Digital PDFs

AA-D027B-TE

March 1980

250 pages

Original

10MB

Document:	VAX/VMS System Manager’s Guide
Order Number:	AA-D027B-TE
Revision:	000
Pages:	250
Original Filename:

OCR Text

VAX/VMS
System Manager's Guide
Order No. AA-00278-TE

March 1980

This document covers the following tasks of VAX/VMS system management:
1) setting up user's accounts, 2) managing public files and volumes, 3) controlling overall system performance, 4) monitoring system activity, and 5) recognizing and dealing with errors and failures.
The aims of this guide are to provide a background for understanding these
tasks and to provide rules and guidelines for performing them.

VAX/VMS
System Manager's Guide
Order No. AA-00278-TE

SUPERSESSION/UPDATE INFORMATION:

This revised document supersedes the
VAX/VMS System Manager's Guide
(Order No. AA-D027 A-TE)

OPERATING SYSTEM AND VERSION:

VAX/VMS V02

SOFTWARE VERSION:

VAX/VMS V02

To order additional copies of this document, contact the Software Distribution
Center, Digital Equipment Corporation, Maynard, Massachusetts 01754

digital equipment corporation · maynard, massachusetts

First Printing, August 1978
Revised, March 1980
The information in this document is subject to change without notice
and should not be construed as a commitment by Digital Equipment
Corporation. Digital Equipment Corporation assumes no responsibility
for any errors that may appear in this document.
The software described in this document is furnished under a license
and may only be used or copied in accordance with the terms of such
license.
No responsibility is assumed for the use or reliability of software on
equipment that is not supplied by DIGITAL or its affiliated companies.

@ 1978, 1980 by Digital Equipment Corporation

The postage prepaid READER'S COMMENTS form on the last page of this
document requests the user's critical evaluation to assist us in
preparing future documentation.
The following are trademarks of Digital Equipment Corporation:

DIGITAL
DEC
PDP
DECUS
UNIBUS
COMPUTER LABS
COMTEX
DDT
DEC COMM
ASSIST-11
VAX
DECnet
DATATRIEVE

DECsystem-10
DECtape
DIBOL
EDUSYSTEM
FLIP CHIP
FOCAL
IND AC
LAB-8
DECSYSTEM-20
RTS-8
VMS
!AS
TRAX

MASSBUS
OMNIBUS
OS/8
PHA
RSTS
RSX
TYPESET-8
TYPESET-11
TMS-11
ITPS-10
SB!
PDT

CONTENTS

Page
PREFACE
SUMMARY OF TECHNICAL CHANGES

xiii
xv

CHAPTER

INTRODUCTION

1-1

1.1
1.2
1.3

WHAT IS SYSTEM MANAGEMENT?
TASKS THAT THE SYSTEM MANAGER PERFORMS
COMPONENTS OF THE VAX/VMS OPERATING SYSTEM

1-1

PART

USER ACCOUNTS

CHAPTER

USER AUTHORIZATION FILE

2-1

2.1
2.2
2.2.1
2.2.2
2.2.3
2.2.4
2.2.5
2.2.n
2.2.7
2.2.8
2.2.9
2.2.10
2.3
2.4
2.5

STRUCTURE OF THE UAF
AUTHORIZE UTILITY
Invoking AUTHORIZE
ADD Command
DEFAULT Command
EXIT Command
HELP Command
LIST Command
MODIFY Command
REMOVE Command
SHOW Command
Messages
OPERATIONS
USER DATA AREAS
UAF VERSION 2.0 UPGRADE

2-1
2-2
2-8
2-8
2-9
2-10
2-10
2-10
2-12
2-12
2-14
2-17
2-18
2-19

PROTECTION AND CONTROL

3-1

3.1
3.2
3.2.1
3.2.2
3.2.3

3-1
3-2

3•n

USER IDENTIFICATION CODE
PROTECTION OF DATA STRUCTURES AND DEVICES
Files and File-Structured Volumes
Nonshareable Non-File-Structured Devices
Mailboxes
Shared Pages in Memory (Global Sections)
Common Event Flags
Group Logical Name Table
FORMING GROUPS
PASSWORDS
PRIVILEGES
SECURING SYSTEM DATA AND DEVICES

3-7
3-7
3-7

RESOURCE CONTROL

4-1

4.1
4.1.1

LIMITS
AST Queue Limit (ASTLM)
Buffered I/O Count Limit (BIOLM)
Buffered I/O Byte Count Limit (BYTLM)
CPU Time Limit (CPUTIME)
Direct I/O Count Limit (DIOLM)

CHAPTER

3.2.4
3.2.5

3.2.6
3.3
3.4
3.5

CHAPTER

4 .1. 2
4 .1. 3
4 .1. 4
4 .1. 5

iii

1-2
1-2

2-11

3-4
3-5
3-5

3-n
3-n
3-n
3-n

4-1

4-2
4-2
4-2
4-3
4-3

CONTENTS

Page
4 .1. ()

4.3.8
4.3.9
4.3.10
4.3.11
4.3.12
4.3.13
4.3.14
4.3.15
4 .3 .in
4.3.17
4.3.18
4.3.19
4.3.20
4.3.21
4.3.22
4.3.23
4.3.24
4.3.25
4.3.2n
4.3.27
4.3.28
4.3.29
4.3.30
4.4

Open File Limit (FILLM)
Paging File Limit (PGFLQUOTA)
Subprocess Creation Limit (PRCLM)
Timer Queue Entry Limit {TQELM)
Working Set Default (WSDEFAULT)
Working Set Quota (WSQUOTA)
PRIORITY
PRIVILEGES
ACNT Privilege
ALLSPOOL Privilege
ALTPRI Privilege
BUGCHK Privilege
BYPASS Privilege
CMEXEC Privilege
CMKRNL Privilege
DETACH Privilege
DIAGNOSE Privilege
EXQUOTA Privilege
GROUP Privilege
GRPNAM Privilege
LOG IO Privilege
MOUNT Privilege
NETMBX Privilege
OPER Privilege
PFNMAP Privilege
PHY IO Privilege
PRMCEB Privilege
PRMGBL Privilege
PRMMBX Privilege
PSWAPM Privilege
SETPRV Privilege
SHMEM Privilege
SYSGBL Privilege
SYSNAM Privilege
SYSPRV Privilege
TMPMBX Privilege
VOLPRO Privilege
WORLD Privilege
ACCOUNTING FOR THE USE OF SYSTEM RESOURCES

PART

OPERATIONAL CONTROL

CHAPTER

MAINTAINING PUBLIC FILES AND VOLUMES

5-1

5.1

FILES-11 DISK STRUCTURE
Index File
Storage Bit Map File
Bad Block File
Master File Directory
Co re Image File
Volume Set List File
Continuation File
Back-Up Log File
Pending Bad Block Log File
Files-11 Structure Level 1 Versus Structure
Level 2

5-2

4 .1. 7
4 .1.8

4 .1.9
4.1.10
4.1.11
4.2
4.3
4.3.l
4.3.2
4.3.3

4.3.4
4.3.5

4.3.11
4.3.7

5 .1.1
5 .1. 2
5 .1. 3
5 .1.4
5 .1. 5
5 .1. 6
5 .1. 7
5 .1.8
5 .1.9

5.1.10

4-3
4-3
4-3
4-4
4-4
4-4
4-4
4-5
4-7
4-7
4-7
4-8
4-8
4-8
4-8

4-9
4-9
4-9
4-9
4-10
4-10
4-10
4-10
4-10
4-11
4-11
4-11
4-12
4-12
4-12
4-13
4-13
4-13
4-13

4-14
4-14
4-14
4-14
4-15

5-2
5-3
5-3
5-3
5-3

5-4
5-4
5-4

5-4
5-4

CONTENTS

Page

5.5
5.n

INITIALIZING PUBLIC VOLUMES
/ACCESSED=n Qualifier
/CLUSTER SIZE=n Qualifier
/EXTENSION=n Qualifier
/HEADERS=n Qualifier
/INDEX=position Qualifier
/MAXIMUM FILES=n Qualifier
/WINDOW=n Qualifier
MOUNTING PUBLIC VOLUMES
/BIND=volume-set-name Qualifier
/PROCESSOR=option Qualifier
MAINTAINING VOLUME INTEGRITY
RMSSHARE UTILITY
BACKING UP PUBLIC VOLUMES

5-8
5-8
5-10

DISK QUOTAS

11-1

n.l

11-1

6.2.4

DISKQUOTA UTILITY
Invoking DISKQUOTA
ADD Command
CREATE Command
DISABLE Command
ENABLE Command
EXIT Command
HELP Command
MODIFY Command
REBUILD Command
REMOVE Command
SHOW Command
USE Command
Error Messages
OPERATIONS
Exceeding the Quota
Suspending Quotas
REBUILD on Mount
Restrictions on Other System Operations

n-9
fi-9
n-10
11-10
11-10

INSTALLING KNOWN IMAGES

7-1

7.1
7.2
7.3
7.3.1
7.3.2
7.3.3

EXECUTABLE AND SHAREABLE IMAGES
KNOWN FILE LISTS
INSTALL UTILITY
Invoking INSTALL
Installing Known Images
Deleting Known Images
Replacing Known Images
Listing Known Images
Specifying and Reading Privileges
Terminating INSTALL
Error Messages
OPERATIONS
Start-Up Procedures
Order of Installation
Privileges
Deleting Known Images and Dismounting
Volumes
Shareable Image Files
Multiport Memory

7-2
7-2
7-3
7-3
7-4
7-5

5.2
5.2.1
5.2.2
5.2.3

5.2.4
5.2.5
5.2.fi

5.2.7
5.3
5.3.1
5.3.2

5.4

CHAPTER

6 .1.1
6 .1. 2
6 .1. 3
6 .1. 4
n.1. 5
6.1.fi
n .1. 7
6 .1.8
6 .1.9
fi.1.10
6.1.11
6.1.12
fi.1.13
fi. 2
6.2.1

6.2.2
fi.2.3
CHAPTER

7.3.4

7.3.5
7.3.n
7.3.7
7.3.8

7.4
7.4.1
7.4.2

7.4.3
7.4.4
7.4.5

7.4.6

5-5

5-n
5-fi

5-n
5-h
S-'1
5-'1

5-7
5-7
5-7
5-7

'1-3
{)-3

11-3

n-tl

11-4
n-4
F}-4
6-4
e:\-5

11-5

n-n
{:)-()

e:\-7

7-5
7-fi

7-9
7-9
7-9
7-10
7-11

7-11
7-11
7-11
7-12

7-12

CONTENTS

Page
CHAPTER

START-UP COMMAND PROCEDURES

8.1
SITE-INDEPENDENT START-UP COMMAND PROCEDURE
8 .1.1
Housekeeping Chores
8 .1. 2
Logical Name Assignments
Symbolic Debugger
8.1.2.1
8.1.2.2
COBOL Programs
PASCAL Programs
8.1.2.3
8.1.2.4
RSX-llM Programs
System Processors
8.1.2.5
8 .1. 3
Known Images
8 .1.4
I/O Devices and Drivers
8 .1. 5
VAX-11 RMS File Sharing
8 .1.6
Termination of the Procedure
8.2
SITE-SPECIFIC START-UP COMMAND PROCEDURE
System Disks
8.2.1
8.2.2
Logical Names
Device Characteristics
8.2.3
Queues
8.2.4
8.2.5
Known Images
8.2.6
System Dump Analyzer
Operator's Log File
8.2.7
8.2.8
Standard Batch Jobs
8.2.9
Manually Connected Devices and Multiport
8.2.10
8.2.11

CHAPTER

Memory Units
Secondary Paging and Swapping Files
Announcement

8-1
8-1
8-1
8-2
8-2
8-2
8-2
8-2
8-3
8-3
8-4
8-4
8-4
8-5
8-5
8-5
8-5

8-n
8-n
8-n
8-7
8-7
8-7
8-7
8-7

BATCH AND PRINT JOBS

9.1

9-2
SPOOLING
Establishing Spooled Devices
9-3
Turning Off Spooling
9-4
9-4
BATCH JOBS
Creating Batch Queues
9-5
Starting Batch Queues
9-5
9-6
Stopping Batch Queues
Deleting Batch Queues
9-6
9-6
Batch Versus Interactive Jobs
9-6
Guides to Setting Up Batch Queues
PRINT QUEUES
9-8
Creating Print Queues
9-9
Starting Print Queues
9-10
Stopping Print Queues
9-10
Deleting Print Queues
9-10
Assigning a Named, or Logical, Print Queue
to a Printer
9-10
Deassigning a Named, or Logical, Print Queue
from a Printer
9-11
Vertical Page Size
9-11
Forms Control
9-11
Printer Characteristics
9-12
Guides to Setting Up Print Queues and
Spooled Line Printers
9-12
TERMINAL QUEUES
9-19

9 .1.1
9 .1. 2
9.2
9.2.1
9.2.2
9.2.3
9.2.4
9.2.5
9.2.n
9.3
9.3.1
9.3.2
9.3.3
9.3.4
9.3.5
9.3.6
9.3.7
9.3.8
9.3.9

9.3.10
9.4

9-1

CONTENTS

Page
CHAPTER

ERRORS AND OTHER SYSTEM EVENTS

10.l
ERROR LOG
10.l.l
SYE Utility
10.1.2
Operations
10.1.2.1 Using Error Reports
10.1.2.2 Maintaining the Error Log Files
10.2
OPERATOR'S LOG FILE
10.3
REPORTING SOFTWARE PROBLEMS
PART

III

SYSTEM CONFIGURATION

CHAPTER

TUNING CONSIDERATIONS

PRETUNING CONSIDERATIONS
11.1
Hardware Resources
11.1.1
Workload Distribution
11.1.2
Sharing Code
11.1. 3
BACKGROUND DISCUSSION
11.2
TOOLS
11. 3
System Parameters
11.3.1
user Authorization File
11.3.2
DCL Commands
11.3.3
DISPLAY Utility
11.3.4
SMALL SYSTEMS
11.4
Working Set Size
11.4.1
11.4.1.1 Initial Working Set Limits
11.4.1.2 Adjustments
Page Cache
11.4.2
11.4.2.l Cache Sizes
11.4.2.2 Cluster Size
Priority, Quantum, and Compute-Bound
11.4.3
Swapping Rate
11.4.3.1 Very Large Working Sets
11.4.3.2 Compute-Bound Programs
11.4.4
System Requirements
11.4.4.1 Nonpaged Dynamic.Pool
11.4.4.2 System Working Set
11.4.4.3 VAX-11 RMS Buffers
11.5
LARGE SYSTEMS
11.5.l
Header Cache
11.5.2
Directory Cache
11.5.3
Quota Cache
11.5.4
Map, File Identification, and Extent Caches
11.5.5
System Directory Cache
11.5.6
Multiple ACPs
CHAPTER

10-1
10-1
10-1
10-3
10-3
10-4
10-5
10-n

11-1

11-2
11-2
11-3
11-3
11-4
11-n
· 11-n

11-7
11-7
11-8
11-9

11-9
11-10

11-11
11-12
11-12
11-13
11-13
11-14
11-14
11-14
11-15
11-15
11-15
11-lh
11-H
11-17
11-17

11-17
11-17
11-17

SYSTEM PARAMETERS

12-1

12.1
12.1.1
12.1.2
12.1.3
12.1.4
12.1.5
12.1.6
12.1.7

MAJOR PARAMETERS
PFCDEFAULT Parameter
GBLSECTIONS Parameter
GBLPAGES Parameter
MAXPROCESSCNT Parameter
SYSMWCNT Parameter
BALSETCNT Parameter
IRPCOUNT Parameter

12-8
12-8

vii

12-9

12-9
12-9
12-10
12-10
12-10

CONTENTS

Page
12.1.8
WSMAX Parameter
12.1.9
NPAGEDYN Parameter
12.1.10
PAGEDYN Parameter
12.1.11
VIRTUALPAGECNT Parameter
12.1.12
QUANTUM Parameter
Modified Page Writer Parameters
12.1.13
12.2
SYS PARAMETERS
12.2.1
KFILSTCNT Parameter
12.2.2
PROCSECTCNT Parameter
12.2.3
MINWSCNT Parameter
12.2.4
INTSTKPAGES Parameter
SPTREQ Parameter
12.2.5
12.2.6
PFRATL Parameter
PFRATH Parameter
12.2.7
WSINC Parameter
12.2.8
12.2.9
WSDEC Parameter
12.2.10
AWSMIN Parameter
12.2.11
AWSMAX Parameter
AWSTIME Parameter
12.2.12
12.2.11
EXTRACPU Parameter
12.2.14
MAXSYSGROUP Parameter
MAXBUF Parameter
12.2.15
12.2.16
DEFMBX Parameters
12.2.17
FREELIM Parameter
12.2.18
XFMAXRATE Parameter
12.2.19
LAMAPREGS Parameter
12.2.20
REALTIME SPTS Parameter
12.2.21
CLISYMTBL Parameter
BUGREBOOT Parameter
12.2.22
12.2.23
CRDENABLE Parameter
DUMPBUG Parameter
12.2.24
12.2.25
BUGCHECKFATAL Parameter
12.2.26
POOL PAGING Parameter
12.2.27
SBIERRENABLE Parameter
12.2.28
SETTIME Parameter
12.2.29
SYSPAGING Parameter
12.2.30
UAFALTERNATE Parameter
12.2.31
RESALLOC Parameter
TTYSCANDELTA Parameter
12.2.32
TTY SPEED Parameter
12.2.33
TTY-BUF Parameter
12.2.34
12.2.35
TTY DEFCHAR Parameter
TTY-TYPAHDSZ Parameter
12.2.36
12.2.37
TTY PROT Parameter
TTY-OWNER Parameter
12.2.38
JOB PARAMETERS
12.3
MAXPRINTSYMB Parameter
12.3.1
12.3.2
DEFPRI Parameter
12.3.3
JOBLIM Parameters
12.4
ACP PARAMETERS
12.4.1
ACP MULTIPLE Parameter
12.4.2
ACP SHARE Parameter
CACHE Parameters
12.4.3
12.4.3.l ACP MAPCACHE Parameter
12.4.3.2 ACP HDRCACHE Parameter
12.4.3.3 ACP-DIRCACHE Parameter
12.4.3.4 ACP FIDCACHE Parameter

viii

12-11
12-11
12-13
12-13
12-14
12-14
12-14
12-15
12-15
12-15
12-15
12-15
12-1'1
12-ln
12-lfi
12-17
12-17
12-17
12-17
12-17
12-17
12-18
12-18
12-18
12-18
12-18
12-18
12-19
12-19
12-19
12-19
12-19
12-19
12-19
12-20
12-20
12-20
12-20
12-20
12-20
12-21
12-21
12-21
12-21
12-21
12-22
12-22
12-22
12-22
12-22
12-22
12-22
12-23
12-23
12-23
12-23
12-23

CONTENTS

Page

CHAPTER

12-24
12-24
12-24
12-24
12-25
12-25

SYSTEM GENERATION

13-1

SYSGEN UTILITY
13.1
Invoking SYSGEN
13.1.l
AUTOCONFIGURE Command
13.1.2
CONNECT (Hardware) Command
13.1.3
CONNECT (Software) Command
13.1.4
CONNECT (Console) Command
13.1.5
CREATE Command
13.1.fi
DISABLE Command
13.1.7
ENABLE Command
13.1.8
EXIT Command
13.1.9
HELP Command
13.1.10
INSTALL Command
13.1.11
LOAD Command
13.1.12
RELOAD Command
13.1.13
SET (Parameter) Command
13.1.14
SET (Start-Up Command Procedure) Command
13.1.15
SHARE (Initialize) Command
13.1.lF>
SHARE (Connect) Command
13.1.17
SHOW (Parameter) Command
13.1.18
SHOW (Device) Command
13.1.19
SHOW (Start-Up) Command
13.1.20
USE Command
13.1.21
WRITE Command
13.1.22
Error Messages
13.1.23
OPERATIONS
13.2
System Parameters
13.2.1
13.2.1.l Create a Parameter File
13.2.1.2 Modify the System Image
13.2.1.3 Modify the Active System
Devices and Device Drivers
13.2.2
System Files
13.2.3
Start-Up Command Procedure
13.2.4
Multiport (Shared) Memory
13.2.5
CHAPTER

12-23
12-23
12-23
12-23
12-23
12-24

12.4.3.5 ACP EXTCACHE Parameter
12.4.3.fi ACP-EXTLIMIT Parameter
12.4.3.7 ACP-QUOCACHE Parameter
12.4.3.8 ACP-WRITEBACK Parameter
12.4.3.9 ACP SYSACC Parameter
ACP-WORKSET Parameter
12.4.4
ACP-MAXREAD Parameter
12.4.5
ACP-WINDOW Parameter
12.4.'1
ACP-DATACHECK Parameter
12.4.7
ACP BASEPRIO Parameter
12.4.8
ACP-SWAPFLGS Parameter
12.4.9
RMS PARAMETERS
12.5
PQL PARAMETERS
12.6

12-~4

13-1
13-4
13-4
13-5
11-n
13-n
13-n
13-7
13-7
13-8
13-8
13-8
13-9
13-9
13-10
13-10
13-11
13-12
13-13
13-14
13-15
13-15
13-1 F)
13-lFl
13-17
13-17
13-17
13-18
13-18
13-19
13-20
13-22
13-23

DISPLAY UTILITY

14-1

14.1
14.2
14.3
14.4
14.5

INVOKING DISPLAY
CTRL/C COMMAND
CYCLE COMMAND
EXIT COMMAND
FCP COMMAND

14-2
14-2
14-3
14-3
14-3

CONTENTS

Page

14.n
14.7
14.8
14.9
14.10
14 .11
14.12
14 .13
14.14

HELP COMMAND
!ORATES COMMAND
M2 AND MS COMMANDS
PAGE COMMAND
POOL COMMAND
S2 AND SS COMMANDS
TOPUSERS COMMAND
USERS COMMAND
ERROR MESSAGES

INDEX

14-S
14-S
14-7
14-9
14-11
14-12
14-15
14-H
14-18
Index-1

FIGURES
FIGURE

2-1
3-1

3-2
7-1
7-2
7-3
9-1
9-2

9-3

9-4
10-1
10-2
11-1
13-1
13-2

13-3
13-4
14-1
14-2
14-3
14-4
14-S
14-n
14-7
14-8

14-9
14-10

Display Produced By SHOW Command
2-13
Classification of Users with Respect to a File
with a UIC of [100,100]
3-2
Controlling Access to a File
3-4
7-F,
Known Image Display -- Brief Description
Known Image Display -- Full Description
7-7
Global Sections Display
7-8
Setting Up a Spooled Printer and a Print Queue
on a system with One Line Printer
9-13
Setting Up Spooled Printers and Print Queues
on a System with Two Line Printers with the
Same Characteristics
9-15
Setting Up Spooled Printers and Print Queues
on a system with Three Line Printers; Two with
the Same Characteristics and One with Special
9-1')
Characteristics or in a Remote Location
Setting Up Spooled Printers and Print Queues
Adding a Logical Queue to the System with
9-18
Three Line Printers
Operator's Log File
10-5
Software Performance Report (SPR)
10-7
VAX/VMS Memory Configuration
11-4
Display Produced by SHOW Command of SYSGEN
13-14
Display Produced by SHOW /DEVICE Command of
SYSGEN
13-15
Display Produced by SHOW /STARTUP Command of
SYSGEN
13-15
Example of Multiport Memory Structures
13-24
FILE PRIMITIVE STATISTICS Display
14-5
14-7
I/O SYSTEM RATES Display
14-8
TIME IN PROCESSOR MODES Display (M2 Command)
TIME IN PROCESSOR MODES Display (MS Command)
14-9
PAGE MANAGEMENT STATISTICS Display
14-11
14-12
NONPAGED POOL STATISTICS Display
NUMBER OF PROCESSES IN SCHEDULER STATES Display
(S2 Command)
14-14
NUMBER OF PROCESSES IN SCHEDULER STATES Display
(SS Command)
14-15
TOP CPU TIME USERS Display
14-10
14-18
VAX/VMS PROCESSES Display

CONTENTS

Page
TABLES
TABLE

2-1
2-2
2-3
2-4
4-1
6-1
6-2
7-1
7-2
7-3
9-1
12-1
12-2
12-3
12-4
12-5
12-6
12-7
13-1
13-2
13-3
14-1
14-2

AUTHORIZE Command Summary
UAF Field Qualifiers
Initial Values of SYSTEM, FIELD, and SYS TEST
Records
AUTHORIZE Error Messages
VAX/VMS Privileges
DISKQUOTA Command Summary
DISK QUOTA Error Messages
INSTALL Command Summary
Privilege Names and Bit Locations
INSTALL Error Messages
Operator Commands Used in Regulating Spooling
and Queuing
MAJOR Parameters
SYS Parameters (excluding MAJOR parameters)
JOB Parameters
ACP Parameters
RMS Parameters
PQL Parameters
Standard System Parameter Files
SYSGEN Command Summary
Standard Device Types and Drivers
Suggested Sizes for System Files
DISPLAY Command Summary
DISPLAY Error Messages

2-2
2-3
2-n
2-14
4-n
h-2
n-7
7-3
7-9
7-10
9-2
12-3
12-4
12-6
12-6
12-7
12-7
12-8
13-1
13-20
13-20
14-1
14-18

PREFACE

The VAX/VMS System Manager's Guide is both a useful adjunct to a
formal course in VAX/VMS system management and an indispensable
reference book for every VAX/VMS system manager.

MANUAL OBJECTIVES

The objectives of this guide are twofold. The first objective is to
give the reader an understanding of the reasons for performing the
tasks of VAX/VMS system management. The second objective is to show
the reader how to perform the tasks of VAX/VMS system management.

INTENDED AUDIENCE

This guide is intended for any person who has overall responsibility
for controlling the operations of a VAX/VMS installation. The reader
of this guide is most likely a data processing generalist, not
necessarily a programmer and probably not a systems programmer.

STRUCTURE OF THIS DOCUMENT

This system manager's guide consists of three parts, as follows:
Part I:

User Accounts

Understanding the user authorization file
AUTHORIZE utility;
understanding groups,
privileges, and the accounting log file
Part I I :

(UAF);
using the
limits, priorities,

Operational Control

Initializing and mounting public volumes, backing up files on
public
volumes, establishing disk quotas,
installing known
images, maintaining the start-up command procedures, setting up
spooling, creating and controlling batch queues and print queues,
logging and reporting errors
Part I I I :

System Configuration

Understanding system tuning considerations, understanding the
system parameters, using the SYSGEN utility, monitoring the
system

xiii

ASSOCIATED DOCUMENTS

The VAX-11 Inform~tlo~_Directory and Index lists and describes all the
documents the system manager may need to ref er to in the course of
performing system management tasks.
For general background information about the system, see
Summary Descrip~i?n and Gl_~-~~-~!:X: and the VAX/VMS Primer.

the

The following documents may also be useful:
VAX/VMS Command Language User's Guide
VAX/VMS 0 per ~~~T'._~~~~Q:~)-.9~--- VAX/VMS Release Notes
VAX-11 Software Installation Guide
VAX/VMS System M~-~~g~s __9_~-~---~-~c~ye ry P r~~~~ur_e~~~-11Ua 1

xiv

VAX/VMS

SUMMARY OF TECHNICAL CHANGES

Modifications to this manual reflect the following changes to
for Version 2.0:

VAX/VMS

•

UAF format -- The format of the UAF has
changed
from
sequential to indexed. A conversion procedure (Section 2.5)
updates VAX/VMS Version 1.0 UAFs to the new format. No longer
is a TMP file used, and no longer is a new UAF version created
after an update.

