HP Integrity NonStop NS14000 Series Planning Guide

Transcription

1 HP Integrity NonStop NS14000 Series Planning Guide HP Part Number: Published: March 2012 Edition: H06.13 and subsequent Hseries RVUs

2 Copyright 2006, 2012 HewlettPackard Development Company, L.P. Confidential computer software. Valid license from HP required for possession, use or copying. Consistent with FAR and , Commercial Computer Software, Computer Software Documentation, and Technical Data for Commercial Items are licensed to the U.S. Government under vendor s standard commercial license. Warranty The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein. Export of the information contained in this publication may require authorization from the U.S. Department of Commerce. Acknowledgements Microsoft, Windows, Windows NT, Windows XP, and Windows Vista are U.S. registered trademarks of Microsoft Corporation. Intel and Intel Itanium are trademarks of Intel Corporation in the U.S. and other countries. Java is a registered trademark of Oracle and/or its affiliates. Motif, OSF/1, UNIX, X/Open, and the X device is a trademark of X/Open Company Ltd. in the UK and other countries. OSF, OSF/1, OSF/Motif, Motif, and Open Software Foundation are trademarks of the Open Software Foundation in the U.S. and other countries. 1990, 1991, 1992, 1993 Open Software Foundation, Inc. The OSF documentation and the OSF software to which it relates are derived in part from materials supplied by the following: 1987, 1988, 1989 CarnegieMellon University. 1989, 1990, 1991 Digital Equipment Corporation. 1985, 1988, 1989, 1990 Encore Computer Corporation Free Software Foundation, Inc. 1987, 1988, 1989, 1990, 1991 HewlettPackard Company. 1985, 1987, 1988, 1989, 1990, 1991, 1992 International Business Machines Corporation. 1988, 1989 Massachusetts Institute of Technology. 1988, 1989, 1990 Mentat Inc Microsoft Corporation. 1987, 1988, 1989, 1990, 1991, 1992 SecureWare, Inc. 1990, 1991 Siemens Nixdorf Informations systeme AG. 1986, 1989, 1996, 1997 Sun Microsystems, Inc. 1989, 1990, 1991 Transarc Corporation. OSF software and documentation are based in part on the Fourth Berkeley Software Distribution under license from The Regents of the University of California. OSF acknowledges the following individuals and institutions for their role in its development: Kenneth C.R.C. Arnold, Gregory S. Couch, Conrad C. Huang, Ed James, Symmetric Computer Systems, Robert Elz. 1980, 1981, 1982, 1983, 1985, 1986, 1987, 1988, 1989 Regents of the University of California. OSF MAKES NO WARRANTY OF ANY KIND WITH REGARD TO THE OSF MATERIAL PROVIDED HEREIN, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. OSF shall not be liable for errors contained herein or for incidental consequential damages in connection with the furnishing, performance, or use of this material.

3 Contents About This Document...7 Supported Release Version Updates (RVUs)...7 Intended Audience...7 New and Changed Information...7 New and Changed Information for New and Changed Information for New and Changed Information for New and Changed Information for New and Changed Information for New and Changed Information for Document Organization...11 Notation Conventions...12 General Syntax Notation...12 Related Information...14 Documentation...14 Publishing History...14 HP Encourages Your Comments NonStop NS14000 Series System Overview...15 NonStop NS14000 Series Architecture...16 NonStop NS14000 Series Hardware...16 NonStop Blade Element...17 Logical Synchronization Unit (LSU)...20 Versatile I/O (VIO) Enclosure...22 Fibre Channel Disk Module (FCDM)...24 Maintenance Switch (Ethernet)...24 System Console...25 UPS and ERM (Optional)...25 Enterprise Storage System (Optional)...26 Tape Drive and Interface Hardware (Optional)...27 Preparation for Other Server Hardware...28 Component Location and Identification...28 Terminology...28 Rack and Offset Physical Location...29 NonStop Blade Element GroupModuleSlot Numbering...30 LSU GroupModuleSlot Numbering...31 VIO Enclosure GroupModuleSlot Numbering...32 Fibre Channel Disk Module GroupModuleSlot Numbering...33 System Installation Document Packet...34 Configuration Technical Document for the FactoryInstalled Hardware...35 Configuration Forms for the ServerNet Adapters...35 Configuration Form for a ServerNet Cluster Site Preparation Guidelines...36 Modular Cabinet Power and I/O Cable Entry...36 Emergency PowerOff Switches...36 EPO Requirement for NonStop NS14000 Series Servers...36 EPO Requirement for HP 5000 UPS or HP 5500 XR UPS...36 Electrical Power and Grounding Quality...36 Power Quality...37 Grounding Systems...37 Power Consumption...37 Contents 3

4 Uninterruptible Power Supply (UPS)...37 Cooling and Humidity Control...38 Weight...39 Flooring...39 Dust and Pollution Control...39 Zinc Particulates...39 Space for Receiving and Unpacking...39 Operational Space System Installation Specifications...41 Modular Cabinets...41 Monitored SinglePhase PDUs...41 AC Power Feeds, Monitored SinglePhase PDUs...42 Input and Output Power Characteristics, Monitored SinglePhase PDUs...45 Branch Circuits and Circuit Breakers, Monitored SinglePhase PDUs...46 Monitored ThreePhase PDUs...46 AC Power Feeds, Monitored ThreePhase PDUs...47 Input and Output Power Characteristics, Monitored ThreePhase PDUs...50 Branch Circuits and Circuit Breakers, Monitored ThreePhase PDUs...51 Modular ThreePhase PDUs...51 AC Power Feeds, Modular ThreePhase PDUs...52 Input and Output Power Characteristics, Modular ThreePhase PDUs...56 Branch Circuits and Circuit Breakers, Modular ThreePhase PDUs...57 Circuit Breaker Ratings for UPS...58 PDU Strapping Configurations...58 Grounding...58 Enclosure AC Input...59 Enclosure Power Loads...59 Dimensions and Weights...60 Plan View From Above the Modular Cabinet...61 Service Clearances for the Modular Cabinet...61 Unit Sizes...61 Modular Cabinet Physical Specifications...61 Enclosure Dimensions...62 Modular Cabinet and Enclosure Weights With Worksheet...62 Modular Cabinet Stability...63 Environmental Specifications...63 Calculating Specifications for Enclosure Combinations System Configuration Guidelines...68 Internal ServerNet Interconnect Cabling...68 Dedicated Service LAN Cables...68 Length Restrictions for Optional Cables...68 Internal Cable Product IDs...68 NonStop Blade Elements to LSU...69 NonStop Blade Element to NonStop Blade Element...69 LSU to VIO Enclosure and Processor IDs...69 VIO Enclosure ServerNet Connections...71 Fibre Channel Port to Fibre Channel Disk Modules...72 Fibre Channel Port to Tape Devices...72 VIO Enclosure and Disk Storage Considerations...72 Fibre Channel Devices...72 FactoryDefault Disk Volume Locations...74 Configurations for Fibre Channel Devices...74 Configuration Restrictions for Fibre Channel Devices...75 Recommendations for Fibre Channel Device Configurations Contents

5 Example Configurations of the VIO Enclosures and Fibre Channel Disk Modules...76 DaisyChain Configurations...80 Ethernet to Networks...82 Default Naming Conventions Hardware Configurations in Modular Cabinets...84 Minimum, Typical, and Maximum Hardware Configuration...84 Enclosure Locations in Cabinets Maintenance and Support Connectivity...85 Dedicated Service LAN...85 Basic LAN Configuration...86 FaultTolerant Configuration...86 IP Addresses...87 Ethernet Cables...89 SWAN Concentrator Restrictions...89 SystemUp Dedicated Service LAN...90 Dedicated Service LAN Links With Two VIO Enclosures...90 Initial Configuration for a Dedicated Service LAN...91 Operating Configurations for Dedicated Service LANs...92 System Consoles...92 System Console Configurations...92 A Cables for NS14000 Series Systems...94 Cable Types, Connectors, Length Restrictions, and Product IDs...94 Cable Management System...96 B Operations and Management Using OSM Applications...97 SystemDown OSM LowLevel Link...97 AC Power Monitoring...98 OSM Power Fail Support...98 Considerations for RideThrough Time Configuration...99 Considerations for Site UPS Configurations AC PowerFail States C Default Startup Characteristics D NonStop NS14000 Series System Architecture NonStop Blade Complex Processor Element Duplex Processor Triplex Processor Processor Synchronization and Rendezvous Memory Reintegration Failure Recovery for Duplex Processor Failure Recovery for Triplex Processor Index Contents 5

6 Figures 1 Example of a NonStop NS14000 Series System Top AC Power Feed, Monitored SinglePhase PDUs Bottom AC Power Feed, Monitored SinglePhase PDUs Top AC Power Feed When Optional UPS and ERM are Installed Bottom AC Power Feed When Optional UPS and ERM are Installed Top AC Power Feed, Monitored ThreePhase PDUs Bottom AC Power Feed, Monitored ThreePhase PDUs Top AC Power Feed When Optional UPS and ERM are Installed Bottom AC Power Feed When Optional UPS and ERM are Installed Top AC Power Feed, Modular ThreePhase PDUs Bottom AC Power Feed, Modular ThreePhase PDUs Top AC Power Feed When Optional UPS and ERM are Installed Bottom AC Power Feed When Optional UPS and ERM are Installed Example Duplex Configuration Four Fibre Channel Ports, Two FCDMs, Two VIO Enclosures Eight Fibre Channel Ports, Four FCDMs, Two VIO Enclosures Twelve Fibre Channel Ports, Six FCDMs, Two VIO Enclosures Sixteen Fibre Channel Ports, Eight FCDMs, Two VIO Enclosures Basic LAN Configuration FaultTolerant LAN Configuration...87 Tables 1 Example Load Calculations for Cabinet One Example Load Calculations for Cabinet Two...67

7 About This Document This guide describes the HP Integrity NonStop NS14000 series system and provides examples of system configurations to assist you in planning for installation of a new system. The NonStop NS14000 series of servers consists of the NonStop NS14000 server and the NonStop NS14200 server. Supported Release Version Updates (RVUs) This publication supports H06.13 and all subsequent Hseries RVUs until otherwise indicated by its replacement publication. NOTE: For NonStop NS14000 series servers that contain I/O Adapter Module (IOAM) enclosures, refer to the H06.07 (or earlier) version of the NonStop NS14000 Planning Guide for planning, configuration, and maintenance information of the IOAM enclosure. Intended Audience This guide is written for those responsible for planning the installation, configuration, and maintenance of the server and the software environment at a particular site. Appropriate personnel must have completed HP training courses on system support for Integrity NonStop NSseries servers. NOTE: Integrity NonStop NSseries and NonStop Sseries refer to hardware systems. Hseries and Gseries refer to release version updates (RVUs). New and Changed Information These chapters and appendices have been changed for the editions, as indicated. All changes are marked with change bars. New and Changed Information for Changed the following maintenance switch information: Added maintenance switch support for multiple systems in Maintenance Switch (Ethernet) (page 24) Changed the maintenance switch connections in Basic LAN Configuration (page 86) and FaultTolerant Configuration (page 86). Changed the maintenance switch's height, width, and depth in Enclosure Dimensions (page 62). Changed the maintenance switch's weight in Modular Cabinet and Enclosure Weights With Worksheet (page 62). Changed the maintenance switch's heat units in Heat Dissipation Specifications and Worksheet (page 64). Changed the maintenance switch's weight and heat units in Table 1: Example Load Calculations for Cabinet One (page 66). Added HP R5000 UPS in UPS and ERM (Optional) (page 25), AC Power Monitoring (page 98), Modular Cabinet Power and I/O Cable Entry (page 36), EPO Requirement for HP 5000 UPS or HP 5500 XR UPS (page 36), Uninterruptible Power Supply (UPS) (page 37), Modular Cabinets (page 41), Top AC Power Feed When Optional UPS and ERM are Installed (page 43), Bottom AC Power Feed When Optional UPS and ERM are Installed (page 44), Branch Circuits and Circuit Breakers, Monitored SinglePhase PDUs (page 46), Circuit Breaker Ratings for UPS (page 58), Enclosure Power Loads (page 59), Unit Sizes (page 61), Enclosure Dimensions (page 62), Modular Cabinet and Enclosure Weights Supported Release Version Updates (RVUs) 7

8 With Worksheet (page 62), Environmental Specifications (page 63), Operating Temperature, Humidity, and Altitude (page 64), Nonoperating Temperature, Humidity, and Altitude (except R5000 UPS and R5000 ERM) (page 65), Nonoperating Temperature, Humidity, and Altitude for R5000 UPS and R5000 ERM (page 65), Calculating Specifications for Enclosure Combinations (page 65), AC PowerFail States (page 100). Added HP R5000 ERM in UPS and ERM (Optional) (page 25), Uninterruptible Power Supply (UPS) (page 37), Unit Sizes (page 61), Enclosure Dimensions (page 62), Modular Cabinet and Enclosure Weights With Worksheet (page 62), Operating Temperature, Humidity, and Altitude (page 64), Calculating Specifications for Enclosure Combinations (page 65), Operating Temperature, Humidity, and Altitude (page 64), Nonoperating Temperature, Humidity, and Altitude for R5000 UPS and R5000 ERM (page 65). Changed the former tech memo term to Technical Document in Configuration Technical Document for the FactoryInstalled Hardware (page 35). Changed the information for OSM power fail support in Uninterruptible Power Supply (UPS) (page 37). Replaced the table for Power and current specifications for each type of enclosure in Enclosure Power Loads (page 59). Made changes to Hardware Configurations in Modular Cabinets (page 84): Deleted the former table for enclosure locations and referenced the Technical Document for enclosures and component locations Deleted the former Typical Configurations section Deleted the former Example Internal Cabling section Made changes to Maintenance and Support Connectivity (page 85): Changed the introductory RSA information Deleted the former Maintenance Architecture illustration Changed the component connections in Basic LAN Configuration (page 86) Changed the illustration in Basic LAN Configuration (page 86) Changed the illustration in FaultTolerant LAN Configuration (page 87) Added Note to refer to the NonStop System Console Installer Guide in System Consoles (page 92) Changed the supported system console configurations in System Consoles (page 92) Changed the system console documentation in System Consoles (page 92) Deleted the former illustration in One System Console Managing One System (Setup Configuration) (page 92) Deleted the former illustration in One System Console Managing Multiple Systems (page 93) Changed all content in One System Console Managing Multiple Systems (page 93) Deleted the former Cascading Ethernet Switch or Hub Connectivity section Replaced the former Cables appendix with the new Cables for NS14000 Series Systems (page 94) appendix Made changes to Operations and Management Using OSM Applications (page 97): Changed introductory RSA and OSM information Changed the HP R5500 XR UPS and OSM Power Fail Support information in AC Power Monitoring (page 98) Created the new OSM Power Fail Support (page 98) section 8

9 Deleted the former How OSM Power Failure Support Works section Changed ridethrough time and OSM information in Considerations for RideThrough Time Configuration (page 99) Added support for Category 6 unshielded twistedpair (CAT 6 UTP) cables for Ethernet connections Changed these graphics to show the extension bars on the left when a UPS is installed in the modular cabinet: Figure 4: Top AC Power Feed When Optional UPS and ERM are Installed (page 44) Figure 5: Bottom AC Power Feed When Optional UPS and ERM are Installed (page 45) Figure 8: Top AC Power Feed When Optional UPS and ERM are Installed (page 49) Figure 9: Bottom AC Power Feed When Optional UPS and ERM are Installed (page 50) New and Changed Information for Changed the C13 receptacle type from 10A to 12A in these sections: Monitored SinglePhase PDUs (page 41) North America and Japan: 200 to 240 V AC, Monitored SinglePhase PDUs (page 45) International: 200 to 240 V AC, Monitored SinglePhase PDUs (page 46) Monitored ThreePhase PDUs (page 46) North America and Japan: 200 to 240 V AC, Monitored ThreePhase PDUs (page 50) International: 380 to 415 V AC, Monitored ThreePhase PDUs (page 51) Modular ThreePhase PDUs (page 51) North America and Japan: 200 to 240 V AC, Modular ThreePhase PDUs and Extension Bars (page 56) International: 380 to 415 V AC, Modular ThreePhase PDUs and Extension Bars (page 57) Changed the C19 receptacle type from 12A to 16A in these sections: Monitored SinglePhase PDUs (page 41) North America and Japan: 200 to 240 V AC, Monitored SinglePhase PDUs (page 45) International: 200 to 240 V AC, Monitored SinglePhase PDUs (page 46) Monitored ThreePhase PDUs (page 46) North America and Japan: 200 to 240 V AC, Monitored ThreePhase PDUs (page 50) International: 380 to 415 V AC, Monitored ThreePhase PDUs (page 51) Made changes to the Enclosure Dimensions (page 62) table. Made changes to the Modular Cabinet and Enclosure Weights With Worksheet (page 62) table. Added note to Chapter 6: Maintenance and Support Connectivity (page 85) and Appendix B: Operations and Management Using OSM Applications (page 97) indicating that HP Insight Remote Support Advanced (Insight RSA) is the goforward remote support solution for NonStop systems, replacing the OSM Notification Director in both modembased and HP Instant Support Enterprise Edition (ISEE) remote support solutions. New and Changed Information 9

10 New and Changed Information for Made additions and changes to the Internal Cables table. Under Appendix C: Default Startup Characteristics (page 101), added a note stating that the configurations documented here are typical for most sites. Your system load paths might be different, depending upon how your system is configured. To determine the configuration of your system, refer to the system attributes in the OSM Service Connection. You can select this from within the System Load dialog box in the OSM LowLevel Link. New and Changed Information for Added the new modular power distribution units (PDUs) for threephase power configurations. Updated these sections: Modular Cabinets (page 41) Monitored SinglePhase PDUs (page 41) Monitored ThreePhase PDUs (page 46) Modular ThreePhase PDUs (page 51) Circuit Breaker Ratings for UPS (page 58) Dimensions and Weights (page 60), Modular Cabinet and Enclosure Weights With Worksheet (page 62) Under NonStop Blade Element (page 17), added statement: With a maximum of only three NonStop Blade Complexes, the NonStop Blade Elements could be in two or three NonStop Blade Complexes. Under Terminology (page 28), changed definitions for NonStop Blade Complex and NonStop Blade Element. For the NonStop Blade Complex, indicated that an 8processor system employs two or three NonStop Blade Complexes. For the NonStop Blade Element, indicated that each processor element supports a different logical processor numbered 07. Changed figures to show appropriate system console icon. Under Chapter 6: Maintenance and Support Connectivity (page 85), indicated that the HP ISEE callout and callin access is determined by what access method (for example, direct connect or VPN router) you have ordered. Under Modular Cabinets (page 41), added note: For instructions on grounding the modular cabinet's G2 rack using the HP Rack Grounding Kit (AF074A), ask your service provider to refer to the instructions in the HP G2 Series Rack Options Installation Guide. This manual is available at: c /c pdf. Added new procedures to AC Power Monitoring (page 98), including the new subsections, How OSM Power Failure Support Works, Considerations for RideThrough Time Configuration (page 99), and Considerations for Site UPS Configurations (page 100). New and Changed Information for Corrected the operating temperatures under Operating Temperature, Humidity, and Altitude (page 64) and Nonoperating Temperature, Humidity, and Altitude (except R5000 UPS and R5000 ERM) (page 65). Indicated under Default Naming Conventions (page 82) that changing the process names for $ZTCP0 and $ZTCP1 will make some components inaccessible. Changed the weight of the PDU monitored configuration from 303 lbs. to 328 lbs. due to the additional weight of the rack extension. See Modular Cabinet and Enclosure Weights With 10

