) y

Note: During the reboot, /etc/vfstab and the dump configuration are updated to reflect the devicename changes.



2. After the reboot, configure your applications to use new device names, as explained in “Device namechange considerations for MPxIO” on page 5-16.

3. If necessary, edit the /kernel/drv/fp.conf configuration file to verify that the following parameter isset as follows:mpxio-disable=”no”;

Edit the /kernel/drv/scsi_vhci.conf configuration file to verify that the following parameters are setas follows:load-balance=”none”;auto-failback=”enable”;

4. If you made any changes to configuration files in the previous step, save the file, and reboot theserver using the following command:

# shutdown -g0 -y -i6

5. If needed, update the Fibre Channel HBA firmware.6. Create the DS3000, DS4000, or DS5000 logical drives and map them to the Fibre Channel HBA ports

in the Sun servers.

Verifying devices and configuring failover/failback path for the mapped LUNs: To verify devices and configurethe failover path for the mapped LUNs, complete the following steps:1. Verify devices using the cfgadm –al command to display information about the host ports and their

attached devices:


# cfgadm -alAp_Id Type Receptacle Occupant ConditionPCI0 vgs8514/hp connected configured okPCI1 unknown empty unconfigured unknownPCI2 unknown empty unconfigured unknownPCI3 mult/hp connected configured okPCI4 unknown empty unconfigured unknownPCI5 unknown empty unconfigured unknownPCI6 unknown empty unconfigured unknownPCI7 mult/hp connected configured okPCI8 mult/hp connected configured okc0 scsi-bus connected configured unknownc0::dsk/c0t6d0 CD-ROM connected configured unknownc1 fc-private connected configured unknownc1::500000e0106fca91 disk connected configured unknownc1::500000e0106fcde1 disk connected configured unknownc1::500000e0106fcf31 disk connected configured unknownc1::500000e0106fd061 disk connected configured unknownc1::500000e0106fd7b1 disk connected configured unknownc1::500000e0106fdaa1 disk connected configured unknownc1::50800200001d9841 ESI connected configured unknownc2 fc-fabric connected configured unknownc2::201400a0b811804a disk connected configured unusablec2::201400a0b8118098 disk connected configured unusablec2::201700a0b8111580 disk connected configured unusablec3 fc-fabric connected configured unknownc3::201500a0b8118098 disk connected configured unusablec3::201600a0b8111580 disk connected configured unusablec3::202500a0b811804a disk connected configured unusablec4 fc-fabric connected configured unknownc4::200400a0b80f1285 disk connected configured unknownc4::200400a0b8127a26 disk connected configured unusablec5 fc-fabric connected configured unknownc5::200400a0b82643f5 disk connected unconfigured unknownc5::200500a0b80f1285 disk connected configured unknownc5::200500a0b8127a26 disk connected configured unusablec5::200c00a0b812dc5a disk connected configured unknownusb0/1 usb-kbd connected configured okusb0/2 usb-mouse connected configured okusb0/3 unknown empty unconfigured okusb0/4 unknown empty unconfigured ok#

2. You can also display information about the attachment points on a system. In the following example,c0 represents a fabric-connected host port, and c1 represents a private, loop-connected host port. (Usethe cfgadm command to manage the device configuration on fabric-connected host ports.)By default, the device configuration on private, loop-connected host ports is managed by Solaris host.

Note: The cfgadm -1 command displays information about Fibre Channel host ports. Also use thecfgadm -al command to display information about Fibre Channel devices. The lines that include aport World Wide Name (WWN) in the Ap_Id field associated with c0 represent a fabric device. Usethe cfgadm configure and cfgadm unconfigure commands to manage those devices and make themavailable to Solaris hosts.

# cfgadm -lAp_Id Type Receptacle Occupant Conditionc0 fc-fabric connected unconfigured unknownc1 fc-private connected configured unknown

3. Configure the device using the following command:

cfgadm –c configure Ap-Id


The Ap_ID argument specifies the attachment point ID of the configured Fibre Channel devices. ThisID can be the controller number and WWN of a device (for example, c3::50020f230000591d).See the output example in Step 1 on page 5-18. Also, see the cfgadm man page for an explanation ofattachment points.

Note: An Ap_Id with type fc-private cannot be unconfigured. Only type fc-fabric can beconfigured and unconfigured.

4. Use the luxadm probe command to list all mapped LUNs:

# luxadm probeluxadm probeNo Network Array enclosures found in /dev/es

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006ADE452CBC62d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006ADF452CBC6Ed0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE0452CBC7Ad0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE1452CBC88d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE2452CBC94d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE3452CBCA0d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE4452CBCACd0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE5452CBCB8d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE6452CBCC4d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE7452CBCD2d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE8452CBCDEd0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AE9452CBCEAd0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AEA452CBCF8d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AEB452CBD04d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AEC452CBD10d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006AED452CBD1Ed0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006B2A452CC65Cd0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006B2B452CC666d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006B2C452CC670d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006B2D452CC67Ad0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006B31452CC6A0d0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c0t600A0B800011121800006B32452CC6ACd0s2

Node WWN:200400a0b8111218 Device Type:Disk deviceLogical Path:/dev/rdsk/c8t201400A0B8111218d7s2


5. You can then use the luxadm display logical path command to list more details on each mappedLUN, including the number of paths to each LUN. The following example uses a logical path fromthe previous example:

# luxadm display /dev/rdsk/c0t600A0B800011121800006B31452CC6A0d0s2DEVICE PROPERTIES for disk: /dev/rdsk/c0t600A0B800011121800006B31452CC6A0d0s2

Vendor: IBMProduct ID: 1742-900Revision: 0914Serial Num: 1T51207691Unformatted capacity: 1024.000 MBytesWrite Cache: EnabledRead Cache: Enabled

Minimum prefetch: 0x0Maximum prefetch: 0x0

Device Type: Disk devicePath(s):

/dev/rdsk/c0t600A0B800011121800006B31452CC6A0d0s2/devices/scsi_vhci/ssd@g600a0b800011121800006b31452cc6a0:c,rawController /devices/pci@7c0/pci@0/pci@8/SUNW,qlc@0,1/fp@0,0Device Address 201400a0b8111218,1eHost controller port WWN 210100e08ba0fca0Class secondaryState STANDBYController /devices/pci@7c0/pci@0/pci@8/SUNW,qlc@0,1/fp@0,0Device Address 201500a0b8111218,1eHost controller port WWN 210100e08ba0fca0Class primaryState ONLINEController /devices/pci@7c0/pci@0/pci@8/SUNW,qlc@0/fp@0,0Device Address 201400a0b8111218,1eHost controller port WWN 210000e08b80fca0Class secondaryState STANDBYController /devices/pci@7c0/pci@0/pci@8/SUNW,qlc@0/fp@0,0Device Address 201500a0b8111218,1eHost controller port WWN 210000e08b80fca0Class primaryState ONLINE

#

Unconfiguring a failover/failback path: Before you unconfigure a fabric device, stop all activity to the deviceand unmount any file systems on the fabric device. (See your Solaris administration documentation forunmounting procedures.)

To unconfigure a failover/failback path, complete the following steps:1. Run the cfgadm -al command to display information about the host ports and their attached devices.2. Unconfigure the LUN by running the following command:

cfgadm –c unconfigure Ap-Id

Where Ap-Id is the LUN that you want to unconfigure.3. Run the cfgadm -al command again, to verify that the LUN is now unconfigured.4. If necessary, define the file structure using the newfs command. Add entries to the /etc/vfstab file.5. Reboot the server using the following command:

shutdown -g0 -y -i6


Disabling the MPxIO multipath driver: For Solaris 10, unconfigure all devices using the cfgadm –cunconfigure AP-id Ap-id command. Then, run the stmsboot –d command, and accept the default toReboot the system now.

For Solaris 8 and 9, unconfigure all devices using the cfgadm –c unconfigure AP-id Ap-id command, andedit the /kernel/drv/scsi_vhci.conf configuration file to set the value of the mpxio-disable parameter toyes. Reboot the system.

To learn how to revert the patches or the StorEdge software, please see the Sun StorEdge SAN FoundationInstallation Software Guide at the following Web site:

http://docs.sun.com.

Installing the RDAC driver on SolarisThis section describes how to install RDAC on a Solaris host.

Before you begin:

1. Because you cannot run both RDAC and MPxIO, make sure that MPxIO is disabled. Check theconfiguration files (/kernel/drv/scsi_vhci.conf and/or /kernel/drv/fp.conf) and make sure that thevalue of the mpxio-disable parameter to set to Yes.

2. You must install an HBA driver package before you install RDAC. If you have a SAN-attachedconfiguration, you must also modify the HBA's configuration file before you install RDAC. If you failto follow the procedures in this order, problems can occur.

Note:

1. RDAC is only supported on Solaris 8 and 9. (RDAC is not supported on Solaris 10.)2. Modifying failover settings in the HBA's configuration file after installing RDAC requires the removal

of the RDAC from the host.

Steps for installing the RDAC failover driver:

Important: In some configurations, a patch is required for RDAC to function properly. Before you beginthe RDAC installation, check the DS Storage Manager README file for Solaris to find out whether thepatch is required for your specific configuration. In addition, you can find the latest RDAC versions andother important information in the README file. (For steps to find the README file on the Web, see“Finding Storage Manager software, controller firmware, and README files” on page xiii.)

Complete the following steps to install RDAC:1. Insert the Solaris installation CD in the CD-ROM drive.

Note: In this procedure, the installation CD is mounted at /cdrom/SM91. Modify these commands asneeded for your installation.

2. Type the following command to start installing the RDAC package:

# pkgadd -d path/filename.pkg

where path/filename is the directory path and name of the package that you want to install.The installation process begins.Information about packages that can be installed in the specified directory is displayed on thecommand line, as in the following example:


http://docs.sun.com

The following packages are available:

1 RDAC Redundant Disk Array Controller(sparc) version number

Select package(s) you wish to process (or ’all’ to process allpackages). (default:all) [?,??,q]:

3. Type the value of the package you are installing and press Enter.The installation process begins.

4. The software automatically checks for package conflicts. If any conflicts are detected, a message isdisplayed indicating that some files are already installed and are in use by another package.The following prompt is displayed:

Do you want to install these conflicting files [y, n, ?]

Type y and press Enter.The following prompt is displayed:

This package contains scripts which will be executed with super-userpermission during the process of installing this package.

Do you want to continue with the installation of <RDAC>

[y, n, ?]

5. Type y and press Enter.The installation process continues.

6. When the RDAC package has been successfully installed, the following message is displayed:

Installation of <RDAC> was successful.

Ensure that the variables in the configuration files for the JNI adapter cards have been set to thecorrect values.

7. Reboot the Solaris host by typing the following command:


Attention: Any modifications to the persistent bindings in the jnic146x.conf configuration file requiresthe removal of RDAC. After the RDAC is removed you can modify the persistent bindings in thejnic146x.conf file, and then reinstall RDAC.

Complete the following steps to modify the sd.conf or jnic146x.conf files:1. Remove RDAC by typing the following command:

# pkgrm RDAC_driver_pkg_name

where RDAC_driver_pkg_name is the name of the RDAC driver package that you want to remove.2. Verify RDAC drive package removal by typing the following command:

# pkginfo RDAC_driver_pkg_name

where RDAC_driver_pkg_name is the name of the RDAC driver package that you removed.3. Reboot the Solaris host by typing the following command:



4. Modify persistent bindings in the sd.conf file or edit the sd.conf file by typing the followingcommand:

# vi /kernel/drv/jnic146x.conf or sd.conf

When you have finished making changes, run the following command to save the changes:

# :wq

5. Install the RDAC driver package by typing the following command:

# pkgadd -d RDAC_driver_pkg_name

where RDAC_driver_pkg_name is the name of the RDAC driver package that you want to install.6. Verify package installation by typing the following command:

# pkginfo RDAC_driver_pkg_name

where RDAC_driver_pkg_name is the name of the RDAC driver package that you installed.7. Reboot the Solaris host by typing the following command:


Note: You must reboot the host after modifying the jnic146x.conf file, because the jnic146x.conf driveris only read during the boot process. Failure to reboot the host might result in some devices beinginaccessible.

Installing the DMP driverThis section describes how to install Veritas Dynamic Multipathing (DMP), which is a failover driver forSolaris hosts. The DMP failover driver is a feature of Veritas Volume Manager, which is a component ofthe Storage Foundation product from Symantec. While RDAC allows you to have only 32 LUNs, DMPallows you to have up to 256 LUNs.

System requirements:

Ensure that your system meets the following requirements for installing DMP:v Solaris operating systemv Veritas VolumeManager 4.0, 4.1, 5.0, or 5.1v Array Support Library (ASL). (enables Solaris to recognize the DS3000, DS4000, or DS5000 machine

type)

Note: The ASL may be a separate file available from Symantec or it may be integrated with VolumeManager, depending on version of Storage Foundation.

DMP installation overview: Ensure that your system meets the following prerequisites for installingDMP:v The HBAs are installed on the Solaris host.v The parameter settings in the HBA configuration file (for example, qla2300.conf) are modified.v In a SAN environment, bindings are configured.v The zones are created and enabled for the Solaris partition.


v Storage is mapped to the Solaris partition.

Perform the following procedures, in the order listed, to complete the DMP installation:1. “Preparing for Veritas DMP installation”2. “Installing Veritas Storage Foundation Solaris with Veritas Volume Manager and DMP” on page 5-263. “Installing the ASL package” on page 5-26

Preparing for Veritas DMP installation:

Complete the following steps to prepare the host for installing Veritas DMP:1. Choose the Solaris host on which you want to install DMP.2. Manually define the targets and LUNs in the /kernel/drv/sd.conf file by completing the following

steps.By default, the /kernel/drv/sd.conf file defines targets 0, 1, 2, and 3. LUN0 also is defined for targets0, 1, 2, and 3.Notes:

v Each target represents a controller to a subsystem, and each LUN represents a logical drive.v If you are adding additional target or LUN definitions to the /kernel/drv/sd.conf file for an

existing DMP configuration, be sure to reboot the Solaris host.a. Open the /kernel/drv/sd.conf file with the vi Editor, by typing the following command:

# vi /kernel/drv/sd.conf

The file looks similar to the following example:

## Copyright (c) 1992, Sun Microsystems, Inc.## ident "@(#)sd.conf 1.9 98/01/11 SMI"

name="sd" class="scsi" class_prop="atapi"target=0 lun=0;




b. Add additional target and LUN definitions, using the vi Editor. In the following example, it isassumed that the Solaris host is attached to one DS3000, DS4000, or DS5000 subsystem with threeLUNs mapped to the DS3000, DS4000, or DS5000 storage partition. In addition, the access LUNmust be mapped to the partition.


## Copyright (c) 1992, Sun Microsystems, Inc.## ident "@(#)sd.conf 1.9 98/01/11 SMI"





name="sd" class="scsi" target=0 lun=1;name="sd" class="scsi" target=0 lun=2;name="sd" class="scsi" target=0 lun=3;name="sd" class="scsi" target=0 lun=31;name="sd" class="scsi" target=1 lun=1;name="sd" class="scsi" target=1 lun=2;name="sd" class="scsi" target=1 lun=3;name="sd" class="scsi" target=1 lun=31;

c. Save the new entries in the /kernel/drv/sd.conf file, by typing the following command:

# :wq

3. Verify that RDAC is not installed on the host, by typing the following command:

# pkginfo -l RDAC

4. If RDAC is installed, remove it by typing the following command:

# pkgrm RDAC

5. Verify that a host partition has been created.Attention: Set the host type to Solaris with DMP. Failure to do so results in an inability to map formore than the RDAC limit of 32 LUNs and causes other undesired results.

6. Ensure that all of the paths are optimal and are in a preferred path state from the SMclient.7. Install Storage Foundation Solaris including Veritas Volume Manager with DMP. For documentation

see http://www.symantec.com/business/support/.8. Reboot the Solaris host, by typing the following command:


Installing Veritas Storage Foundation Solaris with Veritas Volume Manager and DMP: Before youbegin to install Veritas Storage Foundation Solaris with Veritas Volume Manager and DMP, ensure thatyou have the required license keys. This document does not describe how to install the Veritas product.For documentation see http://www.symantec.com/business/support/.

Installing the ASL package: Complete the following steps to install the ASL package if required:

Note: The VxVM 4.x version of the ASL package is named "SMibmasl", see http://seer.entsupport.symantec.com/docs/284913.htm. For VxVM version 5.0 and later, many ASLs are




http://seer.entsupport.symantec.com/docs/284913.htm


integrated into VxVM and do not need to be installed. For VxVM version 5.0 and later, the ASL packageis named "VRTSLSIasl", see http://seer.entsupport.symantec.com/docs/340469.htm. This followingexample assumes VxVM 4.x is being installed.1. Install the SMibmasl package, by typing the following command:

Tip: You can select either the default (all), or select option 1.

# pkgadd -d SMibmasl_pkg

2. Reboot the Solaris host, by typing the following command:


In addition, see Sysmantec's the Veritas documentation for information about how to complete thefollowing tasks:v Start Veritas Volume Managerv Set up disk groupsv Create volumesv Create file systemsv Mount file systems

Identifying devicesAfter you have installed the multipath driver, or verified that the multipath driver is already installed,you then need to use the SMdevices utility to identify a storage subsystem logical drive associated withan operating system device.

Using the SMdevices utility

Using SMdevices on Windows operating systemsThe SMutil software includes a utility called SMdevices that you can use to view the storage subsystemlogical drive that is associated with a particular operating system device name. This utility is helpfulwhen you want to create drive letters or partitions by using Disk Administrator.

When you finish creating the logical drives on a particular storage subsystem, go to the host that isattached to that storage subsystem, and perform the following steps to use SMdevices on Windows:1. From a DOS or command prompt, change to the directory <installation_directory>\Util,

where installation_directory is the directory in which you installed the SMutil.The default directory is c:\Program Files\IBM_DS4000\Util.

2. Type:

SMdevices

3. Press Enter.

Using SMdevices on UNIX-based operating systemsYou can use SMdevices to map the host-assigned device name for each LUN back to its correspondingDS3000, DS4000, or DS5000 Storage Subsystem device.

In the SMdevices output, you can view the following DS3000, DS4000, or DS5000 Storage Subsysteminformation, as it is shown on SMclient.

Note: The examples in the list refer to the sample SMdevices output.



v Host assigned name (/dev/sdh)v DS3000, DS4000, or DS5000 Storage Subsystem name (DS4500_Storage_Server-A)v Logical drive name Raid-5-0A

v LUN ID (LUN 4)v Preferred controller owner, and whether that controller is currently controlling the logical drive

The following example shows a sample SMdevices output for the DS4500_Storage_Server-A subsystem:

# SMdevicesIBM FAStT Storage Manager Devices, Version 09.12.A5.00Built Fri Jan 14 16:42:15 CST 2005(C) Copyright International Business Machines Corporation,2004 Licensed Material - Program Property of IBM. All rights reserved.

/dev/sdh (/dev/sg10) [Storage Subsystem DS4500_Storage_Server-A,Logical Drive Raid-5-0A, LUN 4, Logical Drive ID<600a0b80000f0fc300000044412e2dbf>, Preferred Path (Controller-A): In Use]

/dev/sdd (/dev/sg6) [Storage Subsystem DS4500_Storage_Server-A,Logical Drive Raid-5-1A, LUN 0, Logical Drive ID<600a0b80000f13ec00000016412e2e86>, Preferred Path (Controller-B): In Use]

/dev/sde (/dev/sg7) [Storage Subsystem DS4500_Storage_Server-A,Logical Drive Raid-0-0A, LUN 1, Logical Drive ID<600a0b80000f0fc30000003c412e2d59>, Preferred Path (Controller-A): In Use]

/dev/sdf (/dev/sg8) [Storage Subsystem DS4500_Storage_Server-A,Logical Drive Raid-1-0A, LUN 2, Logical Drive ID<600a0b80000f0fc30000003e412e2d79>, Preferred Path (Controller-A): In Use]

/dev/sdg (/dev/sg9) [Storage Subsystem DS4500_Storage_Server-A,Logical Drive Raid-3-0A, LUN 3, Logical Drive ID<600a0b80000f13ec00000012412e2e4c>, Preferred Path (Controller-A): In Use]

Identifying devices on AIX hostsThe multipath driver creates the following devices that represent the DS3000, DS4000, or DS5000 StorageSubsystem configuration:

dar The disk array router (dar) device represents the entire array, including the current and thedeferred paths to all LUNs (hdisks).

dac The disk array controller (dac) devices represent a controller within the storage subsystem. Thereare two dacs in the storage subsystem. With MPIO, the dac device will only show up if there is aUTM device assigned.

hdisk Each hdisk device represents an individual LUN on the array.

utm The universal transport mechanism (utm) device is used only with in-band managementconfigurations, as a communication channel between the SMagent and the DS3000, DS4000, orDS5000.

Note: You might see the utm device listed in command output, whether or not you have anin-band management configuration. For example, a utm might be listed when you run the lsattrcommand on a dac.

Performing initial device discoveryComplete these steps to perform the initial device discovery:

Before you begin: Ensure that the DS3000, DS4000, or DS5000 Storage Subsystem has been set up, LUNshave been assigned to the host, and the multipath driver has been installed.1. Type the following command to probe for the new devices:


# cfgmgr -v

Note: In a SAN configuration, the devices do not log into the SAN switch until you run cfgmgr.2. Type the following command:

# lsdev -Cc disk

3. Examine the output of the lsdev -Cc disk command to ensure that the RDAC software recognizes theDS3000, DS4000, or DS5000 logical drives, as shown in the following list:v Each DS4200 logical drive is recognized as an “1814 DS4200 Disk Array Device.”v Each DS4300 logical drive is recognized as an “1722-600 (600) Disk Array Device.”v Each DS4400 logical drive is recognized as an “1742-700 (700) Disk Array Device.”v Each DS4500 logical drive is recognized as an “1742-900 (900) Disk Array Device.”v Each DS4700 logical drive is recognized as an “1814 DS4700 Disk Array Device.”v Each DS4800 logical drive is recognized as an “1815 DS4800 Disk Array Device.”

Important: You might discover that the configuration process has created two dacs and two dars onone DS3000, DS4000, or DS5000 subsystem. This situation can occur when your host is using apartition that does not have any associated LUNs. When that happens, the system cannot associatethe two dacs under the correct dar. If there are no LUNs, the system generates two dacs as expected,but it also generates two dars.The following list shows the most common causes:v You create a partition and attach the LUNs to it, but you do not add the host ports to the partition.

Therefore, the host ports remain in the default partition.v You replace one or more HBAs, but do not update the worldwide name (WWN) of the partition for

the HBA.v You switch the DS3000, DS4000, or DS5000 from one set of HBAs to another as part of a

reconfiguration, and do not update the WWNs.

In each of these cases, resolve the problem, and run cfgmgr again. The system removes the extra dar, ormoves it from the Available state to the Defined state. (If the system moves the dar into the Definedstate, you can then delete it.)

Note: When you perform the initial device identification, the Object Data Manager (ODM) attributes ofeach device are updated with default values. In most cases and for most configurations, the defaultvalues are satisfactory. However, there are some values that can be modified for maximum performanceand availability. See Appendix E, “Viewing and setting AIX Object Data Manager (ODM) attributes,” onpage E-1 for information about using the lsattr command to view attribute settings on an AIX system.


Initial discovery with MPIO

# lsdev -C |grep hdisk10hdisk10 Available 05-08-02 MPIO Other DS4K Array Disk

# lscfg -vpl hdisk10hdisk10 U787F.001.DPM0H2M-P1-C3-T1-W200400A0B8112AE4-L9000000000000MPIO Other DS4K Array Disk

Manufacturer................IBMMachine Type and Model......1814 FAStTROS Level and ID............30393136Serial Number...............Device Specific.(Z0)........0000053245004032Device Specific.(Z1)........

# mpio_get_config -AStorage Subsystem worldwide name: 60ab8001122ae000045f7fe33Storage Subsystem Name = ’Kinks-DS-4700’

hdisk LUN #hdisk2 1hdisk3 2hdisk4 3hdisk5 4hdisk6 5hdisk7 6hdisk8 7hdisk9 8hdisk10 9hdisk11 10

Configuring devicesTo maximize your storage subsystem's performance, you can set the queue depth for your hdisks, disablecache mirroring, use dynamic capacity and dynamic volume expansion, and check the size of your LUNs.

Using the hot_add utilityThe hot_add utility enables you to add new logical drives without restarting the system. The utilityregisters the new logical drives with the operating system so that you can use Disk Administrator tocreate partitions, add device names, and so on. The hot_add utility is part of the SMutil softwarepackage. If you run the program twice and the new logical drives are not displayed in the DiskAdministrator window, you must either run Fibre Channel diagnostics or restart (reboot) the host.

When you finish creating logical drives on a particular storage subsystem, go to the host that is attachedto that storage subsystem and perform the following steps to use the hot_add utility:1. From a DOS or command prompt, change to the directory:

<installation_directory>\Util


2. From a DOS or command prompt, type the following command:hot_add

3. Press Enter. The new logical drives are available through the Disk Administrator.

Using the SMrepassist utilityYou can use the SMrepassist utility to flush cached data for a logical drive.


Important: The FlashCopy drive cannot be added or mapped to the same server that has the base logicaldrive of the FlashCopy logical drive in a Windows 2000, Windows Server 2003, Windows Server 2008, orNetWare environment. You must map the FlashCopy logical drive to another server.

To flush cached data in a logical drive, perform the following steps:1. From a DOS or command prompt, change to the directory:

<installation_directory>\Util


2. Type the following command:smrepassist -f logical_drive_letter:

where logical_drive_letter is the operating system drive letter that was assigned to the diskpartition created on the logical drive.

3. Press Enter.

Stopping and restarting the host-agent softwareYou must stop and restart the host-agent software if you add additional storage subsystems to themanagement domain of the host-agent software. When you restart the service, the host-agent softwarediscovers the new storage subsystems and adds them to the management domain.

Note: If none of the access logical drives are detected after a restart, the host-agent softwareautomatically stops running. Make sure that there is a good Fibre Channel connection from the host tothe SAN to which the storage subsystem is connected. Then restart the host or cluster node so that newhost-agent-managed storage subsystems can be discovered.

Windows 2000To stop and restart the host-agent software, perform the following steps:1. Click Start → Programs → Administrative Tools → Services. The Services window opens.2. Right-click IBM DS Storage Manager Agent.3. Click Restart. The IBM DS Storage Manager Agent stops and then starts again.4. Close the Services window.

Windows Server 2003 and 2008To stop and restart the host-agent software, perform the following steps:1. Click Start → Administrative Tools → Services. The Services window opens.2. Right-click IBM DS Storage Manager Agent.3. Click Restart. The IBM DS Storage Manager Agent stops and then starts again.4. Close the Services window.

Setting the queue depth for hdisk devicesSetting the queue_depth attribute to the appropriate value is important for system performance. If youhave a large DS3000, DS4000, or DS5000 configuration with many logical drives and hosts attached, usethis setting for high performance.

This section provides methods for calculating your system's maximum queue depth, which you can useas a guideline to help you determine the best queue depth setting for your configuration.

Calculating maximum queue depthThe formula for calculating the maximum queue depth for your system depends on which firmwareversion is installed on the controller. Use one of the following formulas to calculate the maximum queuedepth for your system.


Important:

1. The maximum queue depth might not be an optimal setting in all cases. Use the maximum queuedepth as a guideline, and adjust the setting as necessary for your specific configuration.

2. In systems with one or more SATA devices attached, you might need to set the queue depth attributeto a lower value than the maximum queue depth.

v Formulas for controller firmware version 07.10.xx.xx and above

On DS4800 and DS4700/DS4200 storage systems that are running DS3000, DS4000, DS5000controller firmware version 07.10.xx.xx, use the following formulas to determine the maximumqueue depth:

DS4800: 4096 / (number-of-hosts * LUNs-per-host )

For example, a DS4800 system with four hosts, each with 32 LUNs, would have a maximumqueue depth of 32:

4096 / ( 4 * 32 ) = 32

DS4700/DS4200: 2048 / (number-of-hosts * LUNs-per-host )

For example, a DS4700 system or a DS4200 system with four hosts, either with 32 LUNs, wouldhave a maximum queue depth of 16:

2048 / ( 4 * 32 ) = 16

v Formula for controller firmware versions 05.4x.xx.xx, or 06.1x.xx.xx to 06.6x.xx.xx

On DS4000 or DS5000 storage systems that are running DS4000 or DS5000 controller firmwareversions 05.4x.xx.xx, or 06.1x.xx.xx to 06.6x.xx.xx, use the following formula to determine themaximum queue depth:

2048 / (number-of-hosts * LUNs-per-host )

For example, a system with four hosts, each with 32 LUNs, would have a maximum queue depthof 16:

2048 / ( 4 * 32 ) = 16

v Formula for controller firmware version 05.30.xx.xx

On DS4000 or DS5000 storage systems that are running DS4000 or DS5000 controller firmwareversion 05.30.xx.xx or earlier, use the following formula to determine the maximum queue depth:

512 / (number-of-hosts * LUNs-per-host )

For example, a system with four hosts, each with 32 LUNs, would have a maximum queue depthof 4:

512 / ( 4 * 32 ) = 4

Changing the queue depth for WindowsYou can use the QLogic SANsurfer program to modify the Host Adapter Settings and Advanced AdapterSettings preferences from the Windows operating system environment; however, you must reboot theservers for the changes to become effective.

Alternatively, to change the queue depth setting for a QLogic adapter in a Microsoft Windows operatingsystem environment, you must select the Configuration Settings menu in Fast!UTIL and then selectAdvanced Adapter Settings to access the Execution Throttle.

Changing the queue depth for AIXYou can change the queue_depth attribute for AIX by using the chdev -l command, as shown in thefollowing example:


# chdev -l hdiskX -a queue_depth=y -P

where X is the name of the hdisk and y is the queue depth setting.

Note: Use the -P flag to make the changes permanent in the Customized Devices object class.

Steps for disabling cache mirroringBefore you begin: If write cache is enabled, make backups of all data before disabling cache mirroring.

In DS Storage Manager, complete the following steps to disable cache mirroring:1. In the Logical/Physical view of the Subsystem Management window, right-click the logical drive on

which you want to disable cache mirroring, and select Change → Cache Settings.2. In the Change Cache Settings window, clear the Enable write caching with mirroring check box.3. Click OK.

Note: For AIX operating systems, when a LUN is opened that is running with write cache enabled andcache mirroring disabled, an FCP array warning message displays. The warning displays again every 24hours until cache mirroring is enabled again.

Using dynamic capacity expansion and dynamic volume expansionDynamic volume expansion (DVE) increases the size of a logical drive. In order to perform a DVE, theremust be free capacity available on the array. If there is not, you can first perform a dynamic capacityexpansion (DCE) to increases the array's capacity by adding drives.

After you have ensured that there is sufficient free capacity within the array, you can perform a DVEoperation.

Performing a dynamic capacity expansion operation

Before you begin: You can find more information about this procedure in the Storage Manager onlinehelp.

Complete the following steps to increase the capacity on the array by performing a DCE:1. In the Logical/Physical view of the Subsystem Management Window, right-click on an array and

select Add Free Capacity (Drives).2. In the Add Free Capacity (Drives) window, select one or two available drives and click Add.

Performing a dynamic volume expansion operation

Before you begin: Ensure that there is available free capacity within the array. You can check freecapacity availability using DS Storage Manager, in the Logical/Physical view of the SubsystemManagement window. If there is not enough free capacity, and extra drives are available, you can addone or more to the array by performing a dynamic capacity expansion (DCE) operation before youperform the DVE operation.

You can find more information about this procedure in the Storage Manager online help.

Note:

1. You cannot resize the logical drive while the logical drive group is activated in classic or enhancedconcurrent mode.

2. You cannot resize the root logical drive group.


Complete the following steps to increase the size of a logical drive by performing a DVE:1. From the Logical/Physical window of the Subsystem Management window, right-click the logical

drive and select Increase Capacity. The Increase Logical Drive Capacity—Additional Instructionswindow opens.

2. Read the additional instructions and click OK. The Increase Logical Drive Capacity window opens.3. Type the amount by which you want to increase the logical drive, and click OK.

You see a clock icon on every logical drive within the array. You must wait for the process tocomplete before you can begin any host intervention.

Tip: If the storage subsystem is busy, the process might take several hours to complete.4. On the host, rescan the logical drive by typing the following commands:

# cd /sys/block/sdXX/device# echo 1 > rescan

where XX is the device name.5. Check the size of the logical drive using the steps that are described in “Checking LUN size” on page

5-35.6. Remount the logical drive.

Veritas Storage Foundation with SUSE Linux Enterprise ServerScanning for LVM volumes can increase boot time and is not required in the Veritas Storage Foundationenvironment. The LVM scan with SLES 10 sp2 or later should be disabled. Use the following procedure todiasble the LVM scan.

Note:

v In the Veritas Storage Foundation Linux environment, the default host type must be set to 13(LNXCLVMWARE).

v IBM supports the DMP A/P-F ASL/APM only, not the A/P-C ASL.v During boot, before DMP is loaded, I/O probes going to the non-owning controller will generate

timeout errors. These boot-time errors are unavoidable and not significant.

Make the following changes in the /etc/lvm/lvm.conf file:1. Change the line filter = [ "a/.*/" ] to filter = [ "r|/dev/.*/by-path/.*|", "r|/dev/.*/by-id/.*|",

"r|/dev/sd.*|", "a/.*/" ]

2. If the root/swap is in an LVM volume:v Make sure appropriate volumes are scanned by adding your specific device to the filter in step 1.v After the change to /etc/lvm/lvm.conf, run mkinitrd and use the new initrd image for future

boots.

Veritas Storage Foundation 5.0 with Red Hat Enterprise LinuxThe following procedure is required for enabling the rdac module on RHEL 5.3 for Storage Foundation5.0 only. The module has already been integrated into Storage Foundation 5.1 and later. The scsi_dh_rdacmodule provides the support for rdac devices. It eliminates the time delay and some of the errormessages during boot/probe.

Note:

v In the Veritas Storage Foundation Linux environment, the default host type must be set to 13(LNXCLVMWARE).

v IBM supports the DMP A/P-F ASL/APM only, not the A/P-C ASL.


v During boot, before DMP is loaded, I/O probes going to the non-owning controller will generatetimeout errors. These boot-time errors are unavoidable and not significant.

This procedure currently works with the IBM NVSRAM due to the scsi_dh_rdac module being VID/PIDdependent.1. Disable all the DS3000, DS4000, or DS5000 storage ports so that the HBA cannot see them.2. Install Storage Foundation as you normally would, and complete the installation.3. Run mkinitrd to include the scsi_dh_rdac module: mkinitrd $resultant_initrd_image_file

$kernel_version --preload=scsi_dh_rdac. For example, mkinitrd /boot/my_image 2.6.18-118.el5--preload=scsi_dh_rdac,

Note: uname -r gives the kernel version.4. Change your boot loader to use the new initrd image. In Power, the initrd image name is yaboot. In

i386, the image name is grub.5. Shut down the host server.6. Enable the ports of the storage so that the HBA can see the storage7. Boot the host server.

Checking LUN sizeComplete the following steps to check the size of a LUN in AIX.1. Type the following commands:

#cd /sys/block/sdXX# cat size

where XX is the device name.Result: A number displays, as in the following example:

8388608

2. Multiply this number by 512 (bytes) to calculate the size of the LUN, as shown in the followingexample:8388608 * 512 = 4294967296 (~ 4GB)

Result: The result of the calculation is the size of the LUN. In the example, the LUN size isapproximately 4 GB.

Redistributing logical drivesIn a failover condition where logical drives have failed over to their secondary controller path, someconfigurations require a manual intervention to move these drives back after the error condition has beenresolved. This will depend on the host multipath driver that is installed and whether ADT (Auto DriveTransfer) is enabled or not. By default AIX and Windows have ADT disabled, but their multipath driverscan automatically recover. Linux by default has ADT enabled, but the MPP driver can do the sameautorecover, and ADT should be disabled when using that driver.

To manually redistribute logical drives, use the Subsystem Management Window, redistribute logicaldrives to their preferred paths by clicking Advanced → Recovery → Redistribute Logical Drives.

Redistributing logical drives on AIXIf you enabled autorecovery on the AIX host, you do not need to redistribute logical drives manuallyafter a controller failover. However, if you have a heterogeneous host environment, you might need toredistribute logical drives manually. Hosts that do not support some form of autorecovery, or AIX hoststhat have autorecovery disabled, will not automatically redirect logical drives to the preferred paths.


Complete the following steps to manually redistribute logical drives to their paths:1. Repair or replace any faulty components. For more information, see the Installation, User's and

Maintenance Guide for the appropriate DS3000, DS4000, or DS5000 Storage Subsystem.2. Using the Subsystem Management Window, redistribute logical drives to their preferred paths by

clicking Advanced → Recovery → Redistribute Logical Drives.

Note: If a large number of LUNs is configured on the DS3000, DS4000, or DS5000 Storage Subsystem,redistributing logical drives might take 60 minutes or more to complete, depending on how busy thesystem is.

3. Run the fget_config command to verify the active paths, as shown in this example:

# fget_config -l dar0dac0 ACTIVE dac1 ACTIVEdac0-hdisk1dac0-hdisk2dac0-hdisk3dac1-hdisk4dac1-hdisk5dac1-hdisk6dac1-hdisk7dac0-hdisk8

Redistributing logical drives on HP-UXAuto Drive Transfer (ADT) is enabled, by default, on HP-UX hosts. If a failure occurs that initiates acontroller failover, ADT redirects I/O to the available controller. ADT does not require manualredistribution.

Important: If a failure occurs in a heterogeneous host environment, the HP-UX host with ADT enabledwill automatically redistribute its LUNs when the path becomes available. However, you will need tomanually redistribute logical drives on any host that does not have ADT enabled. Failure to do so willleave the subsystem in a Needs Attention state, because hosts that do not support ADT or have ADTdisabled will not automatically redirect I/O to the preferred controller. In this case, Storage ManagerRecovery Guru will indicate which host platform is associated with the LUN that is in a failover state.

Note: DS5000 storage subsystems are not ALUA-compliant. DS5000 subsystems have Target Port GroupSupport (TPGS), which is a similar SCSI protocol that directs I/O to preferred ports. For HP-UX 11.31, thedefault HP-UX host type must be changed to the TPGS host type HPXTPGS.

To turn on TPGS support and change the host type, complete the following steps:1. Change the O/S type for the DS5000 subsystem from HPUX to HPXTPGS.2. Change the load balancing to Default, round-robin.3. Verify that the changes are correct. The following example shows one of the LUNs that has the correct

four active paths and four standby paths:# scsimgr get_info all_lpt -D /dev/rdisk/asm1ai|grep -e STATUS -e ’Open close state’

STATUS INFORMATION FOR LUN PATH : lunpath306Open close state = ACTIVE

STATUS INFORMATION FOR LUN PATH : lunpath344Open close state = STANDBY








4. Use the SAN Fibre Channel switch monitoring tools to verify that the I/O loads are distributedproperly.

Redistributing logical drives on SolarisIf you select Solaris as the host type when you define the host and host port, Auto Drive Transfer (ADT)is disabled on Solaris hosts. In this case, if a failure occurs that initiates a controller failover, you mustmanually redistribute logical drives to their preferred paths.

Complete the following steps to manually redistribute logical drives to their preferred paths:1. Repair or replace any faulty components. For more information, see the Installation, User's and

Maintenance Guide for the appropriate DS3000, DS4000, or DS5000 Storage Subsystem.2. Using the Subsystem Management window, redistribute logical drives to their preferred paths by

clicking Advanced → Recovery → Redistribute Logical Drives.

Resolving disk array errors on AIXThis section shows a list of possible disk array errors that could be reported in the AIX error log. You canview the AIX error log by running the errpt -a command.

You can also check your DS Storage Manager Major Event log (MEL) to find out whether there is anycorrelation between the host, SAN, and the DS3000, DS4000, or DS5000 Storage Subsystem.

You might need to validate your configuration or replace defective hardware to correct the situation.