•

Limits -- PGFLQUOTA is specified in pages rather than 256-page
blocks.
CPUTIME has been implemented.
Former deductible
limits are now pooled. Only CPUTIME is a deductible limit.

•

Log-in flags -- LOCKPWD, a new log-in flag,
from issuing the new SET PASSWORD command.

•

Privileges -- Privileges are now categorized by level of
danger:
none, normal, group, devour, systems, file, all.
Five new privileges exist: BYPASS (bypass UIC protection),
SHMEM (multiport memory), EXQUOTA (exceed disk quotas), PFNMAP
(map to physical pages), and SYSPRV (assume system UIC
status). Privileges on the DEFAULT record include only TMPMBX
and NETMBX; GROUP was deleted.

•

Log-in without a UAF -- Any name and password work if no UAF
exists;
this permits an open system. Any name and password
work on the console if the UAF cannot be accessed;
this
permits the system manager or operator to bypass a file
problem.

•

AUTHORIZE commands -- The LIST and SHOW commands in AUTHORIZE
now
take
a
userspec argument, plus /FULL and /BRIEF
qualifiers. The display format has changed.

•

User data on UAF -- Users can place up to 256 bytes of their
own data on each UAF record using VAX-11 RMS but must specify
the offset where this data begins and observe several other
conventions.

•

Volume sets -- The /BIND qualifier to the MOUNT command
(see
the VAX/VMS Command Language User's Guide) joins volumes into
a volume set.

•

File system caches
The system will now cache extensive
amounts of information concerning a volume's free space, file
identifications, quota file entries, and file headers.
The
system manager, however, must take precautions to ensure that
the cached information is written back to disk.

prevents

users

•

RMSSHARE utility -- The RMSSHARE utility enables VAX-11 RMS
file sharing and sets a maximum count on the number of pages
that can be allocated from the paged dynamic pool to support
this facility.

•

DISKQUOTA utility -- The system manager can establish disk
quotas
with
the
new
DISKQUOTA
utility.
The system
automatically records disk usage and enforces quotas during
file operations. The system now requires the SYSPRV privilege
to change the owner UIC of a file. Relative volume 1 of a
volume set must be online at all times.

•

Shareable images -- Shareable images linked into executable
images without private copies no longer need to be installed
to be executed, as long as they do not contain writeable
non-CRF sections. The system will create private sections for
them.

•

/PR.OTECT qualifier to INSTALL -- The INSTALL utility contains
a new qualifier to designate images containing protected
code -- code that runs in kernel or executive mode -- that may
be called by a user-level image.

•

INSTALL utility displays -- The INSTALL utility displays have
been expanded to show protected images and to show global
sections in multiport memory units.

•

System symbol GBL$file-name -- This new symbol means all
global sections beginning with file-name.
The symbol is
useful for installing shared images in multiport memory units.

•

Site-independent
start-up
command
procedure -- Changes
include:
minor changes to housekeeping chores and known
images; deletion of commands to mount diskettes; addition of
logical names for VAX-11 COBOL-74 and VAX-11 PASCAL; deletion
of logical names for VAX-11 FORTRAN; and addition of commands
to enable VAX-11 RMS file sharing.
FORTRAN logical name
assignments are now implicit.

•

Model site-specific start-up command
procedure -- Changes
include commands to run the system dump analyzer, to connect
multiport memory units, and to install secondary paging and
swapping files.

•

Batch queues -- The system manager can specify (in addition to
a job limit, priority, and swapping mode) a working set
default, working set quota, CPU time default, and CPU time
maximum as qualifiers to the INITIALIZE and START commands.
These options permit working set sizes and CPU time limits for
batch jobs to be independent of those for interactive users.
Users can also specify working set sizes and CPU time limits
on the $JOB card or as qualifiers to the SUBMIT command.

•

Printer characteristics and forms types -- The system manager
can set printer characteristics and forms types as qualifiers
to INITIALIZE and START commands. Print jobs submitted with
characteristics not belonging to the print queue or with a
different forms type are put on a hold status.

xvi

•

System parameters -- New SYS parameters
include
PFRATL,
PFRATH, WSINC, AWSMIN, AWSMAX, AWSTIME
(all for automatic
working set adjustments),
EXTRACPU, XMAXRATE,
LAMAPREGS,
REALTIME SPTS,
CLISYMTBL,
TTY TYPAHDSZ,
TTY PROT,
and
TTY OWNER.
New
ACP
parameters
include
ACP QUOCACHE,
ACP WRITEBACK, and ACP DATACHECK;
in addition, ACP-EXTCACHE,
ACP-EXTLIMIT, and ACP FlDCACHE are now used.
The system now
adjusts working set lTmits automatically according to the page
fault rates for
the working sets and the values of the
automatic
working
set
parameters.
The new and newly
implemented ACP parameters mostly specify
cache
sizes.
Certain parameters are now designated as generative (affecting
structures at system generation)
and dynamic
(able to be
modified during system operation).

SYSGEN utility -- The WRITE ACTIVE option modifies the dynamic
parameters during system operation. The WRITE CURRENT option
modifies the system image on disk (SYS$SYSTEM:SYS.EXE),
which
contains the
initial parameter values for the next bootstrap
operation. SYSGEN now initializes and connects multiport
memory
units.
The
CREATE
command
can
now
create
non-contiguous secondary paging and swapping files on any
volume or volume set.
The new INSTALL command augments
SYS$SYSTEM:PAGEFILE.SYS or SWAPFILE.SYS with a
secondary
paging or swapping file.

xvii

CHAPTER 1
INTRODUCTION

In a nutshell, system management of a VAX/VMS installation entails two
main responsibilities: system performance and system operation. The
system manager:
the

overall

•

Makes decisions that relate to optimizing
performance and efficiency of the system

•

Performs procedures that relate to the overall management
control of the system

and

To make good decisions, the system manager must understand both the
needs of users and the capabilities of the VAX/VMS operating system.
To perform system management procedures well, the system manager must
have a working knowledge of the functions of the VAX/VMS operating
system.
The pattern followed throughout this guide is to discuss the issues
and the facts that will help the system manager make decisions and
then to give general rules and guidelines for performing
the
procedures of system management.
It is not possible to prescribe a precise set of formulas for setting
up and running a VAX/VMS system. System management cannot be done by
rote or from a cookbook.
Rather, the system manager must -- by
combining an understanding of users' needs and system capabilities
with a working knowledge of the functions of VAX/VMS -- work out a
coherent strategy for effective system management.

1.1

WHAT IS SYSTEM MANAGEMENT?

In its most abstract sense, system management means the overall
control of the operations of a computer system for the benefit of the
users of the system. System management is a function that can be
exercised by a single system manager, assisted by system operators, or
it is a function that can be shared by several persons, some or all of
whom may also serve as system operators.
A computer installation exists to serve its users. Ideally, then,
it
should be operated to provide service to all users with efficiency and
economy. This is the challenge of system management.

1-1

INTRODUCTION
1.2

TASKS THAT THE SYSTEM MANAGER PERFORMS

For practical purposes, system management is best defined in terms of
the tasks that the system manager typically performs. The tasks of
system management fall into the following four categories:
•

Getting the system up and running

•

Setting up users' accounts

•

Controlling the operation of the system

•

Configuring the system

f~r

good performance

The first topic (getting the system up and running) is the subject of
the VAX-11 Software Installation.Gu~de.
The remaining three topics
are the sub]ects -oT-HiTs-sysfe_m___m_aiiagerTs- guide.

1.3

COMPONENTS OF THE VAX/VMS OPERATING SYSTEM

Among the first and most important responsibilities of the system
manager is getting the VAX/VMS system up and running. At a minimum,
this means the bootstrap loading of the operating system distributed
by DIGITAL.
Usually, in addition to bootstrapping the VAX/VMS operating system,
getting a system up and running means both customizing some of the
parameters of the operating system and incorporating optional software
into the system. This optional software, which runs under control of
the VAX/VMS operating system, includes VAX-11 FORTRAN, VAX-11 BASIC,
VAX-11 COBOL-74, and other optional products.
For an explanation of the steps that the system manager may have to
take
in getting a VAX/VMS system up and running, see the VAX-11
Software Installation Guide.
The components of the VAX/VMS operating system are cataloged in nine
directories on the system distribution medium.
The names of these
directories and brief descriptions of their contents follow:
[SYSLIB]
• This
directory contains various macro and object libraries.
[SYSMSG]
• This
directory contains system message files.
•

e
•

[SYSMGR]
This directory contains files used in managing
system.

the

operating

[SYSHLP]
This directory contains text files and help libraries for
HELP utility.
[SYSERR]
This is the directory for the error log file

the

(ERRLOG.SYS).

[SYSTEST]
This directory contains files used in testing the functions of
the operating system.

1-2

•

[SYSMAINT]
This directory contains system diagnostic programs.

[SYSUPD]
This directory contains files used in applying system updates.

•

[SYSUPD.EXAMPLES]
This directory contains sample driver programs,
system services, and other source programs.

•

user-written

[SYSEXE]
This directory contains the executable images of most
functions of the operating system.

the

For a complete list of the files contained on the distribution medium,
see the VAX-11 Software Installation Guide.

1-3

PART I
USER ACCOUNTS

The system manager is responsible for setting
meaningful system of accounts for:

and

maintaining

•

Defining the users to the system

•

Defining relationships among the users of the system for
protection, interprocess communication, and accounting

•

Granting some users the privileges to
perform
certain
sensitive system functions, and restricting other users fr?m
performing these functions

•

Setting limits on the use of reusable system resources

•

Giving users priorities in using the system

file

Part I consists of Chapters 2 through 4:
2: User Authorization File -- Discusses the make-up
• Chapter
of
the
user
authorization
file,
provides
operating
instructions for the AUTHORIZE utility
3:
• Chapter
organization

Protection
and
Control -- Discusses
of users for protection, communication,
accounting functions

•

the
and

Chapter 4:
Resource Control -- Defines the
limits
and
privileges available to VAX/VMS users, explains priorities and
the accounting log file

CHAPTER 2
USER AUTHORIZATION FILE

The system manager names users who may log in to the system and places
controls on their activities by maintaining a record for each user in
the user authorization file (UAF).
The full name of the UAF is
SYS$SYSTEM:SYSUAF.DAT.
Its
primary maintenance vehicle is the
AUTHORIZE utility.

2.1

STRUCTURE OF THE UAF

Each record in the UAF includes the following information:
•

Name and password -- Identifies a user to the system at
time

•

User identification code (UIC) -- Identifies a user by a group
number and a member number (see Chapter 3 for more detailed
information)

•

Default file specification -- Provides
directory names for file access

•

Default command
language -- Names
interpreter as DCL or MCR

•

Login flags -- Allows the system manager to inhibit the use of
the CTRL/Y function, restrict users to their default command
interpreters, and/or lock user passwords

•

Priority -- Specifies the base priority of the process created
for the user at login time (see Chapter 4 for more detailed
information)

•

Resources -- Limits the system resources the user may
(see Chapter 4 for more detailed information)

•

Privileges -- Limits activities the user
Chapter 4 for more detailed information)

The software distribution kit provided
contains a UAF of four records:
•

with

default

the

device

default

may
new

command

consume

perform
VAX/VMS

and

(see
system

DEFAULT -- Serves as a template in creating user records in
the UAF.
A new user record is assigned the values of the
DEFAULT record except where the system manager explicitly
overrides those values.
The DEFAULT record can be modified
but cannot be deleted from the UAF.
2-1

USER AUTHORIZATION FILE

AUTHORIZE
•

SYSTEM -- Provides a means for the system manager to log in
with full privileges. The SYSTEM record can be modified but
cannot be deleted from the UAF.

•

FIELD -- Permits DIGITAL field service personnel to check out
a new system. The FIELD record can be deleted once the system
is installed.

•

SYSTEST -- Provides an appropriate environment for running the
User Environment Test Package (UETP). The SYSTEST record can
be deleted once the system is installed.

Typically, the system manager adds, deletes, and modifies records in
the UAF provided with the distribution kit. The system manager can,
however, delete the UAF with the DCL command DELETE (or rename the UAF
with the DCL command RENAME) and create a new UAF with AUTHORIZE. A
new UAF created in this manner contains just two records -- DEFAULT
and SYSTEST.

2.2

AUTHORIZE UTILITY

The system manager runs the AUTHORIZE utility to modify the existing
UAF or to create a new UAF.
Table 2-1 summarizes the AUTHORIZE
commands. The ADD, DEFAULT, and MODIFY commands act upon individual
fields of UAF records through the specification of appropriate
qualifiers.
Table 2-2
lists
the
qualifiers,
describes
the
corresponding fields, and specifies the defaults (as provided in the
DEFAULT record in the software distribution kit). Table 2-3 specifies
the values of the qualifiers in the SYSTEM, FIELD, and SYSTEST records
(as provided with the software distribution kit).
Table 2-1
AUTHORIZE Command Summary
Function

Format
ADD username [qualifier ••• ]

Adds a record to the UAF

DEFAULT qu"ali fier [ ••• ]

Modifies the DEFAULT record

EXIT

Returns the user to DCL
level

HELP

Lists the AUTHORIZE commands

LIST [userspec] [/FULL]

Creates a listing
records

MODIFY username qualifier [ ••• ]

Modifies a record in the UAF

REMOVE username

Deletes a record from the UAF

SHOW userspec [/BRIEF]

Displays UAF records

2-2

file

command

UAF

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Table 2-2
UAF Field Qualifiers
Qualifier

Default

/ACCOUNT=account-name

blanks

Default account name (for
example, a billing name or
number) -- 1-8 characters

/ASTLM=value

AST queue limit (total
asynchronous system trap (AST)
operations and scheduled
wake-up requests that can be
outstanding at one time) -integer; should be minimum of
2

/BIOLM=value

Buffered I/O count limit
(total buffered (for example,
terminal) I/O operations that
can be outstanding at one
time) -- integer; should be
minimum of 2

/BYTLM=value

4096

Buffered I/O byte limit (total
bytes that can be specified
for transfer in outstanding
buffered I/O operations) -integer; should be minimum of
1024

/CL I =c l i -name

DCL

Name of the default command
interpreter -- 1-9 characters;
should be DCL or MCR

/CPUTIME=delta-time

Maximum CPU time that a user's
process can take per
session -- unit of time in
delta format; 0 means
INFINITE

/DEVICE=device-name

blanks

Name of default device (must
be a direct-access device) -1-15 alphanumeric characters;
colon automatically added if
omitted; at log-in time, a
blank value is interpreted as
the equivalent of SYS$SYSDISK

Description

/DIOLM=value

Direct I/O count limit (total
direct (usually disk) I/O
operations that can be
outstanding at one
time) -- integer; should be
minimum of 2
(continued on next page)

2-3

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Table 2-2 (Cont.)
UAF Field Qualifiers
Qualifier

Default

Description

/DIRECTORY=
directory-name

[USER]

Name of default directory -1-31 characters; brackets
must be typed to be included

/ENQLM=value

Unimplemented

/FILLM=value

Open file limit (total files
that can be open at one time,
including active network
logical links) -- integer;
should be minimum of 2

/FLAGS=
( [[NO] DISCTLY]
[, [NO]DEFCLI]
[, [NO] LOCKPWD] )

blanks

Log-in flags to disable
CTRL/Y interrupts (if the
intent of DISCTLY is only to
force execution of the login
command procedure, that
procedure should issue a SET
CONTROL Y command before
exiting}, restrict the user to
the default command
interpreter (prohibit use of
the /CL! qualifier at log~in
time; however, the MCR
command can still be used),
and lock the user's password
(prohibit use of the SET
PASSWORD command) -- DISCTLY,
DEFCLI, and LOCKPWD; NO in
front of a flag name clears
the flag; parentheses may be
omitted for a single flag
name; multiple flag names
must be separated by commas

/LGICMD=filespec

blanks

Name of login command
file -- standard file
specification with defaults as
follows: device is as
specified for /DEVICE,
directory is as specified for
/DIRECTORY, file type is COM
(if command interpreter is
DCL) or CMD (if command
interpreter is MCR)

/OWNER=owner-name

blanks

Name of owner (for billing
purposes, for example) -- 1-32
characters
(continued on next page)

2-4

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Table 2-2 (Cont.)
UAF Field Qualifiers
Qualifier

Default

Description

/PASSWORD=password

USER

Password for logging in -1-31 characters; must be
alphanumeric characters,
dollar signs, or underlines

/PBYTLM=value

Unimplemented

/PGFLQUOTA=value

10000

Paging file limit (total pages
that a user process can use in
the system paging
file) -- integer; should be
minimum of 256

/PRCLM=value

Subprocess creation limit
(total subprocesses that can
exist at one time) -- integer

/PRIO=value

Default base priority for user
processes -- integer; should
be in the range of 0-31; 4 is
the norm for timesharing users

/PRIVILEGES=
([NO] privname [, ••• ])

NETMBX
TMPMBX

Privileges allotted user -names of privileges as
detailed in Chapter 4 (see
Table 4-1 for quick
reference); NO in front of a
privname removes the
privilege; parentheses may be
omitted for a single privname;
multiple privnames must be
separated by commas

/SHRFILLM=value

Unimplemented

/TQELM=value

Timer queue entry limit (total
entries in the timer queue
plus the number of temporary
common event flag clusters
that the user's process can
have at one time) -- integer

/UIC=uic

[200,200]

User identification code as
explained in Chapter 3 -group number and member number
separated by a comma and
enclosed in brackets; each
number must be octal in the
range of 0-377
(continued on next page)

2-5

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Table 2-2 (Cont.)
UAF Field Qualifiers
Qualifier

Default

Description

------ --- ..•. . --- -··--······-·· ·-···-···········---·--···-····· ···---··-··········-············-··· ··---··-········---· ·-···--· ·-·-. ··-- -·-- .................................. ...._.....
.

....

/WSDEFAULT=value

150

Default working set limit
(initial limit of a working
set for the user process) -integer;
should be minimum of
50

/WSQUOTA=value

200

Default working set quota
(maximum to which the user's
process may raise working set
limit) -- integer; should 'be
minimum of 50

Table 2-3
Initial Values of SYSTEM, FIELD,
' and SYSTEST Records
Qualifier

SYSTEM

FIELD

SYSTEST

/ACCOUNT

SYSTEM

FIELD

SYSTEST

/ASTLM

/BI OLM

/BYTLM

20480

10240

20480

/CL!

DCL

/CPUTIME

/DEVICE

blanks

/DIOLM

/DIRECTORY

[SYSMGR]

[SYSMAINT]

[SYSTEST]

/ENQLM

/FILLM

/FLAGS

blanks

/LGICMD

blanks

/OWNER

SYSTEM MANAGER

FIELD SERVICE

SYSTEST-UETP

(Continued on next page)

2-6

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)

Table 2-3 (Cont.)
Initial Values of SYSTEM, FIELD,
and SYSTEST Records
Qualifier

SYSTEM

FIELD

SYSTEST

/PASSWORD

MANAGER

SERVICE

UETP

/PBYTLM

/PGFLQUOTA

10000

/PRCLM

/PRIO

/PRIVILEGES

all

GRPNAM ALL SPOOL
GROUP DIAGNOSE
PRMCEB LOG IO
SETPRV TMPMBX
PRMMBX PHY IO
NETMBX

CMKRNL CMEXEC
SY SN AM GRPNAM
DETACH LOG IO
GROUP PRMCEB
PRMMBX TMPMBX
NETMBX VOL PRO
SYSPRV PHY IO
DIAGNOSE -

/SHRFILLM

/TQELM

/UIC

[001,004]

[001,010]

[001,007]

/USRDATOFF

/WSDEFAULT

150

/WSQUOTA

1024

Typically, the system manager issues the following commands in setting
up a new UAF:
•

MODIFYs -- Modifies the passwords in the SYSTEM, FIELD, and
SYSTEST records (FIELD and SYSTEST can be deleted if no longer
needed)

•

DEFAULT -- Modifies values in the DEFAULT record
DIGITAL-provided values are not appropriate

•

ADDs -- Adds new users

where

Modification of passwords is essential to prevent unauthorized
to system files.

2-7

the

access

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
2.2.l

Invoking AUTHORIZE

The following commands invoke the utility:
$ SET DEFAULT SYS$SYSTEM
$ RUN AUTHORIZE
It is essential that the system manager's default device and directory
be that of the system to access the system UAF.
(Users may create and
maintain UAF files in their own directories, using AUTHORIZE.
Such
files, however, have no effect on system operation.) As long as a UAF
exists, the system responds with the following prompt:
UAF>
The system manager can then enter any of the commands listed in Table
2-1.
These commands follow the standard rules of grammar as specified
in the VAX/~~,§__(;. o_ITl_!rl~!:15~- _f:._~ ng_~~9~____!:!~~_!-~-~....q__':l id~.
If no UAF exists (that is, the system manager has deleted
system issues an error message and the following prompt:

it),

the

Do you want to create a new file?
A response of YES (or Y) results in creation of a new UAF cont~ining a
DEFAULT record and a SYSTEM record. These records are initialized
with the values shown in Tables 2-2 and 2-3.
Use
of
the
AUTHORIZE
utility
requires
write
access
to
SYS$SYSTEM:SYSUAF.DAT, which is normally restricted to users in the
same group as the owner, or users with a system UIC or the SYSPRV
privilege.
The system manager can further restrict access to the UAF
(to just himself, for example) with the DCL command SET PROTECTION.
The system manager should not, however, expand access rights to the
UAF.

2.2.2

ADD Command

ADD adds a user record to the UAF.

It takes the form:

ADD username [qualifier ••• ]
user name
A string of 1 through 12 alphanumer:::ic chara'cters and underlines
(dollar signs are permissible, but usually reserved for system
names)
qualifier
A qualifier as listed in Table 2-2

2-8

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Qualifiers not specified take their values from the DEFAULT record
except that the default password is always USER (no matter what the
password in the DEFAULT record is).
Typically, the system manager
takes
defaults
on
the
limits, priority, privileges, command
interpreter, and sometimes device; as a result, the system manager
types only the password, UIC, directory, owner, account, and sometimes
device. The following example illustrates a typical ADD command:
UAF>ADD ROBIN /PASSWORD=SP0152 /UIC=[Ol4,00n]/DEVICE=DB1 /DIRECTORY=[ROBIN]~OWNER=CHRISTOPHER ROBIN /ACCOUNT=VMS
After the UAF record is successfully added, the system manager should
create a first-level directory for the new user, specifying the device
name, directory name, and UIC of the UAF record. Protection for the
"ordinary" user is normally read, write, execute, and delete access
for system, owner, and group processes, and read and execute access
for world processes.
The following command creates a first-level
directory for user Robin:
$CREATE DBl: [ROBIN] /DIRECTORY /OWNER_UIC=[Ol4,00n]

2.2.3

DEFAULT Command

DEFAULT modifies the DEFAULT record.

It takes the form:

DEFAULT qualifier [ ••• ]
qualifier
A qualifier as listed in Table 2-2
Qualifiers not specified in the command remain unchanged.
The system manager typically modifies the DEFAULT record
when
qualifiers
normally
assigned
to a new user differ from the
most
often
needing
DIGITAL-supplied values.
The
qualifiers
modification are as follows:
•

/CLI

If the command interpreter is MCR

•

/DEVICE

•

/LGICMD
Where automation of initial housekeeping chores
login time is desired

•

/PRIVILEGES -- Where users are normally
privileges than the DIGITAL-supplied ones

If most users have the same default device

given

different

The system manager almost always specifies a login command file.
At
some installations, the system manager simply provides each user
directory with its own login command file by specifying /LGICMD as
follows:
/LGICMD=LOGIN

2-9

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
(The name of the file does not have to be LOGIN, although this
convention
predominates.)
Login
protocol,
then,
becomes the
responsibility of each user.
Another approach provides a common login
directory:

command

file

system

/LGICMD=SYS$SYSDISK: [SYSMGRJSYLOGIN
The system manager maintains a login command file on behalf of all
users.
This command file may, at the end, contain a call to the
user's login file:
$ @LOGIN

(Of course, in this case, all users must name their login com~and
files LOGIN.COM.) In this way, the log-in protocol consists of a
common command procedure followed by
a
user-specific
command
procedure.
The following example illustrates the DEFAULT command:
UAF>DEFAULT /DEVICE=DBl /LGICMD=SYS$SYSDISK: [SYSMGR]SYLOGIN_/PRIVILEGES=(TMPMBX,GRPNAM,GROUP)
After modifying the default value record, the system manager
pencil in the changes in the middle column of Table 2-2.

2.2.4

should

EXIT Command

EXIT returns the system manager to DCL command level.
form:

takes

the

EXIT
The system manager can also return to
<CTRL/Z>.

2.2.5

DCL

command

level

typing

HELP Command

HELP lists and explains the AUTHORIZE commands.

It takes the form:

.HELP

2.2.6

LIST Command

LIST creates a listing file of reports for selected UAF
takes the form:
LIST [userspec]

[/FULL]

2-10

records.

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
userspec
A user name, a wild card character (*), or a
wild card

defaults

UIC;

/FULL
Request for a full report;

omission results in a brief report

The name of the listing file is SYSUAF.LIS and can be printed with the
DCL command PRINT.
Specification of a user name results in a one-user
following example lists a full report for user ROBIN:

report.

The

UAF>LIST ROBIN /FULL
Specification of a wild card character results
users in ascending sequence by user name:
UAF>LIST

reports

for

all

Specification of a UIC results in reports for all users with the UIC.
Wild card characters (*) can be used in specifying the UIC, as
illustrated in the following examples:

Command

Description

LIST [014,006] /FULL

Lists a full report for the user
(or users) with member number OOo
in group 014

LIST [014,*]

Lists a brief report for all users
in group 014, in ascending sequence
by member number

LIST [* ,006]

Lists a brief report for all
with a member number of 006

LIST [*,*]

Lists a brief report for all users,
in ascending sequence by UIC

users

Users with the same UIC are listed in the order that they were added
to the UAF.
Full reports include the details of the limits,
privileges, login flags, and cpmmand interpreter.
Brief reports do
not include the limits, login flags, and command interpreter, and
summarize the privileges. The password is never listed.
See Figure
2-1 under the SHOW command for examples.

2.2.7

MODIFY Command

MODIFY changes values in a user record.
MODIFY username qualifier [ ••• ]
user name
Name of user in the UAF
qualifier
A qualifier as listed in Table 2-2

2-11

It takes the form:

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Values not specified in the command remain unchanged.
example changes the password for user ROBIN:

The

following

UAF>MODIFY ROBIN /PASSWORD=SP0172
Modifications to UAF records do not affect users logged
modifications take effect the next time the user logs in.

2.2.8

in.

The

REMOVE Command

REMOVE deletes a user record from the UAF.