11 Worksheet (page 62), Table 1: Example Load Calculations for Cabinet One (page 66), and Table 2 (page 67). Changed VIO IP Addresses (page 87) from to and to New and Changed Information for The title of this guide has been changed to the NonStop NS14000 Series Planning Guide because it supports both the NonStop NS14000 server and the NonStop NS14200 server. This guide has been streamlined and reorganized to focus on planning information. The NonStop NS14200 server and the NSES processor type has been added to Chapter 1 (page 15). Processor Element (page 105) has been updated to include the NSES processor type. Plan View From Above the Modular Cabinet (page 61) has been changed to show the dimensions of the HP10000 G2 Series rack. Modular Cabinet Physical Specifications (page 61) has been changed to show the dimensions of the HP10000 G2 Series rack. Modular Cabinet Stability (page 63) has been added to discuss the use of cabinet stabilizers. The disk path identifiers in several illustrations have been corrected. See Example Configurations of the VIO Enclosures and Fibre Channel Disk Modules (page 76). Information about ISEE has been added. See Chapter 6 (page 85). Two cables have been added to the list of available cables in Internal Cables. Editorial changes have been made throughout the manual. Document Organization NOTE: For NonStop NS14000 series servers that contain I/O Adapter Module (IOAM) enclosures, refer to the H06.07 (or earlier) version of the NonStop NS14000 Planning Guide for planning, configuration, and maintenance information of the IOAM enclosure. This guide is divided into these chapters: Chapter Chapter 1: NonStop NS14000 Series System Overview (page 15) Chapter 2: Site Preparation Guidelines (page 36) Chapter 3: System Installation Specifications (page 41) Chapter 4: System Configuration Guidelines (page 68) Chapter 5: Hardware Configurations in Modular Cabinets (page 84) Chapter 6: Maintenance and Support Connectivity (page 85) Section : Cable Types, Connectors, Length Restrictions, and Product IDs (page 94) Appendix B: Operations and Management Using OSM Applications (page 97) Contents Provides an overview of the modular hardware components in a NonStop NS14000 series system. Describes power, environmental, and space considerations for your site. Provides installation specifications for fully populated NonStop NS14000 series enclosures. Provides guidelines for implementing the modular hardware. Provides example configurations of the NonStop NS14000 series modular hardware. Describes the maintenance architecture used for NonStop NS14000 series system. Identifies the cables used with NonStop NS14000 series hardware Describes how the OSM applications are used to manage a NonStop NS14000 series system. Document Organization 11

12 Chapter Appendix C: Default Startup Characteristics (page 101) Appendix D: NonStop NS14000 Series System Architecture (page 103) Contents Lists the default startup characteristics for a NonStop NS14000 series system Describes the architecture used in the NonStop NS14000 series system. Notation Conventions General Syntax Notation 12 This list summarizes the notation conventions for syntax presentation in this manual. UPPERCASE LETTERS Uppercase letters indicate keywords and reserved words. Type these items exactly as shown. Items not enclosed in brackets are required. For example: MAXATTACH Italic Letters Italic letters, regardless of font, indicate variable items that you supply. Items not enclosed in brackets are required. For example: filename Computer Type Computer type letters indicate: C and Open System Services (OSS) keywords, commands, and reserved words. Type these items exactly as shown. Items not enclosed in brackets are required. For example: Use the cextdecs.h header file. Text displayed by the computer. For example: Last Logon: 14 May 2006, 08:02:23 A listing of computer code. For example if (listen(sock, 1) < 0) { perror("listen Error"); exit(1); } Bold Text Bold text in an example indicates user input typed at the terminal. For example: ENTER RUN CODE?123 CODE RECEIVED: The user must press the Return key after typing the input. [ ] Brackets Brackets enclose optional syntax items. For example: TERM [\systemname.]$terminalname INT[ERRUPTS] A group of items enclosed in brackets is a list from which you can choose one item or none. The items in the list can be arranged either vertically, with aligned brackets on each side of the list, or horizontally, enclosed in a pair of brackets and separated by vertical lines. For example: FC [ num ]

13 [ num] [ text] K [ X D ] address { } Braces A group of items enclosed in braces is a list from which you are required to choose one item. The items in the list can be arranged either vertically, with aligned braces on each side of the list, or horizontally, enclosed in a pair of braces and separated by vertical lines. For example: LISTOPENS PROCESS { $applmgrname } { $processname } ALLOWSU { ON OFF } Vertical Line A vertical line separates alternatives in a horizontal list that is enclosed in brackets or braces. For example: INSPECT { OFF ON SAVEABEND } Ellipsis An ellipsis immediately following a pair of brackets or braces indicates that you can repeat the enclosed sequence of syntax items any number of times. For example: M address [, newvalue ] ] { } An ellipsis immediately following a single syntax item indicates that you can repeat that syntax item any number of times. For example: "schar " Punctuation Parentheses, commas, semicolons, and other symbols not previously described must be typed as shown. For example: error := NEXTFILENAME ( filename ) ; LISTOPENS SU $processname.#suname Quotation marks around a symbol such as a bracket or brace indicate the symbol is a required character that you must type as shown. For example: "[" repetitionconstantlist "]" Item Spacing Spaces shown between items are required unless one of the items is a punctuation symbol such as a parenthesis or a comma. For example: CALL STEPMOM ( processid ) ; If there is no space between two items, spaces are not permitted. In this example, no spaces are permitted between the period and any other items: $processname.#suname Line Spacing If the syntax of a command is too long to fit on a single line, each continuation line is indented three spaces and is separated from the preceding line by a blank line. This spacing distinguishes items in a continuation line from items in a vertical list of selections. For example: Notation Conventions 13

14 ALTER [ / OUT filespec / ] LINE Related Information Documentation [, attributespec ] Manuals, Hotstuff messages, and other kinds of documentation are available in the NonStop Technical Library (NTL) at Publishing History Part Number Product Version N.A. N.A. N.A. N.A. N.A. Publication Date February 2009 May 2009 August 2009 November 2009 August 2011 HP Encourages Your Comments HP encourages your comments concerning this document. We are committed to providing documentation that meets your needs. Send any errors found, suggestions for improvement, or compliments to: docsfeedback@hp.com Include the document title, part number, and any comment, error found, or suggestion for improvement you have concerning this document. 14

15 1 NonStop NS14000 Series System Overview The Integrity NonStop NS14000 series of servers consists of the NonStop NS14000 server and the NonStop NS14200 server. A NonStop NS14000 series server uses much of the same modular system hardware as other NonStop NSseries servers. It supports up to eight processors using the NonStop NS14000 Series Architecture (page 16), and ServerNet connectivity is extended from the Versatile I/O (VIO) enclosure to the NonStop Blade Complex by way of the logical synchronization unit (LSU). NOTE: For NonStop NS14000 series servers that contain I/O Adapter Module (IOAM) enclosures, refer to the H06.07 (or earlier) version of this manual for planning, configuration, and maintenance information of the IOAM enclosure. The characteristics of a NonStop NS14000 series server: Processor Processor models Main memory Maximum processors Supported processor configurations Maximum Fibre Channel disk modules (FCDMs) Maximum VIO enclosures Maximum embedded I/O connectivity per VIO enclosure Maximum optional expansion I/O connectivity per VIO enclosure ESS support M8201R Fibre Channel to SCSI router support Connection to NonStop ServerNet Clusters Connection to NonStop Sseries I/O Intel Itanium NSED (NonStop NS14000 server) NSES (NonStop NS14200 server) 4 GB or 8 GB per processor 8 2, 4, 6 or 8 8 (32 with eight daisychained configurations containing 4 FCDM per configuration) 2 required (one each for X and Y fabrics) 4 embedded Fibre Channel ports for FCDMs, tape, or ESS connections4 embedded copper Ethernet ports 4 optional Fibre Channel ports for FCDMs, tape, or ESS connectionsor4 optional Ethernet ports (copper/optical) Yes Yes 6780 cluster switch only Not supported Figure 1 (page 16) shows the rear view of an example NonStop NS14000 series system with a duplex processor in a 42U modular cabinet without the optional UPS and ERM. 15

16 Figure 1 Example of a NonStop NS14000 Series System NonStop NS14000 Series Architecture A NonStop NS14000 series server use Intel Itanium processors in duplex or triplex configurations as part of the NonStop NS14000 series architecture. For details of the architecture, see Appendix D (page 103). NonStop NS14000 Series Hardware A large number of enclosure combinations are possible within the modular cabinets of a NonStop NS14000 series server. The applications and purpose of any NonStopNS series server determine the number and combinations of the enclosures and modular cabinets. 16 NonStop NS14000 Series System Overview

17 Standard hardware for a NonStop NS14000 series server includes: NonStop Blade Element (page 17) Logical Synchronization Unit (LSU) (page 20) Versatile I/O (VIO) Enclosure (page 22) Fibre Channel Disk Module (FCDM) (page 24) Maintenance Switch (Ethernet) (page 24) Optional hardware for a NonStop NS14000 series server includes: UPS and ERM (Optional) (page 25) Enterprise Storage System (Optional) (page 26) Tape Drive and Interface Hardware (Optional) (page 27) All NonStop NSseries server components are fieldreplaceable units (FRUs) that can only be serviced by service providers trained by HP. Because of the large number of possible configurations, you calculate the total power consumption, heat dissipation, and weight of each modular cabinet based on the hardware configuration that you order from HP. For site preparation specifications of the modular cabinet and the individual enclosures, see Chapter 3 (page 41). For examples of various combinations of enclosures installed in a modular cabinet, see Chapter 5 (page 84). NonStop Blade Element The NonStop Blade Element enclosure, which is 5U and weighs 112 pounds (46 kilograms), has these physical attributes: Rackmountable Redundant AC power feeds Fronttorear cooling Cable connections at rear (power, reintegration, LSU) with cable management equipment on the rear of the cabinet Each NonStop Blade Element includes these field replaceable units (FRUs): Processor board with up to four Itanium microprocessors, interface ASIC, clock generation (accessed through front of enclosure) Memory board that can hold up to 32 DIMMs, each one 4 GB, for a total memory capacity of 128 GB (accessed through front of enclosure) DIMMs Reintegration board for managing internal memory traffic (accessed from top of enclosure when enclosure is pulled forward on its rails) Blade optics adapter plugin cards (PICs) with two ports; two adapters minimum, eight maximum (accessed from top of enclosure when enclosure is pulled forward on its rails) Redundant cooling fans (accessed from top of enclosure when enclosure is pulled forward on its rails) Redundant V AC power supplies and power cords (accessed from back of of enclosure) NonStop NS14000 Series Hardware 17

18 I/O interface board, NonStop Blade Element I/O interface board for parallel to serial conversion and maintenance logic (requires removal of enclosure from modular cabinet) Front panel with indicator LEDs and power buttons The NonStop Blade Element midplane for logic interconnection and power distribution, which is part of the chassis assembly, is not a FRU. Two NonStop Blade Elements provide up to four processor elements in a highavailability duplex configuration, and four NonStop Blade Elements provide an 8processor duplex system. For a faulttolerant triplex system, three NonStop Blade Elements provide four processors, and six NonStop Blade Elements provide an 8processor triplex system. NOTE: NonStop NS14000 series systems do not support duplex and triplex processors within the same system. This illustration shows the rear of the NonStop Blade Element, equipped with two power supplies and eight Blade optics adapters: The numbers below the optics connectors refer to the processor numbers within the NonStop Blade Element enclosure. Currently, only connectors J0, J1, J2, and J3 are used. The remaining J connectors are for future use by the NonStop Blade Element hardware. CAUTION: To maintain proper cooling air flow, blank panels must be installed in all slots that do not contain Blade optics adapters. Labels for the fiberoptic cable connections can become complex in a large system. Fiberoptic cables for the ServerNet fabric and also the optic cable connections use the same type of optics connector. Because the modular hardware provides considerable flexibility in how the hardware is distributed among multiple cabinets, a single cabinet could contain four NonStop Blade Elements, with each NonStop Blade Element a member of a different NonStop Blade Complex. With a maximum of only three NonStop Blade Complexes, the NonStop Blade Elements could be in two or three NonStop Blade Complexes. To reduce ambiguity in identifying proper cable connections to the NonStop Blade Complex, the identification convention uses combinations of a letter to refer to each connection. A number such as 0, 1, 2, and 3 identifies NonStop Blade Complexes, and a letter such as A, B, or C identifies the NonStop Blade Element. Therefore, a single NonStop Blade Element is identified with an 18 NonStop NS14000 Series System Overview

19 alphanumeric ID, such as A1, A2, A3, and so forth. These IDs refer to the appropriate NonStop Blade Element for proper connection of the fiberoptic cables. The optic cables provide communications between each NonStop Blade Element and the LSU as well as between the LSU and the VIO enclosures on the X fabric and Y fabric. No requirement exists to connect cables from a particular Blade optics adapter on a NonStop Blade Element to a physically corresponding adapter on an LSU. However, to help reduce the complexity of cable connections, HP recommends that you use a physically sequential order of slots for fiberoptic cable connections on the LSU and do not randomly mix the LSU slots. Cable connections to the LSU have no bearing on NonStop Blade Complex number, but HP also recommends you connect NonStop Blade Elements A to the NonStop Blade Element A connection on the LSU. Optic cable connections to the VIO enclosure determine the identification numbers of each NonStop Blade Complex. This simplified example shows connections from a NonStop Blade Element to the LSU and to the VIO enclosure: Front Panel Buttons Button Function Condition Operation Power Hard reset Power is on. Cycle power and reset or reconfigure logic. Power is in standby. Remain in standby. TOC/NMI Soft reset Power is on. Send initialize interrupt to processors, but without reset or reconfiguration of logic. NonStop NS14000 Series Hardware 19

20 Front Panel Indicator LEDs LED Indicator Power Fault Locator State Flashing green Flashing yellow Off Steady amber Off Flashing blue Meaning Power is on; NonStop Blade Element is available for normal operation. NonStop Blade Element is in power mode. Power is off. Hardware or software fault exists. NonStop Blade Element is available for normal operation. System locator is activated. Logical Synchronization Unit (LSU) The LSU is the heart of both the highavailability duplex NonStop Blade Complex and the faulttolerant triplex NonStop Blade Complex. In NonStop NS14000 series systems, each LSU is associated with only one logical processor. The LSU is the gateway to the ServerNet fabrics for the duplex or triplex NonStop Blade Complexes. It ensures that I/O from logical processors is valid before allowing it into the ServerNet fabric. The LSU module compares I/O streams between the output of each NonStop Blade Element to ensure validity of the data. No data leaves the confines of a single NonStop Blade Element without being compared to and validated with the data from the other NonStop Blade Element (duplex processor) or the other two NonStop Blade Elements (triplex processor). The LSU enclosure, which is 4U and weighs 96 pounds (43.5 kilograms) when fully populated, has these physical attributes: Rackmountable Redundant AC power feeds Front to rear cooling Connections for three NonStop Blade Elements Connections for two ServerNet fabrics Cable management and connectivity on the rear of the cabinet Logically the LSU: Implements a fault domain that affects only a single logical processor Has environmental sense and control (ESC) aspects managed by the logical processor Supports singlepoint system poweron The LSU module consists of two types of FRUs: LSU logic board (accessible from the front of the LSU enclosure) LSU optics adapters (accessible from the rear of the LSU enclosure) AC power assembly (accessible from the rear of the LSU enclosure) CAUTION: To maintain proper cooling air flow, blank panels must be installed in all slots that do not contain logic adapter PICs or logic boards. This illustration shows an example LSU configuration as viewed from the rear of the enclosure and equipped with four LSU optics adapter PICs in positions 20 through 23: 20 NonStop NS14000 Series System Overview

21 This illustration shows an example LSU configuration as viewed from the front of the enclosure and equipped with four LSU logic boards in positions 50 through 53: LSU Indicator LEDs LED LSU optics adapter PIC (green LED) LSU optics adapter PIC (amber LED) LSU optics adapter connectors (green LEDs) State Green Off Amber Off Green Meaning Power is on; LSU is available for normal operation. Power is off. Power is in progress, board is being reset, or a fault exists. Normal operation or powered off. NonStop Blade Element optics link or ServerNet link is functional. NonStop NS14000 Series Hardware 21

22 LED LSU logic board (green LED) LSU logic board PIC (amber LED) Off Versatile I/O (VIO) Enclosure State Off Green Off Amber Normal operation or powered off. Meaning Power is off, or a fault exists (amber LED is on). Power is on with LSU available for normal operation. Power is off. Power is in progress, board is being reset, or a fault exists. NOTE: For NonStop NS14000 series servers that contain I/O Adapter Module (IOAM) enclosures, refer to the H06.07 (or earlier) version of this manual for planning, configuration, and maintenance information of the IOAM enclosure. A VIO enclosure, which is 4U, provides connectivity to the processors, Fibre Channel storage devices, and gigabit Ethernet networking. Two VIO enclosures, one required for each ServerNet fabric, are installed in a 19inch rack. Each VIO enclosure contains: Connectivity for up to eight processors, configured in pairs Up to eight Fibre Channel ports for communicating with storage devices such as a Fibre Channel Disk Module (FCDM) (page 24), tape devices, or an Enterprise Storage System (ESS) disk Up to eight copper/optical Ethernet ports used for Ethernet connectivity Connectivity to 6780 ServerNet cluster switches Two fans to provide the cooling for components inside a VIO enclosure Two power supplies with universal AC input to provide power to the components in a VIO enclosure For each VIO enclosure to operate, at least one power supply in each VIO enclosure must be operational. For fault tolerance, you must connect the power supplies in each VIO enclosure to separate power distribution units (PDUs). For example, the power supplies located in slot 15 and slot 18 of each VIO enclosure cannot both be plugged into the left PDU. The two VIO enclosures promote faulttolerant configuration and servicing: Redundant hardware provides fault tolerance if it is divided between the two VIO enclosures. Faulttolerant paths must be configured through the two VIO enclosures. This illustration shows the slot locations and component details for a single VIO enclosure: 22 NonStop NS14000 Series System Overview

