FTX Continuum 400 Series (PA-7100) Technical Service Guide 400 (7100... · FTX Continuum 400 Series...

FTX Continuum 400 Series (PA-7100) Technical Service Guide

(Last Updated 4/7/00)

Revision History

Rev. 1 (12/30/96) - Updates made to original manual, including newinformation on CPU/Memory upgrade procedure, 512-MB memorymodule, disk expansion cabinet, and part numbers).1/14/97 - Added Section 6.1.4 (Updating the Suitcase Label).2/7/97 - Added Section 4.4 (Isolating Memory Faults).2/12/97 - Updated Section 1.4.2 (Disk Configurations)2/26/97 - Updated disk expansion cabinet and tape drive informationin Section 1.3 (Hardware Components) and Section 1.4.1 (SystemConfigurations). Updated Section 1.4.2 (Disk Configurations).7/17/97 - Added D804 information to Section 1.9/25/97 - Updated Section 1.3.10/22/97 - Updated Sections 5 and 6 (removed references tosingle-initiator configurations).11/20/97 - Updated Section 1.4.1.12/22/97 - Updated Section 1.3 (PCI subsection).5/27/98 - Updated Section 1.3 (Peripherals subsection).

Title Page

file:///H|/CSDoc/400/title.htm (1 of 2) [4/7/2000 11:47:16 AM]

9/24/99 - Removed CPU upgrade procedure from Section 6.4/7/00 - Updated Sections 6.2.3 and 9.4

Title Page

file:///H|/CSDoc/400/title.htm (2 of 2) [4/7/2000 11:47:16 AM]

CSD Homepage

Title Page

Notice

Preface

1. Introduction

1.1 Overview●

1.2 Operating System Requirements●

1.3 Hardware Components●

1.4 Configurations●

1.4.1 System Configurations●

1.4.2 Disk Configurations●

1.5 System Specifications●

1.5.1 Physical●

1.5.2 Environmental●

1.5.3 Electrical●

1.6 Connectors and Cables●

1.6.1 PCI Cards●

1.6.2 SCSI Cables●

2. Operating and Software Maintenance Procedures

2.1 Starting the System●

2.1.1 Manual System Startup●

2.1.2 Defining a Boot Search Path●

2.2 Shutting Down the System●

2.3 Rebooting the System●

2.4 Configuring the Console Terminal●

2.5 Console Command Menu●

2.6 System Component Locations and Device Node Names●

2.6.1 System Bus Locations●

2.6.2 I/O PCI Bus Locations●

2.6.3 Virtual SCSI Bus (VSB) Locations●

2.6.4 Pseudo-Device Locations●

2.6.5 Listing of System Component Locations●

2.6.6 Device Node Names●

2.7 Maintenance Procedures●

2.7.1 Preparing to Remove a Suitcase●

2.7.2 Verifying Suitcase Replacement●

2.7.3 Preparing to Remove a PCI Card●

2.7.4 Verifying PCI Card Replacement●

2.7.5 Preparing to Remove a Flash Card●

Table of Contents

file:///H|/CSDoc/400/toc.htm (1 of 4) [4/7/2000 11:47:23 AM]

2.7.6 Verifying Flash Card Replacement●

2.7.7 Preparing to Remove a Disk Drive●

2.7.8 Verifying Disk Drive Replacement●

2.7.9 Preparing to Replace a Tape Drive●

2.7.10 Logically Adding a Tape Drive●

2.7.11 Preparing to Remove a System Base Power Supply●

2.7.12 Verifying System Base Power Supply Replacement●

2.7.13 Determining/Modifying Virtual SCSI Bus (VSB) Configurations●

2.7.14 Updating Flash Cards●

2.7.15 Automatic PROM Burning●

2.8 Veritas Volume Manager (VxVM) Utility●

2.8.1 Glossary of Terms Used With Volume Manager●

2.8.1.1 Disk device address●

2.8.1.2 Disk group●

2.8.1.3 Disk name●

2.8.1.4 Removed disk●

2.8.1.5 Failed disk●

2.8.1.6 Disk partition●

2.8.2 vxdiskadm Utility●

2.8.2.1 Add or initialize one or more●

2.8.2.2 Encapsulate one or more disks●

2.8.2.3 Remove a disk●

2.8.2.4 Remove a disk for replacement●

2.8.2.5 Replace a failed or removed disk●

2.8.2.6 Mirror volumes on a disk●

2.8.2.7 Move volumes from a disk●

2.8.2.8 Enable access to (import) a disk group●

2.8.2.9 Remove access to (deport) a disk group●

2.8.2.10 Enable (online) a disk device●

2.8.2.11 Disable (offline) a disk device●

2.8.2.12 Mark a disk as a hot-spare for a disk group●

2.8.2.13 Turn off the hot-spare flag for a disk●

3. Theory of Operation

3.1 System Bus●

3.2 Suitcase●

3.2.1 CPU-Memory Board●

3.2.1.1 Logic Section●

3.2.1.2 Console Controller Module●

3.2.2 Cooling Fans●

3.2.3 Power Supply●

3.3 System Base●

3.3.1 PCI I/O Subsystem●

Table of Contents


3.3.1.1 PCI Bus●

3.3.1.2 PCI Bridge Card●

3.3.1.3 PCI Adapter Cards●

3.3.2 Disk/Tape Drives●

3.3.3 System Base Power Supplies●

3.4 D81X Disk Expansion Tower/Cabinet●

3.4.1 D810 Disk Expansion Tower●

3.4.2 D811/2 Disk Expansion Cabinet●

3.5 Power Subsystem●

3.5.1 AC Input Module●

3.5.2 Suitcase Power Supply●


3.5.4 Grounding●

3.5.5 Power Fault and Status Reporting●

3.5.6 Power Switches●

3.6 Cooling System●

4. Fault Isolation

4.1 Component LEDs●

4.2 System Logs●

4.3 Component Status●

4.4 Isolating Memory Faults●

4.5 Troubleshooting Procedures●

4.5.1 Non-Critical Fault (System Operational)●

4.5.2 Critical Fault (System not Operational)●

4.6 Diagnostic Process eXecutive (DPX)●

4.7 Service Considerations●


4.7.2 PCI Cards●

4.7.3 System Placement●

5. Disk Expansion Cabinet Installation and Maintenance

5.1 Model Numbers●

5.2 Unpacking the Cabinet●

5.3 Removing the Cabinet from the Pallet●

5.4 Leveling the Cabinet●

5.5 Installing the Skirt Kit●

5.6 Connecting the Disk Drives to the System●

5.7 Inspecting the Cabinet●

5.8 Powering On the Cabinet●

5.9 Initializing the Disk Drives●

5.10 Removal/Replacement Procedures●

Table of Contents


5.10.1 Shelf Fan Removal●

5.10.2 Shelf Fan Replacement●

6. Upgrades

6.1 Memory Upgrade●

6.1.1 Hardware Components●

6.1.1.1 Suitcase●

6.1.1.2 Upgrade Kits●

6.1.2 ESD Requirements●

6.1.3 Upgrade Procedure●

6.1.3.1 Burning the Board ID PROM●

6.1.3.2 Installing Memory Modules●

6.1.3.3 Switching the Operational Status of the Boards●

6.1.4 Updating the Suitcase Label●

6.2 D811/2 Disk Expansion Cabinet Upgrade●

6.2.1 Upgrade Kit Marketing IDs●

6.2.2 Installing Disk Expansion Cabinet Shelves●

6.2.3 Connecting the Shelf Components●

6.2.4 Initializing the Disk Drives●

7. Service Strategies

7.1 Maintenance Strategy●

7.2 Logistics Strategy●

7.3 Training Strategy●

8. Related Documentation

9. Part Numbers

9.1 Suitcase●

9.2 System Base●

9.3 Tape Drives/Modem●

9.4 D81X Disk Expansion Tower/Cabinet●

Table of Contents


NoticeThe information contained in this document is subject to change without notice.

STRATUS COMPUTER, INC. MAKES NO WARRANTY OF ANY KIND WITH REGARD TO THISMATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OFMERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. Stratus Computer, Inc.,shall not be liable for errors contained herein or incidental or consequential damages in connection withthe furnishing, performance, or use of this material.

Software described in Stratus documents (a) is the property of Stratus Computer, Inc., or the third party,(b) is furnished only under license, and (c) may be copied or used only as expressly permitted under theterms of the license.

This document is protected by copyright. All rights are reserved. No part of this document may becopied, reproduced, or translated, either mechanically or electronically, without the prior written consentof Stratus Computer, Inc.

Stratus, Continuum, Continuous Processing, StrataNET, FTX, and the Stratus logo are registeredtrademarks ofStratus Computer, Inc.

XA, XA/R, StrataLINK, RSN, SINAP, Isis, Isis Distributed, RADIO, and the SQL/2000 logo aretrademarks ofStratus Computer, Inc.

Hewlett-Packard and HP are registered trademarks of Hewlett-Packard.IBM PC is a registered trademark of International Business Machines Corporation.Sun is a registered trademark of Sun Microsystems, Inc.UNIX is a registered trademark of X/Open Company, Ltd., in the U.S.A. and other countries.PA/RISC is a trademark of Hewlett-Packard.

All trademarks are the property of their respective owners.

Manual Name: FTX Continuum 400 Series Technical Service Guide

Last Update: February 12, 1997

Stratus Computer, Incorporated

Customer Service Documentation Department111 Powdermill RoadMaynard, MA 01754-3409

Notice

file:///H|/CSDoc/400/notice.htm (1 of 2) [4/7/2000 11:47:29 AM]

WarningThe equipment documented in this manual generates and uses radio frequency energy, which if notinstalled and used in strict accordance with the instructions in this manual, may cause harmfulinterference to radio communications. The equipment has been tested and found to comply with thelimits for a Class A computing device pursuant to Subpart J of Part 15 of FCC rules, which are designedto provide reasonable protection against such interference when operated in a commercial environment.

Operation of this equipment in a residential area is likely to cause interference, in which case the user athis own expense will be required to take whatever measures may be required to correct the interference.

This document contains Stratus Proprietary and Confidential Information. It is provided to you andits use is limited by the terms of your contractual arrangement with Stratus regarding maintenance anddiagnostic tools.

Copyright© 1996 by Stratus Computer, Inc. All rights reserved.

Notice

file:///H|/CSDoc/400/notice.htm (2 of 2) [4/7/2000 11:47:29 AM]

PrefaceThe FTX Continuum 400 Series Technical Service Guide contains technical information which pertainsto the servicing of Continuum 400 Series systems in accordance with Stratus servicing policies.

This document is organized as follows:

Section 1 - Introduction

Section 2 - Operating and Maintenance Procedures

Section 3 - Theory of Operation

Section 4 - Fault Isolation

Section 5 - Disk Expansion Cabinet Installation

Section 6 - Upgrades

Section 7 - Service Strategies

Section 8 - Related Documentation

Section 9 - Part Numbers

Audience

This guide is intended for authorized service personnel who install and maintain Stratus systems, andwho have completed Stratus field-service training courses.

Preface

file:///H|/CSDoc/400/preface.htm [4/7/2000 11:47:37 AM]

1. IntroductionThis section describes the major components and specifications of Stratus Continuum 400 Series(PA-7100) systems. It covers the following topics.

Overview ●

Operating system requirements ●

Hardware components ●

System configurations ●

System specifications ●

Connectors and cables ●

1.1 OverviewThe Continuum 400 Series is an entry-level Stratus RISC system that provides customers with aphysically small, continuously available computer. Its logic architecture is based around the HewlettPackard® PA-RISC HP7100 microprocessor which is available in uni or twin processor designs runningat 96 MHz. The I/O section of the 400 Series is based on the PCI (Peripheral Component Interconnect)I/O bus and uses standard commodity PCI cards. This is the first Continuum Series product to utilize PCItechnology.

The Continuum 400 Series system is customer installable. It must be installed in either a computer roomor unoccupied office environment.

1.2 Operating System RequirementsThe minimum FTX OS releases for the Continuum 400 Series systems are:

FTX 3.1 for uni processor models ❍

FTX 3.1.1 for twin processor models ❍

FTX 3.2 for systems containing disk expansion towers/cabinets or 512-MB memorymodules

❍

The 400 Series is software compatible with the Continuum Series 600/1200 systems.

1.3 Hardware ComponentsThe two major assemblies in the Continuum 400 Series system are the system base and two suitcaseunits. (See Figure 1.) Optional disk expansion towers/cabinets are also available.

Figure 1. Continuum 400 Series System

Introduction

file:///H|/CSDoc/400/section1.htm (1 of 9) [4/7/2000 11:48:13 AM]

The two suitcases are identical. Each houses the following components:

CPU-Memory motherboard (contains CPU/cache modules, memory modules, and a ConsoleController module)

❍

power supply ❍

three CPU cooling fans❍

As stated previously, the CPU/cache modules are available in uni or twin processor designs running at 96MHz. The uni processor contains one logical/two physical CPUs and the twin processor has twological/four physical CPUs. Both the uni and twin designs are available in a low cache 512 KB version(256 KB instruction cache and 256 KB data cache) or high cache 2 MB version (1 MB instruction cacheand 1 MB data cache).

A memory module contains 128 MB (model M128) or 512 MB (model M512) of memory. TheCPU-Memory motherboard can contain from one to four memory modules, for a maximum of 512 MB(using 128-MB modules) or 2 GB (using 512-MB modules). For the most current information onavailable memory refer to the Continuum Memory Subsystem Technical Reference.

Figure 2 shows the locations of the components in the suitcase.

Figure 2. Suitcase Components

Introduction


The system base is the backbone of the system. It contains the following components:

PCI card cage ❍

PCI bus connectors ❍

disk enclosure ❍

power supplies for the PCI bus and disk devices ❍

interconnect cables❍

The PCI card cage houses up to 14 PCI cards on two separate PCI buses. One slot in each PCI enclosureis reserved for the PCI bridge (PCIB) card which is the interface between the CPU and the PCI bus forthat enclosure. The PCIB contains a PCMCIA (PC Memory Card International Association) flash ram(20 MB) card which acts as a boot disk. Another slot in each PCI enclosure is configured for a fast SCSIdual-initiator PCI card that drives the disk devices associated with the PCI bus for that enclosure.

The PCI buses are powered separately so that at least one of the disk PCI cards will always have power.When a PCI enclosure access door is opened to replace a card, its PCI bus is automatically powereddown.

The disk enclosure supports up to eight physical (four logical) 3.5" disk drives for data storage.Additional disk drives are housed in standalone disk towers or expansion cabinets. The disks areinterchangeable between the system base and the disk expansion tower/cabinet.

Introduction


There are no internal tape drives. All tape drives are standalone units.

Figure 3 shows the locations of the components in the system base.

Figure 3. System Base Components

The D810 disk expansion tower supports up to five D800 series disk drives. The system can mirror thedisk drives in two towers. A maximum of two disk expansion towers is supported. The D811/2 diskexpansion cabinet provides even larger storage capacity. It can contain from two to eight disk shelves, inmultiples of two shelves. Each shelf can contain up to five or six disk drives, depending on itsconfiguration. Only one disk expansion cabinet is supported.

The Continuum 400 Series system does not contain a physical control panel. Operating commands areentered at the system console which is connected to the system via the Console Controller module in thesuitcase. The Console Controller functionality includes the calendar clock, the serial ports for the systemconsole, RSN, and an auxiliary RS-232 port for connection to an optional uninterruptable power supply(UPS) or console printer.

Several components in the Continuum 400 Series system contain status LEDs that indicate the conditionof the component. Section 4.1 contains information on interpreting the various status conditions that mayappear on the LEDs.

Continuum 400 Series power supplies are auto-configurable for 115 VAC or 230 VACVAC, singlephase, running at 50 or 60 Hz. The two power supplies in the system base load-share the disk subsystem,but each one supplies power to only one PCI enclosure. The suitcase power supplies are independent of

Introduction


each other, with each one supplying power only to the suitcase in which it is housed.

Continuum 400 Series systems do not have internal battery backup to guard against power failureshutdowns. Backup is optional and is provided by an external UPS manufactured by APC (AmericanPower Conversion) or through the customer's site-wide UPS.

The following tables list model numbers and give a brief description of the CPU-Memory motherboards(suitcases), PCI cards, and disk/tape drives supported in Continuum 400 Series systems.

NOTE: The hardware components shown in the following tables are at the base minimumrevisions approved for operation at the time of publication. For current revisionrequirements and complete revision history, refer to the HQ Service Support Rev Dir.

NOTE: Although their descriptions are similar, sub model 00 boards are not compatiblewith sub model 10 boards. The difference is that the sub model 10 CPU-Memorymotherboards have an updated ASIC chip which is not compatible with sub model 00boards.

Suitcases

Model Description

G811-G814 (Sub Model 00) 96 MHz, Uni processor, 512 KB cache, 128-512 MB memory (128-MBmemory modules)

G811-G814 (Sub Model 10) 96 MHz, Uni processor, 512 KB cache, 128-512 MB memory (128-MBmemory modules)

G815-G818 (Sub Model 00) 96 MHz, Uni processor, 2 MB cache, 128-512 MB memory (128-MBmemory modules)

G815-G818 (Sub Model 10) 96 MHz, Uni processor, 2 MB cache, 128-512 MB memory (128-MBmemory modules)

G821-G824 (Sub Model 00) 96 MHz, Twin processor, 512 KB cache, 128-512 MB memory (128-MBmemory modules)

G821-G824 (Sub Model 10) 96 MHz, Twin processor, 512 KB cache, 128-512 MB memory (128-MBmemory modules)

G825-G828 (Sub Model 00) 96 MHz, Twin processor, 2 MB cache, 128-512 MB memory (128-MBmemory modules)

G825-G828 (Sub Model 10) 96 MHz, Twin processor, 2 MB cache, 128-512 MB memory (128-MBmemory modules)

G835-G838 (Sub Model 00) 96 MHz, Uni processor, 2 MB cache, 512 MB-2 GB memory (512-MBmemory modules)

G835-G838 (Sub Model 10) 96 MHz, Uni processor, 2 MB cache, 512 MB-2 GB memory (512-MBmemory modules)

G845-G848 (Sub Model 00) 96 MHz, Twin processor, 2 MB cache, 512 MB-2 GB memory (512-MBmemory modules)

Introduction


http://www.cac.stratus.com/vos/doc2/rev_dir

G845-G848 (Sub Model 10) 96 MHz, Twin processor, 2 MB cache, 512 MB-2 GB memory (512-MBmemory modules)

PCI Cards

Note: The PCI cards shown in the following table are based on information available at the timeof publication. For current (and more detailed) PCI information refer to the PCI/PMC AdapterTechnical Reference.

Model Description Minimum OSRelease

K138 PCI bridge card FTX 3.1

E525 PCMCIA flash card FTX 3.1

U401-01 4-port, 1-MB synchronouscard (V.36) FTX 3.1

U401-02 4-port, 1-MB synchronouscard (RS-422) FTX 3.1

U401-03 4-port, 1-MB synchronouscard (X.21) FTX 3.1


U402 8-port, 1-MB synchronouscard FTX 3.1


U403-02 4-port, 4-MB synchronouscard (RS-422) FTX 3.3

U403-03 4-port, 4-MB synchronouscard (X.21) FTX 3.3


U404 8-port, 4-MB synchronouscard (RS-232) FTX 3.3

U420 T1/E1/ISDN-PRI card FTX 3.3

U450 8-port asynchronous card FTX 3.3

U501 Fast wide single-endedSCSI card (1 external port) FTX 3.1

U502 Fast wide differential SCSIcard (3 external ports) FTX 3.2

Introduction


U510 Ethernet card (10/100Mbps) FTX 3.1

U512 2-port ethernet card (10/100Mbps) FTX 3.1

U513 1-port ethernet card (10/100Mbps) FTX 3.1

U530 1-port FDDI card (100Mbps) FTX 3.1

U542 1-port ATM card (155 Mbps) FTX 3.4

Peripherals

Note: The disk and tape drives shown in the following table are based on information available atthe time of publication. For current (and more detailed) disk information refer to the ContinuumDisk Drives Technical Reference. The document is also available in PDF format at athttp://www.cac.stratus.com/CSDoc/home/disk.htm. For more information on tape drives, refer tothe DDS DAT Tape Drive Technical Reference. The document is also available in PDF format athttp://www.cac.stratus.com/CSDoc/home/tape.htm.

Model DescriptionMinimumOSRelease

D802 2-GB, 3.5", 7200 rpm,SCSI disk drive FTX 3.1.1

D803 4-GB, 3.5", 7200 rpm,SCSI disk drive FTX 3.1.1

D804 9-GB, 3.5", 7200 rpm,SCSI disk drive FTX 3.2

T304

140-MB, 6250 bpi,SCSI desktop tapedrive (includesstandalone SCSIconversion unit,power cord, andcables).