It takes the form:

REMOVE username
user name
Name of a user in the UAF
The DEFAULT and SYSTEM records
example deletes user ROBIN:

cannot

deleted.

The

following

selected

UAF

records.

takes

UAF>REMOVE ROBIN

2.2.9

SHOW Command

SHOW displays reports
following form:

for

the

SHOW userspec [/BRIEF]
user spec
A user name, wild card character (*), or UIC
/BRIEF
Request for a brief report;

omission results in a full report

Specification of a user name results in a 1-user report.
following example displays a full report for user ROBIN:

The

UAF>SHOW ROBIN
Specification of a wild card character results
users in ascending sequence by user name:
UAF>SHOW

* /BRIEF

2-12

reports

for

all

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Specification of a UIC results in reports for all users with the UIC.
Wild card characters (*) can be used in specifying the UIC, as
illustrated in the following examples:
Description

Command
SHOW [ 014, 0 0 6]

Displays a full report for the user
(or users) with member number 006
in group 014

SHOW [014,*] /BRIEF

Displays a brief report for all
users in group 014, in ascending
sequence by member number

SHOW [*,006] /BRIEF

Displays a brief report for all
users with a member number of OOn

SHOW [*,*] /BRIEF

Displays a brief report for all
users, in ascending sequence by UIC

Users with the same UIC are listed in the order that they were added
to the UAF.
Full reports include the details of the limits,
privileges, login flags, and command interpreter.
Brief reports do
not include the limits, login flags, and command interpreter, and
summarize the privileges. The password is never listed.
Figure 2-1
illustrates the display.

UAF>SHOW HUME /FULL
Username: ROBIN
Owner: CHRISTOPHER ROBIN
Account:
VMS
UIC:
[014,006)
CLI:
DCL
LGICMD: SYS$SYSDISK:[SYSMGR]SYLOGIN
Default Directory: DBl:[ROBIN]
Login Flags:
PRIO:
4 BYTLM:
4096 BIOLM:
6
PRCLM:
2 PBYTLM:
0 DI OLM:
6
ASTLM:
10 WSDEFAULT:
150 FILLM:
20
ENQLM:
0 WSQUOTA:
200 SHRFILLM:
0
TQELM:
10 PGFLQUOTA:
10000 CPU:
no limit
Privileges:
TMPMBX NETMBX
UAF>SHOW [014,*] /BRIEF
OWNER
USERNAME
UIC
Account Privs Pri Default Directory
CHRISTOPHER ROBIN
ROBIN
Normal 4 DB!: [ROBIN)
[014,006) VMS
DAVID J. JONES
JONES
[ 014 , 011 ] VMS
Normal 4 DB!: [JONES)
BOBBY SCHMAUTZ
SCHMAUTZ
[ 014 , 112 ] VMS
Normal 4 DBl:[SCHMAUTZ]

Figure 2-1 Display Produced By SHOW Command

2-13

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)

2.2.10

Messages

Table 2-4 lists the messages
alphabetical order).

issued

the

AUTHORIZE

utility

(in

Table 2-4
AUTHORIZE Error Messages
Message

Remedy

command is not unique

Enter command name in its entirety.

connect error

Respond to
follows.

device name too long
to add trailing ":"

Restrict device name to 15 characters.

do you want to create
a new file?

Respond YES (or Y)
Section 2.2.1.

or NO

(or N).

See

error during show

Respond to
follows.

VAX-11

RMS

message

that

error generating
listing file

Respond to
follows.

VAX-11

RMS

message

that

error in UIC
specification

Enter as [group,member] where group and
member are octal values in the range
000-377,
inclusive, or the wild card
character (*)

error in user
specification

Enter a

error in value
specification

Enter the value according to the specifications of Tables 2-1 and 2-2.

get error

Respond to
follows.

invalid command

Enter a valid command
Table 2-1.

invalid option name

Enter a valid qualifier as shown in
2-2.

Table

invalid password syntax

Enter 1-31 alphanumeric, dollar sign
or underline (_) characters.

($),

invalid privilege name

Enter a valid privilege name as
Table 4-1 (Chapter 4).

VAX-11

RMS

message

that

user
name, wild card character
LIST
or
SHOW
or
UIC.
See
specifications above.
(*),

VAX-11

RMS

message

name

that

shown

(continued on next page)

2-14

USER AUTHORIZATION FILE

AUTHORIZE (Cont.)
Table 2-4 (Cont.)
AUTHORIZE Error Messages

..--------------·----.--------- ----------------- - - - - - - - - - .
Message

Remedy
-----------~--------

you

that

Respond to
follows.

invalid username; username already exists

Enter a unique user name.

listing file SYSUAF.LIS
complete

Do not respond -- informational message.

missing argument for
option

Enter
2-2.

missing username

Enter a user name as listed in Table 2-1.

missing user
specification

Enter a user specification as listed in
Table 2-1.

no modifications made

Do not respond -- informational message.

no user matched
specification

Enter a user specification that matches
at least one existing user account.

option name not unique

Enter at least the first
of a qualifier.

four

privilege name
not unique

Enter at least the first
of a privilege name.

four characters

put error

Respond to the
follows.

quoted string missing
end quote

Enter the command with both quotes.

string too long
for field

Restrict character string to length specified in Table 2-2.

superfluous parameter
detected

Do not enter the superfluous text.

The DEFAULT record may
not be removed

Do not respond

The SYSTEM record may
not be removed

Do not respond -- informational message.

UAF modified

Do not respond -- informational message.

unable to add user
record

Respond to the
follows.

"Do

message

invalid response

qualifier

VAX-11

listed

RMS

Table

characters

message

that

informational message.

VAX-11

RMS

message

that

(continued on next page)

2-15

USER AUTHORIZATION FILE

AUTHORIZE {Cont.)
Table 2-4 {Cont.)
AUTHORIZE Error Messages
Message

Remedy

unable to create
SYSUAF.DAT

Respond to the
follows.

unable to delete
record

Respond
follows.

unable to obtain
DEFAULT record

Respond to the
VAX-11
RMS message that
follows.
If DEFAULT record cannot be
accessed, AUTHORIZE cannot be used to
modify the UAF.

unable to open
SYSUAF.DAT

Respond to the
follows.

VAX-11

RMS

message

that

unable to update
record

Respond to the
follows.

VAX-11 RMS

message

that

username does not
exist

Enter an existing user name.

username syntax error

Enter alphanumeric, underline, and
sign characters only.

username too long

Enter a user
characters.

user record removed

Do not respond

informational message.

user record
successfully added

Do not respond

informational message.

user record updated

Do not respond -- informational message.

value too large
for field

Enter a smaller value for the qualifier,
as specified in Table 2-2.

Warning: ADD command
does not use DEFAULT
record password

Do not respond -- informational message.

writing listing file

Do not respond -- informational message.

VAX-11

message

that

to the VAX-11 RMS message

that

name

RMS

longer

~ollar

than

---------------------------~

For explanations of the VAX-11 RMS messages, see
Message~_____and Recove
Procedures Manual.

2-16

the

VAX/VMS

System

USER AUTHORIZATION FILE
2.3

OPERATIONS

During normal log-in activities, the UAF
(that is, the file
SYS$SYSTEM:SYSUAF.DAT) controls user access of the system:
•

named

User validation -- To use the system at all, a user must state
(1) a name for which a record exists in the UAF and (2) the
password associated with that name. Otherwise, the following
message appears:
user validation error

•

Level of Access -- A user who successfully logs in receives
UIC access rights, a default file specification, a default
command interpreter, login flags, a default base priority,
resource limits, and privileges as stated in the UAF.

AUTHORIZE can run concurrently with user login operations as long as
VAX-11 RMS file sharing is in effect (that is, RMSSHARE has been run).
A slight chance exists that a user login operation might fail on a
record lock, but the user need only try again to remedy the situation.
The system manager can also implement the
normal login activities:
•

following

alternatives

Open system -- The system manager implements an open system by
deleting
(or renaming)
the UAF. When no UAF exists, a user
may log in with any name and password. The system assigns the
following values to such a user:
Name -- JOB CONTROL
UIC -- [010,040]
Command interpreter -- DCL
Login flags -- None
Priority -- Value of SYSGEN parameter DEFPRI
Resources -- Values of PQL SYSGEN parameters
Privileges -- All
The process name is normally the name of the device
the user logged in.

•

which

Inaccessible UAF -- If the UAF is locked or disabled, the
system manager
or operator -- can log in on the console
terminal with any name and password.

2-17

USER AUTHORIZATION FILE
•

Alternate UAF -- At bootstrap time, the system manager can
select an alternate UAF named SYS$SYSTEM:SYSUAFALT.DAT by
using a system generation file in which the UAFALTERNATE
parameter has a value of 1, or by changing this value to 1
during a conversational bootstrap operation.
(See Chapter 13
for creating system generation files.) Naturally, a file named
SYSUAFALT in UAF format must exist at bootstrap time.
The
system manager can create or modify such a file with the
following commands:
$ SET DEFAULT SYS$SYSTEM
$ ASSIGN SYSUAFALT SYSUAF
$ RUN AUTHORIZE
Any time after booting the system with an alternate UAF, the
system manager can switch to the standard UAF (that is,
SYS$SYSTEK:SYSUAF.DAT) with the following command:
$ DEASSIGN SYSUAF /SYSTEM
(The system initially changes to the alternate UAF
by
assigning SYSUAF as a logical name for SYSUAFALT.) If the
UAFALTERNATE parameter is set at bootstrap time and no
alternate UAF exists, the system behaves as an open system.
The system manager can also switch to the alternate UAF
bootstrap time with the following command:

after

$ ASSIGN SYSUAFALT SYSUAF /SYSTEM
(For this mode of operation, the alternate UAF
name in the range of 1 to 9 characters.)

2.4

can

have

any

USER DATA AREAS

Users can access UAF records by means of VAX-11 RMS for storage and
retrieval of up to 255 bytes of user data. The format of a UAF record
is defined in the module $UAFDEF in SYS$LIBRARY:LIB.MLB. The records
can be accessed sequentially or through the following keys:
•

User name -- The primary key is the user name (as specified to
AUTHORIZE), a character field of size UAF~S USERNAME located
at relative offset UAF$T USERNAME.

•

UIC -- The secondary key is the UIC, located at relative
offset UAF$L UIC, consisting of two binary subfields of one
word each. Tne subfields are named UAF$W_GRP
(high-order
word) and UAF$W_MEM (low-order word).

•

Member -- The tertiary key is the member field of the UIC,
located at relative offset UAF$W_MEM, consisting of one binary
word.

AUTHORIZE uses the alternate keys to display UAF records in UIC order.
The tertiary key may be dropped in future releases.

2-18

USER AUTHORIZATION FILE
To place data in a UAF record, the user should take the following
steps, which are designed to protect programs against future changes
to the format of the UAF:
1.

Read the UAF record.

Check the value of UAF$W USRDATOFF. If it is zero, insert
the current size of tne record, as found in the VAX-11 RMS
record access block (RAB), into UAF$W USRDATOFF.
(In VAX/VMS
Version 2.0, the system initializes this field to zero.
Inserting the current size of the record has the effect of
placing the user data at the end of the record. However,
future changes to the UAF might require the system to fix the
location of the user data. In this event, the system would
initialize UAF$W USRDATOFF and the user must not change its
value.)
-

Insert the user data at the relative offset pointed to by
UAF$W USRDATOFF.
The data must take the form of a counted
string -- the first byte must specify the number of bytes
that follow and the total bytes must not exceed 256.

Modify the size of the record in the RAB to reflect the
addition of the user data area, unless the current size of
the record exceeds the end of the user data area
(that is,
the contents of UAF$W USRDATOFF plus the length of the user
data).
(In Version 2.0~ the user data area, when specified
as outlined above, resides at the end of the record, so that
modification of the RAB is essential.
It is possible,
however, that a future version of AUTHORIZE might fix ·the
size (at 256 bytes) and location of the user data area, and
locate additional system data after the user data. In this
event, the user must not modify the size of the record.)

Update the record.

The user can now access the counted string by
referring
to
UAF$W USRDATOFF.
For additional updates, the user does not modify
UAF$W-USRDATOFF (the user should not modify UAF$W USRDATOFF if its
value- is not zero), but does modify the first byte of the counted
string and the record size in the RAB (withstanding the considerations
for future changes) if the record size changes. The user should lock
records while accessing them.
The format of the UAF record and the way in which the system modifies
it is subject to change without notice.
(For format changes, however,
a utility will be provided to update old UAFs.) Adherance to the above
guidelines will minimize reprogramming in the event of system changes.
pg

2.5

UAF VERSION 2.0 UPGRADE

The format of the UAF has changed for VAX/VMS Version 2.0 from
sequential
to
indexed
organization.
The installation upgrade
procedure (see the VAX-11 Software Installation Guide) automatically
upgrades SYS$SYSTEM:SYSUAF.DAT to the proper format.
The system
manager can upgrade other UAFs by invoking the command procedure
SYS$SYSTEM: [SYSUPD]CVTUAF as follows:
$ SET DEFAULT SYS$SYSTEM
$ @SYS$SYSDISK: [SYSUPD]CVTUAF

2-19

USER AUTHORIZATION FILE
The old UAF must be accessible to the command procedure. If the old
system volume cannot be mounted on a second disk drive (for example,
in a single-disk system), the system manager must be sure to include
the old UAF with the user files that are saved on another medium at
installation time.
In response to a prompt from the command procedure, the system manager
types the name of the device on which the old UAF resides.
Execution of the command procedure requires read access to the old UAF
and write access to SYS$SYSTEM.
In Version 2.0, the UAF is accessed with file sharing enabled, and
updates to the UAF are made on a per-record basis, obviating the need
for a temporary UAF and a new version of the UAF after each AUTHORIZE
run.
Updates become effective as soon as AUTHORIZE commands are
entered, not after the termination of AUTHORIZE.
(For this reason,
the system manager should not enter temporary values with the intent
of fixing them later in the run.)

2-20

CHAPTER 3
PROTECTION AND CONTROL

The system manager assists in data protection and interprocess control
primarily in the following ways:

3.1

•

User identification code (UIC) -- By assigning each user a UIC
that reflects the user's relationships with other users

•

Passwords -- By changing the passwords of the SYSTEM, FIELD,
and SYSTEST records in the UAF, and by assigning confidential
passwords to users of the system

•

Privileges
need them

those

who

•

System data and devices -- By applying proper protection
system data files, interactive terminals, and card readers

By strictly limiting

privileges

USER IDENTIFICATION CODE

A UIC consists of two numbers in the range of 0 through 377 (without
the radix indicators), separated by a comma and enclosed in brackets.
The first number is the group number of the UIC, and the second number
is the member number.
For example, [320,101] is a UIC whose group
number is 320 octal and whose member number is 101 octal.
The system manager explicitly assigns a UIC to each user in the UAF.
The system implicitly assigns this UIC to the detached process created
for the user at login time. User processes pass on this UIC to any
subprocesses
they
create.
Detached processes created by user
processes require explict assignment of a UIC.
Processes can further assign UICs to files,
file-structured volumes,
devices, mailboxes, shared pages in memory (global sections), common
event flags, and logical name tables. For most users, the assignment
of UICs to these structures is automatic and transparent. Suppose,
for example, that a user with UIC [320,101] logs in and issues the
following command:
$ CREATE SCRATCH.RNO

The system assigns a UIC of [320,101] to
default directory.

3-1

SCRATCH.RNO

the

user's

PROTECTION AND CONTROL
3.2

PROTECTION OF DATA STRUCTURES AND DEVICES

For purposes of protection, four different categories
with respect to data structures and devices:

user

exist

•

Owner -- A user whose UIC is identical with the UIC of the
data structure or device. The owner of a file is ordinarily
the creator of the file.

•

Group -- A user whose group number is identical with the group
number of the data structure or device.

•

System
A user whose group number is between 0 and the value
of the MAXSYSGROUP system parameter
(usually 10 octal),
inclusive.

•

World -- Any user. This category includes users who are not
owners of the data structure or device in question, are not
members of the same group as the owner, and are not system
users.

Figure 3-1 illustrates the relationships of these four types
to a file whose UIC is [100,100].

WORLD
(All UICs)

g = Group Number
m = Member Number

Figure 3-1 Classification of Users with Respect
to a File with a UIC of [100,100]

3-2

user

PROTECTION AND CONTROL
Users are permitted four types of access to protected data
and devices:

structures

•

Read

•

Write

•

Execute (or physical
devices)

I/O

for

shareable

non-file-structured

•

Delete (or
devices)

I/O

for

shareable

non-file-structured

logical

Each category of user is permitted or denied each type of access
through association of the data structure or device with a 16-bit
protection mask of the following form:
DE LET&
EXECUTE
WRITE
READ

WORLD
15
14
13
12

GROUP OWNER
11
7
10
6
9

SYSTEM
3
2
1

5
4

When a bit is on, that category of user is denied that type of access
to the protected data structure or device. For example, if bits 12,
13, and 15 of a file's protection mask are on, world users are denied
read, write, and delete access to that file.
DCL commands permit the user to set values for protection masks in the
following form:
(user-category:access-categories[, ••• ])
user-category
OWNER, GROUP, SYSTEM, or WORLD
access-categories
Any combination of R (read), W (write), E (execute;
physical I/O), or D (delete; or L for logical I/O)

for

R, W, E, and D represent the access categories permitted the user
category (that is, the categories for which the mask bits are set
off). In the following example, a user creates a file that denies
read, write, and delete access to the world:
$ CREATE SCRATCH.RNO -

/PROTECTION=(SYSTEM:RWED, -OWNER:RWED,GROUP:RWED,WORLD:E)
System and I/O services require specification of the mask as four
hexadecimal digits. The low-order digit specifies system access, the
next digit specifies owner access, the next digit specifies group
access, and the high-order digit specifies world access. In the
following example, a user creates a mailbox that denies read and write
access to the world:
$CREMBX S PRMFLG=#O,CHAN-MBXCHAN,MAXMSG=MBUFLENBUFQU0=#384,PROMSK=#~X3000,LOGNAM=MBLOGNAM

The protection mask is specified as hexadecimal
low-order bits of the high-order digit are on.

3000 -- the

The following sections describe the protection provided
the data structures and devices.

3-3

for

each

two
of

PROTECTION AND CONTROL
Files and File-Structured Volumes

3.2.1

The default Files-11 protection is to deny write and delete access to
the world and to allow all other categories of access.
A user can
reset the default Files-11 protection:

$ SET PROTECTION-(WORLD:E)

/DEFAULT

Default protection for the world permits execute
the other user categories do not change).
The default protection is applied to all files
unless explicitly overridden:

access

only

created

(while

the

user

$ CREATE SCRATCH.RNO /PROTECTION-(SYSTEM:RWED,_OWNER:RWED,GROUP:RWED,WORLD:RWE)
On an explicit specification, all categories must be included.
The protection on a file
created:

can

also

changed

after

the

file

$ SET PROTECTION=(WORLD=R) SCRATCH.RNO
The world has only read access to SCRATCH.RNO (while
categories do not change).

the

other

user

Figure 3-2 illustrates how protection works to control access
to
file that has a UIC of (100,100] and default Files-11 protection.

Application Program
Development Group,
g= 100,
Consisting of
Three Users

File Created by User
with UIC of (100,100)
(the Owner) with Default
Files-11 Protection

+
r--------,
GROUP
I
OWNER
(100,100)

RWED

(7, 10]

System
Users

AWED

[100,101]

UIC=

RWED

(1,20]

[100,100]

~_fl_WED

[200, 10]
One of All Other
Users - the World

L ________ J
Types of Access:

Read

W: Write
E: Execute
D: Delete

Figure 3-2 Controlling Access to a File
3-4

PROTECTION AND CONTROL

The scheme for protecting file-structured volumes is similar to the
one for protecting files, except that execute access to a volume gives
a user the right to create.files on the volume.
Users who have been assigned the VOLPRO privilege can override the
protection of file-structured volumes and users who have been assigned
the BYPASS privilege can bypass the UIC protection mechanism for all
files. In addition, the SYSPRV privilege grants users the same access
rights to files and file-structured volumes as system users.
(See
Chapter 4 for descriptions of all privileges.)

3.2.2

Nonshareable Non-File-Structured Devices

Nonshareable devices not used for storing files
(for
example,
terminals and card readers) can be protected from allocation by
setting or not setting the read bits in the protection masks
ass0ciated with the devices.
(The other bits have no meaning in
protection masks for nonshareable
non-file-structured
devices.)
Protection can be applied to all terminals at bootstrap time with the
TTY PROT and TTY OWNER system parameters (see Chapter 12) and on a
device-by-device-basis during system operation with the SET PROTECTION
/DEVICE command {see the VAX/VMS
Operator's
Guide).
Setting
protection on a terminal does not inhibit logins, but restricts the
processes that can accept and display data on the termianl.
Users who have been assigned the BYPASS privilege can bypass the UIC
protection mechanism for all devices. In addition, users with the
SYSPRV privilege have the same access rights to devices as system
users.

3.2.3

Mailboxes

Mailboxes {both temporary and permanent) are protected by a code, or
mask, that is similar to the code used in protecting files. As with
files, four types of user {defined by UIC) can gain access to a
mailbox: owner, group, system, and world. However, only two types of
access -- read access and write access -- are meaningful to users of a
mailbox.
Thus, when creating a mailbox, the user can specify read
access and write access to the mailbox separately for each type of
user.
Because temporary mailboxes are customarily used for interprocess
communication between cooperating processes that have the same group
number, they are of special interest here. The logical names of these
mailboxes are entered in the group logical name table. Temporary
mailboxes thus have two layers of protection. First, easy access to
the logical names of temporary mailboxes is granted only to users who
have the same group number as the creator of the mailbox; other users
have no such easy access. Second, by use of the protection mask, the
creator of the temporary mailbox grants additional security to the
mailbox.
As a rule, users who are not in the same group as the
creator are totally excluded from using the mailbox.
Furthermore, the creator of a temporary mailbox can discriminate
between owners and other group members in granting read access and
write access to a temporary mailbox.
For example, owners may be
allowed only read access or only write access to the mailbox, but
other members of the group may be allowed both read access and write
access to the mailbox.

3-5

PROTECTION AND CONTROL
Shared Pages in Memory (Global Sections)

3.2.4

All shared pages in memory (global sections) are protected by a code,
or mask, that is the same as the one used to protect other files.
Thus, the creator of a global section can specify read access, write
access, execute access, and delete access separately for each of the
four types of user.
Of special interest are group global sections; such a section can be
shared only by processes that have the same group number as the global
section. This group number is that of the creator of the global
section.
Thus, group global sections have two layers of protection.
First, if a process is to map to a global section, the group numbers
of the process and of the global section must be identical. Second,
by using the protection mask, the creator of a group global section
grants additional security to that global section.

Common Event Flags

3.2.5

Common event flags, which are used in establishing communication and
synchronization among cooperating processes in the same group, are
protected by UIC.
When a common event flag cluster is created, it takes on the UIC of
the process that requested its creation and, thus, the UIC of the user
whose process created the cluster. To control access to a common
event flag cluster, a protection indicator is used with the UIC. If
the value of this indicator is 0 (the default), both read access and
write access to the cluster are extended to all processes and users in
the owner's group. If the value of this indicator is 1, both read
access and write access to the cluster are restricted to owners of the
cluster and to subprocesses that have been created on behalf of these
owners.

Group Logical Name Table

3.2.6

Entries in the group logical name table are protected by group number.
When a logical name and its equivalence name are entered into the
group logical name table, the group number of the process that created
the logical name is associated with the entry. Access to an entry in
the group logical name table is thus restricted to users and processes
that have the same group number as the entry.

3.3

FORMING GROUPS

In setting up a group, the system manager aims toward two goals:
•

Sharing -- Users who typically share data and/or control
another's processes should be assigned to the same group.

•

Protection -- Users who should not have access to each other's
data or control over each other's processes should be assigned
to separate groups.

3-6

one

PROTECTION AND CONTROL

In addition to group number, user privilege determines the ability
one process to control another process or access its data:

•

No privilege -- A process needs no special privilege to access
its own data structures or control its own subprocesses.

•

GROUP privilege -- A process needs GROUP privilege to control
processes belonging to other members of the same group.

•

WORLD privilege -- A process needs WORLD privilege to
processes belonging to users outside one's group.

control

The importance of properly setting up groups of users should not be
underestimated. As the system is used more and more and relationships
among users, their processes, and their data structures become
increasingly complex, redefining groups becomes more difficult.
Normally, the system manage· should not assign the same UIC to more
than one user, and the status of system user should be limited to the
SYSTEM, FIELD, and SYSTEST accounts.

3.4

PASSWORDS

It is essential that the system manager change the passwords for the
SYSTEM, FIELD, and SYSTEST records to values that are reasonably
obscure, are difficult to deduce, and exceed
six
characters.
Similarly, the system manager should afford users of the system
individ~al protection by assigning obscure
passwords and suggesting
that users change their passwords from time to time.

3.5

PRIVILEGES

Chapter 4 discusses privileges in detail. The system manager should
not assign privileges above the category of normal except where (1)
the user demonstrates a need for the privilege and (2) the user is not
likely to misuse the privilege.
In addition, the system manager
should not install user's images with amplified privileges (see
Chapter 7) unless the above conditions are met.

3.6

SECURING SYSTEM DATA AND DEVICES

The system manager (UIC [1,4]) should own all
system
files.
Protection on all system executable and library files should permit
only read and execute access to the world.
Protection on the
following files should permit no access to the world:
SYS$SYSTEM:PAGEFILE.SYS
SYS$SYSTEM:SWAPFILE.SYS
SYS$SYSTEM:SYSDUMP.DMP
SYS$SYSTEM:SYSUAF.DAT
SYS$SYSTEM:RMTNODE.DAT
These files contain sensitive data, allowing a user with
to them a degree of control over the system.

read

access

The software distribution kit provides the required protection for
these files.
However, the system manager could easily compromise a
sensitive system file by copying it without applying the proper

3-7

PROTECTION AND CONTROL

protection.
always:

prevent

such a situation, the system manager should

•

Copy utilities -- Use the VAX-11 utilities DSCl and DSC2.
These utilities preserve the protection assigned to files.
If
another utility (RMS-11 BACKUP or RESTORE, for example) is
used,
the
system manager should immediately check the
protection on the copied file (DIRECTORY /PROTECTION command)
and apply the proper protection (SET PROTECTION command) as
needed.

•

COPY command
Log in as the system manager (UIC [1,4]) and
set the protection explicitly when copying system files (that
is, use the /PROTECTION qualifier).

These procedures require a conscious effort on the part of the system
manager to
(1)
realize that a sensitive system file is being copied
and (2) actively ensure that the copy is protected the same as the
original.
The system manager should protect interactive terminals and card
readers from allocation by processes other than system processes. If
general access to these devices is permitted, a user's process could
tie up another user's terminal, or -- worse -- monitor another user's
input to pick up such sensitive data as the user's password.
The
system manager can restrict the allocation of all terminals to system
processes through the following system parameter specifications:
PROT -- Specify the value of TTY PROT as %Xlll0.
This
• TTY
protection mask allows only system processes to allocate
terminals.
•

TTY OWNER -- Specify the value of TTY OWNER
UIC-of the system manager.

