Bull Cedoc | DPS7000/XTA NOVASCALE 7000 CDA 7 5630 | Product manual | Bull Cedoc DPS7000/XTA NOVASCALE 7000 CDA 7 5630 Product manual

Product Manual
Hardware: CDA
REFERENCE
77 A1 69UP 00
DPS7000/XTA
NOVASCALE 7000
CDA 7 5630
DPS7000/XTA
NOVASCALE 7000
CDA 7 5630
Product Manual
Hardware: CDA
September 1999
BULL CEDOC
357 AVENUE PATTON
B.P.20845
49008 ANGERS CEDEX 01
FRANCE
REFERENCE
77 A1 69UP 00
The following copyright notice protects this book under Copyright laws which prohibit such actions as, but not
limited to, copying, distributing, modifying, and making derivative woks.
Copyright
Bull SAS 1999
Printed in France
Suggestions and criticisms concerning the form, content, and presentation of this
book are invited. A form is provided at the end of this book for this purpose.
To order additional copies of this book or other Bull Technical Publications, you
are invited to use the Ordering Form also provided at the end of this book.
Trademarks and Acknowledgements
We acknowledge the right of proprietors of trademarks mentioned in this book.
Intel® and Itanium® are registered trademarks of Intel Corporation.
Windows® and Microsoft® software are registered trademarks of Microsoft Corporation.
UNIX® is a registered trademark in the United States of America and other countries licensed exclusively through
the Open Group.
Linux® is a registered trademark of Linus Torvalds.
The information in this document is subject to change without notice. Bull will not be liable for errors contained
herein, or for incidental or consequential damages in connection with the use of this material.
Preface
Scope and
Objectives
The CDA 7 subsystem is a very reliable high performance Integrated Cached Disk
Array (ICDA) designed for online data storage. The CDA 7 houses both controller
and data storage capabilities in a single cabinet.
As part of its effort to continuously improve and enhance the performance and
capabilities of the Bull product line, Bull from time to time releases new revisions
of CDA 7 5630 hardware and microcode. Therefore, some functions described in
this manual may not be supported by all revision of CDA 7 microcode or hardware
presently in use.
The document concerns the CDA 7 5630 product, which will be called CDA 7 in
the rest of the document.
Where to Get Help
To obtain technical support for your CDA 7 5630, call your local Bull Competence
Center.
Intended
Readers
This manual is part of the documentation set for the CDA 7 5630 product.
This manual describes the CDA 7 5630 features and operations.
CAUTION:
The CDA 7 5630 contains no user-serviceable parts, so it should not be
opened for any reason by untrained persons. If the CDA 7 is in need of
repair, only qualified personnel familiar with safety procedures for
electrical equipment and the CDA 7 should access components inside the
unit.
This manual is intended for the storage administrator, system programmer, or
operator who is involved in acquiring, managing, or operating the CDA 7
subsystem.
77 A1 69UP Rev00
iii
CDA 7 5630 Product Manual
Prerequisites
For GCOS 7 systems, the CDA 7 subsystem is available with two disk technologies
giving two different disk capacities: 18 GB and 36 GB.
18 GB refers to the formatted disk capacity. This is equivalent to an unformatted
disk capacity of 23 GB.
36 GB refers to the formatted disk capacity. This is equivalent to an unformatted
disk capacity of 47 GB.
Structure
iv
This manual is structured as follows:
Chapter 1
Introducing CDA 7 provides an overview of the
CDA 7 5630, highlighting its performance and
reliability features, and describes hardware and
software options for the unit.
Chapter 2
CDA 7 Hardware introduces the hardware
components of the CDA 7 5630. It describes its main
components, the function of the operator panel, and the
types of host channels and devices to which CDA 7
can attach.
Chapter 3
CDA 7 Input/Output Operations discusses
integrated cached disk arrays, I/O operation, and cache
management.
Chapter 4
Performance Features describes the CDA 7
performance features, how they will affect overall
performance, and how to use these features to get the
best performance from CDA 7.
Chapter 5
Managing Critical Data discusses the CDA 7 features
that affect data availability and reliability.
Appendix A
Power Sequences provides step-by-step instructions
for powering the CDA 7 5630 on and off.
Appendix B
EPS Installation Requirements covers the tasks you
need to perform when planning or verifying the
physical configuration of CDA 7 in your system or
creating I/O addressing schemes.
Glossary
defines terms used in this manual.
Index
contains an index to help you access the information of
interest directly.
77 A1 69UP Rev00
Preface
Bibliography
Other EMC publications include:
Symmetrix Enterprise Storage Platform Product Guide
...................................................................... P/N 200-999-556, EMC Corporation
Symmetrix Open Systems Host Environment Product Guide, Volumes I and II
...................................................................... P/N 200-999-563, EMC Corporation
Symmetrix Remote Data Facility Product GuideP/N 200-999-554, EMC Corporation
SRDF Host Component Product Guide........... P/N 200-999-561, EMC Corporation
Symmetrix Data Migration Facility Product Guide
...................................................................... P/N 200-999-559, EMC Corporation
Symmetrix MultiHost Transfer Facility Product Guide
...................................................................... P/N 200-999-567, EMC Corporation
Open Symmetrix Manager-Base Component, Product Guide
...................................................................... P/N 200-999-555, EMC Corporation
Open Symmetrix Manager-SRDF Component Product Guide
...................................................................... P/N 200-999-565, EMC Corporation
Symmetrix Manager Product Guide................ P/N 200-999-555, EMC Corporation
Symmetrix High Availability Environment Product Guide
...................................................................... P/N 200-999-566, EMC Corporation
Symmetrix Multiple Mirror Facility MVS Batch Utility Product Guide
...................................................................... P/N 200-999-574, EMC Corporation
Other Bull publications include:
DPS 7000 - Installation Guide - Vol 2 - Peripherals ............................. 77 A1 66US
Bull DPS 7000 - System Repair Manual ............................................... 77 A7 72US
CDA/X Product Manual ....................................................................... 96 A1 70EC
CDA 7 Maintenance Manual...............................................................77 A7 48UU
CDA 7 5630 Site Preparation............................................................... 77 A1 70UP
Syntax
Notation
Bull uses the following convention for CDA 7 5630 identification:
The CDA 7 5630 uses:
• 3.5-inch 10 000 RPM, disk devices having a formatted capacity of 18 GBytes,
• 3.5-inch 7200 RPM, having a formatted capacity of 36 GBytes
Bull uses the following conventions for notes, cautions, and warnings.
NOTE:
A note presents information that is important, but not hazard-related.
77 A1 69UP Rev00
v
CDA 7 5630 Product Manual
DANGER:
A danger warning contains information essential to avoid a hazard that can
cause severe personal injury, death, or substantial property damage if you ignore
the warning.
vi
77 A1 69UP Rev00
Table of Contents
1.
2.
Introducing CDA 7
1.1
Versions of the Product............................................................................................... 1-1
1.2
CDA 7 Overview ......................................................................................................... 1-2
1.2.1 CDA 7 Capacities........................................................................................... 1-3
1.2.2 Channel Connectivity and Host Integration ..................................................... 1-4
1.2.3 Performance Features.................................................................................... 1-5
1.2.4 Availability Features ....................................................................................... 1-5
1.2.5 Serviceability Features ................................................................................... 1-6
1.3
CDA 7 Options............................................................................................................ 1-7
1.3.1 Hardware Option ............................................................................................ 1-7
1.3.1.1 Phone Multiplexer........................................................................... 1-7
1.3.2 Software Options............................................................................................ 1-8
1.3.2.1 CDA 7 Enterprise Storage Platform (ESP) ...................................... 1-8
1.3.2.2 CDA 7 Remote Data Facility (SRDF) .............................................. 1-8
CDA 7 Hardware
2.1
Major Components...................................................................................................... 2-1
2.1.1 Component Location ...................................................................................... 2-3
2.1.2 CDA 7 Block Diagrams................................................................................... 2-5
2.2
Operator Panel ........................................................................................................... 2-6
2.2.1 Disk Director Display ...................................................................................... 2-6
2.2.2 Channel Director Display ................................................................................ 2-7
2.3
Disk Devices and SCSI Disk Emulation ....................................................................... 2-7
2.3.1 FBA Data and Command Format ................................................................... 2-8
2.3.2 Logical Volume Structure for GCOS 7 environment ........................................ 2-8
2.4
Directors and Cache ................................................................................................... 2-9
2.4.1 Ultra SCSI Channel Director........................................................................... 2-9
2.4.2 Fibre Channel Director ................................................................................... 2-9
77 A1 69UP Rev00
vii
CDA 7 5630 Product Manual
2.4.3
2.4.4
2.4.5
2.5
3.
4.
5.
Channel Attachments................................................................................................ 2-11
2.5.1 Ultra SCSI Channel Attachement ................................................................. 2-11
2.5.2 Fibre Channel Attachments .......................................................................... 2-12
CDA 7 Input/Output Operations
3.1
ICDA Operation........................................................................................................... 3-1
3.1.1 Cache Management ....................................................................................... 3-2
3.1.1.1 LRU Algorithm ................................................................................ 3-3
3.1.1.2 Prefetch Algorithm.......................................................................... 3-4
3.1.2 Caching Techniques....................................................................................... 3-4
3.2
Elements of a CDA 7 I/O Operation............................................................................. 3-6
3.2.1 I/O Response Time ........................................................................................ 3-6
3.2.2 Read Operations ............................................................................................ 3-9
3.2.2.1 Read Hit ....................................................................................... 3-10
3.2.2.2 Read Miss .................................................................................... 3-11
3.2.3 Write Operations .......................................................................................... 3-11
3.2.3.1 Fast Write..................................................................................... 3-13
3.2.3.2 Delayed Fast Write ....................................................................... 3-14
Performance Features
4.1
Cache ......................................................................................................................... 4-1
4.2
Multiple Channel Directors .......................................................................................... 4-1
4.3
Parallel Processing ..................................................................................................... 4-1
4.4
Fast Write Capabilities ................................................................................................ 4-2
4.5
Dynamic Mirror Service Policy..................................................................................... 4-2
4.6
RPS Miss Elimination.................................................................................................. 4-2
4.7
Channel Speed ........................................................................................................... 4-2
Managing Critical Data
5.1
viii
SRDF Remote Link Director ........................................................................... 2-9
Disk Director ................................................................................................ 2-10
Cache .......................................................................................................... 2-10
CDA 7 Data Management Overview............................................................................ 5-1
5.1.1 CDA 7 Reliability and Availability Features ..................................................... 5-1
5.1.2 CDA 7 Data Integrity Protection Features ....................................................... 5-2
5.1.3 Data Protection Options ................................................................................. 5-3
5.1.3.1 RAID-1 Option (Mission Critical/ Business Critical).......................... 5-3
5.1.3.2 CDA 7 Remote Data Facility (SRDF) .............................................. 5-3
77 A1 69UP Rev00
A.
B.
5.2
Reliability and Availability Features ............................................................................. 5-4
5.2.1 Reliable Components ..................................................................................... 5-4
5.2.2 Redundant Power Subsystem ........................................................................ 5-4
5.2.3 System Battery Backup .................................................................................. 5-4
5.2.3.1 CDA 7 Power Failure on GCOS 7 Channels Connected to
GCOS 7 ......................................................................................... 5-5
5.2.4 Dual-Initiator Feature...................................................................................... 5-5
5.2.5 Non-disruptive Component Replacement........................................................ 5-9
5.2.6 Microcode Upgrades and Loads ................................................................... 5-10
5.2.6.1 Non-disruptive Microcode Upgrades and Loads ............................ 5-10
5.2.6.2 Dynamic Reconfigurations ............................................................ 5-11
5.2.6.3 Online SCSI-to-Fibre Channel Migration ....................................... 5-11
5.3
Data Integrity Protection............................................................................................ 5-12
5.3.1 Error Checking, Correction, and Data Integrity Protection ............................. 5-12
5.3.1.1 Parity............................................................................................ 5-13
5.3.1.2 ECC ............................................................................................. 5-13
5.3.1.3 LRC.............................................................................................. 5-13
5.3.2 Disk Error Correction and Error Verification .................................................. 5-13
5.3.3 Cache Error Correction and Error Verification ............................................... 5-14
5.4
Data Protection Guidelines........................................................................................ 5-14
5.5
Mirroring ...................................................................................................................5-15
5.5.1 Write Operations with Mirroring .................................................................... 5-15
5.5.2 Read Operations with Mirroring .................................................................... 5-15
5.5.3 Error Recovery with Mirroring ....................................................................... 5-16
5.5.4 Mirroring Advantages ................................................................................... 5-16
5.6
SRDF for GCOS 7 .................................................................................................... 5-16
Power Sequences
A.1
Powering Up the CDA 7 ............................................................................................. A-1
A.2
Routinely Powering Down CDA .................................................................................. A-2
A.3
Emergency Powering Down ....................................................................................... A-2
A.4
Powering Up After an Emergency Shutdown .............................................................. A-3
ESP Installation Requirements
B.1
CDA 7 Hardware Checklist......................................................................................... B-1
B.2
Open Checklists for ESP Connection ......................................................................... B-2
B.3
SCSI Cable Worksheet .............................................................................................. B-4
B.4
ULTRA SCSI Channel Adapters................................................................................. B-5
77 A1 69UP Rev00
ix
CDA 7 5630 Product Manual
B.5
SCSI Cable Precautions............................................................................................. B-5
Glossary
Index
x
77 A1 69UP Rev00
Table of Graphics
Figures
1-1.