FTX 3.1

T403-004

4-GB, IBM 3480/3490media-compatible,tape drive(10-cartridge)

FTX 3.1.1

T404

4-GB, 3490Emedia-compatible,tape drive(10-cartridge)

FTX 3.4(Q4/98)

Introduction


T801 4-mm DDS2 DAT tapedrive FTX 3.1

T8024-mm DDS2 DAT tapedrive with 6-cartridgeautoloader

FTX 3.1

T804 525-MB/1.2-GB QICtape drive FTX 3.3

T805 12-GB, DDS3 DAT tapedrive (1 cartridge) FTX 3.3

T80672-GB, DDS3 DAT tapedrive (6-cartridgeautoloader)

FTX 3.3

V103 Terminal

V105 Terminal FTX 3.2

1.4 Configurations

1.4.1 System Configurations

The hardware configuration requirements and restrictions for the various Continuum 400 Series modelsare shown in the following table.

NOTE: The configurations shown in the following table are based on information available at the timeof publication. For current (and more detailed) configuration information refer to the StratusConfiguration Specification Document No. ES-000096 on the Stratus internal WWW site athttp://webserv.hw.stratus.com/ (go to Records, then Stratus Configuration Specifications). Thedocument is available in Word or PDF format.

Component Model 412 Model 415 Model 422 Model 425

CPU-Memorymotherboard(suitcase)

G811-G814G815-G818

G835-G838G821-G824

G825-G828

G845-G848

CPU module Uni 96 MHz Uni 96 MHz Twin 96 MHz Twin 96 MHz

Cache 512 KB 2 MB 512 KB 2 MB

No. logical CPUs 1 1 2 2

Introduction


http://websurfer.sw.stratus.com/groups/hweng.dir/index.htm

No. memorymodules(128-MB or512-MB)

Min. = 2

Max. = 8

Min. = 2

Max. = 8

Min. = 2

Max. = 8

Min. = 2

Max. = 8

Duplexedmemory (128-MBmemorymodules)

Min. = 128MB

Max. = 512MB

Min. = 128MB

Max. = 512MB

Min. = 128MB

Max. = 512MB

Min. = 128MB

Max. = 512MB

Duplexedmemory (512-MBmemorymodules)*

NA

Min. = 512MB

Max. = 2 GB

NA

Min. = 512MB

Max. = 2 GB

No. PCI adaptersMin. = 2

Max. = 14

Min. = 2

Max. =14

Min. = 2

Max. = 14

Min. = 2

Max. = 14

K138U401U402U501U502U510

2 required8 max.4 max.2 required2 max.8 max.

2 required8 max.4 max.2 required2 max.

Introduction


2. Operating and Software MaintenanceProceduresThis chapter explains the basic operating procedures used for system operation and maintenance. Topics coveredinclude the following:

Starting the system ●

Shutting down the system ●

Rebooting the system ●

Configuring the console terminal ●

Console command menu ●

System component locations ●

Maintenance procedures ●

Veritas Volume Manager (VxVM) utility●

There is no physical control panel on Continuum 400 Series systems. Operating commands are entered at thesystem console which is connected to the system via the Console Controller module in the suitcase.

2.1 Starting the System

When the system is powered up, it displays a PROM: prompt on the system console. The system waitsapproximately 10 seconds before proceeding with an automatic boot process in order to provide the option ofperforming a manual boot.

The automatic boot device path(s) is defined in the Console Controller prom code. If the boot path has not beenburned into the Console Controller prom, the system will wait at the PROM: prompt until a boot command is

entered. To get a listing of PROM commands and syntax, enter the help command at the PROM: prompt.

2.1.1 Manual System Startup

Perform the following procedure to boot the system manually.

NOTE: The system can be booted from the flash card in either card cage 2 or 3.

1. Power on the system and press any key when the following message is displayed:

Hit any key to enter manual boot mode, elsewait for auto boot

2. When the PROM: prompt appears, enter the following command to boot the system using the flashcard in card cage 2.

boot 2

Operating & Software Maintenance Procedures


NOTE: To boot from the flash card in card cage 3, enter boot 3 instead.

3. Once the system finds the boot device, it loads the bootloader and displays the bootloader prompt

$$. To start the boot process, enter the following command:

go

NOTE: If the go command does not work, enter the ramgo command. ramgoloads and boots the kernel from a ramdisk instead of using the root file system. For moreinformation on the ramgo command, refer to FTX System Administrator's Guide: GeneralServices (R455X).

The system console displays a series of messages as the booting progresses. At the conclusion of thesemessages, the system indicates it is in single user mode and prompts you to proceed as follows.

INIT: SINGLE USER MODE

Type Ctrl-d to proceed with normal startup,(or give root password for system maintenance):

4. Press CTRL-D to proceed with normal startup.

5. When the system prompts you to specify a system startup mode, enter 2 for multi-user mode.

6. When the system displays the login prompt, log in as root.

2.1.2 Defining a Boot Search Path

The Console Controller contains a path partition that specifies the boot device search path used for automaticbooting. The default boot path can be changed in the standby Console Controller using the hwmaintburnprom command.

Perform the following procedure to define a search path to automatically boot the system.

1. Boot the system if it is not already booted.

2. Enter the following command to determine the boot device (flash card) locations:

hwmaint ls flash

3. Record the slot information for each flash card (2000 or 3000) found in the slotloccolumn.

4. Enter the following command to determine which Console Controller is on standby.

hwmaint ls ccbd



The information in the State and Slotloc fields shows which Console Controller is on

standby (stby) and its slot location (either cc0 or cc1).

5. Enter the following command to change to the /etc/prom_code directory.

cd /etc/prom_code

This directory contains a file named bootpath which is a default configuration file supplied byStratus. It can be modified, or a new configuration file added, using vi or another editor. Refer to theFTX System Administrator's Guide: General Services (R455X) for more information.

6. Burn the boot path information found in the bootpath file (or new configuration file) into thestandby Console Controller by entering the following command.

hwmaint burnprom -p path -f pathname cc number

where pathname is the full path name to the configuration file

(/etc/prom_code/bootpath for the default file) and number is the slot location

for the standby Console Controller as determined in Step 4, which is either 0 or 1.

NOTE: You must specify the standby Console Controller. An error messagewill be displayed if you specify the online board.

A rotating clock symbol (/) is displayed while the command is executing.

7. Switch the status of the two Console Controllers (the online board becomes standby and vice versa)by entering the following command.

hwmaint switchover cc number

where number is the slot location for the standby Console Controller as determined in

Step 4, which is either 0 or 1.

8. Enter the following command to update the old online board (now the standby board) to the currentlevel of the updated board (now the online board).

hwmaint burnprom -p sync cc number

where number still refers to the standby Console Controller, but because of the switchover thatoccurred in Step 7, this is now the opposite slot number from the one in Step 7. (For example, if thestandby board was in slot 0 in Step 7, it is now in slot 1.)

NOTE: It is important that the correctly configured board is the online board before thesync command is executed. Otherwise, the updated board will be changed back to theconfiguration of the unchanged board.



2.2 Shutting Down the SystemThis section describes how to perform an orderly shutdown of the system.

1. Login as root.

2. Check to see if any users are on the system by entering the who -H command.

3. If there are users on the system, enter the wall command.

Then enter a message (on one or more separate lines) announcing the shutdown. End the message by pressingCTRL-D.

Example:

The system will shut down in 5 minutes.

Please log off.

<CTRL-d>

4. To power down the system from multiuser state:

Enter the following command:

shutdown -i0

where -i0 changes the system state to state 0 (off).

The system will prompt you with questions about the reason for the shutdown. You will then be prompted toverify the shutdown. Enter y to continue with the shutdown.

CAUTION: Do not turn off power to the machine until the shutdown procedure is completely finishedor damage to data or components might result. It is safe to power it off when the following message isdisplayed:

The system is down.

To power down the system from the single user state:

Enter the following command:

shutdown -y -i0 -g0

where the -y option assumes "yes" to all prompts, -i0 shuts down the system to state 0 (off), and -g0defines the grace period as 0 seconds.

CAUTION: Do not turn off power to the machine until the shutdown procedure is completely finishedor damage to data or components might result. It is safe to power it off when the following message isdisplayed:



The system is down.

2.3 Rebooting the SystemThis section describes how to reboot the system.

1. Login as root.

2. Check to see if any users are on the system by entering the who -H command.

3. If there are users on the system, enter the wall command followed by a message (on one or more separatelines) announcing the shutdown. End the message by pressing the CTRL and D keys simultaneously.

Example:

wall

The system will be coming down in 5 minutes.

Please log off.

<CTRL-d>

4. To halt and reboot the system, enter the following command:

shutdown -i6

where -i6 indicates state 6, meaning stop and reboot.

2.4 Configuring the Console TerminalPerform the following procedure to configure the console terminal.

1. Consult the terminal manual for specific configuration information or instructions.

2. Execute the following commands at the console or (preferably) put them into the /etc/profilefile.

TERM=terminal_typeexport TERMtput inittabs

where TERM establishes the terminal type, tput initializes the terminal, and tabs sets tabs.



2.5 Console Command MenuThe Console Controller supports a command menu that allows the user to issue key machine managementcommands to the system from the system console.

NOTE: There is no immediate power-off menu option. If conditions require that the system be powered offimmediately, turn off the two power switches on the rear of the system base.

To access the console command menu, press the CTRL and BREAK keys simultaneously. This puts the consoleinto command mode and displays the following menu:

help...........displays command list.shutdown..........begin orderly system shutdown.restart_cpu.......force CPU into kernel dump/debug mode.reset_bus.........send reset to system.status............report state of system indicator lamps.history...........display switch closure history.quit, q...........exit the front panel command loop.. .............display firmware revision.

The following describes the actions of each command:

help - Displays the menu list on the screen.

shutdown - Initiates an immediate orderly system shutdown.

restart_cpu - Issues a broadcast interrupt to all CPU boards in the system. Generates asystem dump or enters the kernel debugger, depending on the autodumpsetting.

reset_bus - Issues a reset command to all boards on the main system bus. Immediatelykills all system activities and reboots the system. This command should beexecuted only when the system is inoperable and the restart_cpucommand does not solve the problem.

status - Reports the status of system indicator lamps.

history - Displays a list of the most recently entered console commands.

quit, q - Exits the console command menu and returns the console to its normal mode.

. - Displays the current firmware version number.

2.6 System Component Locations and Device NodeNamesThis section describes how to determine and specify the physical location of hardware components.

Most devices in Continuum 400 Series systems are assigned two identifiers: a slot location and a device node name.The slot location is used to uniquely identify the physical location of a hardware component or the virtual mappinglogical location of a virtual SCSI bus (VSB). The following slot location scheme is used:



System bus slot locations identify suitcases (and corresponding CPU-Memory boards) and PCI card cages(and associated bridge (PCIB) cards).

●

I/O PCI adapter card and port slot locations identify the corresponding PCI card cage slots and the ports oneach card (if any).

●

VSB slot locations identify VSB locations for supported devices.●

Pseudo-device locations identify components that do not have precise slot locations, such as buses, powersupplies, and Console Controllers. The slot locations for these components are identified by abbreviations.The location can also include a partial numeric address.

●

A device node name is an entry into the device drivers that allows a program to perform I/O operations on thespecified device (normally a disk).

The slot location for a component and the device node name are not necessarily equivalent. For example, there arecomponents that have a slot location but no device node (e.g., a fan), and there are device nodes that do not map toa specific hardware entity (pseudo devices).

2.6.1 System Bus Locations

The system bus connects the CPU-Memory boards in the suitcases with the PCIB cards in the card cages. TheCPU/Memory board (and corresponding suitcase) slot locations are 0 and 1. The PCIB (and corresponding cardcage) slot locations are 2 and 3.

2.6.2 I/O PCI Bus Locations

The PCI card cage slot numbering convention is as follows:

PCIB slot # Constant (0) Card Cageslot # Port #

2 or 3 always 0 0-7 Physical portlocation on PCI card

Example: 2 0 7 1 specifies port 1 on the U501 SCSI controller card in slot 7 of card cage #2.

2.6.3 Virtual SCSI Bus (VSB) Locations

A VSB is a pair of SCSI adapters operating in parallel with each other. For system redundancy the system can useeither board to access the external SCSI devices.

The VSB mapping scheme describes the logical location of a physical SCSI bus. The VSB can consist of either oneor two SCSI controllers located anywhere in the system, as long as they are physically connected to the same SCSIbus.

Devices on the bus are accessed through the VSB slot location. In the system base, four VSBs are defined off theU501 SCSI controllers: dual-initiated buses 0 and 1 for disk drives and single-initiated buses 2 and 3 for tapedrives. Six VSBs (4-9) are defined off the U502 SCSI controllers for disk drives in the disk expansion towers andcabinets.

The SCSI ID is the bus address associated with a device.

The VSB slot location numbering convention is as follows:



VSB # SCSI ID # Logical unit number(LUN)

0-31 0-5, 8-150 for most devices (tape driveautoloaders use both 0 and1)

Base system:0-1 for disk drives (dual-initiated)

0-3 0

Base system:2-3 for tape drives (single-initiated)

1-7 0, 1

Expansion towers/cabinets:4-9 for disk drives

1-5, 8-13 0

Example: 1 2 0 specifies the VSB location of a disk drive (in the system base) with SCSI ID of 2 and LUN 0on VSB 1.

The following example shows the physical location, SCSI ID, and VSB location of some disk drives on VSB 4.These disk drives are in a disk expansion tower or on the first shelf in a pair in a disk expansion cabinet.

Slot

#7 6 5 4 3 2 1 0

Type of

Device

PS1

PS2

Disk5

Disk4

Disk3

Disk2

Disk1

SCSIConv.Unit

SCSI

ID---- ---- 5 4 3 2 1 ----

VSB

Location---- ---- 4 5

04 40

4 30

4 20

4 10

2.6.4 Pseudo-Device Locations

The slot locations for these components are identified by abbreviations. The location can also include a partialnumeric address, for example, a logical SCSI cabinet.

The following are some examples of pseudo-device locations:

Component Slot Location

Console Controller in suitcase0

cc 0



Calendar clock cc 0 5

System base power supply A dbpa

2.6.5 Listing of System Component Locations

The following table shows the slot locations assigned to various system components (when viewed from the front ofthe system):

Slot Location Component

sc0 suitcase (right side)

sc1 suitcase (left side)

0 CPU-Memory board (suitcase 0)

1 CPU-Memory board (suitcase 1)

cc 0 console controller on CPU board 0

cc 1 console controller on CPU board 1

cc [0|1] 0 primary console device

cc [0|1] 1 remote service network (RSN) device

cc [0|1] 2 secondary console device

cc [0|1] 5 clock

2 PCIB/card cage (right side)

2 0 PCI Bus

3 0 PCI Bus

2 0 0 0 PCMCIA flash card

2 0 1 PCI card 1

2 0 2 PCI card 2

2 0 3 PCI card 3

2 0 4 PCI card 4

2 0 5 PCI card 5

2 0 6 PCI card 6

2 0 7 PCI card 7 (single-ended SCSI controller)

2 0 7 0 SCSI controller port 0



3 PCIB/card cage (left side)

3 0 0 0 PCMCIA flash card



3 0 1 PCI card 1

3 0 2 PCI card 2

3 0 3 PCI card 3

3 0 4 PCI card 4

3 0 5 PCI card 5

3 0 6 PCI card 6

3 0 7 PCI card 7 (single-ended SCSI controller)




vsb 0 virtual SCSI bus from 2 0 7 1 & 3 0 7 2

vsb 0 0 0 disk drive 0/0




vsb 1 virtual SCSI bus from 2 0 7 2 & 3 0 7 1





vsb 2 virtual SCSI bus from 2 0 7 0 (tape drive)

vsb 3 virtual SCSI bus from 3 0 7 0 (tape drive)

vsb 4 virtual SCSI bus from 2 0 3 0 and 3 0 3 0(dual-initiator)

vsb 5

virtual SCSI bus from 2 0 3 1(single-initiator)

virtual SCSI bus from 2 0 3 1 and 3 0 3 1(dual-initiator)

vsb 6


virtual SCSI bus from 2 0 3 2and 3 0 3 2(dual-initiator)

vsb 7 virtual SCSI bus from 2 0 6 0 and 3 0 6 0(dual-initiator)

vsb 8





vsb 9



dbpa system base power supply (right side)

dbpb system base power supply (left side)

To view a listing of the system components, enter the following command:

hwmaint ls |pg

The display is shown in the following format.

Sample Screen:

●



2.6.6 Device Node Names

Device node names have the following format: c#a#d#s# where c# specifies the logical slot number of the device,a# specifies the VSB number, d# is the SCSI ID, and s# is the partition number (always 14).. The followingconvention is used for device node names:

c# a# d# s#

always c0for VSBs

a0-a31 d1-d5,d8-d15 always s14

For example, the device name for the VSB in location vsb 1 3 0 is c0a1d3.

The following examples show the physical location, SCSI ID, VSB location, and device node name of each of thedisks on VSB 4. These disk drives are in the first pair of shelves in the first disk expansion cabinet.

Shelf # 1

Slot

#7 6 5 4 3 2 1 0



Type of

Device

PS1

PS2 Disk 5 Disk 4 Disk 3 Disk 2 Disk 1

SCSIConv.Unit

SCSI

ID---- ---- 5 4 3 2 1 ----

VSB

Location---- ---- 4 5 0 4 4 0 4 3 0 4 2 0 4 1 0 ----

DeviceNodeName

---- ---- c0a4d5s14 c0a4d4s14 c0a4d3s14 c0a4d2s14 c0a4d1s14 ----

Shelf # 2

Slot

#7 6 5 4 3 2 1 0

Type of

Device

PS1

PS2 Disk 11 Disk 10 Disk 9 Disk 8 Disk 7 Disk 6

SCSI

ID---- ---- 13 12 11 10 9 8

VSB

Location---- ---- 4 13 0 4 12 0 4 11 0 4 10 0 4 9 0 4 8 0

Device

NodeName

c0a4d13s14 c0a4d12s14 c0a4d11s14 c0a4d10s14 c0a4d9s14 c0a4d8s14

2.7 Software Maintenance ProceduresThis section provides a set of software procedures used to maintain a Continuum 400 Series system.

2.7.1 Preparing to Remove a Suitcase

A suitcase is hot pluggable (can be removed without disrupting the system or entering any commands) if it wasduplexed and the other suitcase is now running simplexed. The amber LED on the failed suitcase should be on andits yellow and green LEDs should be off. The failed suitcase must be replaced with an identically configuredsuitcase.



Verify the state of the failed suitcase as follows:

1. Note the number (either 0 or 1) on the front label of the failed suitcase.

2. Enter the following command to verify the state of the suitcase:

hwmaint ls slot_number

where slot_number is the number on the suitcase front label.

The information in the State and Code columns of the display will show information on thefailed suitcase.

2.7.2 Verifying Suitcase Replacement

After the replacement suitcase is installed, the system automatically tests it and duplexes it with the other suitcase.When testing is complete, the amber LED will turn off and the green LED will come on. Perform the following toverify proper operation of the suitcase:

1. Enter the hwmaint ls slot_number command. Check the State column to verify that

the replacement suitcase is now online. It should show as onln.

2. Enter the following command to update the date to ensure that it is the same on both Console Controllersin the suitcase.

date mmddHHMMyyyy

where mm specifies the month, dd is the day of the month, HH is the hour (24-hour system), MM is

the minute, and yyyy is the year.

NOTE: The month, day, and year can be omitted; if so, the current values are assumed.

2.7.3 Preparing to Remove a PCI Card

When an I/O card cage is opened, its PCI bus is automatically powered down and all PCI cards housed within thecard cage are logically removed from the system. Before opening the card cage, verify the location and state of thefailed PCI card as follows:

1. Check the LEDs on the PCI card cage slot where the PCI card is located.

2. If the red LED is on, enter one of the following commands:

hwmaint ls bd

hwmaint ls model_number

where model_number is the model number for that card (for example, u510)



The information in the Slotloc, State, and Code columns of the display will showinformation on the failed PCI card.

2.7.4 Verifying PCI Card Replacement

After the replacement PCI card is installed and the card cage door is closed, the system automatically tests andbrings all the cards in that card cage online. To verify proper operation of the PCI cards, enter the hwmaintls model_number command and look in the Slotloc, State, and Code columns of thedisplay to verify that the replacement card is now online.

2.7.5 Preparing to Remove a Flash Card

The flash card is hot pluggable (can be removed without disrupting the system or entering any commands).However, you should verify its location by entering the following command:

hwmaint ls flash

The information in the Slotloc, State, and Code columns of the display will show information on the

failed flash card. The shot location for a flash card is either 2000 or 3000.

2.7.6 Verifying Flash Card Replacement

After the replacement flash card is installed, verify proper operation by entering the hwmaint ls flashcommand again and look in the Slotloc, State, and Code fields of the display to verify that thereplacement card is now online.

2.7.7 Preparing to Remove a Disk Drive

A disk drive is hot pluggable (can be removed without disrupting the system) if its amber LED is blinking. Performthe following procedure before removing a failed disk drive.

1. Log in as superuser (su).

2. Check the state of the disk to be replaced by entering the following command:

hwmaint ls disk

The screen output lists the status of all configured disk drives on the system. The following sampleoutput is a partial listing of the disk drives in a system. It shows two disk drives in the system diskenclosure and six disk drives in a dual-initiator configuration in an expansion tower or cabinet.