Note: For more information about troubleshooting, see the Installation, User's and Maintenance Guide foryour DS3000, DS4000, or DS5000 Storage Subsystem.v FCP_ARRAY_ERR1 ARRAY OPERATION ERROR

A permanent hardware error involving the disk array media.v FCP_ARRAY_ERR2 ARRAY OPERATION ERROR

A permanent hardware error.v FCP_ARRAY_ERR3 ARRAY OPERATION ERROR

A permanent error detected by the array adapter.v FCP_ARRAY_ERR4 ARRAY OPERATION ERROR

A temporary error within the array, communications, adapter, and so on.v FCP_ARRAY_ERR5 UNDETERMINED ERROR

An undetermined error has occurred.v FCP_ARRAY_ERR6 SUBSYSTEM COMPONENT FAILURE

A degradation condition has occurred other than a disk drive.v FCP_ARRAY_ERR7 CONTROLLER HEALTH CHECK FAILURE

A health check on the passive controller has failed.v FCP_ARRAY_ERR8 ARRAY CONTROLLER SWITCH

One array controller has become unavailable, so I/O has moved to the other controller.v FCP_ARRAY_ERR9 ARRAY CONTROLLER SWITCH FAILURE

An array controller switch has failed.v FCP_ARRAY_ERR10 ARRAY CONFIGURATION CHANGED

A logical unit has been moved from one controller to the other (most likely by the action of analternate host).


v FCP_ARRAY_ERR11 IMPROPER DRIVE TYPE FOR DUAL ACTIVE MODE

This error should not be possible on the 2102 array, and exists for history reasons only.FCP_ARRAY_ERR11 might be reused for a different error in the future.

v FCP_ARRAY_ERR12 POLLED AEN FAILURE

An automatic error notification has failed.v FCP_ARRAY_ERR13 ARRAY INTER-CONTROLLER COMMUNICATION FAILURE

The controllers are unable to communicate with each other. This could result from one of thecontrollers being rebooted while the error log was being generated. However, it could be a much moreserious error that indicates a problem with the Fibre Channel connections.

v FCP_ARRAY_ERR14 ARRAY DRIVE FAILURE

A serious or unrecoverable error has been detected on a physical disk within the DS3000, DS4000, orDS5000 subsystem. A system engineer might be able to obtain the exact cause from an analysis of thesense data.

v FCP_ARRAY_ERR15 CACHE BATTERY LOW/DATA LOSS POSSIBLE

If a controller card is replaced, it is likely that the cache batteries will be flat. It can take two days forthe cache batteries to be fully recharged. During this time errors are logged in the error log. Do notreplace the controller.

v FCP_ARRAY_ERR16 CACHE BATTERY CHARGE BELOW 87.5%

If a controller card is replaced, it is likely that the cache batteries will be flat. It can take two days forthe cache batteries to be fully recharged. During this time errors are logged in the error log. Do notreplace the controller.

v FCP_ARRAY_ERR17 WORLDWIDE NAME CHANGED

A controller has changed worldwide names (most likely either it was replaced without placing it in thereset state first, or the cabling was changed so that a different controller with the same SCSI ID is onthe loop).

v FCP_ARRAY_ERR18 RESERVATION CONFLICT

An operation failed because the disk array logical drive (LUN) is reserved by another host.v FCP_ARRAY_ERR19 SNAPSHOT VOLUME'S REPOSITORY FULL

The repository capacity limit has been reached. To resolve this error you can increase the repositorycapacity.

v FCP_ARRAY_ERR20 SNAPSHOT OPERATION STOPPED BY ADMIN

The FlashCopy (snapshot) operation has been disabled or stopped. To resolve this error you canre-create the FlashCopy.

v FCP_ARRAY_ERR21 SNAPSHOT REPOSITORY METADATA ERROR

There was a problem with the metadata of the FlashCopy (snapshot) repository during the FlashCopyoperation. To resolve this error you can recreate the FlashCopy.

v FCP_ARRAY_ERR22 REMOTE VOL MIRRORING: ILLEGAL I/O ORIGIN

The primary logical drive received I/O from a remote array, or the secondary logical drive receivedI/O from other than the primary logical drive. To resolve this error you can try the operation again.

v FCP_ARRAY_ERR23 SNAPSHOT OPERATION NOT ALLOWED

The repository capacity limit has been reached, so the FlashCopy (snapshot) operation has failed. Toresolve this error you can delete or recreate the FlashCopy.

v FCP_ARRAY_ERR24 SNAPSHOT VOLUME'S REPOSITORY FULL

The repository capacity limit has been reached. To resolve this error you can delete or recreate theFlashCopy (snapshot).

v FCP_ARRAY_ERR25 CACHED DATA WILL BE LOST IF CONTROLLER FAILS

This message is a warning that a disk array logical drive (LUN) is running with write cache enabledand cache mirroring disabled. The warning displays when the LUN is opened, and it displays againevery 24 hours until cache mirroring is enabled again.


If a controller failure or a power down occurs while the LUN is running in this mode, data that is inthe write cache (but not written to the physical disk media) might be lost. This can result in corruptedfiles, file systems, or databases.

v FCP_ARRAY_ERR26 LOGICAL VOLUME IS WRITE PROTECTED

The status of the logical drive is read-only. The probable reason is that it is a secondary logical drive ofa FlashCopy, VolumeCopy, or remote mirror pair. Check which relationship applies to the logical drive.– For FlashCopy, a status of read-only on the secondary logical drive usually indicates that the

repository is full.– For VolumeCopy, both the primary and secondary logical drives are read-only during the copy. The

secondary logical drive is read-only when the copy is stopped but the copy pair had not beendeleted.

– For remote mirroring, the secondary logical drive is always read-only, as long as the mirror is active.v FCP_ARRAY_ERR27 SINGLE CONTROLLER RESTARTED

The subsystem is operating as a single controller, and an error has been repaired. The error might havebeen a communication or hardware problem, or it might have occurred because a LUN was moved toa controller that does not have a path to the current hostIf this is a dual-controller subsystem, find the reason that the subsystem is operating insingle-controller mode, and resolve the problem. Possible reasons include the following:– An HBA, switch port, switch, DS3000, DS4000, or DS5000 port or DS3000, DS4000, or DS5000

controller was unavailable during the last system reboot or the last time the cfgmgr command wasrun.

– A user removed a path (dac) as part of a Fibre Channel adapter hot swap operation.v FCP_ARRAY_ERR28 SINGLE CONTROLLER RESTART FAILURE

The subsystem is operating as a single controller, and the error has not been repaired. There is aproblem with the path between this host and the subsystem or with the subsystem itself. The host hasattempted to communicate with the subsystem and that communication has failed.If the number of retries that is specified in the ODM attribute switch_retries is reached, the I/O is failedback to the user.Repair the error. Then, if this is a dual-controller subsystem, find the reason that the subsystem isoperating in single-controller mode, and resolve that problem. Possible reasons include the following:– An HBA, switch port, switch, DS3000, DS4000, or DS5000 port or DS3000, DS4000, or DS5000

controller was unavailable during the last system reboot or the last time the cfgmgr command wasrun.

– A user removed a path (dac) as part of a Fibre Channel adapter hot swap operation.

A new errorlog DISK_ERR7 has been created to notify that a path has been designated as failed due to apredetermined number of IO errors that occurred on the path. This is normally preceded with other errorlogs that represent the actual error that occurred on the path.

Replacing hot swap HBAsThis section describes the procedure for hot-swapping Fibre Channel host bus adapters (HBAs) on aSystem p server.

If this procedure is not followed as documented here, loss of data availability can occur. IBMrecommends that you read and understand all of the steps in this section before you begin the HBA hotswap procedure.

Known issues and restrictions for AIX

Please note the following known issues and restrictions when you perform a hot swap operation:

Caution: Any deviations from these notes and procedures might cause a loss of data availability.


v The autorecovery attribute of the dar must be set to no. Autorecovery is a dynamically set feature thatcan be turned back on after the hot swap procedure is complete. Failure to disable autorecovery modeduring a hot swap procedure can cause loss of access to data.

v Do not redistribute logical drives to the preferred path until you verify that the HBA replacementsucceeded and that the subsequent configuration was performed correctly. Redistributing the logicaldrives before verifying successful hot swap and configuration can cause a loss of access to data.

v The only supported hot swap scenario is the following operation:– Replacing a defective HBA with the same model HBA, and in the same PCI slot.

Do not insert the defective HBA into any other system, even if the HBA is found not to actually bedefective. Always return the HBA to IBM.

Important: No other variations of replacement scenarios are currently supported.v Hot swap is not supported in single-HBA configurations.

Preparing for the HBA hot swap for AIX

Complete the following procedures to prepare for the hot swap:

Collecting system data: In preparation for the hot swap procedure, complete the following steps tocollect data from the system:1. Type the following command:

# lsdev -C |grep fcs

The output is similar to the following example:

fcs0 Available 17-08 FC Adapterfcs1 Available 1A-08 FC Adapter

2. Type the following command:

# lsdev -C |grep dac

The output is similar to the following example:

dac0 Available 17-08-02 1815 DS4800 Disk Array Controllerdac1 Available 1A-08-02 1815 DS4800 Disk Array Controller

3. Type the following command for each of the fcs devices:

# lscfg -vpl fcsX

where X is the number of the fcs device. The output looks similar to the following example:


lscfg -vpl fcs0fcs0 U0.1-P1-I1/Q1 FC Adapter

Part Number.................09P5079EC Level....................ASerial Number...............1C21908D10Manufacturer................001CFeature Code/Marketing ID...2765FRU Number..................09P5080Network Address.............10000000C92D2981ROS Level and ID............02C03951Device Specific.(Z0)........2002606DDevice Specific.(Z1)........00000000Device Specific.(Z2)........00000000Device Specific.(Z3)........03000909Device Specific.(Z4)........FF401210Device Specific.(Z5)........02C03951Device Specific.(Z6)........06433951Device Specific.(Z7)........07433951Device Specific.(Z8)........20000000C92D2981Device Specific.(Z9)........CS3.91A1Device Specific.(ZA)........C1D3.91A1Device Specific.(ZB)........C2D3.91A1Device Specific.(YL)........U0.1-P1-I1/Q1

PLATFORM SPECIFIC

Name: fibre-channelModel: LP9002Node: fibre-channel@1Device Type: fcpPhysical Location: U0.1-P1-I1/Q1

4. Type the following command:

# lsdev -C |grep dar

The output looks similar to the following example:

# dar0 Available 1815 DS4800 Disk Array Routerdar1 Available 1815 DS4800 Disk Array Router

5. Type the following command to list the attributes of each dar found on the system:

# lsattr -El darX

where X is the number of the dar. The output looks similar to the following example:

lsattr -El dar0act_controller dac0,dac2 Active Controllers Falseall_controller dac0,dac2 Available Controllers Falseheld_in_reset none Held-in-reset controller Trueload_balancing no Dynamic Load Balancing Trueautorecovery no Autorecover after failure is corrected Truehlthchk_freq 600 Health check frequency in seconds Trueaen_freq 600 Polled AEN frequency in seconds Truebalance_freq 600 Dynamic Load Balancing frequency in seconds Truefast_write_ok yes Fast Write available Falsecache_size 1024 Cache size for both controllers Falseswitch_retries 5 Number of times to retry failed switches True


Verifying that autorecovery is disabled: Before you perform the hot swap, you must complete thefollowing steps to ensure that autorecovery is disabled on every dar that is involved with the HBA youwant to hot swap:1. Identify all the dac(s) that are involved with the HBA by typing the following command:

# lsdev -C|grep 11-08

The output looks similar to the following example:

# lsdev -C|grep 11-08fcs0 Available 11-08 FC Adapterfscsi0 Available 11-08-01 FC SCSI I/O Controller Protocol Devicedac0 Available 11-08-01 1742 (700) Disk Array Controllerhdisk1 Available 11-08-01 1742 (700) Disk Array Devicehdisk3 Available 11-08-01 1742 (700) Disk Array Devicehdisk5 Available 11-08-01 1742 (700) Disk Array Devicehdisk7 Available 11-08-01 1742 (700) Disk Array Devicehdisk8 Available 11-08-01 1742 (700) Disk Array Device

2. Consult the lsattr command output that you collected in step 5 on page 5-41 of the procedure“Collecting system data” on page 5-40. In the lsattr output, identify the dar(s) that list the dacs youidentified in step 1 of this procedure.

3. For each dar that you identified in step 2, type the following command:

# lsattr -El darX |grep autorecovery

where X is the number of the dar. The output looks similar to the following example:

# lsattr -El dar0 |grep autorecoveryautorecovery no Autorecover after failure is corrected True

4. In the lsattr command output, verify that the second word is no. If the second word is set to yes, thenautorecovery is currently enabled.

Important: For each dar on which autorecovery is enabled, you must disable it by setting theautorecovery ODM attribute to no. See “Using the lsattr command to view ODM attributes” on pageE-5 to learn how to change attribute settings. Do not proceed with the hot swap procedure until youcomplete this step and verify that autorecovery is disabled.

Replacing the hot swap HBA for AIX

Note: If this procedure is not followed as documented here, loss of data availability can occur. IBMrecommends that you read and understand all of the steps in this section before you begin the HBA hotswap procedure.

Complete the following steps to replace the hot swap HBA:1. Place the HBA that you want to replace into the Defined state by typing the following command:

# rmdev -Rl fcsX

where X is the number of the HBA. The output is similar to the following example:

rmdev -Rl fcs0fcnet0 Defineddac0 Definedfscsi0 Definedfcs0 Defined

For Linux operating systems, type the following command to identify the PCI Hotplug slot:


# drslot_chrp_pci -i -s slot-name

where slot-name is the name of the slot for the HBA you are replacing. (Example:U7879.001.DQD014E-P1-C3)The light at slot slot-name begins flashing, and this message displays:

The visual indicator for the specifiedPCI slot has been set to the identifystate. Press Enter to continue orenter x to exit.

2. In the AIX smit menu, initiate the process that is required for the HBA hot swap by selecting smit →Devices → PCI Hot Plug Manager → Replace/Remove a PCI Hot Plug Adapter.

3. In the Replace/Remove a PCI Hot Plug Adapter window, select targeted HBA. A window displaysthat contains instructions for replacing the HBA.

4. Replace the HBA by following the smit instructions.

Note: Do not reinstall the Fibre Channel cable at this time.5. If the steps in this procedure are completed successfully up to this point, you obtain the following

results:v The defective HBA is removed from the system.v The replacement FC HBA is powered on.v The associated fcsX device is in the Defined state.

Before continuing, verify that these results have been obtained.6. Install the Fibre Channel loop back on the replacement HBA.7. Place the HBA into the Active state by typing the following command:

# cfgmgr

Note: The new HBA is placed in the default group. If the default group has hdisks assigned to itthen the HBA will generate a new dar and dac, which will cause a split. Issue the rmdev command toremove the new dar and dac after mapping the WWPN.

8. Verify that the fcs device is now available by typing the following command:


9. Verify or upgrade the firmware on the replacement HBA to the appropriate level by typing thefollowing command:

# lscfg -vpl fcsX

where X is the number of the fcs.10. Record the 16-digit number that is associated with Network Address, as it was displayed in the

output of the command you used in step 9. This Network Address number will be used in the nextprocedure, manually map the replacement HBA's WWPN to the storage subsystem(s).

11. Place the HBA back into the Defined state by typing the following command:

# rmdev -Rl fcsX

When you have completed this procedure, continue to the next procedure, “Mapping the new WWPN tothe DS3000, DS4000, or DS5000 Storage Subsystem for AIX and Linux” on page 5-48.


Replacing IBM host bus adapters on a Linux operating systemThis section provides requirements and procedures for replacing IBM host bus adapters in Series pservers using PCI Hotplug tools.

Requirements:

PCI Hotplug tools:

Ensure that the following tools are installed in the /usr/sbin directory:v lsslot

v drslot_chrp_pci

If these tools are not installed, complete the following steps to install them:1. Ensure that rdist-6.1.5-792.1 and compat-2004.7.1-1.2 are installed from the SLES 9 media.2. Download the PCI Hotplug Tools rpm files from the following Web site:

http://www14.software.ibm.com/webapp/set2/sas/f/lopdiags/3. On the Web page, select the appropriate link for your operating system. Download and

install the following rpm files:v librtas-1.3.1-0.ppc64.rpmv rpa-pci-hotplug-1.0-29.ppc64.rpm

4. Type the following command to install each rpm file:

# rpm -Uvh <filename>.rpm

where <filename> is the name of the rpm file.

PCI core:The PCI core must be loaded on the system. Type the following command to verify:

# ls -l /sys/bus/pci/slots

If the PCI core is loaded, the output will look similar to the following:

elm17c224:/usr/sbin # ls -l /sys/bus/pci/slotstotal 0drwxr-xr-x 8 root root 0 Sep 6 04:29 .drwxr-xr-x 5 root root 0 Sep 6 04:29 ..drwxr-xr-x 2 root root 0 Sep 6 04:29 0000:00:02.0drwxr-xr-x 2 root root 0 Sep 6 04:29 0000:00:02.4drwxr-xr-x 2 root root 0 Sep 6 04:29 0000:00:02.6drwxr-xr-x 2 root root 0 Sep 6 04:29 0001:00:02.0drwxr-xr-x 2 root root 0 Sep 6 04:29 0001:00:02.6drwxr-xr-x 2 root root 0 Sep 6 04:29 control

If the /sys/bus/pci/slots directory does not exist, then the PCI core is not loaded.

rpaphp driver:The rpaphp driver must be loaded on the system. Type the following command to verify:

ls -l /sys/bus/pci/slots/*

If the rpaphp driver is loaded, the output will look similar to the following:


http://www14.software.ibm.com/webapp/set2/sas/f/lopdiags/

elm17c224:/usr/sbin # ls -l /sys/bus/pci/slots/*/sys/bus/pci/slots/0000:00:02.0:total 0drwxr-xr-x 2 root root 0 Sep 6 04:29 .drwxr-xr-x 8 root root 0 Sep 6 04:29 ..-r--r--r-- 1 root root 4096 Sep 6 04:29 adapter-rw-r--r-- 1 root root 4096 Sep 6 04:29 attention-r--r--r-- 1 root root 4096 Sep 6 04:29 max_bus_speed-r--r--r-- 1 root root 4096 Sep 6 04:29 phy_location-rw-r--r-- 1 root root 4096 Sep 6 04:29 power

Listing information about the I/O slots: Before you replace an HBA using PCI Hotplug, you can use thelsslot tool to list information about the I/O slots. This section describes how to use lsslot, and providesexamples.

Syntax for the lsslot command: Use the lsslot tool as follows:v Syntax: lsslot [ -c slot | -c pci [ -a | -o]] [ -s drc-name ] [ -F delimiter ]

v Options:

No optionsDisplays all DR slots

-c slot Displays all DR slots

-c pci Displays all PCI Hotplug slots

-c pci -aDisplays all available (empty) PCI Hotplug slots

-c pci -oDisplays all occupied PCI Hotplug slots

-F Uses delimiter to delimit columns

Listing PCI Hotplug slots using the lsslot command: This section shows the command lines you can use tolist PCI Hotplug slots.

Note: In the Device(s) columns of the command-line outputs, the PCI devices in the slots are listed asfollows: xxxx:yy:zz.t. (For example: 0001:58:01.1)

List all PCI Hotplug slots: Type the following command to list all PCI Hotplug slots:

# lsslot -c pci -a

The resulting output looks similar to the following:

# Slot Description Device(s)U7879.001.DQD014E-P1-C1 PCI-X capable, 64 bit, 133MHz slot EmptyU7879.001.DQD014E-P1-C2 PCI-X capable, 64 bit, 133MHz slot 0002:58:01.0U7879.001.DQD014E-P1-C3 PCI-X capable, 64 bit, 133MHz slot 0001:40:01.0U7879.001.DQD014E-P1-C4 PCI-X capable, 64 bit, 133MHz slot EmptyU7879.001.DQD014E-P1-C5 PCI-X capable, 64 bit, 133MHz slot EmptyU7879.001.DQD014E-P1-C6 PCI-X capable, 64 bit, 133MHz slot 0001:58:01.00001:58:01.1

List all empty PCI Hotplug slots: Type the following command to list all empty PCI Hotplug slots:

# lsslot -c pci -a



# Slot Description Device(s)U7879.001.DQD014E-P1-C1 PCI-X capable, 64 bit, 133MHz slot EmptyU7879.001.DQD014E-P1-C4 PCI-X capable, 64 bit, 133MHz slot EmptyU7879.001.DQD014E-P1-C5 PCI-X capable, 64 bit, 133MHz slot Empty

List all occupied PCI Hotplug slots: Type the following command to list all occupied PCI Hotplug slots:

# lsslot -c pci -o


# Slot Description Device(s)U7879.001.DQD014E-P1-C2 PCI-X capable, 64 bit, 133MHz slot 0002:58:01.0U7879.001.DQD014E-P1-C3 PCI-X capable, 64 bit, 133MHz slot 0001:40:01.0U7879.001.DQD014E-P1-C6 PCI-X capable, 64 bit, 133MHz slot 0001:58:01.00001:58:01.1

Show detailed information about a particular device: Select a device number from the output of #lsslot -c pci -o, as seen in the previous output example, and type the following command to showdetailed information about that particular device:

# lspci | grep xxxx:yy:zz.t

where xxxx:yy:zz.t is the number of the PCI Hotplug device. The resulting output looks similar to thefollowing:

0001:40:01.0 Ethernet controller: Intel Corp. 82545EM GigabitEthernet Controller (Copper) (rev 01)

Replacing a PCI Hotplug HBA

Replacing an HBA: Complete the following procedures to replace a PCI Hotplug HBA by using thedrslot_chrp_pci command.

Note: In these procedures, the variable slot-name refers to the slot that contains the HBA that you arereplacing.

Attention: Before you remove the HBA, you will need to remove the Fibre Channel cable that is attachedto the HBA. The Fibre Channel cable must remain unattached for at least five minutes to ensure that allI/O activity is transferred to the alternate path. Failure to remove the Fibre Channel cable can lead toundesirable results.

1. Identify the PCI Hotplug slot: Type the following command to identify the PCI Hotplug slot:

# drslot_chrp_pci -i -s slot-name

where slot-name is the name of the slot for the HBA you are replacing. (Example: U7879.001.DQD014E-P1-C3)

The light at slot slot-name begins flashing, and this message displays:


2. Hot unplug the HBA from the slot: Complete the following steps to hot unplug (remove) the HBA:


1. Remove the Fibre Channel cable that is connected to this HBA, and wait for failover to complete.2. After failover is complete, type the following command:

# drslot_chrp_pci -r -s slot-name

This message displays:


3. Press Enter. This message displays:

The visual indicator for the specifiedPCI slot has been set to the action state.Remove the PCI card from the identified slotand press Enter to continue.

4. Press Enter.5. Physically remove the HBA from the slot.6. Type the following command to verify that the slot is empty:

# lsslot -c pci -s slot-name

If the slot is empty, the resulting output looks similar to the following:

# Slot Description Device(s)U7879.001.DQD014E-P1-C3 PCI-X capable, 64 bit, 133MHz slot Empty

3. Hot plug the HBA into the slot: Complete the following steps to hot plug the HBA into the slot.1. Type the following command:

# drslot_chrp_pci -a -s slot-name

This message displays:


2. Press Enter. This message displays:

The visual indicator for the specifiedPCI slot has been set to the action state.Insert the PCI card into the identified slot,

connect any devices to be configuredand press Enter to continue. Enter x to exit.

3. Insert the new HBA into the slot.4. Type the following command to verify that the slot is no longer empty:

# lsslot -c pci -s slot-name

If the slot is not empty, the resulting output looks similar to the following:


# Slot Description Device(s)U7879.001.DQD014E-P1-C3 PCI-X capable, 64 bit, 133MHz slot 0001:40:01.0

Mapping the new WWPN to the DS3000, DS4000, or DS5000 Storage Subsystemfor AIX and LinuxFor each DS3000, DS4000, or DS5000 Storage Subsystem that is affected by the hot swap, complete thefollowing steps to map the worldwide port name (WWPN) of the HBA to the storage subsystem:1. Start DS Storage Manager and open the Subsystem Management Window.2. In the Mapping View of the Subsystem Management window, select Mappings → Show All Host Port

Information. The Host Port Information window displays.3. Find the entry in the Host Port Information window that matches the WWPN of the “defective” HBA

(the HBA that you removed), and record the alias name. Then, close the Host Port Informationwindow.

4. In the Mapping View, select the alias name of the HBA host port that you just recorded.5. Select Mappings → Replace Host Port. The Replace Host Port window opens.6. In the Replace Host Port window, verify that the current HBA Host Port Identifier, which is listed at

the top of the window, exactly matches the WWPN of the HBA that you removed.7. Type the 16-digit WWPN, without the : (colon), of the replacement HBA in the New Identifier field,

and click OK.

When you have completed these steps continue to the next procedure, “Completing the HBA hot swapprocedure.”

Completing the HBA hot swap procedureComplete the following steps, for both AIX and Linux, to finish replacing the hot swap HBA:1. AIX and Linux: Remove the Fibre Channel loop back plug, and insert the Fibre Channel cable that

was previously attached to the HBA that you removed.2. AIX and Linux: If an HBA is attached to a Fibre Channel switch, and the zoning is based on WWPN,

modify the zoning information to replace the WWPN of the former HBA with the WWPN of thereplacement HBA.

Note: Skip this step if the HBA is directly attached to the DS3000, DS4000, or DS5000 StorageSubsystem, or if the Fibre Channel switch zoning is based on port numbers instead of WWPNs. Ifyou do need to modify the zoning, failure to correctly do so will prevent the HBA from accessing thestorage subsystem.

3. Linux: If RDAC is installed, type the following command to recognize the new HBA:

# mppBusRescan

4. AIX: Run the cfgmgr command. (Run cfgmgr at this time to allow the HBA to register its WWPN inthe Fibre Channel switch.)

5. AIX: Type the following commands to verify that the replaced fcsX device and its associated dac(s)are placed in the Available state:


lsdev -C |grep dac

6. AIX: Type the following step to verify that no additional dar(s) have been created and that theexpected dar(s) are in the Available state.

Note: With MPIO the only time you have a dac device is when the UTM LUN is assigned.


# lsdev -C |grep dar

Caution: The presence of additional dar(s) in the lsdev output indicates a configuration problem. Ifthis occurs, do not continue this procedure until you correct the problem, Loss of data availabilitycan occur.

7. AIX: For each dar, type the following command to verify that affected dar attributes indicate thepresence of two active dac(s):

# lsattr -El darX|grep act_controller

where X is the number of the dar.The output looks similar to the following:

lsattr -El dar0|grep act_controlleract_controller dac0,dac2 Active Controllers False

Caution: If two dacs are not reported for each affected dar, loss of data availability can occur. Do notcontinue this procedure if two dac(s) are not reported for each dar. Correct the problem beforecontinuing.

8. AIX: Manually redistribute volumes to preferred paths.9. AIX: Verify that disks stay on preferred path by using one or both of the following methods:

Using the AIX systemRun the mpio_get_config -Av command, and verify that drives are on expected path.

Using Storage ManagerIn the Enterprise Management Window, verify that the storage subsystem(s) are Optimal. Ifthey are not Optimal, verify that any drives that are part of the subsystems involved withhot swap process are not listed in the Recovery GURU.

10. AIX: If necessary, enable autorecovery of the affected dar(s) at this time. (See Appendix E, “Viewingand setting AIX Object Data Manager (ODM) attributes,” on page E-1 to learn how to changeattribute settings.)

Result: The Fibre Channel HBA hot swap is now complete.



Chapter 6. Working with full disk encryption

This chapter describes the capabilities and advantages of full disk encryption (FDE) disk drives and howto implement security on DS5000 storage systems that are equipped with FDE disks. The following topicsare addressed:v “FDE disk drives”v “FDE security authorizations” on page 6-11v “FDE key terms” on page 6-12v “Configuring DS5000 disk encryption with FDE drives” on page 6-13v “Frequently asked questions” on page 6-31

FDE disk drivesFull disk encryption (FDE) disk drives enable you to significantly reduce the security vulnerabilities ofstored data. FDE disk drives that adhere to the Trusted Storage Group (TCG) enterprise securitysubsystem class specification are National Security Agency qualified and provide unparalleled securitywith government-grade encryption.

Note: No single security implementation can effectively secure all levels of data against all threats.

Different technologies are required to protect data that is stored on hard disk drives against differentthreats. FDE drives ensure the security of stored data through the following methods:v Securing stored data against a breach: If an unauthorized user gains possession of a disk drive that

contains encrypted data (the drive is removed from the data center or powered down) the data isprotected.

v Permanently erasing stored data: Secure erase provides fast, permanent erasure of data on drives thatare planned for reuse or disposal.

FDE drives secure data against threats when the drive eventually leaves the owner's control but cannotprotect data from threats that occur within the data center or on the network. If an attacker gains accessto a server and can access an unlocked drive, the attacker can read the clear text that comes from thedrive. Remember that drive-level encryption technology does not replace the access controls of the datacenter; rather, it complements them.

Securing data against a breachDrives with the full disk encryption technology are security capable. Each FDE drive comes from thefactory in Security Capable (security disabled) state. In this state, the FDE drives behave exactly like thenon-FDE drives. The data that is stored on them is not protected when the drives are removed from thestorage subsystem. They can be moved from one storage subsystem to another without the having beunlocked with a security key file. They can also be used as part of a RAID array that is composed ofnon-encrypting (non-FDE) disks. However, a RAID array that is composed of Security Capable FDE andnon-FDE drives cannot be converted into a secured RAID array at a later time, leaving the data on theFDE drives unprotected if they are removed from the storage subsystem.

The IBM DS storage subsystem controllers can apply security to every FDE drive in a RAID array that iscomposed entirely of FDE drives. The controller firmware creates a security key and activates theencryption function of the drive, which causes each FDE disk drive to randomly generate an encryptionkey that is embedded on the disk. When security is enabled, the FDE drive automatically performs fulldisk encryption:


v When a write operation is performed, clear text enters the disk and is encrypted before it is written tothe media, using the disk encryption key.

v When a read operation is performed, encrypted data that is read from the media is decrypted before itleaves the drive.

During normal operation, whether the FDE drive is in Security Capable or Security Enabled state, itbehaves the same as a non-encrypting disk to the storage subsystem. A security-enabled FDE drive isconstantly encrypting data. Disk encryption cannot be accidentally turned off. The disk encryption key isgenerated by the drive itself, is stored on the disk, never leaves the disk, and is unique to that drivealone. To ensure that security is never compromised, an encrypted version of the encryption key is storedon the disk drive only. Because the disk encryption key never leaves the disk, you might not have toperiodically change the encryption key, the way a user might periodically change the operating-systempassword.

Creating a security keyWith full disk encryption the process of securing a drive is simple and consists of enabling security onthe DS5000 and then securing the specific security-capable RAID arrays where the data is stored.

Enabling security on the DS5000 storage subsystem is a one-time process, unless the you decide at a laterdate to change the security key. Separate security keys are not required for each individual drive. Toenable security on the DS5000, you must purchase an IBM DS Disk Encryption premium feature key andenable the feature in the DS5000 subsystem, using the instructions that come with the premium featurekey entitlement kit.

A security key must then be generated for the storage subsystem. The process of creating a security keyrequires you to the enter the security key identifier, the pass phrase, and the security key file name andlocation. The controllers create the security key and use this key to secure the security-enabled FDEdrives. The controllers use one security key to unlock all of the security-enabled FDE drives in thestorage subsystem even though each FDE drive has its own unique encryption key. An encrypted versionof the security key is obfuscated in the storage subsystem to maintain security from hackers. You cannotsee the security key directly, but its encrypted version is saved in a backup file at a location that youspecify. In addition to the saved location that you specify, the storage manager also saves a copy of thefile in the default location ...\IBM_DS\client\data\securityLockKey in a Microsoft Windowsenvironment or in /var/opt/SM/securityLockkey in AIX, Linux, Solaris, and HP-UX environments.

The security key file contains the security key identifier and the encrypted security key. The pass phraseis not stored anywhere in the storage subsystem or in the security key file. The controller uses the passphrase to encrypt the security key before it exports the security key to the security key file. The securitykey identifier is stored in the security key file so that you can identify which storage subsystem thesecurity key file is for.

Attention: The pass phrase is used only to protect the security key in the security key file. Anyone whocan access the Subsystem Management window can save a copy of the security key file with a new passphrase. Set a storage subsystem password for each of the DS5300 or DS5100 storage subsystems, whichrequires you to provide a password when any configuration changes are made, including creating andchanging the security key. See “Storage subsystem password protection” on page 3-10 for instructions forsetting the subsystem password.

The security key file provides protection against a corrupted security key or the failure of both controllersin the storage subsystem. The security key file is also needed to unlock security-enabled FDE driveswhen they are moved from one storage subsystem to another. In these cases, the security-enabled FDEdrives remain locked until the drives are unlocked by the security key that is stored in the security keyfile. To decrypt the security key in the security key file, you must provide the same pass phrase that wasentered when the security key file was generated. The drive then determines whether its security key andthe security key that was provided by the storage subsystem are the same. If they are the same, data canbe read from and written to the security-enabled FDE drives.


After the storage subsystem controller creates the security key, the RAID arrays can be changed from astate of Security Capable to a state of Security Enabled. The Security Enabled state requires the RAIDarray FDE drives to be unlocked during the drive power up process using the security key to access thedata that is stored on the drives. Whenever power is applied to the drives in a RAID array, the drives areall placed in Security Locked state. They are unlocked only during drive initialization with the storagesubsystem security key. The Security Unlocked state makes the drives accessible for the read and writeactivities. After they are unlocked, the drives remain unlocked until the power is removed from thedrives, the drives are removed and reinserted in the drive bays, or the storage subsystem power iscycled.

After a drive is secured, if it is ever powered down or removed, the drive becomes locked. Theencryption key within that drive will not encrypt or decrypt data, making the drive unreadable until it isunlocked by the controllers.

After authentications are established and security is enabled on an array, the encryption of writeoperations and decryption of read operations that takes place inside the FDE drive are not apparent tothe user or to the DS5000 controllers. However, if a secured drive is lost, removed, or stolen, the drivebecomes locked, and the data that is stored on the disk remains encrypted and unreadable. Because anunauthorized user does not have the security key file and pass phrase, gaining access to the stored datais impossible.

Figure 6-1. Security-enabled FDE drives: With the correct authorizations in place, the reading and writing of dataoccurs in Unlocked state

Chapter 6. Working with full disk encryption 6-3

Changing a security keyWhen you change a security key, a new security key is generated by the controller firmware. The newsecurity key is obfuscated in the storage subsystem, and you cannot see the security key directly. Thenew security key replaces the previous key that is used to unlock the security-enabled FDE drives in thestorage subsystem. The controller negotiates with all of the security-enabled FDE drives for the new key.However, an n-1 version of the security key is also stored in the storage subsystem for protection in casesomething prevents the controllers from completing the negotiation of the new security key with thesecurity-enabled FDE drives (for example, loss of storage subsystem power during the key changeprocess). If this happens, you must change the security key so that only one version of the security key isused to unlock drives in a storage subsystem. The n-1 key version is stored in the storage subsystem only.It cannot be changed directly or exported to a security key file.

A backup copy of the security key file is always generated when you change a security key and shouldbe stored on some other storage medium in case of controller failure, or for transfer to another storagearray. You participate in creation of the security key identifier, the pass phrase, and the security key filename and location when you change the security key. The pass phrase is not stored anywhere in thestorage subsystem or in the security file. The controller uses the pass phrase to encrypt the security keybefore it exports the security key to the security key file.

Security key identifierFor additional protection, the security key that is used to unlock FDE drives is not visible to the user. Thesecurity key identifier is used to refer to a security key value instead. You can see the security keyidentifier during operations that involve the drive security key file, such as creating or changing thesecurity key.

Figure 6-3 on page 6-5 shows an example of the security key identifier field when you are performing achange security key operation.

Figure 6-2. A security-enabled FDE drive is removed from the storage subsystem: Without correct authorizations, astolen FDE disk cannot be unlocked, and the data remains encrypted


The Change Security Key Complete window shows that the security key identifier that was written to thesecurity key file has a random number appended to the security key identifier you entered in Figure 6-3and the storage subsystem worldwide identifier. Figure 6-4 on page 6-6 shows an example of the randomnumber part of the security key identifier.

Figure 6-3. Changing the security key


The security key identifier field in the FDE Drive Properties window includes a random number that isgenerated by the controller when you create or change the security key. Figure 6-5 on page 6-7 shows anexample of the random number. The random number is currently prefixed with 27000000. If all of thesecured FDE drives in the storage subsystem have the same value in the security key identifier field, theycan be unlocked by the same security key identifier. Note that the Security Capable and Secure fields inthe drive Properties window show whether the drive is secure capable and whether it is in Secure (Yes)or Unsecured (No) state.

Figure 6-4. Changing the security key - Complete


Figure 6-6 on page 6-8 shows an example of the security key identifier that is displayed in the fileinformation field when you select a security key back up file to unlock the secured drives in the storagesubsystem. The security key identifier or LockKeyID, shown in the file information field, contains the

Figure 6-5. Drive properties - Secure FDE drive


characters that you entered in the security key identifier field when you created or changed the securitykey along with the storage subsystem worldwide identifier and the randomly-generated number thatappears in the security key identifier of all secured FDE drives. This information is delimited by a colon.For example, LockKeyID:

Passw0rdplus3:600a0b800029ece6000000004a2d0880:600a0b800029ed8a00001aef4a2e4a73

contains the following information:v the security key identifier that you specified, for example Passw0rdplus3

v the storage subsystem worldwide identifier, for example 600a0b800029ece6000000004a2d0880

v a randomly-generated number 600a0b800029ed8a00001aef4a2e4a73

Figure 6-7 on page 6-9 shows an example of the drive properties for an unsecured FDE drive. Note thatthe security key identifier field for an unsecured FDE drive is populated with zeros. Note also that the

Figure 6-6. Select file - LockKeyID


Security Capable field value is yes and the Secure field value is no, indicating that this is a securitycapable but unsecured FDE drive.

Figure 6-7. Drive properties - Unsecured FDE drive


Unlocking secure drivesYou can export a RAID array with its security-enabled FDE drives to a different storage subsystem. Afteryou install those drives in the new storage subsystem, you must unlock the security-enabled FDE drivesbefore data can be read from or written to the drives. The security key on the new storage subsystem willbe different and will not unlock the drives. You must supply the security key from a security key file thatyou saved from the original storage subsystem. In addition, you must provide the pass phrase that wasused to encrypt the security key to extract the security key from the security key file. After you unlockthe drives by using the security key in the security key file, the controller negotiates the existing securitykey for these drives so that only one version of the security key is used to unlock drives in a storagesubsystem.

You do not have to provide the security key file to unlock the security-enabled drives in a storagesubsystem every time the storage subsystem power is cycled or the drives are removed and reinserted inthe same storage subsystem, because the controllers always keep a copy of the current and the previous(n-1) values of the security key to unlock these drives. However, if the drives are removed from thesubsystem and the security key is changed more than two times in the same subsystem, the controllerswill not have the security key to unlock the drives when they are reinserted in the same storagesubsystem.

Attention: Always back up the data in the storage subsystem to secured tape to prevent loss of datadue to malicious acts, natural disasters, abnormal hardware failures, or loss of the FDE security key.

Secure eraseSecure erase provides for the protection of FDE drives from security threats when they are eventuallyretired, returned, discarded, or repurposed. As these drives are moved from the data center or reused, itis critical that the data on the disks be permanently erased and not vulnerable to recovery. Discardeddrives might still have residual data that can be reconstructed by an unauthorized user. Secure eraseprotects against this threat by cryptographically erasing the data.

The traditional methods that are used to permanently erase data often prove to be expensive and slowand might not provide the highest level of data erasure. Traditional methods might also put the drivesbeyond your control and therefore subject to a data breach. Secure erase provides the followingadvantages compared to traditional methods:v Immediate, cryptographic data erasurev Lower overall costsv A higher level of media sanitation as per the National Institute of Standard and Technology (NIST)

Secure erase with FDE drives allows for immediate erasure of data without requiring that the drive beremoved from the data center. With just a few clicks, you can quickly reuse or discard a drive. Costsavings with secure erase are realized because drives are not destroyed but instead are erased to be usedagain and again. This eliminates the need to destroy a drive, while still securing warranty and expiredlease returns or enabling drives to be reused securely. Per the NIST, secure erase is considered a type ofdata purging, which is regarded as a higher level of data sanitation than traditional methods.

Secure erase prompts the FDE drive to permanently erase the current the encryption key and replace itwith a new randomly generated encryption key within the drive. The new drive encryption key is usedto encode and decode all data on the disk. After the encryption key is changed, any data that waspreviously written to the disk becomes unintelligible. Data that was encrypted with the previousencryption key is unintelligible when it is decrypted with the new encryption key. This includes all bits,headers, and directories. The data is completely and permanently inaccessible.


FDE security authorizationsThe following table identifies and describes the authorization parameters that are used to implementsecurity on DS5000 storage systems.

Table 6-1. Security authorizations

Parameter DescriptionWhere is it located andmanaged? How is it generated?

Encryption key The encryption key is used toencrypt and decrypt data onthe FDE disk drive.

Is stored on and is managedby the FDE disk drive:

v Is never transferred from thedrive

v Each drive has its ownunique encryption key

The encryption key isgenerated when the drive ismanufactured and thenregenerated at the customersite (by a command from thecontroller to the drive) toensure that the key was notcompromised prior to use.

Security key The security key is needed tounlock the encryption key forencrypting and decrypting tooccur. One security key iscreated for all FDE drives onthe storage subsystem. Thesecurity is sometimesreferred to as the lock key.

Is stored on and is managedby the controller. A singlesecurity key is synchronizedfor all controllers in a storagesubsystem.

The security key is generatedby the storage subsystem andis encrypted and hidden in thesubsystem.