%Xl000~,

the

Protection on noninteractive terminals can be reset individually with
SET PROTECTION /DEVICE commands.
(These commands should be placed in
the site-specific start-up command procedure; see Chapter 8.) The SET
PROTECTION /DEVICE command should also be used to protect card readers
from allocation by nonsystem processes.

3-8

CHAPTER 4
RESOURCE CONTROL

This chapter contains detailed descriptions of the resource control
attributes that the system manager assigns to a user when creating a
record in the UAF:
•

Limits on the use of reusable system resources

•

Base priority used in scheduling the process that
creates for the user

•

Privileges of using restricted and sensitive system functions

the

system

In addition, this chapter discusses the accounting log file.

4.1

LIMITS

Limits are set on system resources that can be reused, for example,
the amount of memory that a process can have in use for queued I/O
requests. Each user of the system is limited in the use of system
resources.
The system manager sets up these limits when the user is
defined to the system. Most limits restrict the use of system memory.
Limits can be either pooled, deductible, or nondeductible in the way
in which a process shares its allotment of a resource with the
subprocesses that it creates. If the limit on the use of a resource
is pooled, a process and created subprocesses share the total limit on
a first-come first-served basis.
If the limit on the use of a
resource is deductible, a subprocess is allotted a portion of the
total limit; the portion given to the subprocess is deducted from the
total limit.
If the limit is nondeductible, the subprocess is
allotted the total limit of the creating process;
there is no
deduction from the allotment of the creating process.
In creating a UAF record, the system manager
following limits:

• AST queue limit (ASTLM)
• Buffered I/O count limit (BI OLM)
• Buffered I/O byte count limit (BYTLM)
• CPU time limit (CPUTIME)
• Direct I/O count limit (DIOLM)
• Open file limit (FILLM)
4-1

assigns

values

the

RESOURCE CONTROL
•

Paging file limit (PGFLQUOTA)

•

Subprocess creation limit (PRCLM)

•

Timer queue entry limit (TQELM)

•

Default working set size (WSDEFAULT)

•

Working set quota (WSQUOTA)

These limits are described in the following sections.
Usually, the
system manager simply assigns the default values for these limits.
However, see Chapter 11 for a discussion of WSDEFAULT and WSQUOTA.

AST Queue Limit (ASTLM)

4.1.l

The AST queue limit (ASTLM) limits the sum of the following:
•

The number of asynchronous system trap (AST) requests
user's process can have outstanding at one time

that

•

The number of scheduled wake-up requests that a user's process
can have outstanding at one time

This limit affects not only all system services that accept an AST
address as an argument, but also the Schedule Wakeup system service.
ASTLM is a nondeductible limit with a suggested minimum value of 2.

4.1.2

Buffered I/O Count Limit (BIOLM)

The buffered I/O count limit (BIOLM) limits the number of
buffered I/O operations permitted a user's process.

outstanding

A buffered I/O operation is an I/O operation in which the data
transfer takes place from an intermediate buffer in the system pool,
not from a process-specified buffer. Buffered operations for a user
process include terminal I/O, card reader input, and unspooled printer
output. During a buffered I/O operation, the pages containing the
process-specified buffer that is given as an argument to the Queue I/O
Request system service need not be locked in memory.
BIOLM is a nondeductible limit with a suggested minimum value of 2.

4.1.3

Buffered I/O Byte Count Limit (BYTLM)

The buffered I/O byte count limit (BYTLM) limits the amount of
space that a user's process can use.
This buffer space is used for buffered
creation of temporary mailboxes.

I/O

operations

and

buffer
for

BYTLM is a pooled limit with a suggested minimum value of 1024.

4-2

the

RESOURCE CONTROL
4.1.4

CPU Time Limit (CPUTIME)

The CPU time limit {CPUTIME) limits the amount of CPU time
user's process can use per interactive session or batch job.

that

The time must be specified in delta format -- hh:mm:ss.ss.
CPUTIME is a deductible limit with a suggested minimum
milliseconds {but not across sessions or jobs).

4.1.5

value

Direct I/O Count Limit (DIOLM)

The direct I/O count limit {DIOLM) limits the number
direct I/O operations permitted a user's process.

outstanding

A direct I/O operation is an I/O operation in which the data transfer
takes place directly from a process-specified buffer. Direct I/O
operations for a user process typically include disk and tape I/O.
The pages containing this buffer are locked in memory by the operating
system during the direct I/O operation.
DIOLM is a nondeductible limit with a suggested minimum value of 2.

4.1.6

Open File Limit (FILLM)

The open file limit {FILLM) limits the number of files that a user's
process can have open at one time. This limit includes the number of
network logical links that can be active at the same time.
FILLM is a pooled limit with a suggested minimum value of 2.

4.1.7

Paging File Limit (PGFLQUOTA)

The paging file limit {PGFLQUOTA) limits the number of pages that
user's process can use in the system paging file.

the

The paging file provides temporary disk storage for pages
of memory by a memory management operation.

out

forced

PGFLQUOTA is a pooled limit with a suggested minimum value of 256.

4.1.8

Subprocess Creation Limit (PRCLM)

The subprocess creation limit
(PRCLM)
subprocesses a user's process can create.

limits

the

number

The process that is created when a user logs in to the system can in
turn create subprocesses. These subprocesses are all accountable to
the user and share the resources allotted to the initial process.
PRCLM is a pooled limit with a suggested minimum value of O.

4-3

RESOURCE CONTROL
4.1.9

Timer Queue Entry Limit (TQELM)

The timer queue entry limit (TQELM} limits the sum of the following:
•

The number of entries that a user's process can
timer queue

•

The number of temporary common
user's process can have

This limit does not
clusters.

govern

the

event

creation

flag

have

clusters

permanent

the

that

event

flag

Timer queue entries are used in time-dependent scheduling;
common
event flags are used in synchronizing activities among groups of
cooperating processes.
TQELM is a pooled limit with a suggested minimum value of O.

4.1.10

Working Set Default (WSDEFAULT)

The working set default (WSDEFAULT} sets the initial working set limit
for a user's process.
WSDEFAULT is a nondeductible limit with a suggested minimum value of
SO.
If the value specified exceeds the value of WSQUOTA (below}, the
lesser value is used.

4.1.11

Working Set Quota (WSQUOTA)

The working set quota
(WSQUOTA}
restricts the size to which the
working set limit of a user's proc~ss can be expanded.
This
enlargement of the working set limit (the initial size of which is
determined by WSDEFAULT) is accomplished by use of the Adjust Working
Set Limit system service or by the SET WORKING_SET command.
WSQUOTA is a nondeductible limit with a suggested minimum value of SO.
If the value specified exceeds the value of the WSMAX system
generation parameter, the lesser value is used.

4.2

PRIORITY

A user's priority is the base priority used in scheduling the process
that the system creates for the user. There are 32 levels of software
priority in the VAX/VMS system, 0 through 31. The highest priority is
31;
the lowest is O.
The range of priorities for timesharing
processes is 1 through lS; the range for real-time processes is ln
through 31.
Processes with real-time priorities are scheduled strictly according
to base priority;
in other words, the executable real-time process
with the highest base priority is executed first.
Processes with
timesharing priorities are scheduled according to a slightly different
principle to promote overlapping of computation and I/O activities.

4-4

RESOURCE CONTROL

In the user's account record of the UAF, the default value of a user's
priority is 4;
for practical purposes, the minimum value is 1. The
priority for timesharing users should remain at
the
default.
Attempting to give some users an edge over other users by varying
priorities usually results in ragged performance, as the system reacts
sharply to even small priority differences. The system manager should
never specify a value
over
31
(system
operation
will
be
unpredictable).

4.3

PRIVILEGES

Privileges restrict the performance of certain system activities to
certain
users.
These
restrictions
protect the integrity of
performance of the operating system and thus the integrity of service
provided users.
The system manager should grant privileges to each
user on the basis of two factors:
(1) whether the user has the skill
and experience to use the privilege without disrupting the system and
(2) whether the user has a legitimate need for the privilege.
Privileges fall into seven categories according to the damage that the
user possessing them could cause the system:
•

None - No privileges

•

Normal - Minimum privileges to effectively use the system

•

Group - Potential to interfere with members of the same group

•

Devour - Potential to devour noncritical system-wide resources

•

System - Potential to interfere with normal system operation

•

File - Potential to compromise file security

•

All - Potential to control the system

A user cannot execute an image that requires a privilege the user does
not possess unless the image is installed as a known image with the
privilege in question.
(See Chapter 7 for instructions on installing
known images.) Execution of a known image with privileges temporarily
(for the duration of the image's execution) grants those privileges to
the user process executing the image.
A user's privileges are recorded in the user's UAF record in a n4-bit
privilege vector.
When a user logs in to the system, the user's
privilege vector is stored in the header of th~ user's process.
In
this way, the user's privileges are passed on to the process created
for the user. A user with the SETPRV privilege can modify the
privilege vector with the SET PROCESS /PRIVILEGES command.
Table 4-1 lists the privileges by category and gives brief, general
definitions of them. The following sections describe each privilege
in detail in alphabetical order.

4-5

RESOURCE CONTROL
Table 4-1
VAX/VMS Privileges
·--·--------.---·---·----------

Category

Privilege

Activity Permitted

- - + --------··------·---·-----------·--------········---------···
----- -+----- ----·------ --- -!

None

None requiring privileges

Normal

MOUNT

Execute mount volume QIO

NETMBX

Create network connections

TMPMBX

Create temporary mailbox

GROUP

Control processes in the same
group

ACNT

Disable accounting

ALLSPOOL

Allocate spooled devices

BUGCHK

Make bugcheck error log
entries

EXQUOTA

Exceed disk quotas

GRPNAM

Insert group logical names in
the name table

PRMCEB

Create/delete permanent common
event flag clusters

PRMGBL

Create permanent global
sections

PRMMBX

Create permanent mailboxes

SHMEM

Create/delete structures in
shared memory

!------·-------·-

Group
t----·--------

Devo u r

----··--------·

System

---·-·-·····--··----········--·-------·--·--··-------····-

ALTPRI

Set base priority higher than
allotment

OPER

Perform operator functions

PSWAPM

Change process swap mode

WORLD

Control any process

DIAGNOSE

Diagnose devices

SYSGBL

Create system-wide global
sections

VOL PRO

Override volume protection

t----·-·······---··---······---·--

F i le s

t----··--············---··-··--·····

All

-----·-···----········-·-------+------------------·

BYPASS

Disregard UIC protection

CMEXEC

Change to executive mode

CMKRNL

Change to kernel mode

-·---------

(continued on next page)

4-6

RESOURCE CONTROL

Table 4-1 (Cont.)
VAX/VMS Privileges
Category

Privilege

Activity Permitted

1-----------T---------- --- --·------- - - t - - - - - - - - - - - - - - -

All (Cont.)

4.3.1

DETACH

Create detached processes

LOG IO

Issue logical I/O requests

PFNMAP

Map to specific physical pages

PHY IO

Issue physical I/O requests

SETPRV

Enable any privilege

SYS NAM

Insert system logical names in
the name table

SYSPRV

Attain system user status

ACNT Privilege

Only a user who has the ACNT privilege can create subprocesses or
detached processes in which accounting is disabled. Thus, only such a
privileged user can issue the RUN command with the /NOACCOUNTING
qualifier or inhibit accounting in the $CREPRC system service.

4.3.2

ALLSPOOL Privilege

The ALLSPOOL privilege allows the user's process to allocate a spooled
device by executing the Allocate Device system service, or the user is
allowed to allocate a spooled device by using the ALLOCATE command.
The Allocate Device system service lets a process allocate, or
reserve, a device for its exclusive use. A shareable mounted device
cannot be allocated.
The system manager should grant this privilege only to users who need
to perform logical or physical I/O operations to a spooled device.
Ordinarily, the privilege of allocating a spooled device is granted
only to symbionts.

4.3.3

ALTPRI Privilege

The ALTPRI privilege allows the user's process to (1) increase its own
base priority and
(2) set the base priority of another process to a
value higher than its own base priority.
The base priority is increased by executing the Set Priority system
service or the SET PROCESS/PRIORITY command. As a rule, this system
service lets a process set its own base priority or the base priority
of another process. However, one process can only set the priority of
a second process if (1) the second process is a subprocess of the
first or (2) the first process has process control privilege (GROUP or
WORLD) over the second.

4-7

RESOURCE CONTROL

With the same privilege a process can create a process with a base
priority higher than its own. Such a process is created by using an
optional argument to the Create Process system service or to the RUN
command.
The system manager should not grant this privilege widely;
if many
users have the unrestricted ability to set base priorities, the fair
and orderly scheduling of processes for execution can easily be
disrupted.

4.3.4

BUGCHK Privilege

The use of this privilege should be restricted to system
supplied by DIGITAL that uses the VAX/VMS Bugcheck Facility.

4.3.5

software

BYPASS Privilege

The BYPASS privilege allows the user's process read, write,
and delete access to all files, byp~ssing UIC protection.

execute,

The system manager should grant this privilege with extreme caution,
as it suspends all file protection. It should be reserved for use by
well tested, reliable programs and command procedures.
SYSPRV (see
below)
is adequate for interactive users, as it ultimately grants
access to all files, while still providing access checks.

4.3.6

CMEXEC Privilege

The CMEXEC privilege allows the user's process to execute
Mode to Executive system service.

the

Change

This system service lets a process change its access mode to
executive, execute a specified routine, and then return to the access
mode that was in effect before the system service was called.
While
in executive mode, the process is allowed to execute the Change Mode
to Kernel system service.
The system manager should grant this privilege only to users who need
to gain access to protected and sensitive data structures and internal
functions of the operating system. If many users have unrestricted
access to sensitive data structures and functions, the operating
system and service to other users can easily be disrupted.
Such
disruptions can include failure of the system, destruction of the data
base, and exposure of confidential information to
unauthorized
persons.

4.3.7

CMKRNL Privilege

The CMKRNL privilege allows the user's process to execute
Mode to Kernel system service.

the

Change

This system service lets a process change its access mode to kernel,
execute a specified routine, and then return to the access mode that
was in effect before the system service was called.

4-8

RESOURCE CONTROL
The system manager should grant this privilege only to users who need
to execute privileged instructions or who need to gain access to the
most protected and sensitive data structures and functions of the
operating system.
If many users have unrestricted use of privileged
instructions and unrestricted access to sensitive data structures and
functions,
the operating system and service to other users can easily
be disrupted. Such disruptions can include failure of the system,
destruction of the data base, and exposure of confidential information
to unauthorized persons.

4.3.8

DETACH Privilege

The DETACH privilege allows the user's process to create detached
processes by executing the Create Process system service. Detached
processes remain in existence even after the user who created them has
logged off the system.
An example of a detached process is the process created by the
for a user when the user logs in to the system.

system

There is no restriction on the UIC that can be specified for a
detached process.
Thus, there are no restrictions on the files and
directories to which a detached process can gain access.

4.3.9

DIAGNOSE Privilege

The DIAGNOSE privilege allows the user to run online diagnostic
programs and to intercept and copy all messages that are written to
the error log file.

4.3.10

EXQUOTA Privilege

The EXQUOTA privilege allows the space taken by the user's files on
given disk volumes to exceed any usage quotas set for the user (as
determined by UIC) on those volumes.

4.3.11

GROUP Privilege

The GROUP privilege allows the user's process to affect other
processes in its own group by executing the following process control
system services: Suspend Process, Resume Process, Delete Process, Set
Priority, Wake, Schedule Wakeup, Cancel Wakeup, and Force Exit. The
user's process is also allowed to examine other processes in its own
group by executing the Get Job/Process Information system service.
The user with the GROUP privilege can issue SET QUEUE, DELETE/ENTRY,
STOP/ENTRY, and SET PROCESS commands for other processes in its group.
The GROUP privilege is not needed for a process to exercise control
over, or to examine, subprocesses that it created. The system manager
should, however, grant this privilege to users who need to share data
and whose processes need to cooperate.

4-9

RESOURCE CONTROL
4.3.12

GRPNAM Privilege

The GRPNAM privilege allows the user's process
the logical name table of the group to which
to delete names from that table by the use of
name system services: Create Logical Name and

to insert names into
the process belongs and
the following logical
Delete Logical Name.

In addition, the privileged user can use the ASSIGN and DEFINE
commands to add names to the group logical name table, the DEASSIGN
command to delete names from the table, and the /GROUP qualifier of
the MOUNT command to share volumes among group members.
This privilege should not be granted to all users of the system
because it allows the user to create an unlimited number of group
logical names. When many users have the unrestricted ability to
create group logical names, excessive use of system dynamic memory can
degrade system performance.

4.3.13

LOG_IO Privilege

The LOG IO privilege allows the user's process to execute the Queue
I/O Request system service to perform logical-level I/O operations.
Usually, user I/O requests are handled indirectly by use of an I/O
package such as VAX-11 Record Management Services.
However, to
increase their control over I/O operations and to improve the
efficiency of I/O operations, skilled users sometimes prefer to handle
directly the interface between their process and a system I/O driver
program.
They can do this by executing the Queue I/O Request system
service;
in many instances, the operation called
for
is
a
logical-level I/O operation.
The system manager should grant this privilege only to users who need
it. If this privilege is given to many users who have no need for it,
the operating system and service to other users can easily be
disrupted.
Such
disruptions
can
include the destruction of
information on the system device, the destruction of user data, and
the exposure of confidential information to unauthorized persons.

4.3.14

MOUNT Privilege

The MOUNT privilege allows the user's process to execute the mount
volume QIO function. The use of this function should be restricted to
system software supplied by DIGITAL.

4.3.15

NETMBX Privilege

The NETMBX privilege allows the user to perform functions related to a
DECnet computer network.

4.3.16

OPER Privilege

The OPER privilege allows the user to use the Operator Communication
Manager
(OPCOM) process, as follows: to reply to users' requests, to
broadcast messages to all terminals logged in, to designate terminals
as operators' terminals and specify the types of messages to be

4-10

RESOURCE CONTROL
displayed on these operators' terminals, and to initialize and control
the log file of operators' messages. In addition, this privilege lets
the user set devices spooled, create and control both batch queues and
print queues, and initialize and mount public volumes.
The system manager should grant this privilege only to special
users -- the operators of the system. These are the users who respond
to the requests of ordinary users, who tend to the needs of the
system's peripheral devices
(mounting reels of tape and changing
printer forms), and who attend to all the other day-to-day chores of
system operation.
(A non-privileged user can log in on the console
terminal to respond to operator requests, for example, to mount a
tape.)

4.3.17

PFNMAP Privilege

The PFNMAP privilege allows the user's process to map to specific
pages of physical memory or I/O device registers, no matter who is
using the pages or registers.
The system manager should exercise caution in granting this privilege.
If many users have unrestricted access to physical memory, the
operating system and service to other users can easily be disrupted.
Such disruptions can include failure of the system, destruction of the
data base, and exposure of confidential information to unauthorized
persons.

4.3.18

PHY_IO Privilege

The PHY IO privilege allows the user's process to execute the Queue
I/O Request system service to perform physical-level I/O operations.
Usually, users' I/O requests are handled indirectly by use of an I/O
package such as VAX-11 Record Management Services.
However, to
increase their control over I/O operations and to improve the
efficiency of their applications, skilled users sometimes prefer to
handle directly the interface between their process and a system I/O
driver program.
They can do this by executing the Queue I/O Request
system service; in many instances, the operation called for is a
physical-level I/O operation.
The system manager should grant the PHY IO privilege only to users who
need it;
in fact,
this privilege -should be granted even more
carefully than the LOG IO privilege (see Section 4.3.13).
If this
privilege is given to many users who have no need for it, the
operating system and service to other users can easily be disrupted.
Such disruptions can include the destruction of information on the
system device, the destruction of user data, and the exposure of
confidential information to unauthorized persons.

4.3.19 . PRMCEB Privilege
The PRMCEB privilege allows the user's prbcess to create or delete a
permanent common event flag cluster by executing the Associate Common
Event Flag Cluster or Delete Common Event Flag Cluster system service.

4-11

RESOURCE CONTROL

Common event flag clusters enable cooperating processes to communicate
with each other and thus provide the means of synchronizing their
execution.
This privilege should not be granted to all users of the system
because it allows the user to create an unlimited number of permanent
common event flag clusters. A permanent cluster remains in the system
even after the creating process has been terminated and continues to
use up a portion of system dynamic memory. When many users have the
unrestricted ability to create permanent common event flag clusters,
the excessive use of system dynamic memory can degrade system
performance.

4.3.20

PRMGBL Privilege

The PRMGBL privilege allows the user's process to create global
sections by executing the Create and Map Section system service. In
addition, the user with this privilege (plus the CMKRNL and SYSGBL
privileges) can use the INSTALL utility.
Global sections are shared
structures
which
can
be
mapped
simultaneously in the virtual address space of many processes. All
processes see the same code or data. Global sections are used for
reentrant subroutines or data buffers.
The system manager should grant this privilege with care.
If
permanent global sections are not explicitly deleted, they tie up
space in the global section and global page tables, which are limited
resources.

4.3.21

PRMMBX Privilege

The PRMMBX privilege allows the user's process to create or delete a
permanent mailbox by executing the Create Mailbox and Assign Channel
system service or the Delete Mailbox system service.
Mailboxes are buffers in virtual memory that are treated as if they
were record-oriented I/O devices.
A mailbox is used for general
interprocess communication.
The PRMMBX privilege should not be granted to all users of the system.
Permanent mailboxes are not automatically deleted when the creating
processes are deleted and, thus, continue to use up a portion of
system dynamic memory.

4.3.22

PSWAPM Privilege

The PSWAPM privilege allows the user's process to control whether it
can be swapped out of the balance set by executing the Set Process
Swap Mode system service. Not only must a process have this privilege
to lock itself in the balance set (that is, to disable swapping), but
also to unlock itself (that is, to enable swapping).
With the same privilege, a process can create a process that is locked
in the balance set (process swap mode disabled) by using an optional
argument to the Create Process system service or, when the RUN command
is used to create a process, by using a qualifier of the RUN command.

4-12

RESOURCE CONTROL

The system manager should grant this privilege only to users who need
to lock a process in memory for performance reasons. Typically, this
will be a real-time process. If many users have the unrestricted
ability to lock processes in the balance set, physical memory can be
held unnecessarily and thereby degrade system performance.

4.3.23

SETPRV Privilege

The SETPRV privilege allows the user's process to create processes
whose privileges are greater than its own, by executing the Create
Process system service with an optional argument or by issuing a RUN
command to create a process.
A user with this privilege can also
execute the SET PROCESS /PRIVILEGES command to grant itself any
desired privilege.
The system manager should exercise the same caution in granting
SETPRV privilege as in granting the CMEXEL or CMKRNL privilege.

4.3.24

the

SHMEM Privilege

The SHMEM privilege allows the user's process to create global
sections and mailboxes {permanent and temporary) in multiport memory,
if the process also has appropriate PRMGBL, PRMMBX, SYSGBL, and TPMBX
privileges.
Just as in local memory, the space required for a
multiport memory temporary mailbox counts against the buffered I/O
byte count limit {BYTLM) of the process.

4.3.25

SYSGBL Privilege

The SYSGBL privilege allows the user's process to create system global
sections by executing the Create and Map Section system service. In
addition, the user with this privilege {plus the CMKRNL and PRMGBL
privileges) can use the INSTALL utility.
The system manager should exercise caution in granting this privilege.
System global sections require space in the global section and global
page tables, which are limited resources.

4.3.26

SYSNAM Privilege

The SYSNAM privilege allows the user's process to insert names into
the system logical name table and to delete names from that table by
using the following logical name system services: Create Logical Name
and Delete Logical Name.
In addition, the user with this privilege can use the ASSIGN and
DEFINE commands to add names to the system logical name table, and can
use the DEASSIGN command to delete names from the table.
The system manager should grant this privilege only to the system
operator or to system programmers who need to define system logical
names {such as names for user devices, library directories, and the
system directory).
For example, to mount a system volume, which
entails defining a system logical name, the system operator must have
the SYSNAM privilege.

4-13

RESOURCE CONTROL
4.3.27

SYSPRV Privilege

The SYSPRV privilege allows the user to assume the file access rights
of a system user, and to change the owner UIC and protection of a
file. Even if a file is protected against system access, the user
with the SYSPRV privilege can simply change its protection to gain
access.
The system manager should exercise caution in granting this privilege.
If many users have system access rights, the operating system and
service to others can easily be disrupted.
Such disruptions can
include failure of the system, destruction of the data base, and
exposure of confidential information to unauthorized persons.

4.3.28

TMPMBX Privilege

The TMPMBX privilege allows the user's process to create a temporary
mailbox by executing the Create Mailbox and Assign Channel sy~tem
service.
Mailboxes are buffers in virtual memory that are treated as if they
were record-oriented I/O devices.
A mailbox is used for general
interprocess communication. Unlike a permanent mailbox, which must be
explicitly deleted, a temporary mailbox is deleted automatically when
no longer allocated by any process.
The system manager should usually grant this privilege to all users of
the
system
to
facilitate
interprocess communication.
System
performance is not likely to be degraded by permitting the creation of
temporary mailboxes, because their number is controlled by limits on
the use of system dynamic memory.

4.3.29

VOLPRO Privilege

The VOLPRO privilege allows the user to (1)
initialize a previously
used volume with an owner UIC different from the user's own UIC;
(2)
override the expiration date on a non-owned tape or disk volume;
(3)
mount a non-owned Files-11 volume with the /FOREIGN qualifier; and
(4) override the owner UIC protection of a volume.
The VOLPRO
privilege only permits control over volumes the user can mount or
initialize. Volumes mounted with the /SYSTEM qualifier are safe from
the user with the VOLPRO privilege as long as the user does not also
have the SYSNAM privilege.
The system manager should exercise extreme caution in granting the
VOLPRO privilege.
If many users can override volume protection, the
operating system and service to others can be disrupted.
Such
disruptions can include failure of the system, destruction of the data
base, and exposure of confidential information to
unauthorized
persons.

4.3.30

WORLD Privilege

The WORLD privilege allows the user's process to affect other
processes both inside and outside its group by executing the following
process control system services:
Suspend Process, Resume Process,
Delete Process, Set Priority, Wake, Schedule Wakeup, Cancel Wakeup,
and Force Exit.
The user's process is also allowed to examine

4-14

RESOURCE CONTROL

processes outside its own group by executing the Get Job/Process
Information system service. The user with the WORLD privilege can
issue SET QUEUE, DELETE /ENTRY, STOP /ENTRY, and SET PROCESS commands
for all other processes.
To exercise control over subprocesses that it created or to examine
these subprocesses, a process needs no special privilege. To affect
or to examine other processes inside its own group, a process needs
only the GROUP privilege. But to affect or examine processes outside
its own group, a process needs the WORLD privilege.