Sample Screen:

Modelx Description State Code Flg Serial BDRv PcRv Slotloc

d80300 4GB SCSI Disk Drive onln - --- 126442 - 0018 vsb 0 0 0d80300 4GB SCSI Disk Drive onln - --- 122362 - 0018 vsb 1 0 0d80300 4GB SCSI Disk Drive onln - --- 243442 - 0018 vsb 4 1 0d80300 4GB SCSI Disk Drive onln - --- 138542 - 0018 vsb 4 2 0d80300 4GB SCSI Disk Drive onln - --- 126442 - 0018 vsb 4 3 0d80300 4GB SCSI Disk Drive onln - --- 287442 - 0018 vsb 5 1 0



d80300 4GB SCSI Disk Drive onln - --- 123957 - 0018 vsb 5 2 0d80300 4GB SCSI Disk Drive onln - --- 295742 - 0018 vsb 5 3 0

Examine the information in the Slotloc and State columns of the display. The Statecolumn indicates whether a disk drive is online (onln) or offline (ofln). Refer to Section 2.6.3for a description of how to determine the vsb location.)

3. Determine the device node name of the disk to be removed by issing the following command.

vxdisk list

The vxdisk command lists information on all disk drives in the system. The following sample outputlists disk drives in the system base and in a disk expansion tower or cabinet.

Sample Screen:

The DEVICE column of the vxdisk list output compares to the Slotloc column of

the hwmaint ls disk output. ( If necessary, refer to Section 2.6.6 for a description of howto relate the device node name to the vsb location.)

4. Determine the slot location of the failed disk the using the information provided by the DEVICEcolumn of the vxdisk command output.

5. If you are replacing an online, simplexed disk drive, move the contents of the disk to another diskby entering:

vxassist move volume-name !disk_to_be_removed

6. Enter the following command to invoke the main menu of the VERITAS Volume Manager utility.For an explanation of the VERITAS Volume Manager utility, refer to Section 2.8 or the FTX System'sAdministrator's Guide: General Services (R455X).

vxdiskadm

7. From the main menu, select:



4 Remove a disk for replacement

This menu-driven interface guides you through the process of logically removing a disk drive from thesystem. If you are unsure of how to respond to any prompts, enter the command list to see the diskname or device.

CAUTION: If the disk contents are not mirrored, you will lose data byperforming this procedure. If data will be lost, you will see the followingconsole message: data will be lost.

When the disk has been logically removed, the following message is displayed.

Removal of disk disk_name completedsuccessfully.

8. Quit out of the VERITAS Volume Manager, then enter the following command:

vxdisk list

The following sample output lists some disk drives in a tower or cabinet.

Sample Screen:

Check the STATUS column. The disk will be listed as removed.

9. Delete the disk from its SCSI bus by entering:

hwmaint del /dev/rdsk/device_name

Be sure to include all characters, including the s14, in the device name.

10. To check the status of the disk, enter the following command:

vxdisk list

The disk STATUS will be listed as offline, and the amber light on the disk will flash in3-blink intervals when the disk has spun down.

2.7.8 Verifying Disk Drive Replacement



After the replacement disk drive is installed, the system automatically tests the it and brings it online. When the diskdrive is online, its green light goes off, and the console displays the following message:

Disk drive on-line (disk_VSB)

1. Invoke the VERITAS Volume Manager utility by entering:

vxdiskadm

2. From the main menu, select:

5 Replace a failed or removed disk

The interface prompts you through the disk-adding process.

NOTE: When the interface prompts for the device, do not include the s14.

When the disk has been logically added, the following message is displayed.

Replacement of disk device_name... completedsuccessfully.

3. Verify that the disk drive is online by entering:

vxdisk list

The disk STATUS should be listed as online.

4. Restore any volumes that were disabled by the disk drive failure. Refer to the FTX SystemAdministrator's Guide: Volume Management (R479X) for an explanation on how to restore diskvolumes.

For more information on commands pertaining to disk maintenance, refer to the FTX SystemAdministrator's Guide: General Services (R455X).

2.7.9 Preparing to Replace a Tape Drive

Tape drives are not hot pluggable devices. Perform the following steps to suspend operation on the virtual SCSI bus(VSB) associated with the failed tape drive.

1. Enter the following command to determine the VSB associated with the failed tape drive.

hwmaint ls tape

The screen output lists the status of all configured tape drives on the system.

Sample Screen:



The Modelx and Description columns show the type of tape drive. The first number in the

Slotloc column lists the tape drive's SCSI bus. In the example shown, the offline (ofln) tape drive is

on virtual SCSI bus (vsb) 3.

CAUTION: The next step suspends access to all devices attached to the VSB. Do notcontinue until all active I/O on the bus stops. If the bus is suspended while I/O activity isoccurring, data could be lost.

2. Enter the following command to suspend operation on the target VSB.

hwmaint stop vsb n

Where n is the slot number for the target SCSI bus identified in step 1 (for example, hwmaintstop vsb 3 suspends operation on VSB 3).

3. Enter the hwmaint ls vsb n command to verify that the State column changed from

ofln to susp.

2.7.10 Logically Adding a Tape Drive

When the replacement drive has been installed, perform the following to bring the VSB back online.

1. Activate the VSB by entering the following command:

hwmaint start vsb n

2. Enter the hwmaint ls vsb n command to verify that the State column changed fromsusp to onln.

3. The system automatically adds the new tape device. To verify that the device is configured as part ofthe system, enter the following:

hwmaint ls tape

2.7.11 Preparing to Remove a System Base Power Supply

A system base power supply is hot pluggable (can be removed without entering any commands) if its red LED ison. However, you should verify its fault status by entering the hwmaint ls power command.

The information in the Slotloc, State, and Code columns of the display will show information on thefailed power supply.

2.7.12 Verifying System Base Power Supply Replacement

After the replacement power supply is installed, it is ready and operating properly when the green LED is on. AllPCI cards are automatically brought back online once the power returns. To verify that each card is online, enter thehwmaint ls bd command and check the status of the cards.



2.7.13 Determining/Modifying Virtual SCSI Bus (VSB) Configurations

vsbconf is the command used to change the VSB configuration. It allows the system administrator to add,define, or delete VSBs. It can be run in either in command line or interactive mode.

To determine the defined VSBs on the system, login as root and enter the vsbconf -L command.

The following display is a sample output from this command for a system with a disk expansion tower or cabinet.

Sample Screen:

The Virtual SCSI Buses: section identifies the slot locations of the primary (Ctrl 1) and

secondary (Ctrl 2) controllers for each SCSI bus. In this example, six VSBs (0-5) are defined. VSBs 0, 1, 4,and 5 are dual-initiated (have both primary and secondary controllers) while VSBs 2 and 3 are single-initiated (haveonly a primary controller).

In the SCSI Controllers: section, the controllers described as SCSI Ctlr (W/SE) are theU501 controllers that support the disk drives in the system base and the tape drives. The controllers described asSCSI Ctlr (W/DIFF) are the U502 controllers that support the disk drives in the disk expansiontower/cabinet.



Stratus provides a default VSB configuration on new systems. The configuration is defined in the/etc/default/vsbconf file. The vsbconf utility allows the user to add, modify, or delete aVSB from the system.

CAUTION: Do not add, modify, or delete VSBs unless you are certain the changes are necessary and correct.Incorrectly defining a VSB can leave the system in an inconsistent or unusable state.

To use the utility in interactive mode, enter vsbconf at the prompt.

The following menu is displayed:

Each option (except Exit vsbconf) has one or more submenus for that task. Use the up and down arrowkeys to highlight an option and press the Return key to select that option. Help messages appear at the bottom ofthe screen as you progress through the selection process.

The following describes the selections:

Add a new VSB - use this selection to add a new VSB to the system.

Add a controller to a VSB - use to add a new or additional controller to an existing VSB.

Configure a VSB - make changes to an existing VSB.

Display configuration - display the current active VSB configuration.

Remove a controller - delete a SCSI controller from an existing VSB.

Remove a VSB - delete an existing VSB completely.

Exit vsbconf - exit.

2.7.14 Updating Flash Cards

The flash card contains an image of the /stand file system that can be updated at any time using the

flashadm command.

At system startup, FTX boots from the flash card and assumes it contains the correct version of the /stand file

system. FTX compares the images of /stand on the flash card (from which the system booted) and the rootdisk. If they are not identical, FTX updates the root disk version to match the flash card version.



At system shutdown or reboot, FTX checks to see whether an idbuild has occurred (e.g., to produce a new

kernel) since the last boot. If yes, it automatically updates /stand on the flash card from which the systembooted to match the root disk. The other flash card is never automatically updated.

NOTE: It is possible that a new version of the kernel will not boot. Therefore, it is essential that aknown version of the kernel be kept on backup.

Writing /stand takes some time and the system is not available while this is occurring during the boot or

shutdown process. Therefore, /stand should be updated using the flashadm command whenever itchanges while the system is still in multiuser mode so that system availability is not affected.

To write the current version of the /stand file system to flash card memory, perform the following procedure.

1. Verify the location and state of the flash cards by entering the following command.

hwmaint ls flash

The information in the Slotloc and State fields of the display will show information on the

flash cards. The possible shot locations for a flash card are 2000 and 3000.

These slot locations correspond to device node pathnames for the flash cards. The following is thecorrespondence:

/dev/rflash/c2a0d0 is the device node pathname that corresponds to slot location

2000.

/dev/rflash/c3a0d0 is the device node pathname that corresponds to slot location

3000.

CAUTION: Always have a version of /stand that can boot the system on a backupflash card. If both flash cards are updated with a corrupt version, there is no way to bootthe system.

2. Move the write protect switch on the flash cards to the unprotected (down) position.

3. Enter the following to update the first flash card.

flashadm -d device_node1

where device_node1 is the device node pathname for the first flash card, as determined inStep 1.

4. To update the second flash card, enter the following command.

flashadm -d device_node2

where device_node2 is the device node pathname for the second flash card.

5. When finished, move the write protect switch on the flash cards to the protected (up) position.



2.7.15 Automatic PROM Burning

PROM code versions can be set for the CPU-Memory boards and Console Controllers by adding special entries inthe /etc/confinfo file. When a new board is inserted, FTX checks to see whether its PROM code version

matches the version specified in the /etc/confinfo file. If the versions do not match, FTX burns the new

board as specified by the appropriate entry in /etc/confinfo.

Perform the following to set PROM code versions for the CPU-Memory boards and Console Controllers.

1. Add the following line in the /etc/confinfo file to set a PROM code version for a newlyinstalled Console Controller:

abrninfo: slotloc = any, model = e594, abrnfile= sync;

This entry specifies that FTX "sync" any newly installed Console Controller to the online ConsoleController.

2. Before setting a PROM code version for a newly installed CPU-Memory board, locate the newPROM code file in the /etc/prom_code directory.

CPU-Memory board PROM code file names have the following format: G8XXSfwVV.Vcofwhere G8XX is the board identification number, S is the submodel compatibility number, fw means

firmware, VV.V is the revision number, and cof is the file type.

3. Add the following line in the /etc/confinfo file to set the PROM code version for aCPU-Memory board:

abrninfo: slotloc = any, modelx = G8XX,abrnfile = /etc/prom_code/G8XXSfwVV.Vcof

This entry specifies that FTX check the PROM code version of any newly installed CPU-Memoryboard (G8XX) and burn the version specified onto the board if it is not already the resident version./etc/prom_code/G8XXSfwVV.Vcof is the full pathname of the PROM code file.

The following is a sample CPU-Memory board PROM code file: G8XX0fw46.0cof.

4. To immediately update the system with the new information, enter the following command.

telconf

This utility validates the contents of the /etc/confinfo file and directs FTX to read andrespond to any changes made in that file.



2.8 Veritas Volume Manager (VxVM) UtilityVxVM is a virtual disk management utility that provides features such as disk mirroring, striping, I/O analysis,online administration, and disk management capabilities. It also provides protection against data loss due to diskfailure, enhances system performance by allowing file systems to be spread across multiple disks, and providesperformance tuning tools.

Physical disks, once identified to the system, are placed under control of the volume manager and data contained onthem is backed up using VxMirror. VxMirror is the default management tool which is a subset of VxVM. Thevxmirror utility mirrors the contents of a specified disk or disk group (collection of disks that share acommon configuration) for data redundancy. It provides such features as soft partitioning (concatenation), andmirroring of the root file system.

The volume manager builds virtual devices called volumes on top of the physical disk. Applications and filesystems deal with volumes as virtual disk devices similar to regular disk partitions. A volume's configuration can bechanged without disruptions to the application or file systems that are using the volume. Physical disks are usuallyunder volume manager control. They are accessed by using their device names. A device name is in the formatc#t#d#, where c# is the controller, t# is the SCSI ID, and d# is the disk unit number. Physical disks aredivided into one or more partitions. The volume manager uses virtual partitions, rather than hard partitioning.

A virtual disk can be two physical disk drives (called partners) that operate as a single disk. Virtual disks are sonamed because, to users and system utilities, a virtual disk appears to be a single disk drive.

When a disk is inserted, the system detects the insertion and the device is automatically added to the system if noentry for it is found in the EDT (disk table). If the device exists, validation is performed on the disk model numberand serial number to assure that the same unit has been reinserted. Disks that fail validation will not come to theonline state. The hwmaint add command can be used to force an add of a disk if it is not seen by the system.

Although data is preserved if a disk is pulled during I/O operations, it is potentially disruptive to the system to doso. Data will be preserved on the mirror, but a pulled and re-inserted disk becomes a hot-spare. Because of this, theprescribed procedure is to use either the hwmaint stop or hwmaint delete command beforeremoval. Stopping a disk is viewed as a temporary measure with the expectation that the disk will come back intoservice in the near future. When it is re-inserted, the volume manager will bring it back online as a hot-spare. If adisk is to be removed from the system permanently, it should be deleted.

For detailed information on the Veritas Volume Manager Utility, refer to the FTX System Administrator's Guide:Volume Management (R479X).

2.8.1 Glossary of Terms Used With Volume Manager

2.8.1.1 Disk device address

The disk device address is a system-dependent name that defines the hardware address of a disk. Disk devices forthis operating system are defined by a controller number, a target address number, and a disk number. These are inthe form c<controller>a<target>d<disks>.

For disks attached to a SCSI disk controller, the target number is the target ID defined by switch settings on thedisk, ranging from 0 to 7. Some SCSI devices can attach several disks under a single target ID. For such devices,each disk is assigned a different disk number, ranging from 0 to 7.



2.8.1.2 Disk group

A disk group is a collection of disks that are maintained with a common configuration. Volumes are defined withina disk group, and can be composed only from disks in that disk group. One disk group, named rootdg, is

provided on all systems. Additional disk groups can be created. Disk groups other than rootdg can be movedbetween two systems by disconnecting the disks from the first system and attaching them to the second.

2.8.1.3 Disk name

A disk in a disk group has special identifiers stored on it by the Volume Manager. These identifiers are used toidentify a physical disk independent of its location. A disk can be moved between controllers, or to different targetIDs on a controller. The Volume Manager will recognize that the disk has moved without additional intervention.

Because disks can be moved, the Volume Manager requires a location-independent name, assigned by the systemadministrator, to identify a particular physical disk within a disk group. This name does not change if the disk ismoved, even if the disk group is moved between two systems.

2.8.1.4 Removed disk

A disk can be removed from a disk group temporarily, either with the intention of adding the same disk back to thedisk group or of replacing the disk with a new physical disk. Removal for replacement is a special type of removal.Typically, a disk is removed from a disk group if the disk fails and you want to replace it. If a disk is removed forreplacement, the disk name is retained as a place-holder, which retains the configuration of subdisks on the disk.When the disk is replaced, data for the subdisks is recovered from other sources, where possible.

2.8.1.5 Failed disk

If a disk is physically removed from the system without telling the Volume Manager, or if a cabling failure or othermajor disk failure prevents the Volume Manager from being able to read the special identifiers stored on a disk,then the Volume Manager will not be able to match up a disk name within a disk group to a particular physical disk.If this happens, the disk will not be usable and is considered to be "failed."

Failed disks can be replaced, just like removed disks. However, if the failure condition goes away (for example, ifthe cable is reconnected), then a reboot will handle the fix without additional user intervention.

2.8.1.6 Disk partition

Disk partitioning splits a disk into pieces for different uses. Partitions represent file systems, swapping areas, orareas for use by applications that use raw disk directly, such as some database products that do not use file systems.Typically, the system must be rebooted to recognize changes made to the list of partitions on a disk.

The Volume Manager does not use disk partitioning to split a disk for different uses. Instead, the Volume Manageruses configuration objects called subdisks, which are a more general concept that can be created and removed on arunning system. The Volume Manager slices a UNIX partition into internally-defined subdisks, which itdynamically creates and deletes based on user Volume Manager commands.

2.8.2 vxdiskadm Utility



Many disk management tasks can be performed interactively using the vxdiskadm utility. This section gives a

brief description of the functions performed by vxdiskadm.

To access the utility, enter vxdiskadm at the prompt. The following screen appears.

The list selection displays a list of all disks on the system. It also lists removed or failed disks. Removed and faileddisks are disconnected from specific devices.

You can also choose to list detailed information about the disk at a specific disk device address. This detailincludes, among other things, the disk group of which the disk is a member, even if that disk group is currentlydisabled.

Selecting list from the menu displays a screen similar to the following:

Sample Screen:



When you enter the device address of a device (e.g., c0a1d1), a screen similar to the following appears.

Sample Screen:



The list selection can also be accessed within any of the other main menu selections. The following subsectionsdescribe the other selections on the vxdiskadm main menu.

2.8.2.1 Add or initialize one or more disks

This menu operation is used to add or initialize formatted disks. SCSI disks are already formatted, so nopre-formatting is needed.

You are prompted for the disk device address (c#t#d#). The operation then prompts for a disk group name anddisk name for the disks. A disk group name of `none' may be specified which will cause the disks to be left asunassigned replacement disks. If a new disk group name is specified, the disk group will be created and the disksadded to the new disk group. By default, disks are added to the root disk group, rootdg. If a disk group isspecified for the disks, you may specify that the disks be designated as hot-spares for the disk group. Hot-sparingmeans that if a disk in the disk group fails, the disk marked as a hot spare will be brought online as an automaticreplacement.

If the disk being added has not been initialized already, the disk is partitioned and initialized for use with theVolume Manager. The disk will be labeled and the volume table of contents (VTOC) created.

If a disk is found to already contain non-Volume Manager partitioning, you are asked whether the disk should beencapsulated. Encapsulation will turn each partition of the disk into a volume. (See Section 2.8.2.2.) A disk should



be encapsulated if it contains file systems or data that should not be overwritten. If encapsulation is not desired for adisk, the disk can be initialized as a new disk for use by the Volume Manager. For new disks, all space on the diskbecomes free space in the disk's disk group. The menu operation interactively allows encapsulation or initializationto be decided for all the disks specified.

2.8.2.2 Encapsulate one or more disks

This menu operation is used to place under Volume Manager control one or more disks that were added to thesystem before installing the Volume Manager.

You are prompted for the disk device address (c#t#d#). The operation then prompts for a disk group name anddisk name for the disks. The disk will be added to the given disk group with the specified disk name. The disk isthen examined in search of partitions that are used for file systems, or for other purposes. Volumes will be createdto replace disk partitions as the means of accessing the data. If the encapsule operation cannot determine thepurpose for a partition automatically, you will be asked what to do with it. You can choose to replace the partitionwith a volume, keep the partition as it is, or remove the partition.

You will be required to reboot the system if any partitions are in use for mounted file systems or for runningapplications. You may have to modify application configuration files to use volumes (rather than direct diskdevices) to access the disk partitions. File system mount information in /etc/vfstab will be adjustedautomatically.

2.8.2.3 Remove a disk

This menu operation is used to remove a disk from a disk group. You are prompted for the name of a disk toremove. You cannot remove a disk if any volumes use storage on the disk. If any volumes are using storage on thedisk, you have the option of asking the Volume Manager to move that storage to other disks in the disk group.

NOTE: You cannot remove the last disk in a disk group using this operation. If you wish to use all theremaining disks in a disk group for some purpose, you should disable (deport) the disk group. You willthen be able to reuse the disks.

2.8.2.4 Remove a disk for replacement

Use this menu operation to remove a physical disk from a disk group, while retaining the disk name. This changesthe state for the disk name to a "removed" disk. If there are any initialized disks that are not part of a disk group,you will be given the option of using one of these disks as a replacement.

2.8.2.5 Replace a failed or removed disk

Use this menu operation to specify a replacement disk for a disk that you removed with the Remove a disk forreplacement menu entry, or that failed during use. You are prompted for a disk name to replace and a disk deviceto use as a replacement.

You can choose an uninitialized disk (in which case the disk will be initialized) or you can choose a disk that youhave already initialized using the Add or initialize a disk menu operation.