Security keyidentifier

The security key identifier ispaired with the security keyto help you remember whichkey to use for secureoperations. The security keyidentifier can be left blank oryou may specify up to 189alphanumeric characters.

The security key identifier isstored in a special area of thedisk:

v Can always be read from thedisk

v Can be written to the diskonly if security has beenenabled and the drive isunlocked

User-specified alphanumericcharacter string. Thesubsystem adds the storagesubsystem worldwideidentifier and a randomlygenerated number to thecharacters that are entered.

Figure 6-8. Secure erase process


Table 6-1. Security authorizations (continued)

Parameter DescriptionWhere is it located andmanaged? How is it generated?

Pass phrase The pass phrase is used toencrypt the security key andthe security key identifier.The pass phrase is auser-specified alphanumericcharacter string, eightcharacters minimum, 32characters maximum. It mustcontain at least one number,one lowercase letter, oneuppercase letter, and onenonalphanumeric character(such as <>&@+-). Spaces arenot allowed and it is casesensitive.

User-specified alphanumericcharacter string, not storedanywhere on the subsystem orin the security key file. Thepass phrase is used to encryptthe security key when it isexported in the security keyfile. It is also used to decryptthe key in the security filewhen it is used to importsecurity-enable FDE drives intoa storage subsystem.

User-specified alphanumericcharacter string.

Security key file File where the security keyidentifier is saved along withthe encrypted security key.

File name and location aredetermined by theadministrator. In addition tothe administrator-specifiedlocation, the storage manageralso saves a copy of thesecurity key backup file in thedefault location. SeeAppendix L, “FDE bestpractices,” on page L-1 formore information.

Generated by the storagesubsystem after you initiate acreate security key, changesecurity key, or save securitykey operation.

FDE key termsThis table defines key terms that are used throughout this chapter.

Table 6-2. Full disk encryption key terms

Term Description

FDE Full disk encryption, a custom chip or ASIC (application specific integrated circuit) onthe disk drive that requires a security key to allow encryption and decryption to begin.FDE disk drives encrypt all the data on the disk. The secured drive requires that asecurity key be supplied before read or write operations can occur. The encryption anddecryption of data is processed entirely by the drive and are not apparent to the storagesubsystem.

Secure erase Permanent destruction of data by changing the drive encryption key. After secure erase,data that was previously written to the disk becomes unintelligible. This feature takesadvantage of FDE disk security capabilities to erase data by changing the encryption keyto a randomly generated value. Because the encryption key never leaves the drive, thisprovides a secure erase. After secure erase, the drive becomes unlocked, allowinganyone to read or write to the disk. Secure erase is sometimes referred to as drivereprovisioning.

Local key management Management of the security keys and key linkage between the storage subsystem andthe FDE drives.


Table 6-2. Full disk encryption key terms (continued)

Term Description

Locked The state that a security-enabled FDE drive enters when it has been removed from andthen reinserted in the storage subsystem, or when the storage subsystem powered off.When storage subsystem power is restored, the drive remains in the Locked state. Datacannot be written to or read from a locked disk until it is unlocked by the controller,using the security key. If the controller does not have the security key, the security keyfile and its pass phrase are required to unlock the drives for read and write operations.

Repurposing/Reprovisioning

Changing a drive from being in Secured state to Unsecured state so that the drive can bereused. Reprovisioning the drive is accomplished by secure erase.

Secure array An array on security-enabled FDE drives.

Security-capable drive An FDE drive that is capable of encryption but is in Unsecured state (security notenabled).

Security-enabled drive An FDE drive with security enabled. The security-enabled FDE drive must be unlockedusing the security key during the power up process before read or write operations canoccur.

Unlocked The state of a security-enabled FDE drive in which data on the disk is accessible forread and write operations.

Configuring DS5000 disk encryption with FDE drivesThis section provides the procedures for enabling full disk encryption and creating secure arrays on theDS5000 storage subsystem. To configure disk encryption with FDE disks, perform the following tasks:1. Install the FDE drives (see “Installing FDE drives”).2. Enable the DS5000 disk encryption feature (see “Enabling the DS5000 full disk encryption feature” on

page 6-14).3. Create an array and enable array security (see “Securing a RAID array” on page 6-16).

A security-enabled FDE drive becomes locked when it is powered down or removed from the storagesubsystem. To unlock a locked drive, see “Unlocking disk drives” on page 6-21.

With the DS5000, drives can be migrated as a complete array into another DS5000 storage subsystem. Tomigrate a secure array, see “Migrating disk drives” on page 6-23.

Installing FDE drivesTable 6-3 lists the FDE disk drives that the IBM DS5000 supports, as of the date of this document. See theIBM System Storage DS3000, DS4000, or DS5000 Hard Drive and Storage Expansion Enclosure Installation andMigration Guide and the DS5000 Interoperability Guide for installation procedures and the most up-to-datesupport information.

Table 6-3. DS5000 supported FDE drives

FDE disks supported for the DS5000

v 4 GBps Fibre Channel, 146.8 GB/15k

v 4 GBps Fibre Channel, 300 GB/15k

v 4 GBps Fibre Channel, 450 GB/15k

Note: If the FDE drive is in Security Enabled state and you do not want to preserve the data on thedrive, perform a secure erase on each drive before you use it as part of a new RAID array. Secure eraseforces the drive to generate a new encryption key, places the drive in Unsecured state, and ensures thatany data that was previously stored on the disk is erased. See “Secure erase” on page 6-10 for moreinformation.


Enabling the DS5000 full disk encryption featureThe full disk encryption optional premium feature must be enabled on the DS5000, using the instructionsthat come with the IBM DS Disk Encryption premium feature key entitlement kit. To verify that full diskencryption is enabled, on the Setup page, select View/Enable Premium Features. In the PremiumFeatures and Feature Pack Information window, Drive Security: Enabled indicates that the full diskencryption optional feature is enabled.

Enabling full disk encryption includes creating the security authorizations that you will need later tounlock a secured FDE drive that has been turned off or removed from the storage subsystem. Theseauthorizations include the security key identifier, a pass phrase, and the security key file. The securityauthorizations apply to all the FDE drives within the storage subsystem and are critical in case a drivemust be unlocked during the drive power-up process.

To create the security authorizations for full disk encryption, complete the following steps:1. From the IBM System Storage DS ES window, click Storage Subsystem, click Drive Security, and

click Create Security Key.

2. Enter a security key identifier, the security key file name and location, and a pass phrase in the CreateSecurity Key window:


v Security key identifier: The security key identifier is paired with the subsystem worldwideidentifier and a randomly generated number and is used to uniquely identify the security key file.The security key identifier can be left blank or can be up to 189 characters.

v Pass phrase: The pass phrase is used to decrypt the security key when it is read from the securitykey file. Enter and record the pass phrase at this time. Confirm the pass phrase.

v Security key backup file: Click Browse next to the file name to select the security key file nameand location, or enter the value directly in the field. Click Create Key.

Note: Save the security key file to a safe location. The best practice is to store the security key filewith your key management policies. It is important to record and remember where this file isstored because the security key file is required when a drive is moved from one storage subsystemto another or when both controllers in a storage subsystem are replaced at the same time.

3. In the Create Security Key Complete window, record the security key identifier and the security keyfile name; then, click OK. The authorizations that are required to enable security on FDE drive in theDS5000 are now in place. These authorizations are synchronized between both controllers in theDS5000 storage subsystem. With these authorizations in place, arrays on the FDE drives in the storagesubsystem can be secured.Attention: For greater security, store more than one copy of the pass phrase and security key file. Donot specify the default security file directory as the location to store your copy of the security key file.If you specify the default directory as the location to save the security key file, only one copy of thesecurity key file will be saved. Do not store the security key file in a logical drive that is mappedfrom the same storage subsystem. See Appendix L, “FDE best practices,” on page L-1 for moreinformation.


Securing a RAID arrayAn array is secured when the FDE drives in the array are security enabled. The FDE drives in a securedarray become locked if they are powered down or removed from the storage subsystem.

All drives in the array must be security-capable FDE drives with security not enabled. The array mustnot contain any FlashCopy base logical disks or FlashCopy repository logical disks. Base logical disks andFlashCopy logical disks can be written to the disks only after security is enabled.

To create a RAID array and then secure it, complete the following steps:1. Create a RAID array from the FDE drives that are available in the storage subsystem and then secure

it. From the Setup page, click Configure Storage Subsystem.

2. In the Select Configuration Task window, click Manual (advanced), click Create arrays and logicaldrives, and then click OK.


3. In the Create Arrays and Logical Drives window, select Create a new array using unconfiguredcapacity. If other (non-FDE) drive types are also installed in the DS5000, be sure to select only FibreChannel FDE drives. Click Next to continue.

4. Use the Create Array wizard to create the array. Click Next to continue.


5. In the Array Name & Drive Selection window, enter an array name (for example, Secure_Array_1).Note that the Create a secure array check box has been preselected in this window. Clear the Createa secure array check box and select Manual (Advanced) under Disk selection choices. Click Next tocontinue.

Note: The Create a secure array check box is displayed and selected only if the full disk encryptionpremium feature is enabled. If you select this check box when you create an array, the array that iscreated will be secured, and the Manual (Advanced) option is not needed to secure the array.

6. Configure drives for the array in the Manual Drive Selection window:a. Select a RAID level (for example, RAID 5).b. From the Unselected drives list, select the security-capable drives that you want to use and click

Add to add them to the Selected drives list (for example, select the disk drives in slots 2 through6 from enclosure 8).

c. Click Calculate Capacity to calculate the total capacity of the selected drives.d. Click Finish to complete the array.


Note: These drives are not yet secure. They are secured later in the process.

7. In the Array Created window, click OK to acknowledge successful creation of the array.

8. A new wizard opens, prompting you to create logical drives in the array. Use the wizard to createthe desired logical drives. After the logical drives are created, continue to the next step. SeeChapter 4, “Configuring storage,” on page 4-1 for more information about creating logical drives.

9. Secure the array that you have created:a. In the Subsystem Management window, click the Logical/Physical tab.

Note: The blue dots below the disk icons on the right side of the window indicate which diskscompose the array.


b. To enable security on the array, right-click the array name; then, click Secure Drives.

c. In the Confirm Array Drive Security window, click Yes to secure the array.

Note:


1) If you move a drive to a separate storage subsystem or if you change the security key morethan two times in the current subsystem while the drive is removed from the subsystem, youmust provide the pass phrase, the security key, and the security key file to unlock the driveand make the data readable.

2) After an array is secured, the only way to remove security is to delete the array. You canmake a volume copy of the array and save it to other disks so that the data can continue tobe accessed.

10. In the Subsystem Management window, click the Logical/Physical tab, and note that the array issecured, as indicated by the lock symbol to the left of the array name.

Unlocking disk drivesA security-enabled FDE drive becomes locked when it is powered down or removed from the storagesubsystem. This is a key feature of DS5000 disk encryption and FDE drives; the locked state makes thedata unreadable to unauthorized users.

Because the controllers always keep a copy of the current and the previous security key, the security keyfile is not needed every time the storage subsystem power is cycled or a drive is removed and reinsertedin the same storage subsystem. However, if a drive is moved to another storage subsystem, or if thesecurity key in the same storage subsystem is changed more than two times while the drive is removedfrom the storage subsystem, the pass phrase and security file are required to unlock the drive.

Note: Security-enabled FDE drives remain unlocked during firmware updates or while components arereplaced. The only time these drives are locked is when they are turned off or removed from thesubsystem.

To unlock a locked FDE drive, complete the following steps:1. In the Subsystem Management window, click the Logical/Physical tab.


2. Right-click the drives that you want to unlock; then, click Unlock.

Note: If there are multiple drives to be unlocked, you have to select only one drive. The storagemanager automatically lists all of the drives that are locked in the storage subsystem and checks eachdrive against the supplied security key file to see whether it can use the key in the security key file.

3. In the Unlock Drives window, the locked drives that you selected are listed. To unlock these drives,select the security key file, enter the pass phrase, and then click Unlock. The storage subsystem uses


the pass phrase to decrypt the security key from the security key file. The storage subsystem thencompares the decrypted security key to the security key on the drive and unlocks all the drives forwhich the security key matches.

Note: The authentication process occurs only when the drive is in Locked state because the drive waspowered on after a power-down event. It does not repeat with each read and write operation.

4. In the Unlock Drives Complete window, click OK to confirm that the drives are unlocked. Theunlocked drives are now ready to be imported.

Migrating disk drivesWith the DS5000, you can migrate drives as a complete storage subsystem into another DS5000 storagesubsystem by using existing disk group migration techniques. User data remains intact on the disksbecause configuration metadata is stored on every drive in the DS5000. FDE security-enabled drives canalso be migrated and remain secure with a few additional steps that are described in this section.

Note:

1. The following procedure covers only the additional data migration steps that are required for securearrays. For complete information and procedures, see the IBM System Storage DS3000, DS4000, orDS5000 Hard Drive and Storage Expansion Enclosure Installation and Migration Guide.

2. The following data migration steps also apply when you replace both controllers in the storagesubsystem. All drives in that subsystem must be included. Partial migrations are not supported when


you replace both controllers. A security file will be needed in this case because you might not havemanagement access to the storage subsystem to export the current security key if both of thecontrollers are to be replaced.

1. Make sure that the Drive Security feature is enabled on the storage subsystem to which you aretransferring the drives.

2. Save the security key that is used to unlock the drives in the existing storage subsystem into asecurity key file before you remove the drives from the existing storage subsystem.

3. Export the security key, pass phrase, and the security key file. The security key file then can betransferred from one storage subsystem to another.a. In the Subsystem Management window, click Storage Subsystem, click Drive Security, and click

Save Security Key File.

b. In the Save Security Key File - Enter Pass Phrase window, select a file save location, and enter andconfirm the Pass phrase; then, click Save.


4. After you replace the existing storage subsystem controller enclosure with a new controller enclosure,unlock the security-enabled FDE drives before you import the RAID arrays:a. Click the Logical/Physical tab in the Subsystem Management window.b. Right-click the drives that you want to unlock; then, click Unlock.


c. Select the security key file for the selected drives and enter the pass phrase that you entered whensaving the security key back up file; then, click Unlock.


Erasing disk drivesSecure erase provides a higher level of data erasure than other traditional methods. When you initiatesecure erase with the DS Storage Manager software, a command is sent to the FDE drive to perform acryptographic erase. A cryptographic erase erases the existing data encryption key and then generates anew encryption key inside the drive, making it impossible to decrypt the data. After the encryption key ischanged, any data that was written to the disk that was encrypted with the previous encryption key isunintelligible. This includes all bits, headers, and directories.

After secure erase takes place on the drive, the following actions occur:v The data becomes completely and permanently inaccessible, and the drive returns to the original

factory state.v Drive security becomes disabled and must be re-enabled if it is required.

Before you initiate secure erase, the security-enabled FDE drive must be unlocked, and the array that it isassigned to must be deleted.

Attention: You must back up the data in the security enable FDE drives to other drives or to secure tapebefore you secure erase an FDE drive if you want to access the data at a later time. All data on the diskwill be permanently and irrevocably erased when the secure erase operation is completed for asecurity-enabled FDE drive. Do not perform this action unless you are sure that you want to erase thedata. The improper use of secure erase will result in lost data.1. Before the drives can be secure erased, you must delete the RAID array the drives are associated with

and return the drives to Unassigned status:a. Click the Logical/Physical tab in the Subsystem Management window.


b. Right-click the array name; then, click Delete.

c. When you are prompted to select the array that you want to delete, click the array name and clickDelete.

d. To confirm that you want to delete the array, enter yes in the field and click OK.


e. Wait for the array deletion process to be completed. When you receive the confirmation Processed1 of array(s) – Complete, click OK.

2. Return to the Logical/Physical tab in the Subsystem Management window.

3. Select the drive on which you want to perform a secure erase. You can select more than one drive tobe erased by holding down the Ctrl key. In the top menu bar, click Drive; then, click Secure Erase.


4. To confirm that you want to permanently erase all data on the disk, enter yes in the field and clickOK. These drives can now be repurposed or discarded.

Global hot-spare disk drivesIf a disk drive fails in the DS5000 storage subsystem, the controller uses redundant data to reconstruct thedata on the failed drive on a global hot-spare drive. The global hot-spare drive is automaticallysubstituted for the failed drive without intervention. When the failed drive is eventually replaced, thedata from the hot-spare drive is copied back to the replacement drive.

Hot-spare drives must meet the array hot-spare requirements. The following drive types are required forhot-spare drives when secure-capable arrays are configured. If a drive does fail, the DS Storage Managerautomatically determines which hot-spare drive to substitute according to the type of the failed drive.


v For an array that has secured FDE drives, the hot-spare drive should be an unsecured FDE drive of thesame or greater capacity. After the unsecured FDE hot-spare drive is used as a spare for a failed drivein the secured RAID array, it is Security Enabled.

v For an array that has FDE drives that are not secured, the hot-spare drive can be either an unsecuredFDE drive or a non-FDE drive.

Note: If an unsecured FDE hot-spare drive was used as a spare for a non-secured FDE array and thearray was secured after the data was copied back, the unsecured FDE hot-spare drive remains unsecured,exposing the data in the drive if it is removed from the storage subsystem.

An unconfigured secured FDE drive cannot be used as a global hot-spare drive. If a global hot spare is asecured FDE drive, it can be used as a spare drive only in secured arrays. If a global hot-spare drive is anunsecured FDE drive, it can be used as a spare drive in secured or unsecured arrays with FDE drives, oras a spare drive in arrays with non-FDE drives. You must secure erase the FDE drive to change it toUnsecured state before it can be used as a global hot-spare drive. The following error message isgenerated if you assign an unconfigured secured FDE drive as a global hot spare.

Return code: Error 2 - The operation cannot complete because either (1) the current state of acomponent does not allow the operation to be completed, (2) the operation has been disabled inNVSRAM (example, you are modifying media scan parameters when that option (offset 0x31, bit 5)is disabled), or (3) there is a problem with the storage subsystem. Please check your storagesubsystem and its various components for possible problems and then retry theoperation.Operation when error occurred: PROC_assignSpecificDrivesAsHotSpares

When a global hot-spare drive is used as a spare for a failed drive in a secure array, it becomes a secureFDE drive and remains secure as long as it is a spare in the secure array. After the failed drive in thesecure array is replaced and the data in the global hot-spare drive is copied back to the replaced drive,the global hot-spare drive is automatically reprovisioned by the controllers to become an unsecured FDEglobal hot-spare drive.

As a best practice in a mixed disk environment that includes non-security capable SATA drives,non-security-capable Fibre Channel drives, and FDE Fibre Channel drives (with security enabled or notenabled), use at least one type of global hot-spare drive (FDE Fibre Channel and a SATA drive) at thelargest capacity within the array. If a secure-capable FDE Fibre Channel and SATA hot-spare drive areincluded, all arrays are protected.

Follow the standard hot-spare drive configuration guidelines described in “Configuring hot-sparedevices” on page 4-4). Hot-spare configuration guidelines are the same for FDE drives.

Log filesThe DS Storage Manager major events log (MEL) includes messages that describe any security changesthat are made in the storage subsystem.

Frequently asked questions

Securing arraysv Can I change an unsecured array with FDE drives to a secured array?

– Yes, the steps to complete this process are described in “Securing a RAID array” on page 6-16. TheDS5000 Encryption feature must be enabled and the security key file and pass phrase alreadyestablished. See “Enabling the DS5000 full disk encryption feature” on page 6-14 for moreinformation.

v When I enable security on an array, will the data that was previously written to that array be lost orerased?


– No, unless you perform a secure erase on the array disk drives, this data remains intact.v Can I change a secured array with FDE drives to an unsecured array?

– No, this is not a supported option. After an unsecured array is changed to a secure array, it cannotbe changed back to an unsecured array without destroying the data in the security-enabled FDEdrives. Use VolumeCopy to copy the secure data to an unsecured array, or back up the data to asecured tape. If you volume copy the secure data to an unsecured array, you must physically securethe drives. Then you must delete the original array and secure erase the array drives. Create a newunsecured array with these drives and use VolumeCopy to copy the data back to the original drivesor restore the data from secure tape.

v If I have an array with FDE drives that is secured, can I create another array that uses these samedrives and not enable security? Does the storage subsystem have a control so that this does not occur?– No, these are not supported functions. Any logical drive that is part of an array must be secured,

because the drives on which it is stored are security enabled.v When a secure array is deleted, does disk security remain enabled?

– Yes. The only way to disable security is to perform a secure erase or reprovision the drives.v If I create a new array on a set of unassigned/unconfigured security-enabled FDE disks, will they

automatically become secure?– Yes.

Secure erasev With secure erase, what can I erase, an individual drive or an array?

– Secure erase is performed on an individual drive. You can not erase a secured drive that is part ofan array; you must first delete the array. After the array is deleted and the drives becomeunassigned, you can erase multiple disks in the same operation by holding the Ctrl key while youselect the drives that are to be secure erased.

v If I want to use only the secure erase feature, do I still have to set up a security key identifier and passphrase?– Yes. The full disk encryption feature must be enabled before you can use secure erase.

v After secure erase is performed on a drive, is security enabled or disabled on that drive?– The drive is returned to Secure Capable (unsecured) state after a secure erase. Security is disabled

on the drive.v If I inadvertently secure erase a drive, is it possible to recover the data in the drive?

– No. After a drive is secure erased, the data in the drive is not recoverable. You must recover the lostdata from a back up copy. Back up the data in secure drives before you secure erase the drives.

Security keys and pass phrasesv Can I get to the security keys through the DS Storage Manager or controller?

– No, the security key is obfuscated in the storage subsystem. Only an encrypted version of the keycan be exported to a security key file, using the save security key operation. The actual security keyis not available for viewing. Implement prudent security features for the storage subsystem. TheDS5000 Storage Manager forces a strong password, but administrator access should have stringentcontrols in place.

v If I lose a drive that is unlocked or security disabled, can that data be accessed even though the data isencrypted?– Yes. Because security has not been enabled on the drive, it remains unlocked, and the data is

accessible.v If my security key falls into the wrong hands, can I change it without losing my data?

– Yes, the drive can be re-keyed, using the procedure to change the security key.


Premium featuresv How do I ensure that my mirrored data is secure? What is a best practice for protecting data at the

remote site?– Secure your data with security-enabled FDE drives at both the primary and the secondary sites.

Also, you must ensure that the data is protected while it is being transferred between the primaryand secondary sites.

v Can I use VolumeCopy to copy a secured logical unit number to a unsecured one? If so, what preventssomeone from doing that first, and then stealing the unsecured copy?– Yes. To prevent someone from stealing the data by using this method, implement prudent security

features for the storage subsystem. The DS5000 Storage Manager forces a strong password, butadministrator access must also have stringent controls in place.

v Can FlashCopy and VolumeCopy data be secured?– Yes. For FlashCopy, the FlashCopy repository logical drive must be secured if the target FlashCopy

data is secured. The DS5000 storage manager enforces this rule. Similarly, if the source array of theVolumeCopy pair is secured, the target array of the VolumeCopy pair must also be secured.

Global hot-spare drivesv Can I use an unconfigured FDE drive as a global hot-spare drive?

– Yes, but only if the drive is unsecured (security not enabled). Check the status of the unconfiguredFDE drive. If the drive is Secure, it must be secure erased or reprovisioned before you can use it asa global hot-spare drive.

v If the hot-spare drive in a secured array is an unsecured FDE drive, does this drive automaticallybecome secured when a secured FDE drive fails and that data is written to the hot-spare drive?– Yes. When the failed drive is removed from the RAID group, a rebuild is automatically started to

the hot-spare drive. Security is enabled on the hot-spare drive before the rebuild is started. A rebuildcannot be started to a non-FDE drive for a secure array. After the failed drive in the secured array isreplaced and the data in the global hot-spare drive is copied back to the replaced drive, the globalhot-spare drive is automatically reprovisioned by the controllers to become an unsecured FDE globalhot-spare drive.

Boot supportv Is there a special process for booting from a security-enabled drive?

– No. The only requirement is that the storage subsystem must be running (which is required in anybooting process).

v Are FDE drives susceptible to cold boot attacks?– No. This issue applies more to the server side, because an individual can create a boot image to gain

access to the server. This does not apply to FDE drives. FDE drives do not use the type of memorythat is susceptible to a cold boot attack.

Locked and unlocked statesv When does a security-enabled drive go into Locked state?

– The drive becomes locked whenever the disk is powered down. When the FDE drive is turned offor disconnected, it automatically locks down the data on the disk.

Backup and recoveryv How can I ensure that my archived data is secure?

– Securing archived data is beyond the scope of this document. See the Storage Networking InterfaceAssociation (SNIA) recommendations for secure tape backup. See Appendix L, “FDE best practices,”on page L-1 for specific references.


Otherv Is DACstore information still written to the disk?

– Yes. However, if the drive is secured, it must be unlocked by the controller first before the DACstoreinformation can be read. In the rare event that the controller security key is corrupted or bothcontrollers are replaced, a security key file must be used to unlock the drive.

v Is data on the controllers cache secure with FDE and IBM Disk Encryption? If not, are there any bestpractices here?– No. This is a security issue of physical access to the hardware. The administrator must have physical

control and security of the storage subsystem itself.v If I have secure-capable disks but have not purchased the IBM Disk Encryption premium feature key,

can I still recognize secure-capable disks from the user interface?– Yes. This information is available from several windows in the DS Storage Manager interface.

v What about data classification?– See the SNIA best practices for more information about data classification. See Appendix L, “FDE

best practices,” on page L-1 for specific references.v Can I use both FDE and non-FDE drives if I do not secure the drives?

– Yes. However, using both FDE and non-FDE drives is not a cost-effective use of FDE drives. Anarray with both FDE and non-FDE drives cannot be converted into a secure array at a later time.

v Do FDE disk drives have lower usable capacity because the data is encrypted or because capacity isneeded for the encryption engine and keys?– No. There is no capacity difference between non-FDE and FDE disk drives (1 GB unencrypted = 1

GB encrypted).


Chapter 7. Configuring and using Support Monitor

The IBM DS Storage Manager Profiler Support Monitor tool is a component of the IBM DS StorageManager version 10.60.x5.17 and later. In addition to the DS Storage Manager Profiler Support Monitorcode, the Apache Tomcat web server and MySQL database software packages are installed as part of thetool.

If you call IBM Support with a critical-event problem, the IBM DS Storage Manager Profiler SupportMonitor tool ensures that IBM Support can get the information that they need about the state of aDS3000, DS4000, or DS5000 storage subsystem prior to the storage subsystem critical event.

The IBM DS Storage Manager Profiler Support Monitor tool performs the following functions:v Automatically installs as part of the IBM DS Storage Manager installation.v Automatically collects the support bundle through the computer TCP connection. The default is to

collect the support bundle daily at 2 a.m. The support data bundle is a compressed file of the followingitems:– Collect all support data (CASD) bundle– Storage subsystem configuration file– SOC counts– RLS counts– ESM state capture

v Automatically manages the collected support bundles. It saves only the last five collected support databundles and deletes older support data bundles.

v Provides a Web-based interface for selecting the appropriate support data bundle to send to IBMSupport.

Note: No user configuration or interaction is required unless the customer wants to change the defaultoperating behavior of the Support Monitor tool.

Use the information in this chapter to configure and use the DS Storage Manager Support Monitor tool.See Chapter 3, “Installing Storage Manager and Support Monitor,” on page 3-1 for information aboutinstalling and the Support Monitor tool.

This chapter contains the following topics:v “Overview of the Support Monitor interface”v “Scheduling collection of the support bundle” on page 7-3v “Sending the support bundle to IBM Support” on page 7-3v “Collecting the support bundle manually” on page 7-5v “Using the Support Monitor log window” on page 7-6v “Solving Support Monitor problems” on page 7-8

Overview of the Support Monitor interface

The Support Monitor Web interface is described in the following sections. The Web interface consists ofthe following two components:v Console area on the right side of the screenv Icons are used in the Support Monitor Web interface


Note: The images in this section might differ slightly from what you see on your screen.

Console area

The console area of the Support Monitor interface shows the primary content for the function or item youselect in the navigation tree.

Icons

The meanings of the Support Monitor icons are described in the following table.

Table 7-1. Support Monitor icon meanings

Icon Icon meaning and function

The last support data collection was successful, or theresource is operational.

The last support data collection was unsuccessful, or theresource is not operational.

The support data for this resource has not yet beencollected.

Figure 7-1. Console area


Table 7-1. Support Monitor icon meanings (continued)

Icon Icon meaning and function

Click to schedule support data collection.

Click to send the support data.

Click this icon to initiate support data collection for astorage subsystem immediately.

The software cannot collect support data for thisresource.

Scheduling collection of the support bundle

Use the information in this section to change the time and frequency of the support bundle collectionschedule.

You can set the support bundle collection frequency to never, daily, weekly, or monthly. You also canspecify the time of day that the support bundle is collected. The support bundle should be collected onlywhen the storage subsystems are not undergoing heavy usage or performing critical tasks.

If there are multiple storage subsystems in the monitored list, the collection schedule should be modifiedto stagger the support bundle collection events among the monitored storage subsystems. Limit thenumber of simultaneous support bundle collections to a maximum of three storage subsystems.Depending on the complexity of the subsystem configuration, the controller workloads, and the Ethernetnetwork activities during the time of the support bundle collection, and the size of the captured logs, itmight take 30 minutes or more to collect the support bundle for a given storage subsystem. Duringoperating conditions when the hosts have light I/O workloads, the time required to collect the supportbundle data for a subsystem is between 5 and 10 minutes.

To save disk space, the Support Monitor will keep only the last five support bundles that are collected fora given storage subsystem. This value cannot be changed. To prevent the support bundles from beingdeleted, they must be copied to a different directory from the support bundle saved directory. If the DSStorage Manger software was installed using the default directory, this directory is C:\Program Files...\IBM_DS\IBMStorageManagerProfiler Server\support in Windows operating-system environments. InUnix-type operating-system environments, this directory is /opt/IBM_DS/IBMStorageManagerProfiler_Server/support.

To configure the support bundle collection schedule, complete the following steps:

1. Click the Calendar icon ( ) for the storage subsystem schedule you want to change. The ScheduleSupport Data Collection window is displayed.

2. Click the radio button for the applicable collection frequency.3. Select the time setting for the collection.4. Click Save to save the schedule settings.

Sending the support bundle to IBM Support

If you experience storage subsystem problems, IBM might request to have the collected support bundlessent to IBM Support for troubleshooting.

To send one or more support bundles to IBM Support, complete the following steps:

Chapter 7. Configuring and using Support Monitor 7-3

1. Click the Envelope icon ( ) for the storage subsystem support bundle you want to send. The SendSupport Data screen, similar to the following illustration, is displayed.

Note: If the e-mail server information has not been entered in the Server tab in the Administrationmenu of the Support Monitor interface, the Server Setup window is displayed, similar to the one inthe following illustration. Enter the settings for the e-mail server and user e-mail address in theE-mail Server and E-mail From fields, and click Save.


2. On the Send Support Data screen, select one or more SOC and RLS change log files, as requested byIBM Support, from the Available box and click the right arrow button to move the selected supportbundle to the Selected box. If you click the double right arrow button, you move all of the availablesupport bundles to the Selected box. If necessary, use the left arrow or double left arrow buttons tomove one or all of the support bundles in the Selected box back to the Available box.

3. Enter the Problem Management Record (PMR) number and the e-mail address that were provided toyou by the IBM Hardware and Software Support representative. To find the phone number for yourcountry, go to http://www.ibm.com/planetwide/ and click the name of your country.

4. Type your e-mail address in the cc: field so that a copy of the same e-mail that is sent to IBM Supportis also sent to you for your records.

5. In the Details box, type any information that might help IBM Support identify the source of thee-mail. For example, type the PMR number, company information, contact name, telephone number,and a one-line summary of the problem.

6. Click the Send button to send the support data bundle to IBM Support.

Collecting the support bundle manually

Support Monitor enables you to manually collect a support bundle for a storage subsystem.

To collect a support bundle manually for a storage subsystem, click the Life Preserver icon ( ) for thestorage subsystem. A message opens that states that the support bundle will be collected in thebackground. The Last Collection column is automatically updated when the operation completes.


http://www.ibm.com/planetwide/

Using the Support Monitor log window

To view the running log of the DS Storage Manager Profiler Support Monitor log, click the View buttonon the upper right corner of the Support Monitor window. A View Module Log File window with thename mod.sys.support.Support is displayed in the Existing Modules field. In this window, you canspecify the number of log entries to display (up to 1,000 lines) and the frequency of the data refresh.

See the following table for more information about the messages that you might see in the SupportMonitor log window.

Table 7-2. Support Monitor messages and descriptions

Message type Message text and description

Support Monitor module online intializing <num> DeviceClients

This message shows the number of storage subsystemsbeing monitored plus one more for Support Monitor.

DeviceClient created:deviceType--><type>deviceIdent--><id>status--><status>

After the client is created, this variable logs informationabout each storage subsystem.

attempting to start <num> DeviceClients

This message shows that each device client was startedand initialized using the initializing DeviceClientscommand.

not starting DeviceClient (<deviceClient name>)sincestatus is set to <status>

This message shows that when the status is anythingother than online, the client does not start.

Registration

This message appears when a storage subsystem monitorregistration key is created for Support Monitor. Themodule status is set to online, and the registration key iscreated for the Support Monitor device to register withthe Storage Manager Profiler server.

Support Monitor module offline stopping <num> DeviceClients

This message appears when the configuration file isupdated with new storage subsystem information, andthe module is temporarily placed offline. The modulethen returns to online status to refresh the information.

<id> supportinfo - stopping ClientProxy

This message shows that a specific client is stopped.

Discovery Discovery (<id>)

This message appears when the device id is assignedfrom the DS Storage Manager Profiler server.


Table 7-2. Support Monitor messages and descriptions (continued)


General discovery messages discovery(<id>): discovering arrays/smtp on <time>sec intervals

This message shows that the discovery data isestablished on a scheduled frequency.

discovery(<id>): discovering arrays/smtp fromon-demand request

This message shows that the discovery data isestablished through a user-initiated action.

discovery(<id>): discovering process completed in<time> secs

This message indicates that the discovery process iscomplete.

Storage subsystem discovery Storage array discovery discovery(<id>): new arraydiscovered-->Name: <arrayName>, IP 1: <ip ofcontroller 1>, IP 2: <ip of controller 2>

This message shows that the storage subsystem is addedto the Enterprise Management Window of the StorageManager Client program.

discovery(<id>): no new arrays discovered

This message appears when the discovery is initiated butno new storage subsystems are found.

discovery(<id>): unmanaged array detected-->Name:<arrayName>, IP 1: <ip of controller 1>, IP 2: <ipof controller 2>

This message shows that the storage array is removedfrom the Enterprise Management Window of the StorageManager Client program.

discovery(<id>): no unmanaged arrays detected

This message appears when the discovery is initiated,but no storage subsystems are removed from theEnterprise Management Window of the Storage ManagerClient program.

SMTP discovery discovery(<id>): discovered smtp server info (<smtpserver>) and email from info (<email from>)

This message shows that the SMTP server informationand the e-mail address is parsed from the StorageManager Client program.


Table 7-2. Support Monitor messages and descriptions (continued)


Support bundle collection retry related message <array name> stopping periodic support capture,since previous <num> attempts have failed

If the scheduled support bundle collection failed “num"times for storage subsystem "array name", the SupportMonitor will stop attempting to collect the supportbundle collection for that storage subsystem.

<array name> retrying support capture since lastattempt failed. Retry attempt <num> of <num>

This message appears when a scheduled capture fails forstorage subsystem "array name" after retrying “num”times.

Scheduled support bundle collection message <array name> started periodic support data Capture

This message appears when a scheduled data collectionis started.

On-demand support bundle collection message <array name> started on-demand support data Capture

This message appears when a user-initiated datacollection is started.

Solving Support Monitor problems

This section contains information to help you solve some of the problems you might have with yoursoftware. Table 7-3 contains problem descriptions, possible problem causes, and suggested actions. Usethis information, in addition to the DS Storage Manager Recovery Guru in the Subsystem Managementwindow, to solve problems with your software.

Always use the DS Storage Manager client to diagnose storage subsystem problems and componentfailures and find solutions to problems that have definite symptoms.

Table 7-3. Problem index

Problem Possible cause Possible solution

Data is not being collectedon monitored subsystems

There is a problem with theStorage Manager client TCPconnection. The operation ofStorage Monitor is dependenton the Storage Manager clientTCP connection.

Make sure that Storage Manager is able to access datafrom the storage subsystem.

A Storage Manager clientsession is not running. AStorage Manager client sessionmust be active and runningwhen you use Storage Monitor.

Make sure that the client is active and running.

A user has disabled datacollection for the storagesubsystem.

Open the Support Monitor console and make surethat support data collection was not disabled for thestorage subsystem in question.


Table 7-3. Problem index (continued)


A networked storagesubsystem is not in the listof monitored storagesubsystems.

The missing storage subsystemhas not been detected by thesoftware.

Re-scan for devices in the Storage Monitor window.

Storage Monitor has not beenconfigured with unique namesfor each storage subsystem.

Make sure that all storage subsystems have uniquenames in the Storage Manager EnterpriseManagement window.

There are too many storagesubsystems defined in theStorage Manager EnterpriseManagement window.

The data collection process of the Storage Monitor ismulti-threaded, with a polling mechanism in place tofind the maximum number of storage subsystems atpre-defined timing intervals. The polling mechanismis not sequential.

For example, if the upper limit of storage subsystemsfrom which the Storage Monitor can find data is 20,and 60 storage subsystems are defined, 20 threads aregathered immediately while the data from theremaining 40 storage subsystems are gathered only asresources become available.

The application will notstart.

A user stopped one or moreservices manually.

Make sure that all of the required services started.

In a Windows operating system, click AdministrativeTools > Computer Management > Services(Start/Stop), and make sure the following serviceswere started:

v ProfilerCollector

v ProfilerMaintenance

v ProfilerEventReceiver

v ProfilerPoller

v ProfilerWebserver (Tomcat Apache)

v MySQL

In a Unix operating system, execute the command/etc/init.d/profiler start to start the application,or the command /etc/init.d/profiler stop to stopthe application.

E-mail notifications are notworking correctly.

E-mail notifications are notconfigured correctly.

Make sure that e-mail notifications meet the followingconditions:

v SMTP server is set up correctly

v SMTP server is operational

v The connection from Storage Monitor server to theSMTP server is operational

Support Monitor cannot beinstalled.

An existing MySQL databaseor Apache Web server softwarewas not removed before theinstallation of DS StorageManager and Support Monitor,or there is not enough space onthe hard drive to install the DSStorage Manager and SupportMonitor.

Review the installation log for the possible causes oferrors and correct them as required.


Table 7-3. Problem index (continued)


The Support Monitorconsole is not responding

There are either networkproblems or the IP address ofthe management station waschanged.

Complete the following steps:

v Check for network problems.

v Check the current IP address of the managementstation on which the Support Monitor is installed. Ifit is different than the IP address that wasestablished when the Support Monitor was installedand configured, you must either change the IP backto the one that was configured initially or removeand reinstall the Support Monitor software.


Appendix A. Using the IBM System Storage DS3000, DS4000,and DS5000 Controller Firmware Upgrade Tool

Tool overviewThe following information describes how to use the IBM System Storage DS3000, DS4000, or DS5000Controller Firmware Upgrade Tool to upgrade your DS4800, DS4700 or DS4200 Express® controllerfirmware from 06.xx to 07.xx.

Note: With version 10.50 of Storage Manager, the Controller Firmware Upgrade Tool has become part ofthe Enterprise Management window (EMW) and is no longer a separate tool.

You are now required to download the firmware files from the Web and place them in a directory, whichthe upgrade tool can browse. Previous versions of the upgrade tool had the firmware files included.

Caution:

Before using the IBM System Storage DS3000, DS4000, or DS5000 Controller Firmware Upgrade Tool, it isimportant that all data for DS4800, DS4700, and DS4200s be completely backed up and that existingsystem configurations be saved. Once the tool has completed an upgrade, controllers cannot be returnedto previous firmware version levels. The Controller Firmware Upgrade Tool is to be used only whenmigrating DS4800, DS4700, and DS4200 controllers from version 06.xx to version 07.xx. This tool is notintended to replace, nor should be used to perform, standard upgrades for controller, ESM, or drivefirmware. (To perform a standard controller, ESM, or drive firmware upgrade, please see “Downloadingcontroller firmware, NVSRAM, ESM firmware” on page 3-15.)

You must perform the upgrade offline. You should perform the overall installation of Storage Managerinto an existing host environment online. For most failover drivers to take effect, you must reboot thehost.

You must make sure that all devices have an Optimal status before you download firmware. You can usethe Healthcheck utility to assist with this. You must also check the current firmware level.

Attention: Potential loss of data access—Make sure the firmware you download is compatible with theStorage Manager software that is installed on your storage system. If non-compatible firmware isdownloaded, you might lose access to the drives in the storage system, so upgrade Storage Manager first.

Do not make changes to your configuration or remove drives or enclosures during the upgrade process.