2-1.
2-2.
2-3.
2-4.
3-1.
3-2.
3-3.
3-4.
3-5.
3-6.
3-7.
3-8.
3-9.
3-10.
3-11.
3-12.
5-1.
5-2.
5-3.
B-1.
CDA 7 Cabinet ............................................................................................................ 1-2
CDA 7 (Interior View) .................................................................................................. 2-2
CDA 7 Block Diagram ................................................................................................. 2-5
CDA 7 Operator Panel ................................................................................................ 2-6
ULTRA SCSI Attachment .......................................................................................... 2-11
Host Cache Use.......................................................................................................... 3-2
LRU Data Flow ........................................................................................................... 3-3
CDA 7 Cache Use....................................................................................................... 3-4
I/O Response Time ..................................................................................................... 3-6
Types of CDA 7 I/O Operations................................................................................... 3-7
Destaging Operation ................................................................................................... 3-8
Read Operations......................................................................................................... 3-9
Read Hit.................................................................................................................... 3-10
Read Miss................................................................................................................. 3-11
Write Operations ....................................................................................................... 3-12
Fast Write ................................................................................................................. 3-13
Delayed Fast Write ................................................................................................... 3-14
CDA 7 Dual-Initiator .................................................................................................... 5-7
Data Record Format for Conventional Disk................................................................ 5-12
CDA 7 Data Record Format ...................................................................................... 5-12
Ultra SCSI Director Channel Designations................................................................. B-6
Tables
1-1.
2-1.
5-1.
B-1.
B-2.
B-3.
CDA 7 (RAID-1 Mirroring) Capacities .......................................................................... 1-3
Cylinders per Logical Volume/or Split Physical Devices............................................... 2-8
Data Protection Options ............................................................................................ 5-14
CDA 7 Host Checklist................................................................................................. B-1
Open System Server Host Checklist (1/2) .................................................................. B-2
Cable Worksheet ....................................................................................................... B-4
77 A1 69UP Rev00
xi
CDA 7 5630 Product Manual
xii
77 A1 69UP Rev00
Introducing CDA 7
1. Introducing CDA 7
This chapter provides an overview of the Bull CDA 7 and highlights its
performance and reliability features. This chapter also describes hardware and
software options.
• CDA 7 Overview
• Channel Configurations
• CDA 7 Options
1.1
Versions of the Product
For GCOS 7 systems, the CDA 7 subsystem is available with two disk technologies
giving two different disk capacities: 18 GB and 36 GB.
18 GB refers to the formatted disk capacity. This is equivalent to an unformatted
disk capacity of 23 GB.
36 GB refers to the formatted disk capacity. This is equivalent to an unformatted
disk capacity of 47 GB.
NOTE:
The rest of the document will refer to the18 GB and 36 GB versions.
In the GCOS 7 environment only 18 GB disks or 36 GB disks are used. In the
Open Systems environment, when the CDA 7 cabinet is shared either 18 GB disks
or 36 GB disks can be used.
77 A1 69UP Rev00
1-1
CDA 7 5630 Product Manual
1.2
CDA 7 Overview
CDA 7 is a reliable high performance Integrated Cached Disk Array (ICDA)
designed for on-line data storage. The following figure provides a front view of the
exterior of the CDA 7 model.
Figure 1-1.
1-2
CDA 7 Cabinet
77 A1 69UP Rev00
Introducing CDA 7
1.2.1
CDA 7 Capacities
The CDA 7 is available with various numbers of disk devices and storage
capacities.
CDA 7 can host both system and user data. In a GCOS 7 environment, diskmodules are 18 formatted capacity. The 18 GB physical disks have to be split into
two 9 GB logical disks. The larger capacity models are particularly suitable in a
shareable context with Open Systems.
Table 1-1 outlines the CDA 7 model capacities when the CDA 7 is attached to a
DPS 7000 system host. The capacities are presented based on the method of data
protection: RAID-1 Mirroring.
Table 1-1.
CDA 7 (RAID-1 Mirroring) Capacities
Total
Formatted
Capacity
of User
Disks
in GB
Unformatted
Capacity of
User Disks
in GB
Accessible
by GCOS 7
Formatted
Capacity
of User
Disks
in GB
Accessible
by GCOS 7
No of
Physical
Disks
No of
User
Disks
Cache
(MB)
Total
unformatted
Capacity of
User Disks
in GB
MSPD056-0000
4
2
1024
46
36
46
36
MSPD057-0000
4
2
2048
46
36
46
36
MSPD058-0000
4
2
4096
46
36
46
36
MSPD059-0000
12
6
1024
138
108
138
108
MSPD060-0000
12
6
2048
138
108
138
108
MSPD061-0000
12
6
4096
138
108
138
108
MSPD062-0000
20
10
2048
230
180
230
180
MSPD063-0000
28
14
2048
322
252
322
252
MSPD064-0000
32
16
2048
368
288
368
288
4
2
-
46
36
0 or 46 *
0 or 36 *
MSUD052-0000
4
2
-
46
36
0 or 46 *
0 or 36 *
MSUD053-0000
4
2
-
46
36
46
36
Marketing
Identifier
Cabinets
Disks options for
initial configurations
MSUD050-0000
Disks Add-Ons
* this value depends on the configuration which receives the Add-On.
77 A1 69UP Rev00
1-3
CDA 7 5630 Product Manual
1.2.2
Channel Connectivity and Host Integration
The CDA 7 supports connectivity
• To the Bull DPS 7000/GCOS 7 through Ultra SCSI channels,
• To Open Systems, when Enterprise Storage Platform (ESP) software is added,
through Ultra SCSI or Fibre channels.
The CDA 7 channel directors are available with the following ports:
• 4-port Ultra SCSI directors,
• 2-port Fibre Channel directors.
CDA 7 supports the Operating System Versions GCOS 7-V7 TS 7560, GCOS 7-V8
TS 8560, GCOS 7-V9 TS9662 and up.
NOTE:
When connected to CDA 7 on Ultra SCSI Channel Director via Ultra SCSI
Adapter, Host Systems can use Fast-Wide or Ultra Differential SCSI interfaces.
1-4
77 A1 69UP Rev00
Introducing CDA 7
1.2.3
Performance Features
CDA 7 offers improved performance over conventional Storage Control Units
(SCU). These CDA 7 features enhance performance and increase throughput:
• Large non-volatile cache
• Asynchronous I/O
• Multiple storage directors
• Parallel processing
• 100% Fast Write capabilities
• Rotation Position Sense (RPS) miss elimination
• Segmented device-level buffer
• Ultra SCSI channel speeds up to 40 MB/sec
• Fibre Channel speeds up to 100 MB/sec
1.2.4
Availability Features
CDA 7 maintains data integrity and maximizes system availability with these
features:
•
•
•
•
•
•
•
•
•
•
•
•
77 A1 69UP Rev00
Redundant architecture
Full system battery backup
Dual-initiator
Redundant power subsystem
Remote Data Facility (SRDF) option
Mirroring option
Non-disruptive component repair (under conditions)
Non-disruptive microcode upgrades (Limited Case)
Cache error correction and error verification
Disk error correction and error verification
Environmental Fault Reporting
Error checking, correction and data integrity protection
1-5
CDA 7 5630 Product Manual
1.2.5
Serviceability Features
Every CDA 7 unit has an integrated service processor that continuously monitors
the CDA 7 environment. The service processor communicates with the Bull
Competence Centers via a customer-supplied, direct phone line. It automatically
dials the Customer Support Center whenever CDA 7 detects a component failure or
environmental violation. A Bull Product Support Engineer at the Bull Competence
Center can also run diagnostics remotely via the service processor to determine the
source of a problem and potentially resolve it before the problem becomes critical.
CDA 7 has a modular design with a low parts count for quick component
replacement should a failure occur. This low parts count minimizes the number of
failure points.
The CDA 7 features non-disruptive replacement of its major components,
including:
• Channel director cards
• Disk director cards
• Cache cards
• Disk devices
• Cooling fan modules
• Communications cards
• Operator panel
• Power supplies
• Service processor
• Battery
1-6
77 A1 69UP Rev00
Introducing CDA 7
1.3
CDA 7 Options
Bull offers the following hardware and software options for the CDA 7. Consult
your Bull Systems Engineer for the latest information on these options.
1.3.1
Hardware Option
1.3.1.1
Phone Multiplexer
Bull offers a Phone Multiplexer for domestic customer sites with multiple CDA 7
units (Integrated Remote Adaptor). The CDA 7 Phone Multiplexer consists of a
PBX with a Direct Inward Systems Access (DISA) board set that switches
incoming calls to the correct CDA 7 unit. The Phone Multiplexer allows any mix
of CDA units to use a common telephone line to communicate with the Bull
Competence Center. Each CDA 7 unit has its own extension in the Phone
Multiplexer network.
77 A1 69UP Rev00
1-7
CDA 7 5630 Product Manual
1.3.2
Software Options
The following software options are used:
• Enterprise Storage Platform (ESP)
• Remote Data Facility (SRDF)
1.3.2.1
CDA 7 Enterprise Storage Platform (ESP)
CDA 7 ESP software enabler is a software option that allows simultaneous storage
and access of mainframe data and open systems data on the same CDA 7 system.
Data must reside on separate physical disks for each platform in the system. Data
from multiple system sources may also co-exist on the same CDA 7, both with
other open systems data and with mainframe data. Each homogenous operating
system environment stores its data on its own physical disks, separate from the
physical disks used for data storage by other operating systems.
To implement ESP in your environment, contact your Bull Sales Engineer.
For an example of HACMP in a Sagister/Escala environment, refer to the CDA/X
Product Manual.
NOTE:
Refer to the Symmetrix Enterprise Storage Platform Product Guide for more
information on CDA 7 ESP. For information on cluster systems, refer to the
Symmetrix High Availability Environment Product Guide.
1.3.2.2
CDA 7 Remote Data Facility (SRDF)
The CDA 7 SRDF option is a mechanism that maintains a mirror image of data at a
logical volume level in two CDA 7 subsystems that can be located in physically
separate sites.
Within the Bull DPS 7000/GCOS 7 environment, the SRDF for GCOS 7 solution is
aimed at providing a back-up solution in the situation where a production site is out
of work due to a major disaster event. Refer to § SRDF for GCOS 7.
1-8
77 A1 69UP Rev00
CDA 7 Hardware
2. CDA 7 Hardware
This chapter describes the main hardware components of the CDA 7 including:
• Major Components
• Operator Panel
• Disk Devices and SCSI Disk Emulation
• Directors and Cache
• Channel Attachments
2.1
Major Components
The CDA 7 is a disk subsystem that houses all Storage Control Unit functions and
Disk in a single cabinet. This section describes:
• Component Location
• CDA 7 Block Diagrams
77 A1 69UP Rev00
2-1
CDA 7 5630 Product Manual
The following figures show the location of the main components of the CDA 7 .
DISK
DEVICES
FAN
ASSEMBLIES
SERVICE
PROCESSOR
ADAPTER
CARDS
REAR
Open Position
BATTERY
SUBSYSTEM
Closed
Position
DIRECTOR AND
CACHE CARDS
FRONT
POWER
SUBSYSTEM
INTERNAL
ETHERNET
HUB
Figure 2-1.
2-2
CDA 7 (Interior View)
77 A1 69UP Rev00
CDA 7 Hardware
2.1.1
Component Location
These components have the following functions:
Cooling Fan Modules
Contains fans for maintaining air circulation and cooling the unit internally.
Card Cage and Backplane
Contains eight slots to accommodate director, cache, and adapter cards. The
directors and cache cards connect to the front of the backplane. The adapter cards
connect to the rear of the backplane.
Disk Devices
Contain up to 32 x 3.5-inch disk-devices for data storage.
Bus and Tags Connector Panel
Connectors for bus and tag cables that serve as the external interface to mainframe
hosts.
Battery Subsystem
Maintains power for three minutes to the entire subsystem if AC power fails.
Power Subsystem
Two power supplies provide +5V, +12V, and +24V power to the CDA 7
components.
Ethernet Hub
The Ethernet board allows the CDA 7 to communicate with each disk director or
channel director. It is located just below the service processor.
77 A1 69UP Rev00
2-3
CDA 7 5630 Product Manual
Integrated Service Processor
Downloads the CDA 7 configuration to the directors and provides diagnostic and
maintenance utilities for CDA 7. It connects to the CDA 7 subsystem via an
RS-232 interface and uses an external modem for communicating with the Bull
Competence Center when CDA 7 detects an error condition.
Dual Power Cords
EMC offers dual power cords on the CDA 7. This allows you to connect the power
subsystem to two dedicated or isolated power lines.
2-4
77 A1 69UP Rev00
CDA 7 Hardware
2.1.2
CDA 7 Block Diagrams
The following figure illustrates the interconnection of the major components of the
CDA 7 .