2.8.2.6 Mirror volumes on a disk



This menu option prompts for a disk, by media name. It then prompts for a destination disk within the same diskgroup, also by media name. Specifying no destination disks indicates that any disk is suitable. The operationproceeds by calling vxmirror to mirror the volumes.

Mirroring volumes from the boot disk produces a disk that can be used as an alternate boot disk. This is done bycalling the vxrootmir command.

2.8.2.7 Move volumes from a disk

Use this menu operation to move any volumes that are using a disk onto other disks. Use this menu immediatelyprior to removing a disk, either permanently or for replacement.

NOTE: Simply moving volumes off a disk, without also removing the disk, does not prevent volumesfrom being moved onto the disk by future operations. For example, using two consecutive moveoperations may move volumes from the second disk to the first.

2.8.2.8 Enable access to (import) a disk group

This menu operation is used to enable access by this system to a disk group. If you wish to move a disk group fromone system to another, you must first disable (deport) it, on the original system. Then, move the disk betweensystems and enable (import) the disk group. You will be prompted for the disk device of a disk containing the diskgroup. All enabled (online) disks will be searched to locate disks that are in the same disk group as the givendevice.

2.8.2.9 Remove access to (deport) a disk group

Use this menu operation to disable access to a disk group that is currently enabled (imported) by this system.Deport a disk group if you intend to move the disks in a disk group to another system. Also, deport a disk group ifyou want to use all of the disks remaining in a disk group for some new purpose.

You are prompted for the name of a disk group. You are also asked if the disks should be disabled (offlined). Forremovable disk devices on some systems, it is important to disable all access to the disk before removing the disk.

2.8.2.10 Enable (online) a disk device

If you move a disk from one system to another during normal system operation, the Volume Manager will notrecognize the disk automatically. Use this menu operation to tell the Volume Manager to scan the disk to find whatdisk it is, and to determine if this disk is part of a disk group. Also, use this operation to re-enable access to a diskthat was disabled by either the disk group deport operation or the disk device disable (offline) operation.

2.8.2.11 Disable (offline) a disk device

Use this menu operation to disable all access to a disk device through the Volume Manager. This operation can beapplied only to disks that are not currently in a disk group. Use this operation if you intend to remove a disk from asystem without rebooting.

NOTE: Many systems do not support disks that can be removed from a system during normaloperation. On such systems, the offline operation is seldom useful.

2.8.2.12 Mark a disk as a hot-spare for a disk group

This option sets up a disk to be used as a hot-spare device for its disk group. A hot-spare disk can be used to



automatically replace a disk that has failed. No space can be used on a disk that is marked as a hot-spare.

2.8.2.13 Turn off the hot-spare flag for a disk

This option removes a disk from those that can be used as a hot-spare and returns its space to the general pool ofavailable space.



3. Theory of OperationThis section contains an overview of the theory of operation for the Continuum 400 Series systems. Itprovides information on how the system operates and includes a description of each of the followingmajor assemblies/subsystems.

System bus●

Suitcase●

System base●

Disk expansion tower/cabinet●

Power subsystem●

Cooling subsystem●

Figure 4 is a high-level architectural view of the Continuum 400 Series system.

Figure 4. Continuum 400 Series System Block Diagram

Therory of Operation


3.1 System BusThe Continuum 400 Series system bus (XBus) and its interfaces perform the following major functions:

Transfer information between CPU-Memory boards and PCI Bridge (PCIB) cards❍

Transfer information between two CPU-Memory boards❍

Provide fault detection and isolation for the bus and the boards on the bus❍



The XBus is actually four distributed point-to-point buses that are not duplicated. Each board in thesystem connects to two of the buses. This is a major difference from the Continuum Series 600/1200system bus which is a single logical split transaction multiplexed address/data bus that is duplicated andshared by all boards.

Although the physical implementation of the XBus uses four point-to-point buses, the protocol isimplemented with a single bus view. On any given cycle, there is no more than one transaction on theinterconnect. This transaction may be on all four buses in the case of a duplexed operation or on only twoof the buses for a simplexed operation.

The XBus features a 32-bit bus interface width, completely synchronous operation, cache consistencysupport, and a single logical Application Specific Integrated Circuit (ASIC) interface. Deviceconfigurations are limited to two CPU-Memory boards and two PCI bridge cards.

Bus signals are divided into two categories, transaction-based data signals and control signals. As shownin Figure 5, the data signals are implemented as four bidirectional point-to-point buses and the controlsignals are separate unidirectional buses. The bus signals are protected through map ram, parity, andloopback checking. The control signals are protected through ECC and loopback checking.

Figure 5 shows the XBus point-to-point bus interconnections in the Continuum 400 Series system.

Figure 5. Continuum 400 Series System Bus Interconnections

A CPU-to-CPU transaction is accomplished by means of a peer-to-peer operation in which thetransaction is broken into two separate transactions since the XBus does not have fully interconnecteddata buses. In the first transaction, the CPU transmits a send to the PCIB. In the second, the PCIBtransmits information to the CPU.

The XBus physical implementation makes use of a direct ASIC-to-ASIC connection technology. The businterface is accomplished through the use of two 391-pin PGA package ASICs. The XBus uses two



different styles of connectors. The CPU-Memory boards in the suitcase use the SL-100 AMP connector.The PCIBs in the system base use the multiple sourced futurebus connector. A single 24 MHz clock isused for clocking data and control onto and off the bus.

A new feature of the XBus is the error protocol. Unlike the Series 600/1200 system bus, the XBusattempts not only to detect bus errors, but also to diagnose the source.

3.2 SuitcaseOne of the two major assemblies in the Continuum 400 Series system is the suitcase. There are twoidentical suitcases mounted on the system base, each with a positive latching mechanism that aligns andlocks it to the base to ensure proper connector mating. The suitcase-to-base connector is a 300-pin AMPSL100 connector with integral grounding. A carrying handle is integrated into the top surface of thesuitcase.

The operating status and fault conditions of the suitcase are displayed on LEDs located on the front bezelof each suitcase.

The suitcase houses the following components/subassemblies:

CPU-Memory board❍

cooling fans❍

power supply❍

3.2.1 CPU-Memory Board

The CPU-Memory board is actually a motherboard containing the logic section (Hewlett PackardPA/RISC 7100 CPU/cache modules and memory modules), and the Console Controller module (withserial connections for the console, RSN, and UPS). The Console Controller serial connections are cabledto external connectors on the rear of the system base.

3.2.1.1 Logic Section

The PA7100 chip is a highly integrated CPU that provides support for integer, floating point, andmemory management functions all on the same integrated circuit. The chip supports a large externalcache with separate RAMs for instructions and data.

The CPU/cache modules are available in uni or twin processor designs running at 96 MHz. The uniprocessor contains one logical/two physical CPUs and the twin processor has two logical/four physicalCPUs. Both the uni and twin designs are available in a low cache 512 KB version (256 KB instructioncache and 256 KB data cache) or high cache 2 MB version (1 MB instruction cache and 1 MB datacache).

A memory module contains 128 MB or 512 MB of memory. Memory configurations range from one tofour memory modules, for a maximum of 512 MB (using 128-MB modules) or 2 GB (using 512-MBmodules).

As in earlier Stratus designs, the CPU-Memory board contains a C-side and D-side which are compared



side-to-side to detect on-board errors. A partner board runs in lockstep and a detected failure on eitherboard causes the board to go offline. The CPU subsystem and system bus interface are fully duplicatedand compared.

The CPU-Memory board design uses a small number of high density CMOS ASICs which significantlyreduces the number of components. The board is designed around a single motherboard with a range ofCPU modules and DRAM configurations.

The Interface ASIC on the CPU-Memory board connects the system bus (XBus) with the localpoint-to-point bus (IBus) going to the CPU-Memory Control ASIC on the CPU module. In addition toserving as a bus interface, the Interface ASIC provides fault detection and isolation for the memoryboard. The Interface ASIC gate array provides the CPU interface and memory control.

The common logic is made up of the following sub-blocks.

XBus IN block (XBI)❍

XBus OUT block (XBO)❍

arbitration logic❍

board states logic❍

error checking logic❍

common register control❍

phase control❍

ping logic❍

reset logic❍

The IBus IN unit takes data from the XBus (XBI) to the IBus. The IBus OUT unit takes data from theIBus to the XBus (XBO). The register unit contains all of the registers that are specific to the InterfaceASIC.

Figure 6 is a block diagram of the CPU-Memory board and its interfaces.

Figure 6. CPU-Memory Block Diagram



An ID PROM on the CPU-Memory board contains the following top level data: part number, revisionlevel, and serial number. This information is set according to the CPU/memory configuration, and acombination of the CPU-Memory board, CPU/cache module, and memory module revision levels. TheCPU-Memory board serial number never changes, regardless of part number or revision changes.

3.2.1.2 Console Controller Module

Each Console Controller module operates independently of the rest of the CPU-Memory board on whichit is located. It functions as a centralized controller for the entire system performing the following criticalfunctions:

supports three async ports located in the rear of the suitcase: a system console port, a RSNport, and an auxiliary port for a UPS connection or console printer

❍

serves as a central collection point for Maintenance and Diagnostics (M&D) services❍

controls and monitors the main power supply for the system❍

provides a console command (front panel) interface❍

contains the hardware calendar/clock and NVRAM used to store data across boot loads❍

contains an ID PROM which stores information such as modelx, serial number, etc.❍

In addition to the above, the Console Controller includes burnable PROM partitions that contain code forthe following: board-level diagnostics, online and standby board operations. The diagnostics and boardoperations code (both online and standby) are burned onto the board at the factory. To update this code, anew firmware file can be burned into these partitions.

The Console Controller also contains a burnable PROM data partition that stores console port



configuration information (bits per character, baud rate, stop bits, and parity) and certain system responsesettings. The defaults can be reset by entering the appropriate information and reburning the partition.

The Console Controller also contains a burnable PROM data partition that stores information on wherethe system should look for a bootable device when it attempts to boot automatically. (However, the init6, shutdown -i6, and reboot -a commands do not use the Console Controller; they take informationstored in the kernel to find the bootable device.)

The Console Controllers are logically paired (not fully duplexed) so that one board is online while theother is on standby status. The online Console Controller is active on the Console Controller bus andcommunicates with other components in the system. The standby Console Controller is active on theConsole Controller bus, but cannot communicate with the rest of the system; it is electrically isolatedfrom all external devices except the Console Controller bus. In the event of a hardware failure of theonline Console Controller, the hardware automatically performs a "failover" operation to the standbyConsole Controller. All ports on the new online Console Controller are initialized and I/O operations canresume. Failover can be manually initiated by breaking or deleting the online Console Controller.

3.2.2 Cooling Fans

The suitcase is cooled using three tube-axial fans located in the rear. The fans evacuate air from thesuitcase by creating negative pressure within the enclosure which draws fresh air in the front anddischarges heated air out the rear. A suitcase fan failure results in a fault condition bringing down thatparticular suitcase.

The PCI cards in the system base are cooled by the exhaust air from the suitcase cooling fans.

Temperature information is acquired by thermal sensors in the suitcase and is used to drive the fan speed.There is no fan speed (high/low) control accessible to the user in 400 Series systems.

For a more detailed description of the cooling system, refer to Section 3.6.

3.2.3 Power Supply

The suitcase power supply is a non-fault tolerant supply that takes in AC and produces power directlyusable by the logic boards and cooling fans. If the power supply fails, the suitcase goes down.

The power for each suitcase is provided by a single power supply. However, the CPU/memory/logicfunction is architected as a 2N redundancy system, so the overall function, including power, is 2Nredundancy.

For a more detailed description of the suitcase power supply, refer to Section 3.5.2.

3.3 System BaseThe following are the major components in the system base:

PCI I/O subsystem❍

disk drives (8 max.)❍



system base power supplies (2)❍

At the rear of the system base are four ports labeled Cabinet Bus, RSN, Console, and RemoteConsole/UPS. The Cabinet Bus port is an RS-485 port and is reserved for future use. The other threeports are RS-232 ports.

3.3.1 PCI I/O Subsystem

The major components in the PCI I/O subsystem are the PCI bus, two PCI bridge (PCIB) cards, and up to14 PCI adapter cards. The PCIBs act as the interface between the system bus (XBus) and the PCI buses.They isolate the PCI cards from the suitcase, so that a suitcase failure does not adversely affect PCI cardavailability.

3.3.1.1 PCI Bus

The PCI bus is an industry standard bus developed by Intel. It is a fully synchronous 32-bit multiplexedaddress/data bus running at 24 Mhz. The bus is a high bandwidth, low latency local bus operating at 5VDC. It supports PCI adapter cards that are 12.8" x 4.2" in size.

The system base contains four logical PCI buses. Each bus has four PCI slots, with the first slot in two ofthe buses reserved for the PCI bridge (PCIB) card and its onboard PCMCIA flash card. This results in atotal of 14 slots for the PCI adapter cards. The I/O backplane provides signal, power and groundinterconnection between the two logic suitcases and the two I/O sections of the system.

Each PCI card is accessible directly from either suitcase via the PCI bus. Logic in the suitcase providesfault tolerance over the PCI bus and protects against failures of the cards.

The PCI buses are powered separately so that at least one of the disk PCI cards will always have power.When a PCI enclosure access door is opened, its PCI bus is automatically powered down. (There is aninterlock that allows only one half of the bus to have its door open at a time.) When the access door isreclosed, the card cage comes back online and a full PCI configuration sequence is initiated.

Online fault detection and isolation includes identifying failing PCI adapters, identifying properlyfunctioning PCI adapters when a failure is indicated, and, when possible, determining cable failures ordisconnected cables. Online fault detection tools consist of internal/external loopback tests andoff-the-shelf network tools.

Each PCI card is in one of three possible states: broken, offline, or online simplex. The state is shown bya row of LEDs on the card cage above each card.

A broken board is basically isolated from the system. It will respond to writes, but will not respond toreads. A broken board usually indicates a hardware failure which requires either a reset or repair.

An offline board is a good board which did not transition to the online state. It will respond to all writesand will respond to board I/O reads. However, in the offline state, a board will not respond to memoryreads. Typically, the offline state is a transitional state, used by the board that is still initializing orrunning diagnostics.

An online simplex board is a fully functional running board, but not running in lockstep with a partner



board. In this state, a board will respond to all operations of any kind, including read access.

3.3.1.2 PCI Bridge Card

The PCI bridge (PCIB) card is the interface and control unit for the PCI cards. The system consists oftwo PCIBs, each of which acts as the interface between the CPU and one logical PCI bus. The systembus (XBus) provides the interconnection between the CPU-Memory boards and the PCIB. The PCIBscontain two ASICs, which each talk to a PCI bus and a FLASH RAM for OS boot.

The PCIB boards do not run in lockstep mode and cannot be duplexed. Their normal operational state isonline simplex. However, if all disk drives are duplicated on different controllers and comm is duplicatedvia software in both PCI card cages as needed, opening a PCI card cage assess door on a running systemshould cause no problems. (See Section 3.3.2.)

Flash card

A 20 MB removable PCMCIA flash card resides on each PCIB card. The function of the flash card is toboot the system. The CPU PROM makes the flash card look like a read-only disk during the bootprocess. In order to boot the kernel with only 20 MB, file compression is employed to shrink the size ofthe kernel.

One major difference between the Continuum 400 Series FTX 3.1 software and earlier releases is that theroot file system is on a disk and the boot device is the PCMCIA flash card. The device on which the rootfile system resides is now specified by an environment variable contained in the /stand/conffile. The flash card is also used for cold boot situations when no initialized FTX root file system exists ondisk. This task was previously handled by booting from a tape that contained a ramdisk image of the rootfile system. For cold boot support on Continuum 400 Series systems, the BFS file system located on theflash card contains a ramdisk image as a file.

At system startup, FTX boots from the flash card and assumes this is the correct version of /stand.FTX compares the images of /stand on the flash card (from which the system booted) and the rootdisk. If they are not identical, FTX updates the root disk version to match the flash card version. Atsystem shutdown or reboot (system states 0, 5, or 6), FTX checks to see whether an idbuild has

occurred since the last boot. If yes, it updates /stand on the flash card (from which the systembooted) to match the root disk version.

The flash card is laid out in a similar fashion to a standard FTX disk. The FTX disk layout contains arecognizable physical disk label (i.e., the "Stratus" label) at block 0, which contains pointers and extentinfo to other important areas of the disk image (e.g., the bootloader, the FTX label, etc.). In order toascertain data integrity, the flash card has a software imposed CRC algorithm. The flash card itself isdivided up into 5120 4-KB blocks. Each block has a 32-bit CRC character stored for it in an arrayoccupying the last five 4-KB blocks of the flash card. In order to support updating the flash card, whichcan only be done in 128-KB (32 4-KB blocks) chunks, the label and the CRC "partitions" are reserved in128-KB module chunks.

Figure 7 is a pictorial diagram of the FTX PCMCIA flash card layout.



Figure 7. FTX PCMCIA Flash Card Layout

3.3.1.3 PCI Adapter Cards

The PCI adapter cards are designed for +5 VDC and a 32-bit bus. The system supports a maximum of 14PCI cards. Each PCI slot in the card cage contains three LEDs which indicate the state of the board.

The following PCI cards are supported in Continuum 400 Series systems.

U401 Programmable Asynchronous/Synchronous (PAS) Adapter (IBM Artic960)

The U401 PAS adapter is based on the Intel i960 RISC coprocessor with 4 KB cache, 1 KB SRAM, anda clock rate of 25 MHz. It can operate at data rates up to 96 Mbps on Continuum 400 Series systems.

The U401 contains four ports and supports the following protocols:

EIA-232-D 38.4 Kbps❍

EIA-530 2.048 Mbps (max. 3 ports)❍

ISO-4902 (V.36) 2.048 Mbps❍



ISO-4903 (X.25) 2.048 Mbps❍

The U401 uses a Signetics SC26C562 dual universal serial communications controller (DUSCC).

Other features include:

32-bit PCI transfers❍

internal buffering provides PCI data rate to 132 Mbps❍

IBM Brighton memory controller allows DMA by host❍

memory controller also provides memory protection❍

packet DRAM (for data storage)❍

instruction DRAM (for instruction storage)❍

128 KB ROM supports POST code for coprocessor and application interface board (AIB)❍

U501 Single-ended SCSI Adapter (DPT PM3224W)

The U501 single-ended SCSI adapter is a fast wide 16-bit SCSI host adapter and RAID (RedundantArray of Inexpensive Drives) controller. It contains three physical SCSI ports, two with internalconnectors, and one external. The internal connectors are used for driving the integrated disk drives. Theexternal connector is for connecting an optional tape drive.

The U501 contains a MC68EC030 32-bit processor which operates at 40 MHz. The SCSI peripheralinterface is implemented using a FAS366 SCSI controller chip. All SCSI bus drivers are built-in andconform to the SCSI-2 specifications.

The U501 supports basic disk and tape I/O via single and dual SCSI interfaces. Dual-initiators providetwo paths to access SCSI devices to assure availability.

U502 Differential SCSI Adapter

The U502 PCI card is a high-performance fast/wide 16-bit differential SCSI adapter that connectsexternal D81X disk expansion towers and cabinets to the PCI bus in a Continuum 400 series system. Itcontains three external ports marked A (port 0), B (port 1), and C (port 2).

The U502 is based on a 32-bit bus architecture that optimizes data flow between the system and the diskdrives, resulting in the highest possible throughput with low CPU utilization. It contains 8 MB cache. Aflash PROM is used to store program code. The U502's synchronous SCSI data transfer rate is 10 MB/sin Fast 8-bit mode and 20 MB/s in Fast/Wide (16-bit) mode.

The U502 provides termination for single-initiator configurations. It does not provide termination fordual-iniator configurations.

U510 10/100MB Ethernet Adapter (Intel EtherExpress PRO/100)

The U510 PCI card is a high performance Ethernet LAN adapter manufactured by Intel that optimizesdata flow between the system and the network, resulting in the highest possible throughput with lowCPU utilization.

The U510 provides support for multiple network protocols (TCP/IP, OSI, and Netware), FTX RNI, andStratus M&D services.



U510 features include:

32-bit PCI direct bus mastering❍

uses shared memory structure in host memory and copies data directly to/from host memory❍

10 Mbps regular Ethernet based on IEEE 802.3❍

100 Mbps fast Ethernet based on IEEE 802.3u❍

external interface supports 100Base-TX (RJ-45 connection)❍

automatically senses speed of the port to which it is attached and configures itself properly❍

network wiring for 100Base-TX / unshielded twisted pair (UTP) Category 5, data grade wire❍

network wiring for 10Base-T / UTP Category 3, 4, or 5 data grade wire❍

U530 FDDI Adapter (Rockwell Network System's RNS 2200)

The U530 PCI card provides a high performance communications interface between a Fiber DistributedData Interface (FDDI) Local Area Network (LAN) and the Continuum 400 Series system.

The U530 is based on the FasTRACC Advanced Communication Controller, which optimizes data flowbetween the system and the network, resulting in the highest possible throughput with low CPUutilization.