4.4

ACCOUNTING FOR THE USE OF SYSTEM RESOURCES

For accounting purposes, the VAX/VMS system keeps records of the use
of system resources.
These records are kept in the accounting log
file SYS$SYSDISK: [SYSMGR]ACCOUNTNG.DAT, which is updated each time an
accountable· process terminates, each time a print job is completed,
and each time a login failure occurs. In addition, users can send
messages to be inserted into the accounting log file.
Accounting records contain cumulative accounts of the resources used
either by processes or subprocesses set up for users or by print
symbionts that print out files for users.
Each accounting record
contains two fields -- user name and account name -- that identify the
user and establish the connection between the accounting record and a
user of the system. These fields correspond to similar fields of the
user's account record in the user authorization file (UAF).
Using the detailed accounting records provided by the system, the
system manager or a system programmer can devise programs for
reporting on the use of system resources and for billing for their
use.
Because the users of system resources are identified in two ways,
reports on the use of system resources and bills for the use of system
resources can be prepared in either of two ways: by user name or by
account name.
The accounting log
operating system
chronologically in
characteristics of

file is created and opened automatically when the
is initialized.
Accounting records are arranged
this file.
The following list summarizes the
the accounting log file:

•

File name: ACCOUNTNG.DAT (this file is not an ASCII
hence, it must be formatted before it is printed)

•

Residence:

the system device

Directory:

[SYSMGR]

•

File organization:

•

Record length:

variable length

•

Record types:

six

sequential

4-15

file;

RESOURCE CONTROL

The six types of records correspond to the six conditions that cause
records to be written to the file. These record types and their
corresponding
codes
as
defined
in
the
macro
$ACCDEF
in
SYS$LIBRARY:LIB.MLB (in parentheses) are as follows:
1.

Records written
(ACC$K _ INTTRM)

Records written when batch processes terminate (ACC$K_BATTRM)

Records written when subprocesses
terminate (ACC$K_PRCTRM)

Records written when print jobs are completed (ACC$K_PRTJOB)

Records written when login failures occur (ACC$K_LOGTRM)

Records written when users' messages
accounting log file (ACC$K_INSMSG)

when

interactive

processes

detached

are

sent

terminate

processes

the

For a detailed description of the records of the accounting log file,
see the discussion of the Send Message to Accounting Manager ($SNDACC)
system service in the VAX/VMS S_yst~~-§~_~y_!~--~~_B_~-~-~-!-~_r_:l_C::~~~!'lual.
To suppress the accounting function and thus avoid accounting for the
use of system resources requires privilege. Only a user who has the
ACNT privilege can create subprocesses or detached processes in which
accounting is disabled.
The /NOACCOUNTING qualifier of the RUN
command is used to disable all accounting in a created process.
A user with OPER privilege can selectively disable various kinds of
accounting system-wide by using the /DISABLE qualifier of 17he SET
ACCOUNTING command. Usually, this is a task of the system operator.
See the VAX/VMS Operator's Guide for a full description of the SET
ACCOUNTING command.
-------------------As records are entered in the accounting log file, all but print job
completion records are immediately flushed to disk. This precaution
guarantees the integrity of the file and the completeness
of
accounting data even if the system fails.
Normally, the accounting log file is closed at the end of a billing
period.
The current version of the accounting log file is closed and
a new version of the file is created and opened. As a rule, this is
the system operator's job, done by use of the SET ACCOUNTING command.
If an attempt to write to the accounting log file results in an error,
the file is closed automatically and a new copy is created and opened.

4-16

PART II
OPERATIONAL CONTROL

The system manager is responsible
operational control over the system:

for

the

following

aspects

•

Initializing and mounting public volumes

•

Enabling VAX-11 RMS file sharing

•

Backing up public volumes

•

Establishing usage
volumes

•

Installing frequently used, shareable, and
as known images

•

Modifying the site-independent start-up command procedure

•

Creating a site-specific start-up command procedure

•

Establishing spooled devices

•

Controlling batch and print queues

•

Logging errors and reporting software problems

quotas

for

individual

users

public

executable

images

Part II consists of Chapters 5 through 10:
•

Chapter 5: Maintaining Public Files and Volumes -- Discusses
procedures for handling public volumes and provides operating
instructions for the RMSSHARE utility.

•

Chapter n:
usage and
utility.

•

Chapter 7: Installing Known Images -- Discusses installation
of selected executable and shareable images as known images,
and the creation of permanent, system global sections for
certain known images. Provides operating instructions for the
INSTALL utility.

•

Chapter 8: Start-up Command Procedures -- Defines the content
of the site-independent start-up command procedure supplied by
DIGITAL, and suggests the content for the site-dependent
start-up command procedure.

Disk Quotas -- Defines capabilities to limit disk
provides operating instructions for the DISKQUOTA

9: Batch and Print Queues -- Discusses the use cf
• Chapter
spooled
devices,
and
common
procedures
for creating
(initializing), deleting, starting, and
batch queues.
•

stopping

device

and

Chapter 10: Errors and Other System Events -- Discusses error
logging procedures and provides operating instructions for the
SYE utility. Outlines the software performance report.

CHAPTER 5
MAINTAINING PUBLIC FILES AND VOLUMES

Public volumes, also called system volumes, are file-structured disk
volumes that contain public files.
Public files are files that must
be available to most, if not all, users.
Public volumes can also
contain files that users create for their own private use or for
general use. Thus, as long as UIC-based file protection permits it,
all users have access to public volumes and to the files on them.
Public volumes can contain
supplied by DIGITAL:

the

following

kinds

files

•

The operating system itself in
related to the operating system

•

Utility programs in executable form

•

Diagnostic and test programs
related to these programs

•

Various system libraries such as macro libraries, object
module libraries, shared run-time libraries, and error message
libraries

•

Text files such as help files

•

Optional software in executable form, plus
and other files

executable

public

executable

form

and

files

form

and

files

libraries

In addition, the system manager can include on public volumes files
that are unique to an installation. These typically are files that
must be accessible to many if not all users of the installation.
Finally, the system manager can permit any user to create and store
files on a public volume. Depending on their file protection, these
files can be of general or restricted accessibility. This use of a
public volume, however, is subject to limitation: a user is free to
create, catalog, and store files on a public volume only if volume
protection permits, if the user has write access to a directory on the
volume, and if disk quotas permit. As a rule, the system manager
creates a default disk file directory on a public volume for each user
authorized to use the system (see Chapter 2).
Knowing how to initialize and mount public volumes is
to managing a system of public files and volumes.

5-1

prerequisite

MAINTAINING PUBLIC FILES AND VOLUMES
5.1

FILES-11 DISK STRUCTURE

The VAX/VMS system recognizes two disk file structures:
Files-11
Structure Level 1 and Files-11 Structure Level 2. Files-11 Structure
Level 2 is the default disk structure of the VAX/VMS system, and
Files-11 Structure Level 1 is a structure used by DIGITAL's RSX-llM,
RSX-llD, RSX-llM-PLUS, and !AS operating systems.
Nine files control the structure of a Files-11 Structure Level 2
volume.
Only five of these files are used for a Files-11 Structure
Level 1 volume. These files, which are referred to as reserved files,
are as follows:
1.

Index file (levels 1 and 2)

Storage bit map file (levels 1 and 2)

Bad block file (levels 1 and 2)

Master file directory (levels 1 and 2)

Core image file (levels 1 and 2)

Volume set list file (level 2 only)

Continuation file (level 2 only)

Back-up log file (level 2 only)

Pending bad block log file (level 2 only)

All of the files listed
directory, [O,O].

5.1.1

above

are

cataloged

the

master

file

Index File

Every Files-11 volume has an index file, which is created when the
volume is initialized. This index file identifies the volume to the
operating system as a Files-11 structure and contains the access data
for all files on the volume. The index file, which is listed in the
master file directory as INDEXF.SYS;l, contains
the
following
information:
•

Bootstrap block -- The volume's bootstrap block is virtual
block number 1 of the index file.
If the volume is a system
volume, this block contains a bootstrap program that loads the
operating system into memory. If the volume is not a system
volume, this block contains a program that displays a message
that the volume is not the system device but a device that
contains users' files only.

•

Home block -- The home block establishes the specific identity
of the volume, providing such information as the volume name
and protection, the maximum number of files allowed on the
volume, and the volume ownership information. The home block
is virtual block number 2 of the index file.

•

Back-up home block -- The back-up home block is a copy of the
home block.
It permits the volume to be used even if the
primary home block is destroyed.

5-2

MAINTAINING PUBLIC FILES AND VOLUMES

•

Back-up index file header -- The back-up index file header
permits recovery of data on the volume if the index file
header goes bad.

•

Index file bit map -- The index file bit map controls the
allocation of file headers and thus the number of files on the
volume. The bit map contains a bit for each file header that
is allowed on the volume. If the value of a bit for a given
file header is O, a file can be created with this file header.
If the value is 1, the file header is already in use.

•

File headers
The largest part of the index file is made up
of file headers.
Each file on the volume has a file header,
which describes such properties of the file as file ownership,
creation date and time, file protection, and location of the
file's blocks. The file header contains all the information
needed for gaining access to the file.

5.1.2

Storage Bit Map File

The storage bit map file controls the available space on a volume;
this file is listed in the master file directory as BITMAP.SYS;l. It
contains a storage control block, which
consists
of
summary
information intended to optimize the Files-11 space allocation, and
the bit map itself, which lists the availability of individual blocks.

5.1.3

Bad Block File

The bad block file, which is listed in the master file directory as
BADBLK.SYS;l, contains all the bad blocks on the volume. The system
detects bad disk blocks dynamically and prevents their reuse once the
files to which they are allocated have been deleted.

5.1.4

Master File Directory

The master file directory
(MFD)
itself is listed in the MFD as
000000.DIR;l.
The MFD, which is the root of the volume's directory
structure, lists the reserved files that control the volume structure
and may list both users' files and users' file directories. Usually,
however, the MFD is used to list the reserved files and users'
file
directories;
users seldom enter files in the MFD, even on private
volumes. In fact, on a private volume, it is most convenient for a
user to create a directory that has the same name as the user's
default directory on a system disk. For an explanation of users' file
directories and file specifications, see the VAX/VMS Command Lanquage
User's Guide.

5.1.5

Core Image File

The core image file is listed in the MFD as CORIMG.SYS;l.
used by VAX/VMS.

5-3

not

MAINTAINING PUBLIC FILES AND VOLUMES
5.1.6

Volume Set List File

The volume set list file is listed in the MFD as VOLSET.SYS;l.
This
file is used only on relative volume 1 of a volume set. The file
contains a list of the labels of all the volumes in the set.

5.1.7

Continuation File

The continuation file is listed in the MFD as CONTIN.SYS;l. This file
is used as the extension file identifier when a file crosses from one
volume to another volume of a loosely coupled volume set.
This file
is reserved for future use.

5.1.8

Back-Up Log File

The back-up log file is listed in the MFD as BACKUP.SYS;l. This file
contains a history of the back-ups done to the volume. This file is
reserved for future use.

5.1.9

Pending Bad Block Log File

The pending bad block log file is listed in the MFD as BADLOG.SYS;l.
This file contains a list of suspected bad blocks on the volume that
are not listed in the bad block file.

5.1.10

Files-11 Structure Level l Versus Structure Level 2

Files-11 Structure Level 2, a compatible superset of Structure Level
1, is the preferred disk structure on VAX/VMS for reasons of
performance and reliability. At volume initialization time
(see the
INITIALIZE command in the VAX/VMS Co~ma~p Language User's Guide),
Structure Level 2 is the default.
Structure Level l should be
specified only for volumes that must be transportable to RSX-llM,
RSX-110, RSX-llM-PLUS, and IAS systems, as these systems support only
that structure level.
Additionally, the system manager may be
required to handle Structure Level 1 volumes transported to VAX/VMS
from one of the above systems.
Where Structure Level l volumes are in use on the system, the
manager should bear in mind the following limitations on them:

system

No
hierarchies
of
directories
and
• Directories
subdirectories, and no ordering of directory entries (that is,
the file names) in any way (RSX-llM, RSX-llD, RSX-llM-PLUS,
and !AS systems do not support subdirectories and alphabetical
directory entries)
•

Wild cards -- Wild card characters only for complete fields of
file specifications (for example, SYS$SYSTEM:*USER.PAR is
illegal, while SYS$SYSTEM:*.PAR is legal)

•

Disk quotas -- Not supported

•

Multi-volume files -- Not supported

5-4

MAINTAINING PUBLIC FILES AND VOLUMES
•

Placement control -- Not supported

•

Caches -- No
caching
of
file
header
identification slots, or extent entries

•

System disk -- Cannot be a Structure Level 1 volume

•

Clustered allocation -- Not supported

•

Back-up home block -- Not supported

•

Protection code E -- Meaning of extend, rather than execute

•

File versions -- Limited to 32,767, rather than being set by
the
/VERSION LIMIT qualifier of CREATE /DIRECTORY, which
defaults to approximately 60

blocks,

file

Future enhancements to VAX/VMS will be based primarily on Structure
Level 2, so that further restrictions on Structure Level 1 volumes may
be incurred.
As explained in the VAX-11 Utilities Reference Manual, the system
manager must use DSCl to back up Structure Level 1 volumes and DSC2 to
back up Structure Level 2 volumes, and VFYl to verify Structure Level
volumes and VFY2 to verify Structure Level 2 volumes.

5.2

INITIALIZING PUBLIC VOLUMES

The purpose of initializing a disk volume is to delete all old
information from the volume and to impart to the volume a Files-11
structure that the operating system recognizes.
This structure
prepares a volume to receive data and permits the operating system to
locate data stored on the volume.
The VAX/VMS Command Language User's Guide contains information about
initializing volumes;
see {l) the description of the INITIALIZE
command in Part II and {2) Chapter 3, "Disk and Tape Volumes."
The following guidelines for initializing public volumes supplement
information presented in the VAX/VMS Command Language User's Guide.
In initializing a public volume {by using the qualifier /SYSTEM), the
system manager may need to use one or all of the following qualifiers
of the DCL command INITIALIZE:
•

/ACCESSED=n

•

/INDEX=position

/CLUSTER_SIZE=n

•

/MAXIMUM_FILES=n

•

/EXTENSION=n

/WINDOW=n

•

/HEADERS=n

As described below, selecting appropriate values for n and selecting
the appropriate position for the /INDEX qualifier often involve making
trade-offs.

5-5

MAINTAINING PUBLIC FILES AND VOLUMES
5.2.l

/ACCESSED=n Qualifier

The /ACCESSED=n qualifier provides an estimate of the number of
directories expected to be in use concurrently on a volume. The file
system keeps this number of directory file control blocks in system
space for ready access on the basis of which directories were most
recently used. The result is a substantial reduction of overhead in
directory
operations.
For volumes mounted /SYSTEM, the SYSGEN
parameter ACP SYSACC overrides this value.

5.2.2

/CLUSTER_SIZE=n Qualifier

The /CLUSTER SIZE=n qualifier specifies the fundamental unit of
allocation
(expressed in blocks) on a volume.
In selecting the
cluster size, wasted space at the end of files is traded off against
the size of the volume storage bit map, which must contain one bit for
each cluster on the volume (or one block for each 409~ clusters).

5.2.3

/EXTENSION=n Qualifier

The /EXTENSION=n qualifier specifies the default number of blocks
allocated for extending files on a volume.
This value is less
important on the VAX/VMS system than on the RSX-llM, RSX-llD,
RSX-llM-PLUS, and IAS systems, because VAX-11 Record Management
Services use an adaptive algorithm maximized against /EXTENSION.
The
value of this qualifier should be an even multiple of /CLUSTER_SIZE.

5.2.4

/HEADERS=n Qualifier

The /HEADERS=n qualifier specifies the number of file headers to be
allocated initially to the index file.
The primary advantage of
preallocating file headers is that they will then be located near the
storage map file (usually in the middle of the disk). This placement
of file headers helps reduce head motion during file manipulation.
This value should be estimated conservatively, because space allocated
to headers cannot later be made available for file storage.

5.2.5

/INDEX=position Qualifier

The /INDEX=position qualifier specifies the location of the index file
on a volume.
The default position (MIDDLE) results in minimum head
motion during file processing. The position BEGINNING should be used
if the disk is to contain only one or a few very large contiguous
files.
(The Disk Save and Compress (DSC) Utility positions the index
at BEGINNING.)

5.2.6

/MAXIMUM_FILES=n Qualifier

The /MAXIMUM FILES=n qualifier specifies the maximum number of files
that a volume can contain. The default value is fairly liberal. A
closer estimate of it helps optimize the dynamic allocation of the
index file;
once set, however, the maximum number of files for a
volume cannot be increased.
Note that each directory and each
extension header of a multiheader file counts as a file against this
maximum value.
5-6

MAINTAINING PUBLIC FILES AND VOLUMES

5.2.7

/WINDOW=n Qualifier

The /WINDOW=n qualifier specifies the number of map pointers in a
default file access window. This value is the number of extents of a
file to which access can be gained without the cost of file system
overhead.

5.3

MOUNTING PUBLIC VOLUMES

The purpose of mounting a volume or volume set is to establish a
relationship between the volume or volume set, the device(s) on which
the volume is physically mounted, and one or more processes that can
gain access to the volume.
The VAX/VMS Command Language User's Guide contains information about
mounting volumes;
see
(1)
the description of the MOUNT command in
Part II and (2) Chapter 3, "Disk and Tape Volumes."
The following guidelines for mounting disk
supplement information presented in the
User's Guide.

volumes
VAX/VMS

for public use
Command Language

In mounting a public volume (by using the qualifier /SYSTEM), the
.system manager may need to use one or all of the qualifiers /ACCESSED,
/EXTENSION, and /WINDOW (described in Section 5.2), or the qualifiers
/BIND=volume-set-name and /PROCESSOR=option (described below).

5.3.l

/BIND=volume-set-name Qualifier

The /BIND=volume-set-name qualifier combines two or more volumes into
a volume set, or adds one or more volumes to an existing volume set.
A volume set makes two or more disk volumes appear to the user as one
volume on one device.

5.3.2

/PROCESSOR=option Qualifier

The /PROCESSOR=option qualifier specifies the number of ancillary
control processes
(ACPs) to be used in controlling various public
volumes. Selecting an appropriate option for the /PROCESSOR qualifier
involves making a trade-off.
If the system manager specifies the
option SAME, file system parallelism and performance may be sacrificed
for the sake of saving system space.
Conversely, if the system
manager specifies the option UNIQUE, system space is sacrificed for
the sake of file system parallelism and performance.

5-7

MAINTAINING PUBLIC FILES AND VOLUMES

RM SS HARE
5.4

MAINTAINING VOLUME INTEGRITY

To enhance performance, the system caches in memory information
concerning a volume's free space, file identifications, quota file
entries, and file headers. The system manager determines the degree
of caching with the ACP cache system parameters (see Chapter 12) and
individual users can alter cache sizes on their volumes
with
qualifiers to the MOUNT command
(see the VAX/VMS Command Language
User's Guide). The system writes the information in the caches to the
disk when - the disk is dismounted or the system is shut down.
Naturally, removal of a disk before the caches are written back loses
any changes made to the information in the caches. Therefore, the
system manager and individual user should:
•

Not write-lock a volume while it is mounted

•

Not remove a volume from a drive until it has been dismounted

•

Not halt the system without performing the orderly
procedure (see the VAX/VMS Operator's Guide)

shut-down

If the user write-locks a volume at mount time,
the
system
additionally applies a software write lock: the user must dismount
and remount the volume to write-enable it.
At mount time, if the system detects that the
back the last time the volume was used,
rebuilds the file information by scanning the
However, file headers for open files may be
(see Section 11.5.1).

5.5

caches were not written
the system automatically
contents of the volume.
partially or wholly lost

RMSSHARE UTILITY

The RMSSHARE utility performs the following functions:
•

Enables the VAX-11 Record Management Services (VAX-11 RMS)
file
sharing
capability
by
initializing file sharing
structures in paged dynamic memory (system SO space), and sets
the maximum number of pages that the structures can occupy.
The system imposes an absolute maximum of one-half the total
space in paged dynamic memory. The VAX-11 RMS file sharing
capability must be enabled each time the operating system is
booted.

•

If VAX-11 RMS file sharing has already been enabled, displays
figures on allowable and actual usage, and permits the
resetting of the maximum number of pages that the file sharing
structures can occupy.

The system manager, or any user with the CMEXEC privilege, invokes the
utility with the following command:
$ RUN SYS$SYSTEM:RMSSHARE

5-8

MAINTAINING PUBLIC FILES AND VOLUMES

RMSSHARE (Cont.)
If VAX-11 RMS file sharing
following message:

not

enabled,

RMSSHARE

displays

the

RMS file sharing is not currently enabled •••
Maximum allocation allowed: n
If VAX-11 RMS file sharing is already in effect, RMSSHARE displays the
following message:
RMS file sharing is currently enabled •••
Maximum allocation allowed: n
Number of pages allocated: n
Max pages used: n
Current number of pages in use: n
In either case, RMSSHARE then prompts for a maximum page count:
Enter max pages:
The system manager can either:
•

Enter the maximum number of pages that file sharing structures
can occupy as a positive integer not less than the current
number of pages in use and not greater than the maximum number
of pages allowed

•

Enter the word EXIT (or
terminate the utility

RMSSHARE continues
terminated.

display

simply
and

press

prompt

the
as

return

above

until

If an invalid value is entered for the maximum number of pages
file structures can occupy, the following message appears:

that

Invalid size parameter, set to maximum value:

shown

key}

RMSSHARE allocates the maximum allowable number of pages
the invalid value.

lieu

The system manager should use the following guidelines to estimate the
maximum number of pages required for file sharing structures:
•

System base requirement -- 2 pages

•

Per sequential file being shared -- l page for the first three
sharers and l page for each additional four sharers

•

Per relative file being shared -- l page for the first
sharers and 1 page for each additional four sharers

•

Per indexed file being shared -- 1 page for the
sharers and l page for each additional two sharers

The system manager totals the above pages.

5-9

first

three
two

MAINTAINING PUBLIC FILES AND VOLUMES

RMSSHARE (Cont.)
If more than one locked record per sharer is anticipated, the system
manager should add 16 bytes for each additional locked record. If the
multibuffer count exceeds 1 for a sequential or relative file, or 2
for an indexed file, the system manager should add 36 bytes for each
additional buffer. The grand total is then rounded up to the next
page.
The site-independent start-up command procedure (see Chapter 8) runs
RMSSHARE with a specification of 20 for the maximum number of pages
for file sharing structures. This number should be adjusted according
to the above guidelines. The system manager should also adjust the
size of the paged dynamic pool (PAGEDYN system parameter; see Chapter
12) •

5.6

BACKING UP PUBLIC VOLUMES

To prevent the inadvertent loss or destruction of valuable information
stored on public volumes, the system manager usually establishes a
policy and a schedule for regularly backing up files on public
volumes.
Once such a policy is established, the system operator
usually is responsible for putting it into effect.
The VAX/VMS
Operator's Guide therefore provides the operating procedures for
backing up both selected files and entire volumes.
Just as preserving information on public volumes by backing it up is
usually considered desirable, preserving files on private volumes is
also considered desirable. However, responsibility for backing up the
files on private volumes usually is left to the individual owners of
those files and volumes.
There are two kinds of back-ups of public disk files and volumes:
(1)
selective, or partial, back-ups and (2) system, or all-inclusive,
back-ups. The back-up medium, in either case, can be disk or tape.
Selective back-ups of files chosen by users of the system can be
accomplished with the VAX-11 RMS utility BACKUP. System back-ups, on
the other hand, are usually accomplished wi~h one of the Disk Save and
Compress (DSC) utility programs (DSCl or DSC2), or with the COPY
command.
As explained in the VAX/VMS Operator's Guid~ and in the description of
the Disk Save and Compress ut1l1ty 1n the VAX-11 Utilities Reference
Manual, the difference between the utilities is in the level of the
Files-11 disk file structure that they write to a new disk. DSCl
(which writes disks with Files-11 Structure Level l) is used in
backing up disks that have been initialized with Files-11 Structure
Level l; and DSC2 (which writes disks with Files-11 Structure Level
2}
is used in backing up disks that have been initialized with
Files-II Structure Level 2.
As a rule, selective back-ups are done more frequently than system
back-ups.
Normally, the system manager, after consulting with users
of the system, decides how frequently to back up files and volumes and
how long to retain back-up files and volumes.

5-10

MAINTAINING PUBLIC FILES AND VOLUMES
The following schedule for backing up public disk volumes on magnetic
tape affords adequate protection of data for many installations:
•

Daily -- A selective back-up retained for seven days.
This
schedule requires seven daily tapes that are rotated once a
week.

•

Weekly -- An all-inclusive back-up retained for four weeks.
This schedule requires four weekly tapes that are rotated once
every four weeks.

•

Monthly -- An all-inclusive back-up retained for a year. This
schedule requires twelve monthly tapes that are rotated once a
year.

Despite all precautions, there is always the risk of
Longer retention periods reduce this risk.

5-11

losing

file.

CHAPTER 6
DISK QUOTAS

The system manager limits the amount of space available to individual
users on public volumes (or volume sets) by creating and maintaining
quota files on those volumes. Individual users can similarly restrict
usage on private volumes. Quotas are maintained and enforced on a per
volume basis. Each volume or volume set has its own quota file;
a
volume on which quotas are not maintained has no quota file;
on a
volume set, volume 1 contains the quota file. Each entry in a quota
file includes the following information:
•

UIC -- UIC of a user entitled to maintain files on the volume

•

Usage - Number of blocks on the volume taken up by the
files

•

Quota -- Maximum number of blocks on the volume that the
user's files may take up before an error message is issued

•

Overdraft
Number of blocks over the quota that
files may take up

the

user's

The absolute maximum number of blocks permitted a user on a volume
the sum of the quota and the overdraft.

The system manager (or user maintaining the volume)
identifies UICs
and assigns quotas and overdrafts with the DISKQUOTA utility (Section
6.1). Usage counts are maintained automatically by the system during
normal file activities (Section 6.2).
The name of the quota file is [O,O]QUOTA.SYS on the applicable volume.
A quota file is initialized with an entry for UIC [O,O].
The usage
count for this UIC should not change from 0 -- the UIC should own no
files.
Its quota and overdraft, however, serve as defaults in certain
situations, and so should be set to values most likely to be assigned
as quotas and overdrafts to other UICs.
A quota file requires one block
entries.

6.1

secondary

storage

for

each

DISKQUOTA UTILITY

The system manager (or any user maintaining a volume) can run the
DISKQUOTA utility to control the usage of disk volumes. Table 6-1
summarizes the DISKQUOTA commands by format and function.

6-1

DISK QUOTAS

DISKQUOTA
Table 6-1
DISKQUOTA Command Summary
Format
ADD uic

Function

[/PERMQUOTA=quota]
[/OVERDRAFT=quota-plus]

Adds an entry to the quota
file

CREATE

Creates a quota file for a
volume
that
does
not
currently contain one

DISABLE

Suspends quota
on a volume

operations

ENABLE

Resumes quota
on a volume

operations

EXIT

Returns the user
command level

HELP

Lists
the
commands

MODIFY uic

[/PERMQUOTA=quota]
[/OVERDRAFT=quota-plus]

DCL

DISKQUOTA

Changes an entry in the
quota file

REBUILD

Reconstructs
the
usage
counts for all entries

REMOVE uic

Deletes an entry from
quota file

the

SHOW uic

Displays quotas and
counts

usage

USE device

Specifies the volume to be
acted upon

The sequence of commands for
includes:

creating

new

quota

file

typically

USE -- To name the volume (can be omitted if
the user's default device)

the

volume

CREATE

To create the quota file

MODIFY

To set the quota and overdraft for UIC [O,O]

REBUILD -- To add entries for existing files (can be
on an empty volume)

ADD -- To add entries (one ADD command per
not automatically added during Step 4

entry)

omitted
for

UICs

The sequence of commands for modifying an existing quota file
typically includes (1) a USE command followed by (2) an ADD, MODIFY,
or REMOVE command for each entry being changed.