For information about the current firmware versions, see “Finding Storage Manager software, controllerfirmware, and README files” on page xiii to find out how to access the most recent Storage ManagerREADME files on the Web.

Checking the device health conditionsPerform the following steps to determine the health condition of your device:1. From the Array Management Window in Storage Manager, right-click the storage system. The Storage

Manager software establishes communication with each managed device and determines the currentdevice status.There are six possible status conditions:v Optimal—Every component in the managed device is in the desired working condition.

© Copyright IBM Corp. 2009, 2010 A-1

v Needs Attention—A problem exists with the managed device that requires intervention to correctit.

v Fixing—A Needs Attention condition has been corrected, and the managed device is currentlychanging to an Optimal status.

v Unresponsive—The storage management station cannot communicate with the device, or onecontroller or both controllers in the storage system.

v Contacting Device—The management software is establishing contact with the device.v Needs Upgrade—The storage system is running a level of firmware that is no longer supported by

the storage management software.2. If the status is a Needs Attention status, write down the condition. Contact an IBM Technical Support

representative for fault resolution.

Note: The Recovery Guru in the Storage Manager software also provides a detailed explanation of,and recovery procedures for, the conditions.

Using the upgrade toolFrom the Enterprise Management Window (EMW) toolbar, select Tools. Then select Firmware Upgrade.See 2-2. When the Firmware Upgrade Window appears, any system listed in the EMW will also appearhere. The Firmware Upgrade Tool will also automatically perform it's own diagnostic check on thesesystems to determine if they are healthy enough to perform a controller firmware upgrade.

Note: Please be aware of the following recommendations:v For any condition other than Optimal, IBM recommends that you call the Support Line for assistance.

See “Software service and support” on page xv for additional information.v You can perform only a major release to major release (06.xx. to 07.xx) upgrade, using this tool. DO

NOT attempt to perform this type of firmware upgrade using the Subsystem Management Window(SMW).

v Once you are at the 07.xx firmware level, you do not need to use the firmware upgrade tool. Anyfuture firmware upgrades can be performed by using the SMW.

Select the Help button within the Firmware Upgrade Tool, for additional instructions regarding thefirmware upgrade.

Adding a storage subsystemTo add a storage subsystem using the upgrade tool, perform the following steps:1. Click Add. The Select Addition Method window appears.2. Select either Automatic or Manual.3. Click OK to begin adding storage subsystems.4. To see any issues with the storage system you added that might impede upgrading the firmware,

click View Log.

Downloading the firmware1. Select the storage subsystem you want to activate to enable the download button. From the Enterprise

Management Window tool bar, select Tools, then Upgrade Firmware. The Download Firmwarewindow appears.

2. Select the controller firmware file that you want to download. Click Browse to choose the file from adirectory on your computer or on your network.

3. Select the NVSRAM file you want to download. Click Browse to choose the file from a directory onyour computer or on your network.

A-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

4. Click OK. The firmware starts to download. A status bar appears in the Controller Firmware Upgradewindow.

Viewing the IBM System Storage DS3000, DS4000, and DS5000Controller Firmware Upgrade Tool log fileIf you encounter any problems updating your firmware, perform the following steps to view the log file:1. Click View Log. The View Log window appears. This log documents any issues with the storage

system that might prevent you from updating the firmware.2. If any issues are documented in the log, correct those issues before you try to download the firmware.

Appendix A. Using the IBM System Storage DS3000, DS4000, and DS5000 Controller Firmware Upgrade Tool A-3

A-4 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Appendix B. Host bus adapter settings

This section covers the default settings for a variety of host bus adapters (HBAs) suitable for use withDS3000, DS4000, and DS5000 Storage Subsystems for Windows, Linux on Intel, VMware ESX, andNetWare operating systems. All other operating systems and platforms should use the default values, orthose recommended by their respective product documentation.

Please refer to the README file that is included in the Fibre Channel host bus adapter BIOS or devicedriver package for any up-to-date changes to the settings.

An HBA is used to connect servers to Fibre Channel topologies. Its function is similar to that provided bynetwork adapters to access LAN resources. The device driver for an HBA is typically responsible forproviding support for a Fibre Channel topology, whether point-to-point, loop, or fabric.

The Fast!UTIL feature enables you to access and modify an adapter's default settings to optimize itsperformance.

See also: For detailed HBA support information, see www.ibm.com/systems/support/storage/config/ssic.

Setting host bus adaptersIt is often necessary to adjust the settings of your HBA to match the capabilities of your device. Thissection describes how to access those settings to make the necessary adjustments.

Accessing HBA settings through Fast!UTILThe Fast!UTIL feature provides access to host bus adapter settings. To access this feature, simultaneouslypress and hold the Alt and Q keys or the Ctrl and Q keys during BIOS initialization. It may take a fewseconds for the Fast!UTIL menu to appear. If more than one board is installed, Fast!UTIL prompts you toselect a board to configure. After you change adapter settings, Fast!UTIL reboots your system to load thenew parameters. Upon entering Fast!UTIL, the following selections are available on the Fast!UTILOptions menu:v Configuration Settingsv Loopback Testv Select Host Adapter

You can access the host bus adapter settings through the Configuration Settings menu in Fast!UTIL.

Note: Alternatively, you can also use the QLogic SANsurfer program to modify the Host adaptersettings and Advanced adapter settings preferences from the Microsoft Windows operating systemenvironment. You must reboot the servers for the changes to become effective.

Accessing host bus adapter settingsAccess the host bus adapter settings through the Configuration Settings menu in Fast!UTIL and selectAdapter Settings. The default host bus adapter settings for the FC2-133 HBA are as follows:

Host Adapter BIOSWhen this setting is Disabled, the ROM BIOS on the FC2-133 HBA is Disabled, freeing space inupper memory. This setting must be Enabled if you are booting from an Fibre Channel disk driveattached to the FC2-133 board. The default is Disabled.

© Copyright IBM Corp. 2009, 2010 B-1



Frame SizeThis setting specifies the maximum frame length supported by the FC2-133 HBA. The default sizeis 2048, which provides maximum performance for F-Port (point-to-point) connections.

Loop Reset DelayAfter resetting the loop, the firmware refrains from initiating any loop activity for the number ofseconds specified in this setting. The default is 5 seconds.

Adapter Hard Loop IDThis setting forces the adapter to attempt to use the ID specified in the Hard Loop ID setting. Thedefault is Enabled.

Hard Loop IDIf the Adapter Hard Loop ID setting is Enabled, the adapter attempts to use the ID specified inthis setting. The default ID is 125. It is recommended to set this ID to a unique value from 0-125if there are more than one adapters connected to a FC-AL loop and the Adapter Hard Loop IDsetting is Enabled.

Spin Up DelayWhen this bit is set, the BIOS will wait up to five minutes to find the first drive. The defaultsetting is Disabled.

Connection OptionsThis setting defines the type of connection (loop or point to point) or connection preference. Thedefault is 2, which is loop preferred unless point-to-point.

Fibre Channel Tape SupportThis setting enables FCP-2 recovery. The default is Enabled. It is recommended to change thissetting to Disabled if the HBA is not connected to a tape device.

Data RateThis setting determines the data rate. When this setting is 0, the FC2-133 HBA runs at 1 Gbps.When this setting is 1, the FC2-133 HBA runs at 2 Gbps. When this setting is 2, Fast!UTILdetermines what rate your system can accommodate and sets the rate accordingly. The default is2 (auto-configure).

Advanced Adapter SettingsAccess the following advanced host bus adapter settings through the Configuration Settings menu inFast!UTIL and select Advanced Adapter Settings. The default settings for the FC2-133 HBA are asfollows:

Execution ThrottleThis setting specifies the maximum number of commands executing on any one port. When aport's execution throttle is reached, no new commands are executed until the current commandfinishes executing. The valid options for this setting are 1-256. The default is 255.

LUNs per TargetThis setting specifies the number of LUNs per target. Multiple LUN support is typically forredundant array of independent disks (RAID) boxes that use LUNs to map drives. The default is0. For host operating systems other than Microsoft Windows, one may need to change this settingto a value other 0 to allow the host seeing more than one logical drive from the DS3000, DS4000,or DS5000 Storage Subsystem.

Enable LIP ResetThis setting determines the type of loop initialization process (LIP) reset that is used when theoperating system initiates a bus reset routine. When this setting is yes, the driver initiates a globalLIP reset to clear the target device reservations. When this setting is no, the driver initiates aglobal LIP reset with full login. The default is No.

Enable LIP Full LoginThis setting instructs the ISP chip to log in, again, to all ports after any LIP. The default is Yes.

B-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Enable Target ResetThis setting enables the drivers to issue a Target Reset command to all devices on the loop whena SCSI Bus Reset command is issued. The default is Yes.

Login Retry CountThis setting specifies the number of times the software tries to log in to a device. The default is30 retries.

Port Down Retry CountThis setting specifies the number of seconds the software retries a command to a port returningport down status. The default is 30 seconds. For the Microsoft Windows servers in MSCSconfiguration, the Port Down Retry Count BIOS parameter must be changed from the default of30 to 70.

Link Down TimeoutThis setting specifies the number of seconds the software waits for a link down to come up. Thedefault is 60 seconds.

Extended Error LoggingThis setting provides additional error and debug information to the operating system. Whenenabled, events are logged into the Windows NT® Event Viewer. The default is Disabled.

RIO Operation ModeThis setting specifies the reduced interrupt operation (RIO) modes, if supported by the softwaredriver. RIO modes allow posting multiple command completions in a single interrupt. Thedefault is 0.

Interrupt Delay TimerThis setting contains the value (in 100-microsecond increments) used by a timer to set the waittime between accessing (DMA) a set of handles and generating an interrupt. The default is 0.

QLogic host bus adapter settings

Note: The BIOS settings under the Windows column are the default values that are set when theadapters are ordered from IBM as IBM FC-2 (QLA2310), FC2-133 (QLA2340) and System Storage DS3000,DS4000, or DS5000 single-port and dual-port 4 Gbps (QLx2460 and QLx2462) FC host bus adapters. If theadapters are not from IBM, the default BIOS may not be the same as the ones defined in the MicrosoftWindows column. There is one exception, the default setting for Fibre Channel tape support is enabled.

Table B-1 covers the default settings for IBM fibre channel FC-2 and FC2-133 (QLogic adapter modelsQLA2310 and QLA2340) host bus adapter settings (for BIOS V1.35 and later) by operating system as wellas the default registry settings for Microsoft Windows operating systems. DS3000, DS4000, or DS5000products require BIOS V1.43 or later for these adapters. In addition, these settings are also the defaultBIOS settings for the newer DS3000, DS4000, or DS5000 4 Gbps single and dual-port host bus adapters(QLogic adapter models QLx2460 and QLx2462). The 4 Gbps host bus adapters adapter BIOS version is1.12 or later. See the appropriate README file for the latest updates to these values.

Table B-1. Qlogic model QLA234x, QLA24xx, QLE2462, QLE2460, QLE2560, QLE2562

Item Default WMWare W2KW2K3/W2K8 Solaris

LINUXMPP

LINUXDMMP Netware

BIOS settings

Host Adapter settings

Host Adapter BIOS Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Frame Size 2048 2048 2048 2048 2048 2048 2048 2048

Loop Reset Delay 5 5 8 8 8 8 8 8

Appendix B. Host bus adapter settings B-3

Table B-1. Qlogic model QLA234x, QLA24xx, QLE2462, QLE2460, QLE2560, QLE2562 (continued)


LINUXMPP

LINUXDMMP Netware

Adapter Hard Loop ID– (only recommendedfor arbitrated looptopology).

Disabled Enabled Enabled Enabled Enabled Enabled Enabled Enabled

Hard Loop ID (shouldbe unique for eachHBA) – (onlyrecommended forarbitrated looptopology).

0 1251 1251 1251 1251 1251 1251 1251

Spin-up Delay Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Connect Options 2 2 2 2 2 2 2 2

Fibre Channel TapeSupport

Disabled Disabled3 Disabled3 Disabled3 Disabled3 Disabled3 Disabled3 Disabled3

Data Rate 2 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto)

Advance Adapter Settings

Execution Throttle 16 256 256 256 256 256 256 256

LUNs per Target 8 0 0 0 0 0 0 32

Enable LIP Reset No No No No No No No No

Enable LIP Full Login Yes Yes Yes Yes Yes Yes Yes Yes

Enable Target Reset Yes Yes Yes Yes Yes Yes Yes Yes

Login Retry Count 8 30 30 30 30 30 30 30

Port Down RetryCount (5.30 controllerfirmware and earlier)

8 30 30 30 30 12 12 70

Port Down RetryCount

8 70 DS3K:144

DS4K/5K: 702

DS3K: 144DS4K/5K:

702

70 DS3K: 70DS4K5K:

35

10 70

Link Down Timeout 30 60 DS3K:144DS4K/5K: 60

DS3K:144DS4K/5K:

60

60 DS3K:144DS4K/5K: 60

NA 60

Extended ErrorLogging

Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

RIO Operation Mode 0 0 0 0 0 0 0 0

Interrupt Delay Timer 0 0 0 0 0 0 0 0

IOCB Allocation 256 256 256 256 256 256 256 256

>4 GB Addressing Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Drivers Load RISCCode

Enabled Enabled Enabled Enabled Enabled Enabled Enabled Enabled

Enable DatabaseUpdates

No No No No No No No No

Disable Database Load No No No No No No No No

Fast Command Posting Disabled Enabled Enabled Enabled Enabled Enabled Enabled Enabled

Extended Firmware Settings (1.34 and Earlier)




LINUXMPP

LINUXDMMP Netware

Extended ControlBlock

Enabled Enabled Enabled Enabled Enabled Enabled Enabled Enabled

RIO Operation Mode 0 0 0 0 0 0 0 0

Connection Options 2 2 2 2 2 2 2 2

Class 2 Service Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

ACK0 Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Fibre Channel TapeSupport

Enabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Fibre Channel Confirm Enabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Command ReferenceNumber

Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Read Transfer Ready Disabled Disabled Disabled Disabled Disabled Disabled Disabled Disabled

Response Timer 0 0 0 0 0 0 0 0

Interrupt Delay Timer 0 0 0 0 0 0 0 0

Data Rate 2 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto) 2 (Auto)

REGISTRY SETTINGS5

(HKEY_LOCAL_MACHINE→System→CurrentControlSet→Services→QL2300→Parameters→Device)

LargeLuns 1 1 N/A N/A N/A N/A N/A N/A

MaximumSGList 0x21 0xff 0xff 0xff N/A N/A N/A N/A

O/S REGISTRY SETTINGS5

(HKEY_LOCAL_MACHINE→System→CurrentControlSet→Services→QL2300→Parameters→Device) underDriverParameter variable (note: DriverParameter is of type REG_SZ and the following parameters are added tothe DriverParameters string. Do not create a separate key for each of the parameters.)Note: Prior to QLogic driver versions 9.1.x.x, the variable name used was DriverParameters instead ofDriverParameter.

UseSameNN 1 1 1 1 N/A N/A N/A N/A

BusChange (SCSIPortminiport 9.0.1.60 andearlier – does not applyto 9.1.1.11 and newer)

2 N/A 0 0 N/A N/A N/A N/A

TimeOutValue 4(REG_DWORD)

0x3C 0x78 DS3K:xA0

DS4K/5K: x78

DS3K:xA0

DS4K/5K:x78

N/A N/A N/A N/A

REGISTRY SETTINGS5

(HKEY_LOCAL_MACHINE→SYSTEM→CurrentControlSet→Services→<FAILOVER>→parameters: Where<FAILOVER>=Rdacdisk for MPPor RDAC installations or <FAILOVER>=mppdsm, ds4dsm, md3dsm, sx3dsm,csmdsm, or tpsdsm for MPIO installations. Mppdsm is for the generic version, your instllation could bedifferent.)

SynchTimeOut(REG_DWORD)

0x78 N/A DS3K:xA0

DS4K/5K: x78

DS3K:xA0

DS4K/5K:x78

DisableLunRebalance(Only applies tocluster configuration.Yuma 1.0 and later.

0x00 N/A 0x03 0x03




LINUXMPP

LINUXDMMP Netware

SuSE 7.3 specific modifications:

v Offset 0x11 in the Linux region (6) of the array controller’s NVSRAM must be changed from the default of 0x20 to0x7f. The following can be run from the script engine.

– Set controller[a] HOSTNVSRAMByte[6,0x11]=0x7f;

– Set controller[b] HOSTNVSRAMByte[6,0x11]=0x7f;

v The Qlogic driver source must be modified to reflect the symbolic link used by SuSE.

– vi makefile

– find OSVER and change it from OSVER=linux-2.4 to OSVER=linux

– Save and quit

Red Hat Linux Advanced Server 2.1 / SuSE Linux Enterprise Server 8.0 (6.x series failover driver [with no RDAC]only). The following should be appended to the HBA driver’s option string in /etc/modules.conf: ql2xretrycount=60ql2xsuspendcount=40

If you are running with the QLogic Inbox driver the string “options qla2xxx qlport_down_retry=144”(PB1-3) or“options qla2xxx qlport_down_retry=70”(PB4-6) should be added in /etc/modprobe.conf (for RHEL) or/etc/modprobe.conf.local (for SLES). For all prior (RH3/4 SLES8/9) Linux versions (and out-of- box drivers) thestring “options qla2xxx qlport_down_retry=72”(PB1-3) or “options qla2xxx qlport_down_retry=35”(PB4-6) should beadded instead.

Note:

1. This setting must be changed to a unique AL-PA value if there is more than one FC device in the FC-AL loop.

2. For larger configurations with heavy I/O loads or in Microsoft cluster service (MSCS) environment, this valuemay be increased..

3. This setting should be changed to enable or supported when the HBA is connected to a tape device only. It isrecommended to set it to Disabled when connecting to a DS3000, DS4000, or DS5000 Storage Subsystem.

4. In certain storage subsystem maximum configuration installations, it may be required to set the TimeOutValue to120 (decimal). Changing this value to a higher value might affect your application especially when it requires thedisk I/O completion acknowledgement within a certain amount of time.

5. Registry settings can be accessed by clicking Start, select Run..., type regedit into the Open: field, and then clickOK.

Attention: Exercise caution when changing the Windows registry. Changing the wrong registry entry or makingan incorrect entry for a setting can introduce an error that prevents your server from booting up or operatingcorrectly.

Note: The BIOS settings under the Windows column are the default values that are set when theadapters are ordered from IBM as IBM FAStT (QLA2200) FC host bus adapters. If the adapters are notfrom IBM, the default BIOS may not be the same as the ones defined in the Microsoft Windows column.There is one exception, the default setting for Fibre Channel tape support is enabled.

Table B-2 covers the default settings for various IBM DS3000, DS4000, or DS5000 FC host bus adapters(QLogic adapter QL220x) models (for BIOS V1.81) by operating system. See the appropriate README filefor the latest updates to these values.

Table B-2. QLogic model QL220x (for BIOS V1.81) host bus adapter settings by operating system

Item Windows Linux NetWare

NT 2000 / Server 2003

BIOS settings

Host Adapter settings


Table B-2. QLogic model QL220x (for BIOS V1.81) host bus adapter settings by operating system (continued)

Host Adapter BIOS Disabled Disabled Disabled Disabled

Frame Size 2048 2048 2048 2048

Loop Reset Delay 5 5 8 5

Adapter Hard Loop ID Enabled Enabled Enabled Enabled

Hard Loop ID (should beunique for each HBA)

1251 1251 1251 1251

Spin Up Delay Disabled Disabled Disabled Disabled

Advanced adapter settings

Execution Throttle 256 256 256 256

>4 Gbyte Addressing Disabled Disabled Disabled Disabled

LUNs per Target 0 0 0 32

Enable LIP Reset No No No No

Enable LIP Full Login Yes Yes Yes Yes

Enable Target Reset Yes Yes Yes Yes

Login Retry Count 30 30 30 30

Port Down Retry Count 30 30 12 302

IOCB Allocation 256 256 256 256

Extended Error Logging Disabled Disabled Disabled Disabled

Extended Firmware Settings

Extended Control Block Enabled Enabled Enabled Enabled

RIO Operation Mode 0 0 0 0

Connection Options 3 3 3 3

Class 2 Service Disabled Disabled Disabled Disabled

ACK0 Disabled Disabled Disabled Disabled

Fibre Channel Tape Support Supported3 Supported3 Supported3 Supported3

Fibre Channel Confirm Disabled Disabled Disabled Disabled

Command Reference Number Disabled Disabled Disabled Disabled

Read Transfer Ready Disabled Disabled Disabled Disabled

Response Timer 0 0 0 0

Interrupt Delay Time 0 0 0 0

Registry settings4 (HKEY_LOCAL_MACHINE → System → CurrentControlSet → Services → QL2200 → Parameters →Device)

LargeLuns 1

MaximumSGList 0x21 0x21

Registry settings4 (HKEY_LOCAL_MACHINE → System → CurrentControlSet → Services → Disk)

TimeOutValue4

(REG_DWORD)0x3C 0x3C

Registry settings4 (HKEY_LOCAL_MACHINE → System → CurrentControlSet → Services → QL2200 → Parameters →Device) under the DriverParameter variable

BusChange 0


Table B-2. QLogic model QL220x (for BIOS V1.81) host bus adapter settings by operating system (continued)

Note:

1. This setting must be changed to a unique AL-PA value if there is more than one FC device in the FC-AL loop.

2. For larger configurations with heavy I/O loads, it is recommended to change this value to 70.

3. This setting should be changed to enable or supported when the HBA is connected to a tape device only. It isrecommended to set it to Disabled when connecting to DS3000, DS4000, or DS5000 Storage Subsystem.

4. Registry settings can be accessed by clicking Start, select Run..., type regedit into the Open: field, and then clickOK.

Attention: Exercise caution when changing the Windows registry. Changing the wrong registry entry or makingan incorrect entry for a setting can introduce an error that prevents your server from booting up or operatingcorrectly.

JNI and QLogic host bus adapter settingsThe following tables detail settings for the various host bus adapter (HBA) cards for Sun Solaris.

Note: JNI host bus adapters are supported only on Solaris 8 and 9. They are not supported on Solaris 10.

JNI HBA card settingsThe JNI cards are not plug-and-play with autoconfiguration. Instead, you might need to change thesettings or bindings.

Configuration settings for FCE-1473/FCE-6460/FCX2-6562/FCC2-6562These JNI HBAs (FCE-1473, FCE-6460, FCX2-6562, and FCC2-6562) are supported with all currentlysupported levels of DS3000, DS4000, or DS5000 controller firmware.

Important: For all settings that are listed in Table B-3, you must uncomment the line. This is true for bothdefault settings and for settings that you must change.

Table B-3. Configuration settings for FCE-1473/FCE-6460/FCX2-6562/FCC2-6562

Original value New value

FcLoopEnabled = 1FcLoopEnabled = 0 (for non-loop; auto-topology)FcLoopEnabled = 1 (for loop)

FcFabricEnabled = 0FcFabricEnabled = 0 (for non-fabric; auto-topology)FcFabricEnabled = 1 (for fabric)

FcEngHeartbeatInterval = 5 Same as original value (in seconds).

FcLinkUpRecoveryTime =1000

Same as original value (in milliseconds).

BusRetryDelay = 5000 Same as original value (in milliseconds).

TargetOfflineEnable = 1TargetOfflineEnable = 0 (Disable)TargetOfflineEnable = 1 (Enable)

FailoverDelay = 30; FailoverDelay = 60 (in seconds).

FailoverDelayFcTape = 300 Same as original value (seconds).

TimeoutResetEnable = 0 Same as original value.

QfullRetryCount = 5 Same as original value.

QfullRetryDelay = 5000 Same as original value (in milliseconds).


Table B-3. Configuration settings for FCE-1473/FCE-6460/FCX2-6562/FCC2-6562 (continued)


LunRecoveryInterval = 50 Same as original value (in milliseconds).

FcLinkSpeed = 3 Same as original value. (This value [auto-negotiate] is the recommended setting.)

JNICreationDelay = 1 JNICreationDelay = 10 (in seconds).

FlogiRetryCount = 3 Same as original value.

FcFlogiTimeout = 10 Same as original value (in seconds).

PlogiRetryCount = 3 Same as original value.

PlogiControlSeconds = 30 Same as original value (in seconds).

LunDiscoveryMethod = 1 Same as original value (LUN reporting).

CmdTaskAttr = 0CmdTaskAttr = 0 (Simple Queue)CmdTaskAttr = 1 (Untagged)

automap = 0 automap = 1 (Enable)

FclpEnable = 1 FclpEnable = 0 (Disable)

OverrunFailoverCount = 0 Same as original value.

PlogiRetryTime = 50 Same as original value.

SwitchGidPtSyncEnable = 0 Same as original value.

target_throttle = 256 Same as original value.

lun_throttle = 64 Same as original value.

Add these settings. target0_hba = “jnic146x0”;

target0_wwpn = “<controller wwpn>”

target1_hba = “jnic146x1”;

target1_wwpn = “<controller wwpn>”

Note: You might need to run the /etc/raid/bin/genjniconf reconfigure script from the Solaris shell.

# /etc/raid/bin/genjniconf

Configuration settings for FCE-1063/FCE2-1063/FCE-6410/FCE2-6410These JNI HBAs (FCE-1063, FCE2-1063, FCE-6410, and FCE2-6410) are supported with all currentlysupported levels of DS3000, DS4000, or DS5000 controller firmware.

Note: For all settings that are listed in Table B-4, you must uncomment the line. This is true for bothdefault settings and for settings that you must change.

Table B-4. Configuration settings for FCE-1063/FCE2-1063/FCE-6410/FCE2-6410


FcLoopEnabled = 1FcLoopEnabled = 0 (for non-Loop)FcLoopEnabled = 1 (for Loop)

FcFabricEnabled = 0FcFabricEnabled = 0 (for non-fabric)FcFabricEnabled = 1 (for fabric)


Table B-4. Configuration settings for FCE-1063/FCE2-1063/FCE-6410/FCE2-6410 (continued)


FcPortCfgEnable = 1FcPortCfgEnable = 0 (port reconfiguration not required)FcPortCfgEnable = 1 (port reconfiguration required)

FcEngHeartbeatInterval = 5 Same as original value (in seconds).

FcLrrTimeout = 100 Same as original value (in milliseconds).

FcLinkUpRecoverTime =1000

Same as original value (in milliseconds).

BusyRetryDelay = 5000 Same as original value (in milliseconds).

FailoverDelay = 30; FailoverDelay = 60;

TimeoutResetEnable = 0 Same as original value.

QfullRetryCount = 5 Same as original value.

QfullRetryDelay = 5000 Same as original value (in milliseconds).

loRecoveryDelay = 50 Same as original value (in milliseconds).

JniCreationDelay = 5; JniCreationDelay = 10;

FlogiRetryCount = 3 Same as original value.

PlogiRetryCount = 5 Same as original value.

FcEmIdEndTcbTimeCount =1533

Same as original value.

target_throttle = 256 Same as original value. (Default throttle for all targets.)

lun_throttle = 64 Same as original value. (Default throttle for all LUNs.)

automap = 0automap = 0 (persistence binding)automap = 1 (automapping)

Add these settings. target0_hba = “jnic146x0”;

target0_wwpn = “controller wwpn”

target1_hba = “jnic146x1”;

target1_wwpn = “controller wwpn”

v You might need to run the /etc/raid/bin/genjniconf reconfigure script from the Solaris shell.


v Set portEnabled = 1; only when you see JNI cards entering non-participating mode in the/var/adm/messages file. Under that condition:1. Set FcPortCfgEnabled = 1;

2. Restart the host.3. Set FcPortCfgEnabled = 0;

4. Restart the host again.When you have done so, check /var/adm/messages to be sure that it sets the JNI cards to Fabric orLoop mode.

Configuration settings for FCI-1063This JNI HBA (FCI-1063) is supported only in configurations with DS3000, DS4000, or DS5000 controllerfirmware versions 05.4x.xx.xx, or earlier.


Note: For all settings that are listed in Table B-5, you must uncomment the line. This is true for bothdefault settings and for settings that you must change.

Table B-5. Configuration settings for FCI-1063


scsi_initiator_id = 0x7d Same as original value.

fca_nport = 0; fca_nport = 1 (for the fabric) / fca_nport = 0 (for the loop)

public_loop = 0 Same as original value.

target_controllers = 126 Same as original value.

ip_disable = 1; Same as original value.

ip_compliant = 0 Same as original value.

qfull_retry_interval = 0 Same as original value.

qfull_retry_interval = 1000 Same as original value (in milliseconds)

failover = 30; failover = 60 (in seconds)

failover_extension = 0 Same as original value.

recovery_attempts - 5 Same as original value.

class2_enable = 0 Same as original value.

fca_heartbeat = 0 Same as original value.

reset_glm = 0 Same as original value.

timeout_reset_enable = 0 Same as original value.

busy_retry_delay= 100; Same as original value. (in milliseconds)

link_recovery_delay = 1000; Same as original value. (in milliseconds)

scsi_probe_delay = 500; scsi_probe_delay = 5000 (in milliseconds; 10 milliseconds resolution)

def_hba_binding = “fca-pci*”;def_hba_binding = “nonjni”; (for binding)def_hba_binding = “fcaw”; (for non-binding)

def_wwnn_binding = “$xxxxxx” def_wwnn_binding = “xxxxxx”

def_wwpn_binding = “$xxxxxx” Same as the original entry.

fca_verbose = 1 Same as the original entry.

Will be added by reconfigure script name=“fca-pci” parent=“physical path” unit-address=“#”

Will be added by reconfigure script target0_hba=“fca-pci0” target0_wwpn=“controller wwpn”;

Will be added by reconfigure script name=“fca-pci” parent=“physical path”unit-address=“#”

Will be added by reconfigure script target0_hba=“fca-pci1” target0_wwpn= “controller wwpn”;

Note: You might need to run the /etc/raid/bin/genjniconf reconfigure script from the Solaris shell.


Configuration settings for FC64-1063This JNI HBA (FC64-1063) is supported only in configurations with DS3000, DS4000, or DS5000 controllerfirmware versions 05.4x.xx.xx, or earlier.

Important: For all settings that are listed in Table B-6 on page B-12, you must uncomment the line. Thisis true for both default settings and for settings that you must change.


Table B-6. Configuration settings for FC64-1063


fca_nport = 0; fca_nport =1;

ip_disable = 0; ip_disable=1;

failover = 0; failover =30;

busy_retry_delay = 5000; busy_retry_delay = 5000;

link_recovery_delay = 1000; link_recovery_delay = 1000;

scsi_probe_delay = 5000; scsi_probe_delay = 5000;

def_hba_binding = “fcaw*”;Direct attached configurations:def_hba_binding = “fcaw*”;

SAN-attached configurations:def_hba_binding = “nonJNI”;

def_wwnn_binding = “$xxxxxx” def_wwnn_bindindef_hba_ binding = “nonjni”; g = “xxxxxx”

def_wwnn_binding = “$xxxxxx” Same as the original entry.

Will be added by reconfigurescript

name=“fcaw” parent=“<physical path>”unit-address=“<#>”


target0_hba=“fcaw0” target0_wwpn=“<controller wwpn>”;


name=“fcaw” parent=“<physical path>”unit-address=“<#>”


target0_hba=“fcaw0” target0_wwpn= “<controller wwpn>”;

Note: You might need to run the /etc/raid/bin/genscsiconf reconfigure script from the shell prompt.

# /etc/raid/bin/genscsiconf

QLogic HBA card settingsThe QLogic cards are not plug-and-play with autoconfiguration. Instead, you need to change the settingsor bindings, as described in Table B-7 on page B-13.

Note: In Table B-7 on page B-13, the HBA is identified as hba0. However, you need to modify thesettings on both QLogic HBA cards: hba0 and hba1.

When you modify the settings on hba1 use the same values that are listed in the table, but change allinstances of hba0 to hba1, as shown in the following example:

HBA card Original value New value

hba0 hba0-execution-throttle=16; hba0-execution-throttle=255;

hba1 hba1-execution-throttle=16; hba1-execution-throttle=255;

In the vi Editor, uncomment and modify the loop attributes of each QLogic HBA card, using the valuesdescribed in Table B-7 on page B-13.


Table B-7. Configuration settings for QL2342

Original value New value Comments

max-frame-length=2048; max-frame-length=2048 Use the default.

execution-throttle=16; execution-throttle=255; Change.

login-retry-count=8; login-retry-count=30; Change.

enable-adapter-hard-loop-ID=0; enable-adapter-hard-loop-ID=1; Change.

adapter-hard-loop-ID=0; adapter-hard-loop-ID=0; Needs to be a uniquenumber.

enable-LIP-reset=0; enable-LIP-reset=0; Use the default.

hba0-enable-LIP-full-login=1; hba0-enable-LIP-full-login=1; Use the default.

enable-target-reset=0; enable-target-reset=0; Use the default.

reset-delay=5 reset-delay=8 Change.

port-down-retry-count=8; port-down-retry-count=70; Change.

maximum-luns-per-target=8; maximum-luns-per-target=0; Change.

connection-options=2; connection-options=2; Use the default.

fc-tape=1; fc-tape=0; Change.

loop-reset-delay = 5; loop-reset-delay = 8; Change.

> gbyte-addressing = disabled; > gbyte-addressing = enabled; Change.

link-down-timeout = 30; link-down-timeout = 60; Change.

Connecting HBAs in an FC switch environmentThere are two primary zoning schemes you can use when you connect Fibre Channel host bus adapters(HBAs) in host servers to DS3000, DS4000, or DS5000 Storage Subsystem host ports in a Fibre Channelswitch environment. In a one-to-one zoning scheme, each HBA port is zoned to one controller host port.In a one-to-two zoning scheme, each HBA port is zoned to two controller host ports.

As a general rule, the HBA and the storage subsystem host port connections should be zoned tominimize the possible interactions between the ports in a SAN fabric environment. A one-to-one zoningscheme, though not required, minimizes interactions because it connects one HBA port to just one serverhost port. However, the zoning scheme you choose depends on your host-storage SAN fabric topologyand the capabilities of your Fibre Channel switches.

Depending on your host-storage SAN fabric topology Fibre Channel switch capabilities, you canimplement one of the two following zoning schemes in Figure B-1 on page B-14 and Figure B-2 on pageB-14.

Note: For more information about zoning best practices and requirements, see the Fibre Channel SwitchHardware Reference Guide or other documentation that came with the Fibre Channel switch. For links toswitch documentation on the IBM Web site, go to:

www.ibm.com/servers/storage/support/san/index.html


http://www.ibm.com/servers/storage/support/san/index.html

In this zoning scheme (denoted by the translucent bar), one HBA port is zoned to one controller host port.

FC Switch FC Switch

Server

HBA 1 HBA 2

Storage subsystem

C1(Controller 1)

C2(Controller 2)

Figure B-1. One-to-one zoning scheme

In this zoning scheme (denoted by the translucent bars), one HBA port is zoned to two controller host ports.

FC Switch FC Switch

Server

HBA 1 HBA 2

Storage subsystem

C1(Controller 1)

C2(Controller 2)

Figure B-2. One-to-two zoning scheme


Appendix C. Using a DS3000, DS4000, or DS5000 with aVMware ESX Server configuration

DS Storage Manager software is not currently available for VMware ESX Server operating systems.Therefore, to manage DS3000, DS4000, and DS5000 Storage Subsystems with your VMware ESX Serverhost, you must install the Storage Manager client software (SMclient) on a Windows or Linuxmanagement workstation. (This can be the same workstation that you use for the browser-based VMwareESX Server Management Interface.)

For additional information about using a DS3000, DS4000, or DS5000 Storage Subsystem with a VMwareESX Server host, see “VMware ESX Server restrictions” on page C-3.

You can also refer to the System Storage Interoperation Center at the following Web site:

www.ibm.com/systems/support/storage/config/ssic

Sample configuration

Figure C-1 shows a sample VMware ESX Server configuration.

Ethernet

ESX server

Fibre-channelI/O path

Controller

Controller

Storage subsystems

Management station ESX server

Ethernet

SJ001150

Software requirements

This section describes the software that is required to use a VMware ESX Server host operating systemwith a DS3000, DS4000, or DS5000 Storage Subsystem.

Figure C-1. Sample VMware ESX Server configuration

© Copyright IBM Corp. 2009, 2010 C-1


Management station

The following software is required for the Windows or Linux management station:1. SM Runtime (Linux only)2. SMclient (Linux and Windows)

Host (VMware ESX Server)

The following software is required for VMware ESX Server:v VMware ESX Server (with DS3000, DS4000, or DS5000 controller firmware version 07.1x.xx.xx)v VMware ESX Server-supplied driver for the Fibre Channel HBAsv VMware ESX Server-supplied QLogic driver failover setupv VMware ESX Server Tools (installed on all virtual machines using DS3000, DS4000 or DS5000 logical

drives)

Earlier versions of VMware ESX Server:

1. VMware ESX Server 2.1 was supported with DS4000 and DS5000 controller firmware version06.12.xx.xx only.

2. VMware ESX Server 2.0 was supported with DS4000 and DS5000 controller firmware version05.xx.xx.xx only.

Clustering: If you intend to create a cluster configuration, you must use Microsoft Cluster Servicessoftware, in addition to the host software requirements listed in this section.

Note: VMware ESX Server 2.5 and higher comes with a Distributed Resource Scheduler and highavailability for clustering, which allows you to aggregate several hosts' resources into one resource pool.(A DRS cluster is implicitly a resource pool.)

For information about Windows clustering with VMware ESX Server, see the ESX Server 2.5 InstallationGuide at the following Web site: http://www.vmware.com/support/pubs/.

Hardware requirements

You can use VMware ESX Server host servers with the following types of DS3000, DS4000, and DS5000Storage Subsystems and expansion units. For additional information, you can refer to the System StorageInteroperation Center at the following Web site:


Note: For general DS3000, DS4000, and DS5000 requirements, see Chapter 1, “Preparing for installation,”on page 1-1.

DS5000 Storage Subsystems

v DS5300v DS5100

DS4000 Storage Subsystems

v DS4100 (Dual-controller units only)v DS4200v DS4300 (Dual-controller and Turbo units only)v DS4400v DS4500

C-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

http://www.vmware.com/support/pubs/


v DS4700v DS4800

DS5000 storage expansion units

v EXP5000

DS4000 storage expansion units

v EXP100v EXP420 ( with DS4200 only)v EXP500v EXP700v EXP710v EXP810

VMware ESX Server restrictionsSAN and connectivity restrictions

v VMware ESX Server hosts support host-agent (out-of-band) managed DS3000, DS4000, andDS5000 configurations only. Direct-attach (in-band) managed configurations are not supported.

v VMware ESX Server hosts can support multiple host bus adapters (HBAs) and DS3000, DS4000,and DS5000 devices. However, there is a restriction on the number of HBAs that can beconnected to a single DS3000, DS4000, or DS5000 Storage Subsystem. You can configure up totwo HBAs per partition and up to two partitions per DS3000, DS4000, or DS5000 StorageSubsystem. Additional HBAs can be added for additional DS3000, DS4000, and DS5000 StorageSubsystems and other SAN devices, up to the limits of your specific subsystem platform.

v When you are using two HBAs in one VMware ESX Server, LUN numbers must be the samefor each HBA attached to DS3000, DS4000, or DS5000 Storage Subsystem.

v Single HBA configurations are allowed, but each single HBA configuration requires that bothcontrollers in the DS3000, DS4000, or DS5000 be connected to the HBA through a switch. Ifthey are connected through a switch, both controllers must be within the same SAN zone asthe HBA.

Attention: Having a single HBA configuration can lead to loss of access data in the event of apath failure.

v Single-switch configurations are allowed, but each HBA and DS3000, DS4000, or DS5000controller combination must be in a separate SAN zone.

v Other storage devices, such as tape devices or other disk storage, must be connected throughseparate HBAs and SAN zones.

Partitioning restrictions

v The maximum number of partitions per VMware ESX Server host, per DS3000, DS4000, orDS5000 Storage Subsystem, is two.

v All logical drives that are configured for VMware ESX Server must be mapped to an VMwareESX Server host group.

Note: Currently, a VMware ESX Server-specific host type is not available for DS3000, DS4000,or DS5000 Storage Subsystems. If you are using the default host group, ensure that the defaulthost type is LNXCLVMWARE.

v In a DS4100 Storage Subsystem configuration, you must initially assign the LUNs to ControllerA, on the lowest-numbered HBA. After the LUNs are formatted, you can change the path toController B. (This restriction will be corrected in a future release of ESX Server.)

v Assign LUNs to the ESX Server starting with LUN number 0.

Appendix C. Using a DS3000, DS4000, or DS5000 with a VMware ESX Server configuration C-3

v Do not map an access (UTM) LUN to any of the ESX Server hosts or host groups. Access(UTM) LUNs are used only with in-band managed DS3000, DS4000, and DS5000configurations, which VMware ESX Server does not support at this time.