Other features include:

supports both single attached station (SAS) and dual attached station (DAS) configurations❍

provides programmable control of an external Optical Bypass Relay (OBR) for DASconfigurations

❍

supports transmission speeds of 100 Mbps and both asynchronous and synchronous FDDItransmission

❍

features two ports (A and B) to support dual attached configurations❍

maximum fiber lengths are 15 km for single mode and 2 km for multi mode❍

U540 ATM Adapter (Interphase 5515)

The Interphase 5515 PCI adapter is a non-intelligent single port ATM (Asynchronous Transfer Mode)controller. It has the following features:

transfer rate of 155 Mbps (SONET OC-3c)❍

32-bit, zero wait-state, up to 64-byte burst PCI DMA master transfers❍

1 MB standard buffer packet memory❍

full support of AAL 5 (ATM Adaptation Layer 5) to insure consistent high-speed datacommunications with all other industry standard ATM networks

❍

supports UNI 3.0 (User to Network Interface) ATM signaling for generalized switchcompatibility

❍

compliant with ATM Forum standards for layer signaling❍

supports LAN Emulation 1.0 for compatibility with existing networks❍

both Permanent Virtual Circuit (PVC) and Switched Virtual Circuit (SVC) software drivers❍



PCI Local Bus 2.0 compliant, transmitting at 132 Mbps❍

hardware based Segmentation and Reassembly (SAR) functions❍

3.3.2 Disk/Tape Drives

The disk drives reside in a disk enclosure in the system base directly below the logic suitcases (and inoptional expansion towers and expansion cabinets which are described later). The drives are standard3.5" hard drives housed in a commodity off-the-shelf enclosure. The maximum number of disk drivesthat can be housed in the disk enclosure is eight. The disks are cooled by the fans in the system basepower supplies located behind the disks.

The disk drives in the system base are controlled by the U501 PCI SCSI controllers configured asdual-initiators. Virtual SCSI buses (VSBs) provide the ability to have dual-initiated SCSI buses acrossany SCSI controller in the system. The VSB is a mapping scheme that describes the logical location of aphysical SCSI bus. This virtual bus consists of one or two controller ports that are connected to acommon physical SCSI bus. The SCSI bus manager (SBM) decides which controller in the pair is themaster and routes I/O requests appropriately. Devices on the bus are accessed through the VSB slotlocation.

SCSI buses (SB) 1 and 2 coming from the U501 PCI SCSI controllers are connected to the disk enclosureand drive virtual SCSI buses (VSB) 0 and 1 as shown in Figure 8. The SCSI buses are connected toalternate slots in the disk enclosure to allow a simple pairing of mirrored disks that are being driven fromseparate buses. This scheme allows a single SCSI controller to reach all drives on both the SCSI buses inthe event of a failed SCSI controller. Tape drives are connected to single-initiated SCSI buses (VSB 2and 3).

Figure 8. SCSI Buses for Disk/Tape Drives in System Base



Each disk drive has a SCSI ID of 0, 1, 2, or 3 which is determined by the location of the disk drive withinthe disk enclosure. Tape drive SCSI IDs are configured by the user and must be unique in order to avoidbreaking the bus. A maximum of four tapes are supported.

Disk drives are hot-pluggable and therefore can be removed and replaced without powering down anypart of the system.

3.3.3 System Base Power Supplies

There are two physical power supplies located in the system base. They supply redundant power to thedisk drives and simplexed power to their respective PCI cards located in the system base.

The power in the system base has a dual power architecture. The power to the disks and the non-faulttolerant clock is configured as full 2N fault-tolerant. This provides the same level of hardwareavailability to the disks that is provided for the CPU/memory/logic in the suitcases. The power for thePCI I/O cards, however, is a simplexed (non fault-tolerant) architecture.

Each system base power supply contains fault-tolerant outputs for the system clock and disks, and nonfault-tolerant outputs for the PCI card cages. In the case of a system base power supply failure, the PCIcard cage associated with the failure goes down, but the disks and system clock remain online.

The power supplies are cooled using fans internal to the supply drawing air through the disk drives, intothe front of the power supply and exhausting the air out the rear of the system. Power supply fan speed iscontrolled by the operating system. The fan speed is set to HI if any of several environmental alarmconditions are reported.

Refer to Section 3.5.3 for more information on the system base power supplies.

3.4 Disk Expansion Tower/CabinetSince the system base holds only eight disk drives, any additional drives are housed in standalone diskexpansion towers or expansion cabinets which contain all of the power, cooling and support structure forthe disks. Disk drives are interchangeable between the system base and the expansion towers/cabinets.

3.4.1 D810 Disk Expansion Tower

The D810 disk expansion tower consists of a vertical free-standing modular storage cabinet with anintegral SCSI bus backplane and 9-slot chassis (called a disk shelf). The backplane distributes disk power(+5V, +12V) and fan power (+12V), and routes the SCSI bus to each disk slot.

The disk shelf contains the following components:

· up to five disk drives

· personality unit

· SCSI conversion unit

· two power supplies



· AC distribution unit

· two dual speed fans

The personality unit provides the connectors necessary to connect external SCSI buses to the disk shelf.Currently, the personality unit is available only in the single-ended SCSI bus version which contains twoexternal SCSI connectors that interface to the internal single-ended SCSI bus. Other functions of thepersonality unit include daisy-chaining of the disk shelves, providing device address assignments (SCSIIDs), cooling fan speed control, fan fault indicators, over-temperature indication, and active, automaticSE SCSI bus termination. The personality unit contains a DIP switch used to set SCSI bus addresses. Theswitches must be set correctly in order to prevent SCSI bus contention due to redundant addressing. (SeeSection 1.7.1.)

The SCSI conversion unit converts the single-ended SCSI bus to differential SCSI (and vice-versa) andhas built in terminators for both buses. It supports 16-bit WIDE bus transfer rates up to 20 Mbps andextends the SCSI bus length up to 12meters. The SCSI conversion unit occupies a slot in the disk shelfwithout consuming a SCSI ID on either SCSI bus.

To provide fault tolerance, the disk expansion tower contains two power supplies, each of which is anauto-ranging 150-watt AC/DC power supply. Input range for the power supply is 90-264 VAC, 47/63Hz. The output power is +5 VDC @ 15 A, +12 VDC @ 9 A. The power supplies have over-voltage,over-current, and over-temperature protection.

Input AC power is routed through the AC distribution unit to the power supplies. The AC distributionunit contains one universal AC input connector (100-240 VACVDC, 50-60 Hz), three AC output powercords (two 2.0A in front, one 4.0A in rear), and a power ON/OFF switch. The two AC output powercords located in the front plug into the two power supplies. The output power cord in the rear can be usedto provide AC power to another expansion tower.

The two dual speed fans located in the rear cool the expansion tower by moving air from the front of theunit through the disk shelf and out the rear of the tower. Air separation panels are used to properly routeheated air through the tower. These panels are small foam baffles that mount at the top and bottom of thetower to prevent heated air from being drawn back into the disk shelves. The fans can operate at low orhigh speed. A failed fan is reported by the Shelf_OK signal.

The system uses two disk expansion towers to mirror the disk drives for fault tolerance. These can be ineither a single-initiator or dual-initiator cable configuration.

In a single-initiator configutration, the disk drives in the disk expansion towers are controlled by theU502 PCI SCSI controllers configured as single-initiators. SCSI buses coming from the U502 PCI SCSIcontrollers are connected to the disk expansion towers and drive virtual SCSI buses VSB 5 and VSB 8 asshown in Figure 9. Port 1 on the U502 in slot 3 in card cage 2 drives VSB 5, and Port 1 on the U502 inslot 3 in card cage 3 drives VSB 8.

NOTE: Port 0 on the U502 card cannot be used in a single-initiator configuration.

Figure 9. Single-initiator Configuration (Disk Expansion Towers)



In a dual-initiator configutration, the disk drives in the disk expansion towers are controlled by the U502PCI SCSI controllers configured as dual-initiators. SCSI buses coming from the U502 PCI SCSIcontrollers are connected to the disk expansion towers and drive VSB 4 and VSB 5 as shown in Figure10. Port 0 on each U502 drives VSB 4, and the port 1 drives VSB 5.

Figure 10. Dual-initiator Configuration (Disk Expansion Towers)



3.4.2 D811/2 Disk Expansion Cabinet

The Continuum 400 Series system can contain one or two D811/2 disk expansion cabinets, each of whichcontains the following components:

· two to eight disk shelves (adjacent shelves are paired)

· two cable distribution units (CDUs)

· two fan assemblies

Each shelf in the cabinet contains the same components as a disk expansion tower disk shelf, except forthe second shelf in a pair, which does not contain a SCSI conversion unit. Therefore, each of thesesecond shelves can contain up to six disk drives.

To provide fault tolerance, the disk expansion cabinet contains two CDUs. The CDUs provide theconnections necessary to distribute AC power to the disk shelves in the cabinet. They also provide surgeand spike protection, along with a circuit breaker to control the incoming AC power.

One CDU supplies AC power to the first power supply in each of the disk shelves. The other CDUsupplies AC power to the second power supply in each shelf. If one of the CDUs or its power supplyfails, the shelf remains powered by the other power supply. This configuration provides complete power



redundancy to the disk shelves' devices.

The disk drives in the disk expansion cabinet(s) are controlled by the U502 PCI SCSI controllersconfigured as single or dual-initiators. In a single-initiator configutration, the disk drives in the diskexpansion cabinet(s) are controlled by the U502 PCI SCSI controllers configured as single-initiators.SCSI buses coming from the U502 PCI SCSI controllers are connected to the disk expansion cabinet anddrive VSB 5 and VSB 8 as shown in Figure 11. Port 1 on the U502 in slot 3 in card cage 2 drives VSB 5,and Port 1 on the U502 in slot 3 in card cage 3 drives VSB 8. VSB 6 and VSB 9 would be driven off Port2 on the U502s, respectively, and would connect to the next two pairs of shelves (5/6 and 7/8) in asimilar manner.

NOTE: Port 0 on the U502 card cannot be used in a single-initiator configuration.

Figure 11. Single-iniator Configuration (Disk Expansion Cabinet)



In a dual-initiator configutration, the disk drives in the disk expansion cabinets are controlled by theU502 PCI SCSI controllers configured as dual-initiators. SCSI buses coming from the U502 PCI SCSIcontrollers are connected to the disk expansion cabinet and drive VSB 4 and VSB 5 as shown in Figure12. Port 0 on each U502 drives VSB 4, and Port 1 drives VSB 5. The next pair of shelves (5/6) wouldconnect to port 2 which would drive VSB 6. Shelves 7/8 would require another pair of U502s whichwould be installed in slot 6 in each card cage and would drive VSB 7. VSB 8 and VSB 9 would also bedriven from this pair of cards.

Figure 12. Dual-initiator Configuration (Disk Expansion Cabinet)



3.5 Power SubsystemThe power subsystem is composed of three main assemblies: dual AC input modules, the suitcase powersupplies, and the system base power supplies.

A fully configured Continuum 400 Series system consumes approximately 2400 watts of power duringpeak normal operation. The system operates from a standard office power source of 85 VAC to 264 VACrunning at 50 to 60 Hz.

Each suitcase contains a power supply that takes universal AC input voltages and supplies logic levelvoltages to the suitcase. A failure in a suitcase power supply results in a suitcase failure, but not a systeminterruption.

There are two I/O power supplies in the system base. Each power unit supplies wired-OR'ed power to thedisks and non fault tolerant power to half of the I/O subsystem. In the case of a power supply failure, halfof the PCI slots fail, but the system remains up.

The Continuum 400 Series does not support any built in powerfail capabilities. A Stratus-qualifiedexternal uninterruptable power supply (UPS) manufactured by APC (American Power Conversion) isavailable for customers who do not already have a site-wide system. The UPS provides AC power to thesystem, and communicates with it via a serial interface attached to one of the Console Controller asyncports at the rear of the system base. The Console Controller firmware communicates with the UPS forboth powerfail indication and for power shut off.

Presently, all Continuum 400 Series systems are powered from AC sources.

Figure 13 is a block diagram of the Continuum 400 Series power architecture.

Figure 13. Continuum 400 Series Power Block Diagram



3.5.1 AC Input Module

Each AC input module provides two line-filtered and switched AC outputs to the system base andsuitcase. The input to the module is in the range of 85 VAC to 264 VAC single-phase, 50 to 60Hz, with amaximum continuous input current of 15A rms.

System power is switched on and off via switches on the AC input modules.

Figure 14 is a block diagram of the AC input module.

Figure 14. AC Input Module Block Diagram



3.5.2 Suitcase Power Supply

There are four separate power outputs required by the CPU-Memory board and the cooling fans: +5VDC, +3.3 VDC and -5 VDC for the motherboard; and +12 VDC for the cooling fans.

The power for each suitcase is provided by a single power supply. However, the CPU/memory/logicfunction is designed as a 2N redundancy system, so that the overall function, including power is 2Nredundancy.

The connection from the power supply to the CPU-Memory board is made via 3 cable assemblies whichplug into the power supply and the CPU-Memory board. The first cable assembly provides the +5 VDCpower signal, the second provides the +3.3 VDC and the -5 VDC power signals, and the third providesthe +12 VDC power signal and the power supply STATUS and SHUTDOWN signals. The power isdistributed on the CPU-Memory board from the three power supply connectors via power planes withinthe board assembly.

Each suitcase is mechanically interlocked with a latching mechanism such that the suitcase will be firmlylocked to the system base. Each latching mechanism utilizes an interlock switch that disables the suitcasepower supply until the latching mechanism is fully engaged (locked). This interlock is provided toprevent full power arcing of the AC input connections.


The system base power supply subsystem consists of two power supplies. These power supplies powerthe disks and the PCI subsystem.

The power to the disks is configured as full 2N fault-tolerant, thereby providing the same level ofhardware availability to the disks as is provided for the CPU/memory/logic in the suitcases. The powerfor the PCI I/O cards is a simplexed (non fault-tolerant) architecture. Therefore, when an I/O powersupply fails, only half of the I/O loses power, and all of the disks remain functional. It is not possible toperform live insertion on the PCI bus.

There are five separate voltage levels required by the PCI I/O cards and the hard disk drives: +5 VDC,+12 VDC and -12 VDC for the PCI cards; +5 and +12 VDC for the disk drives. The system utilizes5 V PCI cards only, so a +3.3 VDC bias source is not required.



The connection from the power supply to the PCI bus backplane is two blind mating connectorassemblies on the rear of the power supply. The power signals are all contained within the firstconnector, and the power is distributed to the PCI backplane via a cable assembly. The remainder of thepower supply signals are contained in the remaining connector and are distributed to the PCI backplanevia a cable assembly.

The power signals are distributed on the PCI bus backplane from the power supply connector via powerplanes within the backplane board assembly.

PCI slots #0 through #7 and slots #8 through #15 are mechanically interlocked with an access door suchthat only one set of slots may be accessed at any time. Additionally, each access door utilizes afield-effect override switch that asserts a shutdown command to the associated system base power supplywhen the door is opened so that PCI cards may be inserted or removed without power applied. There is amanual interlock to prevent both covers from being open at the same time.

Also included in the system base power supply is an interlock switch that disables the power supply untilthe removal handle/latching mechanism is fully engaged. This prevents arcing of the AC input contactsduring removal/insertion into the system and implements a warm-plug strategy for the system base powersupply.

3.5.4 Grounding

All the CRUs in a Continuum 400 Series system provide for multiple low impedance earthings of chassisand ground planes to the system base chassis. Logic ground is bonded to chassis ground using solid metalstandoffs around the periphery of the CPU-Memory board and the PCI backplane.

The system obtains an earth reference ground via the two power line cords attached to the system base.Similarly, each peripheral connected to the system obtains ground through its line cords. For eachperipheral device, a high-integrity safety-ground conductor must be installed as part of the wiring system(in accordance with U.S. national electric code NFPA 70, or the equivalent). The international safetystandard (EN60-950) for electronic data processing (EDP) equipment also requires a ground conductor.Depending upon local conditions, ground potentials may differ between the system base and anyperipherals connected to the system base. Some peripherals, such as the disk expansion cabinet, mustalso be connected to the system base with a grounding strap. All grounds in the system must return to thesame reference point in the power distribution system, as close as possible to zero (0) volts potentialrelative to earth reference ground.

A Star Ground is often used to obtain the same earth reference ground. (See Figure 15.) Each earthreference ground, such as the system base ground, is returned separately to a common point where a zero(0) volt earth ground exists. The star ground ensures that all equipment is at the same potential and thatno noise or safety problems associated with an unpredictable or uncharacterized grounding system willoccur.

Figure 15. Star Ground Example



3.5.5 Power Fault and Status Reporting

Fault/status reporting is provided from both the suitcase and system base power supplies via an outputsignal indicating any of the following faults within the power supply:

any DC output Overvoltage or Undervoltage by +/-5% of the Nominal Value❍

any DC Output Overcurrent by at least +10% of Maximum Current❍

temperature of any power supply component >110º C (230º F)❍

low fan speed (system base power supply only)❍

The status output will be logic 1 to indicate normal power supply operation and logic 0 (e.g., 0 VDC) toindicate the presence of any fault condition. Both of the power supplies will, upon sensing any faultcondition, latch-off until the input power is cycled from ON to OFF to ON.

At power on, each power supply POWER STATUS output transitions from logic 0 to logic 1 only afterthe last power supply output has entered into its specification range (+/- 5% of the nominal value).

The status output from each suitcase power supply terminates on the CPU-Memory board within thesuitcase with which it is shared as a signal STATUS. This is a low true output signal with the occurrenceof a fault (e.g., it is logic 0 when a fault has occurred within the suitcase power supply).

The status outputs from each system base power supply terminates on the PCIB card located in slots #0and #8 on the PCI backpanel as a signal STATUS_OUT_H and STATUS_OUT_L. These arecomplimentary logic signals, where STATUS_OUT_H is a low true output signal with the occurrence ofa fault (e.g., it is logic 0 when a fault has occurred within the system base power supply), andSTATUS_OUT_L is the logical compliment of this signal.

3.5.6 Power Switches

System power is turned on and off via the two ON/OFF switches on the rear of the system base. Theseswitches are part of the AC input module.



3.6 Cooling SystemContinuum 400 Series systems utilize an air evacuation cooling system that works by creating negativepressure inside the system which in turn draws air in through the front of the machine and exhausts it outthe rear. (See Figure 16.)

The CPU-Memory board is cooled by three tube axial fans arranged vertically at the rear of the suitcase.The lower-most fan also cools the suitcase power supply by drawing ambient temperature air through thepower supply from the front (the end opposite the AC input) and exhausting the air at the rear of thepower supply case.

The PCI cards in the system base are cooled via the discharge air from the lower-most suitcase fan andby natural convection. Cooling air for the other components in the system base is provided by the twofans at the rear of the system base power supplies (one fan per power supply). These fans draw air inthrough the front of the system base, then pass it through the disk enclosure before entering into thepower supplies after which it is discharged out the rear of the system. A minimum flow rate through thesystem base of approximately 100 CFM (cubic feet per minute) is required.

All cooling fans in the Continuum 400 Series system are variable speed fans. The fans are integrated intothe CRUs that they cool. Therefore, a fan failure results in a CRU failure.

The Continuum 400 Series system is equipped with a series of temperature sensing ICs which are locatedon the CPU-Memory board. These sensors monitor incoming ambient air temperature which is indicativeof the room ambient temperature in which the machine is operating.

Figure 16. Continuum 400 Series System Air Flow



Each disk shelf in the disk expansion tower/cabinet contains two dual speed cooling fans. The fans aremonitored by the personality unit via the logic signals on the shelf backplane.

All fans operate at low speed except when certain events occur. The following table lists the events andthe corresponding responses

.

Event Response

Room ambient temperature > 86º F All fans in the machine go to highspeed

Suitcase broken or removed All fans in the suitcase go to highspeed

Suitcase fan failure All other fans in the suitcase goto high speed

System base power supply broken orremoved

The other system base powersupply fan goes to high speed

System base power supply fan failure The other system base powersupply fan goes to high speed

Disk expansion tower/cabinet fanfailure

The other disk shelf fan goes tohigh speed



Room ambient temp. > 32+2º C (90+3ºF)

Both fans on disk shelf go to highspeed

In each case, once the event has been rectified, the cooling system returns to normal operation.



4. Fault IsolationThe isolation of Continuum 400 Series hardware system failures is accomplished through the use ofLEDs, error messages, log files, and diagnostics. Maintenance software monitors all systemmalfunctions. If a malfunction is transient, the hardware component restarts automatically. If amalfunction is permanent, the part is taken out of service and its status LED indicates a failure.

Remote diagnostics is provided via the Remote Service Network (RSN) which automatically reportsfailures to the Stratus CAC. The CAC then notifies the customer and diagnoses the cause of the failure.

The following topics are covered in this section:

Component LEDs●

System logs●

Component status●

Isolating Memory Faults●

Troubleshooting procedures●

Diagnostic Process eXecutive (DPX)●

Service considerations●

4.1 Component LEDsThe suitcases, boards, disk drives, and system base power supplies all contain status LEDs that indicatethe condition of the components. Figure shows the interpretations of the various status conditions thatmay appear on the LEDs.