A B (C D)
CHANNEL
CHANNEL
DIRECTORS DIRECTOR
(FRONT END)
TELEPHONE
SERVICE
PROCESSOR
MEMORY
M1
CHANNELS
A B (C D)
A B (C D)
A B (C D)
CHANNEL
DIRECTOR
CHANNEL
DIRECTOR
CHANNEL
DIRECTOR
MEMORY
M2
X-BUS
OPERATOR
PANEL
CACHE
Y-BUS
PRIMARY
AC
DISK
DIRECTOR
4
4
POWER
SUPPLY
BACKUP
BATTERY
DISKS
DISK
DIRECTOR
4
4
DISKS
EPC
REDUNDANT
FAN
SUBSYSTEM
DISK ARRAY
(BACK END)
DISKS
Figure 2-2.
77 A1 69UP Rev00
DISKS
CDA 7 Block Diagram
2-5
CDA 7 5630 Product Manual
2.2
Operator Panel
This section describes the functions of the various operator panel components.
The CDA 7 operator panel is located at the top of the front door. It has two
displays: one for the disk directors and one for the channel directors. The operator
panel shows the current activity of each channel director. The following figure
shows an example of the operator panel.
Active/Ready LEDS
POWER
A
a
DA 1
DIR3
DIR14
DIR 14
DIR16
DIR15
BA
b a
BA
b a
BA
b a
DIR 3
DA 2
EN
Ready
Active
DIR 15
B
b
DIR 16
RESET ENABLE
DIS
DISABLE
Disk Director
Enable/Disable Switches
Button A
Button B
Channel Director
Enable/Disable Switches
Figure 2-3.
2.2.1
CDA 7 Operator Panel
Disk Director Display
The disk director display is located on the left side of the operator panel.
The disk director display has an Enable/Disable switch for each disk director. This
switch places the disk director in an on-line or off-line state to the host system.
When a switch is in a disabled position, the host system sees all disk devices
physically connected to that disk director in a not ready (“Intervention Required”)
state the hex display on the disk director indicates “0F”.
2-6
77 A1 69UP Rev00
CDA 7 Hardware
2.2.2
Channel Director Display
The channel director display is located on the right side of the operator panel.
There is a separate display for each channel director.
The operator panel displays the current activity and status of each channel director.
The switches control physical channel on-line/off-line activity.
The channel director display components function as follows for the channel
directors:
POWER
Light indicating that the CDA 7 operator panel is powered on.
READY/ACTIVE LED
Light indicating channel interface state. When the Ready/Active LED is lit, that
channel interface is on-line to the host system. When the Ready/Active LED is
flashing, I/O operations are in progress. For remote link directors (SRDF), these
LEDs have no meaning.
ENABLE/DISABLE SWITCHES
Switch for placing a channel interface in the on-line or off-line state.
2.3
Disk Devices and SCSI Disk Emulation
The CDA 7 uses 3.5-inch disk devices, each having an unformatted storage
capacity of 23 GB (18 GB formatted). The CDA 7 can contain up to 32 diskdevices.
When using a SCSI interface, the CDA 7 system appears as standard SCSI disk
devices with data stored in fixed-block architecture (FBA) format.
The following paragraphs describe:
• the FBA Data and Command Format
• the Logical Volume Structure
77 A1 69UP Rev00
2-7
CDA 7 5630 Product Manual
2.3.1
FBA Data and Command Format
FBA disk devices store data in fixed sized blocks (typically 512 bytes). A disk
device using FBA format is viewed as a large array of blocks. The physical
position of the block (cylinder and track) is usually not significant to the host.
When requesting disk access for read or write, the host addresses the device by the
logical block address (LBA) of the starting block and a count of the total blocks
needed. CDA 7 SCSI channel directors and disk directors control access to cache
and disk devices, responding to host requests as a standard SCSI disk device.
2.3.2
Logical Volume Structure for GCOS 7 environment
The channel directors interact with cache. Therefore, there is no mechanical
meaning to cylinders, tracks, and heads on the CDA 7 logical volume from the
front end point of view.
However, CDA 7 uses a logical geometry definition for its logical volume
structure. This geometry is reflected in the SCSI mode sense data available to the
host.
CDA 7 uses the following logical volume structure:
• Each logical volume has N cylinders
• Each cylinder has 15 tracks (heads)
• Each track has 64 blocks of 512 bytes
Therefore, a CDA 7 logical volume with N cylinders has a usable block capacity
of:
N * 15 * 64
“N” for each volume is defined during CDA 7 configuration.
Table 2-1.
Cylinders per Logical Volume/or Split Physical Devices
Physical Disk Device Size
Unformatted capacity of 23 GB
Formatted capacity 18 GB (3.5")
2 Logical Volumes per Disk Device
18,414 Cylinders x 2
8.42 GB2 x 2 (9.05 GB10 x 2)
To calculate the size of the logical volume:
# of Cylinders * Heads * Blocks * 512
(N * 15 * 64 * 512)
2-8
77 A1 69UP Rev00
CDA 7 Hardware
2.4
Directors and Cache
This section describes the CDA 7 directors and cache capacity.
The channel directors and cache manage the storage control functions. The disk
directors handle the data storage functions.
2.4.1
Ultra SCSI Channel Director
The Ultra SCSI director is a single card that occupies one slot on the CDA 7
backplane. Each SCSI director interfaces to the host channels via a SCSI adapter
connected to the opposite side of the backplane.
The SCSI director has four differential wide interfaces for connection to host
systems and one high speed path to cache. The SCSI director contains two
advanced microprocessors that process commands and data from the host and
manage access to cache. They support data transfer rates up to 40 MB/sec with a
host over each channel interface simultaneously. On each SCSI channel director,
data transfers from the host to cache occur as four concurrent operations. The
CDA 7 can contain two or four SCSI directors.
2.4.2
Fibre Channel Director
The Fibre channel director is a single card that occupies one slot on the CDA 7
blackplane. Each Fibre Channel director interfaces to the host channels via a Fibre
Channel adapter connected to the opposite side of the backplane.
These adapters provide the connectivity between the host channel and the Fibre
Channel directors (FC-0 layer of the Fibre Channel standard). Each adpater
provides transceiver connections for two Fibre Channel ports.
The Fibre Channel director supports a data transfer rate up to 100MB/sec with a
host over each channel interface simultaneously.
The CDA 7 can contain two Fibre Channel directors.
2.4.3
SRDF Remote Link Director
The Remote Link Director (RLD) is a 4-port serial channel director microcode
configured as the link between CDA 7 units in a CDA Remote Data Facility
(SRDF) configuration. CDA requires two RLDs when used in the SRDF for
GCOS 7 solution.
77 A1 69UP Rev00
2-9
CDA 7 5630 Product Manual
NOTE:
For more information, call your Bull Sales Engineer.
2.4.4
Disk Director
The CDA 7 disk director provides an interface between cache and the disk devices.
The CDA 7 contains two disk directors. Each director, with its two advanced
microprocessors, supports up to sixteen 23 or 47 GB 3.5-inch disk devices— one to
four disk devices per SCSI bus (C, D, E, and F).
The CDA 7 supports the following:
Number of Disk Directors
Maximum Number of Disk Devices
Number of Buses per Disk Director
Maximum Number of Drives per Bus
Maximum Number of Drives per Director
CDA 7
2
32
4
4
16
Each disk director provides an alternate path to the disk devices of its disk director
pair. That is, should the primary path through a disk director to a disk device fail,
CDA 7 accesses that device through the other disk director in the pair via its
dual-initiator function. Refer to Dual-Initiator Feature later in this document for
more information on this data availability feature.
2.4.5
Cache
Memory is a crucial component of the CDA 7 system. All read and write
operations transfer data to or from cache. Any transfers between the host processor,
channel directors and cache are achieved at much greater electronic speeds than
transfers involving disks. CDA 7 optimizes data movement between disk and
cache, resulting in the highest performance possible.
Each cache memory card has two cache buses, x and y, with each having a 360 MB
bandwidth for a total processing bandwidth of 720 MB per second. The CDA 7 has
two slots dedicated to cache memory. The subsystem supports a maximum of 8,192
MB of cache. Individual cache memory cards are available in 512 MB, and 1 GB, 2
GB, and 4 GB sizes
2-10
77 A1 69UP Rev00
CDA 7 Hardware
2.5
Channel Attachments
2.5.1
Ultra SCSI Channel Attachement
CDA 7 can attach to Ultra SCSI channels and Fibre Channels, or a mix of channel
types. The physical connection to a CDA 7 SCSI channel interface occurs at the
connectors on the channel adapters.
The following figure illustrates the type of SCSI attachments.
DIFFERENTIAL
WIDE
SCSI HCST
ULTRA FAST-WIDE
SCSI
DIRECTOR
Figure 2-4.
ULTRA SCSI Attachment
When connecting to hosts with differential Ultra SCSI interface, use differential
P-cables that are designed for wide SCSI applications (see labeling on cable).
With the Ultra SCSI adapters, there are special connecting cables that have a bare
part that must be installed in the clamp grounding of the CDA. The switchable port
terminations must be set to ENABLE and the corresponding LED is on. The
TERMPOWER is set to VCC (the down position).
77 A1 69UP Rev00
2-11
CDA 7 5630 Product Manual
2.5.2
Fibre Channel Attachments
The CDA 7 Fibre Channel adapter provides an interface between the director and
host channels. Each Fibre Channel adapter is located at the rear of the backplane,
opposite its corresponding channel director. These adapters provide the
connectivity between the host channels and the Fibre Channel directors (FC-0 layer
of the Fibre Channel standard). Each adapter provides transceiver connections for
two Fibre Channel ports (four directors, eight Fibre Channel ports).
2-12
77 A1 69UP Rev00
3. CDA 7 Input/Output Operations
This chapter describes input/output operations between the CDA 7 and a host
system.
• Integrated Cached Disk Array Operation
• Elements of a CDA 7 I/O Operation
3.1
ICDA Operation
Integrated Cached Disk Array operation is based on the principle that the currently
active data at any given time is relatively small when compared to the total
subsystem storage capacity. If this data is in cache, there is a significant
improvement in performance. The success of an Integrated Cached Disk Array
subsystem is based upon satisfaction of the following two characteristics:
Locality of Reference
If a given piece of information is used, there is a high probability that a nearby
piece of information will be used shortly thereafter.
77 A1 69UP Rev00
3-1
CDA 7 5630 Product Manual
Data Reuse
If a given piece of information is used, there is a high probability that it will be
reused shortly thereafter.
This cache principle has been in use for years on host systems (CPU and storage
devices). The following figure illustrates this type of host cache use. The cache
used in this manner is often a high speed, high cost storage unit used as an
intermediary between the CPU and main storage.
HOST
CACHE
CPU
Figure 3-1.
3.1.1
MEMORY
Host Cache Use
Cache Management
In CDA 7, the channel directors and disk directors share cache. This cache
management is transparent to the host operating system. CDA 7 manages cache
using a Least Recently Used (LRU) algorithm and a Prefetch algorithm. CDA 7
dynamically switches between the LRU and Prefetch algorithms to maintain data in
cache based on the data access patterns.
3-2
77 A1 69UP Rev00
CDA 7 Input/Output Operations
3.1.1.1
LRU Algorithm
The directors replace the least recently used data in cache with new data when
needed. The LRU algorithm optimizes cache performance through high hit ratios
by ensuring that the most appropriate data is stored in cache. The following figure
illustrates data flow with the LRU algorithm.
NEWLY STAGED
CACHE SLOT
PROMOTED
TO TOP
READ or WRITE HIT
CACHE
CAPACITY
OLDEST SLOT
Figure 3-2.
LRU Data Flow
Each time a read hit or write hit occurs, CDA 7 marks that cache slot as most
recently used and promotes it to the top of the LRU list. For each write, a
written-to flag is set on the initial write to each cache block and is cleared when the
cache block is "destaged". At the bottom of the LRU list is the least recently used
cache slot.
77 A1 69UP Rev00
3-3
CDA 7 5630 Product Manual
3.1.1.2
Prefetch Algorithm
The Prefetch algorithm dynamically detects sequential data access patterns to the
disk devices. The directors improve the hit ratio of these accesses by promoting
blocks from the disk devices to cache slots before a read miss can occur.
The Prefetch algorithm can stage between a single block and a full track of data to
a cache slot. The LRU algorithm adds the cache slot to the top of the LRU list. To
make room for these tracks, CDA 7 removes slots at the bottom of the list from
cache.
3.1.2
Caching Techniques
CDA 7 uses the same cache principle as host systems, but with enhanced caching
techniques. The following figure illustrates cache use in CDA 7.
ICDA
DISKS
CACHE
DISK
DIRECTOR
CHANNEL
DIRECTOR
DIRECTORY
Figure 3-3.
CDA 7 Cache Use
CDA 7 directors handle either channel or disk operations. Channel directors attach
to the CPU channels as well as to cache. Disk directors attach to cache as well as
the disk drives. The CDA 7 directors perform the following functions:
• The channel director handles I/O requests from the host. It accesses the
directory in cache to determine if the request can be satisfied out of cache. The
directory contains information on each cache slot and the blocks within each
slot. It maintains data in cache based on the data access patterns.