23 Because each VIO enclosure is a single module, two VIO enclosures are required to provide faulttolerant paths. The paths must be configured through both VIO enclosures. For example, you could configure these paths through the Fibre Channel ports located on the two VIO enclosures: Path Group Module Slot.Port Primary Primary backup Mirror Mirror backup These paths all exist within the same group. But they are divided between two VIO enclosures, so the configuration is faulttolerant. For additional information, see Fibre Channel Devices (page 72). NonStop NS14000 Series Hardware 23

24 Fibre Channel Disk Module (FCDM) The FCDM is a rackmounted enclosure that contains: Up to 14 Fibre Channel arbitrated loop disk drives (enclosure front) Environmental monitoring unit (EMU) (enclosure rear) Fibre Channel arbitrated loop (FCAL) modules (enclosure rear) An FCDM connects to the Fibre Channel ports on a VIO enclosure. You can daisychain together up to four FCDMs to reduce the number of Fibre Channel ports required. Up to eight daisychain configurations are supported when using all embedded and optional Fibre Channel ports for a maximum of 32 FCDMs. This illustration shows a fully populated Fibre Channel disk module: Maintenance Switch (Ethernet) The ProCurve maintenance switch includes management features that the NonStop NS14000 series system requires and provides communication between the NS14000 series system through the VIO enclosures, the optional UPS, and the system console running HP NonStop Open System Management (OSM). The maintenance switch includes enough ports to support multiple systems. The maintenance switch mounts in the modular cabinet, but no restrictions exist for its placement. This illustration shows an example of two maintenance switches installed in the top of a cabinet: 24 NonStop NS14000 Series System Overview

25 Each NonStop NS14000 series system requires multiple connections to the maintenance switch. For more information, refer to the connections described in Basic LAN Configuration (page 86) and FaultTolerant Configuration (page 86). System Console A system console is an HP approved personal computer (PC) running maintenance and diagnostic software for NonStop systems. When supplied with a new NonStop system, system consoles have factoryinstalled HP and thirdparty software for managing the system. You can install software upgrades from the HP NonStop System Console Installer CD. Some system console hardware, including the PC system unit, monitor, and keyboard, can be mounted in the NonStop NS14000 series modular cabinet. Other PCs are installed outside the modular cabinet and require separate provisions or furniture to hold the PC hardware. For more information, refer to System Consoles (page 92). UPS and ERM (Optional) An uninterruptible power supply (UPS) is optional but recommended where a site UPS is not available. You can use any UPS that meets the modular cabinet power requirements for all enclosures being powered by the UPS; however, to take advantage of OSM Power Fail Support (page 98), you must use the HP R5000 UPS. For information about the requirements for installing a UPS other than the HP R5000 UPS in a NonStop NS14000 series system, see Uninterruptible Power Supply (UPS) (page 37). Cabinet configurations that include an HP R5000 UPS or HP R5500 XR UPS also have one extended runtime module (ERM). An ERM is a battery module that extends the overall batterysupported system run time. A second ERM can be added for even longer batterysupported system run time. Adding an HP R5000 UPS or R5500 XR UPS to a modular cabinet in the field requires removing the PDU on the right side of the modular cabinet and installing HP extension bars that are compatible with the UPS. A factoryinstalled UPS ships with the HP extension bars already installed on the right side of the modular cabinet. The PDU and extension bars are oriented inward, facing the components within the modular cabinet. Both the UPS and the ERM are 3U and must reside in the bottom of the cabinet. NOTE: Retrofitting a system in the field with a UPS and ERMs will likely require moving all installed enclosures in the modular cabinet to provide space for the new hardware. One or more of the enclosures that formerly resided in the modular cabinet might be displaced and therefore have to be installed in another modular cabinet that would also need a UPS and ERMs installed. Additionally, lifting equipment might be required to lift heavy enclosures to their new location. NOTE: The AC input power cord for the HP R5000 UPS or HP R5500 XR UPS is routed to exit the modular cabinet at either the top or bottom rear corners of the cabinet, depending on what is ordered for the site power feed, and the large output receptacle is unused. This illustration shows the location of an HP R5000 UPS and an ERM in a modular cabinet: NonStop NS14000 Series Hardware 25

26 This illustration shows the location of an R5500 XR UPS and an ERM in a modular cabinet: For power and environmental requirements, planning, installation, and emergency poweroff (EPO) instructions for the HP R5000 UPS or HP R5500 XR UPS, refer to the documentation shipped with the UPS. Enterprise Storage System (Optional) An Enterprise Storage System (ESS) is a collection of magnetic disks, their controllers, and a disk cache in one or more standalone cabinets. ESS connects to the NonStop NSseries systems either directly via Fibre Channel ports on the VIO enclosures (direct connect) or through a separate storage area network (SAN) using a Fibre Channel SAN switch (switched connect). For more information about these connection types, see your service provider. NOTE: The Fibre Channel SAN switch power cords might not be compatible with the modular cabinet PDU. Contact your service provider to order replacement power cords for the SAN switch that are compatible with the modular cabinet PDU. 26 NonStop NS14000 Series System Overview

27 High availability and a faulttolerant configuration for the Fibre Channel ports on the VIO enclosures are similar to the configurations required for Fibre Channel disk drives, as explained in VIO Enclosure and Disk Storage Considerations (page 72). Cables and switches vary, depending on whether the connection is direct, switched, or a combination: Connection Direct connect Switched Combination of direct and switched LCLC Cables 2 Fibre Channel ports 4 Fibre Channel ports 2 Fibre Channel ports for each direct connection4 Fibre Channel ports for each switched connection Fibre Channel Switches 0 1 or more 1 This illustration shows an example of connections between two VIO enclosures and an ESS via separate Fibre Channel switches: For fault tolerance, the primary and backup paths to an ESS logical device (LDEV) must go through different Fibre Channel switches. Some storage area procedures, such as reconfiguration, can cause the affected switches to pause. If the pause is long enough, I/O failure occurs on all paths connected to that switch. If both the primary and the backup paths are connected to the same switch, the LDEV goes down. Refer to the documentation that accompanies the ESS. Tape Drive and Interface Hardware (Optional) For an overview of tape drives and the interface hardware, see Fibre Channel Port to Tape Devices (page 72). For a list of supported tape devices, refer to the NonStop NS14000 Hardware Installation Manual. NonStop NS14000 Series Hardware 27

28 Preparation for Other Server Hardware This guide provides the specifications only for the NonStop NS14000 series modular cabinet and enclosures identified earlier in this chapter. For site preparation specifications for other HP hardware that will be installed at the site with the NonStop NS14000 series servers, consult with your HP account team. For site preparation specifications relating to hardware from other manufacturers, refer to the documentation for those devices. Component Location and Identification Terminology Topics discussed in this subsection are: Terminology (page 28) Rack and Offset Physical Location (page 29) NonStop Blade Element GroupModuleSlot Numbering (page 30) LSU GroupModuleSlot Numbering (page 31) VIO Enclosure GroupModuleSlot Numbering (page 32) Fibre Channel Disk Module GroupModuleSlot Numbering (page 33) These are terms used in locating and describing components: Term Cabinet Rack Rack Offset Group Module Slot (or Bay or Position) Port GroupModuleSlot (GMS) NonStop Blade Complex Definition Computer system housing that includes a structure of external panels, front and rear doors, internal racking, and dual PDUs. Structure integrated into the cabinet into which rackmountable components are assembled. The physical location of components installed in a modular cabinet, measured in U values numbered 1 to 42, with 1U at the bottom of the cabinet. A U is 1.75 inches (44 millimeters). A subset of a system that contains one or more modules. A group does not necessarily correspond to a single physical object, such as an enclosure. A subset of a group that is usually contained in an enclosure. A module contains one or more slots (or bays). A module can consist of components sharing a common interconnect, such as a backplane, or it can be a logical grouping of components performing a particular function. A subset of a module that is the logical or physical location of a component within that module. A connector to which a cable can be attached and which transmits and receives data. A notation method used by hardware and software in NonStop systems for organizing and identifying the location of certain hardware components. A set of two or three NonStop Blade Elements, identified as A, B, or C, and their associated LSUs. Each NonStop Blade Complex usually has four logical processors. An 8processor system employs two or three NonStop Blade Complexes.See NonStop Blade Complex (page 103). 28 NonStop NS14000 Series System Overview

29 Term NonStop Blade Element LSU Definition A physical portion of a logical processor containing up to four processor elements, Each processor element supports a different logical processor numbered 07. A component of the system that synchronizes the processor elements of a logical processor and validates all output operations from each processor element before passing the output to the ServerNet fabric. On NonStop NS14000 series systems, locations of the physical and logical modular components are identified by: Physical location: Rack number Rack offset Logical location: group, module, and slot (GMS) notation as defined by their position on the ServerNet rather than the physical location OSM uses GMS notation in many places, including the Tree view and Attributes window, and it uses rack and offset information to create displays of the server and its components. For example, in the Tree view, OSM displays the location of a power supply in a VIO enclosure in group 100, module 2, slot 15 in this form: Power Supply ( ) Rack and Offset Physical Location Rack name and rack offset identify the physical location of components in a NonStop NS14000 series system. The rack name is located on an external label affixed to the rack, which includes the system name plus a 2digit rack number. Rack offset is labeled on the rails in each side of the rack. These rails are measured vertically in units called U, with one U measuring 1.75 inches (44 millimeters). The rack is 42U, with 1U located at the bottom and 42U at the top. The rack offset is the lowest number on the rack that the component occupies. This example shows the location of NonStop Blade Element A in rack 1 with an offset of 3U and NonStop Blade Element B with an offset of 8U: Component Location and Identification 29

30 NonStop Blade Element GroupModuleSlot Numbering Processor group: 400 through 401 relates to NonStop Blade Complex 0 through 1. Example: group 401 = NonStop Blade Complex 1 Module: 1 through 3 relates to the processor NonStop Blade Element ID A through C. Example: module 2 = NonStop Blade Element B Slot: 71 through 78 relates to location of the Blade optics adapter. Example: Slot 72 = Blade optics adapter in slot 72 Port: J0 through J7 or K0 through K7 relates to the two optics ports in a specific slot. A number of GMS configurations are possible in the modular NonStop NS14000 series system. This table shows the default numbering for the logical processors: Logical Processors Group (NonStop Blade Complex) Module (NonStop Blade Element) Slot (Optics) Port 03* (A) 2 (B) 3 (C) Blade optics adapters 18 (software identified as slots 7178) J0J7, K0K7 * Logical processor 0 must be in NonStop Blade Complex 0 (group 400). All other processors can be in any userdefined group. This illustration shows GMS numbering for a NonStop Blade Element enclosure: 30 NonStop NS14000 Series System Overview

31 LSU GroupModuleSlot Numbering This table shows the default numbering for the LSUs: Item Group (NonStop Blade Complex) 1 Module I/O Position (Slot) Individual LSU J set Optics adapter (rear side, slots 2027) 2 Logic board (front side, slots 5057) Individual LSU K set Not used at this time 1 See NonStop Blade Element GroupModuleSlot Numbering (page 30). Group and module numbers correspond to the logical processor number (0, 1, 2, 3) and NonStop Blade Element (A, B, C) as determined by the ServerNet connection to the VIO enclosure. This illustration shows an example LSU configuration equipped with four optic adapters (rear side) in slots 20 through 23 and four LSU logic boards (front side) in positions 50 through 53: Component Location and Identification 31

32 VIO Enclosure GroupModuleSlot Numbering A NonStop NS14000 series system supports a single pair of VIO enclosures, identified as group 100: VIO Enclosure Item Ports Group Module Slot X Fabric Y Fabric Displayed by OSM Displayed on Chassis Fibre Channel ports Processor ports (processors 47) Not supported 4 3 Not supported 5 5 ServerNet cluster ports 1 6 6a Ethernet ports (optical) C and D (10/100/1000 Mbps) 6 6b Ethernet ports (copper) A, B (10/100 Mbps)C, D (10/100/1000 Mbps) 7 7a Ethernet ports (optical) C, D (10/100/1000 Mbps) 7 7b Ethernet ports (copper) A, B (10/100 Mbps)C, D (10/100/1000 Mbps) 7 7c Fibre Channel ports Processor ports (processors 03) , 18 Power supplies 16, 17 Fans This illustration shows the slot locations for the VIO enclosure: 32 NonStop NS14000 Series System Overview

33 Fibre Channel Disk Module GroupModuleSlot Numbering This table shows the default numbering for the Fibre Channel disk module: VIO Enclosure FCDM Group Module Slot Port Shelf Slot Item X fabric; 1 or 7c 1, if daisy 3 Y fabric chained); 1 if single disk module FCDM Disk drives bays Transceiver A1 Transceiver A2 Transceiver B1 Component Location and Identification 33

34 VIO Enclosure FCDM Group Module Slot Port Shelf Slot Item 92 Transceiver B2 93 Left FC_AL board 94 Right FC_AL board 95 Left power supply 96 Right power supply 97 Right blower 98 Left blower 99 EMU The form of the GMS numbering for a disk in a Fibre Channel disk module is: This example shows the disk in bay 03 of the Fibre Channel disk module that connects to the Fibre Channel port in the VIO enclosure group 100, module 2, slot 1, port 1: System Installation Document Packet To keep track of the hardware configuration, internal and external communications cabling, IP addresses, and connect networks, assemble and retain as the systems records an Installation Document Packet. This packet can include: Configuration Technical Document for the FactoryInstalled Hardware (page 35) Configuration Forms for the ServerNet Adapters (page 35) System Console (page 25) 34 NonStop NS14000 Series System Overview

35 Configuration Technical Document for the FactoryInstalled Hardware Each new NonStop NS14000 series system includes a document called a Technical Document. It serves as the physical location and connection map for the system and describes: Each cabinet included with the system Each hardware enclosure installed in the cabinet Cabinet U location of the bottom edge of each enclosure Each ServerNet cable with: Source and destination enclosure, component, and connector Cable part number Source and destination connection labels Configuration Forms for the ServerNet Adapters Fibre Channel ports and Ethernet ports on a VIO enclosure provide the functionality of Fibre Channel ServerNet adapters (FCSAs) and Gigabit Ethernet 4port ServerNet adapters (G4SAs) in a NonStop NS14000 series system. To add Fibre Channel and Ethernet configuration forms to your Installation Document Packet, ask your service provider to provide the necessary forms from the Versatile I/O Manual and follow any associated planning instructions. Configuration Form for a ServerNet Cluster The configuration form for installing a ServerNet cluster is located in the ServerNet Cluster Manual. System Installation Document Packet 35

36 2 Site Preparation Guidelines This chapter describes power, environmental, and space considerations for your site. Modular Cabinet Power and I/O Cable Entry Power and I/O cables can enter the Integrity NonStop NS14000 series server from either the top or the bottom rear of the modular cabinets, depending on how the cabinets are ordered from HP and the routing of the AC power feeds at the site. NonStop NS14000 series cabinets can be ordered with the AC power cords for the PDUs exiting either: Top: Power and I/O cables are routed from above the modular cabinet. Bottom: Power and I/O cables are routed from below the modular cabinet NOTE: If your system includes the optional rackmounted HP R5000 XR UPS or the HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. The R5500 XR UPS AC input power cord is routed as described above, and the large output receptacle is unused. For information about modular cabinet power and cable options, refer to Modular Cabinets (page 41). Emergency PowerOff Switches Emergency Power Off (EPO) switches are required by local codes or other applicable regulations when computer equipment contains batteries capable of supplying more than 750 voltamperes (VA) for more that five minutes. Systems that have these batteries also have internal EPO hardware for connection to a site EPO switch or relay. In an emergency, activating the EPO switch or relay removes power from all electrical equipment in the computer room (except that used for lighting and firerelated sensors and alarms). EPO Requirement for NonStop NS14000 Series Servers NonStop NS14000 series servers without an optional UPS (such as an HP R5500 XR UPS) installed in the modular cabinet do not contain batteries capable of supplying more than 750 voltamperes (VA) for more that five minutes, so they do not require connection to a site EPO switch. EPO Requirement for HP 5000 UPS or HP 5500 XR UPS The rackmounted HP 5500 XR UPS that can be optionally installed in a modular cabinet contains batteries and has an EPO circuit. For site EPO switches or relays, consult your HP site preparation specialist or electrical engineer regarding requirements. If an EPO switch or relay contactor is required for your site, contact your HP representation or refer to the HP UPS R5000 User Guide or the HP UPS R5500 XR Models User Guide for connector and wiring. This guide is available at Electrical Power and Grounding Quality Proper design and installation of a power distribution system for a NonStop NS14000 series server requires specialized skills, knowledge, and understanding of appropriate electrical codes and the limitations of the power systems for computer and data processing equipment. For power and grounding specifications, refer to Chapter 3: System Installation Specifications (page 41). 36 Site Preparation Guidelines

37 Power Quality This equipment is designed to operate reliably over a wide range of voltages and frequencies, described in Enclosure AC Input (page 59). However, damage can occur if these ranges are exceeded. Severe electrical disturbances can exceed the design specifications of the equipment. Common sources of such disturbances are: Fluctuations occurring within the facility s distribution system Utility service lowvoltage conditions (such as sags or brownouts) Wide and rapid variations in input voltage levels Wide and rapid variations in input power frequency Electrical storms Large inductive sources (such as motors and welders) Faults in the distribution system wiring (such as loose connections) Computer systems can be protected from the sources of many of these electrical disturbances by using: A dedicated power distribution system Power conditioning equipment Lightning arresters on power cables to protect equipment against electrical storms For steps to take to ensure proper power for the servers, consult with your HP site preparation specialist or power engineer. Grounding Systems The site building must provide a power distribution safety ground/protective earth for each AC service entrance to all NonStop server equipment. This safety grounding system must comply with local codes and any other applicable regulations for the installation locale. For proper grounding/protective earth connection, consult with your HP site preparation specialist or power engineer. Power Consumption In NonStop NS14000 series systems, the power consumption and inrush currents per connection can vary because of the unique combination of enclosures housed in the modular cabinet. Thus, the total power consumption for the hardware installed in the cabinet should be calculated as described in Enclosure Power Loads (page 59). Uninterruptible Power Supply (UPS) Modular cabinets do not have builtin batteries to provide power during power failures. To support system operation through a power failure, Integrity NonStop NS14000 series servers require either an optional UPS (such as the HP R5000 UPS or HP R5500 XR UPS) installed in each modular cabinet or a site UPS to support system operation through a power failure. If using the HP R5000 UPS or HP R5500 XR UPS, you can take advantage of OSM support for a planned orderly shutdown at a predetermined time in the event of an extended power failure. A timely and orderly shutdown prevents an uncontrolled and asymmetric shutdown of the system resources from depleted UPS batteries. For additional information, see OSM Power Fail Support (page 98). You can order an optional HP R5000 UPS for each modular cabinet to supply power to the enclosures within that cabinet. Up to two extended runtime modules (ERMs) can be included with the HP R5000 UPS or HP R5500 XR UPS to extend the power backup time. If you add an HP R5000 UPS or HP R5500 XR UPS to a modular cabinet in the field, the PDU on the right side is Uninterruptible Power Supply (UPS) 37