6-2

DISK QUOTAS

DISKQUOTA (Cont.)
6.1.l

Invoking DISKQUOTA

The following command invokes the utility:
$ RUN SYS$SYSTEM:DISKQUOTA
The system responds with the following prompt:
DISKQ>
The user can then enter any of the commands
listed in Table n-1.
These commands
follow the standard rules of grammar as specified in
the VAX/VMS Command Language User's Guide.

6.1.2

ADD Command

ADD adds an entry to the quota file.

It takes the form:

ADD uic [/PERMQUOTA=quota]
[/OVERDRAFT=quota-plus]
uic
A valid UIC
quota
A positive integer that specifies a quota for the specified
defaults to the value of quota in the entry for [O,O]

UIC;

quota-plus
A positive integer that specifies an overdraft for the specified
UIC;
defaults to the value of quota-plus in the entry for [O,O]
The usage count for a new entry is initialized to O.
In the following example, the user sets the quota for UIC [300,211] to
200 and the overdraft to 50 (for an absolute limit of 250 blocks):
DISKQ>ADD [300,211) /PERMQUOTA=200 /OVERDRAFT=50
The ADD command requires write access to the quota file.

6.1.3

CREATE Command

CREATE creates a new quota file.

It takes the form:

CREATE
A quota file must not already be present on the volume being used.
The CREATE command should be immediately followed by a MODIFY command
for UIC [O,O] with /PERMQUOTA and /OVERDRAFT set to reasonable default
values.
The CREATE command requires write access to the volume's MFD;
and
either the SYSPRV privilege, a system UIC, or ownership of the volume.

6-3

DISK QUOTAS

DISKQUOTA (Cont.)
6.1.4

DISABLE Command

DISABLE suspends the maintenance
volume. It takes the form:

and

enforcement

quotas

UIC,

DISABLE
The DISABLE command requires the SYSPRV privilege, a
ownership of the volume.

6.1.5

system

ENABLE Command

ENABLE resumes quota operations on a volume.

It takes the form:

ENABLE
The ENABLE command requires the SYSPRV privilege,
ownership of the volume.

6.1.6

system

UIC,

EXIT Command

EXIT returns the user to DCL command level.

It takes the form:

EXIT
The system manager can also return to
<CTRL/Z>.

6.1.7

DCL

command

level

typing

HELP Command

HELP lists and explains the DISKQUOTA commands.

It takes the form:

HELP [command [qualifier]]
command
Name of a DISKQUOTA command
qualifier
Name of a DISKQUOTA command qualifier

6.1.8

MODIFY Command

MODIFY changes an entry in the quota file.

It takes the form:

MODIFY uic [/PERMQUOTA=quota]
[/OVERDRAFT=quota-plus]
uic
A UIC with an entry on the quota file
quota
A positive integer that specifies a quota for the specified UIC

6-4

DISK QUOTAS

DISKQUOTA (Cont.)
quota-plus
A positive integer that specifies an overdraft for the
UIC
If quota is less than the current usage count, a warning
issued.
(The new quota does take effect, however.)

specified
message

In the following example, the user sets the permanent quota
for
(300,211] to 300 blocks, while making no change to the overdraft:

UIC

DISKQ>MODIFY (300,211] /PERMQUOTA=300
The MODIFY command requires write access to the quota file.

6.1.9

REBUILD Command

REBUILD reconstructs the usage counts for all entries on
It takes the form:

the

volume.

REBUILD
The REBUILD command automatically adds entries for files owned by UICs
with no entries in the quota file, setting their quotas and overdrafts
to the values of the defaults in UIC [O,O].
During the time that the
REBUILD
command
is executing,
file
activity on the volume
is
frozen -- no files can be created, deleted,
extended,
or truncated.
For this
reason, the command should be used judiciously, normally in
the following situations:
•

Established files -- When a quota file is created on a volume
with existing files, the REBUILD command should be run before
the ADD commands.
REBUILD adds entries and
records
the
existing usage for all UICs with detected usage.

•

Usage not updated -- If usage counts were not updated for some
period due to suppression or suspension, REBUILD should be run
to correct the usage.

The REBUILD command requires write access
to the quota
file;
and
either the SYSPRV privilege, a system UIC, or ownership of the volume.

6.1.10

REMOVE Command

REMOVE deletes an entry from the quota file.

It takes the form:

REMOVE uic
uic
A UIC with an entry in the quota file
If the usage count for the UIC is not O, a warning message is
(The UIC is removed, however.) UIC [O,O] cannot be removed.
The following example deletes UIC [300,211]

from the quota file:

DISKQ>REMOVE [300,211]
The REMOVE command requires write access to the quota file.

6-5

issued.

DISK QUOTAS

DISKQUOTA (Cont.)
6.1.11

SHOW Command

SHOW displays quotas, overdrafts, and
form:

usage

counts.

specifying

the

takes

the

UIC,

SHOW uic
uic
A UIC with an entry in the quota file
A wild card character (*) can be used
illustrated in the following examples:
Command

Description

SHOW [ 300, 211]

Show user 211 in group 300

SHOW (300,*]

Show all users in group 300

SHOW [*, 211]

Show all users with a
member number of 211

SHOW [*,*]

Show all users

The SHOW command requires no privileges to show one's own quota,
overdraft, and usage count, but otherwise requires read access to the
quota file.

6.1.12

USE Command

USE specifies the volume to be acted upon.

It takes the form:

USE device
device
A physical device name or a logical name equated
device name

physical

If USE is not specified, DISKQUOTA uses the user's default device.
USE can be specified more than once during a session to work with
quota files on more than one volume.
Any volume in a volume set can be specified.
operated on, however, is relative volume 1.

The

volume

actually

The following examples specify the volume to be acted upon as
(1) a
volume on a physical device and (2) a volume on the physical device
equated to a logical name:
1.

DISKQ>USE DMA2:

DISKQ>USE X2 RESEARCH DATA

6-6

DISK QUOTAS

DISKQUOTA (Cont.)
6.1.13

Error Messages

Table 6-2 lists the error messages issued by the DISKQUOTA utility (in
alphabetical order).
Table 6-2
DISKQUOTA Error Messages
---------------------------Remedy
Message
! - - - - - - - - - - - - - - - - - - - - - - - ··-+--··
-·------··----ambiguous command
Spell out the command name
more completely.

ambiguous qualifier

Spell out the qualifier name
more completely.

cannot allocate sufficient
memory

Increase the VIRTUALPAGCNT
system parameter (with the
SYSGEN utility) so that the
REBUILD command has sufficient
space to build the usage
table.

cannot examine quota file entry

Respond to VAX-11 RMS message
that follows.

command syntax error

Format the command according
to the specifications of Table

n-1.
device is not a local device

Enter a device name (or
logical name that translates
to a device name) that does
not contain a node name.
DISKQUOTA operates only on
local devices.

error closing quota file

Respond to VAX-11 RMS message
that follows.

error creating quota file

Respond to VAX-11 RMS message
that follows.

error initializing quota file

Respond to VAX-11 RMS message
that follows.

failed help library index init

Respond to VAX-11 RMS message
that follows.

failed to access index file on
relative volume n

Respond to VAX-11 RMS message
that follows.

failed to access help text

Respond to VAX-11 RMS message
that follows.
(continued on next page)

6-7

DISK QUOTAS

DISKQUOTA (Cont.)
Table 6-2 (Cont.)
DISKQUOTA Error Messages
Message

Remedy

failed to access quota file

Respond to VAX-11 RMS message
that follows.

failed to add quota file entry

Respond to VAX-11 RMS message
that follows.

failed to disable quota file

Respond to VAX-11 RMS message
that follows.

failed to enable quota file

Respond to VAX-11 RMS message
that fol lows

failed to lock volume

Respond to VAX-11 RMS message
that follows.

failed to open help library

Respond to VAX-11 RMS message
that follows.

failed to read home block on
relative volume n

Respond to VAX-11 RMS message
that follows.

failed to remove quota file
entry

Respond to VAX-11 RMS
that follows.

failed to unlock volume

Respond to VAX-11 RMS message
that follows.

invalid UIC

Enter a correctly formatted
UIC.

I/O error reading commands

Respond to VAX-11 RMS message
that follows.

I/O error reading file header
k on relative volume n

Respond to VAX-11 RMS message
that follows. The value of k
is the file number of the file
in error.

I/O error reading index file
bitmap on relative volume n

Respond to VAX-11 RMS message
that follows.

I/O error reading quota file

Respond to VAX-11 RMS message
that follows.

logical name is recursively
defined

Enter the actual name or a
logical name that can be
translated within 10
iterations.

~----·-·-·---·-···-··-·-····--.--

mess~ge

-.--..-·-..-· ·-·-·-.................................... .

.........

(continued on next page)

6-8

DISK QUOTAS

DISKQUOTA {Cont.)
Table 6-2 (Cont.)
DISKQUOTA Error Messages
Remedy

Message
no device currently selected

Enter another USE command
specifying a valid device
name.

uic has n blocks in use

Do not respond -inf ormat i ona l message. The
specified UIC is over quota
due to the MODIFY or DELETE
operation just performed.

unrecognized command

Enter a valid command.

unrecognized qualifier

Enter a valid qualifier.

volume set has too many volumes
to handle

Reduce the number of volumes
in the volume set to 255 or
less, or do not use disk
quotas.

_____________________ __________________

__,_

___,

Where correction of an entered value is indicated, the entire command
must be reentered. For explanations of the VAX-11 RMS messages, see
the VAX/VMS System Messages and Recovery Procedures Manual.

6.2

OPERATIONS

During normal use of a volume with a quota file, the system
automatically updates the usage counts as users create, delete,
extend, and truncate files. Users without entries in the quota file
are not allowed to create files or allocate space on the volume,
unless they have the EXQUOTA privilege.

6.2.1

Exceeding the Quota

If an operation to add a new file or expand a current file will put a
user's usage count over the quota, the system prohibits the operation
and issues the following message:
disk quota exceeded
If the rejected operation is an extension of a file opened for write,
a user with an overdraft can perform the operation by retrying it.
Operations to extend the file will succeed until usage exceeds the sum
of the quota and the overdraft. At this point, the system reissues
the above message and prohibits further extensions to the file.
To create new
ove rd raft) •

files,

user's

usage

must

below

Quota restrictions are not enforced for users with
privilege. Howeve'r, their usage counts a re maintained.
6-9

quota
the

(not

EXQUOTA

DISK QUOTAS
6.2.2

Suspending Quotas

The DISABLE command in DISKQUOTA
(Section 6.1.4)
suspends quota
operations on a volume. The ENABLE command (Section 6.1.5) lifts the
suspension.
In addition,
quota operations on a volume can be
suspended at mount time by specifying the /NOQUOTA qualifier in the
DCL command MOUNT.
To discontinue quota operations on a volume,
the user executes the
DISABLE command, exits from DISKQUOTA, and deletes the QUOTA.SYS file.

6.2.3

REBUILD on Mount

When a volume is mounted with quota operations enabled and that volume
was not properly dismounted the last time it was used, the system
performs an automatic REBUILD operation.
This action ensures that the
quota file accurately reflects usage of the disk in the event that
quota operations were suspended, the system failed,
the volume was
physically removed before being dismounted,
or the WRITE PROTECT
button was pushed.

6.2.4

Restrictions on Other System Operations

The following restrictions must be observed whether or not disk quotas
are being used:
•

File ownership -- Because a change in file ownership consumes
the
resources of another user, changing the owner UIC of a
file requires the SYSPRV privilege.

•

Volume sets -- Relative volume l must be online at all times.

6-10

CHAPTER 7
INSTALLING KNOWN IMAGES

The system manager enhances the performance of selected executable and
shareable images by installing them as known images with the INSTALL
utility. Known images can be assigned the following attributes:
•

Permanently open -- Directory information on the image file
remains permanently resident, eliminating the usual directory
search required to locate a file.
The cost of keeping an
image file permanently open is 192 bytes of nonpaged dynamic
memory.

•

Header resident -- The header of the image file (native images
only) remains permanently resident, saving one disk I/O
operation per file access, at a cost of less than one page of
paged dynamic memory.
The image must also be declared
permanently open.

•

Privileged -- Amplified privileges are temporarily assigned to
any process running the image
(executable images only),
permitting the process
to
bypass
its
UAF
privilege
restrictions during execution of the image. In this way,
"normal"
users
can
run
programs
that
require
higher-than-normal privileges.

•

Protected -- A shareable image contains protected code, that
is, code that runs in kernel or executive mode but that can be
called by a user-level image.

•

Shared -- More than one user can access the read-only and
non-copy-on-reference (non-CRF) read/write sections of the
image concurrently, so that only one copy of those sections
ever need be in physical memory.
(CRF sections always require
a separate copy for each process.) The image must also be
declared permanently open.

•

Writeable -- When a shared non-CRF writeable section
is
removed from physical memory (for paging reasons or because no
processes are referencing it), it is written back to the image
file.
Any updates made by using processes, therefore, are
preserved (while the initial values are lost). The image must
also be declared shared.

7-1

INSTALLING KNOWN IMAGES
7.1

EXECUTABLE AND SHAREABLE IMAGES

Most images installed as known images are likely to be executable
images.
An executable image is one linked with the /EXECUTABLE
qualifier.
The system manager can also install shareable images as known images.
A shareable image is one linked with the /SHAREABLE qualifier. A
shareable image must subsequently be linked into an executable image
to be used.
Shareable images must not be confused with known images installed with
the shared qualifier:
•

Shareable images -- A shareable image is not copied into the
executable images that link with it (as long as the options
file does not specify /SHAREABLE=COPY or GSMATCH=NEVER).
Thus, only one copy of the shareable image need be on disk, no
matter how many executable images have linked with it.

•

Shared images -- The shared attribute can be assigned to, or
withheld from, any known image -- shareable or executable.
Its assignment results in the creation of permanent, system
global
sections.
Execution of non-CRF global sections
requires only one copy per secti-0n to be in physical memory,
no matter how many processes are running the image to which
the sections belong. Global sections are created for CRF
sections, but the sections are not shared in memory.

When an image is not installed, or is installed without the shared
attribute, each process running the image requires private sections in
memory.
(As of Version 2.0, a shareable image linked to an executable
image without specification of the /SHAREABLE=COPY qualifier or
GSMATCH=NEVER in the options file need not be installed to be
executed.
At image execution time, the system will create private
sections from the shareable image.
The only exception is that a
shareable image containing a writeable non-CRF section must be
installed as a known image with the shared and writeable attributes.)
The number of images that can be installed with the shared qualifier
is restricted by the GBLPAGES and GBLSECTIONS system parameters (see
Chapter 12).

7.2

KNOWN FILE LISTS

The system defines known images on internal data structures called
known file lists. Each known file list contains entries for all known
images whose device, directory, and file type are identical.
For
example, all known images with the file name DBAl: [MAIN] file-name.EXE
would be on one known file list, while all known images with the file
name DBAl:[TEST]file-name.EXE would be on another known file list.
The number of known file lists is restricted by the
generation parameter (see Chapter 12).

7-2

KFILSTCNT

system

INSTALLING KNOWN IMAGES

INSTALL
7.3

INSTALL UTILITY

The system manager runs the INSTALL utility to install and
known images. Table 7-1 summarizes the INSTALL commands.

maintain

Table 7-1
INSTALL Command Summary
Format

Function

file-spec [/OPEN]
[/HEADER RESIDENT]
[/PRIVILEGED[=(priv-name, ••• )]]
[/PROTECTED]
[/SHARED]
[/WRITEABLE]

Installs an image as a known
image with qualifiers as
specified

file-spec /DELETE

Deletes an image as a known
image

file-spec /REPLACE

Associates a known image with
the latest version of the
image file

[file-spec] /LIST
[file-spec] /FULL
/GLOBAL

Displays descriptions of known
images and global sections

7.3.1

Invoking INSTALL

The following command invokes the utility:
$RUN SYS$SYSTEM:INSTALL
The system responds with the following prompt:
INSTALL>
The system manager can then enter any of the commands listed in Table
7-1. These commands follow the standard rules of grammar as specified
in the VAX/VMS Command Language User's Guide.
The file specification must name an existing executable
image. Defaults are applied as follows:
•

Device and directory

•

File type -- EXE

SYS$SYSTEM

7-3

shareable

INSTALLING KNOWN IMAGES

INSTALL (Cont.)
A version number must not be specified. The highest existing version
is
always
used.
Specification of a version number produces
unpredictable results (because file look-ups for known images ignore
version numbers).
Use of the INSTALL utility requires the
privileges.

7.3.2

CMKRNL,

PRMGBL,

and

SYSGBL

Installing Known Images

The system manager installs a known image with the following command:
file-spec [/OPEN]
[/HEADER RESIDENT]
[/PRIVILEGED[=(priv-name, ••• )]]
[/PROTECTED]
[/SHARED]
[/WRITEABLE]
file-spec
File specification of the image being installed
/OPEN
For installation of a permanently open known image
/HEADER RESIDENT
For installation of a known image with a permanently resident
header
(native mode images only); the image is made permanently
open even if /OPEN is not specified
/PRIVILEGED[=(priv-name, ••• )]
For installation of a known image with privileges (executable
images only);
the image is made permanently open even if /OPEN
is not specified; privilege names are specified as shown in
Table 7-2 (Section 7.3.6); if no privilege names are specified,
all privileges are assigned;
parentheses can be omitted for
specification of one privilege
/PROTECTED
For installation of a known image with protected code
/SHARED
For installation of a shared known image;
causes creation of
global sections for the image; the image is made permanently
open even if /OPEN is not specified
/WRITEABLE
For installation of a writeable known image;
also specified, this qualifier is ignored

if /SHARED

not

The following example installs a permanently open, shared known image:
INSTALL> DBAl:[MAIN]STATSHR /OPEN /SHARED

7-4

INSTALLING KNOWN IMAGES

INSTALL (Cont.)
The next example
privileges:

installs

permanently

open

known

image

with

INSTALL> GRPCOMM /OPEN /PRIVILEGES=(GROUP,GRPNAM)
Any process running GRPCOMM receives the GROUP and GRPNAM privileges
for the duration of the execution of GRPCOMM. The full name of
GRPCOMM is assumed to be SYS$SYSTEM:GRPCOMM.EXE.

7.3.3

Deleting Known Images

The system manager deletes a known image with the following command:
file-spec /DELETE
file-spec
File specification of an image installed as a known image
The image's entry on the known file list and any global sections
created for the image are deleted. The image itself (that is, the
image file) remains unaffected. Writeable non-CRF global sections are
written back to disk upon deletion of their known images.
The following example deletes a known image:
INSTALL> DBAl:[MAIN]STATSHR /DELETE

7.3.4

Replacing Known Images

·The system manager replaces a known image with the following command:
file-spec /REPLACE
file-spec
File specification of an image installed as a known image
The image's entry on the known file list becomes associated
latest version of the image file.

with

the

The replace function cannot be used to assign additional attributes to
a known image.
If an image is installed without a resident header,
for example, a specification of /REPLACE /HEADER RESIDENT does not add
this attribute.
The system manager ~ust deleie the known image and
reinstall it with the desired attributes.
The following example replaces a known image:
INSTALL> GRPCOMM /REPLACE
The full name of the
SYS$SYSTEM:GRPCOMM.EXE.

file

specification

7-5

assumed

INSTALLING KNOWN IMAGES

INSTALL {Cont.)
7.3.5

Listing Known Images

The system manager displays information on
sections with the following commands:
•

[file-spec] /LIST

•

[file-spec] /FULL

•

/GLOBAL

known

images

and

global

file-spec
File specification of an image installed as a known image
/LIST
For display of a 1-line description of the specified known image
or
(if no file specification is made) all known images; see
Figure 7-1
$ RUN SYS$SYSTEM:INSTALL
INSTALL>/LIST

Shm
Shm
Shm

DBAO: [SYSEXE]COPY.EXE;l
DBAO: [SYSEXE]BLISS32.EXE;8
DBAO: [SYSEXE]SET.EXE;2

Shar Head
Open
Shar Head 8
Open
Shar Head Priv Open

DBAO: [SYSEXE]SOS.EXE;l
DBAO: [SYSEXE]PIP.EXE;l

Shar

Open Compat
Open Compat

DBAl:[MAIN]STATSHR.EXE;3

Shar

Open

Pro

Noexe

O
8

File specification of known image

Known image has permanently resident header

Known image has privileges

Known image is permanently open

Known image runs in compat i bi 1 i ty mode

Known image contains protected code

€)

Known image is installed in multiport memory

Known image is not executable -- that is, it is a shareable image

Known image is shared

Figure 7-1

Known Image Display -- Brief Description

7-6

INSTALLING KNOWN IMAGES

INSTALL (Cont.)
/FULL
For display of a multiline description of the specified known
image or
(if no file specification is made) all known images;
see Figure 7-2
/GLOBAL
For display of all global sections (whether or not the
was created as a result of INSTALL); see Figure 7-3

section

$
SYS$SYSTEM:INSTALL
INSTALL>USERS /FULL
DBAO: [SYSEXE] USERS. EXE; 1
Head Pr iv Open 0
# of times run
30 8
Privilege bits
00010100,000000006)
Entry adr/size
8006EBF0/64 G)
Window adr/size
800Al7A0/48 CD
Header adr/size
8006A8FC/368 (i»

Same as for brief description (see Figure 7-1)

Number of times image has been run
Privileges allotted to known image;
see
Table
7-2
translation; appears only if image installed with privileges

G) Address (in hexadecimal) of image's entry on known file

list

for
and

the size of the entry in bytes

Address (in hexadecimal) of image's directory window and the size
of
the window in bytes;
appears only if image installed
permanently open

(i» Address (in hexadecimal) of image's resident header and

the size
of the header in bytes;
appears only if image installed with
permanently open resident header

Not shown in example - if image runs in compatibility mode, second
line reads: Compatibility type = 0000
Figure 7-2

Known Image Display -- Full Description

7-7

INSTALLING KNOWN IMAGES

INSTALL (Cont.)
$ RUN SYS$SYSTEM:INSTALL
INSTALL>/GLOBAL

LBRSHR 004
CRFSHR-003

0
System Global Sections

(01000001) WRT CRF
(010003E8) WRT CRF

PRM SYS
PRM SYS

(i)
Pagcnt/Refcnt=l/O
Pagcnt/Refcnt=l/l

27 Global Sections Used, 828 Global Pages Used, 32n8 Global Pages Unused

Global Sections in Multiport Memory "SHM"
VMSRTL 001
(010003FC)
- Basepfn/Pagcnt=00000000/7~
Port 0 PTE Refcnt=35 41)

PRM SYS

Creator Port=O

659 Global Pages Used, 1329 Global Pages Unused
15 Global Sections Used, 17 Global Sections Unused 41)

Display of global sections in local memory

Name of global section

Version number (in hexadecimal) of global section;
high-order
byte
(01
in CRFSHR 003)
contains maJor
identification and
low-order
bytes
(0003E8
in
CRFSHR_003)
contain
minor
identification

The global section is writeable

0
0

The global section is copy-on-reference
The global section
global section

permanent;

TMP

indicates

temporary

and

group

The global section is system-wide;
indicate a group-wide section

(i)

Number of pages in the section and number of times pages from the
section have been referenced

Number of global sections created, number of global
and number of global pages unused in local memory

Display of global sections in shared multiport memory unit;
one
such display appears for each attached multiport memory unit;
the multiport memory unit 2 in the example is named SHM

Number of the port from which the global section was created

Base page frame number
the section

GRP

(in hexadecimal) and number

Figure 7-3

pages

- - - - - ····-····--------------"'

Global Sections Display

7-8

number

used,

pages

INSTALLING KNOWN IMAGES

INSTALL (Cont.)
G)

Page table entry (PTE) reference count;
port where the reference count is not O

one

appears

for

each

Number of global pages used, number of global pages unused,
number of global sections used, and number of global sections
unused in shared memory
Figure 7-3 (Cont.) Global Sections Display

7.3.6

Specifying and Reading Privileges

Table 7-2 names the 30 currently available privileges and identifies
the bit in the privilege vector with which each is associated. The
names are used with the /PRIVILEGED qualifier to install known images
with privileges.
The bit locations are used to interpret the
privilege bits displayed (in hexadecimal) by the /FULL qualifier. The
table lists the bits from low order to high order as they exist in the
first longword shown in the display.
(Two bits of the first longword
and the entire second longword are currently unused.) A display of
00000001,00000000 means the known image has the CMKRNL privilege;
a
display of 00000002 means the known image has the CMEXEC privilege; a
display of 00000003 means the known image has the CMKRNL and CMEXEC
privileges; and so on.
Table 7-2
Privilege Names and Bit Locations
···-

Bit

Name

Bit

Name
-

0
1
2
3
4
5

6
7

7.3.7

CMKRNL
CMEXEC
SY SN AM
GRPNAM
ALLSPOOL
DETACH
DIAGNOSE
LOG IO

---------

8
9
10
11
12
13
14
15

Bit
·-···

GROUP
ACNT
PRMCE B
PRMMB x
PSWAP M
ALTPR I
SETPR v
TMPMB x

-- ·-·-·

Name

Bit

Name

WORLD
MOUNT
OPER
EXQUOTA
NETMBX
VOL PRO
PHY IO
BUGCHK

24
25
26
27
28
29
30
31

PRMGBL
SYSGBL
PFNMAP
SHMEM
SYSPRV
BYPASS
unused
unused

16
17
18
19
20
21
22
23

Terminating INSTALL

The system manager terminates INSTALL and returns to DCL command level
by typing <CTRL/Z>.

7.3.8

Error Messages

Table 7-3 lists the
alphabetical order).

messages

issued

7-9

the

INSTALL

utility

(in

INSTALLING KNOWN IMAGES

INSTALL (Cont.)
Table 7-3
INSTALL Error Messages
Message

Remedy

ambiguous qualifier
or privilege name

Enter the name in its entirety.

command line too long

Shorten the command line or continue
with a hyphen.

error creating global
sections

Submit an SPR. This message
an internal error.

indicates

invalid qualifier or
privilege name

Enter a valid name

known file not found

Check
for
specification.
help.

name already in use

Either use another file specification;
use the REPLACE function (if updating an
existing image); or delete the existing
image with the name to be used.

no room to add a new
known file list

Either delete existing known images, ·or
increase the KFILSTCNT and/or NPAGEDYN
system
parameters
(and
reboot the
system).

unable to make image
header

Increase the size of the PAGEDYN system
parameter (and reboot the system)
if a
lack of space in the paged dynamic pool
is suspected.
However, image headers
exceeding one block in size cannot be
made resident, and this could be the
problem.
(Large headers result from
linking many shareable images.)

shown

Table

7-1.

the
The

correct
file
LIST function can

The INSTALL user also receives error messages from the Create and Map
Section ($CRMPSC) system service and the image activator ($IMGACT).