NOTE: The status LEDs for the PCI cards are located on the PCI card cage directly aboveeach slot.

Figure 17. Component LEDs

Fault Isolation


The LED codes are as follows:

Fault - The component has been removed from service by the operating system. It is broken and itspartner is usually displaying simplexed status.

●

Running (Duplexed) - The component is fully functional and is partnered.●

Do NOT Pull (Simplexed) - The component is fully functional but is not partnered.●

Testing - Occurs upon power up or when a component is hot-plugged into the system.●

NOT Configured - The component doesn't have a compatible PROM revision with its partner.●

Partially Broken (OK to Pull) - A part within the component has failed.●

Partially Broken (Do NOT Pull) - The component and its partner are partially broken.●

The disk expansion tower/cabinet LEDs identify error conditions or failures caused by the majorcomponents in the disk shelves. They are located on the power supplies, SCSI conversion units,

Fault Isolation


personality units, and disk drives. The disk drives' LEDs operate in the same manner as in the systembase.

Each power supply in an expansion tower/cabinet contains two LEDs. The top/left LED is the shelfstatus (fan) LED and the bottom/right LED is the power supply status LED.

The personality unit contains two LEDs, each of which is a status LED for the fans on the disk shelf. Theleft/bottom LED is for Fan 2 (bottom/left); the right/top LED is for Fan 1 (top/right). These LEDs areused in combination with the shelf status LED on the power supply to determine fan faults and overtemperature conditions.

NOTE: If the personality unit or internal SCSI conversion unit fails, these failures will beindicated by the complete termination of disk I/O on the affected disk expansion tower orshelf (or partial shelves) on the disk expansion cabinet. The status LEDs do not indicatefailure of these components.

Figure 18 shows the interpretations of the various status conditions that may appear on the LEDs.

Figure 18. D81X Disk Expansion Tower/Cabinet LEDs

4.2 System LogsSystem logs contain information concerning major system events and the time they occurred which canhelp detect and evaluate system problems. Failures are logged in the online system error log and are

Fault Isolation


displayed on the system console.

The elgpost command is used to read daily error messages from the /var/errlog file. To examine today'serror log, enter the following command:

/etc/elgpost |pg

NOTE: To exit the command at any time, enter :q.

To examine the error log for another day, enter the following:

/etc/elgpost /var/errlog.date |pg

where the date is entered in the yymmdd format.

For a detailed explanation of the system error messages, refer to FTX System Messages Manual (R475X).

4.3 Component StatusThe hwmaint ls command lists the components in the system and identifies any components that havebeen removed from service. The list serves as a simple troubleshooting tool to verify that all thecomponents are present and shows their status (in/out of service). Refer to Section 2.6.5 to view a samplehwmaint ls output screen.

If a component is not functioning properly, an out-of-service code is displayed in the code field. Theout-of-service code is repeated at the bottom of the display with a short description of the problem.

The CPU-Memory board and Console Controller module in the suitcase can be in one of the states shownin the following table:

State Description

online duplexBoard is fully functioning and running in lockstep(CPU-Memory board) or logically paired (Console Controller)operation with its partner board.

online simplexBoard is fully functioning, but it is not running in lockstep(CPU-Memory board) or logically paired (Console Controller)operation with a partner board. The partner board can be offline, broken, or missing.

offlineBoard is operationally sound, but it has not transitioned intothe online state. Typically, the offline state is a transitionstate used by the board while it is initializing or runningdiagnostics.

brokenBoard is not operational and is essentially isolated from thesystem. A broken board usually indicates a hardware failurethat requires either a reset or repair.

The PCI adapter cards in the PCI I/O card cage can be in one of the states shown in the following table.

Fault Isolation


State Description

online simplex Board is fully functioning, but it is not running in lockstep orstandby operation with a partner board.

offlineBoard is operationally sound, but it has not transitioned intothe online state. Typically, the offline state is a transition stateused by the board while it is initializing or running diagnostics.

brokenBoard is not operational and is essentially isolated from thesystem. A broken board usually indicates a hardware failurethat requires either a reset or repair.

Because PCI cards cannot support Stratus duplexing, they do not have an online duplex state. However,certain ones can be paired in a fault-tolerant configuration by defining an appropriate virtual SCSI bus asdescribed in Section 3.3.2.

4.4 Isolating Memory FaultsThis section describes how to decode memory ECC error messages and identify which memory moduleon a CPU-Memory board has failed.

4.4.1 FTX

Memory modules are mapped into the address space starting as base address 0 or base address 2GB - ie.0x00000000 or 0x80000000.

A given CPU-Memory board is either populated with 128-MB (M702) or 512-MB (M713) memorymodules. The following tables indicate the mapping of memory modules to physical locations on theCPU-Memory board.

128-MB memory module identification for base address 0x00000000

MemoryModule # Address Range Physical Location on

CPU-Memory Board

0 0x00000000-0x07FFFFFF J8/J9

1 0x08000000-0x0FFFFFFF J12/13

2 0x10000000-0x17FFFFFF J10/J11

3 0x18000000-0x1FFFFFFF J14/J15

128-MB memory module identification for base address 0x80000000


CPU-Memory Board

0 0x80000000-0x87FFFFFF J8/J9

Fault Isolation


1 0x88000000-0x8FFFFFFF J12/13

2 0x90000000-0x97FFFFFF J10/J11

3 0x98000000-0x9FFFFFFF J14/J15

512-MB Memory Module identification for Base address 0x00000000


CPU-Memory Board

0 0x00000000-0x1FFFFFFF J8/J9

1 0x20000000-0x3FFFFFFF J12/13

2 0x40000000-0x5FFFFFFF J10/J11

3 0x60000000-0x7FFFFFFF J14/J15

512-MB Memory Module identification for Base address 0x80000000


CPU-Memory Board

0 0x80000000-0x9FFFFFFF J8/J9

1 0xA0000000-0xBFFFFFFF J12/13

2 0xC0000000-0xDFFFFFFF J10/J11

The following is a sample output from the elgpost -f command indicating a memory module failureon a system with 128-MB memory modules running FTX 3.2.

Sample Command:

# elgpost -f errlog.961030 | grep -i ecc

Sample Output:

Oct 30 16:29:36.21 [FTS:I FTS_ECC_ERR ,cpu1] g748 03 CPU boardsingle bit ecc error at 0x1e4f7a50 syndrome 0x1f




Oct 30 16:46:40.21 [FTS:I FTS_ECC_ERR ,cpu0] g748 03 CPU board

Fault Isolation


single bit ecc error at 0x1e4f7a50 syndrome 0x1f

The errors at location 0x1e4f7a50 indicate a failure on memory module 3 since the address is between0x18000000 and 0x1FFFFFFF.

4.5 Troubleshooting ProceduresWhen a fault occurs, several things can happen. If it is a non-critical fault, the system will continue toprocess data. If it is a critical fault, the system (or a subsystem) may be inoperative. Whentroubleshooting the system, determine the fault first by using the LEDs and screen messages. Then,verify that the component(s) is out of service using the hwmaint ls command (Section 2.6.5), error logs,and diagnostic tests. This flow is shown Figure 19.

Figure 19. Troubleshooting Flow

When troubleshooting system faults, use the following process:

Check the console terminal for fault information.●

Locate the failed component(s) using the LEDs.●

Verify the component is bad by using software commands/error logs.●

Run system diagnostics tests (DPX).●

Remove and replace the failed component.●

Check to make sure the problem is resolved.●

4.5.1 Non-Critical Fault (System Operational)

A non-critical fault is when there is a bad component in the system, but the system is still functioning.Some components, especially boards may have a transient or hard failure. A transient failure is atemporary hardware failure that causes the board's red light to illuminate, but the board or softwarecorrects the failure. It does not place the board out of service unless the error count occurring on the

Fault Isolation


board is very high. Inspect the /var/errlog file for transient failures. Analyze the type and frequency ofthe failure to determine if board replacement is necessary.

Hard failures cause the red LED on a failed board to remain on, and the board cannot be restored toservice. Replace boards experiencing a hard failure after verifying that the board is not operational usingthe hwmaint ls bd command. (See Section 2.6.5.) If necessary, add the board using the hwmaint addslot_location command.

4.5.2 Critical Fault (System not Operational)

A critical fault is when the system is partially or totally inoperable. To isolate a critical system problem,use the same troubleshooting procedure described in Section 4.4. If you're unable to isolate and repair thesystem, call the Customer Assistance Center (CAC).

4.6 Diagnostic Process eXecutive (DPX)DPX is a test harness that provides a user interface for executing system-level diagnostic tests. DPX canautomatically assess the current system configuration with regards to the available diagnostic tests andexecute a single or group of tests based on this information. This allows the user to test systeminteractions using one or more tests.

The System Diagnostic User's Guide (HR065) contains descriptions of the diagnostic tests that can beexecuted on Continuum 400 Series systems using the DPX interface. It also provides procedures forstarting DPX and executing the diagnostics tests.

4.7 Service ConsiderationsMost of the replaceable units in the Continuum 400 Series system are customer replaceable. Refer to theFTX Continuum 400 Series: Operation and Maintenance Guide (R442X) for CRU replacementprocedures.


A system base power supply that has failed should be left in place until it is replaced. The reason for thisis that when a system base power supply is removed, an air flow bypass hole is created which results inless air being drawn through the disk enclosures. This is because it is easier for air to enter through thehole left by the vacant power supply than be drawn through the disks.

4.7.2 PCI Cards

PCI card replacement should be done with great care since the mechanical hardware that is removedduring card replacement is susceptible to being dropped down between the card slots and falling throughholes in the PCI backplane.

Fault Isolation


4.7.3 System Placement

The Continuum 400 Series system should never be placed with its rear facing a wall. A minimumclearance of two feet is recommended.

The system should never be placed behind the exhaust air of any other electronic equipment in the office.Make sure that none of the air inlets or exhausts are blocked in any way.

Fault Isolation


5. Disk Expansion Cabinet Installation andMaintenanceThis section describes the procedure for installing an optional D811/2 disk expansion cabinet and removal/replacementprocedures for cabinet FRUs. Model D811 is the 120V disk expansion cabinet and D812 is the 240V version. The followingtopics are covered in this section:

Model Numbers●

Unpacking the cabinet●

Removing the cabinet from the pallet●

Leveling the cabinet●

Installing the skirt kit●

Connecting components●

Inspecting the cabinet●

Powering on the cabinet●

Removal/Replacement procedures●

5.1 Model NumbersD811/2 disk expansion cabinets are installed in a dual-initiator configuration. Refer to Section 3.4.2 for a detailed descriptionof dual-initiator configurations.

The following table lists the disk expansion cabinet model numbers for dual-initiator configurations.

Model Number Description

D811-020 120V disk expansion cabinet(dual-initiator configuration)

D812-020 240V disk expansion cabinet(dual-initiator configuration)

5.2 Unpacking the CabinetThe disk expansion cabinet is packed in a corrugated carton attached to a wooden shipping pallet, as shown in Figure 26.Unpack the cabinet as follows.

NOTE: Before unpacking the cabinet, inspect the shipping carton for signs of external damage. Report anydamage to the local carrier and to the CAC.

CAUTION: Failure to thermally stabilize preconfigured storage subsystems may damage drive media orassociated electronics when the unit is turned on. Environmental stabilization begins when the equipment isplaced in the room in which it is to be installed.

1. Remove the cover, the fasteners, and corrugated board from the pallet.

2. Remove the cartons containing the ramps and skirt kit and set them aside.

3. Cut the shipping straps. Some cabinets are packaged in a plastic barrier bag. If the cabinet arrives in a plastic bag, leave thebag in place until the cabinet has adjusted to the local temperature and humidity.

Disk Expansion Cabinet Installation and Maintenance


4. Once the cabinet is unpacked, examine the front and rear doors, right and left side panels, top panel, and undercarriage forany apparent damage. Report such problems immediately.

5. Retain the shipping container and all packing materials.

Figure 26. Shipping Container Contents

5.3 Removing the Cabinet from the PalletPerform the following procedure to remove the cabinet from the pallet.

1. Remove any packing material remaining on the pallet.

2. Remove the two unloading ramps from the carton and inspect them.

WARNING: Serious personal injury may result if correct safety precautions are not taken during the unpackingprocedure. All personnel should wear safety glasses. The ramps, ramp side rails, and metal hardware should beinspected for the following defects:- cracks more than 25 percent of ramp depth, either across or lengthwise down the ramp.- knots/knotholes going through the thickness of ramp and greater than 50% of ramp width.- loose, missing, or broken ramp side rails.- loose, missing, or bent metal hardware.If any of these defects are present, do not use the ramp. Investigate alternate means of removing the cabinet ororder a new ramp.

3. Attach the ramps by fitting the metal prongs into the holes on the pallet, as shown in Figure 27. Make sure that the arrowson the ramps match up with the arrows on the pallet.

Figure 27. Shipping Pallet Ramp Installation



4. Extend the ramps to their full length.

5. See Figure 28 for the location of the shipping bolts. Remove the bolts.

6. Remove the shipping brackets, shown in Figure 28, from the cabinet levelers and set them aside.

Figure 28. Shipping Bolts and Brackets

WARNING: The levelers must be raised fully for the cabinet to roll easily down the unloading ramps. Failure tofully raise the levelers may cause the cabinet to tip over as you roll it off the pallet or down the ramp.

7. Loosen the leveler locking nuts and screw the four cabinet levelers all the way up into the cabinet as shown in Figure 29.



WARNING: At least two people are required to unload the cabinet from the shipping pallet. Failure to usesufficient personnel may result in injury or equipment damage.

8. Carefully roll the cabinet off the pallet and down the ramps to the floor as shown in Figure 29.

Figure 29. Removing the Disk Expansion Cabinet from the Pallet

9. Secure any loose cables and roll the cabinet to its final installation position.

WARNING: Use extreme caution when rolling the cabinet across the floor. Failure to raise all leveler feet andto provide a clear path for the cabinet's casters may result in the cabinet tipping over and injury to personnel.

5.4 Leveling the CabinetPerform the following procedure to level the cabinet in the location where it will be installed.

1. Loosen the locknuts on all four leveler feet as shown in Figure 30.

Figure 30. Leveler Foot Adjustment



2. Turn each leveler hex nut clockwise until the leveler foot contacts the floor.

3. Adjust all four leveler feet until the cabinet is level and the load is removed from all casters. Verify that the casters spinfreely.

4. Tighten the locknuts on all four leveler feet.

5.5 Installing the Skirt KitThe skirt kit is packaged separately inside the corrugated carton with the cabinet. Installation of the skirt is optional. Performthe following procedure to install the skirt kit around the base of the cabinet.

1. Unpack the skirt kit carton and use Figure 31 to identify the right, left, front, and rear skirts.

Figure 31. Skirt Installation

2. Position the skirts next the cabinet as shown in Figure 31.

3. The fasteners on the skirts consist of small pins with flat and barbed sides. Using a Phillips screwdriver, turn the fasteners



on each skirt until the flat sides face up.

4. Position each skirt such that the fasteners mate with the receptacles on the cabinet's base.

5. Using a Phillips screwdriver, push each fastener straight into its mating receptacle on the cabinet base.

With the fasteners locked in place, a small amount of play allows the skirts to be adjusted slightly up or down for properalignment.

5.6 Connecting the Disk Drives to the SystemThe disk expansion cabinet is shipped with the following components installed on each shelf:

one personality unit●

two power supplies●

one SCSI conversion unit (unless the shelf is the second in a pair)●

disk drives (up to five on shelves containing a SCSI conversion unit; up to six on shelves not containing a SCSIconversion unit)

The cabinet can contain up to four pairs of shelves (1/2, 3/4, 5/6, and 7/8). Only the first shelf in a pair contains a SCSIconversion unit. (See Figure 32.)

Figure 32. Shelves Containing SCSI Conversion Units

The disks in the expansion cabinet are connected in a dual-initiator configuration.

Each pair of shelves in the disk expansion cabinet is controlled by a U502 differential SCSI PCI card(s) in the systembase. A port on the U502 card is cabled to the SCSI conversion unit in the top shelf in each pair. The personality unitson the paired shelves are connected to each other with another SCSI cable.

The following is the procedure for configuring the personality units and connecting the disk drives in the diskexpansion cabinet to the system. It assumes that both shelves in each pair contain power supplies and disk drives, andthat one shelf contains a SCSI conversion unit.

1. Squeeze the locking tabs on the personality units that are in the shelves containing SCSI conversion units and pullthe personality units part way out of the shelves. Make sure all the dip switches on each of these personality units is set

●



to the OFF (down) position. (See Figure 33.)

2. Squeeze the locking tabs on the personality units that are in the shelves that do not contain SCSI conversion units.Pull the personality units part way out of the shelves. Make sure dip switches 1, 2, and 3 are set to the ON (up) positionand switches 4, 5, 6, and 7 are set to the OFF (down) position.

Figure 33. Personality Unit Switch Settings and Cable Connections

3. Connect an AW-B15900-XX differential SCSI cable to the front of each SCSI conversion unit. (See Figure 34.)

Figure 34. Connecting a SCSI Cable to the SCSI Conversion Unit



4. Route the differential SCSI cables as shown in Figure 35.

Figure 35. Routing the Differential SCSI Cables

5. Install a U502 PCI differential SCSI card in slot 3 in each of the card cages in the system base. (See Figure 36.)

NOTE: The U502 PCI cards must be installed in slot 3 in each card cage in order for the vsbconf commandto automatically configure the virtual SCSI buses (VSBs) for the disk drives in the cabinet. If a second pair ofU502 cards is used, they should be installed in slot 6 in each card cage.

Figure 36. Installing a U502 PCI Differential SCSI Card

6. On the rear of the system base, remove the upper screws from the left and right edges of the system base, as shownin Figure 37. The screw holes will be used to mount a wire-frame cable tray to the rear of the system base. The cabletray supports the dual-initiator cables.

Figure 37. Removing Screws from the Rear of the System Base



7. Remove the cable tray (MS-002571) from its carton. Because of the weight of the dual-initiator cables, this traymust be installed prior to connecting the dual-initiator cables.

Align the mounting holes on the sides of the cable tray with the screw holes on the rear of the system base, as shown inFigure 38. Firmly tighten the captive thumb screws to attach the cable tray to the rear of the system base.

NOTE: If cables are currently connected to the rear of the system base, place the cable tray under the cablesbefore attempting to align the cable tray with the screw holes.

Figure 38. Attaching the Cable Tray to the System Base



8. Using the components in the dual-initiator cable kit (AX-E66000), create a dual-initiator cable assembly asdescribed in the following steps.

(a) Attach two AW-B17900 differential SCSI cables to two tri-link adapters as shown in Figure 39.

(b) Connect the two tri-link connectors using a AW-B17800 SCSI jumper cable. You can attach the cable toeither female port on the tri-link connector.

(c) Connect a JT-000015 terminator to the unused port on one of the tri-link connectors.

(d) Attach the other end of the AW-B15900-XX differential SCSI cable (that is connected to the SCSIconversion unit on shelf #1 in the disk expansion cabinet) to the unused port on the other tri-link connector.

9. Repeat Step 8 for the second pair of shelves in the expansion cabinet.

Figure 39. Dual-Initiator Cable Assembly



10. Connect the other end of each AW-B17900 differential SCSI cable in the first cable assembly to port 0 (labeled"A") on the U502 PCI cards, as shown in Figure 40.

11. Connect the other end of each AW-B17900 differential SCSI cable in the second cable assembly to port 1 (labeled"B") on the U502 PCI cards.

Figure 40. Connecting Dual-Initiator Cable Assemblies to U502 PCI Cards



12. At the rear of the cabinet, make sure the circuit breakers on both cable distribution units (CDUs) are in the OFF(down) position. (See Figure 41.)

13. Connect the primary power cords to the CDUs.

Figure 41. Connecting the Power Cords to the CDUs



5.7 Inspecting the CabinetPerform the following procedure to inspect the cabinet installation.

1. Make sure that all hardware within the cabinet is fastened securely, and that there are no loose pieces present in thecabinet interior.

2. Make sure that all four leveler feet are lowered to support the full weight of the cabinet, and that the cabinet is level.

3. Make sure that there are no obstructions to the airflow from the shelf blowers. (The side panels may need to beremoved to check the shelf blowers.)

4. Check the identification label on the rear of the cabinet to verify that the cabinet is configured to accept the poweravailable at the site.

5. Make sure that all AC power cards connected from the shelves and cabinet fans to the cabinet power supplies arefirmly seated in their connectors at both ends.

6. Make sure that all internal signal cables are firmly seated in their connectors at both ends.

7. Make sure that all disk drives are seated firmly in their shelves.

8. Make sure that all necessary external interface cables are installed and firmly seated in their connectors.

5.8 Powering On the CabinetPerform the following procedure to turn on the power for the cabinet.

1. Plug the primary power cords from the CDUs into the appropriate power receptacles.

2. At the rear of the cabinet, switch the circuit breaker on each CDU to the ON (up) position.

3. Verify that all shelf blowers are operating and that both status indicators on each shelf power supply are illuminated.