38 replaced with HP extension bars. The extension bars are oriented inward, facing the components within the cabinet. For complete information and specifications on the HP R5000 UPS or HP R5500 XR UPS, contact your HP representation or refer to the HP UPS R5000 Models User Guide or the HP UPS R5500 XR Models User Guide available at If you install a UPS other than the HP R5000 UPS or HP R5500 XR UPS in each modular cabinet of a NonStop NS14000 series system, these requirements must be met to insure the system can survive a total AC power fail: The UPS output voltage can support the HP PDU input voltage requirements. The UPS phase output matches the PDU phase input. Both 1phase and 3phase output UPSs are supported. Both 1phase and 3phase input HP PDUs are supported. The UPS output can support the targeted system in the event of an AC power failure. Calculate each cabinet load to insure the UPS can support a proper ridethrough time in the event of a total AC power failure. NOTE: A UPS other than the HP R5000 UPS or HP R5500 XR UPS will not be able to utilize the OSM Power Fail Support function. If your applications require a UPS that supports the entire system or even a UPS or motor generator for all computer and support equipment in the site, you must plan the site s electrical infrastructure accordingly. Cooling and Humidity Control Do not rely on an intuitive approach to design cooling or to simply achieve an energy balance that is, summing up to the total power dissipation from all the hardware and sizing a comparable air conditioning capacity. Today s highperformance servers use semiconductors that integrate multiple functions on a single chip with very high power densities. These chips, plus highpowerdensity mass storage and power supplies, are mounted in ultrathin server and storage enclosures, and then deployed into modular cabinets in large numbers. This higher concentration of devices results in localized heat, which increases the potential for hot spots that can damage the equipment. Additionally, variables in the installation site layout can adversely affect air flows and create hot spots by allowing hot and cool air streams to mix. Studies have shown that above 70 F (20 C), every increase of 18 F (10 C) reduces longterm electronics reliability by 50%. Cooling airflow through each enclosure in the NonStop NS14000 series server is fronttoback. Because of high heat densities and hot spots, an accurate assessment of air flow around and through the server equipment and specialized cooling design is essential for reliable server operation. For an airflow assessment, consult with your HP cooling consultant or your heating, ventilation, and air conditioning (HVAC) engineer. NOTE: Failure of site cooling with the server continuing to run can cause rapid heat buildup and excessive temperatures within the hardware. Excessive internal temperatures can result in full or partial system shutdown. Ensure that the site s cooling system remains fully operational when the server is running. Because each modular cabinet houses a unique combination of enclosures, use the Heat Dissipation Specifications and Worksheet (page 64) to calculate the total heat dissipation for the hardware installed in each cabinet. For air temperature levels at the site, refer to Operating Temperature, Humidity, and Altitude (page 64). 38 Site Preparation Guidelines

39 Weight Flooring Because modular cabinets for NonStop NS14000 series servers house a unique combination of enclosures, total weight must be calculated based on what is in the specific cabinet, as described in Modular Cabinet and Enclosure Weights With Worksheet (page 62). NonStop NS14000 series servers can be installed either on the site s floor with the cables entering from above the equipment or on raised flooring with power and I/O cables entering from underneath. Because cooling airflow through each enclosure in the modular cabinets is fronttoback, raised flooring is not required for system cooling. The site floor structure and any raised flooring (if used) must be able to support the total weight of the installed computer system as well as the weight of the individual modular cabinets and their enclosures as they are moved into position. To determine the total weight of each modular cabinet with its installed enclosures, refer to Modular Cabinet and Enclosure Weights With Worksheet (page 62). For your site s floor system, consult with your HP site preparation specialist or an appropriate floor system engineer. If raised flooring is to be used, the design of the NonStop NS14000 series modular cabinet is optimized for placement on 24inch floor panels. Dust and Pollution Control NonStop NS14000 series servers do not have air filters. Any computer equipment can be adversely affected by dust and microscopic particles in the site environment. Airborne dust can blanket electronic components on printed circuit boards inhibiting cooling airflow and causing premature failure from excess heat, humidity, or both. Metallically conductive particles can short circuit electronic components. Tape drives and some other mechanical devices can experience failures resulting from airborne abrasive particles. For recommendations to keep the site as free of dust and pollution as possible, consult with your heating, ventilation, and air conditioning (HVAC) engineer or your HP site preparation specialist. Zinc Particulates Over time, fine whiskers of pure metal can form on electroplated zinc, cadmium, or tin surfaces such as aged raised flooring panels and supports. If these whiskers are disturbed, they can break off and become airborne, possibly causing computer failures or operational interruptions. This metallic particulate contamination is a relatively rare but possible threat. Kits are available to test for metallic particulate contamination, or you can request that your site preparation specialist or HVAC engineer test the site for contamination before installing any electronic equipment. Space for Receiving and Unpacking Identify areas that are large enough to receive and to unpack the system from its shipping cartons and pallets. Be sure to allow adequate space to remove the system equipment from the shipping pallets using supplied ramps. Also be sure adequate personnel are present to remove each cabinet from its shipping pallet and to safely move it to the installation site. WARNING! A fully populated cabinet is unstable when moving down the unloading ramp from its shipping pallet. Arrange for enough personnel to stabilize each cabinet during removal from the pallet and to prevent the cabinet from falling. A falling cabinet can cause serious or fatal personal injury. Ensure sufficient pathways and clearances for moving the server equipment safely from the receiving and unpacking areas to the installation site. Verify that door and hallway width and height as well as floor and elevator loading will accommodate not only the server equipment but also all required Weight 39

40 personnel and lifting or moving devices. If necessary, enlarge or remove any obstructing doorway or wall. All modular cabinets have small casters to facilitate moving them on hard flooring from the unpacking area to the site. Because of these small casters, rolling modular cabinets along carpeted or tiled pathways might be difficult. If necessary, plan for a temporary hard floor covering in affected pathways for easier movement of the equipment. For physical dimensions of the server equipment, refer to Dimensions and Weights (page 60). Operational Space When planning the layout of the server site, use the equipment dimensions, door swing, and service clearances listed in Dimensions and Weights (page 60). Because location of the lighting fixtures and electrical outlets affects servicing operations, consider an equipment layout that takes advantage of existing lighting and electrical outlets. Also consider the location and orientation of current or future air conditioning ducts and airflow direction and eliminate any obstructions to equipment intake or exhaust air flow. Refer to Cooling and Humidity Control (page 38). Space planning should also include the possible addition of equipment or other changes in space requirements. Depending on the current or future equipment installed at your site, layout plans can also include provisions for: Channels or fixtures used for routing data cables and power cables Access to air conditioning ducts, filters, lighting, and electrical power hardware Communications cables, patch panels, and switch equipment Power conditioning equipment Storage area or cabinets for supplies, media, and spare parts 40 Site Preparation Guidelines

41 3 System Installation Specifications All specifications provided in this chapter assume that each enclosure in the modular cabinet is fully populated; for example, a NonStop Blade Element with four processors and maximum memory. The maximum current for each AC service depends on the number and type of enclosures installed in the modular cabinet. Power, weight, and heat loads are less when enclosures are not fully populated; for example, a NonStop Blade Element with fewer processors or a Fibre Channel disk module with fewer disks. Modular Cabinets The modular cabinet is a EIA standard 19inch, 42U, rack for mounting modular components. The modular cabinet comes equipped with front and rear doors and includes a rear extension that makes it deeper than some industrystandard racks. The Power Distribution Units (PDUs) are mounted along the rear extension without occupying any Uspace in the cabinet and are oriented inward, facing the components within the modular cabinet. NOTE: For instructions on grounding the modular cabinet's G2 rack using the HP Rack Grounding Kit (AF074A), ask your service provider to refer to the instructions in the HP G2 Series Rack Options Installation Guide. This manual is available at: bc/docs/support/supportmanual/c /c pdf. Depending on your NS14000 power configuration, your system uses one of these PDU types: Monitored SinglePhase PDUs Monitored ThreePhase PDUs (page 46) Modular ThreePhase PDUs (page 51) NOTE: If your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Each PDU is wired to distribute the load segments to its receptacles. CAUTION: If you are installing NonStop NS14000 series enclosures in a modular cabinet, balance the current load among the available load segments. Using only one of the available load segments, especially for larger systems, can cause unbalanced loading and might violate applicable electrical codes. Connecting the two power plugs from an enclosure to the same load segment causes failure of the hardware if that load segment fails. Monitored SinglePhase PDUs Two monitored singlephase power distribution units (PDUs) are installed to provide redundant power outlets for the components mounted in the modular cabinet. The PDUs are oriented inward, facing the components within the modular cabinet. Each PDU is 60 inches long and has 39 AC receptacles, three circuit breakers, and an AC power cord. The PDU is oriented with the AC power cord exiting the modular cabinet at either the top or bottom rear corners of the cabinet, depending on the sites power feed needs. For information about specific characteristics for PDUs factoryinstalled in monitored singlephase cabinets, refer to AC Power Feeds, Monitored SinglePhase PDUs (page 42) Input and Output Power Characteristics, Monitored SinglePhase PDUs (page 45) Branch Circuits and Circuit Breakers, Monitored SinglePhase PDUs (page 46) Modular Cabinets 41

42 Each singlephase PDU in a modular cabinet has: 36 AC receptacles per PDU (12 per segment) IEC 320 C13 12A receptacle type 3 AC receptacles per PDU (1 per segment) IEC 320 C19 16A receptacle type 3 circuitbreakers These PDU options are available to receive power from the site AC power source: 200 to 240 V AC, singlephase for North America and Japan 200 to 240 V AC singlephase for International Each PDU distributes the site AC power as single phase 200 to 240 V AC to the 39 outlets for connecting the power cords from the components mounted in the modular cabinet. AC Power Feeds, Monitored SinglePhase PDUs Power can enter the NonStop NS14000 series server from either the top or the bottom rear of the modular cabinets, depending on how the cabinets are ordered from HP and the AC power feeds are routed at the site. NonStop NS14000 series server cabinets can be ordered with the AC power cords for the PDU installed either: Top: Power and I/O cables are routed from above the modular cabinet. Bottom: Power and I/O cables are routed from below the modular cabinet For information on the modular cabinets, refer to Modular Cabinets (page 41). The AC power feed cables for the PDUs are mounted to exit the modular cabinet at either the top or bottom rear corners of the cabinet depending on what is ordered for the site power feed. Top AC Power Feed, Monitored SinglePhase PDUs Figure 2 shows the AC power feed cables on PDUs for AC feed at the top of the cabinet: Figure 2 Top AC Power Feed, Monitored SinglePhase PDUs 42 System Installation Specifications

43 Bottom AC Power Feed, Monitored SinglePhase PDUs Figure 3 shows the AC power feed cables on PDUs for AC feed at the bottom of the cabinet and the AC power outlets along the PDU. These power outlets face in toward the cabinet: Figure 3 Bottom AC Power Feed, Monitored SinglePhase PDUs Top AC Power Feed When Optional UPS and ERM are Installed If your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear right side and four extension bars on the rear left side. The PDU and extension bars are oriented inward, facing the components within the modular cabinet. To provide redundancy, components are plugged into the rightside PDU and the extension bars. Each extension bar is plugged into the UPS. Figure 4 shows the AC power feed cables for the PDU and UPS for AC power feed from the top of the cabinet when the optional UPS and ERM are installed. Also see UPS and ERM (Optional) (page 25). Monitored SinglePhase PDUs 43

44 Figure 4 Top AC Power Feed When Optional UPS and ERM are Installed Bottom AC Power Feed When Optional UPS and ERM are Installed If your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear right side and four extension bars on the rear left side. The PDU and extension bars are oriented inward, facing the components within the modular cabinet. To provide redundancy, components are plugged into the rightside PDU and the extension bars. Each extension bar is plugged into the UPS. Figure 5 shows the AC power feed cables for the PDU and UPS for AC power feed from the bottom of the cabinet when the optional UPS and ERM are installed. Also see UPS and ERM (Optional) (page 25). 44 System Installation Specifications

45 Figure 5 Bottom AC Power Feed When Optional UPS and ERM are Installed Input and Output Power Characteristics, Monitored SinglePhase PDUs The cabinet includes two monitored singlephase PDUs. North America and Japan: 200 to 240 V AC, Monitored SinglePhase PDUs The North America and Japan PDU power characteristics are: PDU input characteristics 200 to 240 V AC, single phase, 40A RMS, 3wire 50/60Hz NonNEMA Locking CS8265C, 50A input plug 6.5 feet (2 m) attached power cord PDU output characteristics 3 circuitbreakerprotected 20A load segments 36 AC receptacles per PDU (12 per segment) IEC 320 C13 12A receptacle type 3 AC receptacles per PDU (1 per segment) IEC 320 C19 16A receptacle type Monitored SinglePhase PDUs 45

46 International: 200 to 240 V AC, Monitored SinglePhase PDUs The international PDU power characteristics are: PDU input characteristics 200 to 240 V AC, single phase, 32A RMS, 3wire 50/60Hz IEC309 3pin, 32A input plug 6.5 feet (2 m) attached harmonized power cord PDU output characteristics 3 circuitbreakerprotected 20A load segments 36 AC receptacles per PDU (12 per segment) IEC 320 C13 12A receptacle type 3 AC receptacles per PDU (1 per segment) IEC 320 C19 16A receptacle type Branch Circuits and Circuit Breakers, Monitored SinglePhase PDUs Modular cabinets for the NonStop NS14000 series system contain two PDUs. In cabinets without the optional rackmounted UPS, each of the two PDUs requires a separate branch circuit of these ratings: Region North America and Japan International 1 Category D circuit breaker is required. Volts (PhasetoPhase) 200 to to 240 Amps (see following CAUTION ) CAUTION: Be sure the hardware configuration and resultant power loads of each cabinet within the system do not exceed the capacity of the branch circuit according to applicable electrical codes and regulations. NOTE: If your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Branch circuit requirements vary by the input voltage and the local codes and applicable regulations regarding maximum circuit and total distribution loading. Select circuit breaker ratings according to local codes and any applicable regulations for the circuit capacity. Note that circuit breaker ratings vary if your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR Integrated UPS (see Circuit Breaker Ratings for UPS (page 58)). Monitored ThreePhase PDUs Two monitored threephase power distribution units (PDUs) are installed to provide redundant power outlets for the components mounted in the modular cabinet. The PDUs are oriented inward, facing the components within the modular cabinet. Each PDU is 60 inches long and has 39 AC receptacles, three circuit breakers, and an AC power cord. The PDU is oriented with the AC power cord exiting the modular cabinet at either the top or bottom rear corners of the cabinet, depending on the sites power feed needs. 46 System Installation Specifications

47 For information about specific characteristics for PDUs factoryinstalled in monitored threephase cabinets, refer to AC Power Feeds, Monitored ThreePhase PDUs (page 47) Input and Output Power Characteristics, Monitored ThreePhase PDUs (page 50) Branch Circuits and Circuit Breakers, Monitored ThreePhase PDUs (page 51) Each threephase monitored PDU in a modular cabinet has: 36 AC receptacles per PDU (12 per segment) IEC 320 C13 12A receptacle type 3 AC receptacles per PDU (1 per segment) IEC 320 C19 16A receptacle type 3 circuitbreakers These PDU options are available to receive power from the site AC power source: 200 to 240 V AC, threephase delta for North America and Japan 380 to 415 V AC, threephase wye for International Each PDU distributes the site AC power as threephase 200 to 240 V AC to the 39 outlets for connecting the power cords from the components mounted in the modular cabinet. AC Power Feeds, Monitored ThreePhase PDUs Power can enter the NonStop NS14000 series server from either the top or the bottom rear of the modular cabinets, depending on how the cabinets are ordered from HP and the AC power feeds are routed at the site. NonStop NS14000 series server cabinets can be ordered with the AC power cords for the PDU installed either: Top: Power and I/O cables are routed from above the modular cabinet. Bottom: Power and I/O cables are routed from below the modular cabinet The AC power feed cables for the PDUs are mounted to exit the modular cabinet at either the top or bottom rear corners of the cabinet depending on what is ordered for the site power feed. Top AC Power Feed, Monitored ThreePhase PDUs Figure 6 shows the AC power feed cables on PDUs for AC feed at the top of the cabinet: Monitored ThreePhase PDUs 47

48 Figure 6 Top AC Power Feed, Monitored ThreePhase PDUs Bottom AC Power Feed, Monitored ThreePhase PDUs Figure 7 shows the AC power feed cables on PDUs for AC feed at the bottom of the cabinet and the AC power outlets along the PDU. These power outlets face in toward the cabinet: Figure 7 Bottom AC Power Feed, Monitored ThreePhase PDUs Top AC Power Feed When Optional UPS and ERM are Installed If your system includes the optional rackmounted HP UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. The PDU and extension 48 System Installation Specifications

49 bars are oriented inward, facing the components within the modular cabinet. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Figure 8 shows the AC power feed cables for the PDU and UPS for AC power feed from the top of the cabinet when the optional UPS and ERM are installed: Figure 8 Top AC Power Feed When Optional UPS and ERM are Installed Bottom AC Power Feed When Optional UPS and ERM are Installed Figure 9 shows the AC power feed cables for the PDU and UPS for AC power feed from the bottom of the cabinet when the optional UPS and ERM are installed: Monitored ThreePhase PDUs 49

50 Figure 9 Bottom AC Power Feed When Optional UPS and ERM are Installed Input and Output Power Characteristics, Monitored ThreePhase PDUs The cabinet includes two monitored threephase PDUs. North America and Japan: 200 to 240 V AC, Monitored ThreePhase PDUs The North America and Japan PDU power characteristics are: PDU input characteristics 200 to 240 V AC, 3 phase delta, 30A, 4wire 50/60Hz NEMA L1530 input plug 6.5 feet (2 m) attached power cord PDU output characteristics 3 circuitbreakerprotected 13.86A load segments 36 AC receptacles per PDU (12 per segment) IEC 320 C13 12A receptacle type 3 AC receptacles per PDU (1 per segment) IEC 320 C19 16A receptacle type 50 System Installation Specifications