7.4

OPERATIONS

Certain operational considerations come into play
are installed and used.

7-10

when

known

images

INSTALLING KNOWN IMAGES
7.4.1

Start-Up Procedures

Known file lists only last while the system is up. If the system is
shut down or fails for any reason, all known images must be
reinstalled after the system is booted. For this reason, the siteindependent
start-up
command
procedure, SYS$SYSTEM:STARTUP.COM,
includes an INSTALL run that installs certain system programs as known
images.
The
system manager is encouraged to include in the
site-specific
start-up
command
procedure,
SYS$SYSDISK:[SYSMGR]SYSTARTUP.COM,
an
INSTALL
run
to
install
additional images that are run frequently, that are usually run
concurrently by several processes, or that require special privileges.
{See Chapter 8 for information on the start-up command procedures.)

7.4.2

Order of Installation

In local memory, installing less frequently used images first and more
frequently used images last {on each known file list) enhances
run-time performance.
In multiport memory, installing the more
frequently used images first enhances run-time performance.

7.4.3

Privileges

Images to be installed with privileges
/NODEBUG and /NOTRACE qualifiers to
security.

should be linked with the
maintain system integrity and

Installing a shareable image with privileges does not assign those
privileges to executable images linked with it. The executable images
must be installed with privileges.

7.4.4

Deleting Known Images and Dismounting Volumes

System operations are affected by two characteristics of known images:
•

Deletion -- A known image is not deleted as soon as the
/DELETE qualifier is applied. The deletion occurs only after
all processes using the image have released it.

•

Dismounting -- A volume cannot be dismounted while
file lists associated with it contain entries.

any

known

To dismount a volume, then, the system manager must not only delete
all known images associated with it, but must wait for all processes
using those images to release them and for the system to write
writeable images back to their files.

7-11

INSTALLING KNOWN IMAGES
7.4.5

Shareable Image Files

At execution time, a shareable image must reside in the directory
SYS$SHARE (which is [SYSLIB] on the system disk), or the file
specification of the shareable image must be assigned to the file
name.
For example, if the file specification of a shareable image is
DBAl:[TEST]STATSHR.EXE, the user must assign that file specification
to the logical name STATSHR before running any executable image that
calls STATSHR:
$ DEFINE STATSHR DBAl: [TEST]STATSHR
The file type defaults to EXE.
were
SYS$SHARE:STATSHR.EXE,
necessary.

If the file specification for STATSHR
no
assignment
statement would be

Likewise, one shareable image can be substituted for another without
requiring the calling executable image to relink. The user simply
assigns the file specification of the new shareable image to the file
name of the old shareable image. The following statement assigns
DBAl:[MAIN]STATSHR.EXE as the shareable image for executable images
calling STATSHR:
$ DEFINE STATSHR DBAl: [MAIN]STATSHR
Again the file type defaults to EXE.
If the new image is installed with the shared qualifier, executable
images linked against the old image will be mapped to global sections
for the new image. Otherwise, they will be mapped to private sections
for the new image.
As demonstrated in the example, the old and new images can have the
same name, but must reside in different directories. The system
manager should not substitute one version of a file for another in the
same directory.

7.4.6

Multiport Memory

To install a shared image so that the global sections will reside in a
multiport memory unit, the system manager issues a DEFINE command (an
ASSIGN command could also be used) in the format:
DEFINE GBL$file-name shmem-name:file-name
The following example ensures that any global sections created for an
image whose file name is STATSHR reside in the multiport memory unit
whose logical name is SHRMEMl:
$ DEFINE GBL$STATSHR SHRMEMl:STATSHR
$RUN SYS$SYSTEM:INSTALL
INSTALL> STATSHR /OPEN /SHARED

7-12

CHAPTER 8
START-UP COMMAND PROCEDURES

The software
procedures:

distribution

kit

contains

two

start-up

command

•

SYS$SYSTEM:STARTUP.COM -- Commands that, in general, must be
executed at initialization time for any VAX/VMS system to run
properly

•

SYS$SYSDISK: [SYSMGR] :SYSTARTUP.COM -- An empty file, called by
STARTUP.COM,
that
the
system
manager
can load with
site-specific initialization commands

The system manager can tailor portions of the site-independent command
procedure
(STARTUP.COM) and should load the site-specific command
procedure (SYSTARTUP.COM). These tasks can be accomplished with any
text editor.
The system manager should generally put site-specific
commands in SYSTARTUP.COM rather than modifying STARTUP.COM, as a new
site-independent start-up command file is provided with each major
release of VAX/VMS.

8.1

SITE-INDEPENDENT START-UP COMMAND PROCEDURE

STARTUP.COM is automatically executed immediately after the operating
system has been booted. The command procedure includes commands for
performing housekeeping chores, assigning system-wide logical names,
installing known images, building the I/O data base and loading the
I/O drivers, enabling VAX-11 RMS
file
sharing,
calling
the
site-specific start-up command procedure, and logging out.

8.1.1

Housekeeping Chores

The first two commands in STARTUP.COM ensure th3t execution of the
command procedure occurs without the commands being echoed on the
terminal and without interruption on an error condition:
$ VERIFY = 'F$VERIFY(O)
$ SET NOON
The third command sets the default directory to the
system executable images:
$ SET DEFAULT [SYSEXE]

8-1

location

the

START-UP COMMAND PROCEDURES
8.1.2

Logical Name Assignments

STARTUP.COM makes system-wide logical name assignments for various
system components. The system manager can edit STARTUP.COM to remove
the logical name assignments for components not being used at the site
or to add logical name assignments for other components (such as
VAX-11 BLISS-32 and VAX-11 BASIC}.
NOTE
The logical name assignments for FOR005,
FOR$ACCEPT,
and FOR$READ to SYS$INPUT,
and FOR006, FOR$PRINT, and FOR$TYPE to
SYS$0UTPUT
are
embedded
in VAX-11
FORTRAN, and
need
not
be
stated
explicitly.

8.1.2.l Symbolic Debugger - The VAX-11 Symbolic Debugger requires the
following logical name assignments:
$ ASSIGN/SYSTEM SYS$INPUT: DBG$INPUT:
$ ASSIGN/SYSTEM SYS$0UTPUT: DBG$0UTPUT:

8.1.2.2 COBOL Programs - VAX-11 COBOL-74
following logical name assignments:
$ ASSIGN/SYSTEM SYS$INPUT:
$ ASSIGN/SYSTEM SYS$0UTPUT:
$ ASSIGN/SYSTEM SYS$ERROR:
$ ASSIGN/SYSTEM SYS$INPUT:
$ ASSIGN/SYSTEM SYS$INPUT:
$ ASSIGN/SYSTEM SYS$0UTPUT:
$ ASSIGN/SYSTEM SYS$0UTPUT:

programs

require

the

require

the

COB$ INPUT
COB$0UTPUT
COB$CONSOLE
COB$CARDREADER
COB$PAPERTAPEREADER
COB$ LINEPRINTER
COB$PAPERTAPEPUNCH

8.1.2.3 PASCAL Programs - VAX-11
following logical name assignments:

PASCAL

programs

$ ASSIGN/SYSTEM SYS$INPUT:
PAS$INPUT
$ ASSIGN/SYSTEM SYS$0UTPUT:
PAS$0UTPUT

8.1.2.4 RSX-llM Programs - RSX-llM compatibility mode programs
(such
as BAD, SOS, and PIP} require the following logical name assignments:
$ ASSIGN/SYSTEM
$ ASSIGN/SYSTEM
$ ASSIGN/SYSTEM
$ ASSIGN/SYSTEM
$ ASSIGN/SYSTEM
$ ASSIGN/SYSTEM

'F$LOG("SYS$DISK"}
'F$LOG("SYS$DISK"}
'F$LOG("SYS$DISK"}
'F$LOG("SYS$DISK")
'F$LOG("SYS$DISK")
'F$LOG("SYS$DISK"}

8-2

LB:
LBO:
WK:
WKO:
SP:
SPO:

START-UP COMMAND PROCEDURES
8.1.2.5 System Processors - The language processors,
the
VAX-11
Linker,
the
image activator,
and the help processor require the
following logical name assignments:
$ ASSIGN/SYSTEM 'F$LOG("SYS$DISK") [SYSLIB] SYS$LIBRARY:
$ASSIGN/SYSTEM 'F$LOG("SYS$DISK") [SYSMSG] SYS$MESSAGE:
$ASSIGN/SYSTEM 'F$LOG("SYS$DISK") [SYSHLP] SYS$HELP:

8.1.3

Known Images

STARTUP.COM installs certain system executable
images that are run
frequently, that are usually run concurrently by several processes, or
that require special privileges:
$ RUN SYS$SYSTEM:INSTALL
SETPO
/PRIVILEGED={CMKRNL)
PRTSMB
/OPEN /SHARED
!NIT
/PRIVILEGED=(CMKRNL,PHY IO,SYSPRV)
VMOUNT
/PRIVILEGED=(CMKRNL,DETACH,BYPASS,SETPRV;ALTPRI ,TMPMBX,WORLD,GROUP,EXQUOTA,ACNT,PHY IO,BUGCHK,MOUNT)
DISMOUNT /PRIVILEGED=(CMKRNL,EXQUOTA,BUGCHK)
SUBMIT
/PRIVILEGED={TMPMBX)
REQUEST /PRIVILEGED=(TMPMBX)
MAIL
/PRIVILEGED=(SYSPRV,OPER,GROUP,WORLD) /OPEN
DISPLAY /PRIVILEGED=(CMKRNL,CMEXEC) /OPEN/HEADER RESIDENT /SHARED
LOGINOUT /PRIVILEGED=(CMKRNL,CMEXEC,TMPMBX,EXQUOTA,SYSPRV)
SYS$SHARE:DCLTABLES /OPEN /HEADER RESIDENT /SHARED
SYS$SHARE:DCLINTPRT /OPEN /HEADER-RESIDENT /SHARED
DCL
/OPEN /HEADER RESIDENT /SHARED
SET
/PRIVILEGED=(CMKRNL,SYSPRV,TMPMBX) /OPEN/HEADER RESIDENT /SHARED
SHOW
/PRIVILEGED=(CMKRNL,CMEXEC,TMPMBX,NETMBX,OPER) /OPEN /HEADER /SHARED
sos
/OPEN /SHARED
SYS$SHARE:RSXSHR /OPEN /SHARED /HEADER RESIDENT
SYS$SHARE:RSXUSR /OPEN /SHARED /HEADER-RESIDENT
RSX
/OPEN /HEADER RESIDENT /SHARED
BACKTRANS /OPEN /HEADER_RESIDENT /SHARED
SYS$LIBRARY:DEBUG /OPEN /SHARED
SYS$LIBRARY:TRACE /OPEN /SHARED
SYS$SHARE:VMSRTL /OPEN /SHARED /HEADER RESIDENT

SYS$SHARE:LBRSHR /OPEN /SHARED /HEADER_RESIDENT
SYS$SHARE:CRFSHR /OPEN /SHARED /HEADER_RESIDENT
SYS$SHARE:SUMSHR /OPEN /SHARED /HEADER_ RESIDENT
The system manager should review the above
installed images and
eliminate any that are not used or are used infrequently.
If
necessary, the executable images should be rearranged in reverse order
of frequency of use -- that is, the most frequently used images should
be installed last.

8-3

START-UP COMMAND PROCEDURES
8.1.4

I/O Devices and Drivers

STARTUP.COM automatically connects devices physically attached to
system and loads their I/O drivers:

the

$ RUN SYS$SYSTEM:SYSGEN
AUTOCONFIGURE ALL

For installations that require permanent residence of the console
block storage device, the system manager must explicitly connect the
device by including the following command in the SYSGEN utility run:
CONNECT CONSOLE [/DRIVERNAME=device-driver]
See Chapter 13 for a detailed discussion of the SYSGEN utility.
The console block storage device should also be mounted with a command
of the form:
$ MOUNT /SYSTEM /PROT=(SYSTEM:RWLP)CSAl: CONSOLE
Failure to mount the console block storage device with appropriate
protection permits random users to mount and access it, as the device
is in RT-11 format and has no file protection.

8.1.5

VAX-11 RMS File Sharing

STARTUP.COM enables VAX-11 RMS file sharing with a maximum page
of 20:

count

$ RUN SYS$SYSTEM:RMSSHARE
20
If the system manager estimates a different maximum page count (see
Section 5.4), that figure should replace 20. In particular the system
manager should ensure that this value does not exceed half the size of
paged dynamic memory.

8.1.6

Termination of the Procedure

STARTUP.COM calls the site-specific start-up command procedure, sets
the number of users that can log in at one time, and logs the
site-independent command procedure off:
$ @[SYSMGR]SYSTARTUP.COM
$ SET LOGIN /INTERACTIVE=64
$ LOGOUT/BRIEF
If necessary, the system manager can replace 64 with a user limit that
is more reasonable for the installation.

8-4

START-UP COMMAND PROCEDURES
8.2

SITE-SPECIFIC START-UP COMMAND PROCEDURE

SYSTARTUP.COM, called from STARTUP.COM, performs any commands the
system manager wan~s to place there. Typically, these commands mount
system disks,
assign logical names,
set the characteristics of
terminals and other devices, establish and start queues, install known
images, run the System Dump Analyzer, purge the operator's log file,
submit batch
jobs that are run at the time the system is initialized
and that are periodically resubmitted, manually connect devices and
multiport memory units, install secondary paging and swapping files,
and announce that the system is up and running. The commands shown in
the following sections are provided as models;
they should not be
copied line for line.

8.2.1

System Disks

The followi'ng commands are typical of those used
disks:
$ MOUNT /SYSTEM DBAl:
$ MOUNT /SYSTEM DBA2:
$ MOUNT /SYSTEM DBB2:

mounting

system

BUILDATA
BILLING
FORTRAN

See Chapter 5 for more information on mounting system disks.

8.2.2

Logical Names

The system manager can assign system-wide logical names in addition to
the
logical names assigned in the site-independent start-up command
procedure:
$ASSIGN/SYSTEM 'F$LOG("SYS$SYSDISK") SYSDSK
$ ASSIGN/SYSTEM SYSD$: [SYSMGR] SYSMGR
The lexical function 'F$LOG("SYS$SYSDISK") returns the name of the
system device.
See Chapter 2 of the VAX/VMS Command Language User's
Guide for more detailed information on logical name assignments.

8.2.3

Device Characteristics

The system manager uses a series of SET commands to establish
characteristics of the terminals and other devices on the system:
$
$
$
$
$

SET TERMINAL
SET TERMINAL
SET TERMINAL
SET PRINTER
SET DEVICE

TTC2:
TTDl:
TTD4:
LPAO:
LPAO:

the

/SPEED=300/LA36/PERM
/SPEED=9600/PERM
/SPEED=l200/PERM
/LOWER /NOCR
/SPOOLED

Note that the /SPEED qualifier sets both transmission and reception
speeds to the same value.
Printer characteristics (SET PRINTER and
SET DEVICE above) must be set prior to establishing queues for the
printers.

8-5

START-UP COMMAND PROCEDURES
8.2.4

Queues

The first time the system is booted, batch and print queues must be
initialized and started. Whenever the system is rebooted, the queues
must just be started.
The following commands provide for both
contingencies by testing $STATUS:
START /QUEUE LPAO
IF $STATUS THEN GOTO ENDLPAO
INITIALIZE /QUEUE /FLAG LPAO
START /QUEUE LPAO
START /QUEUE LPBO
IF $STATUS THEN GOTO ENDLPBO
INITIALIZE /QUEUE /FLAG LPBO
START /QUEUE LPBO

$
$

$ ENDLPAO:
$
$
$

$ ENDLPBO:
See Chapter 9
queues.

8.2.5

for

information

initializing

and

starting

Known Images

system manager often installs user and system programs (in
addition to the ones installed in the site-independent start-up
command procedure) so that they can be located quickly, shared, or
provided privileges:

~he

$ RUN SYS$SYSTEK:INSTALL
BLISS32 /OPEN /SHARED /HEADER RESIDENT
COPY /OPEN
LINK /OPEN
USERS /PRIV=WORLD
MACR032 /OPEN /SHARED
TALK /PRIV=(OPER,SYSNAM,WORLD,GROUP,PRMMBX)
DIRECTORY /OPEN
The most frequently used images should be
memory,
but first in shared memory.
information on installing known images.

8.2.6

installed last
See Chapter 7

in local
for mo re

System Dump Analyzer

Each time the system is booted, the system manager should run the
System Dump Analyzer (in case the system failed the last time it was
running):
$ RUN SYS$SYSTEM:SDA
SYS$SYSTEM:SYSDUMP.DMP
COPY [SYSERR]SYSDUMP.DMP
SET OUTPUT LPAO:SYSDUMP.LIS
SHOW CRASH
SHOW STACK/ALL
SHOW SUMMARY
SHOW PROCESS /PCB /PHD /REGISTERS
SHOW SYMBOL/ALL
EXIT

If further information is required, the system manager can invoke the
System Dump Analyzer for an interactive session upon completion of
startup.
See the VAX/VMS ~~~-!!'1...Pump~r.::!~1..Y_?_~_!___ .!3_e:_.~-~-!:~E~-~-e: __ _!:!anual.

8-6

START-UP COMMAND PROCEDURES
Operator's Log File

8.2.7

Each time the system is booted, the system manager should purge
but the last two or three versions of the operator's log file:

all

$ PURGE /KEEP=2 [SYSMGR]OPERATOR.LOG

Standard Batch Jobs

8.2.8

Some sites may have batch jobs that are submitted at system start-up
time and that resubmit themselves to run at intervals as long as the
system is running. For such jobs, the SUBMIT command is used in the
start-up file:
$ SUBMIT SYS$SYSTEM:LOGJOBS

Manually Connected Devices and Multiport Memory Units

8.2.9

The system manager runs the SYSGEN utility to connect devices not
automatically connected in STARTUP.COM and to initialize or connect
multiport memory units:
$ RUN SYS$SYSTEM:SYSGEN
CONNECT NET /NOADAPTER /DRIVER=NETDRIVER
SHARE MPMl SHR MEM 1 /!NIT

8.2.10

Secondary Paging and Swapping Files

The system manager runs the SYSGEN utility to install secondary paging
and swapping files:
$ RUN
INSTALL
INSTALL

8.2.11

SYS$SYSTEM:SYSGEN
DRAS: [SYSTEM]PAGEFILE.SYS /PAGEFILE
DRAS: [SYSTEM]SWAPFILE.SYS /SWAPFILE

Announcement

The last command in SYSTARTUP.COM typically announces to all terminals
that the system is up and running:
$ REPLY /ALL /BELL "VAX/VMS System Initialized"

8-7

CHAPTER 9
BATCH AND PRINT JOBS

System performance can be greatly influenced by how the system manager
establishes spooled devices, creates and controls input and output
queues, and controls batch and print jobs.
Typically, the system
manager of a VAX/VMS operating system is responsible for performing
the following four closely related functions:
•

Establishing spooling of input and output -- The VAX/VMS
operating system supports input spooling of batch job files
from card readers and transparent spooling of output files for
line printers and terminals. Using DCL commands, the system
manager specifies which output devices are to be spooled.
Section 9.1 describes spooling and the use of DCL commands to
establish spooled devices.

•

Controlling batch
jobs -- Section
9.2
describes
batch
processing and the use of DCL commands to control batch jobs.

•

Controlling print jobs
Section 9.3 describes queuing output
to line printers and the use of DCL commands to control print
jobs.

•

Creating and controlling terminal queues -- See Section 9.4.

The system manager need not learn all the inner workings of spooling
and queuing. However, a working knowledge of how to establish spooled
devices and how to create and control queues is essential for the
system manager to keep the system running efficiently. The kind of
working knowledge that the system manager needs presupposes
a
familiarity with the DCL commands listed in Table 9-1. The use of
these commands is restricted to users who have operator privilege
(OPER), typically, the system manager and system operators. The
VAX/VMS Operator's Guide fully describes these commands.
In addition, three other DCL commands play a role in
batch and print jobs:

the

control

(or files)
QUEUE -- Displays information about a file
• SHOW
queued for batch execution or for output. No privilege is
needed to use this command.
•

SET QUEUE -- Changes the attributes of a
queued for batch execution or for output.

file

(or

files)

Ordinarily, no privilege is needed to use this command;
operator privilege
(OPER)
is needed, however, to use the
command to:
- Modify queued jobs entered by a member of another group
- Increase the queue priority of a job
9-1

BATCH AND PRINT JOBS
e

DELETE/ENTRY

Deletes jobs from queues.

No privilege is needed to delete entries queued by oneself;
world (WORLD) or operator privilege (OPER) is needed, however,
to use this command to delete a queued job entered by a member
of another group, and group privilege (GROUP) is needed to
delete a queued job entered by another member of the same
group.
These commands are described
Lang u a ~~--~-5-~ r__'__~_g:-i i ~~-.

fully

the

VAX/VMS

Command

Table 9-1
Operator Commands Used in Regulating Spooling and Queuing
~---------·-·····-·--···.--·-···---·······---·--·

····---···----·--·-······--.----·- -----·-····-··········---·---·---·--···-··------- ----·--·--····-~·-·-·-···----

Command

Function

- - · · - - - - - -----------·----··--···•·· ······-·······-··---·--------·- ----+--------···--·--·--····-··· --·-. --·-·--·-·-·--·-----····

SET DEVICE/SPOOLED

Establishes spooled printers
or terminals, and assigns
queues to them

SET DEVICE/NOSPOOLED

Turns off spooling of
printers or terminals

INITIALIZE/QUEUE

Creates queues

DELETE/QUEUE

Deletes queues

START/QUEUE

Starts queues

STOP/QUEUE

Stops queues

ASSIGN/QUEUE

Assigns queues to devices

DEASSIGN/QUEUE

Deassigns queues from
devices

ASSIGN/MERGE

Empties queues of jobs and
places them in other queues

STOP/ABORT

Aborts printing of files
currently being printed

STOP/REQUEUE

Stops the printing of jobs
currently being printed and
requeues them at the end

9.1

SPOOLING

Spooling is the technique of using secondary storage to buffer data
passing between slow I/O devices
(such as line printers and card
readers) and physical memory. The slow devices, which can be either
the ultimate sources or the ultimate destinations of buffered I/O
data, are called spooled devices; the secondary storage devices are
called intermediate devices.
The system manager establishes the spooled devices;
to all other
users, and their programs, the mechanism of spooling is transparent.

9-2

BATCH AND PRINT JOBS

Input spooling makes input from a spooled device
(such as a card
reader) available for processing by placing it into a file on an
intermediate device
(such as a disk).
Input spooling is used
principally to create, from card reader input, batch input files on
disk. After they are spooled to disk, batch jobs are queued for
processing according to their priority.
Output spooling makes output from the processor available
for
transmission to a spooled device (such as a line printer) by placing
it into files on an intermediate device
(such as a disk).
Output
spooling is used principally to create printer output files on disk.
After they are spooled to disk, print jobs are queued for printing
according to their priority.
As a rule, programs demand input and produce output at irregular
intervals during their execution. If programs were allowed to read
directly from slow devices and to write directly to slow devices, the
execute time of programs would be limited by the speed of the slow
devices. If a process output directly to a printer, the process would
be tied up for the time it took to print the listing. Also, other
processes needing the printer would have to wait.
The actual transfer of inputs from a spooled device to an intermediate
device or the transfer of outputs from an intermediate device to a
spooled device is carried out by processes called symbionts.
Input symbionts read input at the speed of the input spooled device
and buffer it in a file on the intermediate device. Later, when the
input is needed, it is read directly from the file on the intermediate
device rather than from the spooled device. While one set of input
data is being processed, the input symbiont is free to read another
set of input data into another file on the intermediate device.
Output symbionts read data from an intermediate device and write the
data to an output spooled device at the speed of that output device.
The data on the intermediate device is generated by programs that
produce outputs directly into files on the intermediate device. The
I/O waiting time of programs is thus minimized. When an output file
is complete, it is queued for printing by an output symbiont. As with
input symbionts, there is an overlapping here.
While an output
symbiont is printing a file stored temporarily on an intermediate
device, another program can be producing another output file on the
intermediate device.

9.1.1

Establishing Spooled Devices

Card readers are spooled by default. To use a card reader without
spooling, users must allocate the reader before making it ready to
read a card deck. By default, also, the queue SYS$BATCH is used to
queue spooled jobs. Thus, no special command is needed to establish
card readers as spooled devices.
On the other hand, the operator command SET DEVICE must be used to
establish a line printer or a terminal as a spooled device. The use
of this command is restricted to users with the OPER privilege.
The
VAX/VMS Operator's Guide describes the SET DEVICE command in detail.

9-3

BATCH AND PRINT JOBS

Typically, the system manager must decide which devices to include in
the system's basic complement of spooled devices. Often, the system
manager sets up devices for spooling by making entries in the system
start-up command procedure.
At a minimum, the system manager should see that at least one line
printer is set spooled when the system is started up. In a system
with only one line printer, this is the default system printer.
The
system manager need not set a card reader spooled, because card
readers are spooled by default.
Depending on system configuration and anticipated operational needs,
more spooled devices can be established at start-up. Moreover, in the
course of normal operations (to meet special operational needs), the
system manager or the operator can define still other devices as
spooled devices without having to reboot the system.
Normally, all
line printers should be spooled.
Finally, and most important, on a system with both spooled input
devices and spooled output devices, the system manager must create and
start at least one batch queue to handle spooled input and one output
queue to handle output for each spooled output device.

9.1.2

Turning Off Spooling

The system manager or operator can, as necessary, turn off spooling to
spooled printers and terminals by use of the SET DEVICE command.

9.2

BATCH JOBS

Batch jobs can be submitted to
execution in two ways:

the

VAX/VMS

system

and

queued

for

•

As batch job files submitted by use of the $JOB command (see
the VAX/VMS Command Language User's Guide) from a card reader.
These batch job-""IITes---~ire--~s~pooiecf onto disk by an input
symbiont and placed in a batch queue according to their
priority. Unless the $JOB card specifies otherwise, the name
of this batch queue is SYS$BATCH (by default). From the batch
queue, batch jobs are selected for execution.

•

As command procedure disk files submitted by use of the SUBMIT
command (see the VAX/VMS Command Language User's Guide).
These files are also placed in a batchque1.ie-·and selected for
execution according to their priority. Again, by default, the
name of this batch queue is SYS$BATCH.

Batch jobs cannot be executed unless at least one batch queue has been
created on the system and unless that queue has been started. By
default, this is the batch queue SYS$BATCH.
In the VAX/VMS system, many jobs, or streams, can be executed at the
same time from each of several batch queues. Thus, the system manager
can create and start several batch queues at once and can specify the
number of jobs, or streams, that can be executed at the same time from
each queue.