• The disk director manages accesses to the disk drives. It performs a background
operation that "destages" “written-to” blocks to disk.
3-4
77 A1 69UP Rev00
CDA 7 Input/Output Operations
CDA 7 uses these techniques to enhance its performance:
Split Director Functions
The director operations are split into two functional
parts, the channel director and disk director. The
channel director services requests from the host. The
disk director services requests between cache and disk.
Splitting the director functions eliminates the
processing overhead and cache locking associated with
Control Units that perform both functions.
High Speed Cache Memory
Cache memory speed is greater than the total speed of
all components (e.g., the directors) that access it.
Disk Microprocessor and Buffer
Each disk device has its own microprocessor and
buffer. This intelligence is brought to the actuator
level providing parallel processing of data. These
features add another level of caching and improve
overall performance.
Multiple Disk Directors
A maximum of 16 disks (CDA 7 models) attach to
each disk director. For more information on director
and disk combinations, refer to Disk Director earlier in
this document.
Sequential Access Patterns
Access patterns can be sequential, random, or a
combination of both. When a miss occurs on a
sequential access pattern, the number of blocks
brought into cache is increased, thus improving the hit
rate (requested data is in cache).
77 A1 69UP Rev00
3-5
CDA 7 5630 Product Manual
3.2
Elements of a CDA 7 I/O Operation
All I/O operations require a certain response time. An I/O request begins with the
application issuing an I/O command and ends when the transfer completes. The
time lapse from I/O request to transfer completion is the I/O response time.
3.2.1
I/O Response Time
The I/O response time can be divided into a host queuing time, a command connect
time, a disconnect time, and a data connect time, as shown in the following figure.
HOST
QUEUING
TIME
COMMAND
CONNECT
TIME
DISCONNECT
TIME
DATA
CONNECT
TIME
DEVICE SERVICE TIME
I/O RESPONSE TIME
Figure 3-4.
I/O Response Time
The Host Queuing Time is the time the request is in the host queue before it is
dispatched on the SCSI bus.
The Command Connect Time is the length of time the channel is transferring a
SCSI command.
The Disconnect Time is the length of time involving device seek and latency. At
this time the SCSI bus can be used by other devices.
NOTE:
In case of a cache hit in an I/O request, the Disconnect Time requirement is
eliminated.
The Data Connect Time is the length of time the channel is transferring data.
There are four basic types of CDA 7 I/O operations:
• Read Hit
• Read Miss
• Fast Write
• Delayed Fast Write
3-6
77 A1 69UP Rev00
CDA 7 Input/Output Operations
1 Cache Directory Search - Hit
2 Transfer to Host
3 Update Directory
CHANNEL
1
2
CHANNEL
1
3
DIRECTORY
3
CHANNEL
DIRECTOR
CACHE
1 Cache Directory Search - Miss
2 Position R/W Head, Stage
Data to Cache
3 Transfer to Host
4 Update Directory
DIRECTORY
4
CHANNEL
DIRECTOR
CACHE
DISKS
DISK
DIRECTOR
DISKS
2
DISK
DIRECTOR
READ HIT
1 Cache Directory Search - Hit
2 Transfer to Cache
3 Update Directory
4 Asynchronous Destage
CHANNEL
2
READ MISS
CHANNEL
4
1
DIRECTORY
3
CHANNEL
DIRECTOR
CACHE
1
CHANNEL
DIRECTOR
FAST WRITE
Figure 3-5.
77 A1 69UP Rev00
CACHE
3
DISK
DIRECTOR
4
DIRECTORY
5
DISKS
DISK
DIRECTOR
1 Cache Directory Search Cache is Full
2 Destage Page
3 Update Cache/Directory
4 Transfer to Cache
5 Update Directory
6 Asynchronous Destage
DISKS
2, 6
DELAYED FAST WRITE
Types of CDA 7 I/O Operations
3-7
CDA 7 5630 Product Manual
In addition to the four types of I/O operation mentioned, CDA 7 performs a
background operation that "destages" "written-to" blocks back to disk. The
following figure illustrates this "destaging" operation.
CHANNEL
1 Destage Block(s)
2 Up date Directory
DIRECTORY
CHANNEL
DIRECTOR
CACHE
1
2
DISKS
DISK
DIRECTOR
Figure 3-6.
Destaging Operation
CDA 7 performs read operations from cache. CDA 7 always caches write
operations. This cache operation is transparent to the host operating system. A
read operation causes the channel director to scan the cache directory for the
requested data. If the requested data is in cache, the channel director transfers this
data immediately to the channel with a channel end and device end (or a SCSI
good ending status). If the requested data is not in cache, the disk director transfers
the data from the disk device to the cache and the channel director transfers the
requested data from the cache to the channel.
3-8
77 A1 69UP Rev00
CDA 7 Input/Output Operations
3.2.2
Read Operations
There are two types of read operations: read hit and read miss. The following
figure illustrates the data flow for read operations.
CHANNEL
CHANNEL
CHANNEL
DIRECTOR
CHANNEL
DIRECTOR
CACHE
CACHE
READ HIT
DISKS
READ MISS
Figure 3-7.
77 A1 69UP Rev00
Read Operations
3-9
CDA 7 5630 Product Manual
3.2.2.1
Read Hit
In a read hit operation (following figure), the requested data resides in cache. It is
transferred directly to the host computer through the channel interface.
CONNECT TIME
OVERHEAD
TOTAL SERVICE TIME
Figure 3-8.
Read Hit
Since the data is in cache, there are no mechanical delays due to seek, latency, and
RPS miss.
3-10
77 A1 69UP Rev00
CDA 7 Input/Output Operations
3.2.2.2
Read Miss
In a read miss operation, the requested data is not in cache and must be retrieved
from a disk device. While the channel director creates space in the cache, the disk
director reads the data from the disk device. The disk director stores the data in
cache as the channel director simultaneously transfers the cached data to the host.
If the requested data is in the process of being pre-fetched (sequential read ahead),
the miss is considered to be a "short miss." If the requested data is not in the
process of being fetched, it goes on a list to be fetched and the miss is considered to
be a "long miss."
CONNECT TIME
OVERHEAD
DISCONNECT TIME
TOTAL SERVICE TIME
Figure 3-9.
Read Miss
Since the data is not in cache, CDA 7 must search for the data on disk and then
transfer it to the channel. This adds seek and latency times to the operation.
During the disconnect time, other commands can be executed on other devices on
the bus, or commands can queue to the same device.
3.2.3
Write Operations
For write operations, the channel director sends a device end and channel end (or a
SCSI good ending status) to the channel as soon as the data is in cache and verified
as error-free so the host system can continue processing without having to wait for
CDA 7 to write the data to disk. The disk director asynchronously "destages" the
data to disk.
There are two types of write operations:
• Fast Write
• Delayed Fast Write
77 A1 69UP Rev00
3-11
CDA 7 5630 Product Manual
The following figure illustrates the data flow for write operations.
CHANNEL
CHANNEL
DIRECTOR
CACHE
ASYNCHRONOUS
DESTAGE
DISKS
READ WRITE
Figure 3-10.
3-12
Write Operations
77 A1 69UP Rev00
CDA 7 Input/Output Operations
3.2.3.1
Fast Write
In a fast write operation, the channel director stores the data in cache and sends a
channel end and device end (or a SCSI "good ending" status) to the host computer.
The disk director then asynchronously "destages" the data from cache to the disk
device.
CONNECT TIME
OVERHEAD
TOTAL SERVICE TIME
Figure 3-11.
Fast Write
Since CDA 7 writes the data directly to cache and not to disk, there are no
mechanical delays due to seek, latency, and RPS miss.
77 A1 69UP Rev00
3-13
CDA 7 5630 Product Manual
3.2.3.2
Delayed Fast Write
In a delayed fast write operation, CDA 7 cannot store the data in cache until space
in cache becomes available. A delayed fast write condition can only happen if the
cache allocated for writes is currently full with data yet to be "destaged" to the disk
devices. The default cache write ceiling is 80% of the cache.
CONNECT TIME
OVERHEAD
DISCONNECT TIME
TOTAL SERVICE TIME
DELAY
Figure 3-12.
NORMAL
FAST WRITE
Delayed Fast Write
Should this situation occur, CDA 7 disconnects the channel directors from the
channels and waits for the disk directors to "destage" the Fast Write data in cache
to disk. When space in cache is available, CDA 7 reconnects the channel directors
to their channels and completes the write operation. CDA 7 continues to process
read operations during delayed fast writes.
3-14
77 A1 69UP Rev00
4. Performance Features
CDA 7 offers improved performance over conventional Storage Control Units and
disk device designs. The CDA 7 features described below allow high cache hit
ratios and less processing overhead, reducing response time and improving
throughput.
4.1
Cache
A cache size of up to 8,192 MB and intelligent caching algorithms greatly improve
hit ratios and overall subsystem response time. CDA 7 caches all read and write
operations making them transparent to the host operating system.
4.2
Multiple Channel Directors
CDA 7 contains multiple channel directors, Ultra SCSI or fibre channel type, each
supplying an independent path to cache or disk from the host system. The
CDA 7 has two or four channel directors.
Each Ultra SCSI director supports up to four SCSI channels.
Each fibre channel director supports two Fibre channel ports.
4.3
Parallel Processing
Each channel director and disk director has two resident microprocessors, and each
disk device has one resident microprocessor. These microprocessors make use of
advances in parallel processing to reduce processing time and improve throughput.
77 A1 69UP Rev00
4-1
CDA 7 5630 Product Manual
4.4
Fast Write Capabilities
CDA 7 caches write operations, eliminating the need to write data to the disk
immediately. This results in faster response times and improved overall subsystem
performance. The write ceiling is 80% of the usable cache. Channel directors and
disk directors dynamically allocate cache space between reads and writes
depending on I/O activity.
4.5
Dynamic Mirror Service Policy
The CDA 7 Dynamic Mirror Service Policy (DMSP) is an enhancement to the
adaptive algorithms in the CDA 7 architecture that improves the performance of
read operations in mirrored environments. The improved system performance is a
result of CDA 7 balancing the load between physical disk drives and disk directors,
and minimizing actuator movement.
To achieve this improved performance, CDA 7 measures and tracks I/O activities
of logical volumes, physical volumes and disk directors. Then, based on these
measurements, CDA 7 directs read operations for mirrored data to the appropriate
mirror that results in the best overall performance of the CDA 7 . As the access
patterns and workloads change, the dynamic algorithm analyzes the new workloads
and adjusts the service policies as needed.
4.6
RPS Miss Elimination
In CDA 7, each disk device has a dedicated microprocessor and segmented data
buffer that can temporarily store data until the disk director is ready to read or write
data. This eliminates rotational positional sensing (RPS) misses that occur in
conventional disk device when the heads are positioned over the desired sector, but
the channel path is not ready for read or write operations. The segmented data
buffer of the disk device allows multiple operations to occur to the head/disk
assemblies.
4.7
Channel Speed
CDA 7 Ultra SCSI channels transfer data at speeds up to 40 MB/sec. The data
transfer rate is host dependent. CDA 7 supports cable lengths of up to 82 feet (25
m) to attach to most host systems.
4-2
77 A1 69UP Rev00
5. Managing Critical Data
This chapter discusses the CDA 7 features and options that affect data availability
and reliability.
• CDA 7 Data Management Overview
• Reliability and Availability Features
• Data Integrity Protection
• Data Protection Guidelines
• Mirroring
5.1
CDA 7 Data Management Overview
CDA 7 has many features and options to ensure a high degree of system and data
availability.
Many of these features and options are built into the CDA 7 design. Other
availability options may be purchased separately and implemented in the CDA 7
operation.
5.1.1
CDA 7 Reliability and Availability Features
The CDA 7 design offers the following reliability and availability features:
•
•
•
•
•
77 A1 69UP Rev00
High reliability components
Redundant power subsystem
System battery backup
Dual-Initiator feature
Non-disruptive maintenance and microcode upgrades
5-1
CDA 7 5630 Product Manual
These basic CDA 7 features provide protection against loss of system and data
availability due to a power loss or failed component. A redundant design allows
CDA 7 to remain online and operational during component repair. For example, if
a power supply fails, the remaining power supplies share the load until the failed
component is replaced. The system battery backup prevents any loss of data due to
a power failure.
The Dual-Initiator feature offers data availability protection against a CDA 7 disk
management component failure. With dual-initiator, each member of a disk
director pair shadows the functions of the other disk director. That is, each disk
director can service any or all of the devices attached to the disk director with
which it is paired. This feature does not, however, provide data availability in the
event of a disk device failure. Should CDA 7 detect a disk management hardware
failure, CDA 7 automatically reads from or writes to the disk devices it was unable
to communicate with via the other disk director in the pair.
5.1.2
CDA 7 Data Integrity Protection Features
The CDA 7 is designed with these data integrity features:
• Error checking, correction, and data integrity protection
• Disk error correction and error verification
• Cache error correction and error verification
• Periodic system checks
Error verification prevents temporary errors from accumulating and resulting in
permanent data loss. CDA 7 also looks at the error verification frequency as a
signal of a potentially failing component.
The periodic system check tests all components as well as microcode integrity.
CDA 7 reports errors and environmental conditions to the host system as well as
the Bull Competence Center.