51 International: 380 to 415 V AC, Monitored ThreePhase PDUs The international PDU power characteristics are: PDU input characteristics 380 to 415 V AC, 3phase Wye, 16A RMS, 5wire 50/60Hz IEC309 5pin, 16A input plug 6.5 feet (2 m) attached harmonized power cord PDU output characteristics 3 circuitbreakerprotected 16A load segments 36 AC receptacles per PDU (12 per segment) IEC 320 C13 12A receptacle type 3 AC receptacles per PDU (1 per segment) IEC 320 C19 16A receptacle type Branch Circuits and Circuit Breakers, Monitored ThreePhase PDUs Modular cabinets for the NonStop NS14000 series system that use a threephase power configuration with monitored threephase PDUs contain two PDUs. In cabinets without the optional rackmounted UPS, each of the two PDUs requires a separate branch circuit of these ratings: Region North America and Japan International 1 Category D circuit breaker is required. Volts (PhasetoPhase) 200 to to 415 Amps (see following CAUTION ) CAUTION: Be sure the hardware configuration and resultant power loads of each cabinet within the system do not exceed the capacity of the branch circuit according to applicable electrical codes and regulations. NOTE: If your system includes the optional rackmounted HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Branch circuit requirements vary by the input voltage and the local codes and applicable regulations regarding maximum circuit and total distribution loading. Select circuit breaker ratings according to local codes and any applicable regulations for the circuit capacity. Select circuit breaker ratings according to local codes and any applicable regulations for the circuit capacity. Note that circuit breaker ratings vary if your system includes the optional rackmounted HP R5500 XR Integrated UPS. Modular ThreePhase PDUs Two threephase modular power distribution units (PDUs) are installed to provide redundant power outlets for the components mounted in the modular cabinet. Each 1U rackmounted modular PDU comes with four modular PDU extension bars. The PDUs are oriented facing each other within the rack. Each PDU has 28 AC receptacles, six circuit breakers, and an AC power cord. The PDU is oriented with the AC power cord exiting the modular cabinet at either the top or bottom rear corners of the cabinet, depending on the site's power feed needs. Modular ThreePhase PDUs 51

52 For information about specific characteristics for PDUs factoryinstalled in modular threephase cabinets, refer to AC Power Feeds, Modular ThreePhase PDUs (page 52) Input and Output Power Characteristics, Modular ThreePhase PDUs (page 56) Branch Circuits and Circuit Breakers, Modular ThreePhase PDUs (page 57) Each threephase modular PDU in a modular cabinet has: 28 AC receptacles per PDU (7 per extension bar) IEC 320 C13 12A receptacle type 6 circuitbreakers These PDU options are available to receive power from the site AC power source: 200 to 240 V AC, threephase delta for North America and Japan 380 to 415 V AC, threephase wye for International Each PDU distributes site threephase power to 34 singlephase 200 to 240 V AC outlets for connecting the power cords from the components mounted in the modular cabinet. AC Power Feeds, Modular ThreePhase PDUs Power can enter the NonStop NS14000 series server from either the top or the bottom rear of the modular cabinets, depending on how the cabinets are ordered from HP and the AC power feeds are routed at the site. NonStop NS14000 series server cabinets can be ordered with the AC power cords for the PDU installed either: Top: Power and I/O cables are routed from above the modular cabinet. Bottom: Power and I/O cables are routed from below the modular cabinet The AC power feed cables for the PDUs are mounted to exit the modular cabinet at either the top or bottom rear corners of the cabinet depending on what is ordered for the site power feed. Top AC Power Feed, Modular ThreePhase PDUs Figure 10 shows the threephase modular PDUs with AC feed at the top of the cabinet. 52 System Installation Specifications

53 Figure 10 Top AC Power Feed, Modular ThreePhase PDUs Bottom AC Power Feed, Modular ThreePhase PDUs Figure 11 shows the power feed cables on modular threephase PDUs with AC feed at the bottom of the cabinet and the output connections for the threephase modular PDU. Modular ThreePhase PDUs 53

54 Figure 11 Bottom AC Power Feed, Modular ThreePhase PDUs Top AC Power Feed When Optional UPS and ERM are Installed If your system includes the optional rackmounted HP UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. The PDU and extension bars are oriented inward, facing the components within the modular cabinet. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Figure 12 shows the AC power feed cables for the PDU and UPS for AC power feed from the top of the cabinet when the optional UPS and ERM are installed. Also see UPS and ERM (Optional) (page 25). 54 System Installation Specifications

55 Figure 12 Top AC Power Feed When Optional UPS and ERM are Installed Bottom AC Power Feed When Optional UPS and ERM are Installed Figure 13 shows the AC power feed cables for the PDU and UPS for AC power feed from the bottom of the cabinet when the optional UPS and ERM are installed. Also see UPS and ERM (Optional) (page 25). Modular ThreePhase PDUs 55

56 Figure 13 Bottom AC Power Feed When Optional UPS and ERM are Installed Input and Output Power Characteristics, Modular ThreePhase PDUs The cabinet includes two modular threephase PDUs. North America and Japan: 200 to 240 V AC, Modular ThreePhase PDUs and Extension Bars The North America and Japan PDU power characteristics are: PDU input characteristics 200 to 240 V AC, 3phase delta, 30A, 4wire 50/60Hz NEMA L1530 input plug 12 feet (3.6 m) attached power cord PDU output characteristics 6 IEC 320 C19 receptacles per PDU with 20A circuitbreaker labels (L1, L2, L3, L4, L5, and L6) Extension bar input characteristics 200V to 240 V AC, 3phase delta, 16A RMS, 4wire 50/60Hz IEC 320 C20 input plug 6.5 feet (2.0 m) attached power cord Extension bar output characteristics 7 IEC 320 C13 receptacles per PDU with 12A maximum per outlet 56 System Installation Specifications

57 International: 380 to 415 V AC, Modular ThreePhase PDUs and Extension Bars The international PDU power characteristics are: PDU input characteristics 380 to 415 V AC, 3phase Wye, 16A RMS, 5wire 50/60Hz IEC309 5pin, 16A input plug 12 feet (3.6 m) attached power cord PDU output characteristics 6 AC IEC 320 C19 receptacles per PDU with 20A circuitbreaker labels (L1, L2, L3, L4, L5, and L6) Extension bar input characteristics 200V to 240 V AC, 3phase delta, 16A max RMS, 4wire 50/60Hz IEC 320 C20 input plug 6.5 feet (2.0 m) attached power cord Extension bar output characteristics 7 AC IEC 320 C13 receptacles per PDU with 12A maximum per outlet Branch Circuits and Circuit Breakers, Modular ThreePhase PDUs Modular cabinets for the NonStop NS14000 series system that use a threephase power configuration with modular threephase PDUs contain two PDUs. In cabinets without the optional rackmounted UPS, each of the two PDUs requires a separate branch circuit of these ratings: Region North America and Japan International 1 Category D circuit breaker is required. Volts (PhasetoPhase) 200 to to 415 Amps (see following CAUTION ) CAUTION: Be sure the hardware configuration and resultant power loads of each cabinet within the system do not exceed the capacity of the branch circuit according to applicable electrical codes and regulations. NOTE: If your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Branch circuit requirements vary by the input voltage and the local codes and applicable regulations regarding maximum circuit and total distribution loading. Select circuit breaker ratings according to local codes and any applicable regulations for the circuit capacity. Select circuit breaker ratings according to local codes and any applicable regulations for the circuit capacity. Note that circuit breaker ratings vary if your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR Integrated UPS. Modular ThreePhase PDUs 57

58 Circuit Breaker Ratings for UPS These ratings apply to systems with the optional rackmounted HP R5000 UPS Integrated UPS that is used for a singlephase power configuration: Version Operating Voltage Settings Power Out (VA/Watts) Input Plug UPS Input Rating 1 North America and Japan 200/208 2, 220/230/ /4500 L630P Dedicated 30 Amp International 200/208/220/230 3, /4500 IEC Amp Dedicated 30 Amp If set at 200/208 Then 5000/ The UPS input requires a dedicated (unshared) branch circuit that is suitably rated for your specific UPS. 2 Factorydefault setting 3 Factorydefault setting For further information and specifications on the HP R5000 UPS, refer to the HP R5000 UPS User Guide. This guide is available at SupportManual/c /c pdf These ratings apply to systems with the optional rackmounted HP R5500 XR Integrated UPS that is used for a singlephase power configuration: Version Operating Voltage Settings Power Out (VA/Watts) Input Plug UPS Input Rating 1 North America and Japan 200/208 2, 220, 230, /4500 L630P Dedicated 30 Amp International 200, 230 3, /4500 IEC Amp Dedicated 30 Amp If set at 200/208 Then 5000/ The UPS input requires a dedicated (unshared) branch circuit that is suitably rated for your specific UPS. 2 Factorydefault setting 3 Factorydefault setting For further information and specifications on the R5500 XR UPS, refer to the HP R5500 UPS User Guide. This guide is available at: PDU Strapping Configurations Grounding PDUs are available in four static strapping configurations that are factoryinstalled in a modular cabinet. The specific PDU strapping configuration for a particular site depends on the type and voltage of AC power at the intended installation site for the system. A safety ground/protective earth conductor is required for each AC service entrance to the NonStop server equipment. This ground must comply with local codes and any other applicable regulations. 58 System Installation Specifications

59 Enclosure AC Input Enclosures (NonStop Blade Element, VIO enclosure, and so forth) require: Specification Nominal input voltage Voltage range 1 Nominal line frequency Frequency ranges Value 200/208/220/230/240 V AC RMS V AC 50 or 60 Hz 4753 Hz or 5763 Hz Number of phases 1 Voltage range for the VIO enclosure is V AC, and for the maintenance switch is V AC. 1 Each PDU is wired to distribute the load segments to its receptacles. Factoryinstalled enclosures are connected to the PDUs for a balanced load among the load segments. CAUTION: If you are installing NonStop NS14000 series system enclosures in a modular cabinet, balance the current load among the available load segments. Using only one of the available load segments, especially for larger systems, can cause unbalanced loading and might violate applicable electrical codes. Connecting the two power plugs from an enclosure to the same load segment causes failure of the hardware if that load segment fails. Enclosure Power Loads The total power and current load for each modular cabinet depends on the number and type of enclosures installed in it. Therefore, the total load is the sum of the loads for all enclosures installed. For examples of calculating the power and current load for various enclosure combinations, refer to Calculating Specifications for Enclosure Combinations (page 65). In normal operation, the AC power is split equally between the two PDUs in the modular cabinet. However, if one of the two AC power feeds fails, the remaining AC power feed and PDU must carry the power for all enclosures in that cabinet. NOTE: If your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Power and current specifications for each type of enclosure are: Enclosure Type AC Power Lines per Enclosure 1 Typical Power Consumption (VA) Maximum Power Consumption (VA) Peak Inrush Current (amps) NonStop Blade Element chassis Processor assembly, 2P NA NA Processor assembly, 4P NA NA LSU enclosure VIO enclosure Fibre Channel Disk Module (no disk) Fibre Channel disk drive NA NA Enclosure AC Input 59

60 Enclosure Type AC Power Lines per Enclosure 1 Typical Power Consumption (VA) Maximum Power Consumption (VA) Peak Inrush Current (amps) Rackmounted system console (NSCR4) Rackmounted system console (NSCR110) Rackmounted keyboard and monitor M Maintenance switch M Maintenance switch Half of the plugs for an enclosure must be connected to the leftside PDU and the other half connected to the rightside PDU or extension bars (if the optional UPS is installed). PDUs must be supplied from separate branch circuits. 2 Measured with 4 LSU Voter Logic & Optics installed and active. Each LSU Voter Logic & Optics consumes 55 W. 3 Measured with PICs installed in slots 1, 2, 5, 6a/7a, 7b, and 7c; PIC in slot 7c had two external connections. Sysex was running. Fans were running at full speed. M88014 consumes 20 W M88024 consumes 12 W M88404 consumes 7 W 4 These components have only one plug. If a UPS is installed in the modular cabinet, the component's plug must be connected to the extension bars on the right side of the modular cabinet. Dimensions and Weights This subsection provides information about the dimensions and weights for modular cabinets and enclosures installed in a modular cabinet and covers these topics: Service Clearances for the Modular Cabinet (page 61) Unit Sizes (page 61) Modular Cabinet Physical Specifications (page 61) Enclosure Dimensions (page 62) Modular Cabinet and Enclosure Weights With Worksheet (page 62) 60 System Installation Specifications

61 Plan View From Above the Modular Cabinet Service Clearances for the Modular Cabinet Unit Sizes Aisles: 6 feet (182.9 centimeters) Front: 3 feet (91.4 centimeters) Rear: 3 feet (91.4 centimeters) Enclosure Type Modular cabinet NonStop Blade Element LSU VIO enclosure Fibre Channel disk module Maintenance switch (Ethernet) R5000 UPS or R5500 XR UPS (singlephase power) R5000 ERM or R5500 XR ERM (extended runtime module) Rackmount console Height (U) Modular Cabinet Physical Specifications Item Height Width Depth Weight in. cm in. cm in. cm Modular Cabinet (HP G2 Series rack Depends on the enclosures installed. Refer Dimensions and Weights 61

62 Item Height Width Depth Weight in. cm in. cm in. cm with extension, doors, and side panels) Rack to Modular Cabinet and Enclosure Weights With Worksheet (page 62). Front door Leftrear door Rightrear door Shipping (palletized) Enclosure Dimensions Enclosure Type Height Width Depth in cm in cm in cm NonStop Blade Element Processor switch LSU VIO enclosure Fibre Channel disk module Maintenance switch (Ethernet) Rackmount console system unit Rackmount console system unit with keyboard and display R5000 UPS R5000 ERM R5500 XR UPS R5500 XR ERM Modular Cabinet and Enclosure Weights With Worksheet The total weight of each modular cabinet is the sum the weights of the cabinet plus each enclosure installed in it. Use this worksheet to determine the total weight: Enclosure Type Number of Enclosures Weight lbs kg Total lbs kg 42U (singlephase with modular PDUs)* U (threephase with modular PDUs)* U (threephase with monitored PDUs)* NonStop Blade Element Processor switch LSU System Installation Specifications

63 Enclosure Type Number of Enclosures Weight lbs kg Total lbs kg VIO enclosure Fibre Channel disk module Maintenance switch (Ethernet) Rackmount console system unit with keyboard and display R5000 UPS R5000 ERM R5500 XR UPS (singlephase power) R5500 ERM Total * Modular cabinet weight includes the PDUs and their associated wiring and receptacles. For examples of calculating the weight for various enclosure combinations, refer to Calculating Specifications for Enclosure Combinations (page 65). Modular Cabinet Stability Cabinet stabilizers are required when you have less than four cabinets bayed together. NOTE: Cabinet stability is of special concern when equipment is routinely installed, removed, or accessed within the cabinet. Stability is addressed through the use of leveling feet, baying kits, fixed stabilizers, and/or ballast. For information about best practices for cabinets, see: HP G2 Series Rack User Guide Best practices for HP Series and HP G2 Series Racks Environmental Specifications This subsection provides information about environmental specifications and covers these topics: Heat Dissipation Specifications and Worksheet (page 64) Operating Temperature, Humidity, and Altitude (page 64) Nonoperating Temperature, Humidity, and Altitude (except R5000 UPS and R5000 ERM) (page 65) Nonoperating Temperature, Humidity, and Altitude for R5000 UPS and R5000 ERM (page 65) Cooling Airflow Direction (page 65) Typical Acoustic Noise Emissions (page 65) Tested Electrostatic Immunity (page 65) Dimensions and Weights 63

64 Heat Dissipation Specifications and Worksheet Enclosure Type Number Installed Unit Heat (Btu/hour with single AC line powered) Unit Heat (Btu/hour with both AC lines powered) Total (BTU/hour) 4processor NonStop Blade Element with 16 GB memory boards processor NonStop Blade Element 32 GB memory boards LSU (with four LSU boards) VIO enclosure Fibre Channel disk module Maintenance switch (Ethernet) Rackmount console system unit with keyboard and display Measured with PICs installed in slots 1, 2, 5, 6a/7a, 7b, and 7c; PIC in slot 7c had two external connections. Sysex was running. Fans were running at full speed. 2 Measured with 14 disk drives installed and active. 3 Maintenance switch has only one plug. If a UPS is installed in the modular cabinet, the maintenance switch plug must be connected to the extension bars on the right side of the modular cabinet. Operating Temperature, Humidity, and Altitude Specification Operating Range 1 Recommended Range 1 Maximum Rate of Change per Hour Temperature (all except Fibre Channel disk module, R5000 UPS, and R5000 ERM) 41 to 95 F (5 to 35 C) 68 to 77 F (20 to 25 C) 9 F (5 C) Repetitive 36 F (20 C) Nonrepetitive Temperature (Fibre Channel disk module) 50 to 95 F (10 to 35 C) 1.8 F (1 C) Repetitive 5.4 F(3 C) Nonrepetitive Temperature (R5000 UPS and R5000 ERM) 50 to 104 F (10 to 40 C) Humidity (all except R5000 UPS and R5000 ERM) 15% to 80%, noncondensing 40% to 50%, noncondensing 6%, noncondensing Humidity (R5000 UPS and R5000 ERM) 20% to 80%, noncondensing Altitude (all except R5000 UPS and R5000 ERM) 2 0 to 10,000 feet (0 to 3,048 meters) Altitude (R5000 UPS and R5000 ERM) 0 to 6,600 feet (0 to 2,000 meters) 1 Operating and recommended ranges refer to the ambient air temperature and humidity measured 19.7 in. (50 cm) from the front of the air intake cooling vents. 2 For each 1000 feet (305 m) increase in altitude above 10,000 feet (up to a maximum of 15,000 feet), subtract 1.5 F (0.83 C) from the upper limit of the operating and recommended temperature ranges. 64 System Installation Specifications