Failover restrictions

v You must use the VMware ESX Server failover driver for multipath configurations. Otherfailover drivers (such as RDAC) are not supported in VMware ESX Server configurations.

v The default failover policy for all DS3000, DS4000, and DS5000 Storage Subsystems is nowMRU (most recently used).

v Use the LNXCLVMWARE host type in VMware ESX Server configurations (2.0 and higher).The LNXCLVMWARE host type automatically disables Auto Drive Transfer (ADT).

Interoperability restrictions

v DS4100 and DS4300 single-controller Storage Subsystems are not supported with VMware ESXServer hosts. (DS4100 and DS4300 dual-controller Storage Subsystems are supported.)

v EXP700 storage expansion units are not supported with DS4800Storage Subsystems. You mustupgrade to EXP710 storage expansion units.

Other restrictions

v Dynamic Volume Expansion is not supported for VMFS-formatted LUNs.v For information about availability of DS Copy Service features that are supported VMware ESX

Server 2.5 Server and higher configurations, contact your IBM support representative.v Recommendation: Do not boot your system from a SATA device.

Other VMware ESX Server host information

For more information about setting up your VMware ESX Server host, see the documentation andREADME files that are maintained at the following Web site:

www.vmware.com/support/pubs/

For information about installing a VMware ESX Server operating system on an IBM server, see the IBMsupport Web site at:

www-03.ibm.com/systems/i/advantages/integratedserver/vmware/

Configuring storage subsystems for VMware ESX ServerBefore you can configure storage subsystems, you must physically configure the host server, SAN fabric,and DS3000, DS4000, or DS5000 controllers; assign initial IP addresses to the controllers; and installSMclient on the Windows or Linux management station. See Chapter 4, “Configuring storage,” on page4-1 for storage subsystem configuration procedures.

Cross connect configuration for VMware connections

A cross-connect Storage Area Network (SAN) configuration is required when VMware hosts areconnected to DS3000, DS4000, or DS5000 storage arrays. Each Host Bus Adapter (HBA) in a VMware hostmust have a path to each of the controllers in the DS storage array. Figure C-2 on page C-5 shows thecross connections for VMware server configurations.



http://www-03.ibm.com/systems/i/advantages/integratedserver/vmware/

FC Switch FC Switch

Server 1 Server 2


Storage subsystem


Notes® on mapping LUNs to a storage partition

See “Mapping LUNs to a storage partition” on page 4-12 for procedures that describe how to map theLUNs to a partition. This section contains notes about LUN mapping that are specific to VMware ESXServers.

When you map you LUNs on VMware ESX Server, note the following:v It is recommended that you always map the LUNs using consecutive numbers, starting with LUN 0.

For example, map LUNs to numbers 0; 1; 2; 3; 4; 5; and so on, without skipping any numbers.v On each partition, you must map a LUN 0.v If your configuration does not require LUN sharing (single or multiple independent ESX Servers, local

virtual cluster), each logical drive must be mapped either directly to a host, or to a host group with asingle host as a member.

v LUN sharing across multiple ESX servers is only supported when you are configuring VMotionenabled hosts or Microsoft Cluster nodes. On LUNs that are mapped to multiple ESX Servers, youmust change the access mode to Shared.You can map the LUNs to a host group for the ESX Servers, so they will be available to all members ofthe host group. For additional information on Windows Clustering with ESX Server, see the ESXInstallation Guide at the following Web site:www.vmware.com/support/pubs/

Steps for verifying the storage configuration for VMwareComplete the following steps to help you verify that your storage setup is fundamentally correct and thatyou can see the DS3000, DS4000, or DS5000 storage:1. Boot the server.2. On initialization of the QLogic BIOS, press Ctrl+Q to enter the Fast!UTIL setup program.3. Select the first host bus adapter that is displayed in the Fast!UTIL screen.4. Select Host Adapter Settings, and press Enter.

Figure C-2. Cross connect configuration for VMware connections

Appendix C. Using a DS3000, DS4000, or DS5000 with a VMware ESX Server configuration C-5


5. Select Scan Fibre Devices and press Enter. The resulting output is similar to the following:

Scan Fibre Channel LoopID Vendor Product Rev Port Name Port ID128 No device present 0520129 IBM 1742 0520 200400A0b00F0A16 610C00130 No device present131 No device present132 No device present133 No device present134 No device present135 No device present

Note: Depending on how the configuration is cabled, you might see multiple instances.

If you do not see a DS3000, DS4000, or DS5000 controller, verify the cabling, switch zoning, and LUNmapping.


Appendix D. Using DS Storage Manager with high-availabilitycluster services

The high-availability clustering services provided by DS Storage Manager allow application services tocontinue when a hardware or software failure occurs. This system protects you from software failures aswell as from the failure of a CPU, disk, or LAN component. If a component fails, its redundant partnercomponent takes over cluster services and coordinates the transfer between components.

General information

This document does not describe how to install or configure cluster services. Refer to documentation thatis provided with your cluster service products for this information.

Important: The information in this document might not include up-to-date cluster software versionlevels.

For the latest requirements and user information about using DS Storage Manager with cluster services,see the README file that is located in the DS Installation CD for your host operating system, or checkthe most recent README files online.

See “Finding Storage Manager software, controller firmware, and README files” on page xiii forinstructions on finding the README files online.

You can also find more information on the System Storage Interoperation Center, which is maintained atthe following Web site:


Using cluster services on AIX systems

The following sections contain general hardware requirements and additional information about thecluster services.

Important: The information in this document might not show up-to-date cluster software version levels.Check the Storage Manager README file for AIX for up-to-date information about clusteringrequirements. See “Finding Storage Manager software, controller firmware, and README files” on pagexiii for instructions on finding the README file on the Web.

You can also refer to the following Web sites for the most current information about AIX and clustering:


publib.boulder.ibm.com/infocenter/clresctr/index.jsp

High Availability Cluster Multi-Processing

This section contains general requirements and usage notes for High Availability Cluster Multi-Processing(HACMP™) support with DS Storage Manager.

© Copyright IBM Corp. 2009, 2010 D-1



http://publib.boulder.ibm.com/infocenter/clresctr/index.jsp

Software requirementsFor the latest supported HACMP versions, refer to the System Storage Interoperation Center at thefollowing Web site:


Configuration limitationsThe following limitations apply to HACMP configurations:v HACMP C-SPOC cannot be used to add a DS3000, DS4000, or DS5000 disk to AIX using the Add a Disk

to the Cluster facility.v HACMP C-SPOC does not support enhanced concurrent mode arrays.v Single-HBA configurations are allowed, but each single-HBA configuration requires that both

controllers in the DS3000, DS4000, or DS5000 be connected to a switch, within the same SAN zone asthe HBA.

Attention: Although single-HBA configurations are supported, they are not recommended for HACMPenvironments because they introduce a single point-of-failure in the storage I/O path.

v Switched fabric connections between the host nodes and the DS3000, DS4000, or DS5000 storagesubsystem are recommended; however, direct attachment from the host nodes to the DS3000, DS4000,or DS5000 storage subsystem in an HACMP environment is supported, only if all the followingrestrictions and limitations are met:– Only dual-controller DS3000, DS4000, or DS5000 storage subsystem versions are supported for direct

attachment in a high-availability (HA) configuration.– The AIX operating system must be version 05.2 or later.– The HACMP clustering software must be version 05.1 or later.– All host nodes that are directly attached to the DS3000, DS4000, or DS5000 storage subsystem must

be part of the same HACMP cluster.– All logical drives (LUNs) that are surfaced by the DS3000, DS4000, or DS5000 storage subsystem are

part of one or more enhanced concurrent mode arrays.– The array varyon is in the active state only on the host node that owns the HACMP non-concurrent

resource group (which contains the enhanced concurrent mode array or arrays). For all other hostnodes in the HACMP cluster, the enhanced concurrent mode array varyon is in the passive state.

– Direct operations on the logical drives in the enhanced concurrent mode arrays cannot beperformed, from any host nodes in the HACMP cluster, if the operations bypass the LogicalVolumeManager (LVM) layer of the AIX operating system. For example, you cannot use a DDcommand while logged in as the root user.

– Each host node in the HACMP cluster must have two Fibre Channel connections to the DS3000,DS4000, or DS5000 storage subsystem. One direct Fibre Channel connection must be to controller Ain the DS3000, DS4000, or DS5000 storage subsystem, and the other direct Fibre Channel connectionmust be to controller B in the DS3000, DS4000, or DS5000 storage system.

– You can directly attach a maximum of two host nodes in an HACMP cluster to a dual-controllerversion of a DS4100 or DS4300 storage subsystem.

– You can directly attach a maximum of two host nodes in an HACMP cluster to a DS3000, DS4000, orDS5000 storage subsystem. Each host node must have two direct Fibre Channel connections to thestorage subsystem.

Note: In a DS3000, DS4000, or DS5000 storage subsystem, the two direct Fibre Channel connectionsfrom each host node must to independent miniHUBs. Therefore, this configuration requires that fourhost miniHUBs (feature code 3507) be installed in the DS3000, DS4000, or DS5000 storagesubsystem—two host miniHUBs for each host node in the HACMP cluster.

D-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide


Other HACMP usage notesThe following notations are specific to HACMP environments:v HACMP clusters can support from two to 32 servers on each DS3000, DS4000, and DS5000 partition. If

you run this kind of environment, be sure to read and understand the AIX device drivers queue depthsettings that are described in “Setting the queue depth for hdisk devices” on page 5-31.

v You can attach non-clustered AIX hosts to a DS3000, DS4000, or DS5000 that is running DS StorageManager and is attached to an HACMP cluster. However, you must configure the non-clustered AIXhosts on separate host partitions on the DS3000, DS4000, or DS5000.

Parallel System Support Programs and General Parallel File System

This section contains general requirements and usage notes for Parallel System Support Programs (PSSP)and General Parallel File System (GPFS™) support with DS Storage Manager.

Software requirementsFor the latest supported PSSP and GPFS versions, refer to the System Storage Interoperation Center at thefollowing Web site:


Configuration limitationsThe following limitations apply to PSSP and GPFS configurations:v Direct connection is not allowed between the host node and a DS3000, DS4000, or DS5000 Storage

Subsystem. Only switched fabric connection is allowed.v RVSD clusters can support up to two IBM Virtual Shared Disk and RVSD servers for each DS3000,

DS4000, or DS5000 partition.v Single node quorum is not supported in a dual-node GPFS cluster with DS3000, DS4000, or DS5000

disks in the configuration.v Heterogeneous configurations are not supported.

Other PSSP and GPFS usage notesIn GPFS file systems, the following DS3000, DS4000, and DS5000 cache settings are supported:v Read cache enabled or disabledv Write cache enabled or disabledv Cache mirroring enabled or disabled (depending upon the write cache mirroring setting)

The performance benefits of read or write caching depends on the application.

GPFS, PSSP, and HACMP cluster configuration diagramsThe diagrams in this section show both the preferred and failover paths from an HBA pair to a givenlogical drive or set of logical drives.

A preferred path to a logical drive is determined when the logical drive is created and distributed acrossa DS3000, DS4000, or DS5000 controller. The controller to which it is assigned determines which path ispreferred or active for I/O transfer. logical drives can, and in most cases should, be assigned to bothcontrollers, balancing the I/O load across HBAs and DS3000, DS4000, or DS5000 controllers.

Figure D-1 on page D-4 shows a cluster configuration that contains a single DS3000, DS4000, or DS5000storage subsystem, with one to four partitions.

Appendix D. Using DS Storage Manager with high-availability cluster services D-3


Figure D-2 on page D-5 shows a cluster configuration that contains three DS3000, DS4000, or DS5000storage subsystems, with one partition on each storage subsystem.

FC fabric 5

Fabric zone 4

Fabric zone 1

FC fabric 8

FC

SW

FC

SW

RVSD ClusterDS4000 Storage Server

PreferredFailover

Partition 1

Partition 2

Partition 3

Partition 4

Serverpair

1

Serverpair

4

Port 0

Port 1

Port 0

CTLR A

CTLR B

Port 1

LUN 31LUN 0

LUN 31LUN 1

SJ001031

WWPN-1A

WWPN-7A

WWPN-7A

WWPN-8A

WWPN-7B

WWPN-8B

WWPN-1A

WWPN-2A

WWPN-1B

WWPN-2BWWPN-2A

WWPN-8A

WWPN-2B

WWPN-8B

WWPN-1B

WWPN-7B

FCFCHA0

FCHA0

FCHA0

FCHA0

FCHA1

FCHA1

FCHA1

FCHA1

FC

VSD 1

FC

FC

VSD 7

FC

FC

VSD 2

FC

FC

VSD 8

Figure D-1. Cluster configuration with single DS3000, DS4000, or DS5000 storage subsystem—one to four partitions


Figure D-3 on page D-6 shows a cluster configuration that contains four DS3000, DS4000, or DS5000storage subsystems, with one partition on each storage subsystem.

Fabric zone 2

Fabric zone 1

FC

SW

FC

SW

PreferredFailover

Partition 1

Partition 1

Partition 1

Port 1

Port 1

Port 1

Port 1

Port 1

Port 1

Port 0

Port 0

Port 0

Port 0

Port 0

Port 0

CTLR A

CTLR A

CTLR A

CTLR B

CTLR B

CTLR B

LUN 31

LUN 31

LUN 31

LUN 0

LUN 0

LUN 0

SJ001032

WWPN-1A

WWPN-2A

WWPN-1B

WWPN-2B

WWPN-1A

WWPN-1B

WWPN-2A

WWPN-1B

WWPN-1A

WWPN-2A

WWPN-1B

WWPN-2B

WWPN-1A

WWPN-2B

WWPN-1B

WWPN-2B

FCFCHA0

FCHA0

FCHA1

FCHA1

FC

FC

FC

VSD 1

VSD 2

RVSD Cluster

DS4000 Storage Server



Figure D-2. Cluster configuration with three DS3000, DS4000, or DS5000 storage subsystems—one partition perDS3000, DS4000, or DS5000


Figure D-4 on page D-7 shows a cluster configuration that contains two DS3000, DS4000, or DS5000storage subsystems, with two partitions on each storage subsystem.

Fabric zone 2

Fabric zone 1

FC

SW

FC

SW DS4000 Storage Server #2

DS4000 Storage Server #3


PreferredFailover

Partition 1

Partition 1

Partition 1

Partition 1

Port 1

Port 1

Port 1

Port 1

Port 1

Port 1

Port 1

Port 1

Port 0

Port 0

Port 0

Port 0

Port 0

Port 0

Port 0

Port 0

CTLR A

CTLR A

CTLR A

CTLR A

CTLR B

CTLR B

CTLR B

CTLR B

LUN 31

LUN 31

LUN 31

LUN 31

LUN 0

LUN 0

LUN 0

LUN 0

WWPN-1A

WWPN-2A

WWPN-1B

WWPN-2B

WWPN-1A

WWPN-2A

WWPN-1B

WWPN-2B

WWPN-1C

WWPN-2C

WWPN-1D

WWPN-2D

WWPN-1C

WWPN-2C

WWPN-1D

WWPN-2D

FC

FC

FC

FC

FC

FC

FC

FC

VSD 1

VSD 2

RVSD Cluster

WWPN-1A

WWPN-1B

WWPN-1A

WWPN-1B

WWPN-1C

WWPN-1D

WWPN-1C

WWPN-1D


SJ001033

Figure D-3. Cluster configuration with four DS3000, DS4000, or DS5000 storage subsystems—one partition perDS3000, DS4000, or DS5000


Figure D-5 on page D-8 shows an HACMP/GPFS cluster configuration that contains a single DS3000,DS4000, or DS5000 storage subsystem, with one partition.

Fabric zone 3

Fabric zone 1

Fabric zone 2

Fabric zone 4

FC

SW

FC

SW


PreferredFailover

SJ001034

WWPN-1A

WWPN-2A

WWPN-1C

WWPN-2C

WWPN-2A

WWPN-2B

WWPN-1A

WWPN-1B

WWPN-2C

WWPN-2D

WWPN-1C

WWPN-1D

WWPN-1B

WWPN-2B

WWPN-1D

WWPN-2D

WWPN-1A

WWPN-2A

WWPN-1C

WWPN-2C

WWPN-1B

WWPN-2B

WWPN-1D

WWPN-2D

FC

FC

FC

FC

FC

FC

FC

FC

VSD 1

VSD 2

RVSD Cluster

Partition 1

Partition 2

Port 1

Port 1

Port 0

Port 0

CTLR A

CTLR B

LUN 31

LUN 31

LUN 0

LUN 0

Partition 1

Partition 2

Port 1

Port 1

Port 0

Port 0

CTLR A

CTLR B

LUN 31

LUN 31

LUN 0

LUN 0


Figure D-4. RVSD cluster configuration with two DS3000, DS4000, or DS5000 storage subsystems—two partitions perDS3000, DS4000, or DS5000


Figure D-6 on page D-9 shows an HACMP/GPFS cluster configuration that contains two DS3000, DS4000,or DS5000 storage subsystems, with two partitions on each storage subsystem.

Fabric zone 2

Fabric zone 1

FC

SW

FC

SW

HACMP/GPFS ClusterPreferredFailover

Partition 1

Port 1

CTLR A

CTLR B

LUN 31LUN 0

SJ001035

WWPN-1A

WWPN-3A

WWPN-1B

WWPN-3B

WWPN-2A

WWPN-4A

WWPN-32A

WWPN-2B

WWPN-4B

WWPN-32B

Primary:WWPN-1AWWPN-2AWWPN-3AWWPN-4A...WWPN-32A

Failover:WWPN-1BWWPN-2BWWPN-3BWWPN-4B...WWPN-32B

Svr 1

Svr 3

Svr 2

Svr 4

Svr 32


Figure D-5. HACMP/GPFS cluster configuration with one DS3000, DS4000, or DS5000 storage subsystem—onepartition


Using cluster services on HP-UX systems

The information in this document might not show up-to-date cluster software version levels. Check theStorage Manager README file for HP-UX for up-to-date information about clustering requirements. See“Finding Storage Manager software, controller firmware, and README files” on page xiii for instructionson finding the README file online.



You can choose among many configurations when you set up clustering on an HP-UX system. Aminimum configuration consists of two servers that are configured with both a primary and two standbyLANs to establish a heartbeat LAN.

Provide Fibre Channel connections to the storage subsystem through two switches that provide thenecessary redundant data path for the hosts. Ensure that each server has two HP Tachyon host busadapters.

Fabric zone 2

Fabric zone 3

Fabric zone 4

Fabric zone 1

FC

SW

FC

SW

PreferredFailover

SJ0

01

03

6

Svr 1

Svr 3

Svr 32

Svr 2

HACMP/GPFS Cluster

WWPN-1AWWPN-2AWWPN-3AWWPN-32A

WWPN-1AWWPN-2AWWPN-3AWWPN-32A

WWPN-1BWWPN-2BWWPN-3BWWPN-32B

WWPN-1BWWPN-2BWWPN-3BWWPN-32B

WWPN-1CWWPN-2CWWPN-3CWWPN-32C

WWPN-1CWWPN-2CWWPN-3CWWPN-32C

WWPN-1DWWPN-2DWWPN-3DWWPN-32D

WWPN-1DWWPN-2DWWPN-3DWWPN-32D

Partition 1

Partition 1

Partition 2

Partition 2

Port 1

Port 1

Port 1

Port 1

Port 0

Port 0

Port 0

Port 0

CTLR A

CTLR A

CTLR B

CTLR B

LUN 31

LUN 31

LUN 31

LUN 31

LUN 0

LUN 0

LUN 0

LUN 0

WWPN-1AWWPN-1BWWPN-1CWWPN-1D




FC

FC

FC

FC

FC

FC

FC

FC

FC

FC

FC

FC

FC

FC

FC

FC



Figure D-6. HACMP/GPFS cluster configuration with two DS3000, DS4000, or DS5000 storage subsystems—twopartitions per DS3000, DS4000, or DS5000



Using cluster services on Solaris systems

The following sections contain general hardware requirements and additional information about thecluster services.

Important: The information in this document might not show up-to-date cluster software version levels.Check the Storage Manager README file for Solaris for up-to-date information about clusteringrequirements, including the latest supported versions of Veritas Cluster Server. See “Finding StorageManager software, controller firmware, and README files” on page xiii for instructions on finding theREADME file online.



General Solaris requirementsEach Solaris system in the cluster requires the following hardware:v At least three Ethernet ports:

– Two for the private network connections– At least one for the public network connection

v Two Fibre Channel host bus adapters for connection to the storage subsystemv A SCSI connection for operating system disksv Each Veritas Cluster Server system requires at least 128 MB of RAM and 35 MB of free disk space

System dependenciesThis section provides information about RDAC IDs and single points of failure.

RDAC IDs

Add up to eight additional IDs to the /etc/symsm/rmparams file. Complete the following steps to addthem:1. Open the /etc/symsm/rmparams file in the vi Editor by typing the following command:

# vi /etc/symsm/rmparams

2. Modify the Rdac_HotAddIDs line as follows:

Rdac_HotAddIDs:0:1:2:3:4:5:6:7:8

3. Save and close the /etc/symsm/rmparams file.

Single points of failure

When setting up cluster services, it is important to eliminate single points of failure because a singlepoint of failure makes a cluster only as strong as its weakest component. Set up the storage subsystem forshared storage; for example, all the nodes in the cluster must recognize the same storage and the hosttypes must be set correctly.



Appendix E. Viewing and setting AIX Object Data Manager(ODM) attributes

Some of the ODM attributes are for information purposes only. These information-only attributes showhow the DS3000, DS4000, or DS5000 Storage Subsystem is configured or its current state. You can modifyother attributes using SMIT or by using the UNIX chdev -p command.

Attribute definitionsThe following tables list definitions and values of the ODM attributes for dars, dacs and hdisks:v Table E-1: Attributes for dar devices

v Table E-2 on page E-2: Attributes for dac devices

v Table E-3 on page E-3: Attributes for hdisk devices

Note:

1. Attributes with True in the Changeable column can be modified from their default settings.2. Attributes with False in the Changeable column are for informational or state purposes only.

However, some attributes with False in the Changeable column can be modified using DS StorageManager.

3. The lsattr -El (uppercase E, lowercase L) command is another way to determine which attributescan be modified. Attributes that can be modified display True in the last column of the lsattr -Eloutput. You can also display the default values by using the lsattr -Dl command.

Table E-1. Attributes for dar devices

Attribute Definition Changeable (T/F) Possible value

act_controller List of controllers in theactive state at the time ofconfiguration.

False Set at configuration time bythe RDAC software.

all_controller List of controllers thatcomprise this array; usuallythere are two dac devices.

False Set at configuration time bythe RDAC software.

held_in_reset Name of the controller thatwas in the held-in-resetstate at the time ofconfiguration, or none if nocontrollers were in thatstate.

True Set at configuration time bythe RDAC software. Shouldnot be changed.

load_balancing Indicator that showswhether load balancing isenabled (yes) or disabled(no); see the definition ofthe balance_freq attribute formore information.

True Yes or No.Attention: You should onlyset the load_balancingattribute to yes insingle-host configurations.

autorecovery Indicator that showswhether the device returnsthe array to dual-activemode when it detectsproper operation of bothpaths and controllers (yes)or not (no).

True Yes or No. See restrictionson use.

© Copyright IBM Corp. 2009, 2010 E-1

Table E-1. Attributes for dar devices (continued)


hlthchk_freq Number that specifies howoften health checks areperformed, in seconds.

True 1 - 9999. Should not bechanged.

aen_freq Number that specifies howoften polled AEN checksare performed, in seconds.

True 1 - 9999. Should not bechanged.

balance_freq If load_balancing is enabled,number that specifies howoften the system performsload-balancing on the array,in seconds.

True 1 - 9999 - should not bechanged.

fast_write_ok Indicator that showswhether fast-writewrite-caching is availablefor this system (yes) or not(no).

False Yes or No. State of DS3000,DS4000, or DS5000configuration.

cache_size Cache size for bothcontrollers, in megabytes; 0if the sizes do not match.

False 512 or 1024. Set by DS3000,DS4000, or DS5000.

switch_retries Number that specifies howmany times to retry failedswitches, in integers.

True 0 - 255.Default: 5

For most configurations, thedefault is the best setting. Ifyou are using HACMP, itcan be helpful to set thevalue to 0.Attention: You cannot useconcurrent firmwaredownload if you change thedefault setting.

Table E-2. Attributes for dac devices


passive_control Indicator that showswhether this controller wasin passive state at the timeof configuration (yes) or not(no).

False Yes or No. State of DS3000, DS4000,or DS5000 configuration.

alt_held_reset Indicator that showswhether the alternatecontroller was in theheld-in-reset state at thetime of configuration (yes)or not (no).

False Yes or No. State of DS3000, DS4000,or DS5000 configuration.

controller_SN Serial number of thiscontroller.

False Set by DS3000, DS4000, or DS5000.

ctrl_type Type of array to which thiscontroller belongs.

False 1742, 1722, 1742-900. Set by DS3000,DS4000, or DS5000.

cache_size Cache size of this controller,in megabytes.

False 512, 1024. Set by DS3000, DS4000, orDS5000.

E-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Table E-2. Attributes for dac devices (continued)


scsi_id SCSI identifier of thiscontroller.

False Set by SAN, reported by AIX.

lun_id Logical unit number of thiscontroller.


utm_lun_id Logical unit number of thiscontroller, or none if UTM(access logical drives) is notenabled.

False 0 - 31. Set by DS Storage Manager.

node_name Name of the Fibre Channelnode.


location User-defined location labelfor this controller; thesystem does not use thisvalue.

True Set by DS Storage Manager.

ww_name Fibre Channel worldwidename of this controller.


GLM_type GLM type used for thiscontroller.

False High or Low. Set by DS3000, DS4000,or DS5000.

Table E-3. Attributes for hdisk devices


pvid AIX physical volumeidentifier, or none if not set.

False Set by AIX.

q_type Queueing type for thisdevice; must be set tosimple.

False Set by AIX. Must be“simple”.

queue_depth Number that specifies thedepth of the queue basedon system configuration;reduce this number if thearray is returning a BUSYstatus on a consistent basis.

True 1 - 64Note: See “Setting thequeue depth for hdiskdevices” on page 5-31 forimportant informationabout setting this attribute.

PR_key_value Required only if the devicesupports any of thepersistent reserve policies.This attribute is used todistinguish betweendifferent hosts.

True 1-64, or None.Note: You must set thisattribute to non-zero beforethe reserve_policy attributeis set.

reserve_policy Persistent reserve policy,which defines whether areservation methodology isemployed when the deviceis opened.

True no_reservePR_shared,PR_exclusive, orsingle_path

max_transfer Maximum transfer size isthe largest transfer size thatcan be used in sending I/O.

True Numeric value;Default = 1 MBNote: Usually unnecessaryto change default, unlessvery large I/Os requireincreasing the value.

Appendix E. Viewing and setting AIX Object Data Manager (ODM) attributes E-3

Table E-3. Attributes for hdisk devices (continued)


write_cache Indicator that showswhether write-caching isenabled on this device (yes)or not (no); see thedefinition of thecache_method attribute formore information.

False Yes or No.

size Size of this logical drive. False Set by DS3000, DS4000, orDS5000.

raid_level Number that specifies theRAID level of this device.

False 0, 1, 3, 5. Set by DS StorageManager.

rw_timeout Number that specifies theread/write timeout valuefor each read/writecommand to this array, inseconds; usually set to 30.

True 30 - 180. Should not bechanged from default.

reassign_to Number that specifies thetimeout value for FCreassign operations, inseconds; usually set to 120.

True 0 - 1000. Should not bechanged from default.

scsi_id SCSI identifier at the timeof configuration.

False Set by SAN, reported byAIX.

lun_id Logical unit number of thisdevice.

False 0 - 255. Set by DS StorageManager.

cache_method If write_cache is enabled, thewrite-caching method ofthis array; set to one of thefollowing:

v default. Default mode;the word "default" is notseen if write_cache is setto yes.

v fast_write. Fast-write(battery-backed, mirroredwrite-cache) mode.

v fw_unavail. Fast-writemode was specified butcould not be enabled;write-caching is not inuse.

v fast_load. Fast-load(non-battery-backed,non-mirroredwrite-cache) mode.

v fl_unavail. Fast-loadmode was specified butcould not be enabled.

False Default, fast_write,fast_load, fw_unavail,fl_unavail.

prefetch_mult Number of blocks to beprefetched into read cachefor each block read.

False 0 - 100.


Table E-3. Attributes for hdisk devices (continued)


ieee_volname IEEE unique logical drivename identifier for thislogical drive.

False Set by DS3000, DS4000, orDS5000.

Using the lsattr command to view ODM attributesTo view the Object Data Manager (ODM) attribute settings for dars, dacs, and hdisks, use the lsattrcommand, as follows:v To view the default settings, type lsattr -Dl.v To view the attributes that are currently set on the system, type lsattr -El.

The lsattr -El output examples shown in Table E-4, Table E-5, and Table E-6 on page E-6, display theODM attribute settings for a dar, a dac and an hdisk.

Table E-4. Example 1: Displaying the attribute settings for a dar

# lsattr -El dar0act_controller dac0,dac1 Active Controllers Falseaen_freq 600 Polled AEN frequency in seconds Trueall_controller dac0,dac1 Available Controllers Falseautorecovery no Autorecover after failure is corrected Truebalance_freq 600 Dynamic Load Balancing frequency in seconds Truecache_size 128 Cache size for both controllers Falsefast_write_ok yes Fast Write available Falseheld_in_reset none Held-in-reset controller Truehlthchk_freq 600 Health check frequency in seconds Trueload_balancing no Dynamic Load Balancing Trueswitch_retries 5 Number of times to retry failed switches True

Table E-5. Example 2: Displaying the attribute settings for a dac

# lsattr -El dac0GLM_type low GLM type Falsealt_held_reset no Alternate held in reset Falsecache_size 128 Cache Size in MBytes Falsecontroller_SN 1T24594458 Controller serial number Falsectrl_type 1722-600 Controller Type Falselocation Location Label Truelun_id 0x0 Logical Unit Number Falsenode_name 0x200200a0b80f14af FC Node Name Falsepassive_control no Passive controller Falsescsi_id 0x11000 SCSI ID Falseutm_lun_id 0x001f000000000000 Logical Unit Number Falseww_name 0x200200a0b80f14b0 World Wide Name False

Note: Running the # lsattr -Rl <device> -a <attribute> command, will show allowable values for thespecified attribute and is an hdisk attribute list when using MPIO.

Note: In Table E-6 on page E-6, the ieee_volname and lun_id attribute values are shown abbreviated. Anactual output would show the values in their entirety.

Appendix E. Viewing and setting AIX Object Data Manager (ODM) attributes E-5

Table E-6. Example 3: Displaying the attribute settings for an hdisk

lsattr -El hdisk174cache_method fast_write Write Caching method Falseieee_volname 600A0B8...1063F7076A7 IEEE Unique volume name Falselun_id 0x0069...000000 Logical Unit Number Falseprefetch_mult 12 Multiple of blocks to prefetch on read Falsepvid none Physical volume identifier Falseq_type simple Queuing Type Falsequeue_depth 2 Queue Depth Trueraid_level 5 RAID Level Falsereassign_to 120 Reassign Timeout value Truereserve_lock yes RESERVE device on open Truerw_timeout 30 Read/Write Timeout value Truescsi_id 0x11f00 SCSI ID Falsesize 2048 Size in Mbytes Falsewrite_cache yes Write Caching enabled False


Appendix F. DS Diagnostic Data Capture (DDC)

DDC information

Under rare circumstances an internal controller error can force a routine to perform a function referred toas a DDC or Diagnostic Data Capture. When this has occurred, if you open the Enterprise Managementwindow you will notice a red stop sign next to the name of the storage subsystem that has the error (is ina non-optimal state). Click on that storage to bring up the Subsystem Management window. You can thenclick on the Recovery Guru which will show what the issue is, or you can check in the MEL (DS StorageManager Major Events Log), where a critical event will be posted. See “DDC MEL events” on page F-3.

DDC function implementationIn order to assist IBM support in collecting data for troubleshooting certain unusual events in controllerfirmware, the DDC function was implemented.

Note: This function is not implemented with controller firmware code versions that are less then06.12.27.xx level.

How Diagnostic Data Capture works

When the DDC function is implemented, the storage subsystem status will change from Optimal to NeedsAttention due to DDC. This will occur whenever the controllers in the DS subsystem detect unusual eventslike Master abort (due to a bad address accessed by the Fibre Channel chip resulting in a PCI bus error),when the controller is not able to process host I/O requests for an extended period of time (severalminutes), when there is destination device number registry corruption, an EDC (error detection code)error is returned by the disk drives, a quiescence failure for the logical drive owned by the alternatecontroller, or corruption in records related to Storage Partition Management. Once the Needs Attentiondue to DDC flag is set, it will be persistent across the power-cycle and controller reboot, provided thecontroller cache batteries are sufficiently charged. In addition, data reflecting the state of the subsystemcontrollers at the moment in time that the unusual event occurred will be collected and saved until it isretrieved by the user. To clear the Needs Attention due to DDC flag and to retrieve the saved diagnosticdata, see the “Recovery steps.”

Because the current DDC function implementation will hold the DDC data for only one unusual event ata time until the DDC data is saved, the SMcli commands must be performed as soon as possiblewhenever the Needs Attention due to DDC error occurs. This is so the controllers can be ready forcapturing data for any other unusual events. Until the diagnostic data is saved and the Needs Attentiondue to DDC flag is cleared, any occurrences of other unusual events will not trigger the controller tocapture diagnostic data for those events. An unusual event is considered a candidate for a DDC trigger ifa previous DDC trigger is at least 48 hours old or the user has successfully retrieved the previous DDCinformation. In addition, DDC information is only available in the case of a controller that is online. Acontroller that is in service or lock down mode will not trigger a DDC event. After collecting the DDCdata, contact IBM support to report the problem and to enlist assistance in troubleshooting the conditionthat triggered the event.

Recovery stepsFollow steps 1–6 to complete the DDC recovery process:1. Open either the Script Editor from the Enterprise Management Window (EMW), or the Command

Line Interface (CLI).

Note: Refer to the EMW online help for more information on the syntax of these commands.2. To either save or not save the diagnostic data, perform the following step:

© Copyright IBM Corp. 2009, 2010 F-1

Table F-1. Recovery Step 2

If... Then...

You want to save the diagnostic data Go to Step 3.

You do not want to save the diagnostic data Go to Step 5.

3. Execute the following SMcli command:save storageSubsystem diagnosticData file="filename ";

where filename is the location and name of the file that will be saved. The file will be formatted as a.zip file.

Note: Currently, the esm parameter of the command syntax is not supported.4. To work with the diagnostic data, complete the following step:


If... Then...

No error was returned Go to Step 6.

An error was returnedIf... Then...

The error message indicates thatthere was a problem saving the data

Wait 2 minutes and then restart Step3.

The error message indicates thatthere was a problem resetting thedata

Wait 2 minutes and then go to Step 5.

5. Execute the following SMcli command:reset storageSubsystem diagnosticData;


If... Then...

No error was returned Go to Step 6.

An error was returned Wait 2 minutes and then execute the command again.The controllers may need additional time to updatestatus. Note that you may also get an error if thediagnostic data status has already been reset.

Go to Step 6.

6. Select Recheck to rerun the Recovery Guru. The failure should no longer appear in the Summaryarea.

After this process has been successfully completed, the DDC message will automatically be removed, anda recheck of the Recovery Guru will show no entries for DDC capture. If for some reason the data hasnot been removed, the Recovery Guru gives an example of how to clear the DDC information withoutsaving the data. Follow the above procedure using the following command in the script editor:reset storageSubsystem diagnosticData;

F-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

DDC MEL events

When the Diagnostic Data Capture action is triggered by an unusual event, one or more of the followingevents will be posted in the storage subsystems event logs, depending on the user actions.1. Event number: 0x6900.

Description: Diagnostic Data is available.Priority: Critical.This is logged when an unusual controller event triggers the DDC function to store Diagnostic Data.

2. Event number: 0x6901.

Description: Diagnostic Data retrieval operation started.Priority: Informational.This is logged when the user runs the SMcli command to retrieve and save the Diagnostic Data, asdescribed in Step 3 on page F-2.


Description: Diagnostic Data retrieval operation completed.Priority: Informational.This is logged when the Diagnostic Data retrieval and save completes.


Description: Diagnostic Data Needs Attention status/flag cleared.Priority: Informational.This is logged when the user resets the Needs Attention flag due to DDC, using the SMcli commandor Diagnostic Data retrieval and save completes successfully when initiated by the user executing thesave storageSubsystem diagnosticData SMcli command.

Appendix F. DS Diagnostic Data Capture (DDC) F-3

F-4 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Appendix G. The Script Editor

Instead of navigating through the GUI interface to perform storage subsystem management functions, aScript Editor window, as shown in Figure G-1, is provided for running scripted management commands.If the controller firmware version is 5.4x.xx.xx or earlier, some of the management functions that can bedone through the GUI are not implemented through script commands. DS Storage Manager 10.xx inconjunction with controller firmware version 07.xx.xx.xx and higher provides full support of allmanagement functions via SMcli commands.

ds50_001138

Important: Use caution when running the commands in the script window because the Script Editordoes not prompt for confirmation on operations that are destructive such as the Delete arrays, ResetStorage Subsystem configuration commands.

Not all script commands are implemented in all versions of the controller firmware. The earlier thefirmware version, the smaller the set of script commands. For more information about script commandsand firmware versions, see the DS Storage Manager Enterprise Management window.

For a list of available commands and their syntax, see the online Command Reference help.

Figure G-1. The Script Editor window

© Copyright IBM Corp. 2009, 2010 G-1

Using the Script EditorPerform the following steps to open the Script Editor:1. Select a storage subsystem in the Device Tree view or from the Device table.2. Select Tools → Execute Script.3. The Script Editor opens. The Script view and the Output view are presented in the window.

v The Script view provides an area for inputting and editing script commands. The Script viewsupports the following editing key strokes:– Ctrl+A: To select everything in the window– Ctrl+C: To copy the marked text in the window into a Windows clipboard buffer– Ctrl+V: To paste the text from the Windows clipboard buffer into the window– Ctrl+X: To delete (cut) the marked text in the window– Ctrl+Home: To go to the top of the script window– Ctrl+End: To go to bottom of the script window

v The Output view displays the results of the operations.

A splitter bar divides the window between the Script view and the Output view. Drag the splitter barto resize the views.

The following list includes some general guidelines for using the Script Editor:v All statements must end with a semicolon (;).v Each base command and its associated primary and secondary parameters must be separated by a

space.v The Script Editor is not case sensitive.v Each new statement must begin on a separate line.v Comments can be added to your scripts to make it easier for you and future users to understand the

purpose of the command statements.

Adding comments to a scriptThe Script Editor supports the following comment formats:v Text contained after two forward slashes (//) until an end-of-line character is reached

For example:

//The following command assigns hot spare drives.set drives [1,2 1,3] hotspare=true;

The comment //The following command assigns hot spare drives. is included for clarification and isnot processed by the Script Editor.

Important: You must end a comment that begins with // with an end-of-line character, which youinsert by pressing the Enter key. If the script engine does not find an end-of-line character in the scriptafter processing a comment, an error message displays and the script fails.

v Text contained between the /* and */ charactersFor example:

/* The following command assigns hot spare drives.*/set drives [1,2 1,3] hotspare=true;

The comment /*The following command assigns hot spare drives.*/ is included for clarification andis not processed by the Script Editor.

G-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Important: The comment must start with /* and end with */. If the script engine does not find both abeginning and ending comment notation, an error message displays and the script fails.

Appendix G. The Script Editor G-3

G-4 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Appendix H. Tuning storage subsystems

The information in the chapter helps you use data from the Performance Monitor. This chapter alsodescribes the tuning options that are available in DS Storage Manager for optimizing storage subsystemand application performance. Use the Subsystem Management window Performance Monitor to monitorstorage subsystem performance in real time and to save performance data to a file for later analysis. Youcan specify the logical drives and controllers to monitor and the polling interval. Also, you can receivestorage subsystem totals, which is data that combines the statistics for both controllers in an active-activecontroller pair.

Table H-1. Performance Monitor tuning options in the Subsystem Management window

Data field Description

Total I/Os Total I/Os performed by this device since the beginning of the polling session.

Read percentage The percentage of total I/Os that are read operations for this device. Writepercentage is calculated as 100 minus this value.

Cache-hit percentage The percentage of read operations that are processed with data from the cache,rather than requiring a read from the logical drive.

Current® KB per second During the polling interval, the transfer rate is the amount of data, in KB, that ismoved through the Fibre Channel I/O path in one second (also calledthroughput).

Maximum KB per second The maximum transfer rate that is achieved during the Performance Monitorpolling session.

Current I/O per second The average number of I/O requests that are serviced per second during thecurrent polling interval (also called an I/O request rate).

Maximum I/O per second The maximum number of I/O requests that are serviced during a one-secondinterval over the entire polling session.

Load balancingLoad balancing is the redistribution of read/write requests to maximize throughput between the serverand the storage array. Load balancing is very important in high workload settings or other settings whereconsistent service levels are critical. The multi-path driver transparently balances I/O workload, withoutadministrator intervention. Without multi-path software, a server sending I/O requests down severalpaths might operate with very heavy workloads on some paths, while others are not used efficiently.