5-2
77 A1 69UP Rev00
Managing Critical Data
5.1.3
Data Protection Options
Although the CDA 7 has standard features that provide a higher level of data
availability than conventional disk, the options listed below ensure an even greater
level of data recoverability and availability.
You can choose from the following CDA 7 data protection options to match your
critical data requirements:
• RAID-1 Mirroring
• CDA 7 Remote Data Facility (SRDF)
5.1.3.1
RAID-1 Option (Mission Critical/ Business Critical)
Mirroring provides the highest performance, availability, and functionality for all
mission critical and business critical applications. With the Mirroring option,
CDA 7 maintains two identical copies of a logical volume on separate disk devices.
Should CDA 7 be unable to read data from one volume of a mirrored pair, it
immediately retrieves the data from the other logical volume.
5.1.3.2
CDA 7 Remote Data Facility (SRDF)
The CDA 7 SRDF option is a mechanism that maintains a mirror image of data at a
logical volume level in two CDA 7 subsystems that can be located in physically
separate sites.
Within the Bull DPS 7000/GCOS 7 environment, the SRDF for GCOS 7 solution is
aimed at providing a back-up solution in the situation where a production site is out
of work due to a major disaster event. Refer to § SRDF for GCOS 7.
77 A1 69UP Rev00
5-3
CDA 7 5630 Product Manual
5.2
Reliability and Availability Features
CDA 7 has several features that allow it to maintain data integrity and maximize
system availability. This section discusses several features in detail.
5.2.1
Reliable Components
CDA 7 uses components that have a mean time between failure (MTBF) of several
hundred thousand to millions of hours for a minimal component failure rate. A
redundant design allows CDA 7 to remain online and operational during
component repair.
A periodic system check tests all components as well as microcode integrity.
CDA 7 reports errors and environmental conditions to the host system as well as
the Bull Competence Center.
5.2.2
Redundant Power Subsystem
The CDA 7 has a modular power subsystem featuring a redundant architecture that
facilitates field replacement of any of its components without any interruption in
processing.
The CDA 7 power subsystem connects to two dedicated or isolated AC power
lines. If AC power fails on one AC line, the power subsystem automatically
switches to the other AC line.
The two power supplies operate in a redundant parallel configuration. If a power
supply module fails, the remaining power supply continue to share the load. CDA 7
senses the fault and reports it as an environmental error (error code 0472).
5.2.3
System Battery Backup
The entire CDA 7 system is made nonvolatile via an onboard battery backup
system. The battery backup system provides the means for destaging any fast write
data that might be in cache if AC power is lost on both AC power lines to the unit.
In addition to providing nonvolatility to the CDA 7 system, the batteries are fully
capable of powering not only all electronic components, but also all HDAs during
this time. This eliminates emergency power off situations, meaning that the disks
are always powered down in an orderly manner, which extends their useful life
considerably.
5-4
77 A1 69UP Rev00
Managing Critical Data
The backup battery subsystem allows CDA 7 to remain online to the host system
for three minutes in the event of an AC power loss. This three minute window
allows CDA 7 to support frequent power outages because the battery is able to
fully recharge during Initial Microcode Load (IML) time. CDA 7 continually
recharges the battery subsystem whenever it is under AC power.
The battery backup also prevents disk device failures due to the sharp power drops
that occur during unexpected power interrupts. If power is restored before the
battery timer expires, CDA 7 becomes available again without IML.
5.2.3.1
CDA 7 Power Failure on GCOS 7 Channels Connected to GCOS 7
When a power failure occurs, power switches immediately to the backup battery
and CDA 7 continues to operate normally. When the battery timer window elapses,
CDA 7 presents a busy status to prevent the host system from writing or reading
any data at the unit. CDA 7 destages any Fast Write data still in cache to disk,
spins down the disk devices and retracts the heads, and powers down, turning off
the battery at that time. CDA 7 will not respond to SCSI selections after it powers
down.
5.2.4
Dual-Initiator Feature
CDA 7 has a dual-initiator feature that ensures continuous availability of data in
the unlikely event of a CDA 7 disk management hardware failure. It does not
provide data availability in the event of a disk device failure. This feature works
by having two disk directors “shadow” the function of the other. That is, this
feature gives each disk director the capability of servicing any or all of the disk
devices of the disk director with which it is paired when one disk director is
completely servicing its own and the shadowed devices.
With dual-initiator, the disk directors work in pairs. See the figures on the
following pages. Under normal conditions, each disk director services its disk
devices. If the sophisticated fencing mechanisms of CDA 7 detect a disk
management hardware failure, CDA 7 notifies the host and the Bull Competence
Center of the failure, and reads from or writes to the disk devices that its disk
director pair was unable to communicate with without interruption. When the
source of the failure is corrected, CDA 7 returns the I/O servicing of the two disk
directors to their normal state.
77 A1 69UP Rev00
5-5
CDA 7 5630 Product Manual
This sophisticated fencing mechanism determines a disk management hardware
failure at two levels:
5-6
Level 1:
a director discovers through use of its online
self-testing that one of its SCSI paths is failing. The
director signals CDA 7 of this condition and the
director it is paired with automatically takes control of
the failing path.
Level 2:
CDA 7 determines a director has failed and fences it
out. The director’s associated pair automatically takes
control of all its devices.
77 A1 69UP Rev00
Managing Critical Data
TO/FROM
MEMORY
DISK DIRECTOR 1
DISK DIRECTOR 2
PROCESSOR
A
PROCESSOR
B
PROCESSOR
A
PROCESSOR
B
PC PD SC SD
PC PD SC SD
PC PD SC SD
PC PD SC SD
DISKS
DISKS
DISKS
DISKS
DISKS
DISKS
DISKS
DISKS
PC = PRIMARY C BUS
PD = PRIMARY D BUS
SC = SECONDARY C BUS
SD = SECONDARY D BUS
Figure 5-1.
77 A1 69UP Rev00
= PRIMARY SCSI PATHS TO DISKS
= SECONDARY SCSI PATHS TO DISKS
CDA 7 Dual-Initiator
5-7
CDA 7 5630 Product Manual
In the following examples, disk director 1 and disk director 2 are paired. Each
CDA 7 disk director services 16 disk devices.
EXAMPLE 1:
Disk director 1 fails. Disk director 2 automatically performs any I/O operations
with all of the devices normally serviced by disk director 1 as well as all its own
disk devices with no interruption in processing.
❑
EXAMPLE 2:
On disk director 1, processor A, primary SCSI bus C fails. On disk director 2,
processor B, primary SCSI bus D fails. The devices of these failing buses are
serviced as follows:
• The devices of failing bus C, are serviced by disk director 2, processor A,
secondary SCSI bus C.
• The devices of failing bus D are serviced by disk director 1, processor B,
secondary SCSI bus D.
The functioning primary buses on each of the disk directors continue to service
their respective devices.
❑
EXAMPLE 3:
On disk director 2, processor A, primary SCSI buses C and D fail. The devices of
these failing buses are serviced as follows:
• The devices of failing bus C are serviced by disk director 1, processor A,
secondary SCSI bus C.
• The devices of failing bus D are serviced by disk director 1, processor A,
secondary SCSI bus D.
The functioning primary buses of disk director 2 continue to service their
respective devices.
5-8
77 A1 69UP Rev00
Managing Critical Data
In summary, dual-initiator provides the following advantages:
• Ensures continuous data availability if a CDA 7 disk management component
fails
• Provides continuous operation by switching data pathing to the alternate disk
director without interruption when a communications failure occurs with one or
more disk devices
• Re-establishes normal data pathing after repair of the defective component
Dual-initiator provides an additional level of data availability in mirrored
configurations. If CDA 7 is unable to read from or write to one of the devices in a
mirrored pair, CDA 7 automatically uses the other disk device in the pair without
interruption. If CDA 7 fails to communicate with that device also, CDA 7 then
attempts to access the volume through the alternate path provided by the dualinitiator function.
❑
5.2.5
Non-disruptive Component Replacement
CDA 7, with its redundant architecture, supports non-disruptive replacement of
many of its components. This includes online replacement of the directors
(channel directors and disk directors), memory cards, disk devices, adapters
(dual-initiator, serial), power supplies, communications cards, and cooling fans.
This non-disruptive replacement capability allows the Bull Customer Engineer to
install a new component, initialize it if necessary, and bring it online without:
•
•
•
•
•
Disrupting access to unaffected volumes
Powering down the CDA 7 unit
Stopping the operating system
Taking unaffected channel paths offline
Taking devices offline (other than the affected device)
This replacement must be done under controlled conditions, please contact your
Bull Competence Center. For a demonstration of non-disruptive component
replacement, contact your Bull Customer Service representative.
NOTE:
When replacing a Director, all activity on the director must be stopped before
attempting this operation.
77 A1 69UP Rev00
5-9
CDA 7 5630 Product Manual
5.2.6
Microcode Upgrades and Loads
Microcode upgrades, performed by the Product Support Engineers (PSE) at the
Bull Competence Center, allow you to take advantage of enhancements to
performance algorithms, error recovery and reporting techniques, diagnostics, and
microcode fixes.
It is possible to upgrade the microcode from one level to an adjacent level.
Problems can happen when upgrading to a much higher microcode level.
5.2.6.1
Non-disruptive Microcode Upgrades and Loads
Non-disruptive microcode upgrades from one version to the next and interim
microcode updates (loads) are available for CDA 7 systems. Symmetrix takes
advantage of its multiprocessing and redundant architecture to allow for hot
loadability of similar microcode platforms. Release levels can be non-disruptively
loaded without interruption to user access.
During a non-disruptive microcode upgrade, the Product Support Engineer
downloads the new microcode to the service processor. The new microcode loads
into the EEPROM areas within the channel and disk directors and remains idle
until requested for hot load in control storage. The CDA 7 system does not require
manual intervention on the customer’s part to perform this function. All channel
and disk directors remain in an online state to the host processor, thus maintaining
application access. CDA 7 will load executable code at selected "windows of
opportunity" within each director hardware resource until all directors have been
loaded.
Once the executable code is loaded, internal processing is synchronized and the
new code becomes operational. This capability can be utilized to upgrade or to
back down from a release level or interim update.
NOTE:
During a non-disruptive microcode load within a code family, the full
microcode is loaded, which consists of the same base code plus additional
patches.
5-10
77 A1 69UP Rev00
Managing Critical Data
5.2.6.2
Dynamic Reconfigurations
CDA 7 supports dynamic reconfiguration activity without disruption to online
applications, such as:
• Establish/de-establish mirrored pairs.
• Modify channel assignments.
5.2.6.3
Online SCSI-to-Fibre Channel Migration
Beginning with CDA 7 microcode revision 5265, CDA 7 systems with SCSI
channel directors can be upgraded to Fibre Channel directors without taking nonSCSI channels offline and without requiring a backup and restore of data. This
capability allows customers with SCSI channels to take advantage of the
connectivity and distance features offered with Fibre Channel directors. Bull
Customer Engineers use a new utility to perform the migration.
NOTE:
For more information on Fibre Channel migration, consult your Bull sales
representative.
77 A1 69UP Rev00
5-11
CDA 7 5630 Product Manual
5.3
Data Integrity Protection
CDA 7 preserves data integrity by performing extensive error checking and
correction on all data and addresses it passes internally.
5.3.1
Error Checking, Correction, and Data Integrity Protection
In conventional disk, the subsystem adds error checking and correction bytes to
each data record field, as shown in the following figure. It uses these error
checking and correction bytes to check the data and correct it if possible. If it
detects an uncorrectable error, the disk subsystem informs the host that it has
encountered bad data to avoid affecting data integrity.
ECC
BYTES
DATA RECORD
Figure 5-2.
Data Record Format for Conventional Disk
CDA 7, like conventional disk, performs this level of error checking and correction
when it passes data and addresses. CDA 7, however, goes further to ensure that the
information passed belongs to the record specified. It does this by including
additional bytes with the data field of each record. These bytes contain the record
ID and a double LRC (Longitude Redundancy Code) check byte as shown in the
following figure. CDA 7 uses these bytes to check that the data is from the
specified record and alarms the host if it is not. This second level of protection
further ensures data integrity by preventing incorrect data from being transferred.
DATA RECORD
Figure 5-3.
EMBEDDED
ID
LRC
BYTES
ECC
BYTES
CDA 7 Data Record Format
CDA 7 has three levels of error detection. Should an error be undetected at one
level, it will be detected at one of the other levels.
CDA 7 uses the following error correction and detection methods:
• Parity
• Error Checking and Correction (ECC)
• Longitude Redundancy Code (LRC)
5-12
77 A1 69UP Rev00
Managing Critical Data
5.3.1.1
Parity
All data and control paths have parity generating and checking circuitry that verify
hardware integrity at the byte level.
5.3.1.2
ECC
The directors detect and correct single-bit and double-bit errors and report
uncorrectable 3-bit or more errors in cache.
5.3.1.3
LRC
The LRC calculation further assures data integrity. The check bytes are the XOR
(exclusive OR) of the accumulated bytes. Each record in memory also includes its
LRC byte, its physical memory address, and block number.