65 Nonoperating Temperature, Humidity, and Altitude (except R5000 UPS and R5000 ERM) Temperature: Up to 72hour storage: 40 to 151 F (40 to 66 C) Up to 6month storage: 20 to 131 F (29 to 55 C) Reasonable rate of change with noncondensing relative humidity during the transition from warm to cold Relative humidity: 10% to 80%, noncondensing Altitude: 0 to 40,000 feet (0 to 12,192 meters) Nonoperating Temperature, Humidity, and Altitude for R5000 UPS and R5000 ERM Temperature: 13 to 131 F (25 to 55 C) Altitude: 0 to 49,212 feet (0 to 15,000 meters) Cooling Airflow Direction Each enclosure includes its own forcedair cooling fans or blowers. Air flow for each enclosure enters from the front of the modular cabinet and exhausts at the rear. Typical Acoustic Noise Emissions 70 db(a) (sound pressure level at operator position) Tested Electrostatic Immunity Contact discharge: 8 KV Air discharge: 20 KV Calculating Specifications for Enclosure Combinations Figure 14: Example Duplex Configuration shows components installed in 42U modular cabinets. Cabinet weight includes the PDUs and their associated wiring and receptacles. Power and thermal calculations assume that each enclosure in the cabinet is fully populated; for example, a NonStop Blade Element with four processors. The power and heat load is less when enclosures are not fully populated, such as a NonStop Blade Element with fewer processors or less memory, a Fibre Channel disk module with fewer disk drives, or an LSU enclosure with fewer LSUs. AC current calculations assume that one PDU delivers all power. In normal operation, the power is split equally between the two PDUs in the cabinet. However, calculate the power load to assume delivery from only one PDU to allow the system to continue to operate if one of the two AC power sources or PDUs fails. NOTE: If your system includes the optional rackmounted HP R5000 UPS or HP R5500 XR UPS, the modular cabinet will have one PDU located on the rear left side and four extension bars on the rear right side. To provide redundancy, components are plugged into the leftside PDU and the extension bars. Each extension bar is plugged into the UPS. Figure 14: Example Duplex Configuration has eight logical processors with two VIO enclosures, and eight Fibre Channel disk modules installed in two 42U modular cabinets. The weight, power, and thermal calculations for the components in each cabinet are shown in Example Load Calculations for Cabinet One (page 66). For a total thermal load for a system with multiple cabinets, add the heat outputs for all the cabinets in the system. Calculating Specifications for Enclosure Combinations 65

66 Figure 14 Example Duplex Configuration Table 1 Example Load Calculations for Cabinet One Component Quantity Height (U) Weight (lbs) Weight (kg) Voltamps, one feed powered Voltamps, both feeds powered Heat (Btu) NonStop Blade Element LSU VIO enclosure Fibre Channel disk module System Installation Specifications

67 Table 1 Example Load Calculations for Cabinet One (continued) Component Quantity Height (U) Weight (lbs) Weight (kg) Voltamps, one feed powered Voltamps, both feeds powered Heat (Btu) Console Maintenance switch Cabinet Total Table 2 Example Load Calculations for Cabinet Two Component Quantity Height (U) Weight (lbs) Weight (kg) Voltamps, one feed powered Voltamps, both feeds powered Heat (Btu) NonStop Blade Element LSU VIO enclosure Fibre Channel disk module Console Maintenance switch Cabinet Total Calculating Specifications for Enclosure Combinations 67

68 4 System Configuration Guidelines Integrity NonStop NS14000 series systems use a flexible modular architecture. Therefore, almost any configuration of the system s modular components is possible within a few configuration restrictions stated in Enclosure Locations in Cabinets (page 84) and VIO Enclosure and Disk Storage Considerations (page 72). NOTE: For NonStop NS14000 series servers that contain I/O Adapter Module (IOAM) enclosures, refer to the H06.07 (or earlier) version of this manual for planning, configuration, and maintenance information of the IOAM enclosure. For other example configurations, see Chapter 5 (page 84). Internal ServerNet Interconnect Cabling This subsection includes: Dedicated Service LAN Cables (page 68) Length Restrictions for Optional Cables (page 68) Internal Cable Product IDs (page 68) NonStop Blade Elements to LSU (page 69) LSU to VIO Enclosure and Processor IDs (page 69) VIO Enclosure ServerNet Connections (page 71) Fibre Channel Port to Fibre Channel Disk Modules (page 72) Fibre Channel Port to Tape Devices (page 72) Dedicated Service LAN Cables The NonStop NS14000 series system uses Category 5e unshielded twistedpair (CAT 5e UTP) or Category 6 unshielded twistedpair (CAT 6 UTP) Ethernet cables for the internal dedicated service LAN and for connections between the VIO enclosure and the application LAN equipment. Length Restrictions for Optional Cables Maximum allowable lengths of cables connecting to components outside the modular cabinet are: Connection VIO enclosure (Fibre Channel port) to ESS VIO enclosure (Fibre Channel port) to FC switch Cable Type MMF MMF Connectors LCLC LCLC Maximum Length 250 m 250 m Product ID M8900nnn 1 M8900nnn 1 1 nnn indicates the length of the cable in meters. For example, M is 125 meters long; M is 15 meters long. For a complete list of cables, see Cable Types, Connectors, Length Restrictions, and Product IDs (page 94). Although a considerable cable length can exist between the modular enclosures in the system, HP recommends placing all cabinets adjacent to each other and bolting them together, with cable length between each of the enclosures as short as possible. Internal Cable Product IDs For product IDs, see Cable Types, Connectors, Length Restrictions, and Product IDs (page 94). 68 System Configuration Guidelines

69 NonStop Blade Elements to LSU Fiberoptic cables provide communications between each NonStop Blade Element and the LSU as well as between the LSU and the VIO enclosure. The ServerNet X fabric and Y fabric provide the system I/O from the LSU to the VIO enclosure, with Fibre Channel and highspeed Ethernet links providing connections to storage and communications LANs, WANs, and so forth. Cable connections between the NonStop Blade Elements and the LSU optics adapters affect proper processor synchronization and rendezvous. For each logical processor element (PE), the port for a particular PE, such as PE1 (NonStop Blade Element port J0) on DMR NonStop Blade Elements A and B or TMR NonStop Blade Elements A, B, and C, must connect to their respective ports on the LSU optics adapters as shown in the example connection diagrams in LSU to VIO Enclosure and Processor IDs (page 69). Although you can randomly select LSU optics adapters for fiberoptic cable connections, HP recommends connecting these cable to the LSUs in sequential order as shown in the connection diagrams. NonStop Blade Element to NonStop Blade Element Reintegration cables interconnect each of the NonStop Blade Elements within individual duplex or triplex NonStop Blade Complexes using connectors S, T, Q, and R as shown in the illustrations on the next four pages. LSU to VIO Enclosure and Processor IDs Each NonStop Blade Element contains four processor elements, and each element is a part of a numbered NonStop Blade Complex, such as 0, 1, 2, or 3. The maintenance entity (ME) firmware running in the VIO enclosure assigns a number to each processor element based on its connection from the LSU to the ServerNettoprocessor ports in slots 14.1 through 14.4 (processors 0 through 3) and slots 2.1 through 2.4 (processors 4 through 7). Therefore, fiberoptic cable connections from the LSU to the ServerNettoprocessor ports on the VIO enclosure determine the processor number of each NonStop Blade Complex. This table lists the default VIO enclosure slot and port coupling to the processor number: VIO Enclosure Module Slot.Port Fabric Processor Number X X X X X X X X Y Y Y Y Y Y 5 Internal ServerNet Interconnect Cabling 69

70 VIO Enclosure Module Slot.Port Fabric Processor Number 2.3 Y Y 7 The two cabling diagrams on the next pages illustrate the default configuration and connections for a triplex system processor. These diagrams are not for use in installing or cabling the system. For instructions on connecting the cables, see the NonStop NS14000 Hardware Installation Manual. This figure shows example connections to the default configuration of the NonStop Blade Element reintegration links (NonStop Blade Element connectors S, T, Q, R) and ServerNettoprocessor ports 14.1 to 14.4 on the VIO enclosure, which defines triplex processor numbers 0 to 3. Two VIO enclosures are required, one for the Xfabric and the other for the Yfabric: This figure shows example connections of the NonStop Blade Element reintegration links (NonStop Blade Element connectors S, T, Q, R) and ServerNettoprocessor ports 2.1 to 2.4 on the VIO enclosure, which defines triplex processor numbers 4 to 7. Two VIO enclosures are required, one for the Xfabric and the other for the Yfabric: 70 System Configuration Guidelines

71 VIO Enclosure ServerNet Connections ServerNet connections to the system I/O devices (storage disk and tape drive as well as Ethernet communication to networks) radiate out from the VIO enclosures for both the X and Y ServerNet fabrics. (VIO enclosure module 2 provides X fabric connectivity, and VIO enclosure module 3 provides Y fabric connectivity.) I/O configurations in NonStop NS14000 series systems are flexible, with few restrictions. Those few restrictions prevent I/O configurations that compromise fault tolerance or high availability, especially with disk storage as outlined in Configuration Restrictions for Fibre Channel Devices (page 75). Internal ServerNet Interconnect Cabling 71

72 Fibre Channel Port to Fibre Channel Disk Modules Fibre Channel disk modules (FCDMs) can be connected directly to the Fibre Channel ports on a VIO enclosure, with these exceptions: Only configurations with two VIO enclosures are supported. A maximum of eight FCDMs (or up to eight daisychained configurations with each daisychain configuration containing 4 FCDMs) can be connected in the NonStop NS14000 series system. See Example Configurations of the VIO Enclosures and Fibre Channel Disk Modules (page 76). Fibre Channel Port to Tape Devices Fibre Channel tape devices can be connected directly to the Fibre Channel port on a VIO enclosure. With a Fibre Channel tape drive connected to a server, you can use the BACKUP and RESTORE utilities to save data to and restore data from tape. VIO Enclosure and Disk Storage Considerations These considerations apply to a NonStop NS14000 series system: Only configurations with two VIO enclosures are supported. The group number for the VIO enclosures is 100. NonStop Sseries I/O enclosures are not supported. A maximum of eight FCDMs can be connected in the NonStop NS14000 series system because only two VIO enclosures are supported. Two Ethernet connections, one in each VIO enclosure module, that connect to the maintenance entity are necessary to enable the OSM Service Connection and OSM Notification Director. Fibre Channel Devices This subsection includes: FactoryDefault Disk Volume Locations (page 74) Configurations for Fibre Channel Devices (page 74) Configuration Restrictions for Fibre Channel Devices (page 75) Recommendations for Fibre Channel Device Configurations (page 75) Example Configurations of the VIO Enclosures and Fibre Channel Disk Modules (page 76) The only Fibre Channel device used internally with NonStop NS14000 series systems is the Fibre Channel disk module (FCDM). An FCDM and its disk drives are controlled through the Fibre Channel ports located on the VIO enclosures. For more information on the Fibre Channel ports, see Versatile I/O (VIO) Enclosure (page 22). For more information on the Fibre Channel disk module (FCDM), see Fibre Channel Disk Module (FCDM) (page 24). For examples of cable connections between Fibre Channel ports on the VIO enclosures and FCDMs, see Example Configurations of the VIO Enclosures and Fibre Channel Disk Modules (page 76). This illustration shows the location of Fibre Channel ports on a VIO enclosure: 72 System Configuration Guidelines

73 For more information on Fibre Channel ports, see Versatile I/O (VIO) Enclosure (page 22). This illustration shows the locations of the hardware in the Fibre Channel disk module as well as the Fibre Channel port connectors at the back of the enclosure: Fibre Channel disk modules connect to Fibre Channel ports on the VIO enclosure via Fiber Channel arbitrated loop (FCAL) cables. This drawing shows the two Fibre Channel arbitrated loops implemented within the Fibre Channel disk module: Fibre Channel Devices 73

74 For more information on the Fibre Channel disk module, see Fibre Channel Disk Module (FCDM) (page 24). FactoryDefault Disk Volume Locations This illustration shows where the factorydefault locations for the primary and mirror system disk volumes reside in separate Fibre Channel disk modules: This table lists the Fibre Channel groupmoduleslotport (GMSP) and disk groupmoduleshelfbay (GMSB) identification for the factorydefault system disk locations: Disk Volume Name $SYSTEM (primary) $DSMSCM (primary) $AUDIT (primary) $OSS (primary) $SYSTEM (mirror) $DSMSCM (mirror) $AUDIT (mirror) $OSS (mirror) Fibre Channel Port GMSP and and and and and and and and Disk GMSB* Configurations for Fibre Channel Devices The NonStop NS14000 series system supports up to eight Fibre Channel ports and 32 Fibre Channel disk modules are possible when using daisychained configurations, with identification depending on the ServerNet connection to the Fibre Channel ports on the VIO enclosure. 74 System Configuration Guidelines

75 Configuration Restrictions for Fibre Channel Devices To avoid creating configurations that are not faulttolerant or do not promote high availability, these restrictions apply and are invoked by Subsystem Control Facility (SCF): Primary and mirror disk drives in FCDMs cannot connect to the same Fibre Channel loop. Loss of the Fibre Channel loop makes both the primary volume and the mirrored volume inaccessible. This configuration inhibits fault tolerance. Disk drives in different FCDMs within a daisychain connect to the same Fibre Channel loop. The primary path and backup Fibre Channel communication links to a disk drive should not connect to Fibre Channel ports in the same VIO enclosure. Loss of one Fibre Channel port can make multiple disk drives inaccessible on a single Fibre Channel communications path. This configuration is allowed, but only if you override an SCF warning message. The mirror path and mirror backup Fibre Channel communication links to a disk drive should not connect to Fibre Channel ports in the same VIO enclosure. Loss of one Fibre Channel port can make multiple disk drives inaccessible on a single Fibre Channel communications path. This configuration is allowed, but only if you override an SCF warning message. Recommendations for Fibre Channel Device Configurations Recommendations for Fibre Channel Device Configuration Different device types should be on different Fibre channel loops. For example, do not put disk and tape devices on the same Fibre Channel loop. These recommendations apply to VIO Fibre Channel ports and FCDM configurations: Primary FCDM connects to the VIO Fibre Channel port 1. Mirror FCDM connects to the VIO Fibre Channel port 2. FCAL port A1 is the incoming port from a Fibre Channel port on the VIO enclosure or from another FCDM. FCAL port A2 is the outbound port to another FCDM. FCAL port B2 is the incoming port from a Fibre Channel port on the VIO enclosure or from a FCDM. FCAL port B1 is the outbound port to another FCDM. In a daisychain configuration, the ID expander harness determines the enclosure number. Enclosure 1 is always at the bottom of the chain. Fibre Channel ports are located in slot 1of a VIO enclosure. Ethernet ports are located in slots 6b, 6a/7a, and 7b of a VIO enclosure. The group number for a VIO enclosure is 100. With primary and mirror FCDMs in the same cabinet, the primary FCDM resides in a lower U than the mirror FCDM. Fibre Channel disk drives are configured with dual paths. Where possible, Fibre Channel ports and FCDMs are configured with four Fibre Channel ports and four FCDMs for maximum fault tolerance. If Fibre Channel ports are not in groups of four, the remaining Fibre Channel ports and FCDMs can be configured in other faulttolerant configurations, such as with two Fibre Channel ports and two FCDMs or four Fibre Channel ports and three FCDMs. Fibre Channel Devices 75

76 The NonStop NS14000 series system supports only configurations with two VIO enclosures: With four Fibre Channel ports and two FCDMs, the primary Fibre Channel port resides in VIO enclosure module 2, and the backup Fibre Channel port resides in VIO enclosure module 3. (See the example configuration in Figure 15 (page 77).) With eight Fibre Channel ports and four FCDMs, the primary Fibre Channel ports reside in VIO enclosure module 2, and the backup Fibre Channel ports resides in VIO enclosure module 3. (See the example configuration in Figure 16 (page 78).) Daisychain configurations follow the same configuration restrictions and rules that apply to configurations that are not daisychained. (See DaisyChain Configurations (page 80).) FCDMs containing mirrored volumes must be installed in separate daisy chains. Daisychained configurations require that all FCDMs reside in the same cabinet and be physically grouped together. Daisychain configurations require an ID expander harness with terminators for proper FCDM and disk drive identification. Example Configurations of the VIO Enclosures and Fibre Channel Disk Modules This section provides various example configurations between Fibre Channel ports on the VIO enclosures and Fibre Channel disk modules. NOTE: Although it is not a requirement for fault tolerance to house the primary and mirror disk drives in separate FCDMs, the example configurations show FCDMs housing only primary or mirror drives, mainly for simplicity in keeping track of the physical locations of the drives. Four Fibre Channel Ports, Two FCDMs, Two VIO Enclosures (page 77) shows example cable connections between four Fibre Channel ports (two on each VIO enclosure) and the primary and mirror Fibre Channel disk modules. Eight Fibre Channel Ports, Four FCDMs, Two VIO Enclosures (page 78)shows example cable connections between eight Fibre Channel ports (four on each VIO enclosure) and the two sets of primary and mirror Fibre Channel disk modules. Twelve Fibre Channel Ports, Six FCDMs, Two VIO Enclosures (page 79) shows example cable connections between twelve Fibre Channel ports (six on each VIO enclosure) and the three sets of primary and mirror Fibre Channel disk modules. Sixteen Fibre Channel Ports, Eight FCDMs, Two VIO Enclosures (page 80) shows example cable connections between sixteen Fibre Channel ports (eight on each VIO enclosure) and the four sets of primary and mirror Fibre Channel disk modules. 76 System Configuration Guidelines

77 Figure 15 Four Fibre Channel Ports, Two FCDMs, Two VIO Enclosures Fibre Channel Devices 77

78 Figure 16 Eight Fibre Channel Ports, Four FCDMs, Two VIO Enclosures 78 System Configuration Guidelines

79 Figure 17 Twelve Fibre Channel Ports, Six FCDMs, Two VIO Enclosures Fibre Channel Devices 79

80 Figure 18 Sixteen Fibre Channel Ports, Eight FCDMs, Two VIO Enclosures DaisyChain Configurations When planning for possible use of daisychained disks, consider: DaisyChained Disks Recommended Costsensitive storage and applications using lowbandwidth disk I/O. Lowcost, highcapacity data storage is important. DaisyChained Disks Not Recommended Many volumes in a large Fibre Channel loop. The more volumes that exist in a larger loop, the higher the potential for negative impact from a failure that takes down a Fibre Channel loop. Applications with a highly mixed workload, such as transaction data bases or applications with high disk I/O. Requirements for DaisyChain 1 All daisychained Fibre Channel disk modules reside in the same cabinet and are physically grouped together. ID expander harness with terminators is installed for proper Fibre Channel disk module and drive identification. Fibre Channel ports for each Fibre Channel loop are installed in a different VIO enclosure for fault tolerance. Two Fibre Channel disk modules minimum, with four Fibre Channel disk modules maximum per daisy chain. 80 System Configuration Guidelines

81 1 See Fibre Channel Devices (page 72). This illustration shows an example of cable connections between two Fibre Channel ports (one on each VIO enclosure) and four Fibre Channel disk modules in a single daisychain configuration: Two VIO enclosures, four Fibre Channel ports, and four Fibre Channel disk modules with an ID expander do not provide faulttolerant mirrored disk storage. Installing each mirrored disk in the same corresponding FCDM and bay number as its primary disk is not required, but it is recommend to simplify the physical management and identification of the disks. This table lists the Fibre Channel port groupmoduleslotport (GMSP) and disk groupmoduleshelfbay (GMSB) identification for the factorydefault system disk locations in a daisychained configuration: Disk Volume Name $SYSTEM $DSMSCM $AUDIT $OSS Fibre Channel Port GMSP and and and and Disk GMSB* * For an illustration of the factorydefault slot locations for a Fibre Channel disk module, see FactoryDefault Disk Volume Locations (page 74). Fibre Channel Devices 81