The multi-path driver determines which paths to a device are in an active state and can be used for loadbalancing. The load balancing policy uses one of three algorithms: round robin, least queue depth, orleast path weight. Multiple options for setting the load balance policies let you optimize I/O performancewhen mixed host interfaces are configured. The load balancing policies that you can choose depend onyour operating system. Load balancing is performed on multiple paths to the same controller, but notacross both controllers.

Table H-2. Load balancing policies supported by operating systems

Operating System Multi-Path Driver Load Balancing Policy

AIX MPIO Round robin, selectable path priority

Red Hat Enterprise Linux 4 Update7

RDAC Round robin, least queue depth

Solaris MPxIO Round robin

© Copyright IBM Corp. 2009, 2010 H-1

Table H-2. Load balancing policies supported by operating systems (continued)

Operating System Multi-Path Driver Load Balancing Policy

SUSE Linux Enterprise 9 ServicePack 4

RDAC Round robin, least queue depth

Windows MPIO Round robin, least queue depth, least pathweight

Round robin with subset

The round robin with subset I/O load balance policy routes I/O requests, in rotation, to each availabledata path to the controller that owns the volumes. This policy treats all paths to the controller that ownsthe volume equally for I/O activity. Paths to the secondary controller are ignored until ownershipchanges. The basic assumption for the round robin policy is that the data paths are equal. With mixedhost support, the data paths might have different bandwidths or different data transfer speeds.

Least queue depth with subset

The least queue depth with subset policy is also known as the least I/Os or least requests policy. Thispolicy routes the next I/O request to a data path that has the least outstanding I/O requests queued. Forthis policy, an I/O request is simply a command in the queue. The type of command or the number ofblocks that are associated with the command are not considered. The least queue depth with subsetpolicy treats large block requests and small block requests equally. The data path selected is one of thepaths in the path group of the controller that owns the volume.

Least path weight with subset

The least path weight with subset policy assigns a weight factor to each data path to a volume. An I/Orequest is routed to the path with the lowest weight value to the controller that owns the volume. If morethan one data path to the volume has the same weight value, the round-robin with subset path selectionpolicy is used to route I/ O requests between the paths with the same weight value.

Balancing the Fibre Channel I/O loadThe Total I/O data field in the Subsystem Management window is used for monitoring the Fibre ChannelI/O activity to a specific controller and a specific logical drive. This field helps you to identify possibleI/O hot spots.

You can identify actual Fibre Channel I/O patterns to the individual logical drives and compare thosewith the expectations based on the application. If a controller has more I/O activity than expected, movean array to the other controller in the storage subsystem by clicking Array → Change Ownership.

It is difficult to balance Fibre Channel I/O loads across controllers and logical drives because I/O loadsare constantly changing. The logical drives and the data that is accessed during the polling sessiondepends on which applications and users are active during that time period. It is important to monitorperformance during different time periods and gather data at regular intervals to identify performancetrends. The Performance Monitor enables you to save data to a comma-delimited text file that you canimport to a spreadsheet for further analysis.

If you notice that the workload across the storage subsystem (total Fibre Channel I/O statistic) continuesto increase over time while application performance decreases, you might need to add storagesubsystems to the enterprise.

H-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Optimizing the I/O transfer rateThe transfer rates of the controller are determined by the application I/O size and the I/O request rate. Asmall application I/O request size results in a lower transfer rate but provides a faster I/O request rateand a shorter response time. With larger application I/O request sizes, higher throughput rates arepossible. Understanding the application I/O patterns will help you optimize the maximum I/O transferrates that are possible for a given storage subsystem.

One of the ways to improve the I/O transfer rate is to improve the I/O request rate. Use thehost-computer operating system utilities to gather data about I/O size to understand the maximumtransfer rates possible. Then, use the tuning options that are available in DS Storage Manager to optimizethe I/O request rate to reach the maximum possible transfer rate.

Optimizing the Fibre Channel I/O request rateThe Fibre Channel I/O request rate can be affected by the following factors:v The Fibre Channel I/O access pattern (random or sequential) and I/O sizev The status of write-caching (enabled or disabled)v The cache-hit percentagev The RAID levelv The logical-drive modification priorityv The segment sizev The number of logical drives in the arrays or storage subsystemv The fragmentation of files

Note: Fragmentation affects logical drives with sequential Fibre Channel I/O access patterns, notrandom Fibre Channel I/O access patterns.

Determining the Fibre Channel I/O access pattern and I/O sizeTo determine if the Fibre Channel I/O access has sequential characteristics, enable a conservative cacheread-ahead multiplier (for example, 4) by clicking Logical Drive → Properties. Then, examine the logicaldrive cache-hit percentage to see if it has improved. An improvement indicates that the Fibre ChannelI/O has a sequential pattern. Use the host-computer operating-system utilities to determine the typicalI/O size for a logical drive.

Enabling write-cachingHigher Fibre Channel I/O write rates occur when write-caching is enabled, especially for sequential FibreChannel I/O access patterns. Regardless of the Fibre Channel I/O access pattern, be sure to enablewrite-caching to maximize the Fibre Channel I/O rate and shorten the application response time.

Optimizing the cache-hit percentageA higher cache-hit percentage is preferred for optimal application performance and is positivelycorrelated with the Fibre Channel I/O request rate.

If the cache-hit percentage of all logical drives is low or trending downward and you do not have themaximum amount of controller cache memory installed, you might need to install more memory.

If an individual logical drive has a low cache-hit percentage, you can enable cache read-ahead for thatlogical drive. Cache read-ahead can increase the cache-hit percentage for a sequential I/O workload. Ifcache read-ahead is enabled, the cache fetches more data, usually from adjacent data blocks on the drive.In addition to the requested data, this feature increases the chance that a future request for data isfulfilled from the cache, rather than requiring a logical drive access.

Appendix H. Tuning storage subsystems H-3

The cache read-ahead multiplier values specify the multiplier to use for determining how manyadditional data blocks are read into the cache. Choosing a higher cache read-ahead multiplier canincrease the cache-hit percentage.

If you determine that the Fibre Channel I/O access pattern has sequential characteristics, set anaggressive cache read-ahead multiplier (for example, 8). Then examine the logical-drive cache-hitpercentage to see if it has improved. Continue to customize logical-drive cache read-ahead to arrive at theoptimal multiplier. (For a random I/O pattern, the optimal multiplier is 0.)

Choosing appropriate RAID levelsUse the read percentage for a logical drive to determine the application behavior. Applications with ahigh read percentage perform well using RAID-5 logical drives because of the outstanding readperformance of the RAID-5 configuration.

Applications with a low read percentage (write-intensive) do not perform as well on RAID-5 logicaldrives because of the way that a controller writes data and redundancy data to the drives in a RAID-5logical drive. If there is a low percentage of read activity relative to write activity, you can change theRAID level of a logical drive from RAID-5 to RAID-1 for faster performance.

Choosing an optimal logical-drive modification priority settingThe modification priority defines how much processing time is allocated for logical-drive modificationoperations versus system performance. The higher the priority, the faster the logical-drive modificationoperations are completed, but the slower the system I/O access pattern is serviced.

Logical-drive modification operations include reconstruction, copyback, initialization, media scan,defragmentation, change of RAID level, and change of segment size. The modification priority is set foreach logical drive, using a slider bar from the Logical Drive - Properties window. There are five relativesettings on the reconstruction rate slider bar, ranging from Low to Highest. The actual speed of eachsetting is determined by the controller. Choose the Low setting to maximize the Fibre Channel I/Orequest rate. If the controller is idle (not servicing any I/O request rates) it ignores the individuallogical-drive rate settings and processes logical-drive modification operations as fast as possible.

Choosing an optimal segment sizeA segment is the amount of data, in KB, that the controller writes on a single logical drive before writingdata on the next drive. A data block is 512 bytes of data and is the smallest unit of storage. The size of asegment determines how many data blocks it contains. For example, an 8 KB segment holds 16 datablocks, and a 64 KB segment holds 128 data blocks.

Important: In Storage Manager version 7.01 and 7.02, the segment size is expressed in the number ofdata blocks. The segment size in DS Storage Manager is expressed in KB.

When you create a logical drive, the default segment size is a good choice for the expected logical-driveusage. To change the default segment size, click Logical Drive → Change Segment Size.

If the I/O size is larger than the segment size, increase the segment size to minimize the number ofdrives that are needed to satisfy an I/O request. This technique helps even more if you have random I/Oaccess patterns. If you use a single logical drive for a single request, it leaves other logical drivesavailable to simultaneously service other requests.

When you use the logical drive in a single-user, large I/O environment such as a multimedia application,storage performance is optimized when a single I/O request is serviced with a single array data stripe(which is the segment size multiplied by the number of logical drives in the array that are used for I/Orequests). In this case, multiple logical drives are used for the same request, but each logical drive isaccessed only once.


Defragmenting files to minimize disk accessEach time that you access a drive to read or write a file, it results in the movement of the read/writeheads. Verify that the files on the logical drive are defragmented. When the files are defragmented, thedata blocks that make up the files are next to each other, preventing extra read/write head movementwhen retrieving files. Fragmented files decrease the performance of a logical drive with sequential I/Oaccess patterns.

Appendix H. Tuning storage subsystems H-5


Appendix I. Critical event problem solving

When a critical event occurs, it is logged in the Event Log. It is also sent to any e-mail and SNMP trapdestinations that you have configured. The critical event type and the sense key/ASC/ASCQ data areboth shown in the event log details.

If a critical event occurs and you plan to call technical support, you can use the Customer SupportBundle feature to gather and package various pieces of data that can aid in remote troubleshooting.Perform the following steps to use the Customer Support Bundle feature:1. From the subsystem management window of the logical drive that is exhibiting problems, go to the

Advanced menu.2. Select Troubleshooting → Advanced → Collect All Support Data. The Collect All Support Data

window opens.3. Type the name of the file where you want to save the collected data or click browse to select the file.

Click Start.It takes several seconds for the zip file to be created depending on the amount of data to be collected.

4. Once the process completes, you can send the zip file electronically to customer support fortroubleshooting.

Table I-1 provides more information about events with a critical priority, as shown in the SubsystemManagement window event log.

Table I-1. Critical events

Critical event number Sense key/ASC/ASCQ Critical event description and required action

Event 1001 - Channel failed 6/3F/C3 Description: The controller failed a channel and cannotaccess drives on this channel any more. The FRU groupqualifier (byte 26) in the sense data indicates the relativechannel number of the failed channel. Typically thiscondition is caused by a drive ignoring the SCSI protocolon one of the controller destination channels. Thecontroller fails a channel if it issued a reset on a channeland continues to see the drives ignore the SCSI Bus Reseton this channel.

Action: Start the Recovery Guru to access the FailedDrive SCSI Channel recovery procedure. Contact yourIBM technical support representative to complete thisprocedure.

Event 1010 - Impendingdrive failure (PFA) detected

6/5D/80 Description: A drive has reported that a failureprediction threshold has been exceeded. This indicatesthat the drive might fail within 24 hours.

Action: Start the Recovery Guru and click the ImpendingDrive Failure recovery procedure. Follow the instructionsto correct the failure.

© Copyright IBM Corp. 2009, 2010 I-1

Table I-1. Critical events (continued)


Event 1015 - Incorrect modeparameters set on drive

6/3F/BD Description: The controller is unable to query the drivefor its current critical mode page settings or is unable tochange these settings to the correct setting. This indicatesthat the Qerr bit is set incorrectly on the drive specifiedin the FRU field of the Request Sense data.

Action: The controller has not failed yet. Contact yourIBM technical-support representative for the instructionsto recover from this critical event.

Event 1207 - Fibre-channellink errors - thresholdexceeded

None Description: Invalid characters have been detected in theFibre Channel signal. Possible causes for the error are adegraded laser in a gigabit interface converter (GBIC) ormedia interface adapter, damaged or faulty Fibre Channelcables, or poor cable connections between components onthe loop.

Action: In the main Subsystem Management window,click Help → Recovery Procedures. Click Fibre-channelLink Errors Threshold Exceeded for more informationabout recovering from this failure.

Event 1208 - Data ratenegotiation failed

None Description: The controller cannot auto-negotiate thetransfer link rates. The controller considers the link to bedown until negotiation is attempted at controllerstart-of-day, or when a signal is detected after a loss ofsignal.

Action: Start the Recovery Guru to access the Data RateNegotiation Failed recovery procedure and follow theinstructions to correct the failure.

Event 1209 - Drive channelset to Degraded

None Description: A drive channel status was set to Degradedbecause of excessive I/O errors or because a technicalsupport representative advised the arrays administratorto manually set the drive channel status for diagnostic orother support reasons.

Action: Start the Recovery Guru to access the DegradedDrive Channel recovery procedure and follow theinstructions to correct the failure.

Event 150E - Controllerloopback diagnostics failed

None Description: The controller cannot initialize thedrive-side Fibre Channel loops. A diagnostic routine hasbeen run identifying a controller problem and thecontroller has been placed offline. This event occurs onlyon certain controller models.

Action: Start the Recovery Guru to access the OfflineController recovery procedure and follow the instructionsto replace the controller.

Event 150F - Channelmiswire

None Description: Two or more drive channels are connectedto the same Fibre Channel loop. This can cause thestorage subsystem to behave unpredictably.

Action: Start the Recovery Guru to access the ChannelMiswire recovery procedure and follow the instructionsto correct the failure.

I-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide



Event 1510 - ESM canistermiswire

None Description: Two ESM canisters in the same storageexpansion enclosure are connected to the same FibreChannel loop. A level of redundancy has been lost andthe I/O performance for this storage expansion enclosureis reduced.

Action: Start the Recovery Guru to access the ESMCanister Miswire recovery procedure and follow theinstructions to correct the failure.

Event 1513 - IndividualDrive - Degraded Path

None Description: The specified drive channel is experiencingintermittent errors along the path to a single drive or toseveral drives.

Action: Start the Recovery Guru to access the IndividualDrive - Degraded Path recovery procedure and follow theinstructions to recover from this failure.

Event 1600 - Uncertifieddrive detected

None Description: An uncertified drive has been inserted intothe storage subsystem.

Action: Start the Recovery Guru to access the UncertifiedDrive recovery procedure and follow the instructions torecover from this failure.

Event 1601 - Reservedblocks on ATA drivescannot be discovered

None Description: Reserved blocks on the ATA drives are notrecognized.

Action: Contact technical support for instructions onrecovering from this event.

Event 200A - Data/paritymismatch detected onlogical drive

None Description: A media scan operation has detectedinconsistencies between a portion of the data blocks onthe logical drive and the associated parity blocks. Userdata in this portion of the logical drive might have beenlost.

Action: Select an application-specific tool (if available) toverify that the data is correct on the logical drive. If nosuch tool is available, or if problems with the user dataare reported, restore the entire logical drive contents fromthe most recent backup, if the data is critical.

Event 202E - Read driveerror during interruptedwrite

3/11/8A Description: A media error has occurred on a readoperation during interrupted write processing.

Action: Start the Recovery Guru to access theUnrecovered Interrupted Write recovery procedure.Contact your IBM technical support representative tocomplete this procedure.

Event 2109 - Controllercache not enabled - cachesizes do not match

6/A1/00 Description: The controller cannot enable mirroring if thealternate controller cache size of both controllers is notthe same. Verify that the cache size for both controllers isthe same.

Action: Contact your IBM technical supportrepresentative for the instructions to recover from thisfailure.

Appendix I. Critical event problem solving I-3



Event 210C - Controllercache battery failed

6/0C/80 Description: The controller has detected that the batteryis not physically present, is fully discharged, or hasreached its expiration date.

Action: Start the Recovery Guru to access the FailedBattery CRU recovery procedure and follow theinstructions to correct the failure.

Event 210E - Controllercache memory recoveryfailed after power cycle orreset

6/0C/81 Description: Recovery from a data-cache error wasunsuccessful. User data might have been lost.


Event 2110 - Controllercache memory initializationfailed

6/40/81 Description: The controller has detected the failure of aninternal controller component (RAID buffer). The internalcontroller component failure might have been detectedduring operation or during an on-board diagnosticroutine.


Event 2113 - Controllercache battery nearingexpiration

6/3F/D9 Description: The cache battery is within six weeks of itsexpiration.

Action: Start the Recovery Guru to access the BatteryNearing Expiration recovery procedure and follow theinstructions to correct the failure.

Event 211B - Batteriespresent but NVSRAMconfigured for no batteries

None Description: A battery is present in the storagesubsystem but the NVSRAM is set to not includebatteries.


Event 2229 - Drive failed bycontroller

None Description: The controller failed a drive because of aproblem with the drive.

Action: Start the Recovery Guru to access the DriveFailed by Controller procedure and follow theinstructions to correct the failure.

Event 222D - Drivemanually failed

6/3F/87 Description: The drive was manually failed by a user.

Action: Start the Recovery Guru to access the DriveManually Failed procedure and follow the instructions tocorrect the failure.

Event 2247 - Data lost onthe logical drive duringunrecovered interruptedwrite

6/3F/EB Description: An error has occurred during interruptedwrite processing during the start-of-day routine, whichcaused the logical drive to go into a failed state.

Action: Start the Recovery Guru to access theUnrecovered Interrupted Write recovery procedure andfollow the instructions to correct the failure. Contact yourIBM technical support representative to complete thisprocedure.




Event 2248 - Drive failed -write failure

6/3F/80 Description: The drive failed during a write command.The drive is marked failed.

Action: Start the Recovery Guru and follow theinstructions to correct the failure.

Event 2249 - Drive capacityless than minimum

6/3F/8B Description: During drive replacement, the capacity ofthe new drive is not large enough to support all thelogical drives that must be reconstructed on it.

Action: Replace the drive with a larger capacity drive.

Event 224A - Drive haswrong block size

6/3F/8C Description: The drive block size does not match that ofthe other drives in the logical drive. The drive is markedfailed.


Event 224B - Drive failed -initialization failure

6/3F/86 Description: The drive failed either from a Format Unitcommand or a Write operation (issued when a logicaldrive was initialized). The drive is marked failed.


Event 224D - Drive failed -no response at start of day

6/3F/85 Description: The drive failed a Read Capacity or Readcommand during the start-of-day routine. The controlleris unable to read the configuration information that isstored on the drive. The drive is marked failed.


Event 224E - Drive failed -initialization/reconstructionfailure

6/3F/82 Description: The previously-failed drive is marked failedbecause of one of the following reasons:

v The drive failed a Format Unit command that wasissued to it

v The reconstruction on the drive failed because thecontroller was unable to restore it (for example,because of an error that occurred on another drive thatwas required for reconstruction)


Event 2250 - Logical drivefailure

6/3F/E0 Description: The controller has marked the logical drivefailed. User data and redundancy (parity) can no longerbe maintained to ensure availability. The most likelycause is the failure of a single drive in nonredundantconfigurations or a nonredundant second drive in aconfiguration that is protected by one drive.

Action: Start the Recovery Guru to access the FailedLogical Drive Failure recovery procedure and follow theinstructions to correct the failure.

Event 2251 - Drive failed -reconstruction failure

6/3F/8E Description: A drive failed because of a reconstructionfailure during the start-of-day routine.





Event 2252 - Drive markedoffline during interruptedwrite

6/3F/98 Description: An error has occurred during interruptedwrite processing which caused the logical drive to bemarked failed. Drives in the array that did not experiencethe read error go into the offline state and log this error.

Action: Start the Recovery Guru to access theUnrecovered Interrupted Write recovery procedure.Contact your IBM technical support representative tocomplete this procedure.

Event 2254 - Redundancy(parity) and data mismatchis detected

6/8E/01 Description: The controller detected inconsistentredundancy (parity) or data during a parity verification.


Event 2255 - Logical drivedefinition incompatible withALT mode - ALT disabledNote: This event is notapplicable for the DS4800.

6/91/3B Description: Auto-LUN transfer (ALT) works only witharrays that have only one logical drive defined. Currentlythere are arrays on the storage subsystem that have morethan one logical drive defined; therefore, ALT mode hasbeen disabled. The controller operates in normalredundant controller mode, and if there is a problem, ittransfers all logical drives on an array instead oftransferring individual logical drives.


Event 2260 - Uncertifieddrive

ASC/ASCQ: None Description: A drive in the storage subsystem isuncertified.

Action: Start the Recovery Guru to access the UncertifiedDrive recovery procedure.

Event 2602 - Automaticcontroller firmwaresynchronization failed

02/04/81 Description: The versions of firmware on the redundantcontrollers are not the same because the automaticcontroller firmware synchronization failed. Controllerswith an incompatible version of the firmware mightcause unexpected results.

Action: Try the firmware download again. If the problempersists, contact your IBM technical supportrepresentative.

Event 2801 - Storagesubsystem running onuninterruptible powersupply battery

6/3F/C8 Description: The uninterruptible power supply hasindicated that ac power is no longer present and theuninterruptible power supply has switched to standbypower. While there is no immediate cause for concern,you should save your data frequently, in case the batteryis suddenly depleted.

Action: Start the Recovery Guru and click the Lost ACPower recovery procedure. Follow the instructions tocorrect the failure.




Event 2803 - Uninterruptiblepower supply battery - twominutes to failure

6/3F/C9 Description: The uninterruptible power supply hasindicated that its standby power supply is nearingdepletion.

Action: Take actions to stop I/O activity to the controller.Normally, the controller changes from a write-backcaching mode to a write-through mode.

Event 2804 - Uninterruptiblepower supply battery failed

None Description: The uninterruptible power supply batteryhas failed.


Event 2807 - Environmentalservice module failed

None Description: An ESM has failed.

Action: Start the Recovery Guru and click the FailedEnvironmental Service Module CRU recovery procedure.Follow the instructions to correct the failure.

Event 2808 - Enclosure IDnot unique

6/98/01 Description: The controller has determined that there aremultiple storage expansion enclosures with the same IDselected. Verify that each storage expansion enclosure hasa unique ID setting.

Action: Start the Recovery Guru and click the EnclosureID Conflict recovery procedure. Follow the instructions tocorrect the failure.

Event 280A - Controllerenclosure componentmissing

6/3F/C7 Description: A component other than a controller ismissing in the controller enclosure (for example, a fan,power supply, or battery). The FRU codes indicate thefaulty component.


Event 280B - Controllerenclosure component failed

6/3F/C7 Description: A component other than a controller hasfailed in the controller enclosure (for example, a fan,power supply, battery), or an over-temperature conditionhas occurred. The FRU codes indicate the faultycomponent.


Event 280D - Driveexpansion enclosurescomponent failed

6/3F/C7 Description: A component other than a drive has failedin the storage expansion enclosure (for example, a fan,power supply, or battery), or an over-temperaturecondition has occurred. The FRU codes indicate thefaulty component.


Event 280E - Standby powersupply not fully charged

6/3F/CA Description: The uninterruptible power supply hasindicated that its standby power supply is not at fullcapacity.

Action: Check the uninterruptible power supply to makesure that the standby power source (battery) is inworking condition.




Event 280F - Environmentalservice module - loss ofcommunication

6/E0/20 Description: Communication has been lost to one of thedual ESM CRUs in a storage expansion enclosure. Thestorage expansion enclosure has only one I/O pathavailable.


Event 2813 - Minihub CRUfailed

6/3F/C7 Description: Communication with the minihub CRU hasbeen lost. This might be the result of a minihub CRUfailure, a controller failure, or a failure in an internalbackplane communications board. If there is only oneminihub failure, the storage subsystem is still operational,but a second minihub failure could result in the failure ofthe affected enclosures.


Event 2815 - GBIC failed None Description: A gigabit interface converter (GBIC) oneither the controller enclosure or the storage expansionenclosure has failed. If there is only one GBIC failure, thestorage subsystem is still operational, but a second GBICfailure could result in the failure of the affectedenclosures.


Event 2816 - Enclosure IDconflict - duplicate IDsacross storage expansionenclosures

6/98/01 Description: Two or more storage expansion enclosuresare using the same enclosure identification number.


Event 2818 - Enclosure IDmismatch - duplicate IDs inthe same storage expansionenclosure

6/98/02 Description: A storage expansion enclosure in the storagesubsystem contains ESMs with different enclosureidentification numbers.


Event 281B - Nominaltemperature exceeded

6/98/03 Description: The nominal temperature of the enclosurehas been exceeded. Either a fan has failed or thetemperature of the room is too high. If the temperature ofthe enclosure continues to rise, the affected enclosuremight automatically shut down. Fix the problemimmediately, before it becomes more serious. Theautomatic shutdown conditions depend on the model ofthe enclosure.





Event 281C- Maximumtemperature exceeded

6/3F/C6 Description: The maximum temperature of the enclosurehas been exceeded. Either a fan has failed or thetemperature of the room is too high. This condition iscritical and might cause the enclosure to shut down ifyou do not fix the problem immediately. The automaticshutdown conditions depend on the model of theenclosure.


Event 281D - Temperaturesensor removed

6/98/03 Description: A fan CRU containing a temperature sensorhas been removed from the storage subsystem.

Action: Replace the CRU as soon as possible. Start theRecovery Guru and click the Failed or Removed FanCRU recovery procedure and follow the instructions tocorrect the failure.

Event 281E - Environmentalservice module firmwaremismatch

6/98/03 Description: A storage expansion enclosure in the storagesubsystem contains ESMs with different versions offirmware. ESMs in the same storage expansion enclosuremust have the same version firmware. If you do not havea replacement service monitor, call your IBM technicalsupport representative to perform the firmwaredownload.

Action: Start the Recovery Guru and click theEnvironmental Service Module Firmware VersionMismatch recovery procedure. Follow the instructions tocorrect the failure.

Event 2821 - Incompatibleminihub

None Description: An incompatible minihub canister has beendetected in the controller enclosure.

Action: Start the Recovery Guru and click theIncompatible Minihub Canister recovery procedure.Follow the instructions to correct the failure.

Event 2823 - Drive bypassed None Description: The ESM has reported that the drive hasbeen bypassed to maintain the integrity of the FibreChannel loop.

Action: Start the Recovery Guru to access the By-PassedDrive recovery procedure and follow the instructions torecover from this failure.

Event 2827 - Controller wasinadvertently replaced withan ESM

None Description: A controller canister was inadvertentlyreplaced with an ESM canister.

Action: Replace the ESM canister with the controllercanister as soon as possible.




Event 2828 - Unsupportedstorage expansion enclosureselected

None Description: Your storage subsystem contains one ormore unsupported drive enclosures. If all of your driveenclosures are being detected as being unsupported, youmight have a problem with an NVSRAM configurationfile or you might have the wrong version of firmware.This error condition will cause the drives in theunsupported expansion enclosures to be locked out,which can cause the defined arrays or logical drives tofail.

Action: If there are array or logical drive failures, callIBM support for the recovery procedure. Otherwise, Startthe Recovery Guru to access the Unsupported DriveEnclosure recovery procedure and follow the instructionsto recover from this failure.

Event 2829 - Controllerredundancy lost

6/E0/20 Description: Communication has been lost between thetwo controllers through one of the drive loops (channels).

Action: Start the Recovery Guru and see if there areother loss of redundancy problems being reported. Ifthere are other problems being reported, fix those first. Ifyou continue to have redundancy problems beingreported, contact the IBM technical supportrepresentative.

Event 282B - storageexpansion enclosure pathredundancy lost

6/E0/20 Description: A storage expansion enclosure withredundant drive loops (channels) has lost communicationthrough one of its loops. The enclosure has only one loopavailable for I/O. Correct this failure as soon as possible.Although the storage subsystem is still operational, alevel of path redundancy has been lost. If the remainingdrive loop fails, all I/O to that enclosure fails.

Action: Start the Recovery Guru and click the Drive -Loss of Path Redundancy recovery procedure. Follow theinstructions to correct the failure.

Event 282D - Drive pathredundancy lost

6/E0/20 Description: A communication path with a drive hasbeen lost. Correct this failure as soon as possible. Thedrive is still operational, but a level of path redundancyhas been lost. If the other port on the drive or any othercomponent fails on the working channel, the drive fails.

Action: Start the Recovery Guru and click the Drive -Loss of Path Redundancy recovery procedure. Follow theinstructions to correct the failure.

Event 282F - Incompatibleversion of ESM firmwaredetected

None Description: A storage expansion enclosure in the storagesubsystem contains ESM canisters with differentfirmware versions. This error might also be reported if astorage expansion enclosure in the storage subsystemcontains ESM canisters with different hardware.

Action: Start the Recovery Guru to access the ESMCanister Firmware Version Mismatch recovery procedureand follow the instructions to recover from this failure.




Event 2830 - Mixed drivetypes not supported

None Description: The storage subsystem currently containsdrives of different drive technologies, such as FibreChannel (FC) and Serial ATA (SATA). Mixing differentdrive technologies is not supported on this storagesubsystem.

Action: Select the Recovery Guru to access the MixedDrive Types Not Supported recovery procedure andfollow the instructions to recover from this failure.

Event 2835 - Driveexpansion enclosures notcabled together

ASC/ASCQ: None Description: There are drive expansion enclosures in thestorage subsystem that are not cabled correctly; they haveESM canisters that must be cabled together sequentially.

Action: Start the Recovery Guru to access the DriveEnclosures Not Cabled Together recovery procedure andfollow the instructions to recover from this failure.

Event 3019 - Logical driveownership changed due tofailover

None Description: The multipath driver software has changedownership of the logical drives to the other controllerbecause it could not access the logical drives on theparticular path.

Action: Start the Recovery Guru and click the LogicalDrive Not on Preferred Path recovery procedure. Followthe instructions to correct the failure.

Event 4011 - Logical drivenot on preferred path

None Description: The controller listed in the Recovery Guruarea cannot be accessed. Any logical drives that have thiscontroller assigned as their preferred path will be movedto the non-preferred path (alternate controller).

Action: Start the Recovery Guru and click the LogicalDrive Not on Preferred Path recovery procedure. Followthe instructions to correct the failure.

Event 5005 - Place controlleroffline

None Description: The controller is placed offline. This couldbe caused by the controller failing a diagnostic test. (Thediagnostics are initiated internally by the controller or bythe Controller → Run Diagnostics menu option.) Or thecontroller is manually placed Offline using the Controller→ Place Offline menu option.

Action: Start the Recovery Guru and click the OfflineController recovery procedure. Follow the instructions toreplace the controller.

Event 502F - Missing logicaldrive deleted

None Description: The storage subsystem has detected that thedrives that are associated with a logical drive are nolonger accessible. This can be the result of removing allthe drives that are associated with an array or a loss ofpower to one or more storage expansion enclosures.

Action: Start the Recovery Guru and click the MissingLogical Drive recovery procedure. Follow the instructionsto correct the failure.




Event 5038 - Controller inlockout mode

None Description: Both controllers have been placed in lockoutmode for 10 minutes because password authenticationfailures have exceeded 10 attempts within a ten-minuteperiod. During the lockout period, both controllers willdeny all authentication requests. When the 10-minutelockout expires, the controller resets the totalauthentication failure counter and unlocks itself.

Action: Wait 10 minutes and try to enter the passwordagain.

Event 5040 - Place controllerin service mode

None Description: The controller was manually placed inservice mode for diagnostic or recovery reasons.

Action: Start the Recovery Guru to access the Controllerin Service Mode recovery procedure. Use this procedureto place the controller back online.

Event 5405 - Gold Key -mismatched settings

ASC/ASCQ: None Description: Each controller in the controller pair has adifferent NVSRAM bit setting that determines if thecontroller is subject to Gold Key restrictions.

Action: This critical event should not be seen in the IBMDS3000, DS4000, or DS5000 Storage Subsystemconfiguration. This event could be generated if there is aninadvertent swapping of IBM Storage Subsystemcontrollers or drives with non-IBM controllers or drives.Contact IBM Support for the recovery procedure.

Event 5406 - Mixed drivetypes - mismatched settings

ASC/ASCQ: None Description: Each controller in the controller pair has adifferent setting for the NVSRAM bit that controlswhether Mixed Drive Types is a premium feature.

Action: Start the Recovery Guru to access the MixedDrive Types - Mismatched Settings recovery procedureand follow the instructions to correct this controllercondition.

Event 5602 - Thiscontroller's alternate failed -timeout waiting for results

None Description: This controller initiated diagnostics on thealternate controller but did not receive a reply indicatingthat the diagnostics were completed. The alternatecontroller in this pair has been placed offline.


Event 560B - CtlrDiag taskcannot obtain Mode Selectlock

None Description: This controller is attempting to rundiagnostics and could not secure the test area from otherstorage subsystem operations. The diagnostics werecanceled.





Event 560C - CtlrDiag taskon controller's alternatecannot obtain Mode

None Description: The alternate controller in this pair isattempting to run diagnostics and could not secure thetest area from other storage subsystem operations. Thediagnostics were canceled.


Event 560D - Diagnosticsread test failed on controller

None Description: While running diagnostics, the controllerdetected that the information that was received does notmatch the expected return for the test. This couldindicate that I/O is not completing or that there is amismatch in the data that is being read. The controller isplaced offline as a result of this failure.


Event 560E - Thiscontroller's alternate faileddiagnostics read test

None Description: While running diagnostics, the alternate forthis controller detected that the information received doesnot match the expected return for the test. This couldindicate that I/O is not completing or that there is amismatch in the data that is being read. The alternatecontroller in this pair is placed offline.


Event 560F - Diagnosticswrite test failed oncontroller

None Description: While running diagnostics, the alternate forthis controller is unable to write data to the test area.This could indicate that I/O is not being completed orthat there is a mismatch in the data that is being written.The controller is placed offline.


Event 5610 - Thiscontroller's alternate faileddiagnostics write test

None Description: While running diagnostics, the alternate forthis controller is unable to write data to the test area.This could indicate that I/O is not being completed orthat there is a mismatch in the data that is being written.The alternate controller in this pair is placed offline.


Event 5616 - Diagnosticsrejected - configurationerror on controller

None Description: The alternate for this controller isattempting to run diagnostics and could not create thetest area necessary for the completion of the tests. Thediagnostics were canceled.





Event 5617 - Diagnosticsrejected - configurationerror on controller'salternate

None Description: The alternate for this controller isattempting to run diagnostics and could not create thetest area necessary for the completion of the tests. Thediagnostics were canceled.

Action: Contact your IBM technical-supportrepresentative for the instructions to recover from thisfailure.

Event 6101 - Internalconfiguration database full

None Description: Because of the amount of data that isrequired to store certain configuration data, themaximum number of logical drives has beenunderestimated. One or both of the following types ofdata might have caused the internal configurationdatabase to become full:

v FlashCopy logical drive configuration data

v Global/Metro remote mirror configuration data

Action: To recover from this event, you can delete one ormore FlashCopy logical drives from your storagesubsystem or you can remove one or more remote mirrorrelationships.

Event 6107 - The alternatefor the controller isnonfunctional and is beingheld in reset

None Description: A controller in the storage subsystem hasdetected that its alternate controller is nonfunctional dueto hardware problems and needs to be replaced.

Action: Start the Recovery Guru to access the OfflineController recovery procedure and follow the instructionsto recover from this failure.

Event 6200 - FlashCopyrepository logical drivethreshold exceeded

None Description: The FlashCopy repository logical drivecapacity has exceeded a warning threshold level. If thecapacity of the FlashCopy repository logical drivebecomes full, its associated FlashCopy logical drive canfail. This is the last warning that you receive before theFlashCopy repository logical drive becomes full.

Action: Start the Recovery Guru and click the FlashCopyRepository Logical Drive Threshold Exceeded recoveryprocedure. Follow the instructions to correct this failure.

Event 6201 - FlashCopyrepository logical drive full

None Description: All of the available capacity on theFlashCopy repository logical drive has been used. Thefailure policy of the FlashCopy repository logical drivedetermines what happens when the FlashCopy repositorylogical drive becomes full. The failure policy can be set toeither fail the FlashCopy logical drive (default setting) orfail incoming I/Os to the base logical drive.

Action: Start the Recovery Guru and click the FlashCopyRepository Logical Drive Capacity - Full recoveryprocedure. Follow the instructions to correct this failure.




Event 6202 - FailedFlashCopy logical drive

None Description: Either the FlashCopy repository logicaldrive that is associated with the FlashCopy logical driveis full or its associated base or FlashCopy repositorylogical drives have failed due to one or more drivefailures on their respective arrays.

Action: Start the Recovery Guru and click the FailedFlashCopy Logical Drive recovery procedure. Follow theinstructions to correct this failure.

Event 6400 - Dual primarylogical drive

None Description: Both logical drives have been promoted to aprimary logical drive after a forced role reversal. Thisevent might be reported when the controller resets orwhen a cable from an array to a Fibre Channel switch isreinserted after it was removed and the other logicaldrive was promoted to a primary logical drive.

Action: Start the Recovery Guru and click the DualPrimary Logical Drive Conflict recovery procedure.Follow the instructions to correct this failure.

Event 6401 - Dualsecondary logical drive

None Description: Both logical drives in the remote mirrorhave been demoted to secondary logical drives after aforced role reversal. This could be reported when thecontroller resets or when a cable from an array to a FibreChannel switch is reinserted after it was removed andthe other logical drive was promoted to a secondarylogical drive.

Action: Start the Recovery Guru and click the DualSecondary Logical Drive Conflict recovery procedure.Follow the instructions to correct this failure.

Event 6402 - Mirror dataunsynchronized

Not recorded with event Description: This might occur because of I/O errors butthere should be other events associated with it. One ofthe other errors is the root cause, that contains the sensedata. A Needs Attention icon displays on both theprimary and secondary storage subsystems of the remotemirror.

Action: Start the Recovery Guru and click the MirrorData Unsynchronized recovery procedure. Follow theinstructions to correct this failure.

Event 6503 - Remote logicaldrive link down

None Description: This event is triggered when either a cablebetween one array and its peer has been disconnected,the Fibre Channel switch has failed, or the peer array hasreset. This error could result in the Mirror DataUnsynchronized, event 6402. The affected remote logicaldrive displays an Unresponsive icon, and this state willbe selected in the tooltip when you pass your cursor overthe logical drive.

Action: Start the Recovery Guru and click the MirrorCommunication Error - Unable to Contact Logical Driverecovery procedure. Follow the instructions to correct thisfailure.




Event 6505 - WWN changefailed

None Description: Mirroring causes a WWN change to becommunicated between arrays. Failure of a WWN changeis caused by non-I/O communication errors between onearray, on which the WWN has changed, and a peer array.(The array WWN is the unique name that is used tolocate an array on a fibre network. When both controllersin an array are replaced, the array WWN changes). Theaffected remote logical drive displays an Unresponsiveicon and this state will be selected in the tooltip whenyou pass your cursor over the logical drive.

Action: Start the Recovery Guru and click the Unable toUpdate Remote Mirror recovery procedure. Follow theinstructions to correct this failure. The only solution tothis problem is to delete the remote mirror and then toestablish another one.

Event 6600 - Logical drivecopy operation failed

None Description: A logical drive copy operation with a statusof In Progress has failed. This failure can be caused by aread error on the source logical drive, a write error onthe target logical drive, or because of a failure thatoccurred on the storage subsystem that affects the sourcelogical drive or target logical drive.

Action: Start the Recovery Guru and click the LogicalDrive Copy Operation Failed recovery procedure. Followthe instructions to correct this failure.

Event 6700 - Unreadablesector(s) detected - data lossoccurred

None Description: Unreadable sectors have been detected onone or more logical drives and data loss has occurred.

Action: Start the Recovery Guru to access the UnreadableSectors Detected recovery procedure and follow theinstructions to recover from this failure.

Event 6703 - Overflow inunreadable sector database

None Description: The Unreadable Sectors log has been filledto its maximum capacity.

Action: Select the Recovery Guru to access theUnreadable Sectors Log Full recovery procedure andfollow the instructions to recover from this failure.


Appendix J. Additional System Storage DS documentation

The following tables present an overview of the IBM System Storage DS Storage Manager, StorageSubsystem, and Storage Expansion Enclosure product libraries, as well as other related documents. Eachtable lists documents that are included in the libraries and what common tasks they address.

You can access the documents listed in these tables at both of the following Web sites:

www.ibm.com/servers/storage/support/disk/

www.ibm.com/shop/publications/order/

DS Storage Manager Version 10 library

Table J-1 associates each document in the DS Version 10 Storage Manager library with its related commonuser tasks.

Table J-1. DS Storage Manager Version 10 titles by user tasks

Title User tasks

Planning Hardwareinstallation

Softwareinstallation

Configuration Operation andadministration

Diagnosis andmaintenance

IBM System StorageDS Storage ManagerInstallation and HostSupport Guide (alloperating systems)

U U U

IBM System StorageDS Storage ManagerCommand LineInterface and ScriptCommandsProgramming Guide

U U U

IBM System StorageDS Storage ManagerCopy Services User'sGuide

U U U U

IBM System StorageDS4000 FibreChannel and SerialATA IntermixPremium FeatureInstallation Overview

U U U U

© Copyright IBM Corp. 2009, 2010 J-1

http://www.ibm.com/servers/storage/support/disk/

http://www.ibm.com/shop/publications/order/

DS5100 and DS5300 Storage Subsystem library

Table J-2 associates each document in the DS5100 and DS5300 Storage Subsystem library with its relatedcommon user tasks.