5.3.2
Disk Error Correction and Error Verification
The disk directors use idle time to read data and check the polynomial correction
bits for validity. If a disk read error occurs, the disk director reads all data on that
track to CDA 7 cache memory. The disk director writes several worst case patterns
to that track searching for media errors. When the test completes, the disk director
rewrites the data from cache to the disk device, verifying the write operation. The
disk microprocessor maps around any bad block (or blocks) detected during the
worst case write operation, thus skipping defects in the media. If necessary, the
disk microprocessor can reallocate up to 32 blocks of data on that track. To further
safeguard the data, each disk device has several spare cylinders available. If the
number of bad blocks per track exceeds 32 blocks, the disk director rewrites the
data to an available spare cylinder. This entire process is called “error
verification.”
The disk director increments a soft error counter with each bad block detected.
When the internal soft error threshold is reached, the CDA 7 service processor
automatically dials the Bull Competence Center and notifies the host system of
errors via sense data. This feature maximizes data availability by diagnosing
marginal media errors before data becomes unreadable.
77 A1 69UP Rev00
5-13
CDA 7 5630 Product Manual
5.3.3
Cache Error Correction and Error Verification
The disk directors use idle time to periodically read cache, correct single-bit errors
(one hard and one soft), and write the corrected data back to cache. This process is
called “error verification.” When the directors detect an uncorrectable error in
cache, CDA 7 reads the data from disk and takes the defective cache memory block
offline until an Bull Customer Engineer can repair it. Error verification maximizes
data availability by significantly reducing the probability of encountering an
uncorrectable error by preventing bit errors from accumulating in cache.
5.4
Data Protection Guidelines
The CDA 7 data protection options ensure a higher level of data protection,
recoverability, and availability than the standard CDA 7 availability and reliability
features. The following options can be purchased separately and implemented into
the CDA 7 operation:
• Mirroring
• CDA 7 Remote Data Facility (SRDF)
The following table describes the CDA 7 data protection options. You can choose
one or more of these options to match your critical data protection requirements.
Table 5-1.
Data Protection Options
Data Protection Option
Mirroring (RAID-1)
CDA 7 Remote Data
Facility (SRDF)
5-14
Description
Provides the highest level of performance and
availability for all mission critical and business
critical applications by maintaining a duplicate
copy of a volume on two disk devices.
For more information, refer to Mirroring.
Provides a disaster recovery solution by
maintaining a mirror image of data in two
CDA 7 systems which can be in physically
separate locations.
For more information, refer to SRDF.
77 A1 69UP Rev00
Managing Critical Data
5.5
Mirroring
Mirroring provides the highest level of performance and availability for all mission
critical and business critical applications. Mirroring maintains a duplicate copy of
a logical volume on two physical disk devices. CDA 7 maintains these copies
internally by writing all modified data to both devices. The mirroring operation is
transparent to the host.
The mirroring feature designates two logical volumes residing on different physical
devices as a mirrored pair, one volume being mirror-1 and the other volume being
mirror-2. The host views the mirror-1 and mirror-2 volumes as the same logical
volume because each has the same unit address.
5.5.1
Write Operations with Mirroring
CDA 7 handles a write operation to a mirrored logical volume as a normal write
operation. The channel director presents channel end and device end (or a good
ending status) to the channel after data is written to and verified in cache. The disk
directors destage the data to each logical volume in the mirrored pair maintaining
the identical copies of data.
5.5.2
Read Operations with Mirroring
During read operations, if the data is not available in cache, CDA 7 reads the data
from the disk pointed to by its performance algorithm for best system performance.
This performance algorithm tracks path busy information as well as actuator
location and what sector is currently under the disk head in each device. If a data
check occurs on the device being read, CDA 7 automatically reads the data from
the other device.
• Interleave Service Policy shares the read operations of the mirrored pair by
reading tracks from both disk devices in an alternating method, a number of
tracks from M1, and a number of tracks from M2. Interleave is designed to
achieve maximum throughput.
• Split Service Policy differs from Interleave because read operations are
assigned to either the M1 or the M2, but not to both.
• Dynamic Mirror Service Policy (DMSP) utilizes both Interleave and Split for
maximum throughput and minimal head movement. DMSP adjusts each logical
volume dynamically based on access patterns detected.
77 A1 69UP Rev00
5-15
CDA 7 5630 Product Manual
5.5.3
Error Recovery with Mirroring
Should one volume in the mirrored pair fail, CDA 7 automatically uses the other
volume without interruption. CDA 7 notifies the Bull Competence Center of this
condition. When the defective device containing the volume is replaced, CDA 7
re-establishes the mirrored pair and automatically copies data to the new disk.
During the copy operation, CDA 7 gives priority to host I/O requests over copying
to minimize the effect on performance. All new writes take place to both devices.
The time it takes to resynchronize the mirrored pair depends on the I/O activity to
the volume, the disk device, and the disk capacity.
5.5.4
Mirroring Advantages
In summary, mirroring offers the following advantages:
• Improved performance over traditional RAID 1 by supporting 100 percent fast
write, and two simultaneous internal data transfer paths.
• Protection of mission-critical data from any single point of failure.
• Continuous business operation by switching to the alternate disk device of a
mirrored pair without interruption to data availability should loss of access occur
to one of the disk devices in a mirrored pair.
• Assurance that the second copy of data is identical to the first copy.
• Automatic resynchronization of the mirrored pair after repair of the defective
volume.
5.6
SRDF for GCOS 7
The SRDF for GCOS 7 Solution is aimed at providing a complete business
continuance capability during the unlikely event of a data center disaster.
SRDF for GCOS 7 is based on CDA 7 SRDF mechanism allowing for a physical
remote copy of disk volumes and operating
• in campus mode with ESCON-compliant fiber optics link or high speed data
lines,
• in synchronous mode for a full logical synchronisation of mirrored volumes,
• with the two involved CDA 7’s configured in RAID-1.
5-16
77 A1 69UP Rev00
Managing Critical Data
In case of disaster occuring at the Production Site, the back-up procedures entitle
the Back-up Site to take over the application workload and to resume the
production tasks thanks to the remote copy of user’s databases and system data.
SRDF for GCOS 7
• is a very powerful solution with simple back-up procedures which are
transparent for user’s applications,
• avoids the need of time-consuming restore operations,
• is capable to quickly resume the business operations with minimal disruption.
SRDF for GCOS 7 is commercialized in a Project Mode. Contact your Bull Sales
Engineer.
Production
Site
Bull DPS 7000
Back-up
Site
TDS
BATCH
IOF
Bull DPS 7000
Local
R
L
D
CDA 7 5630 Raid-1
DEVELOPMENT,
TEST…
Remote
SRDF
ESCON
R
L
D
CDA 7 5630 Raid-1
RLD = Remote Link Director
77 A1 69UP Rev00
5-17
CDA 7 5630 Product Manual
❑
5-18
77 A1 69UP Rev00
A. Power Sequences
This Appendix provides step-by-step instructions for powering the CDA 7 on and
off. The unit is powered on and off via the power switches on the rear door of the
unit.
•
•
•
•
A.1
Powering Up the CDA 7
Routinely Powering Down CDA
Emergency Shutdown
Powering Up After an Emergency Shutdown
Powering Up the CDA 7
DANGER:
If the CDA 7 is in need of repair, only qualified personnel familiar with
safety procedures for electrical equipment and the CDA 7 should access
components inside the unit.
Perform these steps to power on the CDA 7 after it has been powered off by the AC
power switch.
1.
Make sure all operator panel switches (channel director and disk director) are
in the disable position.
2.
Lift the EPO switch on the rear door to the up (|) position.
3.
Lift the AC power switch on the rear door to the up (|) position. The CDA 7
begins its IML procedure. This procedure takes several minutes to perform.
NOTE:
The power controls are inoperative unless the Emergency Power Off (EPO)
switch on the rear door is in its enabled (up, | position). The EPO switch when
turned off (down, O position) disables all power to CDA 7 including the backup
battery systems.
77 A1 69UP Rev00
A-1
CDA 7 Product Manual
4.
When the IML procedure completes (all directors show "0F" in their LED
displays on their front panels), place the operator panel switches (channel
director and disk director) in their enable position.
When "0F" clears from all director LED displays, the CDA 7 is then available
to the host system(s).
A.2
Routinely Powering Down CDA
Perform these steps to power off the CDA 7 using the AC power switch.
1.
Stop all processes to the CDA 7.
2.
Watch the operator panel and wait for all "Active" LEDs to stop flashing.
NOTE:
The displays stop flashing when the CDA 7 has written all cache data to disk.
3.
Press the AC power switch on the rear door to the down (O) position.
NOTE:
The battery backup system automatically turns on when the CDA 7 detects loss
of AC power. The battery will keep the CDA 7 powered for 3 minutes following
which CDA 7 turns off.
A.3
Emergency Powering Down
If it becomes necessary to power off the CDA 7 immediately:
• Press the red EPO switch on the rear door to the down (O) position.
NOTE:
CDA 7 immediately switches to backup battery power and destages any pending
writes in cache to disk. This destage operation takes approximately 20 seconds
following which CDA 7 powers down.
1
3
3
IMPORTANT:
When there is an emergency power cut, data in the cache might not be
saved on the physical disks.
A-2
77 A1 69UP Rev00
Power Sequences
A.4
Powering Up After an Emergency Shutdown
CAUTION:
This unit contains no user-serviceable parts, so it should not be opened for
any reason by untrained personnel. If the CDA 7 is in need of repair, only
qualified personnel familiar with safety procedures for electrical equipment
and the CDA 7 should access components inside the unit.
When the CDA 7 is powered down via the red EPO switch, a breaker switch on the
power subsystem trips.
Perform these steps to power on the CDA 7 after it has been powered off by the red
EPO switch.
CAUTION:
Always contact the Bull Customer Support Center prior to powering up the
CDA 7 after an emergency shutdown.
1.
Make sure all operator panel switches (channel director and disk director) are
in the disable position.
2.
Lift the EPO switch on the rear door to the up (|) position.
3.
Lift the AC power switch on the rear door to the up (|) position.
NOTE:
The CDA 7 begins its IML procedure. This procedure takes several minutes to
perform.
4.
When the IML procedure completes, all directors show "0F" in their LED
displays.
CAUTION:
Before you enable channel and disk directors from the CDA 7 operator
panel, contact an EMC Product Support Engineer from the EMC
Customer Support Center. Explain that an emergency shutdown was
performed and that you need them to check the condition of the CDA 7
before you enable the channel and disk directors. When the CDA 7 checks
out correctly, proceed to the next step.
5.
Place the operator panel switches (channel director and disk director) in their
enable position.
When "0F" clears from all director LED displays, the CDA 7 is then available
to the host system(s).
77 A1 69UP Rev00
A-3
CDA 7 Product Manual
❑
A-4
77 A1 69UP Rev00
B. ESP Installation Requirements
This Appendix covers the tasks you need to perform when planning or verifying
the physical configuration of CDA 7 in your system or creating I/O addressing
schemes.
B.1
CDA 7 Hardware Checklist
Make sure you discuss with and obtain the following site profile information with
the Customer Engineer and/or Systems Engineer. This information is necessary for
each CDA 7 subsystem you are installing (following table).
Table B-1.
CDA 7 Host Checklist
CDA 7 Model
Total number of physical drives to be configured on the
CDA 7
Physical drive type (size)
Total amount of CDA 7 cache
Number of SCSI channel or Fibre channel directors
Ultra channel director model used
Number of Ultra channels used per director
Fibre channel director model used
Number of fibre port used per director
77 A1 69UP Rev00
B-1
CDA 7 5630 Product Manual
B.2
Open Checklists for ESP Connection
Make sure you discuss with and provide the following host information to the
Customer Engineer and/or Systems Engineer. This information is necessary for
each host you are attaching to the CDA 7 subsystem. Make a copy of this form for
each host you will attach to CDA 7. When you are done you should have a
checklist for each host.
Table B-2.
Open System Server Host Checklist (1/2)
Host Configuration Requirements
Host (CPU) vendor and model number
Host controller type and model number
Memory capacity of host
O/S revision level of host
I/O rate per second expected per host
SCSI adapter used on host, type and model
number
Is this a clustered environment? Which one?
Will devices be shared? Which ones?
Total number of SCSI channels per host, to
which SA ports will they attach?
Total number of Fibre channels per host, to
which FA ports will they attach ?
Specify if any narrow channels used
Number of logical devices needed per SCSI bus
Size of volumes required to be visible to host1
Total customer usable data storage required
Will host-level mirroring be used, which
volumes?
Will CDA 7 Dynamic Sparing (specify number
of spares) be used, which volumes?
Will CDA 7 RAID-1 be used, which volumes?
Will SRDF be used, which volumes?
Data storage utilization per host2
Average transfer size of data
Using raw devices or filesystems?
Size of filesystem
B-2
Host 1
Host 2
77 A1 69UP Rev00
ESP Installation Requirements
Table B-2.
Open System Server Host Checklist (2/2)
Host Configuration Requirements
Will data stripping be used?
What type of data stripping package?
Partitioning?
Partition sizes?
LVM used3
What major applications are to be run?
Database used:
Oracle®/Sysbase®/Informix®/other?