82 Ethernet to Networks The Ethernet ports provide Gigabit connectivity between NonStop NS14000 series systems and Ethernet LANs. The Ethernet port is an end node on the ServerNet and uses either fiberoptic or copper cable for connectivity to user application LANs, as well as for the dedicated service LAN. For more information on the Ethernet ports, see Versatile I/O (VIO) Enclosure (page 22). This illustration show a conceptual example for copper and fiberoptic connectivity to the various LANs: Default Naming Conventions With a few exceptions, default naming conventions are not necessary for the modular resources that make up Integrity NonStop NS14000 series systems. In most cases, users can name their resources at will and use the appropriate management applications and tools to find the location of the resource. However, default naming conventions for certain resources simplify creation of the initial configuration files and automatic generation of the names of the modular resources. When autoconfigure mode is set to ON for storage subsystems, the preconfigured default naming convention is used to generate the names of the Fibre Channel disk drives. Preconfigured default resource names are: Type of Object Naming Convention Example Description Fiber Channel disk drive $FC number $FC10 Tenth Fibre Channel disk drive in the system ESS disk drive $ESS number $ESS20 Twentieth ESS disk drive in the system Modular tape drive $TAPE number $TAPE01 First modular tape drive in the system 82 System Configuration Guidelines

83 Type of Object Naming Convention Example Description Ethernet connection G group module slot G10027 Ethernet connectivity in location Ethernet LIF L group module slot port L10027B LIF for PIF at location B TCP/IP process $ZTC number $ZTC0 First TCP6SAM or TCP/IP process for the system Telserv process $ZTN number $ZTN0 First Telserv process for the system Listener process $LSN number LSN0 First Listener process for the system TFTP process Automatically created by WANMGR None None WANBOOT process Automatically created by WANMGR None None SWAN Concentrator S number S10 Tenth SWAN concentrator in the system On new NonStop systems, only one of each of these processes and names is configured: TCP6SAM $ZTC0 Telserv $ZTCN0 Listener $LNS0 No TFTP or WANBOOT process is configured for new NonStop systems. NOTE: Naming conventions or configurations for the dedicated service LAN TCP/IP are the same as the TCP/IP conventions used with Gseries RVUs and are named $ZTCP0 and $ZTCP1. CAUTION: Do not change the process names for $ZTCP0 and $ZTCP1. Doing so will make some components inaccessible. OSM Service Connection provides the location of the resource by adding an identifying suffix to the names of all the system resources. Other interfaces, such as SCF, also provide means to locate named resources. Default Naming Conventions 83

84 5 Hardware Configurations in Modular Cabinets This chapter shows examples of hardware component configurations for the Integrity NonStop NS14000 series server. A number of other configurations are also possible because of the flexibility inherent to the NonStop advanced architecture and ServerNet. NOTE: Hardware configuration drawings in this appendix represent the physical arrangement of the modular enclosures but do not show location of the PDU junction boxes. Location of the PDU junction boxes for AC power feed can be from above or below the cabinet. Minimum, Typical, and Maximum Hardware Configuration This table shows the minimum, typical, and maximum number of the modular components installed in a system. These typical values might not reflect the system you are planning and are provided only as an example, not as exact values. Enclosure or Component Duplex Processor Minimum Typical Maximum Triplex Processor Minimum Typical Maximum 4processor NonStop Blade Element with 16 DIMMs GB memory quad Processor board with two 1.6 GHz processors 2 3 Processor board with four 1.6 GHz processors LSU logic board and optics adapter VIO enclosure Fibre Channel disk module (32 daisychained) (32 daisychained) Enclosure Locations in Cabinets For details about the location of NonStop NS14000 enclosures and components within a cabinet, including the U location and cabling for enclosures and components, refer to the Technical Document for your system. 84 Hardware Configurations in Modular Cabinets

85 6 Maintenance and Support Connectivity Local monitoring and maintenance of the NonStop NS14000 series system occurs over the dedicated service LAN. The dedicated service LAN provides connectivity between the system console and the maintenance infrastructure in the system hardware. HP Insight Remote Support Advanced is the goforward remote support solution for NonStop servers. It works in conjunction with OSM server software, but replaces the OSM Notification Director in both modembased and HP Instant Support Enterprise Edition (ISEE) remote support solutions. For more information on Insight Remote Support Advanced, see Insight Remote Support Advanced for NonStop, located in the Service Information collection of NTL. Only components specified by HP can be connected to the dedicated LAN. No other access to the LAN is permitted. A maximum of eight NonStop systems can be connected to the dedicated service LAN. The dedicated service LAN uses a ProCurve Ethernet switch for connectivity between the VIO enclosures and the system console. An important part of the system maintenance architecture, the system console is a personal computer (PC) approved by HP to run maintenance and diagnostic software for NonStop NS14000 series systems. Through the system console, you can: Monitor operations on systems using the HP NonStop Open System Management (OSM) interface View manuals and service procedures Run HP Tandem Advanced Command Language (TACL) sessions using terminalemulation software Install and manage system software using the Distributed Systems Management/Software Configuration Manager (DSM/SCM) Make remote requests to and receive responses from a system using remote operation software Dedicated Service LAN A NonStop NS14000 series system requires a dedicated LAN for system maintenance through OSM. Only components specified by HP can be connected to a dedicated LAN. No other access to the LAN is permitted. This subsection includes: Basic LAN Configuration (page 86) FaultTolerant Configuration (page 86) IP Addresses (page 87) Ethernet Cables (page 89) SWAN Concentrator Restrictions (page 89) SystemUp Dedicated Service LAN (page 90) Dedicated Service LAN Links With Two VIO Enclosures (page 90) Initial Configuration for a Dedicated Service LAN (page 91) Operating Configurations for Dedicated Service LANs (page 92) Dedicated Service LAN 85

86 Basic LAN Configuration A basic dedicated service LAN that does not provide a faulttolerant configuration requires connection of these components to the ProCurve maintenance switch installed in the modular cabinet: One connection from the ME ENET port on each of the two VIO enclosures One connection from slot 6B, port A on each of the two VIO enclosure for the OSM Service Connection and OSM Notification Director One connection to each of the optional UPS modules One connection for each system console running OSM This illustration shows a basic LAN configuration with one maintenance switch: Figure 19 Basic LAN Configuration FaultTolerant Configuration Connect one system console to each maintenance switch. Connect the VIO enclosure for the X fabric to one of the maintenance switches and the VIO enclosure for the Y fabric to the other maintenance switch. 86 Maintenance and Support Connectivity

87 Connect one Ethernet port in each VIO enclosure to each maintenance switch. Connect one maintenance switch to an HP extension bar that is powered by the UPS. HP extension bars are installed in place of a PDU on the rear right side of the modular cabinet. CAUTION: To avoid possible conflicts on the LAN: If the configuration includes two maintenance switches, install and configure one switch completely, including assigning its IP address, before you install the other. Only one connection between the maintenance switches is permitted. More than one connection overloads network traffic, rendering the dedicated service LAN unusable. If VIO enclosures will have static IP addresses, configure one completely, including assigning the IP address, before you configure the other. This illustration shows a faulttolerant LAN configuration with two maintenance switches: Figure 20 FaultTolerant LAN Configuration IP Addresses NonStop NS14000 series servers require Internet protocol (IP) addresses for these components that are connected to the dedicated service LAN: VIO enclosure logic boards Maintenance switches Dedicated Service LAN 87

88 System consoles UPSs (optional) These components have default IP addresses that are preconfigured at the factory. You must change these preconfigured IP addresses to addresses appropriate for the LAN environment: Component Group.Module.Slot Default IP Address Used By VIO enclosure OSM LowLevel Link Maintenance switch (ProCurve) (rackmounted) N/A N/A OSM Service Connection Additional switches (rackmounted only) N/A , , n Primary system console (rackmounted or standalone) Backup system console (rackmounted only) N/A N/A OSM LowLevel LinkOSM Service ConnectionOSM Notification Director Up to two additional system consoles (rackmounted only) N/A UPS (rackmounted only) N/A OSM Service Connection UPS (nonrackmounted) N/A TCP/IP processes for OSM: $ZTCP b port A GW: OSM Service Connection $ZTCP b port A GW: OSM Notification Director Whether or not the new system will receive dynamic IP addresses from a Dynamic Host Configuration Protocol (DHCP) server, it is recommended that the IP addresses be reconfigured as either: Static IP Addresses (page 88) Dynamically Assigned IP Addresses (page 89) NOTE: Be aware of possible conflicts with existing operations LANs. This guide cannot predict all possible configurations of existing LANs. If the site where you install the maintenance switch has network administrators and specialists, consult them for information about their current Dynamic Host Configuration Protocol (DHCP) configuration. Static IP Addresses Static IP addresses must be configured manually for each component using values defined for the site. If a DHCP server already exists on the dedicated service LAN, you must configure the static IP addresses to be in the same subnet as the dynamic IP addresses assigned by that server. 88 Maintenance and Support Connectivity

89 Dynamically Assigned IP Addresses Dynamically assigned IP addresses are provided by a DHCP server that assigns an IP address whenever a component is added or moved. If a DHCP server is already present on the network to which these components will be connected, you can use that server. If your network is not equipped with a DHCP server, work with your account team to identify available options. Some guidelines for configuring the DHCP server: Configure the range of IP addresses to be assigned dynamically by the DHCP server to be in the same subnet as existing IP addresses on the LAN and any static IP address included in the dedicated service LAN. IP addresses for these components should remain static: The four IP addresses reserved for the OSM LowLevel Link and Service Connection Do not include these address in the pool of addresses the DHCP uses for address assignment. The UPS (recommended but not required) Static or Dynamic IP Addresses Various components within the dedicated service LAN can have static or dynamic IP addresses. This sample shows the address in an example dedicated service LAN, but any combination is possible: Component Maintenance switches* UPSs (if the system has them)* TCP/IP process $ZTCP0** TCP/IP process $ZTCP1** VIO enclosures System consoles IP Address Static Static Static Static Dynamic Dynamic *If a combination of dynamic and static IP addresses are used in the LAN configuration, HP recommends that the maintenance switches and UPSs (if the system has them) be configured with static IP addresses. **You do not need to reconfigure the IP addresses for the TCP/IP processes. Ethernet Cables Ethernet connections for a dedicated service LAN require CAT 5e or CAT 6 unshielded twistedpair (UTP) cables. SWAN Concentrator Restrictions Isolate any ServerNet wide area networks (SWANs) on the system. The system must be equipped with at least two LANs: one LAN for SWAN concentrators and one for the dedicated service LAN. Most SWAN concentrators are configured redundantly using two or more subnets. Those subnets also must be isolated from the dedicated service LAN. Do not connect SWANs on a subnet containing a DHCP. Dedicated Service LAN 89

90 SystemUp Dedicated Service LAN When the system is up and the OS running, the ME connects to the system s dedicated service LAN using connections from the maintenance switch to ME ENET ports on each VIO enclosure. To use the OSM Service Connection and OSM Notification Director, you must have connections from the maintenance switch to Ethernet ports on each VIO enclosure. For a faulttolerant dedicated service LAN, two Ethernet ports are required; one port located in VIO enclosure module 2 for X ServerNet fabric connectivity, and one port located in VIO enclosure module 3 for Y ServerNet fabric connectivity. When the Ethernet port provides connection to the dedicated service LAN, use the slower 10/100 Mbps PIF A rather than one of the highspeed 1000 Mbps Ethernet ports of PIF C or D. Dedicated Service LAN Links With Two VIO Enclosures This illustration shows the dedicated service LAN cables connected to the Ethernet ports in slot 6B of both VIO enclosures (module 2 and module 3) and to the maintenance switch: 90 Maintenance and Support Connectivity

91 The values in this table show the identification for the Ethernet ports in the preceding illustration: GMS for Ethernet Port Location in VIO Enclosure Ethernet PIF Ethernet LIF TCP/IP Stack IP Configuration b Port A G A L1002R $ZTCP0 IP: GW: Subnet: %hffffff00 Hostname: osmlanx b Port A G A L1003R $ZTCP1 IP: GW: Subnet: %hffffff00 Hostname: osmlany Initial Configuration for a Dedicated Service LAN New systems are shipped with an initial set of IP addresses configured. For a listing of these initial IP addresses, see IP Addresses (page 87). Factorydefault IP addresses for the Ethernet ports are available from your service provider (in the Versatile I/O Manual.) IP addresses for SWAN concentrators are in the WAN Subsystem Configuration and Management Manual. HP recommends that you change these preconfigured IP addresses to addresses appropriate for your LAN environment. You must change the preconfigured IP addresses on: A backup system console if you want to connect it to a dedicated service LAN that already includes a primary system console or other system console Any system console if you want to connect it to a dedicated service LAN that already includes a primary system console Keep track of all the IP addresses in your system so that no IP address is assigned twice. Dedicated Service LAN 91

92 Operating Configurations for Dedicated Service LANs You can configure the dedicated service LAN in several different ways, as described in the OSM Migration and Configuration Guide. HP recommends that you use a faulttolerant LAN configuration. For a faster configuration option for the OSM Service Connection only, see the OSM Migration and Configuration Guide. System Consoles New system consoles are preconfigured with the required HP and thirdparty software. When upgrading to the latest RVU, you can install software upgrades from the HP NonStop System Console Installer CD. NOTE: The NonStop system console must be configured with some ports open. For more information, see the NonStop System Console Installer Guide. Some system console hardware, including the PC system unit, monitor, and keyboard, can be mounted in the cabinet. Other PCs are installed outside the cabinet and require separate provisions or furniture to hold the PC hardware. System consoles communicate with NonStop NS14000 series servers over a dedicated service local area network (LAN) or a secure operations LAN. A dedicated service LAN is required for use of OSM LowLevel Link and Notification Director functionality, which includes configuring primary and backup dialout points (referred to as the primary and backup system consoles, respectively). HP recommends that you also configure the backup dedicated service LAN with a backup system console. For Information About... Refer to... Connecting and configuring system consoles NonStop NS14000 Series Hardware Installation Manual System Console Configurations Several system console configurations are possible: NonStop Dedicated Service LAN Installation and Configuration Guide One System Console Managing One System (Setup Configuration) (page 92) One System Console Managing Multiple Systems (page 93) Primary and Backup System Consoles Managing One System (page 93) Primary and Backup System Consoles Managing Multiple Systems (page 93) One System Console Managing One System (Setup Configuration) The one system console on the LAN must be configured as the primary system console. This configuration can be called the setup configuration and is used during initial setup and installation of the system console and the server. When you use this configuration, you do not need to change the preconfigured IP addresses. The setup configuration is an example of a secure, standalone network. A LAN cable connects the primary system console to the maintenance switch, and additional LAN cables connect the VIO enclosures. The maintenance switch or an optional second maintenance switch allows you to later add a backup system console and additional system consoles. See Basic LAN Configuration (page 86) for an illustration of this configuration. 92 Maintenance and Support Connectivity

93 NOTE: Because the system console and maintenance switch are single points of failure that could disrupt access to OSM, this configuration is not recommended for operations that require high availability or fault tolerance. One System Console Managing Multiple Systems The one system console on the LAN must be configured as the primary system console. Because all servers are shipped with the same preconfigured IP addresses, you must change the IP addresses for the second and subsequent servers before you can add them to the LAN. To add a system to the LAN, see the NonStop Dedicated Service LAN Installation and Configuration Guide. Primary and Backup System Consoles Managing One System Two maintenance switches provide fault tolerance and extra ports for adding system consoles. You must change the preconfigured IP addresses of the second and subsequent system consoles before you can add them to the LAN. See FaultTolerant LAN Configuration (page 87) for an illustration of this configuration. Primary and Backup System Consoles Managing Multiple Systems If you want to manage more than one node from a faulttolerant pair of consoles, you can daisy chain the maintenance switches together. This configuration requires an IP address scheme to support it. Contact your HP service provider to design this configuration. System Consoles 93

94 A Cables for NS14000 Series Systems Cable Types, Connectors, Length Restrictions, and Product IDs TIP: Although a considerable cable length can exist between the modular enclosures in the system, HP recommends that cable length between each of the enclosures be as short as possible. Available cables and their lengths are: Connection From... Cable Type Connectors Length Product ID NonStop Blade MMF LCLC 2 meters M Element to LSU enclosure 3 m M m M m 10 m 15 m 40 m 80 m 100 m M M M M M M NonStop Blade MMF MTPMTP 3.15 meters M Element to NonStop Blade Element 5 m M m M m M m M LSU Enclosure to VIO enclosure MMF LCLC 2 meters 3 m M M m M m M m M m M m M m M m M m M VIO enclosure to VIO MMF MTPMTP 3.15 meters M enclosure (crosslink connection) 5 m M m M m 50 m M M Cables for NS14000 Series Systems

95 Connection From... Cable Type Connectors Length Product ID VIO enclosure (Fibre Channel port) to: MMF LCLC 2 meters M Fibre Channel disk 3 m M module (FCDM) 5 m ESS M FC switch 6 m M m 15 m 40 m 80 m 100 m 125 m 200 m 250 m M M M M M M M M VIO enclosure (Fibre MMF LCLC 2 meters M Channel port) to 6780 cluster switch 3 m M m M m M m M m M m M m M m M Maintenance LAN interconnect CAT 5e UTP RJ45 to RJ meters 3 m M M m M m M Maintenance LAN interconnect CAT 6 UTP RJ45 to RJ meters 1.2 m M M m M m M m M m M m M M m M NOTE: ServerNet cluster connections on NonStop NS14000 series systems follow ServerNet cluster and cable length rules and restrictions. For more information, refer to the ServerNet Cluster Supplement for NonStop NSSeries Systems. Cable Types, Connectors, Length Restrictions, and Product IDs 95

96 Cable Management System NonStop NS14000 series systems include the cable management system (CMS) to protect all power, fiberoptic, and CAT 5e or CAT 6 Ethernet cables within the systems. The CMS maintains a 25 millimeter (1 inch) minimum bend radius for the fiberoptic cables and provides strain relief for all cables. Several NonStop NS14000 series enclosures, specifically the NonStop Blade Element and LSU, integrate CMS provisions for routing and securing the fiberoptic cables to prevent damaging them when the enclosures are moved out and back into the cabinet for servicing. Additionally, cable spools mounted on the inside of the cabinet structure provide the means for looping and containing free lengths of fiberoptic cables to prevent damage. For details on using the CMS, refer to the NonStop NS14000 Hardware Installation Manual. 96 Cables for NS14000 Series Systems