4. Measure the cabinet's leakage current. If the leakage current exceeds 3.5 mA, Stratus recommends that you installpower cables with industrial type B, IEC 950 connectors. For more information on this type of connector, seeContinuum Site Planning: 400 Series (R411).

WARNING: Failure to reduce the leakage current can cause equipment performance degradation and cancreate an electric shock hazard.

5.9 Initializing the Disk DrivesTo initialize the disk drives in a disk expansion cabinet on a running Continuum 400 series system, the virtual SCSIbuses (VSBs) must be configured. Perform the following procedure to configure the VSBs for the disk expansioncabinet.

CAUTION: Make sure the the first pair of U502 cards are installed in slot 3 in the card cages. If there is a second setof U502s, they should be installed in slot 6 in the card cages. If the U502 cards are not installed in this manner, thefollowing software configuration procedure will not work properly.

1. Enter the following command to activate the VSBs.

/etc/default/vsbconf -Z -a dualexp

2. Enter the following command to verify that the U502 cards are online:



hwmaint ls u502

The State column of the screen display should list the U502s as onln.

3. Enter the following command to verify that the disks in the expansion cabinet are online.

hwmaint ls disk

The State column of the screen display should list the disk drives in the expansion cabinet as onln.

If any disks are offline, they will be listed as ofln.

The Slotloc field indicates the VSB locations of the disk drives. Each VSB location is defined by aVSB number, a disk slot location, and a logical unit number (always 0). Disk slot locations in each shelfin the expansion cabinet are numbered from 1 to 5, beginning with the slot next to the SCSI conversionunit.

The following table shows the correspondence between the VSBs and the shelves in the disk expansioncabinet in a dual-initiator configuration. It also shows the SCSI IDs of the drives.

VSB Shelf # SCSI IDs

4 1 1-5

4 2 8-13

5 3 1-5

5 4 8-13

6 5 1-5

6 6 8-13

7 7 1-5

7 8 8-13

CAUTION: The following step must be performed to update the primary flash card with the modified/stand/conf file on the disk. If it is not done, all the VSB information on the disks in the expansion cabinet will belost on the next reboot .

4. If the U502 cards and disk drives in the disk expansion cabinet are online, enter the following command to updatethe primary flash card.

flashadm -avd /dev/rflash/flashcard_location

where flashcard_location is either c2a0d0 for the flash card in card cage 2 or

c3a0d0 for the flash card in card cage 3.

5. Complete the logical disk drive configuration by entering the appropriate Veritas disk management commands asdescribed in the FTX System Administrator's Guide: Volume management (R479X).

5.10. Removal/Replacement Procedures

5.10.1. Shelf Fan Removal

1. Logically delete the disks in the shelf containing the failed fan. (Refer to Section 2.7.7)



2. Remove the locking brackets securing the disk shelf that contains the failed fan. (See Figure 41a.)

3. Slide the shelf out of the cabinet.

Figure 41a. Removing a Cabinet Shelf

4. Use a Phillips screwdriver to remove the safety screw in the upper right or lower left corner of the fan. (See Figure41b.)

5. Press the upper and lower fan mounting tabs together to release the fan.

6. Pull the fan straight out to disconnet it from the shelf power connector.

Figure 41b. Removing a Shelf Fan

5.10.2 Shelf Fan Replacement

1. Align the replacement fan connector and push the fan straight in, making sure that both mounting tabs lock in place.

2. Install the Phillips safety screw.

3. Replace the disk shelf in the cabinet.



4. Install the locking brackets to secure the disk shelf.

5. Verify that the shelf and all other shelf components are operating properly by observing the LEDs.

6. Logically add the disk drives in the shelf. (See Section 2.7.8).



6. UpgradesThe following upgrades are presented in this section:

Memory upgrade ●

Disk expansion cabinet upgrade●

6.1 Memory UpgradeThis section describes how to install memory module upgrades to Continuum 400 Series systems at customer sites. Theprocess consists of three parts:

Use the CPU/Memory Burn ID-PROM tools to update ID PROM. ●

Add memory modules to the suitcase. ●

Switch the operational status of the suitcases to update the partner.●

The minimum FTX operating system release and promcode revision level required for upgrading the suitcase are FTX 3.1and suitcase promcode revision 37 or above.

CAUTION: Make sure the system has been updated to the required operating system version and promcoderevision before performing any part of the upgrade procedure. Failure to do this may result in a system crash

6.1.1 Hardware Components

6.1.1.1 Suitcase

The suitcase houses the CPU/cache and memory modules, the CPU-Memory motherboard (which contains the ConsoleController module), and a power supply. The CPU-Memory motherboard is an assembly that accommodates 96 MHzCPU/cache modules (512 KB or 2 MB cache), and from one to four 128-MB or 512-MB memory modules.

The part number of the finished (top level) assembly is determined by the CPU/cache and memory configurations. The IDPROM for the top level assembly resides on the CPU-Memory motherboard, and contains information on the part number,the revision level, and the serial number. This information is set according to the CPU/cache and memory configuration,and a combination of the CPU-Memory motherboard, CPU/cache module, and memory module revision levels. Thesuitcase serial number never changes, regardless of part number or revision changes.

The following table lists the model numbers of the suitcases used in Continuum 400 Series systems.

NOTE: Although their descriptions are similar, models G811 thru G828 (sub model 00) are not compatible withmodels G811 thru G828 (sub model 10). The difference is that the sub model 10 CPU-Memory motherboards havean updated ASIC chip which is not compatible with sub model 00 boards.

Model Description

G811-G814 (Sub Model 00) 96 MHz, Uni processor, 512 KB cache, 128-512 MBmemory (128-MB memory modules)

G811-G814 (Sub Model 10) 96 MHz, Uni processor, 512 KB cache, 128-512 MBmemory (128-MB memory modules)

G815-G818 (Sub Model 00) 96 MHz, Uni processor, 2 MB cache, 128-512 MBmemory (128-MB memory modules)

G815-G818 (Sub Model 10) 96 MHz, Uni processor, 2 MB cache, 128-512 MBmemory (128-MB memory modules)

G821-G824 (Sub Model 00) 96 MHz, Twin processor, 512 KB cache, 128-512 MBmemory (128-MB memory modules)

Upgrades


G821-G824 (Sub Model 10) 96 MHz, Twin processor, 512 KB cache, 128-512 MBmemory (128-MB memory modules)

G825-G828 (Sub Model 00) 96 MHz, Twin processor, 2 MB cache, 128-512 MBmemory (128-MB memory modules)

G825-G828 (Sub Model 10) 96 MHz, Twin processor, 2 MB cache, 128-512 MBmemory (128-MB memory modules)

G835-G838 (Sub Model 10) 96 MHz, Uni processor, 2 MB cache, 512 MB-2 GBmemory (512-MB memory modules)

G845-G848 (Sub Model 10) 96 MHz, Twin processor, 2 MB cache, 512 MB-2 GBmemory (512-MB memory modules)

Memory Modules

The memory module is a daughter card that contains 128 MB or 512 MB of memory. The suitcase supports one to fourmemory modules, which equals a memory range of 128 MB to 512 MB (using 128-MB modules) or 512 MB to 2 GB(using 512-MB modules).

The memory module is packaged in static protective material. Its serial number, part number, and revision level are markedon the bottom of the module.

6.1.1.2 Upgrade Kits

The following table lists the marketing IDs of the memory upgrade kits and describes the upgrades.

Marketing ID Description

UPMXXXX Memory upgrade (128 MB memory)

UPMXXXX Memory upgrade (512 MB memory)

6.1.2 ESD Requirements

Since many of the components on the CPU-Memory motherboard are particularly susceptible to ESD (Electro-StaticDischarge), the CPU-Memory motherboard must be protected from ESD from the time that the board cover is removeduntil the cover is securely back in place. ESD protection kits must be employed when doing reconfigurations of Continuumsuitcases.

The memory modules must also be protected from ESD before they are removed from their ESD-protected packaging, andwhile being handled.

To prevent equipment damage while handling components, take the following ESD precautions:

A securely fastened ESD wrist strap MUST be worn at all times when removing the components. ●

Avoid touching a component's leads or contacts.●

Set up the ESD protection kit as close to the system as possible. Instructions for setting up the rubberized mat, groundingwrist strap, etc. are supplied with the kit.

6.1.3 Upgrade Procedure

This section describes the steps needed to install a memory upgrade into a customer's system in the field.

The procedure is performed in the following sequence:

Burn the ID PROM on the CPU-Memory motherboard in the first suitcase. ●

Burn the ID PROM on the CPU-Memory motherboard in the second suitcase. ●

Upgrades


Install memory modules on the CPU-Memory motherboard in the first suitcase. ●

Switch the operational status of the two suitcases. ●

Install memory modules on the CPU-Memory motherboard in the second suitcase.●

CAUTION: ESD protection must be maintained for all parts of this process where the ESD covers have beenremoved from the suitcase

Before you begin the procedure, check the memory modules you will be adding to the system. Write down the followinginformation for each component (listed on the bottom of each module) and indicate which suitcase (0 or 1) the componentwill be installed in. This is important because you need to enter the information in the proper suitcase ID prom and placethe correct labels on the suitcase after you have completed the upgrade procedure.

Subassembly model number ●

Sub model number ●

Serial number ●

Revision number●

Artwork revision number (always 0)●

NOTE: The subassembly model number and sub model number are listed within the part number (e.g., if thepart number is AA-M70300, the subassembly model number is M703 and the sub model number is 00).

The following subsections outline the procedure in detail. Be sure to follow the steps in the order they are listed.

6.1.3.1 Burning the Board ID PROM

1. Login as root.

2. Use the update_idprom command to burn one of the CPU-Memory motherboards with the new ID PROMimage.

Sample Command:

/sbin/update_idprom -i /dev/idmem/idmem00

where -i specifies the idmem file and 00 is the slot number.

NOTE: The suitcase must be online for update_idprom to access the CPU-Memorymotherboard ID PROM.

The following screen appears.

Show/Add_/Delete_/List__subassembly/Validate/Write/Exit?

3. To add subassemblies, enter a.

The following screen will be displayed.

[Add] subassembly model?

4. Enter the required information for the memory module you will be installing on the CPU-Memory motherboard.

The following are the subassembly models for memory modules: M128 or M512

NOTE: On older systems the memory modules are M702 or M704.

5. After you have entered the subassembly model and pressed the Return key, you are prompted for the followinginformation on subsequent screens.

Upgrades


[Add] serial number?[Add] submodel?[Add] revision?[Add] art revision?

Fill in the information as requested. When finished, a screen similar to the following will appear.

Add: model=M128 serial=12345 submodel=0 rev=1 art_rev=2[Add] correct?

6. If the information is correct, enter y.


Show/Add_subassembly/Delete_subassembly/Validate/Write/Exit?

7. Repeat the process to add the remaining subassemblies.

8. When finished, enter v to validate the information.

A screen similar to the following appears.

CPU Board modelx G825, 96MHz Clock, 256MB MemoryFru ID and Subassembly info validated.


NOTE: The validate command reads through the subassembly information, verifies that the information is correct,and updates several fields of the idprom according to the subassembly information. An error message will appear ifany of the following are detected:- CPU/cache modules are not all the same type- invalid number of CPU/cache modules- invalid number of memory modules- memory modules are not all the same type- G8XX subassembly is missing- P254 subassembly is missing- subassembly model is unknown

9. If the validation is successful, enter w to write the information.


ID prom written and verified.


10. Enter e to exit.

11. Repeat the procedure to update the CPU-Memory motherboard in the second suitcase.

12. When both CPU-Memory motherboards have had their ID PROM updated, proceed to Section 6.1.3.2.

6.1.3.2 Installing Memory Modules

Upgrades


NOTE: When performing the following procedure, use care installing or reinstalling screws and covers tomaintain FCC compliance.

1. Logically delete the first suitcase to be upgraded. (Omit this step if this is the second suitcase being upgraded, since it isalready deleted.)

Sample Command:

hwmaint delete 0

The green LED on the deleted suitcase will go out and its amber LED will come on.

2. Loosen the captive locking screw on the suitcase locking lever. (See Figure 42.)

3. Slide the locking lever fully toward the unlock symbol.

4. Grasp the suitcase handle and lift the suitcase straight up off the system base.

Figure 42. Unlocking and Removing the Suitcase

Upgrades


5. Grasp the suitcase front cover and pull it straight out from the chassis as shown in Figure 43.

Figure 43. Removing the Suitcase Front Cover

6. Turn the suitcase on its side and remove the four screws securing the suitcase cover to the bottom of thesuitcase. (See Figure 44.)

Figure 44. Removing the Suitcase Cover Bottom Screw

Upgrades


7. Return the suitcase to its upright position and remove the six screws securing the cover to the front of thesuitcase. (See Figure 45.)

Figure 45. Removing the Suitcase Cover Front Screws

8. Remove the six screws securing the cover to the rear of the suitcase. (See Figure 46.)

Figure 46. Removing the Suitcase Cover Rear Screws

Upgrades


9. Pull outward on the bottom of the sides of the suitcase cover and lift it up and off the suitcase as shown inFigure 47.

Figure 47. Removing the Suitcase Cover

10. Remove the eight screws securing the cover to the CPU-Memory motherboard.(See Figure 48.)

Figure 48. Removing the Cover from the CPU-Memory Motherboard

Upgrades


11. Follow the procedure below to physically configure the board by adding memory modules. The followingtable describes the memory module upgrade options and their associated procedures. Refer to Figure 49 for thelocations of the memory modules on the CPU-Memory motherboard.

CAUTION: Be sure to follow all ESD precautions from this point on.

Model Number Upgrade Option Procedure

M128 128 MB to 256 MB Add one M128 memory module (#2)

M128 128 MB to 384 MB Add two M128 memory modules (#2 and #3)

M128 128 MB to 512 MB Add three M128 memory modules (#2, #3, and #4)


M128 256 MB to 512 MB Add two M128 memory modules (#3 and #4)


M512 512 MB to 1 GB Add one M512 memory module (#2)

M512 512 MB to 1.5 GB Add two M512 memory modules (#2 and #3)

M512 512 MB to 2 GB Add three M512 memory modules (#2, #3, and #4)

M512 1 GB to 1.5 GB Add one M512 memory module (#3)

M512 1 GB to 2 GB Add two M512 memory modules (#3 and #4)

M512 1.5 GB to 2 GB Add one M512 memory module (#4)

Figure 49. Memory Module Locations

Upgrades


To install a memory module:

Remove the memory module from its static protective package. 1.

Align the pins on the memory module with the connector on the CPU-Memory motherboard and carefully plug itinto the connector, making sure it is firmly seated. (See Figure 50.)

2.

Install the screw as shown to secure the module to the suitcase frame.3.

Figure 50. Installing a Memory Module

Replace the cover on the CPU-Memory motherboard and reinstall the eight screws.4.

Upgrades


Replace the suitcase cover and reinstall all screws (16).5.

Replace the suitcase on the system base and lock it in place.6.

Upon replacement of the suitcase, the CPU-Memory motherboard will automatically run the on-board power-up self test.The yellow LEDs on the suitcase will blink during the self test. At the end of the self test, the red LED will be on steady,indicating the self test detected a local ID mismatch and the board will not come into service.

If this is the first CPU-Memory motherboard (suitcase) to be updated, proceed to Section 6.1.3.3.

If this is the second motherboard (suitcase), go to Step 2.

7.

Bring the suitcase back online using the hwmaint add command.

Sample Command:

8.

hwmaint add 1

The suitcase should now duplex with its partner and they should both become green-lit.

6.1.3.3 Switching the Operational Status of the Boards

1. Use the hwmaint add -x command to switch the operational status of the first suitcase with the second suitcase(that is, the upgraded suitcase goes online and the partner suitcase is deleted).

Sample Command:

hwmaint add -x 0

where 0 is the slot location of the upgraded suitcase.

In a few minutes the suitcase containing the upgraded CPU-Memory motherboard will have its yellow andgreen LEDs lit (indicating it is online and simplexed), and the other suitcase will become amber-lit.

2. Go to Section 6.1.3.2 and follow the procedure (beginning at step 2) to physically upgrade the second board.

6.1.4 Updating the Suitcase Label

1. In the spaces provided on the label affixed to the top of the suitcase (see diagram below), write in the part numbers(P/N) and serial numbers (S/N) of the memory module(s) that were added to the CPU-Memory motherboard.

2. Install the label insert provided with the upgrade kit as shown.

Upgrades


6.2 D811/2 Disk Expansion Cabinet UpgradeThis section describes how to upgrade a D811/2 disk expansion cabinet to increase storage capacity. The upgradeprocedure consists of increasing the number of disk shelves in the cabinet from four to six or from six to eight.

6.2.1 Upgrade Kit Marketing IDs

The disks in an upgraded expansion cabinet are connected in a dual-initiator configuration. The following table shows themarketing ID and contents of the upgrade kit.

Marketing ID Description Contents

Upgrades


D818-020 Dual-initiator upgrade kit

AX-D65900 2-shelf upgrade kit containing:one AW-B15700-XX SCSI cableone disk shelf (includes personality unit)one internal SCSI conversion unitfour disk shelf power suppliestwo shelf mounting brackets (includes stopbrackets, locking brackets, and mountinghardware)

AW-B15900-XX SCSI cable

AA-E66000 dual-initiator cabling kit

NOTE: Refer to Section 3.4.2 for a detailed description of dual-initiator configurations.

6.2.2 Installing Disk Expansion Cabinet Shelves

NOTE: It is recommended that AC power be removed from the cabinet components while working in thecabinet interior.

Perform the following procedure to upgrade a D811/2 disk expansion cabinet.

1. On the backplane of each shelf at slot 2 (as viewed from the front of the shelf) is a set of 4 jumper pins labeledJ17. Make sure that on shelves that will contain a SCSI conversion unit (shelves 5 and 7) there is a jumper isinstalled across both positions 1 and 2 on J17. On each of the other shelves (shelves 6 and 8) make sure a jumper isinstalled across only position 2 on J17. (See Figure 56.)

Figure 56. J17 Configuration on Shelf Backplane

NOTE: In the following steps, the terms front and rear refer to locations in thecabinet. The terms inner and outer refer to positions on the shelf brackets.

2. Remove the filler panels from the rear of the expansion cabinet where the shelves will be installed. (See Figure 51.)

3. Remove the clip nuts that were securing the filler panel screws.

Figure 51. Removing Filler Panels from the Expansion Cabinet

Upgrades


4. Insert the tab on a stop bracket into the slot at stop position #4 on the shelf mounting bracket as shown in Figure 52.Fasten the stop bracket with a screw.

Figure 52. Installing a Stop Bracket on a Shelf Mounting Bracket

5. Determine the correct rear shelf mounting locations by checking the position of the shelves mounted in the front of thecabinet on the front vertical cabinet rails. The rear shelves will be installed in the same-numbered holes on the rear verticalcabinet rails.

6. Position a shelf mounting bracket at the correct rear mounting holes just behind the outer flange of the rear verticalcabinet rail as shown in Figure 53.

7. Align the bracket mounting studs and mounting tabs with the appropriate holes in the rear vertical cabinet rail and seatthe studs and tabs in the holes.

8. Fasten the bracket to the rear vertical cabinet rail by installing KEP nuts (5/16") on the upper-most and lower-mostmounting studs as shown.

Upgrades


NOTE: Do not install KEP nuts on the two middle studs at this time.

Figure 53. Installing a Shelf Mounting Bracket

9. Repeat Step 4 thru Step 8 to mount the companion shelf bracket on the opposite cabinet rail.

NOTE: To be sure of proper alignment of the shelf mounting brackets, make sure they are mounted in thecorrect mounting holes.

10. Align a disk shelf with its power supply end at the left and its fans facing the cabinet; then slide the shelf into the shelfmounting brackets until it contacts the stop brackets. (See Figure 54.)

NOTE: The shelf should slide smoothly into the shelf mounting brackets. If it binds, remove it and check thealignment of the shelf mounting brackets. The KEP nuts securing the shelf mounting brackets might need to beloosened slightly to allow the brackets to align with the shelf. Retighten the nuts when the shelf mountingbrackets are properly aligned.

11. Once the shelf is positioned within the shelf mounting brackets, install each rear shelf locking bracket on the twomiddle mounting studs and secure it with KEP nuts as shown.

Figure 54. Installing a Disk Expansion Cabinet Shelf

Upgrades


12. Repeat the above procedure to install the remaining shelves.

6.2.3 Connecting the Shelf Components

Each pair of shelves in the disk expansion cabinet is controlled by a U502 differential SCSI PCI card(s) in the system base.A port on the U502 card is cabled to the SCSI conversion unit in the top shelf in each pair. The personality units on thepaired shelves are connected to each other with another SCSI cable.

The following is the procedure for connecting components in the disk expansion cabinet. It assumes that both shelves ineach pair contain power supplies and disk drives, and that one shelf contains a SCSI conversion unit.