Size of database
Database release version
Supply typical high level database schema and
queries
Any patches or modifications related to
IO/SCSI?
Additional comments
Host 1
Host 2
NOTES:
77 A1 69UP Rev00
1.
This is the volume size needed for each volume visible on the SCSI bus.
2.
Percentage of available CDA data capacity used by that host.
3.
Special attention is required when using a Logical Volume Manager
(LVM) and/or data stripping when using hyper-volumes.
B-3
CDA 7 5630 Product Manual
B.3
SCSI Cable Worksheet
The physical connection to a Ultra channel interface occurs at the connectors on
the SCSI adapters located at the rear of the CDA 7 unit.
When connecting to hosts with differential Ultra or differential narrow SCSI
interfaces, use differential P-cables that are designed for wide SCSI applications
(see labeling on cable).
NOTE:
For a list of the Ultra SCSI cables available from Bull, refer to the manual
CDA 7 Site Preparation. You can use the following table as a cable worksheet
form.
Table B-3.
SCSI
Director
1A
1A
1B
1B
2A
2A
2B
2B
3A
3A
3B
3B
4A
4A
4B
4B
B-4
Cable Worksheet
SCSI
Director
Port
a
b
a
b
a
b
a
b
a
b
a
b
a
b
a
b
Cable
Length
Required
Cable Model
Number or Part
Number
Host ID Model
Number
77 A1 69UP Rev00
ESP Installation Requirements
B.4
ULTRA SCSI Channel Adapters
Several Fast-Wide / Ultra SCSI adapters are available in CDA 7, depending on the
host channel connection requirements, such as terminated and non-terminated
channels, and Y-cable cluster configurations.
NOTE:
Contact your Bull Sales Representative to obtain SCSI adapter part numbers.
For cluster configurations, refer to the Symmetrix High Availability Environment
Product Guide.
B.5
SCSI Cable Precautions
When connecting CDA 7 Ultra SCSI channels to host channels it is important to
know CDA 7 Ultra SCSI hardware components and their channel designations.
Each Ultra SCSI director occupies one slot on the CDA 7 backplane. Each director
interfaces to the host channels via a SCSI adapter connected to the opposite side of
the backplane.
The Ultra SCSI director contains two advanced microprocessors. Each
microprocessor has two wide-differential SCSI channels. The following figure
illustrates this interface and the channel designations.
NOTE:
When connecting the CDA 7 SCSI cables to the host channels, notice that the
"A" processor channels are on the bottom and the "B" processor channels are on
the top.
77 A1 69UP Rev00
B-5
CDA 7 5630 Product Manual
Four-Port Dual-Processor, Ultra
SCSI Director
SDA
Operator Panel
Designations
C
Physical
Channel
Designations
SA-1, SIDE B, PORT A
A
B-SIDE
B
D
SA-1, SIDE B, PORT B
HOST
CHANNELS A
SA-1, SIDE A, PORT A
B
X-BUS
A
SA-1, SIDE A, PORT B
B
ADAPTER
Figure B-1.
MIDPLANE
A
B
A-SIDE
DIRECTOR
Ultra SCSI Director Channel Designations
Additional Precautions
Observe the following precautions when connecting SCSI cables to the CDA 7
SCSI channel interface connectors:
• Use the shortest cabling possible between the CDA 7 unit and the host.
• Always check for bent pins on the SCSI cables before connecting them. The
SCSI cable pins bend very easily.
• Always attach SCSI cables to their connectors straight on to minimize the
chance of bending pins.
• The pin shroud on the SCSI cable ends deforms easily. Take care not to drop
cable ends or bang them against objects.
• When attaching the SCSI cable to the connector on the SCSI adapter in the CDA
unit, remember to fasten the screws on the connector to ensure a secure
connection.
• When connecting to hosts with differential Fast-Wide SCSI interfaces, do not
allow the cable length to exceed 82 feet (25 m).
• When connecting to hosts with Ultra SCSI interfaces, do not allow the cable
length to exceed 62 feet (19 m).
• Fibre Channel interfaces extend connectivity up to a maximum of 1600 feet (500
meters).
B-6
77 A1 69UP Rev00
Glossary
This glossary contains terms related to disk storage subsystems. Many of these
terms are used in this manual.
A
Actuator
A set of access arms and their attached read/write heads, which move as an independent component
within a head and disk assembly (HDA).
Adapter
Card that provides the physical interface between the director and disk devices (dual-initiator
adapter), director and parallel channels (Bus & Tag adapter), director, serial channels (serial
adapter), and SCSI host channels (SCSI adapter).
ADT
Automatic Diagnostic Test.
Alternate Track
A track designated to contain data in place of a defective primary track. See also Primary Track.
ANSI
American National Standards Institute. A standards-setting, non-government organization which
develops and publishes standards for voluntary use in the USA.
Asynchronous Transmission
A handshaking protocol that requires each byte be requested and acknowledged before the
transmission of the next byte begins.
77 A1 69UP Rev00
G-1
CDA 7 5630 Product Manual
B
Backplane
Card that accommodates the director, cache, and adapter cards.
Bit
The smallest unit of computer memory. A bit can hold a value of 0 or 1.
Byte
Term that indicates an 8-bit hexadecimal construction.
C
Cache
Random access electronic storage used to retain frequently used data for faster access by the
channel.
Cache Slot
Unit of cache equivalent to one track.
Channel Director
The component in the Symmetrix subsystem that interfaces between the host channels and data
storage. It transfers data between the channel and cache.
Command Descriptor Block (CDB)
The structure used to communicate commands from an initiator to a target. This structure may be 6
bytes, 10 bytes, or 12 bytes in size.
D
Data Availability
Access to any and all user data by the application.
Delayed Fast Write
There is no room in cache for the data presented by the write operation. Therefore, the write is
delayed until there is room in cache.
Destage
The asynchronous write of new or updated data from cache to disk device. See also the term
"Stage".
Device
A uniquely addressable part of the Symmetrix subsystem that consists of a set of access arms, the
associated disk surfaces, and the electronic circuitry required to locate, read, and write data. See
also Volume.
Device Number
The value that logically identifies a disk device in a string.
G-2
77 A1 69UP Rev00
Glossary
Diagnostics
System level tests or firmware designed to inspect, detect, and correct failing components. These
tests are comprehensive and self-invoking.
Differential Interface
An electrical signal configuration that uses a pair of lines for transfer. This configuration (as
compared to a single-ended interface) has a higher tolerance for common-mode noise and little
crosstalk when used with twisted pair cables. It supports cables lengths to 25 meters (82 feet). See
also Single-ended Interface.
Directors
The components in the Symmetrix subsystem that allows Symmetrix to transfer data between the
host channels and disk devices. See also Channel Director and Disk Director.
Disk Director
The component in the Symmetrix subsystem that interfaces between cache and the disk devices.
DMA
Direct Memory Access.
Dual-Initiator
A Symmetrix feature that automatically creates a backup data path to the disk devices serviced
directly by a disk director, if that disk director or the disk management hardware for those devices
fails.
Dynamic Path Reconnect (DPR)
A function that allows disconnected I/O operations with Symmetrix to reconnect over any available
channel path rather than be limited to the one on which the I/O operation was started.
Dynamic Sparing
A Symmetrix feature that automatically transfers data from a failing disk device to an available
spare disk device without affecting data availability. This feature supports all devices in the
Symmetrix subsystem and is used with the Mirroring and SRDF options.
E
EPO
Emergency Power Off.
Error Verification
The process of reading, checking the error correction bits, and writing corrected data back to the
source.
ESP
Enterprise Storage Platform. Symmetrix ESP is a functional enhancement that allows simultaneous
storage and access of mainframe data and open systems data on the same Symmetrix 53xx system.
77 A1 69UP Rev00
G-3
CDA 7 5630 Product Manual
F
Fast Write
In Symmetrix, a write operation at cache speed that does not require immediate transfer of data to
disk. The data is written directly to cache and is available for later destaging.
FBA
Fixed Block Architecture. Disk device data storage format using fixed size data blocks.
Fibre channel
Fibre channel speeds up to 100 MB/sec.
Field Replaceable Unit (FRU)
A component that is replaced or added by service personnel as a single entity.
G
Gigabyte (GB)
109 bytes.
The GB2 value is based on the convention: 1 GB = 1024x1024x1024 bytes.
The GB10 value is based on the convention: 1 GB = 1000x1000x1000 bytes.
H
HACMP
High Availability Clustered Multiprocessing.
Head and Disk Assembly (HDA)
A field replaceable unit in the Symmetrix subsystem containing the disk and actuator.
I
ICDA
Integrated Cached Disk Array.
ICKDSF
See Device Support Facilities program.
Identifier (ID)
A sequence of bits or characters that identifies a program, device, controller, or system.
IML
Initial microcode program loading.
Index Marker
Indicates the physical beginning and end of a track.
G-4
77 A1 69UP Rev00
Glossary
Index Point
The reference point on a disk surface that determines the start of a track.
IOCP
Input/Output Configuration Program.
I/O Device
An addressable input/output unit, such as a disk device.
IPL
Initial Program Loading.
K
Kilobyte (K)
1024 bytes.
L
Least Recently Used Algorithm (LRU)
The algorithm used to identify and make available the cache space by removing the least recently
used data.
Logical Unit
A physical or virtual device addressable through a target. A physical device can have more than
one logical unit.
Logical Unit Number (LUN)
An encoded three-bit identifier for the logical unit.
Logical Volume
A user-defined storage device. In the Symmetrix subsystem, the user can define a physical disk
device as one to eight logical volumes.
Logical Volume Manager (LVM)
The part of the AIX system responsible for disk subsystems. Interaction with the LBM is done via
SMIT menus or via stand-alone commands.
Long Miss
Requested data is not in cache and is not in the process of being fetched.
Longitude Redundancy Code (LRC)
Exclusive OR (XOR) of the accumulated bytes in the data record.
77 A1 69UP Rev00
G-5
CDA 7 5630 Product Manual
M
Media
The disk surface on which data is stored.
Megabyte (MB)
The MB2 value is based on the convention: 1 MB = 1024x1024 bytes.
The MB10 value is based on the convention: 1 MB = 1000x1000 bytes.
Mirrored Pair
A logical volume with all data recorded twice, once on each of two different physical devices.
Mirroring
The Symmetrix option that maintains two identical copies of a designated volume on separate
disks. Each volume automatically updates during a write operation. If one disk device fails,
Symmetrix automatically uses the other disk device.
P
Physical Partition (PP)
A physical partition is the smallest unit of disk space that can be allocated in a volume group in an
AIX environment. Any disk space allocated is an integral number of physical partitions. By
default, a PP is 4 MB in size.
Primary Track
The original track on which data is stored. See also Alternate Track.
Promotion
The process of moving data from a track on the disk device to cache slot.
R
Read Hit
Data requested by the read operation is in cache.
Read Miss
Data requested by the read operation is not in cache.
Reconnect
The function that occurs when a target selects an initiator to continue an operation after a
disconnect.
Reconnection
A reconnection exists from the assertion of the BSY signal in a RESELECTION phase until the
next BUS FREE phase occurs. A reconnection can only occur between a target and an initiator.
G-6
77 A1 69UP Rev00
Glossary
Record Zero
The first record after the home address.
Reselect
The function that occurs when a target disconnects from an initiator in order to perform a timeconsuming function and, then after performing that function, reestablishes the connection.
Reserved
The term used for bits, bytes, fields, and code values that are set aside for future standardization.
S
SCSI
Small Computer System Interface.
SCSI Adapter
A circuit board used to interface the SCSI bus with an internal bus. Used in this manual to identify
the Symmetrix SCSI Adapter that interfaces the SCSI bus from the host(s) to the Symmetrix SCSI
Channel Director. Symmetrix SCSI Adapters, located in the rear card cages, have four SCSI
connectors for attaching cables from the hosts(s).
SCSI Address
The octal representation of the unique address (0-7) assigned to a SCSI device. This address would
normally be assigned and set in the SCSI device during system installation.
SCSI Device
A host computer adapter or peripheral controller or intelligent peripheral that can be attached to the
SCSI bus.
SCSI ID
The bit significant representation of the SCSI address referring to one of the signal lines DB(7-0).
Short Miss
Requested data is not in cache, but is in the process of being fetched.
Single-ended Interface
An electrical signal configuration that uses a single line for each signal and references a ground
path common to the other signal lines. The single-ended configuration is susceptible to common
mode noise and has a maximum cable length of 6 meters. See also Differential Interface.
SIO
Start I/O.
SRDF
Symmetrix Remote Data Facility. SRDF consists of the microcode and hardware required to
support Symmetrix remote mirroring.
77 A1 69UP Rev00
G-7
CDA 7 5630 Product Manual
SRDF for GCOS 7
This is a disaster recovery solution that maintains a mirror image of data in two CDA 7 5330-23
subsystems that are located in physically separate sites.
Stage
The process of writing data from a disk device to cache. See also the term "Destage".
Storage Control Unit (SCU)
The component in the Symmetrix subsystem that connects Symmetrix to the host channels. It
performs channel commands and communicates with the disk directors and cache. See also
Channel Director.
Striping
The process of segmenting logically sequential data and writing the segments to multiple physical
disk devices.