Table J-2. DS5100 and DS5300 Storage Subsystem document titles by user tasks

Title User Tasks

Planning HardwareInstallation

SoftwareInstallation

Configuration Operation andAdministration

Diagnosis andMaintenance

IBM System StorageDS5100 and DS5300Storage SubsystemInstallation, User'sand MaintenanceGuide

U U U U U

IBM System StorageQuick Start Guide,Quick Reference forDS5100 and DS5300Storage Subsystems,and for the EXP5000Storage ExpansionEnclosure

U U U

IBM System StorageDS5000 EXP5000Storage ExpansionEnclosure Installation,User's, andMaintenance Guide

Installing or replacinga DS5000 Cache andFlash Memory Card

U U U

Installing or replacinga DS5000 HostInterface Card

U U U

DS5020 Storage Subsystem library

Table J-3 associates each document in the DS5020 Storage Subsystem library with its related common usertasks.

Table J-3. DS5020 Storage Subsystem document titles by user tasks

Title User Tasks





IBM System StorageDS5020 StorageSubsystemInstallation, User'sand MaintenanceGuide

U U U U U

J-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Table J-3. DS5020 Storage Subsystem document titles by user tasks (continued)

Title User Tasks





IBM System StorageDS5020 Quick StartGuide

U U U


Installing or replacinga DS5000 Cache andFlash Memory Card

U U U

Installing or replacinga DS5000 HostInterface Card

U U U




Title User Tasks






U U U U U

IBM System StorageQuick Start Guide,Quick Reference forthe DS4800

U U U

IBM TotalStorage®

DS4800 ControllerCache Upgrade KitInstructions

U U U

Appendix J. Additional System Storage DS documentation J-3




Title User Tasks






U U U U U

IBM System StorageQuick Start Guide,Quick Reference forthe DS4700 andDS4200, Sections 2,3, and 4 also forinstalling the EXP810and EXP420

U U U


Table J-6 associates each document in the DS4500 (previously FAStT900) Storage Subsystem library withits related common user tasks.


Title User Tasks





IBM TotalStorageDS4500 StorageSubsystemInstallation, User's,and MaintenanceGuide

U U U U U

IBM TotalStorageDS4500 StorageSubsystem CablingInstructions

U U

IBM TotalStorageDS4500 RackMountingInstructions

U U





Title User Tasks





IBM TotalStorageDS4400 FibreChannel StorageServer User's Guide

U U U U U

IBM TotalStorageDS4400 FibreChannel StorageServer Installationand Support Guide

U U U U

IBM TotalStorageDS4400 FibreChannel CablingInstructions

U U




Title User Tasks





IBM TotalStorageDS4300 StorageSubsystemInstallation, User's,and MaintenanceGuide

U U U U U

IBM TotalStorageDS4300 RackMountingInstructions

U U

IBM TotalStorageDS4300 StorageSubsystem CablingInstructions

U U

IBM TotalStorageDS4300 SCU BaseUpgrade Kit

U U

IBM TotalStorageDS4300 SCU TurboUpgrade Kit

U U



Title User Tasks





IBM TotalStorageDS4300 Turbo Models6LU/6LX Upgrade Kit

U U

DS4200 Express Storage Subsystem library

Table J-9 associates each document in the DS4200 Express Storage™ Subsystem library with its relatedcommon user tasks.

Table J-9. DS4200 Express Storage Subsystem document titles by user tasks

Title User Tasks





IBM System StorageDS4200 ExpressStorage SubsystemInstallation, User'sand MaintenanceGuide

U U U U U


U U U




Title User Tasks





IBM TotalStorageDS4100 StorageServer Installation,User's andMaintenance Guide

U U U U U

IBM TotalStorageDS4100 StorageServer Cabling Guide

U



Table J-11 associates each document in the DS3500 Storage Subsystem library with its related commonuser tasks.


Title User Tasks





IBM System StorageDS3500 andEXP3500 Installation,User's, andMaintenance Guide

U U U U U

IBM System StorageDS3500 andEXP3500 RackInstallation and QuickStart Guide

U




Title User Tasks





IBM System StorageDS3400 Installation,User's, andMaintenance Guide

U U U U U


U




Title User Tasks






U U U U U



Title User Tasks






U


“DS3200 Storage Subsystem library” associates each document in the DS3200 Storage Subsystem librarywith its related common user tasks.


Title User Tasks






U U U U U


U

DS5000 Storage Expansion Enclosure documents

Table J-15 associates each of the following documents with its related common user tasks.

Table J-15. DS5000 Storage Expansion Enclosure document titles by user tasks

Title User Tasks






U U U U U

IBM System StorageQuick Start Guide,Quick Reference forDS5100 and DS5300Storage Subsystems,and for the EXP5000Storage ExpansionEnclosure

U U U


Table J-15. DS5000 Storage Expansion Enclosure document titles by user tasks (continued)

Title User Tasks





IBM System StorageDS5000 Hard Driveand StorageExpansion EnclosureInstallation andMigration Guide

U U

IBM System StorageEXP5060 StorageExpansion EnclosureInstallation, User's,and MaintenanceGuide

U U U U U

IBM System StorageQuick Start Guide,Quick Reference forthe EXP5060 StorageExpansion Enclosure

U U U


DS5020 Storage Expansion Enclosure document titles by user tasks associates each of the followingdocuments with its related common user tasks.


Title User Tasks





IBM Storage DS5020System StorageInstallation, User's,and MaintenanceGuide

U U U U U

IBM Storage DS5020Quick Start Guide

U U U

IBM Storage DS5020EXP520 StorageExpansion EnclosureInstallation, User's,and MaintenanceGuide

U U U U U



DS3950 Storage Expansion Enclosure document titles by user tasks associates each of the followingdocuments with its related common user tasks.


Title User Tasks





IBM Storage DS3950System StorageInstallation, User's,and MaintenanceGuide

U U U U U

IBM Storage DS3950Quick Start Guide

U U U

IBM Storage DS3950EXP39520 StorageExpansion EnclosureInstallation, User's,and Maintenance

U U U U U




Title User Tasks






U U U U U


U U U

IBM TotalStorageDS4000 EXP700 andEXP710 StorageExpansion EnclosuresInstallation, User's,and MaintenanceGuide

U U U U U


Table J-18. DS4000 Storage Expansion Enclosure document titles by user tasks (continued)

Title User Tasks





IBM DS4000 EXP500Installation andUser's Guide

U U U U U


U U U U U

IBM System StorageDS4000 Hard Driveand StorageExpansion EnclosuresInstallation andMigration Guide

U U

Other DS and DS-related documents


Table J-19. DS4000 and DS4000–related document titles by user tasks

Title User Tasks





IBM SafetyInformation

U

IBM TotalStorageDS4000 HardwareMaintenance Manual¹

U

IBM System StorageDS4000 ProblemDetermination Guide

U

IBM Fibre ChannelPlanning andIntegration: User'sGuide and ServiceInformation

U U U U

IBM TotalStorageDS4000 FC2-133Host Bus AdapterInstallation andUser's Guide

U U

IBM TotalStorageDS4000 FC2-133Dual Port Host BusAdapter Installationand User's Guide

U U


Table J-19. DS4000 and DS4000–related document titles by user tasks (continued)

Title User Tasks





IBM Netfinity® FibreChannel CablingInstructions

U

IBM Fibre ChannelSAN ConfigurationSetup Guide

U U U U

Note: The IBM TotalStorage DS4000 Hardware Maintenance Manual does not contain maintenance information for theIBM System Storage DS4100, DS4200, DS4300, DS4500, DS4700, or DS4800 storage subsystems. You can findmaintenance information for these products in the IBM System Storage DSx000 Storage Subsystem Installation, User's,and Maintenance Guide for the particular subsystem.


Appendix K. Accessibility

This section provides information about alternate keyboard navigation, which is a DS Storage Manageraccessibility feature. Accessibility features help a user who has a physical disability, such as restrictedmobility or limited vision, to use software products successfully.

By using the alternate keyboard operations that are described in this section, you can use keys or keycombinations to perform Storage Manager tasks and initiate many menu actions that can also be donewith a mouse.

Note: In addition to the keyboard operations that are described in this section, the DS Storage Managerversion 9.14 - 10.10 (and later) software installation packages for Windows include a screen readersoftware interface.

To enable the screen reader, select Custom Installation when using the installation wizard to installStorage Manager 9.14 - 10.10 (or later) on a Windows host/management station. Then, in the SelectProduct Features window, select Java Access Bridge, in addition to the other required host softwarecomponents.

Table K-1 defines the keyboard operations that enable you to navigate, select, or activate user interfacecomponents. The following terms are used in the table:v Navigate means to move the input focus from one user interface component to another.v Select means to choose one or more components, typically for a subsequent action.v Activate means to carry out the action of a particular component.

Note: In general, navigation between components requires the following keys:v Tab - Moves keyboard focus to the next component or to the first member of the next group of

componentsv Shift-Tab - Moves keyboard focus to the previous component or to the first component in the previous

group of componentsv Arrow keys - Move keyboard focus within the individual components of a group of components

Table K-1. DS3000 and DS4000 Storage Manager alternate keyboard operations

Short cut Action

F1 Open the Help.

F10 Move keyboard focus to main menu bar and post first menu; use thearrow keys to navigate through the available options.

Alt+F4 Close the management window.

Alt+F6 Move keyboard focus between dialogs (non-modal) and betweenmanagement windows.

Alt+ underlined letter Access menu items, buttons, and other interface components by usingthe keys associated with the underlined letters.

For the menu options, select the Alt + underlined letter combination toaccess a main menu, and then select the underlined letter to access theindividual menu item.

For other interface components, use the Alt + underlined lettercombination.

Ctrl+F1 Display or conceal a tool tip when keyboard focus is on the toolbar.

© Copyright IBM Corp. 2009, 2010 K-1

Table K-1. DS3000 and DS4000 Storage Manager alternate keyboard operations (continued)

Short cut Action

Spacebar Select an item or activate a hyperlink.

Ctrl+Spacebar(Contiguous/Non-contiguous)AMW Logical/Physical View

Select multiple drives in the Physical View.

To select multiple drives, select one drive by pressing Spacebar, andthen press Tab to switch focus to the next drive you want to select;press Ctrl+Spacebar to select the drive.

If you press Spacebar alone when multiple drives are selected then allselections are removed.

Use the Ctrl+Spacebar combination to deselect a drive when multipledrives are selected.

This behavior is the same for contiguous and non-contiguous selectionof drives.

End, Page Down Move keyboard focus to the last item in the list.

Esc Close the current dialog. Does not require keyboard focus.

Home, Page Up Move keyboard focus to the first item in the list.

Shift+Tab Move keyboard focus through components in the reverse direction.

Ctrl+Tab Move keyboard focus from a table to the next user interface component.

Tab Navigate keyboard focus between components or select a hyperlink.

Down arrow Move keyboard focus down one item in the list.

Left arrow Move keyboard focus to the left.

Right arrow Move keyboard focus to the right.

Up arrow Move keyboard focus up one item in the list.

The publications for this product are in Adobe® Portable Document Format (PDF) and should becompliant with accessibility standards. If you experience difficulties when you use the PDF files and wantto request a Web-based format or accessible PDF document for a publication, direct your mail to thefollowing address:v Information Development

v IBM Corporation

v 205/A015

v 3039 E. Cornwallis Road

v P.O. Box 12195

v Research Triangle Park, North Carolina 27709-2195

v U.S.A.

In the request, be sure to include the publication part number and title.

When you send information to IBM, you grant IBM a nonexclusive right to use or distribute theinformation in any way it believes appropriate without incurring any obligation to you.

K-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Appendix L. FDE best practices

This appendix describes best practices for maintaining security on a DS3000 and DS5000 storagesubsystem that is equipped with self-encrypting disk drives. Topics include:v “Physical asset protection”v “Data backup”v “FDE drive security key and the security key file” on page L-2v “DS subsystem controller shell remote login” on page L-3v “Working with FDE drives” on page L-3v “Replacing controllers” on page L-3v “Storage industry standards and practices” on page L-4

Physical asset protectionFull disk encryption (FDE) protects the data on the disks only when the drives are removed from storagesubsystem enclosures. FDE does not prevent someone from copying the data in the storage subsystemsthrough Fibre Channel host port connections when the drives are unlocked and operating. FDE also doesnot prevent unauthorized management access.

To prevent unauthorized access to the storage subsystem, take the following precautions:v Place the subsystem behind a closed and locked door.v Secure (disable) any unused Fibre Channel or Ethernet switch ports.v Place the storage subsystem management network on a separate subnet from the production Ethernet

network.v Assign a subsystem management password for each storage subsystem that is managed by the IBM DS

Storage Manager client.

The DS Storage Manager caches the subsystem management password and reuses it for every operationthat requires it without prompting you to reenter the password. You must establish strong user accesscontrols in the system that hosts the DS Storage Manager client. User login and password must beenabled, and the host system must be configured so that the screen automatically becomes locked if it isnot used.

The DS storage subsystem comes with a null storage manager client subsystem management password asa default. Assign a storage manager Subsystem Management password to each managed storagesubsystem before you configure drives or arrays. If the storage manager Subsystem Management windowpassword remains null, any computer in the storage subsystem management network can connect to thestorage subsystem and perform malicious acts such as deleting arrays or mapping logical drives tounauthorized hosts.

Data backupAlways back up the data in the storage subsystem to secured tape to prevent loss of data due tomalicious acts, natural disasters, abnormal hardware failures, or loss of the FDE security key.

© Copyright IBM Corp. 2009, 2010 L-1

FDE drive security key and the security key fileThe security key is used to unlock secured full disk encryption (FDE) drives for read and writeoperations. The security key is obfuscated inside the controllers and is used to unlock the drives whenthey are turned on. A backup copy of the encrypted security key is stored in a user-specified locationwhen the key is initially generated or whenever you change the security key. The security key can bechanged only when all of the defined arrays in the storage subsystem are in Optimal state. It is notnecessary to change the security key at the same regular intervals as password changes for the server andFibre Channel/Ethernet switch.

In the rare event that the controllers encounter a corrupted security key, the stored copy of the key isused to unlock the drives and restore the security key that is stored in the controllers. You can also usethe stored copy of the security key to unlock secured FDE drives when you migrate them to anotherstorage subsystem.

The security key that is stored in the security key file is protected by a pass phrase. To protect thesecurity key, the pass phrase and the security key are not stored separately in the security key file. Thesecurity key is wrapped by the hashed pass phrase before it is written to the security key file; then, thepass phrase is discarded. A security key identifier is paired with the security key to help you rememberwhich key to use for a particular storage subsystem. The security key identifier can contain up to 189alphanumeric characters. The storage subsystem worldwide identifier and a randomly generated numberare appended to the security key identifier when it is written to the security key file.

There is no way to unlock a secured FDE drive without the security key. If the security key that is storedin the subsystem controllers becomes corrupted, the only way to recover the security key in thecontrollers is to use the stored security key file. The only way to decrypt the information in the securitykey file to retrieve the security key is to use the pass phrase. If the pass phrase does not match the oneyou specified when you saved the security key file, there is no way to retrieve the security key from thesecurity key file and unlock the secured FDE drives. In this case, the data in the secured FDE drives willbe lost.

When you save the security key file, a copy of the security key file is stored in the location that youspecify and in the default security file directory c:\Program Files\IBM_DS\client\data\securityLockKeyin the Microsoft Windows operating system environment or /var/opt/SM/securityLockkey in the AIX,Linux, Solaris and HP-UX operating system environments. The default directory security key file acts as abackup file in case your copy of the security key file is lost or damaged. If the security key file name thatyou specify is the same as an existing security key file name, the name of the security key file that issaved in the default directory is modified to include the date and time. For example, file nameds5ktop2.slk becomes ds5ktop2_2009_06_01_11_00_36.slk.

By design, the security key can be changed only when all of the defined arrays in the storage subsystemare in Optimal state. The change security key process will terminate if there are any degraded arrays inthe storage subsystem. If there are any disconnected, failed, or exported spun-down drives in the storagesubsystem when the security key is changed, the security key for these drives will not change. Failed orexported spun-down drives will have the n-1 drive security key value, where n is the number of timesthe security key was changed since the drives failed, were exported or spun down. Because these drivesare already turned on, they can become operational even though they have a different security key thanthe rest of the FDE drives in the storage subsystem. However, these drives might not unlock for read andwrite operations after the power is cycled because the controllers do not have the security key for thesedrives. These drives can be unlocked only when the security key is provided.

Take the following precautions to protect the security key and pass phrase:v To prevent situations in which there are multiple security keys in the storage subsystem, do not change

the security key when there are any disconnected, failed, or exported-down drives in the subsystem.

L-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

v If you must change the security key while there are disconnected, failed, or exported spun-downdrives, be sure to first save a security key file for these drives before you change the security key.

v Because the pass phrase is used to decrypt the wrapped security key that is stored in the security keyfile, be careful not to create a pass phrase that can be easily discovered by unauthorized users.

v Record the pass phrase, the security key identifier, and the name and location of the security key fileand store this information in a secure place. If you lose this information, you could lose access to yourdata.

v Do not store the security key file on a logical drive that you created by using the disk drives of thesame DS storage subsystem.

v Do not write the pass phrase on the media that is used to store the security key file.v Before you run the command to export a set of drives from the subsystem, change the security key and

save the security key file and pass phrase for the exported drives in a secure place. Make sure that youcan identify the correct security key file and pass phrase for the exported drives.

v Store more than one copy of the security key file. Do not specify the default security file directory asthe location to store your copy of the security key file. If you specify the default directory, only onecopy of the security key file will be saved.

DS subsystem controller shell remote loginDo not enable remote login to the DS subsystem controller shell. The controller shell is reserved for IBMsupport personnel to perform subsystem recovery and diagnostics procedures. The DS subsystemcontroller shell does not provide prompts to confirm potentially destructive tasks. Mistakes in commandsyntax or the inappropriate use of controller shell commands might result in lost access or lost data.

Working with FDE drivesTo maintain security on FDE drives, take the following precautions:v After you delete an array, secure erase the FDE drives that were part of the array. Secure erase

cryptographically erases the disk and also changes the drive state back to Unsecured.v Use only an FDE global hot-spare drive as a spare for a drive failure in a secured array. A Fibre

Channel FDE global hot-spare drive can be used as a spare drive for Fibre Channel FDE drives insecure or nonsecure arrays and as a spare for a Fibre Channel drive in arrays with non-FDE drives.

v Although you can use an array with non-FDE drives as the repository logical drive for a FlashCopyimage of a secured FDE logical drive or as the target logical drive of a secured VolumeCopy sourceFDE logical drive, do not use non-FDE drives, because arrays with non-FDE drives are not secured.

v In a remote mirroring configuration, even if the primary and secondary logical drives of a mirroredlogical drive pair are secured, the data that is transferred between the primary and secondary storagesubsystems is not encrypted. You must ensure that the data that is transferred between remotemirroring logical drive pairs is protected.

Replacing controllersWhen you replace a controller, take the following precautions:v Do not replace both controller field replaceable units (FRUs) at the same time. If you replace both

controllers at the same time, the new controllers will not have the correct security key to unlock thedrives. If this occurs, you must use the saved security key file to unlock the drives.

v Do not replace a controller FRU while the storage subsystem is powered down. If you replace acontroller FRU with the power off, the new controller and the existing controller will have differentsecurity keys, causing security key mismatch errors. Perform the change security key procedure toresynchronize the security key between controllers.

Appendix L. FDE best practices L-3

v Replace controller FRUs only when the subsystem is powered on and operational. This ensures that thefirmware version, premium feature IDs, and security key in the new controller are synchronized withthe existing controller.

Storage industry standards and practicesSee the following documents for storage and networking industry best practices and guidelines:v Storage Networking Interface Association (SNIA):

– For SNIA key management best practices charts, see http://www.snia.org/images/tutorial_docs/Security/WaltHubis-Best_Practices_Secure_Storage.pdf.

– For SNIA guidance and best practices documents, see http://www.snia.org/forums/ssif/programs/best_practices. Registration is required to download the following SNIA documents:- Best Current Practices: Provides broad guidance for organizations seeking to secure their storage

systems- SSIF Solutions Guide to Data at Rest: Provides baseline considerations and guidance for some of the

factors you should consider when evaluating storage security– For the SNIA Storage Security forum, see http://www.snia.org/forums/ssif/.

v National Institute of Standards and Technology (NIST):– See the NIST Computer Security Division Guidelines for Media Sanitation at http://csrc.nist.gov/

publications/nistpubs/800-88/NISTSP800-88_rev1.pdf.v Trusted Computing Group:

– https://www.trustedcomputinggroup.org/home.

L-4 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

http://www.snia.org/images/tutorial_docs/Security/WaltHubis-Best_Practices_Secure_Storage.pdf

http://www.snia.org/images/tutorial_docs/Security/WaltHubis-Best_Practices_Secure_Storage.pdf

http://www.snia.org/forums/ssif/programs/best_practices

http://www.snia.org/forums/ssif/programs/best_practices

http://www.snia.org/forums/ssif/

http://csrc.nist.gov/publications/nistpubs/800-88/NISTSP800-88_rev1.pdf

http://csrc.nist.gov/publications/nistpubs/800-88/NISTSP800-88_rev1.pdf

https://www.trustedcomputinggroup.org/home

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 inyour area. Any reference to an IBM product, program, or service is not intended to state or imply thatonly 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 isthe user's responsibility to evaluate and verify the operation of any non-IBM product, program, orservice.

IBM may have patents or pending patent applications covering subject matter described in thisdocument. The furnishing of this document does not grant you any license to these patents. You can sendlicense inquiries, in writing, to:

IBM Director of LicensingIBM CorporationNorth Castle DriveArmonk, NY 10504-1785U.S.A.

For license inquiries regarding double-byte (DBCS) information, contact the IBM Intellectual PropertyDepartment in your country or send inquiries, in writing, to:

IBM World Trade Asia CorporationLicensing2-31 Roppongi 3-chome, Minato-kuTokyo 106-0032, Japan

The following paragraph does not apply to the United Kingdom or any other country where suchprovisions are inconsistent with local law: INTERNATIONAL BUSINESS MACHINES CORPORATIONPROVIDES THIS PUBLICATIONS "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSOR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OFNON-INFRINGEMENT, MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Somestates do not allow disclaimer of express or implied warranties in certain transactions, therefore, thisstatement may not apply to you.

This information could include technical inaccuracies or typographical errors. Changes are periodicallymade to the information herein; these changes will be incorporated in new editions of the publications.IBM may make improvements or changes (or both) in the product(s) or program(s) (or both), described inthis publication at any time without notice.

Any references in this information to non-IBM Web sites are provided for convenience only and do not inany manner serve as an endorsement of those Web sites. The materials at those Web sites are not part ofthe materials for this IBM product and use of those Web sites is at your own risk.

Information concerning non-IBM products was obtained from the suppliers of those products, theirpublished announcements or other publicly available sources. IBM has not tested those products andcannot confirm the accuracy of performance, compatibility or any other claims related to non-IBMproducts. Questions on the capabilities of non-IBM products should be addressed to the suppliers ofthose products.

© Copyright IBM Corp. 2009, 2010 M-1

IBM may use or distribute any of the information you supply in any way it believes appropriate withoutincurring any obligation to you.

Some software may differ from its retail version (if available), and may not include user manuals or allprogram functionality.

Trademarks

IBM, the IBM logo, and ibm.com® are trademarks or registered trademarks of International BusinessMachines Corporation in the United States, other countries, or both. If these and other IBM trademarkedterms are marked on their first occurrence in this information with a trademark symbol (® or ™), thesesymbols indicate U.S. registered or common law trademarks owned by IBM at the time this informationwas published. Such trademarks may also be registered or common law trademarks in other countries. Acurrent list of IBM trademarks is available on the Web at “Copyright and trademark information” athttp://www.ibm.com/legal/copytrade.shtml.

Adobe and PostScript® are either registered trademarks or trademarks of Adobe Systems Incorporated inthe United States and/or other countries.

Cell Broadband Engine™ is a trademark of Sony Computer Entertainment, Inc., in the United States, othercountries, or both and is used under license therefrom.

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

Java and all Java-based trademarks are trademarks of Sun Microsystems, Inc., in the United States, othercountries, or both.

Linux is a registered trademark of Linus Torvalds in the United States, other countries, or both.

Microsoft, Windows, and Windows NT are trademarks of Microsoft Corporation 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, or service names may be trademarks or service marks of others.

Important notes

Processor speed indicates the internal clock speed of the microprocessor; other factors also affectapplication performance.

CD or DVD drive speed is the variable read rate. Actual speeds vary and are often less than the possiblemaximum.

When referring to processor storage, real and virtual storage, or channel volume, KB stands for 1024bytes, MB stands for 1 048 576 bytes, and GB stands for 1 073 741 824 bytes.

When referring to hard disk drive capacity or communications volume, MB stands for 1 000 000 bytes,and GB stands for 1 000 000 000 bytes. Total user-accessible capacity can vary depending on operatingenvironments.

Maximum internal hard disk drive capacities assume the replacement of any standard hard disk drivesand population of all hard disk drive bays with the largest currently supported drives that are availablefrom IBM.

M-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

http://www.ibm.com/legal/copytrade.shtml

Maximum memory might require replacement of the standard memory with an optional memorymodule.

IBM makes no representation or warranties regarding non-IBM products and services that areServerProven®, including but not limited to the implied warranties of merchantability and fitness for aparticular purpose. These products are offered and warranted solely by third parties.

IBM makes no representations or warranties with respect to non-IBM products. Support (if any) for thenon-IBM products is provided by the third party, not IBM.

Some software might differ from its retail version (if available) and might not include user manuals or allprogram functionality.

Particulate contamination

Attention: Airborne particulates (including metal flakes or particles) and reactive gases acting alone orin combination with other environmental factors such as humidity or temperature might pose a risk tothe storage subsystem that is described in this document. Risks that are posed by the presence ofexcessive particulate levels or concentrations of harmful gases include damage that might cause thestorage subsystem to malfunction or cease functioning altogether. This specification sets forth limits forparticulates and gases that are intended to avoid such damage. The limits must not be viewed or used asdefinitive limits, because numerous other factors, such as temperature or moisture content of the air, caninfluence the impact of particulates or environmental corrosives and gaseous contaminant transfer. In theabsence of specific limits that are set forth in this document, you must implement practices that maintainparticulate and gas levels that are consistent with the protection of human health and safety. If IBMdetermines that the levels of particulates or gases in your environment have caused damage to thestorage subsystem, IBM may condition provision of repair or replacement of storage subsystem or partson implementation of appropriate remedial measures to mitigate such environmental contamination.Implementation of such remedial measures is a customer responsibility.

Table M-1. Limits for particulates and gases

Contaminant Limits

Particulate v The room air must be continuously filtered with 40% atmospheric dust spot efficiency (MERV 9)according to ASHRAE Standard 52.21.

v Air that enters a data center must be filtered to 99.97% efficiency or greater, using high-efficiencyparticulate air (HEPA) filters that meet MIL-STD-282.

v The deliquescent relative humidity of the particulate contamination must be more than 60%2.

v The room must be free of conductive contamination such as zinc whiskers.

Gaseous v Copper: Class G1 as per ANSI/ISA 71.04-19853

v Silver: Corrosion rate of less than 300 Å in 30 days1 ASHRAE 52.2-2008 - Method of Testing General Ventilation Air-Cleaning Devices for Removal Efficiency by Particle Size.Atlanta: American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.

2 The deliquescent relative humidity of particulate contamination is the relative humidity at which the dust absorbsenough water to become wet and promote ionic conduction.

3 ANSI/ISA-71.04-1985. Environmental conditions for process measurement and control systems: Airborne contaminants.Instrument Society of America, Research Triangle Park, North Carolina, U.S.A.

Notices M-3

M-4 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Glossary

This glossary provides definitions for theterminology and abbreviations used in IBMSystem Storage publications.

If you do not find the term you are looking for,see the IBM Glossary of Computing Terms located atthe following Web site:

http://www.ibm.com/ibm/terminology

This glossary also includes terms and definitionsfrom:v Information Technology Vocabulary by

Subcommittee 1, Joint Technical Committee 1,of the International Organization forStandardization and the InternationalElectrotechnical Commission (ISO/IECJTC1/SC1). Definitions are identified by thesymbol (I) after the definition; definitions takenfrom draft international standards, committeedrafts, and working papers by ISO/IECJTC1/SC1 are identified by the symbol (T) afterthe definition, indicating that final agreementhas not yet been reached among theparticipating National Bodies of SC1.

v IBM Glossary of Computing Terms. New York:McGraw-Hill, 1994.

The following cross-reference conventions areused in this glossary:

See Refers you to (a) a term that is theexpanded form of an abbreviation oracronym, or (b) a synonym or morepreferred term.

See alsoRefers you to a related term.

Abstract Windowing Toolkit (AWT)A Java graphical user interface (GUI).

accelerated graphics port (AGP)A bus specification that gives low-cost 3Dgraphics cards faster access to mainmemory on personal computers than theusual peripheral component interconnect(PCI) bus. AGP reduces the overall cost ofcreating high-end graphics subsystems byusing existing system memory.

access volumeA special logical drive that allows the

host-agent to communicate with thecontrollers in the storage subsystem.

adapterA printed circuit assembly that transmitsuser data input/output (I/O) between theinternal bus of the host system and theexternal fibre-channel (FC) link and viceversa. Also called an I/O adapter, hostadapter, or FC adapter.

advanced technology (AT) bus architectureA bus standard for IBM compatibles. Itextends the XT bus architecture to 16 bitsand also allows for bus mastering,although only the first 16 MB of mainmemory are available for direct access.

agent A server program that receives virtualconnections from the network manager(the client program) in a Simple NetworkManagement Protocol-TransmissionControl Protocol/Internet Protocol(SNMP-TCP/IP) network-managingenvironment.

AGP See accelerated graphics port.

AL_PASee arbitrated loop physical address.

arbitrated loopOne of three existing fibre-channeltopologies, in which 2 - 126 ports areinterconnected serially in a single loopcircuit. Access to the Fibre ChannelArbitrated Loop (FC-AL) is controlled byan arbitration scheme. The FC-ALtopology supports all classes of serviceand guarantees in-order delivery of FCframes when the originator and responderare on the same FC-AL. The defaulttopology for the disk array is arbitratedloop. An arbitrated loop is sometimesreferred to as a Stealth Mode.

arbitrated loop physical address (AL_PA)An 8-bit value that is used to uniquelyidentify an individual port within a loop.A loop can have one or more AL_PAs.

array A collection of fibre-channel or SATA harddrives that are logically grouped together.All the drives in the array are assignedthe same RAID level. An array is

© Copyright IBM Corp. 2009, 2010 N-1

http://www.ibm.com/ibm/terminology

sometimes referred to as a "RAID set." Seealso redundant array of independent disks(RAID), RAID level.

asynchronous write modeIn remote mirroring, an option that allowsthe primary controller to return a writeI/O request completion to the host serverbefore data has been successfully writtenby the secondary controller. See alsosynchronous write mode, remote mirroring,Global Copy,Global Mirroring.

AT See advanced technology (AT) busarchitecture.

ATA See AT-attached.

AT-attachedPeripheral devices that are compatiblewith the original IBM AT computerstandard in which signals on a 40-pinAT-attached (ATA) ribbon cable followedthe timings and constraints of theIndustry Standard Architecture (ISA)system bus on the IBM PC AT computer.Equivalent to integrated drive electronics(IDE).

Auto Drive Transfer (ADT)A function that provides automaticfailover in case of controller failure on astorage subsystem.

ADT See Auto Drive Transfer .

AWT See Abstract Windowing Toolkit.

basic input/output system (BIOS)The personal computer code that controlsbasic hardware operations, such asinteractions with diskette drives, harddisk drives, and the keyboard.

BIOS See basic input/output system.

BOOTPSee bootstrap protocol.

bootstrap protocol (BOOTP)In Transmission Control Protocol/InternetProtocol (TCP/IP) networking, analternative protocol by which a disklessmachine can obtain its Internet Protocol(IP) address and such configurationinformation as IP addresses of variousservers from a BOOTP server.

bridge A storage area network (SAN) device thatprovides physical and transport

conversion, such as Fibre Channel tosmall computer system interface (SCSI)bridge.

bridge groupA bridge and the collection of devicesconnected to it.

broadcastThe simultaneous transmission of data tomore than one destination.

cathode ray tube (CRT)A display device in which controlledelectron beams are used to displayalphanumeric or graphical data on anelectroluminescent screen.

client A computer system or process thatrequests a service of another computersystem or process that is typically referredto as a server. Multiple clients can shareaccess to a common server.

commandA statement used to initiate an action orstart a service. A command consists of thecommand name abbreviation, and itsparameters and flags if applicable. Acommand can be issued by typing it on acommand line or selecting it from amenu.

community stringThe name of a community contained ineach Simple Network ManagementProtocol (SNMP) message.

concurrent downloadA method of downloading and installingfirmware that does not require the user tostop I/O to the controllers during theprocess.

CRC See cyclic redundancy check.

CRT See cathode ray tube.

CRU See customer replaceable unit.

customer replaceable unit (CRU)An assembly or part that a customer canreplace in its entirety when any of itscomponents fail. Contrast with fieldreplaceable unit (FRU).

cyclic redundancy check (CRC)(1) A redundancy check in which thecheck key is generated by a cyclicalgorithm. (2) An error detectiontechnique performed at both the sendingand receiving stations.

N-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

dac See disk array controller.

dar See disk array router.

DASDSee direct access storage device.

data stripingSee striping.

default host groupA logical collection of discovered hostports, defined host computers, anddefined host groups in thestorage-partition topology that fulfill thefollowing requirements:v Are not involved in specific logical

drive-to-LUN mappingsv Share access to logical drives with

default logical drive-to-LUN mappings

device typeIdentifier used to place devices in thephysical map, such as the switch, hub, orstorage.

DHCP See Dynamic Host Configuration Protocol.

direct access storage device (DASD)A device in which access time iseffectively independent of the location ofthe data. Information is entered andretrieved without reference to previouslyaccessed data. (For example, a disk driveis a DASD, in contrast with a tape drive,which stores data as a linear sequence.)DASDs include both fixed and removablestorage devices.

direct memory access (DMA)The transfer of data between memory andan input/output (I/O) device withoutprocessor intervention.

disk array controller (dac)A disk array controller device thatrepresents the two controllers of an array.See also disk array router.

disk array router (dar)A disk array router that represents anentire array, including current anddeferred paths to all logical unit numbers(LUNs) (hdisks on AIX). See also diskarray controller.

DMA See direct memory access.

domainThe most significant byte in the node port(N_port) identifier for the fibre-channel

(FC) device. It is not used in the FibreChannel-small computer system interface(FC-SCSI) hardware path ID. It is requiredto be the same for all SCSI targetslogically connected to an FC adapter.

drive channelsThe DS4200, DS4700, and DS4800subsystems use dual-port drive channelsthat, from the physical point of view, areconnected in the same way as two driveloops. However, from the point of view ofthe number of drives and enclosures, theyare treated as a single drive loop insteadof two different drive loops. A group ofstorage expansion enclosures areconnected to the DS3000 or DS4000storage subsystems using a drive channelfrom each controller. This pair of drivechannels is referred to as a redundantdrive channel pair.

drive loopsA drive loop consists of one channel fromeach controller combined to form one pairof redundant drive channels or aredundant drive loop. Each drive loop isassociated with two ports. (There are twodrive channels and four associated portsper controller.) For the DS4800, driveloops are more commonly referred to asdrive channels. See drive channels.

DRAMSee dynamic random access memory.

Dynamic Host Configuration Protocol (DHCP)A protocol defined by the InternetEngineering Task Force that is used fordynamically assigning Internet Protocol(IP) addresses to computers in a network.

dynamic random access memory (DRAM)A storage in which the cells requirerepetitive application of control signals toretain stored data.

ECC See error correction coding.

EEPROMSee electrically erasable programmableread-only memory.

EISA See Extended Industry Standard Architecture.

electrically erasable programmable read-onlymemory (EEPROM)

A type of memory chip which can retainits contents without consistent electricalpower. Unlike the PROM which can be

Glossary N-3

programmed only once, the EEPROM canbe erased electrically. Because it can onlybe reprogrammed a limited number oftimes before it wears out, it is appropriatefor storing small amounts of data that arechanged infrequently.

electrostatic discharge (ESD)The flow of current that results whenobjects that have a static charge come intoclose enough proximity to discharge.

environmental service module (ESM) canisterA component in a storage expansionenclosure that monitors the environmentalcondition of the components in thatenclosure. Not all storage subsystemshave ESM canisters.

E_port See expansion port.

error correction coding (ECC)A method for encoding data so thattransmission errors can be detected andcorrected by examining the data on thereceiving end. Most ECCs arecharacterized by the maximum number oferrors they can detect and correct.

ESD See electrostatic discharge.

ESM canisterSee environmental service module canister.

automatic ESM firmware synchronizationWhen you install a new ESM into anexisting storage expansion enclosure in aDS3000 or DS4000 storage subsystem thatsupports automatic ESM firmwaresynchronization, the firmware in the newESM is automatically synchronized withthe firmware in the existing ESM.

EXP See storage expansion enclosure.

expansion port (E_port)A port that connects the switches for twofabrics.

Extended Industry Standard Architecture (EISA)A bus standard for IBM compatibles thatextends the Industry StandardArchitecture (ISA) bus architecture to 32bits and allows more than one centralprocessing unit (CPU) to share the bus.See also Industry Standard Architecture.

fabric A Fibre Channel entity whichinterconnects and facilitates logins ofN_ports attached to it. The fabric isresponsible for routing frames between

source and destination N_ports usingaddress information in the frame header.A fabric can be as simple as apoint-to-point channel between twoN-ports, or as complex as a frame-routingswitch that provides multiple andredundant internal pathways within thefabric between F_ports.

fabric port (F_port)In a fabric, an access point for connectinga user's N_port. An F_port facilitatesN_port logins to the fabric from nodesconnected to the fabric. An F_port isaddressable by the N_port connected to it.See also fabric.

FC See Fibre Channel.

FC-AL See arbitrated loop.

feature enable identifierA unique identifier for the storagesubsystem, which is used in the processof generating a premium feature key. Seealso premium feature key.

Fibre Channel (FC)A set of standards for a serialinput/output (I/O) bus capable oftransferring data between two ports at upto 100 Mbps, with standards proposals togo to higher speeds. FC supportspoint-to-point, arbitrated loop, andswitched topologies.

Fibre Channel Arbitrated Loop (FC-AL)See arbitrated loop.

Fibre Channel Protocol (FCP) for small computersystem interface (SCSI)

A high-level fibre-channel mapping layer(FC-4) that uses lower-level fibre-channel(FC-PH) services to transmit SCSIcommands, data, and status informationbetween a SCSI initiator and a SCSI targetacross the FC link by using FC frame andsequence formats.

field replaceable unit (FRU)An assembly that is replaced in itsentirety when any one of its componentsfails. In some cases, a field replaceableunit might contain other field replaceableunits. Contrast with customer replaceableunit (CRU).


FlashCopyA premium feature for that can make aninstantaneous copy of the data in avolume.

F_port See fabric port.

FRU See field replaceable unit.

GBIC See gigabit interface converter

gigabit interface converter (GBIC)A transceiver that performs serial,optical-to-electrical, andelectrical-to-optical signal conversions forhigh-speed networking. A GBIC can behot swapped. See also small form-factorpluggable.

Global CopyRefers to a remote logical drive mirrorpair that is set up using asynchronouswrite mode without the write consistencygroup option. This is also referred to as"Asynchronous Mirroring withoutConsistency Group." Global Copy doesnot ensure that write requests to multipleprimary logical drives are carried out inthe same order on the secondary logicaldrives as they are on the primary logicaldrives. If it is critical that writes to theprimary logical drives are carried out inthe same order in the appropriatesecondary logical drives, Global Mirroringshould be used instead of Global Copy.See also asynchronous write mode, GlobalMirroring, remote mirroring, MetroMirroring.

Global MirroringRefers to a remote logical drive mirrorpair that is set up using asynchronouswrite mode with the write consistencygroup option. This is also referred to as"Asynchronous Mirroring withConsistency Group." Global Mirroringensures that write requests to multipleprimary logical drives are carried out inthe same order on the secondary logicaldrives as they are on the primary logicaldrives, preventing data on the secondarylogical drives from becoming inconsistentwith the data on the primary logicaldrives. See also asynchronous write mode,Global Copy, remote mirroring, MetroMirroring.

graphical user interface (GUI)A type of computer interface that presents

a visual metaphor of a real-world scene,often of a desktop, by combininghigh-resolution graphics, pointing devices,menu bars and other menus, overlappingwindows, icons, and the object-actionrelationship.

GUI See graphical user interface.