1. Connect an AW-B15900-XX differential SCSI cable to the front of each SCSI conversion unit. (See Figure 55.)

Figure 55. Connecting a SCSI Cable to the SCSI Conversion Unit

Upgrades


2. Route the differential SCSI cables as shown in Figure 56.

Figure 56. Routing the Differential SCSI Cables

3. On the rear of the system base, remove the upper screws from the left and right edges of the system base, as shownin Figure 57. The screw holes will be used to mount a wire-frame cable tray to the rear of the system base. The cabletray supports the dual-initiator cables.

Upgrades


Figure 57. Removing Screws from the Rear of the System Base

4. Remove the cable tray (MS-002571) from its carton. Because of the weight of the dual-initiator cables, this traymust be installed prior to connecting the dual-initiator cables.

Align the mounting holes on the sides of the cable tray with the screw holes on the rear of the system base, as shownin Figure 58. Firmly tighten the captive thumb screws to attach the cable tray to the rear of the system base.

NOTE: If cables are currently connected to the rear of the system base, place the cable tray under the cablesbefore attempting to align the cable tray with the screw holes.

Figure 58. Attaching the Cable Tray to the System Base

Upgrades


NOTE: If you are installing the old dual-initiator cable kit (AX-E66000), go to step 5.If you are installing the new one-piece dual- initiator cable (AW-B24600-14), go tostep 10.

5. Using the components in the dual-initiator cable kit (AX-E66000), create a dual-initiator cable assembly asdescribed in the following steps.

(a) Attach two AW-B17900 differential SCSI cables to two tri-link adapters as shown in Figure 59.

(b) Connect the two tri-link connectors using a AW-B17800 SCSI jumper cable. You can attach the cable toeither female port on the tri-link connector.

(c) Connect a JT-000015 terminator to the unused port on one of the tri-link connectors.

(d) Attach the other end of the AW-B15900-XX differential SCSI cable (that is connected to the SCSIconversion unit on shelf #1 in the disk expansion cabinet) to the unused port on the other tri-link connector.

6. Repeat Step 5 for the second pair of shelves in the expansion cabinet.

Figure 59. Dual-Initiator Cable

Upgrades


7. Connect the other end of each AW-B17900 differential SCSI cable in the first cable assembly to port 0 (labeled"A") on the U502 PCI cards, as shown in Figure 60.

8. Connect the other end of each AW-B17900 differential SCSI cable in the second cable assembly to port 1 (labeled"B") on the U502 PCI cards.

Figure 60. Connecting Dual-Initiator Cable Assemblies to U502 PCI Cards

Upgrades


9. Go to step 15.

10. Attach the other end of one of the AW-B15900-XX differential SCSI cables (the one connected to the SCSIconversion unit on shelf #1 of the first pair of shelves in the disk expansion cabinet) to P1 on the first dual-initiatorcable (AW-B24600-14).(See Figure 61.)

11. Attach the other end of the other AW-B15900-XX differential SCSI cable (the one connected to the SCSIconversion unit on shelf #1 of the second pair of shelves in the disk expansion cabinet) to P1 on the seconddual-initiator cable (AW-B24600-14).

12. Connect J1 of the first dual-initiator cable (AW-B24600-14) to port 0 (labeled "A") on one of the U502 PCIcards, as shown in Figure 62. Connect J2 to port 0 on the other U502 card.

13. Connect J1of the second dual-initiator cable (AW-B24600-14) to to port 1 (labeled "B") on the first U502 card.Connect J2 to port 1 on the other U502 card.

14. Attach the cable clamps on the AW-B24600-14 cables to the cable tray.

Figure 61. Connecting the AW-B24600-14 Dual-Initiator Cables to U502 PCI Cards

Upgrades


.

15. At the rear of the disk expansion cabinet, select two adjacent shelves to be paired. Figure 63 shows which shelvesshould be paired. (5/6 and 7/8)

Figure 63. Pairing of Adjacent Shelves

Upgrades


16. Squeeze the locking tabs on the personality units that are in the shelves containing SCSI conversion units andpull the personality units part way out of the shelves. Set all the dip switches on each of these personality units to theOFF (down) position. (See Figure 64.)

17. Attach one end of an AW-B15700 SCSI cable to the JB1 connector on the front of the personality units in theshelves containing SCSI conversion units. (See Figure 64.) Use a flat-bladed screw driver to secure the cable to thepersonality units. Push the personality units back into the shelves.

18. Squeeze the locking tabs on the personality units that are in the shelves that do not contain SCSI conversionunits. Pull the personality units part way out of the shelves. Set dip switches 1, 2, and 3 to the ON (up) position andswitches 4, 5, 6, and 7 to the OFF (down) position.

19. Attach the other end of each the AW-B15700 SCSI cables to the JA1 connector on these personality units. Pushthe personality units back into the shelves.

Figure 64. Personality Unit Switch Settings and Cable Connections

6.2.4 Initializing the Disk Drives

To initialize the disk drives in an upgraded disk expansion cabinet on a running Continuum 400 series system, the virtualSCSI buses (VSBs) must be configured. Perform the following procedure to configure the VSBs for the disk expansioncabinet.

CAUTION: Make sure the the first pair of U502 cards are installed in slot 3 in the card cages. If there is asecond set of U502s, they should be installed in slot 6 in the card cages. If the U502 cards are not installed inthis manner, the following software configuration procedure will not work properly.

1. Enter the following command to activate the VSBs.

/etc/default/vsbconf -Z -a dualexp

2. Enter the following command to verify that the U502 cards are online:

Upgrades


hwmaint ls u502

The State column of the screen display should list the U502s as onln.

3. Enter the following command to verify that the disks in the expansion cabinet are online.

hwmaint ls disk

The State column of the screen display should list the disk drives in the expansion cabinet as

onln. If any disks are offline, they will be listed as ofln.

The Slotloc field indicates the VSB locations of the disk drives. Each VSB location isdefined by a VSB number, a disk slot location, and a logical unit number (always 0). Disk slotlocations in each shelf in the expansion cabinet are numbered from 1 to 5, beginning with the slotnext to the SCSI conversion unit.

The following table shows the correspondence between the VSBs and the shelves in the diskexpansion cabinet in a dual-initiator configuration. It also shows the SCSI IDs of the drives.

VSB Shelf # SCSI IDs

4 1 1-5

4 2 8-13

5 3 1-5

5 4 8-13

6 5 1-5

6 6 8-13

7 7 1-5

7 8 8-13

CAUTION: The following step must be performed to update the primary flash card with the modified /stand/conf fileon the disk. If it is not done, all the VSB information on the disks in the expansion cabinet will be lost on the next reboot .

4. If the U502 cards and disk drives in the disk expansion cabinet are online, enter the following command to update theprimary flash card.

flashadm -avd /dev/rflash/flashcard_location

where flashcard_location is either c2a0d0 for the flash card in card cage 2 or

c3a0d0 for the flash card in card cage 3.

5. Complete the logical disk drive configuration by entering the appropriate Veritas disk managementcommands as described in the FTX System Administrator's Guide: Volume management (R479X).

Upgrades


7. Service Strategies

7.1 Maintenance StrategyThe Continuum 400 Series system is customer installable. Stratus offers a fee-based system start-uppackage for customers that choose to have Stratus perform the installation.

Service is provided under Stratus' current Assured Availability Services support portfolio. TheContinuum 400 Series includes a one year hardware warranty and a 90-day software warranty. Thewarranty level is consistent with Stratus' Comprehensive Service offering.

Customers are expected to replace all parts that are specified as Customer Replaceable Units (CRUs). Allother parts will be replaced by Stratus-trained personnel. Additional customer responsibilities areoutlined in the customer's service agreement.

7.2 Logistics StrategyPrimary spare support in the Americas is provided from the Logistics Support Center in Marlboro, MA.Internationally, there are three Logistics Support Centers located in Ireland, Japan, and Singapore.Ireland provides regional spares support to all countries within Europe. Singapore provides direct supportto Singapore, Hong Kong, and Australia. Singapore provides replenishment support to Japan. All repairsare performed by Ireland and distributed directly to Singapore for the entire region.

7.3 Training StrategyTraining on the Continuum 400 Series system is provided in updates to existing courses offered by theStratus Education Center. The courses are designed to familiarize service providers (Stratus fieldengineers, OEMs, and VARs) with the installation/maintenance requirements of the system.

For complete course information, including schedules, see the Stratus Education (Internal) homepage.

Service Strategies

file:///H|/CSDoc/400/section7.htm [4/7/2000 11:48:59 AM]

8. Related DocumentationThe following manuals contain additional information on installing and maintaining Continuum 400Series systems. It is recommended that you obtain a copy of each of these manuals.

Continuum 400 Series Illustrated Parts Breakdown●

Continuum Series 400/FTX Product Marketing Guide (SG115)●

System Diagnostic User's Guide (HR065) ●

FTX Continuum 400 Series: Site Planning Guide (R411) ●

FTX Continuum 400 Series: Installation Guide (R443X) ●

FTX Continuum 400 Series: Operation and Maintenance Guide (R442X) ●

FTX System Administrator's Guide: General Services (R455X) ●

FTX System Administrator's Guide: Volume Management (R479X) ●

FTX System Administrator's Guide: File Systems (R456X) ●

FTX System Messages Manual (R475X) ●

FTX Commands Reference Manual (R460X) ●

FTX Continuum 400 Series: U401 PCI Card Installation Guide (R437X) ●







FTX Continuum 400 Series: U540 PCI Card Installation Guide (R446X)●

Related Documentation

file:///H|/CSDoc/400/section8.htm [4/7/2000 11:49:03 AM]

9. Part NumbersThe following tables list the part numbers for the Customer Replaceable Units (CRUs), FieldReplaceable Units (FRUs), and Distributor Replaceable Units (DRUs) in the Continuum 400 Seriessystem.

9.1 Suitcase

Description FRU/CRU/DRU Part Number

Uni 96 MHz Suitcase (128 MB memory, 512 KBcache, SUB 00) CRU AA-G81100




Twin 96 MHz Suitcase (128 MB memory, 512 KBcache, SUB 00) CRU AA-G82100




Uni 96 MHz Suitcase (128 MB memory, 2 MB cache,SUB 00) CRU AA-G81500




Twin 96 MHz Suitcase (128 MB memory, 2 MBcache, SUB 00) CRU AA-G82500



Part Numbers




Uni 96 MHz Suitcase (1 GB memory, 2 MB cache,SUB 00) CRU AA-G83600

Uni 96 MHz Suitcase (1.5 GB memory, 2 MB cache,SUB 00) CRU AA-G83700



Twin 96 MHz Suitcase (1 GB memory, 2 MB cache,SUB 00) CRU AA-G84600

Twin 96 MHz Suitcase (1.5 GB memory, 2 MB cache,SUB 00) CRU AA-G84700













Part Numbers









Uni 96 MHz Suitcase (1.5 GB memory, 2 MB cache,SUB 10) CRU AA-G83710




Twin 96 MHz Suitcase (1.5 GB memory, 2 MB cache,SUB 10) CRU AA-G84710


M128 memory module PCB (128 MB) FRU AA-M12800

M512 memory module PCB (512 MB) FRU AA-M51200

CPU/cache module PCB (96 MHz, 512 KB cache) FRU AA-G80200

CPU/cache module PCB (96 MHz, 2 MB cache) FRU AA-G80300

CPU-Memory motherboard DRU AA-G80100

CPU-Memory motherboard (SUB 10 models) DRU AA-G80110

Memory adapter PCB DRU AA-E78300

3V memory power adapter PCB (SUB 10 models) DRU AP-E79500

LED board DRU AP-E78400

LED board cable DRU AW-000424

Suitcase fan DRU MF-000030

Part Numbers


Suitcase fan snap rivet (4 per fan) DRU F9-000193

Suitcase power supply DRU AA-P25300

Suitcase power cable #1 DRU AW-000900



Suitcase ON/OFF switch and cable DRU AW-000903

9.2 System Base

System base (without CRUs) FRU AA-E69900

System base power supply CRU AA-P25400

Disk chassis assembly DRU AX-D80000

SCSI disk chassis cable DRU AW-000908

Power disk chassis cable DRU AW-000909

System backplane PCB DRU AA-E78100

PCI fault display assembly DRU AA-P25600

AC line filter assembly DRU AA-P25500

RS-232 daughter card assembly DRU AA-E45045

Base power cable DRU AW-000905

Base logic cable DRU AW-000906

PCI bridge card CRU AA-K13800

PCMCIA flash card CRU AA-E52500

PCI programmable synchronous card (4 port) CRU AA-U40100

4-port RS-422 cable CRU AW-B16200

4-port X.21 cable CRU AW-B16300

4-port V.35/V.36 cable CRU AW-B16500

PCI programmable synchronous card (8 port) CRU AA-U40200

8-port RS-232 cable CRU AW-B23000

PCI 3-bus fast wide single-ended SCSI card CRU AA-U50100

PCI fast wide differential SCSI card CRU AA-U50200

Part Numbers


PCI 10/100 MB ethernet card CRU AA-U51000

PCI FDDI card* CRU AA-U53000

SCSI PCI cable DRU AW-000944

SCSI PCI cable (contains inline SCSI bus terminators) DRU AW-000952

1 GB 7200 RPM disk drive CRU AA-D80100



System base power cable (domestic, 120 V/20A), length = 1.8m (6 ft) CRU AW-B19008

System base power cable (domestic, 120 V/20A), length = 2.5m (8.2 ft) CRU AW-B19009



System base power cable (domestic, 250 V/20A), length = 2.5m (8.2 ft) CRU AW-B19012


System base power cable (Continental Europe, 250V/20A),length = 2.5 m (8.2 ft) CRU AW-B19014

System base power cable (Continental Europe, 250V/20A),length = 3.7 m (12 ft) CRU AW-B19015

System base power cable (Great Britain/India, 250V/13A),length = 2.5 m (8.2 ft) CRU AW-B19016

System base power cable (Great Britain/India, 250V/13A),length = 3.7 m (12 ft) CRU AW-B19017

System base power cable (Uncommitted/Universal,250V/20A), length = 2.5 m (8.2 ft) CRU AW-B19018

System base power cable (Uncommitted/Universal,250V/20A), length = 3.7 m (12 ft) CRU AW-B19019

System base power cable (Australia/New Zealand, 250V/15A,length = 2.5 m (8.2 ft) CRU AW-B19020

System base power cable (Australia/New Zealand, 250V/15A,length = 3.7 m (12 ft) CRU AW-B19021

System base power cable (Denmark/Switzerland, 250V/15A,length = 2.5 m (8.2 ft) CRU AW-B19022

System base power cable (Denmark/Switzerland, 250V/15A,length = 3.7 m (12 ft) CRU AW-B19023

Part Numbers


System base power cable (Italy, 250V/16A, length = 2.5 m (8.2ft) CRU AW-B19024

System base power cable (Italy, 250V/16A, length = 3.7 m (12ft) CRU AW-B19025

Power cable retainer CRU MS-0002472

9.3 Tape Drives/Modem

4 mm DDS2 DAT tape drive CRU AA-T80100

4 mm DDS2 DAT tape drive with 6-cartridge autoloader CRU AA-T80200

525 MB QIC tape drive CRU AA-T80300

Single-ended SCSI cable - U501 PCI card to T80X tape drive (12 ft) CRU AW-B21000-12

SCSI cable (for daisy chaining T80X tape drives) CRU AW-B20000-02

Single-ended SCSI terminator - T80X tape drives CRU JC-005TRM

6250 BPI 1/2" SCSI tabletop tape drive FRU AA-T20410

Standalone SCSI conversion unit CRU AA-E64300

Differential SCSI cable - Standalone SCSI conversion unit to T204tape drive (30 ft) CRU AW-B24000-30

Differential SCSI cable - Standalone SCSI conversion unit to T204tape drive (60 ft) CRU AW-B24000-60

Differential SCSI terminator - T204 tape drive CRU AX-T20001

Power cable (Contintial Europe, 120V/15A, Type vII G (CEE7),length = 2.5m (8.4 ft) CRU AW-B12800-01

Power cable (Great Britain/India, 250V/10A, Type BS/89/13), length= 2.5m (8.4 ft) CRU AW-B12800-02

Power cable (Italy, 250V/10A, I/3)), length = 2.5m (8.4 ft), length =2.5m (8.4 ft) CRU AW-B12800-03

Power cable (Australia, 250V/10A, SAA/3), length = 2.5m (8.4 ft) CRU AW-B12800-04

Power cable (Switzerland, 250V/10A, SEV 1011 Type 11), length =2.5m (8.4 ft) CRU AW-B12800-05

Power cable (Japan, 125V/15A, NEMA 5-15), length = 2.5m (8.4 ft) CRU AW-B12800-06

Power cable (North America, 120V/15A, NEMA 5-20, length = 2.5m(8.4 ft) CRU AW-B12800-07

Power cable (North America, 120V/15A, NEMA 5-20, length = 1.82m(6 ft) CRU AW-B12800-08

Single-ended SCSI cable - U501 PCI card to standalone SCSIconversion unit CRU AW-B25000-03

RSN modem CRU AA-C41900

Gender adapter (25 position M/M)) CRU JD-025PLG-07

Data cable (B cable), length = 7.6 m (25 ft) CRU AW-B10102-25

Data cable, length = 7.6 m (25 ft) CRU AW-B15200-25

Part Numbers


9.4 D81X Disk Expansion Tower/Cabinet

D810 disk expansion tower: pedestal kit (includes front/rearbezel, door, PDU, covers) FRU AA-D66000

D811 disk expansion cabinet (120V) FRU AA-D66200

D812 disk expansion cabinet (240V) FRU AA-D66100

D811/2 disk expansion cabinet 2-shelf upgrade kit FRU AA-D65900

D81X disk expansion tower/cabinet: disk shelf FRU AX-E64500

D81X disk expansion tower/cabinet: shelf fan CRU AX-E65200

D81X disk expansion tower/cabinet: shelf power supply CRU AX-E64400

D81X disk expansion tower/cabinet: shelf personality unit CRU AX-E65300

D81X disk expansion tower/cabinet: shelf internal SCSIconversion unit CRU AX-E64200

Differential SCSI cable - U502 PCI card to shelf internal SCSIconversion unit (3 m/10 ft) CRU AW-B15800-01




D811 disk expansion cabinet: AC distribution unit (120V, 50/60Hz) FRU AX-E64600

D812 disk expansion cabinet: AC distribution unit (230V, 50/60Hz) FRU AX-E65500

D811/2 disk expansion cabinet: AC cable harness (black) FRU AW-B17300

D811/2 disk expansion cabinet: AC cable harness (grey) FRU AW-B17400

D811 disk expansion cabinet: line cord (120V/20A, domestic,twist-lock NEMA5-20P) CRU AW-B15500

D812 disk expansion cabinet: line cord (240V/16A,international, IEC 309) CRU AW-B15600

D811 disk expansion cabinet: AC fan (120V) FRU AX-E64700

D812 disk expansion cabinet: AC fan (230V) FRU AX-E65600

Part Numbers


D811/2 disk expansion cabinet: SCSI cable (for daisy chainingshelf-to-shelf) CRU AW-B15700

D811/2 disk expansion cabinet: shelf mounting bracket &hardware kit FRU AX-E65400

D810 disk expansion tower: line cord (US/Japan, 120V/15A)length = 3 m/9.8 ft CRU AW-B15400-01

D810 disk expansion tower: line cord (Central Europe,250V/10A) length = 2.5 m/8.2 ft CRU AW-B15400-02

D810 disk expansion tower: line cord (Ireland/UK, 250V/10A)length = 2.5 m/8.2 ft CRU AW-B15400-03

D810 disk expansion tower: line cord (Italy, 250V/10A) length= 2.5 m/8.2 ft CRU AW-B15400-04

D810 disk expansion tower: line cord (Switzerland, 250V/10A)length = 2.5 m/8.2 ft CRU AW-B15400-05

D810 disk expansion tower: line cord (Denmark, 250V/10A)length = 2.5 m/8.2 ft CRU AW-B15400-06

D810 disk expansion tower: line cord (Australia/N. Zealand,250V/10A) length = 2.5 m/8.2 ft CRU AW-B15400-07

D810 disk expansion tower: line cord (Israel, 250V/10A) length= 2.5 m/8.2 ft CRU AW-B15400-08

D810 disk expansion tower: line cord (India/S.Africa,250V/10A) length = 2.5 m/8.2 ft CRU AW-B15400-09

D81X Cable basket CRU MS-0002571

D81X Dual-initiator cable kit (includes AW-B17800,AW-B17900, JT-000015, JC-204CON) CRU AX-E66000

--Differential SCSI jumper cable - tri-link connector to tri-linkconnector (48.3 cm/19 in) CRU AW-B17800

--Differential SCSI adapter cable - U502 PCI card to tri-linkconnector (22.9 cm/9 in) CRU AW-B17900

--SCSI bus terminator CRU JT-000015

--Tri-link connector CRU JC-204CON

D81X Dual-initiator cable CRU AW-B24600-14

Part Numbers


FTX Continuum 400 Series (PA-7100) Technical Service Guide 400 (7100... · FTX Continuum 400 Series...

Documents

Transcript of FTX Continuum 400 Series (PA-7100) Technical Service Guide 400 (7100... · FTX Continuum 400 Series...