Synchronous Transmission
A timing protocol that uses a master clock with a clock period and allowable offset that holds the
sending and receiving devices in the desired phase relationship.
T
Target
A SCSI device that performs an operation requested by an initiator.
Terabyte
The TB2 value is based on the convention: 1 TB = 1024x1024x1024x1024 bytes.
The TB10 value is based on the convention: 1 TB = 1000x1000x1000x1000 bytes.
U
ULTRA SCSI
Ultra SCSI (Small Computer Systems Interface). The transmission rate is 20 MB/s in 8 bit format
and 40 MB/s in 16 bit format.
UNIX
UNIX is an interactive, multi-tasking, multi-user operating system. UNIX is written in “C”
language. There are three types of UNIX files: directories, data files, and special files. A directory
is a file containing certain information about another file. A directory contained within another
directory is a sub-directory. The two most common types of UNIX are BSD (Berkeley Software
Distribution) and System VR4 (developed by AT&T). Most UNIX systems are a “mix” of both
types.
G-8
77 A1 69UP Rev00
Glossary
V
Volume
A general term referring to a storage device. In the Symmetrix subsystem, a volume corresponds to
single disk device.
W
WSP
Wide SCSI Processor.
77 A1 69UP Rev00
G-9
CDA 7 5630 Product Manual
G-10
77 A1 69UP Rev00
Index
A
Active LED 2-7
Actuator
Definition G-1
Adapter
Definition G-1
ADT
Definition G-1
Alternate Track
Definition G-1
ANSI
Definition G-1
Asynchronous Transmission
Definition G-1
Attachments
SCSI Channel 2-11
Availability
Dual-initiator 5-2
Features 1-5, 5-1, 5-4
Non-disruptive Component Replacement
5-9
Redundant Power Subsystem 5-4
Upgrades and Loads 5-10
B
Backplane 2-3
Definition G-2
Battery
Usage in Power Failure
Battery Subsystem 2-3
Bit
Definition G-2
77 A1 69UP Rev00
5-5
Bull Competence Center 1-6, 5-2, 5-4
Bus and Tags Connector Panel 2-3
Business Critical Data 5-3
Byte
Definition G-2
C
Cache
Definition G-2
Error Correction and Verification 5-14
Fast Write Ceiling 3-14
Host Cache Use 3-2
LRU Algorithm 3-3
Management 3-2
Performance Features 4-1
Prefetch Algorithm 3-4
Read Operation 3-9
Slots 2-10
Techniques 3-4
Cache Slot
Definition G-2
Capacities
CDA 7 1-3
Card Cage 2-3
CDA 7
Availability Features 1-5
Cabinet 1-2
Capacities 1-3
Channel Configurations 1-4
Channel Interfaces B-5
Data Integrity Protection 5-2, 5-12
Data Management Overview 5-1
Enterprise Storage Platform (ESP) 1-8
ESP 1-8
i-1
CDA 7 5630 Product Manual
ESP Connection B-2
Hardware Checklist B-1
Major Components 2-1
Mirroring 5-16
Overview 1-2
Performance Features 1-5, 4-1
Phone Multiplexer 1-7
Powering Down A-2
Powering Up A-1, A-3
Reliability and Availability 5-1, 5-4
SCSI Cable Worksheet B-4
Serviceability Features 1-6
Channel
CDA 7 Configurations 1-4
Channel and Device End 3-11
Connection Precautions B-5
Director Display 2-7
Interfaces B-5
Power Failure (Open System) 5-5
SCSI Adapters B-5
SCSI Attachments 2-11
SCSI Director 2-9
Speed 4-2
SRDF Remote Link Director 2-9
Channel Director
Definition G-2
Error Verification 5-14
Function 3-4
Command Connect Time 3-6
Command Descriptor
Definition G-2
Cooling Module 2-3
Correcting Errors 5-12
D
Data Availability
Definition G-2
Data Connect Time 3-6
Data Integrity
Cache Error Correction 5-14
Cache Error Verification 5-14
Periodic System Check Tests 5-2
Protection Features 5-2, 5-12
Protection Guidelines 5-14
Data Management
i-2
Overview 5-1
Data Protection
Mirroring Option 5-3, 5-16
Options 5-3
Data Transfer Rates
SCSI 2-9
Delayed Fast Write
Definition G-2
Operation 3-14
Destaging
Definition G-2
Operation 3-11
Device
Definition G-2
Device Number
Definition G-2
Diagnostics
Definition G-3
Differential
Adapter B-5
Definition G-3
Directors
Definition G-3
Disk 2-10
SCSI Channel 2-9
SRDF Remote Link 2-9
Disconnect Time 3-6
Disk Device 2-3, 2-7
RPS Miss Elimination 3-11
Disk Director 2-10
Definition G-3
Dual-Initiator Operation 5-6
Function 3-4
Operator Panel 2-6
Disk Emulation 2-7
Disk Error Correction 5-13
Disk Error Verification 5-13
DMA
Definition G-3
Dual Power Cords 2-4
Dual-Initiator
Advantages 5-9
Availability Features 5-2
Definition G-3
Description 5-5
Failure Levels 5-6
77 A1 69UP Rev00
Index
Fencing Mechanism 5-6
With Mirrored Configurations 5-9
Dynamic
Path Reconnection G-3
Reconfiguration 5-11
Sparing G-3
Dynamic Mirror Service Policy 4-2
E
ECC 5-13
Emergency
Power Off Procedure A-2
Powering Up After Shutdown A-3
Enable/Disable Switch 2-7
Enterprise Storage Platform (ESP) 1-8
EPO
Definition G-3
EREP File 5-14
Error
Checking 5-12
Checking Cache 5-14
Checking LRC 5-13
Checking Parity 5-13
Correcting Cache 5-14
Correcting Parity 5-13
Correction 5-12
Detection 5-12
ECC 5-13
Verification G-3
Error Verification 5-2
ESP 1-8
Checklist B-2
Definition G-3
Ethernet Hub 2-3
External Modem 2-4
F
Fast Write
Capabilities 4-2
Ceiling 3-14, 4-2
Definition G-4
Operation 3-13
FBA
77 A1 69UP Rev00
Data 2-8
Definition G-4
Fencing Mechanism 5-6
fibre channel
Definition G-4
Fibre Channel Attachments 2-12
Fibre Channel Director 2-9
Field Replaceable Unit
Definition G-4
FRU
Definition G-4
G
GB
Definition G-4
Guidelines
Data Protection 5-14
H
HACMP
Definition G-4
Hardware
Checklist B-1
Hardware Options
Phone Multiplexer
HDA
Definition G-4
1-7
I
I/O
Command Connect Time 3-6
Data Connect Time 3-6
Delayed Fast Write Operation 3-14
Device G-5
Disconnect Time 3-6
Fast Write Operation 3-13
Read Miss Operation 3-11
Response Queuing Time 3-6
Response Time 3-6
ICDA
Definition G-4
ICKDSF
i-3
CDA 7 5630 Product Manual
Definition
G-4
ID
Definition G-4
IML
Definition G-4
Index Marker
Definition G-4
Index Point
Definition G-5
Integrated Cached Disk Array 3-1
IOCP
Definition G-5
IPL
Definition G-5
L
LEDs
Power 2-7
Ready/Active 2-7
Loads
Microcode 5-10
Logical
Unit G-5
Unit Number G-5
Volume G-5
Volume Manager G-5
Long Miss
Definition G-5
LRC
Definition G-5
Description 5-13
Usage 5-12
LRU
Algorithm 3-3
Definition G-5
LUN
Definition G-5
LVM
Definition G-5
Battery Subsystem 2-3
Bus and Tags Connector Panel
Card Cage 2-3
Cooling Modules 2-3
Disk Device 2-3
Dual Power Cords 2-4
Ethernet Hub 2-3
Power Subsystem 2-3
Service Processor 2-4
MB
Definition G-6
Media
Definition G-6
Megabyte
Definition G-6
Mirrored Pair
Definition G-6
Mirroring
Advantages 5-16
Definition G-6
Description 5-15
RAID-1 5-3
Miss Elimination 4-2
Mission Critical Data 5-3
2-3
N
Non-disruptive
Component Replacement
5-9
O
Open System Channels
Power Failure 5-5
Open Systems 1-8
Operator Panel
Disk Director 2-6
Functions 2-6
P
M
Main Components
Backplane 2-3
i-4
Parallel Processing 4-1
Parity
Checking 5-13
Performance Features 1-5
77 A1 69UP Rev00
Index
Cache 4-1
CDA 7 4-1
Fast Write Capabilities 4-2
Multiple Channel Directors 4-1
Parallel Processing 4-1
RPS Miss Elimination 4-2
Physical Partition
Definition G-6
Power
Emergency Power Off A-2
Failure on Open System Channels 5-5
LED 2-7
Powering Down A-2
Powering Up A-1
Powering Up After Emergency Shutdown
A-3
Redundant Subsystem 5-4
Subsystem 2-3
PP
Definition G-6
Prefetch Algorithm 3-4
Primary Track
Definition G-6
Promotion
Definition G-6
Protection
Error Verification 5-2
Guidelines 5-14
Mirroring 5-15
Q
Queuing Time
3-6
R
RAID-1 Mirroring 5-3
Read
Definition of Hit G-6
Definition of Miss G-6
Hits 3-10
Miss Operation 3-11
Operations 3-9
Ready/Active LED 2-7
Reconfiguration
77 A1 69UP Rev00
Dynamic 5-11
Reconnect
Definition G-6
Reconnection
Definition G-6
Record Zero
Definition G-7
Redundant Power Subsystem 5-4
Reliability
Components 5-4
Features 5-1, 5-4
Non-disruptive Component Replacement
5-9
Redundant Power Subsystem 5-4
Upgrades and Loads 5-10
Reselect
Definition G-7
Reserved
Definition G-7
Response Time
I/O 3-6
RPS Miss Elimination 4-2
S
SCSI
Cable Worksheet B-4
Channel Adapters B-5
Channel Attachments 2-11
Channel Director 2-9
Definition G-7
Definition ID G-7
definition of Adapter G-7
Definition of Address G-7
Definition of Device G-7
Disk Emulation 2-7
Path Failure 5-6
SCU
Definition G-8
Segmented Data Buffer(ing) 4-2
Sequential Data Access Patterns 3-4
Service Processor 1-6, 2-4
Serviceability Features 1-6
Short Miss
Definition G-7
Single-ended Interface
i-5
CDA 7 5630 Product Manual
Definition G-7
SIO
Definition G-7
Slots
Cache 2-10
Soft Error Thresholds 5-13
Speed
Channel 4-2
SRDF
Definition G-7
for GCOS 7 G-8
Remote Link Director 2-9
SRDF for GCOS 7 5-16
Stage
Definition G-8
Storage Control Unit
Definition G-8
Storage Directors 4-1
Striping
Definition G-8
Switches
Enable/Disable 2-7
Synchronous Transmission
Definition G-8
System Battery Backup 5-4
System Check
Tests 5-2
U
Ultra SCSI
Channel Adapters B-5
Channel Attachment 2-11
Channel Director 2-9
Definition G-8
UNIX
Definition G-8
Upgrades
Microcode 5-10
V
Volume
Definition
G-9
W
Write
Delayed Fast Write Operation
Fast Write Operation 3-13
Operations 3-11
WSP
Definition G-9
3-14
X
T
Target
Definition
Terabyte
Definition
i-6
XOR
in LRC
5-13, G-5
G-8
G-8
77 A1 69UP Rev00
Technical publication remarks form
Title :
DPS7000/XTA NOVASCALE 7000 CDA 7 5630 Product Manual Hardware: CDA
Reference Nº :
77 A1 69UP 00
Date:
September 1999
ERRORS IN PUBLICATION
SUGGESTIONS FOR IMPROVEMENT TO PUBLICATION
Your comments will be promptly investigated by qualified technical personnel and action will be taken as required.
If you require a written reply, please include your complete mailing address below.
NAME :
COMPANY :
ADDRESS :
Please give this technical publication remarks form to your BULL representative or mail to:
Bull - Documentation Dept.
1 Rue de Provence
BP 208
38432 ECHIROLLES CEDEX
FRANCE
info@frec.bull.fr
Date :
Technical publications ordering form
To order additional publications, please fill in a copy of this form and send it via mail to:
BULL CEDOC
357 AVENUE PATTON
B.P.20845
49008 ANGERS CEDEX 01
FRANCE
CEDOC Reference #
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
_ _
_ _
_ _ _ _
_
[ _ _ ]
[ _ _ ]
Phone:
FAX:
E-Mail:
+33 (0) 2 41 73 72 66
+33 (0) 2 41 73 70 66
srv.Duplicopy@bull.net
Designation
Qty
: The latest revision will be provided if no revision number is given.
NAME:
Date:
COMPANY:
ADDRESS:
PHONE:
E-MAIL:
For Bull Subsidiaries:
Identification:
For Bull Affiliated Customers:
Customer Code:
For Bull Internal Customers:
Budgetary Section:
For Others: Please ask your Bull representative.
FAX:
BULL CEDOC
357 AVENUE PATTON
B.P.20845
49008 ANGERS CEDEX 01
FRANCE
REFERENCE
77 A1 69UP 00
Download PDF