HBA See host bus adapter.

hdisk An AIX term representing a logical unitnumber (LUN) on an array.

heterogeneous host environmentA host system in which multiple hostservers, which use different operatingsystems with their own unique diskstorage subsystem settings, connect to thesame storage subsystem at the same time.See also host.

host A system that is directly attached to thestorage subsystem through a fibre-channelinput/output (I/O) path. This system isused to serve data (typically in the formof files) from the storage subsystem. Asystem can be both a storage managementstation and a host simultaneously.

host bus adapter (HBA)An interface between the fibre-channelnetwork and a workstation or server.

host computerSee host.

host groupAn entity in the storage partitiontopology that defines a logical collectionof host computers that require sharedaccess to one or more logical drives.

host portPorts that physically reside on the hostadapters and are automatically discoveredby the Storage Manager software. To givea host computer access to a partition, itsassociated host ports must be defined.

hot swapTo replace a hardware component withoutturning off the system.

hub In a network, a point at which circuits areeither connected or switched. Forexample, in a star network, the hub is thecentral node; in a star/ring network, it isthe location of wiring concentrators.

IBMSAN driverThe device driver that is used in a Novell

Glossary N-5

NetWare environment to providemultipath input/output (I/O) support tothe storage controller.

IC See integrated circuit.

IDE See integrated drive electronics.

in-bandTransmission of management protocolover the fibre-channel transport.

Industry Standard Architecture (ISA)Unofficial name for the bus architecture ofthe IBM PC/XT personal computer. Thisbus design included expansion slots forplugging in various adapter boards. Earlyversions had an 8-bit data path, laterexpanded to 16 bits. The "ExtendedIndustry Standard Architecture" (EISA)further expanded the data path to 32 bits.See also Extended Industry StandardArchitecture.

initial program load (IPL)The initialization procedure that causes anoperating system to commence operation.Also referred to as a system restart,system startup, and boot.

integrated circuit (IC)A microelectronic semiconductor devicethat consists of many interconnectedtransistors and other components. ICs areconstructed on a small rectangle cut froma silicon crystal or other semiconductormaterial. The small size of these circuitsallows high speed, low power dissipation,and reduced manufacturing costcompared with board-level integration.Also known as a chip.

integrated drive electronics (IDE)A disk drive interface based on the 16-bitIBM personal computer Industry StandardArchitecture (ISA) in which the controllerelectronics reside on the drive itself,eliminating the need for a separateadapter card. Also known as anAdvanced Technology AttachmentInterface (ATA).

Internet Protocol (IP)A protocol that routes data through anetwork or interconnected networks. IPacts as an intermediary between thehigher protocol layers and the physicalnetwork.

Internet Protocol (IP) addressThe unique 32-bit address that specifiesthe location of each device or workstationon the Internet. For example, 9.67.97.103is an IP address.

interrupt request (IRQ)A type of input found on manyprocessors that causes the processor tosuspend normal processing temporarilyand start running an interrupt handlerroutine. Some processors have severalinterrupt request inputs that allowdifferent priority interrupts.

IP See Internet Protocol.

IPL See initial program load.

IRQ See interrupt request.

ISA See Industry Standard Architecture.

Java Runtime Environment (JRE)A subset of the Java Development Kit(JDK) for end users and developers whowant to redistribute the Java RuntimeEnvironment (JRE). The JRE consists ofthe Java virtual machine, the Java CoreClasses, and supporting files.

JRE See Java Runtime Environment.

label A discovered or user entered propertyvalue that is displayed underneath eachdevice in the Physical and Data Pathmaps.

LAN See local area network.

LBA See logical block address.

local area network (LAN)A computer network located on a user'spremises within a limited geographicarea.

logical block address (LBA)The address of a logical block. Logicalblock addresses are typically used inhosts' I/O commands. The SCSI diskcommand protocol, for example, useslogical block addresses.

logical partition (LPAR)A subset of a single system that containsresources (processors, memory, andinput/output devices). A logical partitionoperates as an independent system. Ifhardware requirements are met, multiplelogical partitions can exist within asystem.


A fixed-size portion of a logical volume.A logical partition is the same size as thephysical partitions in its volume group.Unless the logical volume of which it is apart is mirrored, each logical partitioncorresponds to, and its contents are storedon, a single physical partition.

One to three physical partitions (copies).The number of logical partitions within alogical volume is variable.

logical unit number (LUN)An identifier used on a small computersystem interface (SCSI) bus to distinguishamong up to eight devices (logical units)with the same SCSI ID.

loop addressThe unique ID of a node in fibre-channelloop topology sometimes referred to as aloop ID.

loop groupA collection of storage area network(SAN) devices that are interconnectedserially in a single loop circuit.

loop portA node port (N_port) or fabric port(F_port) that supports arbitrated loopfunctions associated with an arbitratedloop topology.

LPAR See logical partition.

LUN See logical unit number.

MAC See medium access control.

management information base (MIB)The information that is on an agent. It isan abstraction of configuration and statusinformation.

man pagesIn UNIX-based operating systems, onlinedocumentation for operating systemcommands, subroutines, system calls, fileformats, special files, stand-alone utilities,and miscellaneous facilities. Invoked bythe man command.

MCA See micro channel architecture.

media scanA media scan is a background processthat runs on all logical drives in thestorage subsystem for which it has beenenabled, providing error detection on thedrive media. The media scan processscans all logical drive data to verify that it

can be accessed, and optionally scans thelogical drive redundancy information.

medium access control (MAC)In local area networks (LANs), thesublayer of the data link control layer thatsupports medium-dependent functionsand uses the services of the physical layerto provide services to the logical linkcontrol sublayer. The MAC sublayerincludes the method of determining whena device has access to the transmissionmedium.

Metro MirroringThis term is used to refer to a remotelogical drive mirror pair which is set upwith synchronous write mode. See alsoremote mirroring, Global Mirroring.

MIB See management information base.

micro channel architecture (MCA)Hardware that is used for PS/2 Model 50computers and above to provide bettergrowth potential and performancecharacteristics when compared with theoriginal personal computer design.

Microsoft Cluster Server (MSCS)MSCS, a feature of Windows NT Server(Enterprise Edition), supports theconnection of two servers into a clusterfor higher availability and easiermanageability. MSCS can automaticallydetect and recover from server orapplication failures. It can also be used tobalance server workload and provide forplanned maintenance.

mini hubAn interface card or port device thatreceives short-wave fiber channel GBICsor SFPs. These devices enable redundantFibre Channel connections from the hostcomputers, either directly or through aFibre Channel switch or managed hub,over optical fiber cables to the DS3000and DS4000 Storage Server controllers.Each DS3000 and DS4000 controller isresponsible for two mini hubs. Each minihub has two ports. Four host ports (twoon each controller) provide a clustersolution without use of a switch. Twohost-side mini hubs are shipped asstandard. See also host port, gigabitinterface converter (GBIC), small form-factorpluggable (SFP).

Glossary N-7

mirroringA fault-tolerance technique in whichinformation on a hard disk is duplicatedon additional hard disks. See also remotemirroring.

model The model identification that is assignedto a device by its manufacturer.

MSCS See Microsoft Cluster Server.

network management station (NMS)In the Simple Network ManagementProtocol (SNMP), a station that runsmanagement application programs thatmonitor and control network elements.

NMI See non-maskable interrupt.

NMS See network management station.

non-maskable interrupt (NMI)A hardware interrupt that another servicerequest cannot overrule (mask). An NMIbypasses and takes priority over interruptrequests generated by software, thekeyboard, and other such devices and isissued to the microprocessor only indisastrous circumstances, such as severememory errors or impending powerfailures.

node A physical device that allows for thetransmission of data within a network.

node port (N_port)A fibre-channel defined hardware entitythat performs data communications overthe fibre-channel link. It is identifiable bya unique worldwide name. It can act asan originator or a responder.

nonvolatile storage (NVS)A storage device whose contents are notlost when power is cut off.

N_portSee node port.

NVS See nonvolatile storage.

NVSRAMNonvolatile storage random accessmemory. See nonvolatile storage.

Object Data Manager (ODM)An AIX proprietary storage mechanismfor ASCII stanza files that are edited aspart of configuring a drive into thekernel.

ODM See Object Data Manager.

out-of-bandTransmission of management protocolsoutside of the fibre-channel network,typically over Ethernet.

partitioningSee storage partition.

parity checkA test to determine whether the numberof ones (or zeros) in an array of binarydigits is odd or even.

A mathematical operation on thenumerical representation of theinformation communicated between twopieces. For example, if parity is odd, anycharacter represented by an even numberhas a bit added to it, making it odd, andan information receiver checks that eachunit of information has an odd value.

PCI local busSee peripheral component interconnect localbus.

PDF See portable document format.

performance eventsEvents related to thresholds set on storagearea network (SAN) performance.

peripheral component interconnect local bus(PCI local bus)

A local bus for PCs, from Intel, thatprovides a high-speed data path betweenthe CPU and up to 10 peripherals (video,disk, network, and so on). The PCI buscoexists in the PC with the IndustryStandard Architecture (ISA) or ExtendedIndustry Standard Architecture (EISA)bus. ISA and EISA boards plug into an IAor EISA slot, while high-speed PCIcontrollers plug into a PCI slot. See alsoIndustry Standard Architecture, ExtendedIndustry Standard Architecture.

polling delayThe time in seconds between successivediscovery processes during whichdiscovery is inactive.

port A part of the system unit or remotecontroller to which cables for externaldevices (such as display stations,terminals, printers, switches, or externalstorage units) are attached. The port is anaccess point for data entry or exit. Adevice can contain one or more ports.


portable document format (PDF)A standard specified by Adobe Systems,Incorporated, for the electronicdistribution of documents. PDF files arecompact; can be distributed globally bye-mail, the Web, intranets, or CD-ROM;and can be viewed with the AcrobatReader, which is software from AdobeSystems that can be downloaded at nocost from the Adobe Systems home page.

premium feature keyA file that the storage subsystemcontroller uses to enable an authorizedpremium feature. The file contains thefeature enable identifier of the storagesubsystem for which the premium featureis authorized, and data about thepremium feature. See also feature enableidentifier.

private loopA freestanding arbitrated loop with nofabric attachment. See also arbitrated loop.

program temporary fix (PTF)A temporary solution or bypass of aproblem diagnosed by IBM in a currentunaltered release of the program.

PTF See program temporary fix.

RAID See redundant array of independent disks(RAID).

RAID levelAn array's RAID level is a number thatrefers to the method used to achieveredundancy and fault tolerance in thearray. See also array, redundant array ofindependent disks (RAID).

RAID setSee array.

RAM See random-access memory.

random-access memory (RAM)A temporary storage location in which thecentral processing unit (CPU) stores andexecutes its processes. Contrast withDASD.

RDACSee redundant disk array controller.

read-only memory (ROM)Memory in which stored data cannot bechanged by the user except under specialconditions.

recoverable virtual shared disk (RVSD)A virtual shared disk on a server nodeconfigured to provide continuous accessto data and file systems in a cluster.

redundant array of independent disks (RAID)A collection of disk drives (array) thatappears as a single volume to the server,which is fault tolerant through anassigned method of data striping,mirroring, or parity checking. Each arrayis assigned a RAID level, which is aspecific number that refers to the methodused to achieve redundancy and faulttolerance. See also array, parity check,mirroring, RAID level, striping.

redundant disk array controller (RDAC)(1) In hardware, a redundant set ofcontrollers (either active/passive oractive/active). (2) In software, a layer thatmanages the input/output (I/O) throughthe active controller during normaloperation and transparently reroutes I/Osto the other controller in the redundantset if a controller or I/O path fails.

remote mirroringOnline, real-time replication of databetween storage subsystems that aremaintained on separate media. TheEnhanced Remote Mirror Option is apremium feature that provides supportfor remote mirroring. See also GlobalMirroring, Metro Mirroring.

ROM See read-only memory.

router A computer that determines the path ofnetwork traffic flow. The path selection ismade from several paths based oninformation obtained from specificprotocols, algorithms that attempt toidentify the shortest or best path, andother criteria such as metrics orprotocol-specific destination addresses.

RVSD See recoverable virtual shared disk.

SAI See Storage Array Identifier.

SA IdentifierSee Storage Array Identifier.

SAN See storage area network.

SATA See serial ATA.

scope Defines a group of controllers by theirInternet Protocol (IP) addresses. A scopemust be created and defined so that

Glossary N-9

dynamic IP addresses can be assigned tocontrollers on the network.

SCSI See small computer system interface.

segmented loop port (SL_port)A port that allows division of afibre-channel private loop into multiplesegments. Each segment can pass framesaround as an independent loop and canconnect through the fabric to othersegments of the same loop.

sense data(1) Data sent with a negative response,indicating the reason for the response. (2)Data describing an I/O error. Sense datais presented to a host system in responseto a sense request command.

serial ATAThe standard for a high-speed alternativeto small computer system interface (SCSI)hard drives. The SATA-1 standard isequivalent in performance to a 10 000RPM SCSI drive.

serial storage architecture (SSA)An interface specification from IBM inwhich devices are arranged in a ringtopology. SSA, which is compatible withsmall computer system interface (SCSI)devices, allows full-duplex packetmultiplexed serial data transfers at ratesof 20 Mbps in each direction.

server A functional hardware and software unitthat delivers shared resources toworkstation client units on a computernetwork.

server/device eventsEvents that occur on the server or adesignated device that meet criteria thatthe user sets.

SFP See small form-factor pluggable.

Simple Network Management Protocol (SNMP)In the Internet suite of protocols, anetwork management protocol that isused to monitor routers and attachednetworks. SNMP is an application layerprotocol. Information on devices managedis defined and stored in the application'sManagement Information Base (MIB).

SL_portSee segmented loop port.

SMagentThe Storage Manager optional Java-basedhost-agent software, which can be usedon Microsoft Windows, Novell NetWare,AIX, HP-UX, Solaris, and Linux onPOWER host systems to manage storagesubsystems through the host fibre-channelconnection.

SMclientThe Storage Manager client software,which is a Java-based graphical userinterface (GUI) that is used to configure,manage, and troubleshoot storage serversand storage expansion enclosures in astorage subsystem. SMclient can be usedon a host system or on a storagemanagement station.

SMruntimeA Java compiler for the SMclient.

SMutilThe Storage Manager utility software thatis used on Microsoft Windows, AIX,HP-UX, Solaris, and Linux on POWERhost systems to register and map newlogical drives to the operating system. InMicrosoft Windows, it also contains autility to flush the cached data of theoperating system for a particular drivebefore creating a FlashCopy.

small computer system interface (SCSI)A standard hardware interface thatenables a variety of peripheral devices tocommunicate with one another.

small form-factor pluggable (SFP)An optical transceiver that is used toconvert signals between optical fibercables and switches. An SFP is smallerthan a gigabit interface converter (GBIC).See also gigabit interface converter.

SNMPSee Simple Network Management Protocoland SNMPv1.

SNMP trap event

An event notification sent by the SNMPagent that identifies conditions, such asthresholds, that exceed a predeterminedvalue. See also Simple NetworkManagement Protocol.

SNMPv1The original standard for SNMP is nowreferred to as SNMPv1, as opposed to


SNMPv2, a revision of SNMP. See alsoSimple Network Management Protocol.

SRAMSee static random access memory.

SSA See serial storage architecture.

static random access memory (SRAM)Random access memory based on thelogic circuit know as flip-flop. It is calledstatic because it retains a value as long aspower is supplied, unlike dynamicrandom access memory (DRAM), whichmust be regularly refreshed. It is however,still volatile, meaning that it can lose itscontents when the power is turned off.

storage area network (SAN)A dedicated storage network tailored to aspecific environment, combining servers,storage products, networking products,software, and services. See also fabric.

Storage Array Identifier (SAI or SA Identifier)The Storage Array Identifier is theidentification value used by the StorageManager host software (SMClient) touniquely identify each managed storageserver. The Storage Manager SMClientprogram maintains Storage ArrayIdentifier records of previously-discoveredstorage servers in the host resident file,which allows it to retain discoveryinformation in a persistent fashion.

storage expansion enclosure (EXP)A feature that can be connected to asystem unit to provide additional storageand processing capacity.

storage management stationA system that is used to manage thestorage subsystem. A storage managementstation does not need to be attached tothe storage subsystem through thefibre-channel input/output (I/O) path.

storage partitionStorage subsystem logical drives that arevisible to a host computer or are sharedamong host computers that are part of ahost group.

storage partition topologyIn the Storage Manager client, theTopology view of the Mappings windowdisplays the default host group, thedefined host group, the host computer,and host-port nodes. The host port, host

computer, and host group topologicalelements must be defined to grant accessto host computers and host groups usinglogical drive-to-LUN mappings.

stripingSplitting data to be written into equalblocks and writing blocks simultaneouslyto separate disk drives. Stripingmaximizes performance to the disks.Reading the data back is also scheduledin parallel, with a block being readconcurrently from each disk thenreassembled at the host.

subnetAn interconnected but independentsegment of a network that is identified byits Internet Protocol (IP) address.

sweep methodA method of sending Simple NetworkManagement Protocol (SNMP) requestsfor information to all the devices on asubnet by sending the request to everydevice in the network.

switch A fibre-channel device that provides fullbandwidth per port and high-speedrouting of data by using link-leveladdressing.

switch groupA switch and the collection of devicesconnected to it that are not in othergroups.

switch zoningSee zoning.

synchronous write modeIn remote mirroring, an option thatrequires the primary controller to wait forthe acknowledgment of a write operationfrom the secondary controller beforereturning a write I/O request completionto the host. See also asynchronous writemode, remote mirroring, Metro Mirroring.

system nameDevice name assigned by the vendor'sthird-party software.

TCP See Transmission Control Protocol.

TCP/IPSee Transmission Control Protocol/InternetProtocol.

Glossary N-11

terminate and stay resident program (TSRprogram)

A program that installs part of itself as anextension of DOS when it is executed.

topologyThe physical or logical arrangement ofdevices on a network. The threefibre-channel topologies are fabric,arbitrated loop, and point-to-point. Thedefault topology for the disk array isarbitrated loop.

TL_portSee translated loop port.

transceiverA device that is used to transmit andreceive data. Transceiver is anabbreviation of transmitter-receiver.

translated loop port (TL_port)A port that connects to a private loop andallows connectivity between the privateloop devices and off loop devices (devicesnot connected to that particular TL_port).

Transmission Control Protocol (TCP)A communication protocol used in theInternet and in any network that followsthe Internet Engineering Task Force (IETF)standards for internetwork protocol. TCPprovides a reliable host-to-host protocolbetween hosts in packed-switchedcommunication networks and ininterconnected systems of such networks.It uses the Internet Protocol (IP) as theunderlying protocol.

Transmission Control Protocol/Internet Protocol(TCP/IP)

A set of communication protocols thatprovide peer-to-peer connectivityfunctions for both local and wide-areanetworks.

trap In the Simple Network ManagementProtocol (SNMP), a message sent by amanaged node (agent function) to amanagement station to report anexception condition.

trap recipientReceiver of a forwarded Simple NetworkManagement Protocol (SNMP) trap.Specifically, a trap receiver is defined byan Internet Protocol (IP) address and portto which traps are sent. Presumably, the

actual recipient is a software applicationrunning at the IP address and listening tothe port.

TSR programSee terminate and stay resident program.

uninterruptible power supplyA source of power from a battery that isinstalled between a computer system andits power source. The uninterruptiblepower supply keeps the system running ifa commercial power failure occurs, untilan orderly shutdown of the system can beperformed.

user action eventsActions that the user takes, such aschanges in the storage area network(SAN), changed settings, and so on.

worldwide port name (WWPN)A unique identifier for a switch on localand global networks.

worldwide name (WWN)A globally unique 64-bit identifierassigned to each Fibre Channel port.

WORMSee write-once read-many.

write-once read many (WORM)Any type of storage medium to whichdata can be written only a single time, butcan be read from any number of times.After the data is recorded, it cannot bealtered.

WWN See worldwide name.

zoningIn Fibre Channel environments, thegrouping of multiple ports to form avirtual, private, storage network. Portsthat are members of a zone cancommunicate with each other, but areisolated from ports in other zones.

A function that allows segmentation ofnodes by address, name, or physical portand is provided by fabric switches orhubs.


Index

Aabout this document xiaccess volume/logical drive 1-5, E-3Adapter settings B-1adding

devices 3-9address

of the IBM director of licensing M-1Advanced adapter settings B-2AIX 5-7

FCP disk array errors 5-37logical drives, redistributing in case of

failure 5-35AIX host

support xvAIX multipath driver 5-7alert notifications, setting up 3-10array 4-5arrays, creating 4-4Arrays, securing, frequently asked

questions 6-31attributes

dac E-5dar E-5hdisk 5-31, E-5LUN 5-31, E-5

audience xiAuto Drive Transfer (ADT) 5-36, 5-37automatic ESM firmware synchronization

defined 3-19Event Monitor requirement 3-19

automatic host discovery 3-8automatic storage subsystem

discovery 3-8

Bbackground media scan 4-19Backup and recovery, frequently asked

questions 6-33Best practices, data backup L-1Best practices, drive security key and the

security key file L-2Best practices, full disk encryption L-1Best practices, iSNS 3-13Best practices, physical asset

protection, L-1Best practices, replacing controllers L-3Best practices, storage industry standards

and practices L-4Best practices, subsystem controller shell

remote login L-3Best practices, working with FDE

drives L-3BIOS settings B-1Boot support, frequently asked

questions 6-33BOOTP server

sample network 1-3, 1-6

Ccache hit

optimizing H-3percentage H-3

cache mirroring 5-39, D-3cache read-ahead, choosing a

multiplier H-3check drive firmware level 3-19client software

package info 3-6cluster services

AIX requirements D-2, D-3HACMP ES and ESCRM D-1hardware requirements D-10HP-UX requirements D-9MC/Service Guard D-9PSSP with GPFS D-3Solaris requirements D-10system dependencies D-1

clusteringVMware ESX Server

configuration C-2collect support bundle manually 7-5concurrent firmware download 3-15,

3-19configuration recovery 4-4configuration types

installing Storage Manager 1-2configuring

devices 5-30direct-attached 1-5hosts 5-1RDAC driver D-10SAN-attached 1-5storage subsystems 1-5

Configuring DS5000 disk encryption withFDE drives 6-13

configuring hot-spare devices 4-4Configuring iSCSI host ports 3-13Configuring Storage Manager

Steps for defining a host group 4-10contamination, particulate and

gaseous M-3controller

transfer rate, optimizing H-3controller firmware

downloading 3-15controllers

IP addresses 1-6Copy Services Guide J-1creating

arrays and logical drives 4-4critical event

problem solving I-1Cross connections

VMware ESX Server C-4

Ddac (disk array controller)

and RDAC 5-28attributes E-5

dar (disk array router)and RDAC 5-28attributes E-5two dars on one subsystem,

causes 5-28data

redundancy 4-5Data backup, full disk encryption L-1DCE (dynamic capacity expansion) 5-33default host type, defining and

verifying 4-9defining default host type 4-9device drivers

description 5-3DMP, installing 5-24downloading latest versions xi, xiiimultipath

installing 5-6RDAC

description 5-15installing 5-22

with HACMP cluster D-3devices

identification 5-28identifying 5-27

devices, configuring hot-spares 4-4DHCP server

sample network 1-3, 1-6DHCP, using 3-13Diagnostic Data Capture

Recovery Guru F-1Script Editor F-1

direct-attached configuration 1-5disk access, minimize H-5disk array controller 5-28disk array router 5-28Disk drives, FDE 6-1Disk drives, global hot spare 6-30Disk drives, global hot spare, frequently

asked questions 6-33Disk drives, migrating secure 6-23Disk drives, unlocking 6-21DMP 5-15DMP DSM driver 5-7documentation

documents xiiiDS Storage Manager J-1DS3200 Storage Subsystem J-8DS3300 Storage Subsystem J-7DS3400 Storage Subsystem J-7DS3500 Storage Subsystem J-7DS4000 J-1DS4000-related documents J-11DS4100 SATA Storage Subsystem J-6DS4200 Express Storage

Subsystem J-6

© Copyright IBM Corp. 2009, 2010 X-1

documentation (continued)DS4300 Fibre Channel Storage

Subsystem J-5DS4400 Fibre Channel Storage

Subsystem J-5DS4500 Storage Subsystem J-4DS4700 Storage Subsystem J-4DS4800 Storage Subsystem J-3DS5000 J-1DS5020 Storage Subsystem J-2DS5100 and DS5300 Storage

Subsystem J-2Sun Solaris 5-15Web sites xiv

download firmware with I/O 3-19downloading controller and NVSRAM

ESM firmware 3-17downloading controller firmware 3-15downloading drive firmware 3-19downloading ESM firmware 3-18downloading NVSRAM firmware 3-15drive firmware

downloading 3-19drive firmware download 3-19drive firmware levels, determining 3-16drive firmware, level 3-19drivers xi, xiiiDS documentation J-1DS Storage Manager

documentation J-1DS3000 J-1DS3200 Storage Subsystem library J-8DS3300 Storage Subsystem library J-7DS3400 Storage Subsystem library J-7DS3500 Storage Subsystem library J-7DS4000

Hardware Maintenance Manual J-11Problem Determination Guide J-11Storage Expansion Enclosure

documentation J-10DS4000 documentation J-1DS4000 Storage Manager

related documents J-11DS4100

Storage Subsystem library J-6DS4200 Express

Storage Subsystem library J-6DS4300






Storage Expansion Enclosuredocumentation J-8

DS5000 documentation J-1DS5020

Storage Subsystem library J-2DS5100 and DS5300

Storage Subsystem library J-2DVE (dynamic volume expansion) 5-33dynamic capacity expansion (DCE) 5-33

Dynamic Multipathing (DMP)description 5-15installing 5-24installing the SMibmasl package 5-26installing Veritas VolumeManager

packages 5-26preparing for installation 5-25system requirements 5-24

dynamic volume expansion (DVE) 5-33

Eenabling

multipath I/O with PV-links 5-10Enabling full disk encryption 6-14Ending an iSCSI session 3-13Enterprise Management window

online help xiiEnterprise Management Window

adding devices 3-9alert notifications 3-10

Erase secure disk drive 6-10Erasing secure disk drives 6-27Erasing secured drives, frequently asked

questions 6-32errors, FCP disk array 5-37errors, media scan 4-20ESM firmware

automatic ESM firmwaredownload 3-18

automatic ESM firmwaresynchronization 3-19

downloading 3-18ESM firmware levels, determining 3-16Ethernet

Solaris requirements D-10Ethernet MAC address, identifying 1-6event log I-1expansion unit firmware levels,

determining 3-16

Ffabric switch environment B-13failover driver

description 5-3failure support

cluster services D-1DMP driver 5-15MPxIO 5-15multipath driver 5-3RDAC driver 5-15redistributing logical drives 5-35,

5-36, 5-37FCP disk array errors 5-37FDE 6-1FDE disk drives 6-1FDE disk drives, configuring full disk

encryption 6-13FDE disk drives, installing 6-13features

disabling 3-23features, premium

enabling 3-21feature enable identifier 3-21feature key file 3-22

Fibre Channel I/Oaccess pattern H-3load balancing H-2request rate optimizing H-3size H-3

Fibre Channel switch zoning B-13files, defragmenting H-5firmware download with I/O 3-19firmware levels, determining 3-16FlashCopy

disk array error messages (AIX) 5-38Frequently asked questions, backup and

recovery 6-33Frequently asked questions, boot

support 6-33Frequently asked questions, full disk

encryption 6-31Frequently asked questions, full disk

encryption premium feature 6-33Frequently asked questions, global

hot-spare drives 6-33Frequently asked questions, locked and

unlocked states 6-33Frequently asked questions, other 6-34Frequently asked questions, secure

erase 6-32Frequently asked questions, securing

arrays 6-31Frequently asked questions, security keys

and pass phrases 6-32full disk encryption 6-1Full disk encryption feature,

enabling 6-14Full disk encryption, best practices L-1Full disk encryption, changing a security

keyChanging a security key, full disk

encryption 6-4Security key, changing 6-4

Full disk encryption, configuring 6-13Full disk encryption, creating a security

keyCreating a security key, full disk

encryption 6-2Security key, creating 6-2

Full disk encryption, data backup L-1Full disk encryption, drive security key

and the security key file L-2Full disk encryption, erasing drives 6-27Full disk encryption, FDE disk

drives 6-1Full disk encryption, frequently asked

questions 6-31Full disk encryption, global hot-spare

drives 6-30Full disk encryption, installing FDE

drives 6-13Full disk encryption, key terms 6-12Full disk encryption, migrating

drives 6-23Full disk encryption, physical asset

protection L-1Full disk encryption, premium features,

frequently asked questions 6-33Full disk encryption, replacing

controllers L-3Full disk encryption, secure erase 6-10

X-2 IBM System Storage DS Storage Manager Version 10: Installation and Host Support Guide

Full disk encryption, securing a RAIDarray 6-16

Full disk encryption, securing dataagainst a breach 6-1

Full disk encryption, securityauthorizations 6-11

Full disk encryption, security keyidentifier 6-4

Full disk encryption, storage industrystandards and practices L-4

Full disk encryption, unlockingdrives 6-21

Full Disk Encryption, unlocking securedrives 6-10

Full disk encryption, working with FDEdrives L-3

Ggaseous contamination M-3General Parallel File System (GPFS) D-3Global hot-spare disk drives 6-30Global hot-spare drives, frequently asked

questions 6-33glossary N-1

HHACMP D-1hardware

Ethernet address 1-4hardware requirements

VMware ESX Server C-2hardware service and support xvHBA in switch environment B-13HBA settings B-1hdisk

attributes 5-31, E-5setting queue depth 5-31verification 5-28

heterogeneous environment 4-11High Availability Cluster

Multi-Processing (HACMP) D-1high-availability cluster services D-1host

VMware ESX Server C-2host and host port

defining 4-12host bus adapters

in a direct-attached configuration 1-5in a SAN-attached configuration 1-5setting host ports 4-1Solaris

JNI settings B-8QLogic settings B-13

host group, defining 4-1, 4-10host port, defined 4-11host software

package info 3-6host table

pre-installation tasks 1-4hosts

configuring 5-1HP-UX

cluster services, requirements D-9

HP-UX (continued)logical drives, redistributing in case of

failure 5-36PV-links 5-10requirements

cluster services D-9

II/O access pattern and I/O size H-3I/O data field H-1, H-2I/O request rate

optimizing H-3I/O transfer rate, optimizing H-3IBM

director of licensing address M-1IBM Safety Information J-11IBM System Storage Productivity

Center xivicons, Support Monitor 7-1identifying

devices 5-27in-band (host-agent) management method

utm device 5-28installation 3-3

completing 3-8preparation 1-1preparing a network 1-2

installingmultipath driver 5-6sequence of 3-6software components

configuration types 1-2Solaris

RDAC driver 5-22Storage Manager

manually 3-6VMware ESX Server

configuration C-1Installing FDE drives 6-13installing Storage Manager and Support

Monitor 3-3installing using a console window 3-5Intel and AMD-based host

support xvinterface, Support Monitor 7-1introduction

DS Storage Manager 1-1IP addresses for DS3000, DS4000, and

DS5000 controllers 1-6IPv6, using 3-14iSCSI host ports, configuring 3-13iSCSI session, viewing or ending 3-13iSCSI settings, managing 3-11iSCSI software initiator considerations,

Microsoft 3-15iSCSI statistics, viewing 3-13iSCSI, changing target

authentication 3-12iSCSI, changing target discovery 3-13iSCSI, changing target

identification 3-13iSCSI, configuring host ports 3-13iSCSI, entering mutual authentication

permissions 3-13iSCSI, network settings 3-14

iSCSI, using supported hardwareinitiators 3-14

iSNS best practices 3-13

JJNI

settings B-8

KKey terms, full disk encryption 6-12

LLinux

dynamic capacity expansion(DCE) 5-33

dynamic volume expansion(DVE) 5-33

Linux hostsupport xv

Linux MPP driver 5-8load balancing E-1

Fibre Channel I/O H-1load_balancing attribute E-1, E-2Locked and unlocked states, frequently

asked questions 6-33LockKeyID, full disk encryption 6-4Log files 6-31log window 7-6logical drives

configuration 4-8creating 4-4creating from free or unconfigured

capacity 4-5expected usage 4-9modification priority setting H-4redistributing in case of failure 5-35,

5-36, 5-37LUNs

attributes 5-31, E-5mapping to a partition 4-12

VMware ESX Server C-5

MMAC address, Ethernet 1-6Major events log 6-31management station 1-4

description 1-1VMware ESX Server C-1

Managing iSCSI settings 3-11mapping

LUNs, to a partition 4-12VMware ESX Server C-5

Maximum Transmission Unit 3-15Maximum Transmission Unit

settings 3-15MC/Service Guard D-9media scan

changing settings 4-19duration 4-22errors reported 4-20overview 4-19

Index X-3

media scan (continued)performance impact 4-20settings 4-21

medical imaging applications 4-7Medium Access Control (MAC) address,

Ethernet 1-6messages, Support Monitor 7-6Microsoft iSCSI Software Initiator

considerations 3-15Migrating secure disk drives 6-23Migration Guide J-1minihubs 1-5MPxIO 5-15multi-user environments 4-7multimedia applications 4-7multipath

DMP, installing on Solaris 5-24installing 5-6MPxIO , using with Solaris 5-15PV-links, using on HP-UX 5-10RDAC

installing on Solaris 5-22redistributing logical drives

AIX 5-35HP-UX 5-36Solaris 5-37

multipath driver 5-7description 5-3installing 5-6

multipath driversafter installation 5-27

Multiplexed I/O (MPxIO) 5-15Mutual authentication permissions, iSCSI

entering 3-13My Support xii

Nnaming storage subsystems 3-10network installation, preparing 1-2Network settings, iSCSI 3-14notes, important M-2notices xi

general M-1NVSRAM firmware

downloading 3-15NVSRAM firmware, downloading 3-15

OObject Data Manager (ODM) attributes

definitions E-1initial device identification 5-28lsattr command E-5viewing and setting E-1

operating systemrequirements 3-1

operating system requirements 3-1Other frequently asked questions 6-34out-of-band (direct) management method

setting IP addresses 1-6overview of heterogeneous hosts 4-11

PParallel System Support Programs

(PSSP) D-3parity 4-5particulate contamination M-3partitioning 4-1Pass phrases and security keys,

frequently asked questions 6-32performance

ODM attribute settings and 5-31performance monitor H-1premium feature

FlashCopy 4-16Full Disk Encryption 4-16key 4-16Remote Mirror Option 4-16

Premium feature, enabling full diskencryption 6-14

Premium feature, full diskencryption 6-1

premium featuresdisabling 3-23enabling 3-21, 3-22feature enable identifier 3-21feature key file 3-22Storage Partitioning

host group 4-1, 4-10Premium features, full disk encryption,

frequently asked questions 6-33preparing a network installation 1-2prerequisites

client software 3-6cluster services D-10HP-UX

cluster services D-9Solaris

cluster services D-10priority setting, modification H-4problem solving, critical event I-1products, developed M-1profile, storage subsystem

saving 3-23PV-links 5-10

QQLogic

settings B-13QLogic adapter settings B-1QLogic SANsurfer xiiqueue depth, setting 5-31

RRAID array, securing 6-16RAID level

application behavior 4-8, H-4choosing 4-8, H-4configurations 4-6data redundancy 4-5described 4-5

RAID-0described 4-6drive failure consequences 4-6

RAID-1described 4-6

RAID-1 (continued)drive failure consequences 4-6



RAID-6dual distributed parity 4-7

RDAC driverdescription 5-3, 5-15IDs D-10Solaris

installing 5-22recover configuration 4-4Recovery Guru

Diagnostic Data Capture F-1redistributing logical drives in case of

failureAIX 5-35HP-UX 5-36Solaris 5-37

Replacing controllers, full disk encryptionbest practices L-3

requirementsclient software 3-6cluster services D-10HP-UX

cluster services D-9operating system 3-1Solaris

cluster services D-10resources

documents xiiiWeb sites xiv

reviewing a sample network 1-3

Ssample network, reviewing 1-3SAN boot

configuring hosts 5-1SAN-attached configuration 1-5schedule support bundle collection 7-3script editor

using G-2Script Editor

adding comments to a script G-2Diagnostic Data Capture F-1window G-1

Secure drives, locked and unlockedstates, frequently asked questions 6-33

Secure drives, unlocking 6-10Secure erase disk drives 6-27Secure erase, FDE disk drive 6-10Secure erase, frequently asked

questions 6-32Securing a RAID array 6-16Securing arrays, frequently asked

questions 6-31Securing data against a breach 6-1Security authorizations, full disk

encryption 6-11Security key identifier, full disk

encryption 6-4


Security keys and pass phrases,frequently asked questions 6-32

segment size, choosing H-4send support bundle to IBM 7-3services offered in the U.S.A. M-1setting

IP addresses 1-6setting up alert notifications 3-10settings 3-15settings, media scan 4-21Simple Network Management Protocol

(SNMP) traps 1-3, 1-6SMagent

software installation sequence 3-6SMclient

software installation sequence 3-6SMdevices utility, using 5-27SMruntime

software installation sequence 3-6SMutil

software installation sequence 3-6SNMP traps 1-3, 1-6software

setting up addresses 1-6Software initiator considerations,

Microsoft iSCSI 3-15software package

multipath driver 5-3RDAC 5-15

software requirementsVMware ESX Server C-1

Solariscluster services requirements D-10DMP 5-24, 5-26logical drives, redistributing in case of

failure 5-37RDAC driver

installing 5-22requirements

cluster services D-10Veritas 5-24, 5-26

SSPC xivSSPC (System Storage Productivity

Center) xivstarting Subsystem Management 3-10storage area network (SAN)

configuration 1-5technical support Web site xv

Storage industry standards andpractices L-4

storage management station 1-4description 1-1

Storage ManagerController Firmware Upgrade Tool

using the tool A-1Enterprise Management Window 2-1installation 3-3installing automatically using the

wizard 3-3installing for use with IBM System

Storage Productivity Center xivintroducing the software 2-1manual installation 3-6Subsystem Management Window 2-4

Storage Manager drive firmware,download 3-19

Storage Manager softwareinstallation sequence 3-6introduction 1-1list of software packages 1-1new terminology xiwhere to obtain xiii

Storage Manager version 10.5x drivefirmware download 3-19

Storage Partitioningand host groups 4-1

storage subsystemcluster services D-1introduction 1-1naming 3-10performing initial automatic

discovery 3-8profile, saving 3-23tuning options available H-1

storage subsystem firmware levels,determining 3-16

storage subsystemstuning H-1

Subsystem controller shell remote login,full disk encryption best practices L-3

Subsystem Management windowevent log I-1online help xii

support bundlecollecting manually 7-5scheduling collection 7-3sending to IBM Support 7-3

Support Monitor 3-5configuring 7-1console area 7-1Enterprise status 7-1icons 7-1installing automatically using the

wizard 3-3interface 7-1log window 7-6messages 7-6sending support bundles 7-3solving problems 7-8support bundle collection

schedule 7-3troubleshooting 7-8using 7-1

Support Monitor log window, using 7-6support notifications xiiswitch

in a SAN-attached configuration 1-5technical support Web site xvzoning 1-5

switch environment B-13System p host

support xvSystem Storage Interoperation Center

(SSIC) xivSystem Storage Productivity Center xivSystem Storage Productivity Center

(SSPC) xivSystem x host

support xv

TTarget authentication, iSCSI

changing 3-12Target discovery, iSCSI changing 3-13Target identification, iSCSI

changing 3-13tasks by document title J-1tasks by documentation title J-1terminology xiTerms, full disk encryption 6-12trademarks M-2transfer rate H-1

Uuniversal transport mechanism 5-28Unlocking disk drives 6-21Unlocking secure drives, full disk

encryption 6-10updates (product updates) xiiupgrade tool

adding a storage subsystem A-2checking device health A-1downloading firmware A-2overview A-1using A-2viewing log file A-3

usingSMdevices utility 5-27

Using DHCP 3-13Using IPv6 3-14Using supported hardware initiators,

iSCSI 3-14utm device 5-28

Vverifying

default host type 4-9Veritas 5-7

Dynamic Multipathing (DMP) 5-15,5-24, 5-25, 5-26

File System 5-24VolumeManager 5-15, 5-24, 5-26

Veritas DMP driver 5-9Viewing an iSCSI session 3-13Viewing iSCSI statistics 3-13VMware ESX Server

cross connections C-4mapping LUNs to a partition C-5

WWeb sites

AIX xvIBM publications center xvIBM System Storage product

information xvlist xivpremium feature activation xivSAN support xvSolaris failover driver info 5-24SSIC xivswitch support xvSystem p xv

Index X-5

Web sites (continued)System Storage Productivity Center

(SSPC) xivSystem x xv

who should read this document xiwindow, Script Editor G-1write caching

enabling H-3

Zzoning B-13zoning switches 1-5


��

Part Number: 59Y7292

Printed in USA

(1P) P/N: 59Y7292

Documents

59Y7292 - IBM DS Storage Manager v10 Installation and Host Support Guide