97 B Operations and Management Using OSM Applications OSM clientbased components are installed on new system console shipments and also delivered by an OSM installer on the HP NonStop System Console (NSC) Installer CD. The NSC CD also delivers all other client software required for managing and servicing NonStop NS14000 series servers. For installation instructions, see the NonStop System Console Installer Guide. OSM serverbased components are incorporated in a single OSM serverbased SPR, T0682 (OSM Service Connection Suite), that is installed on Integrity NonStop NSseries servers running the HP NonStop operating system. For information on how to install, configure and start OSM serverbased processes and components, see the OSM Migration and Configuration Guide. The OSM components are: Product ID T0632 T0633 T0634 Component OSM Notification Director OSM LowLevel Link OSM Console Tools Description No longer needed when using Insight Remote Support Advanced for remote support services, including dialout (see note below). Provides downsystem support Includes: Start menu shortcuts and default home pages for easy access to the OSM Service Connection and OSM Event Viewer (browserbased OSM applications that are not installed on the system console) OSM System Inventory Tool Terminal Emulator File Converter OSM Certificate Tool CLIM Boot Service Configuration Wizard NOTE: HP Insight Remote Support Advanced is the goforward remote support solution for NonStop servers. It works in conjunction with OSM server software, but replaces the OSM Notification Director in both modembased and HP Instant Support Enterprise Edition (ISEE) remote support solutions. For more information on Insight Remote Support Advanced, see Insight Remote Support Advanced for NonStop, located in the Service Information collection of NTL. SystemDown OSM LowLevel Link In NonStop NS14000 series systems, the maintenance entity (ME) in the VIO enclosures provides dedicated service LAN services via the OSM LowLevel Link for both OS coldload, system management, and hardware configuration when hardware is powered up but the OS is not running. LAN connections are direct from the 10/100 Ethernet ports on the VIO enclosures to the maintenance switch that is installed in the modular cabinet. This illustration shows the LAN connections for the OSM LowLevel Link: SystemDown OSM LowLevel Link 97

98 AC Power Monitoring For Integrity NonStop NS14000 series servers, you can use one of the following to provide continued system operation through a power failure. However, to take advantage of OSM Power Fail Support (page 98), you must use the HP R5000 UPS or HP R5500 XR UPS. When used, it is connected to the system s dedicated service LAN via the maintenance switch through which OSM monitors the power state. Optional HP R5000 UPS or HP R5500 XR UPS (with one or two ERMs for additional battery power) Usersupplied UPS installed in each modular cabinet Usersupplied site UPS If a usersupplied UPS is used instead of the HPsupported UPS models mentioned above, there is no OSM notification or support when a power outage occurs, so the ridethrough time that can be configured through SCF does not apply either. Operators are responsible for detecting power transients and outages and developing the appropriate actions, based on the estimated capacity of the site UPS and the power demands made on that UPS. The R5000 UPS or R5500 XR UPS and ERMs installed in modular cabinets do not support any devices that are external to the cabinets. External devices can include tape drives, external disk drives, LAN routers, and SWAN concentrators. Any external peripheral devices that do not have UPS support will fail immediately at the onset of a power failure. Plan for UPS support of any external peripheral devices that must remain operational as system resources. This support can come from a site UPS or individual units as necessary. OSM Power Fail Support When properly configured, OSM provides important power fail support for Integrity NonStop NS14000 series servers. OSM detects power outages and helps you take appropriate actions in the event that the outage lasts longer than the estimated capabilities of your UPS. When OSM detects that one power rail is running on UPS and the other power rail has lost power, it logs an event and begins counting down the ridethrough time configured for the system. The ridethrough time, designed to avoid disruption if the power outage is short enough in duration for a UPS to provide the needed power, is specified through the POWERFAIL_DELAY_TIME attribute of the SCF command ALTER SUBSYS (for more information, see Considerations for RideThrough Time Configuration (page 99)). 98 Operations and Management Using OSM Applications

99 The actions that OSM takes next are directly tied to that specified ridethrough time: If AC power is restored before the ridethrough period ends, the ridethrough countdown terminates and OSM does not initiate a controlled shutdown of I/O operations and processors. If AC power is not restored before the ridethrough period ends, OSM initiates a controlled shutdown of I/O operations and processors by broadcasting a PFAIL_SHOUT message to all processors (the processor running OSM being the last one in the queue) to shut down the system's ServerNet routers and processors in a fashion designed to allow disk writes for items that are in transit through controllers and disks to complete. CAUTION: You should not turn off the UPS as soon as the NonStop OS is down, since the UPS continues to provide power until its supply is exhausted; time which may be needed for disk controllers and disks to complete disk writes. Unlike NonStop Sseries systems, this system requires a system load after a power failure that results in system shutdown. For more information on recovering from power failures, see the NonStop Operations Guide for HSeries and JSeries RVUs. Because this shutdown described above does not include powering off the system or stopping TMF or other applications, you are encouraged to create scripts to shut down database activity before the processors are shut down. With T0682 H02 ACC and later, OSM provides support for automatic execution of those scripts through two new OSMCONF settings: SHUTDOWN_SCRIPT_NAME and SHUTDOWN_SCRIPT_TIME. Through these settings, you tell OSM to execute a shutdown script a specified number of seconds before the ridethrough period ends. If AC power is restored and the countdown is terminated before this time is reached, OSM will not execute the script. For more information on implementing those OSMCONF settings, see the OSM Migration and Configuration Guide. To configure OSM power fail support, you must perform these actions, located under the Power Supply units in either Pswitches or IOAM modules in the OSM Service Connection: Perform the Configure a Power Source as UPS action to configure the power rail (either A or B) connected to the UPS. Perform the Configure a Power Source as AC action, to configure the power rail (either B or A) connected to AC power. You can then perform the Verify Power Fail Configuration action, located under the System object, to verify that power failure support has been properly configured and is in place for the system. For more information on locating or performing these actions, see the OSM Service Connection online help. Considerations for RideThrough Time Configuration Ridethrough time is specified through the POWERFAIL_DELAY_TIME attribute of the SCF command ALTER SUBSYS. For command syntax, see the SCF Reference Manual for the Kernel Subsystem. The goal in configuring the ridethrough time is to allow the maximum time for power to be restored while at the same time allowing sufficient time for completion of disk writes for IOs that passed to the disk controllers before the ServerNet was shut down. Allowing enough time for sufficient completion of these tasks allows for a relatively clean shutdown from which TMF recovery is less timeconsuming and difficult than if all power failed and disk writes did not complete. The maximum ridethrough time for each system will vary, depending on system load, configuration, and the UPS capability. A rackmounted UPS can often supply five minutes of power with some extra capacity for contingencies, provided the batteries are new and fully charged. This five minute figure is an estimate used for illustration in this discussion, not a guarantee for any specific configuration. You must ensure that the battery capacity for a fullypowered system allows for at least two minutes after OSM initiates the orderly shutdown to allow the disk cache to be flushed to nonvolatile media. Assuming the UPS has five minutes of power capacity, you would set the ridethrough time for three minutes (UPS capacity of five minutes minus two minutes for OSM). AC Power Monitoring 99

100 NOTE: OSM does not make dynamic computations based on remaining capacity of the rackmounted UPS. The ridethrough time is statically configured in SCF for OSM use. For example, when power comes back before the initiated shutdown, but then fails again shortly afterward, the UPS has been depleted by some amount and does not last for the ridethrough time until it is fully recharged. OSM does not account for multiple power failures that occur within the recharge time of the rackmounted UPS. Power can be extended by adding ERMs to the configuration; and power for UPS alone can extend beyond five minutes based on power consumption. To extend the ridethrough time beyond three minutes, use this manual and your UPS documentation to calculate the expected power consumption, measure the site power consumption, factor in the ERM, if present, and make adjustments. Also consider air conditioning failures during a real power failure because increased ambient temperature typically causes the fans to run faster, which causes the system to draw more power. By allowing for the maximum power consumption and applying those figures to the UPS calculations provided in the UPS manuals, you can increase the ridethrough time beyond three minutes. Considerations for Site UPS Configurations OSM cannot monitor a site UPS. The SCF configured ridethrough time on a NonStop NS14000 system has no effect if only a site UPS is used. With a site UPS instead of a rackmounted UPS, the customer must perform manual system shutdown if the backup generators cannot be started. It is also possible to have a rackmounted UPS in addition to a site UPS. Since the site UPS can supply a whole computer room or part of that room, including required cooling, from the perspective of OSM, site UPS power can supply the group 100 AC power. The group 100 UPS power configured in OSM, in this case, would still come from a rackmounted UPS (one of the supported models). AC PowerFail States These states occur when a power failure occurs and an optional HP R5000 UPS or HP R5500 XR UPS is installed in each cabinet within the system: System State NSK_RUNNING RIDE_THRU HALTED LOW_POWER POWER_OFF Description NonStop operating system is running normally. OSM has detect a power failure and begins timing the outage. AC power returning terminates RIDE_THRU and puts the operating system back into an NSK_RUNNING state. At the end of the predetermined RIDE_THRU time, if AC has not returned, OSM executes a PFAIL_SHOUT that results in the system going to LOW_POWER. Normal halt condition. Halted processors do not participate in powerfail handling. A normal poweron also puts the processors into the HALTED state. Haltedstate services (HSS) informs the blade element that it is in LOW_POWER state and then waits until optic power to the blade element is removed. Loss of optic power from the blade element occurs, or the UPS batteries suppling the blade elements are completely depleted. When power returns, the system is essentially in a coldboot condition. 100 Operations and Management Using OSM Applications

101 C Default Startup Characteristics NOTE: The configurations documented here are typical for most sites. Your system load paths might be different, depending upon how your system is configured. To determine the configuration of your system, refer to the system attributes in the OSM Service Connection. You can select this from within the System Load dialog box in the OSM LowLevel Link. Each NonStop NS14000 series system ships with these default startup characteristics: $SYSTEM disks residing in a Fibre Channel disk module connected to VIO enclosure group 100 with the disks in these locations: Path VIO Enclosure Fibre Channel Port FCDM Group Module Slot SAC Shelf Bay Primary Backup Mirror Backup Configured system load paths Enabled command interpreter input (CIIN) function If the automatic system load is not successful, additional paths for loading are available in the boot task. Using one load path, the system load task attempts to use another path and keeps trying until all possible paths have been used or the system load is successful. These 16 paths are available for loading and are listed in the order of their use by the system load task: Load Path Description Source Disk Destination Processor ServerNet Fabric 1 Primary $SYSTEMP 0 X 2 Primary $SYSTEMP 0 Y 3 Backup $SYSTEMP 0 X 4 Backup $SYSTEMP 0 Y 5 Mirror $SYSTEMM 0 X 6 Mirror $SYSTEMM 0 Y 7 Mirror backup $SYSTEMM 0 X 8 Mirror backup $SYSTEMM 0 Y 9 Primary $SYSTEMP 1 X 10 Primary $SYSTEMP 1 Y 11 Backup $SYSTEMP 1 X 12 Backup $SYSTEMP 1 Y 13 Mirror $SYSTEMM 1 X 14 Mirror $SYSTEMM 1 Y 15 Mirror backup $SYSTEMM 1 X 16 Mirror backup $SYSTEMM 1 Y 101

102 This illustration shows the system load paths: The command interpreter input file (CIIN) is automatically invoked after the first processor is loaded. The CIIN file shipped with new systems contains the TACL RELOAD * command, which loads the remaining processors. For default configurations of the Fibre Channel ports, Fibre Channel disk modules, and load disks, see Example Configurations of the VIO Enclosures and Fibre Channel Disk Modules (page 76). For system load procedures, see the NonStop NS14000 Hardware Installation Manual. 102 Default Startup Characteristics

103 D NonStop NS14000 Series System Architecture Integrity NonStop NS14000 series systems employ a unique method for achieving fault tolerance in a clustered processor environment, utilizing Intel Itanium microprocessors without cyclebycycle lockstepping. Instead, two or three microprocessors run the same instruction stream concurrently in a loose lockstep process. In loose lockstep: Each microprocessor runs at its own clock rate. Results of each command execution are compared on processor output to the ServerNet fabric. Error recovery and minor indeterminate processing results from one microprocessor do not cause output comparison errors. If the output of one microprocessor is incorrect, it is discarded, and its source microprocessor is taken offline for error handling and correction. The remaining functional hardware continues normal operation until the errant microprocessor is again operational. At that time, all microprocessors are synchronized and then proceed with executing instructions. NonStop NS14000 series systems use modular hardware implemented in enclosures with fiberoptic cabling between these enclosures. These enclosures reside in 19inch modular cabinets. NonStop Blade Complex The basic building block of NS14000 series systems compute engine is the NonStop Blade Complex, which consists of two or three CPU modules called NonStop Blade Elements. Each NonStop Blade Element houses two or four microprocessors called processor elements (PEs). A logical processor consists of one processor element from each NonStop Blade Element plus a logical synchronization unit (LSU). Although a logical processor physically consists of multiple processor elements, with each element contained in a separate enclosure, it is convenient to think of a logical processor as a single entity within the system. Each logical processor has its own memory and its own copy of the operating system, and it processes a single selfchecked instruction stream. NS14000 series logical processors are usually referred to simply as processors. All input and output to and from each NonStop Blade Element goes through a logical synchronization unit (LSU). The LSU interfaces with the ServerNet fabrics and contains logic that compares all output operations of the PEs in a logical processor, ensuring that all NonStop Blade Elements agree on the result before the data is passed to the ServerNet fabrics. A processor with two NonStop Blade Elements (NSBEs) and their associated LSUs makes up the dual modular redundant (DMR) NonStop Blade Complex, which is also referred to as a duplex processor. This duplex processor provides data integrity and system availability that is comparable to NonStop Sseries systems but at considerably faster processing speeds. Three NonStop Blade Elements plus their associated LSUs make up the triple modular redundant (TMR) NonStop Blade Complex, which is referred to as a triplex system. The triplex processor provides the same processing speeds as the duplex processor but also enables hardware fault recovery that is transparent to all but the lowest level of the HP NonStop operating system (OS). This diagram provides an overview of the modular NS14000 series system and shows one NonStop Blade Complex with four processors, two VIO enclosures (the I/O hardware), and links to external I/O and storage: NonStop Blade Complex 103

104 In the event of a processor fault in either a duplex or triplex processor, the failed component within a NonStop Blade Element (processor element, power supply, and so forth) or the entire Blade Element can be replaced while the system continues to run. A single NonStop NS14000 series system can have up to two NonStop Blade Complexes for a total of eight processors. Processors communicate with each other and with the system I/O over dual ServerNet fabrics. In the term ServerNet fabric, the word fabric is significant because it contrasts with the concept of a bus. A bus provides a single, fixed communications path between start and end points. A fabric is a complex web of links between electronic routers that provide a large number of possible paths from one point to another. Two communications fabrics, the X and Y ServerNet, provide redundant, faulttolerant communications pathways. If a hardware fault occurs on one of the ServerNet fabrics, communication continues on the other with hardware fault recovery transparent to all but the lowest level of the OS. Input/output components tap into the ServerNet fabrics through Fibre Channel and Ethernet ServerNet functionality in the VIO enclosures in the preceding diagram of the NonStop NS14000 series system. The Fibre Channel and Ethernet functionality in the VIO enclosures provides the system I/O to Fibre Channel storage devices and gigabit Ethernet communications networks. Although the hardware architecture differs from one series of NonStop servers to another, the NonStop NS14000 series system can be networked with other NonStop systems using the same message system and the same network software. 104 NonStop NS14000 Series System Architecture

105 In summary, these terms describe the NonStop NS14000 series processor: Term Processor element (PE) NonStop Blade Element Logical processor NonStop Blade Complex NonStop Blade Element enclosure Description A single Itanium microprocessor with its associated memory. A PE is capable of executing an individual instruction stream and I/O communication through fiberoptic links. Two or four PEs contained within a single NonStop Blade Element enclosure. One or more PEs from each NonStop Blade Element executing a single instruction stream. A duplex processor has two PEs forming a logical processor. A triplex processor has three PEs. Two NonStop Blade Elements (duplex processor) or three NonStop Blade Elements (triplex processor). A NonStop NS14000 series system includes up to two NonStop Blade Complexes. The power, processor, memory, and I/O hardware for a complete NonStop Blade Element. It mounts in a 19inch modular cabinet. Processor Element Each of the two or four processor elements in a NonStop Blade Element includes: An Intel Itanium microprocessor: NSED processors ship with the NonStop NS14000 server. NSES processors ship with the NonStop NS14200 server. 4 GB or 8 GB of main memory partitioned from DDR SDRAM DIMMs I/O interface with maintenance logic shared with the other PEs in the NonStop Blade Element Interface for fiberoptic I/O communications with the corresponding LSU Memory reintegration logic and fiberoptic links shared with the other PEs in the NonStop Blade Element and used for memory rendezvous between the NonStop Blade Elements This diagram provides an overview of the processor element architecture: Duplex Processor The DMR or duplex processor uses two NonStop Blade Elements, A and B, both with two or four microprocessors. Fiberoptic cables from each NonStop Blade Element connect the PEs to the LSUs. These LSUs then connect to two independent ServerNet fabrics via the VIO enclosures. These connections create communications redundancy in case one of the fabrics fails. For a description of the LSU functions, see Processor Synchronization and Rendezvous (page 107). A duplex processor includes these elements: Processor Element 105

106 Triplex Processor The TMR or triplex processor uses three NonStop Blade Elements, A, B, and C. As with the duplex processor, the fiberoptic cables connect the PEs to the LSUs, and these LSUs then connect to the two independent ServerNet fabrics via the VIO enclosure. Dual ServerNet fabrics create communications redundancy in case one of the fabrics fails. For a description of the LSU functions, see Processor Synchronization and Rendezvous (page 107). A triplex processor includes these elements: 106 NonStop NS14000 Series System Architecture

107 In a TMR or triplex processor, each LSU has inputs from three PEs. As with the duplex processor, the LSU keeps the three PEs in loose lockstep. The triplex processor provides fault tolerance upon failure of a PE by comparing the output from the three NonStop Blade Elements to determine which one is errant. This method eliminates the question of which output is valid when two outputs disagree in a duplex system. Processor Synchronization and Rendezvous Synchronization and rendezvous at the LSUs perform two main functions: Keep the individual PEs in a logical processor in loose lockstep through a technique called rendezvous. Rendezvous occurs to: Periodically synchronize the PEs so they execute the same instruction at the same time. Synchronization accommodates the slightly different clock speed within each PE. Allow each PE to individually and deterministically respond to asynchronous incoming interrupts and then to respond collectively as a single logical processor. Exchange software state information when performing operations that are distributed across PEs; for example, memory reintegration, error handling, and memory scrubbing. Compare output from each PE. If identical, the output is transmitted over the ServerNet fabrics. If the PE outputs are not the same, appropriate actions occur to identify the errant one and to recover from the failure. Under some failure conditions, it can be necessary to stop normal operations of the erring PE. Memory Reintegration Memory reintegration initiates processing in a PE whose operation has been stopped because the NonStop Blade Element diverged or has been replaced. This reintegration requires that all of the Processor Synchronization and Rendezvous 107