9-4

BATCH AND PRINT JOBS

Among the jobs in a batch queue that has been started, the one with
the highest priority is the first candidate for execution. Whether or
not that job is actually started up, however, depends on an evaluation
of the following limits and conditions:
•

The maximum number of batch jobs allowed to be executed from
the queue at the same time. The system manager specifies this
limit with either the INITIALIZE /QUEUE command or the START
/QUEUE command.

•

The maximum number of all jobs allowed to be executed
system at the same time.

•

The number of jobs currently being executed in the system.

the

Hence, the highest priority batch job in a queue is started up only if
both of the following conditions are satisfied:

9.2.1

•

Fewer than the maximum number of
currently running from the queue.

batch

•

The system is not saturated with other jobs.

jobs

allowed

are

Creating Batch Queues

The operator command INITIALIZE /QUEUE is used in creating, or
initializing, a batch queue. The use of this command is restricted to
users with the OPER privilege to execute operator functions.
The
VAX/VMS Operator's Guide describes the INITIALIZE /QUEUE command in
detail.
Typically, the system manager must decide on the number of batch
queues for an installation, on the job limit of each queue, on the
priority of each queue, and on the swap mode of each queue.
Often,
the system manager creates batch queues by making entries in the
system start-up command procedure.
Setting up batch queues is not restricted to start-up time.
In the
course of normal operations, the system manager or operator can create
batch queues as operational needs dictate.

9.2.2

Starting Batch Queues

The execution of batch jobs from a batch queue (dequeuing) can only
take place if the batch queue has been started. The operator command
START /QUEUE starts a batch queue.
The use of this command is
restricted to users with the OPER privilege. The VAX/VMS Operator's
Guide describes the START /QUEUE command in detail.
Typically, the system manager must see that batch queues created by
use of the INITIALIZE /QUEUE command are started. Often, the system
manager starts batch queues by making entries in the system start-up
command procedure.
Starting batch queues is not restricted to start-up time.
In the
course of normal operation, the system manager or operator can start
queues as operational needs dictate.

9-5

BATCH AND PRINT JOBS
9.2.3

Stopping Batch Queues

The system manager or operator can, as necessary, abort a job in a
batch queue or disable all processing from the queue until the queue
is restarted by use of the START /QUEUE command.
The STOP /QUEUE
command is used to stop batch queues. The VAX/VMS Operator's Guide
describes this command in detail.
·--------··----

9.2.4

Deleting Batch Queues

The system manager or operator can delete batch queues, as necessary,
by use of the DELETE /QUEUE command. The VAX/VMS Operator's Guide
describes this command in detail.

9.2.5

··----

Batch Versus Interactive Jobs

The system manager should normally encourage users to submit large
jobs (such as compiling and linking large programs) as batch jobs and
reserve interactive use of the system for jobs that do not require
extensive resources.
A technique toward this end is to (1) restrict
the working set size of interactive jobs by providing small
(for
example, 150) WSDEFAULT and WSQUOTA values in the UAF records and (2)
expand the working set size of batch jobs by providing large (for
example, 512) WSDEFAULT and WSQUOTA values in the START /QUEUE and
INITIALIZE /QUEUE commands. The system manager ~an likewise restrict
and expand time limits on jobs by setting the CPU values.

9.2.6

Guides to Setting Up Batch Queues

The following rules of thumb are useful in setting up
for a system that is predominantly interactive:
1.

Set up one batch queue
default batch queue.

named

SYS$BATCH,

Give SYS$BATCH the following characteristics:
a.

Job limit -- 1 to 4

Priority -- 3 or 4

Swapping mode

Working set default -- 150 to 1024

Working set quota -- 150 to 1024

CPU default

INFINITE

CPU maximum

INFINITE

the

batch
name

swapping enabled (by default)

9-6

queue
of

the

BATCH AND PRINT JOBS
The system manager executes the following command procedure to
and start this queue:

create

$ START/QUEUE SYS$BATCH
$ IF $STATUS THEN GO TO BATCH DONE
$ INITIALIZE/QUEUE/BATCH/JOB LIMIT=l/PRIORITY=3/WSDEFAULT=700/WSQUOTA=700/CPUDEFAULT=IN~INITE/CPUMAXIMUM=INFINITE SYS$BATCH
$ START/QUEUE SYS$BATCH
$ BATCH DONE:
The first START command and the test
existing queue is not initialized.
Normally,
these commands
procedure (see Chapter 8).

are

contained

$STATUS

ensure

that

start-up

command

the

The following rules of thumb are useful in setting up batch queues for
a system that is predominantly a batch system and in which editing is
the principal interactive activity:
1.

Set up three batch queues as follows:
a.

SYS$BATCH -- the default batch queue.

FAST -- a high-priority queue for executing high-priority
jobs that should not be swapped out of memory.

SLOW -- a low-priority background queue for processing
low-priority jobs.
Typically, these are large jobs with
large requirements for physical memory.
Usually,
it
is
uneconomical to swap such jobs out of memory.
The system
operator can adjust the system workload by stopping and
restarting background queues as needed.

Give SYS$BATCH the following characteristics:
a.

Job limit -- 6 to 10

Priority -- 4 (by default)

Swapping mode -- swapping enabled (by default)

Give FAST the following characteristics:
a.

Job limit

-- 1 (by default)

Priority

-- high ( 5, for example)

Swapping mode

-- swapping disabled

Give SLOW the following characteristics:
a.

Job limit

Priority

-- 1 (by default)

-- low ( 3, for example)
Swapping mode -- swapping disabled

9-7

BATCH AND PRINT JOBS
NOTE
This
configuration
should
not
be
attempted on a small system (under lMB).
Additionally, the system manager should
add up the pages required for the batch
working sets and insure that enough
fluid memory remains for interactive
jobs.
Otherwise, the system manager
must reduce the number of batch jobs or
make the FAST and SLOW jobs swappable.
In particular, the system manager must
not let fluid memory drop below the
value of the WSMAX system parameter, or
a deadlock could result.
The system manager executes the following command procedure to
and start these queues:

create

$ START/QUEUE SYS$BATCH
$ IF $STATUS THEN GOTO BATCH DONE
$ INITIALIZE/QUEUE/BATCH/JOB-LIMIT=6 SYS$BATCH
$ START/QUEUE SYS$BATCH
$ BATCH DONE:
$ START7QUEUE FAST
$ IF $STATUS THEN GOTO FAST DONE
$ INITIALIZE/QUEUE/BATCH/PRlORITY=lO/DISABLE SWAPPING FAST
$ START/QUEUE FAST
$ FAST DONE:
$ START/QUEUE SLOW
$ IF $STATUS THEN GOTO SLOW DONE
$ INITIALIZE/QUEUE/BATCH/PRlORITY=3/DISABLE SWAPPING SLOW
$ START/QUEUE SLOW
$ SLOW DONE:
Normally, these commands are contained in the
command procedure (see Chapter 8).

9.3

site-specific

start-up

PRINT QUEUES

Unless a line printer is associated with a physical queue (a queue
that has the same name as the line printer) and unless that queue has
been started, no queued output can occur on that line printer.
Print jobs are queued for processing in one of two ways: without the
direct intervention of a user (that is, implicitly) or with the direct
intervention of a user (that is, explicitly).
An implicitly spooled file is created when a program or DCL command
sends its output to a spooled printer. When an implicitly spooled
print file destined for a spooled printer is closed, the file is
placed in a print queue. Both the spooling of the output file to an
intermediate device and the subsequent queuing of a job consisting of
this file occur without the direct intervention of a user.
By use of the PRINT command, a user can explicitly queue a disk file
or several files for printing. The VAX/VMS Command Language User's
Guide describes the PRINT command in detail. The disk file or--rfres
specified in the PRINT command are queued as a print job; if several
files make up a print job, they will be printed together.

9-8

BATCH AND PRINT JOBS
Print jobs are placed in queues according to
queues can be any one of the following:

their

priority.
with

(that

These

•

Physical device queues -- Queues associated
named for) a specific line printer.

is,

•

Generic queues -- Queues from which files can be printed out
on any available line printer that has correctly matching
characteristics.

or logical, queues -- Queues that are not associated,
• Named,
even indirectly, with any device. To obtain printed output
from a named, or logical, queue, the system manager or
operator must explicitly assign the queue to a printer. The
command ASSIGN /QUEUE is used for this purpose.
From these queues, print jobs are selected for initiation. Among the
jobs in a print queue for a particular printer at any given time, the
job with the highest priority is the one chosen for printing.
By default, print jobs queued by use of the PRINT command are placed
in the queue named SYS$PRINT. Thus, to use the default version of the
PRINT command in a system with only one line printer, the name
SYS$PRINT is equated with the name of the physical line printer. To
use the default version of the PRINT command in a system with several
line printers of matching characteristics, SYS$PRINT is normally
established as the name of a generic queue.
The maximum number of physical device queues that can be printing at
one time is restricted to the value of the MAXPRINTSYMB system
parameter. In particular, system managers of systems configured using
the Buser and 16USER standard system parameter files should note that
the default value of this parameter is 1. See Chapter 12 for further
inf6rmation on system parameters.

9.3.1

Creating Print Queues

The operator command INITIALIZE /QUEUE is used in creating, or
initializing, a print queue. The use of this command is restricted to
users with the OPER privilege to execute operator functions.
The
VAX/VMS Operator's Guide describes the INITIALIZE /QUEUE command in
detail.
Typically, the system manager must decide on the number of print
queues for an installation and on their attributes. Often, the system
manager creates print queues by making entries in the site-specific
start-up command procedure.
Setting up print queues is not restricted to start-up time.
In the
course of normal operations, the system manager or operator can create
print queues as operational needs dictate.

9-9

BATCH AND PRINT JOBS
9.3.2

Starting Print Queues

The initiation of print jobs from a print queue (dequeuing) can only
take place if the print queue has been started. The operator command
START /QUEUE starts a print queue.
The use of this command is
restricted to users who have the OPER privilege to execute operator
functions. The VAX/VMS Operator's Guide describes the START /QUEUE
command in detail-.--A·n-opfT6ns__. that can be specified in the
INITIALIZE /QUEUE command can also be specified in the START /QUEUE
command.
Typically, the system manager must see that print queues created by
use of the INITIALIZE /QUEUE command are started. Often, the system
manager starts print queues by making entries in the site-specific
start-up command procedure.
Starting print queues is not restricted to start-up time.
In the
course of normal operations, the system manager or operator can start
queues as operational needs dictate.

9.3.3

Stopping Print Queues

The system manager or operator can abort a job in a print queue,
suspend the printing of a job currently being printed, or disable
processing from the queue entirely until the queue is restarted by use
of the START /QUEUE command. The STOP /QUEUE command is used to stop
print queues and to suspend printing of jobs. The VAX/VMS Operator's
Guide describes this command in detail.

9.3.4

Deleting Print Queues

The system manager or operator can delete print queues, as necessary,
by use of the DELETE /QUEUE command. The VAX/VMS Operator's Guide
describes this command in detail.

9.3.5

Assigning a Named, or Logical, Print Queue to a Printer

The operator command ASSIGN /QUEUE is used in
assigning,
or
redirecting, a named, or logical, print queue to a printer. The use
of this command is restricted to authorized users with the OPER
privilege to execute operator functions. The VAX/VMS Operator's Guide
describes the ASSIGN /QUEUE command in detail.
To produce printer output, a logical queue must first be assigned to a
printer and then started.
Typically, the print files of a group of low-priority users can be
placed in a logical queue and held there until off-peak hours. Then,
to print the files, the system operator can assign the queue to a line
printer and start the queue.

9-10

BATCH AND PRINT JOBS
9.3.6

Deassigning a Named, or Logical, Print Queue from a Printer

The operator command DEASSIGN /QUEUE is used in deassigning a named,
or logical, print queue from a printer. The use of this command is
restricted to users with the OPER privilege. The VAX/VMS Operator's
Guide describes the DEASSIGN /QUEUE command in detail.

9.3.7

Vertical Page Size

By default, the various system utilities
(including the editors,
compilers, and linker) produce listings with a vertical page size of
66 lines. The system manager may change the vertical page size for
listings produced by the native mode utilities by specifying a numeric
value, an integer in the range of 30 to 99, inclusive, for the system
logical name SYS$LP LINES. The following example changes the vertical
page size for all users to 60 lines per page:
$ DEFINE /SYSTEM SYS$LP_LINES 60
Individual users may change the vertical
process basis.

page

size

group

A vertical page size of less than that specified by SET PRINTER /PAGE
(which defaults to 64) causes a skip to the head of the next form each
time the specified line count is reached.
A greater vertical page
size causes the excess lines to overflow to the next form and then a
skip to head of form when the count is reached.

9.3.8

Forms Control

The system manager or operator specifies the forms type of a print
queue with the /FORMS TYPE qualifier of the INITIALIZE /QUEUE or START
/QUEUE command. If a-user enters a print job with a forms value
(/FORMS qualifier of the PRINT command) different from that of the
queue, the job is placed on a hold status until the forms value of the
queue is set equal to the forms value of the job.
(The operator
should stop the queue, physically change the forms, and start the
queue specifying the new value for the /FORMS_TYPE qualifier.)
The forms type can be specified as a number or an
Alphabetic
codes
must
be
defined
SYS$SYSDISK: [SYSMGR] :FORMSTYPE.DAT, one code per
following format:

alphabetic code.
file
in
the
line,
in
the

% code number comments

The system manager can include lines of 'text in the file if desired.
Only lines beginning with a percent sign are taken as code-number
definitions. The following example defines the code NORMAL as the
number 0:
% NORMAL 0 NORMAL LINE PRINTER PAPER

A specification of /FORMS=NORMAL (or /FORMS=N, /FORMS=NO, and so on)
is interpreted to mean /FORMS=O. Note, however, that the first match
found in FORMSTYPE.DAT prevails with no ambiguity checks made, so that
potentially conflicting names (that is, names starting with the same
letter) should be avoided.

9-11

BATCH AND PRINT JOBS
9.3.9

Printer Characteristics

The system manager or operator specifies the printer characteristics
of a print queue with the /CHARACTERISTICS qualifier of the INITIALIZE
/QUEUE or START /QUEUE command. If a user enters a print job with a
characteristic (/CHARACTERISTICS qualifier of the PRINT command) not
included in those for the queue, the job is placed on a hold status
until the characteristics of the queue are set to include all the
characteristics of the job.
(The operator should stop the queue,
physically change the characteristics of the printer, and start the
queue specifying the new values for the /CHARACTERISTIC qualifier.)
A characteristic can be specified as a number or an alphabetic code.
Alphabetic
codes
must
be
defined
in
the
file
SYS$SYSDISK:[SYSMGR]CHARTYPE.DAT, one code per line, in the following
format:
% code number comments

The system manager can include lines of text in the file if desired.
Only lines beginning with a percent sign are taken as code-number
definitions. The following example defines the code REDINK as the
number 3:
% REDINK 3

Subsequent INITIALIZE /QUEUE or /START /QUEUE and PRINT commands can
use the number 3 or the alphabetic code REDINK to describe one printer
characteristic.
Note, however, that the first match found
in
CHARTYPE.DAT
prevails
with no ambiquity checks made, so that
potentially conflicting names (that is, names starting with the same
letter) should be avoided.

9.3.10

Guides to Setting Up Print Queues and Spooled Line Printers

The following rules of thumb are useful in setting up
print queues and spooled line printers:

and

regulating

Normally, set all line printers spooled.

To produce output on a spooled line printer, initialize a
print queue with the same name as the spooled printer and
start that queue.

If more than one line printer is on the system, enable
generic printing from as many print queues as possible, and
make at least one print queue (SYS$PRINT) a generic queue.
Queues for line printers that are in remote locations, that
use special forms,
or
that
possess
unique
printer
characteristics should not be enabled for generic printing.

To use special forms or apply unique printer characteristics
to a general-purpose queue, stop the queue, physically change
the forms or apply the printer characteristics, and start the
queue
with
appropriate /FORMS TYPE or /CHARACTERISTICS
qualifiers. After the special jobs are printed, stop the
queue, physically reset the forms or printer characteristics,
and start the
queue
with
the
original
/FORMS
or
/CHARACTERISTICS values.

9-12

BATCH AND PRINT JOBS
Figures 9-1 through 9-4 illustrate some of
the
most
common
arrangements of spooled line printers and print queues. These figures
can be used, with the rules of thumb listed above, as guidelines in
setting up spooled line printers and print queues.
NOTE
The commands shown in the
examples
assume
manual
entry
at run time.
Command
procedures
especially
start-up
command
procedures
containing INITIALIZE and START commands
should contain logic to ensure that an
existing queue is not initialized.
See
Chapter 8.
If an existing queue is
initialized, any jobs in that queue are
lost.
Figure 9-1 illustrates a typical spooling
for a system with one line printer.
figure produce the following results:

and queuing configuration
The commands listed in this

The line printer LPAO is set spooled.

System wide, the logical name SYS$PRINT is equated with the
name LPAO. The equivalence of these names is recorded in the
system logical name table.

The print queue LPAO is initialized and started.

All print jobs explicitly directed to the printer LPAO are
placed in the queue LPAO and are printed from that queue.

All print jobs that normally would be placed by default in a
queue named SYS$PRINT {if that queue existed) are actually
placed in the queue LPAO {in this case, the system default
print queue) and are printed from that queue.
COMMANDS

!!> ~:; i::: T x:i i::: 1) I c i::: . . . ~:;r-o 01... i::: ri :=Lr:" ()0 1... i::- t10
!!; t1'.:;'.::; I CJN/~:;y'.::;TCM l... !'.>(10
'.::;Yb!f,i::·1::; I NT
!!> I i'~ I T I (1 I... I Z. [/Cl UEU C / i::· 1... (1 C !.. r:· (1 0
~!> HT t1F~·T /OULU[
!... F'(10

Line Printer

Queue

LPAO
Default Print
Queue

LPAO

Figure 9-1 Setting Up a Spooled Printer and a Print
Queue on a System with One Line Printer
9-13

BATCH AND PRINT JOBS
Figure 9-2 illustrates a typical spooling and queuing configuration
for
a
system
with
two
line
printers that have the same
characteristics. The commands listed in this figure produce the
following results:
1.

The line printer LPAO is set spooled.

The line printer LPBO is set spooled.

The generic queue SYS$PRINT is initialized and started.

Physical queues LPAO and LPBO are initialized
with generic printing enabled by default.

All print jobs explicitly directed by use of the PRINT
command to one of the two printers are placed in the queue
associated with the specified printer.

Print jobs queued by use of the PRINT command without device
specification are placed by default in the generic queue
SYS$PRINT. From the generic queue, jobs are printed on
whichever printer is free, by way of either of the two
physical queues, LPAO or LPBO.

Spooled print files destined either for LPAO or for LPBO are
placed in the generic queue SYS$PRINT, which was associated
with both these printers. From the generic queue, these jobs
are printed on whichever printer is free.

and

started,

Figure 9-3 illustrates a typical spooling and queuing configuration
for a system with three line printers:
two that have the same
characteristics and one that uses special forms, has unique printer
characteristics, or is in a remote location. The configuration shown
in Figure 9-3 is basically the same as the one in Figure 9-2, with the
addition of the spooled printer LPCO and the creation and starting of
the queue LPCO. Because of the special forms, unique characteristics,
or the remote location, printer LPCO is not suited for general
printing.

9-14

BATCH AND PRINT JOBS
COMMANDS
~!i

·'!>
~!:.

:or:: 1) I c i::: ./ ::~. r· ou1.. F :u Li::· ;:-:·1 •o::i
::;r::T nr:: :i: ci:: ./:::::r:·uo1..i::::o 1r:·no
I i··~ I T J r':1 L J :z F /DUL UE:: .:' i::· 1. . ti Ci/. C! CNF r:: J C ~:; Y:::; 'Ii r:· r:: l i'l T

:::. r:: T

1
.)

:!; '.:;T f'.:1F::T /OULU[ :::::y··::::+i::·i::: TNT
'!; INITI l...I:ZE::./UUFUF/.Fl..t1C L.. F·(·,o
:!: :::::T (,r:.:T nur::ur:: 1. . r·t,o
~!>
I i--! I T I I .. J Z F ,. .· UUFU F ..-' r:· L.. (1 Ci 1. . F' I·:
+ : :; T 1"1h'T ciui:::ur:: !.. F>f-: ()
1
.)

Queues

Line Printers

SYS$PRINT

Generic Print
Queue

LPAO
Generic Printing
Enabled

LPAO

LPBO
Generic Printing
Enabled

LPBO

Figure 9-2 Setting Up Spooled Printers and Print
Queues on a System with Two Line Printers with
the Same Characteristics

9-15

BATCH AND PRINT JOBS
COMMANDS

$ SET DEVICE/SPOOLED LPAO
$ SET DEVICE/SPOOLED LPBO
$ SET DEVICE/SPOOLED=LPCO LPCO
$ INITIALIZE/QUEUE/FLAG/GENERIC
$ START/QUEUE SYSSPRINT
$ INITIALIZE/QUEUE/FLAG LPAO
$ START/QUEUE LPAO
$ INITIALIZE/QUEUE/FLAG LPBO
$ START/QUEUE LPBO
$

SYSSPRINT

INITIALIZE/QUEUE/FLAG/NOENABLE_GENERIC_PRINTING LPCO
START/QUEUE LPCO

Queues

Line Printers

SYS$PRINT
Generic Print
Queue

LPAO
Generic Printing
Enabled

LPAO

LPBO
Generic Printing
Enabled

LPBO
LPCO
No Generic
Printing

LPCO

Figure 9-3 Setting Up Spooled Printers and Print Queues on a System
with Three Line Printers; Two with the Same Characteristics and One
with Special Characteristics or in a Remote Location

9-10

BATCH AND PRINT JOBS
Thus, with the following exceptions, the commands listed in Figure 9-3
produce the same results as the commands listed in Figure 9-2:
1.

The line printer LPCO is set spooled.

The physical queue LPCO is
generic printing disabled.

Only print jobs explicitly directed to the printer LPCO are
ever placed in the queue LPCO; no generic printing is ever
done on printer LPCO by way of the queue SYS$PRINT.

initialized

and

started

with

Figure 9-4 adds still another feature to the configuration illustrated
in Figure 9-3. This is a logical queue, which is a named, nongeneric
queue that is not directly associated with any line printer.
When a
logical queue is assigned to a line printer and started, however,
output to a line printer can occur.
Logical queues can be used, for example, to hold print jobs of
low-priority users for printing during off-peak hours. To channel the
print jobs of these users into a logical queue, the name of the
logical queue
(HOLD, for example) must be assigned to the name of
their default print queue (SYS$PRINT).
As shown in Figure 9-4, the INITIALIZE /QUEUE command is used to
initialize the logical queue HOLD. This queue is initialized with
generic printing disabled. When the queue HOLD is assigned to the
printer LPBO and started, the jobs in the queue HOLD are printed on
the line printer LPBO by way of the physical queue LPBO.

9-17

BATCH AND PRINT JOBS
COMMANDS

:!: 1:,r,
nr 1,i:rcc/~:;r:uc:i1. . 1::ci
L,,r·;1•:·
. :::r"i' x:ii::::,,:1cc,':::;r·uu1.1:ci
L.r:.ro
'!; ::::,E:'f {1[ 1.JTCF. '.::;r:.uuL..I :O' !,,F>C(\
i r>C·)
+ Ti•' I ·r J (1 I., J / [ / DUEUE ,''Fl...1.'1 [i ./[1[ i.JE r:: :r c
+. :::, T (1 F: T . . . Ci UEU E '.::; Y'.::; +PF: I NT
:!; l (.1 I T I (l 1. .. I 7 [ / U U E U [ /FL. i'l 1:-!
!.. r:· 1 :) O
$ START/QUEUE
L..PAO
'!; Ir·~ IT J t1L.. I ZE /. UUE U[ .'FL.tiCi
1. .. r:· HO
+. : : ; T (1 h: T . /Cl U EU F L.. r:· B 0
'!; l j\j I T I (1 L.. I l F ,.. uu[ uE /FI.. (1 u / i'! 0 EH (i :c 1... F ... uE 1-l F i:;: J c . . r:· F;: I r-.i T l Hu
$ ::::;T t1F:T ./Ul.JEUE
L..F'CO
'.!; J H I T J (·1 L. I Z F / Cl U EU E /FL (1 U ./ i\I Ct [ i-l (1 :c: I... I : .... C·! [ N E F;: J C ... r:· h: I NT TNU
'~ (1 ::; :; I 1::-; N,/DUE U E
L F> P 0
H0 L. D
~!; '.::: T (1 F;: T / U U EU F
H 0 I... fl

I. .. r: c 0
H 01. .. fl

Line Printers

Queues

SYS$PRINT
Generic Print
Queue

LPAO
Generic Printing
Enabled

LPAO

1-----~-·.-·-~

LPBO
Generic Printing
Enabled

LPBO

LPCO
No Generic
Printing

HOLD
Logical Print
Queue

LPCO

Figure 9-4 Setting Up Spooled Printers and Print Queues
Adding a Logical Queue to the System with Three
Line Printers

9-18

BATCH AND PRINT JOBS
9.4

TERMINAL QUEUES

Terminal queues are print queues destined for terminal devices
(which
are probably being used as remote printers, never interactively).
They are created and controlled by use of the same commands that are
used in creating and controlling regular print queues. Print jobs in
generic print queues are not dequeued to a terminal device queue
unless the generic queue is initialized or started with the /TERMINAL
qualifier.

9-19

CHAPTER 10
ERRORS AND OTHER SYSTEM EVENTS

The system provides several tools for recording and reporting errors
and other system events. These tools include facilities for logging
and reporting system events, logging operator messages, and reporting
problems to DIGITAL. The system also provides facilities for dumping
itself on a failure and reporting the contents of the dump -- see the
VAX/VMS System Dump Analyzer Reference Manual.

ERROR LOG

10.1

The system automatically writes messages to the latest version of a
file named SYS$SYSDISK: [SYSERR]ERRLOG.SYS as the following events
occur:
•

Errors
Device errors, machine
checks,
bus
errors,
synchronous backplane interconnect
(SBI) alerts, soft error
correcting code (ECC) errors, asynchronous write errors, hard
ECC errors

•

Configuration changes -- Volume mounts and dismounts

•

System events -- Cold start-up, warm start-up, crash start-up,
message from Send Message to Error Logger ($SNDERR) system
service, time stamp

The system manager reports the information in the error log by running
the SYE utility.
Since the system continues to log messages to
ERRLOG.SYS and creates a new version of the file if the current one is
locked, the user should rename the file before running SYE against it.
Recommended as the new name is ERRLOG.OLD since SYE uses this name as
a default.

10.1.1

SYE Utility

The system manager runs SYE to selectively report the contents
error log file. The reports are as follows:
•

Roll-up

Reports totals in each category of errors.

For device errors, reports the contents of the
• Cryptic
device registers at the time of each error. The figures are
in hexadecimal without explanation.
•

Brief
Briefly reports the
nature
configuration change, and system event.

each

error,

•

Standard
Fully reports the nature
configuration change, and system event.

each

error,

10-1

ERRORS AND OTHER SYSTEM EVENTS

SYE
The reports, which are 80
terminal.

columns

wide,

can

displayed

the

The following command invokes the SYE utility:
$ RUN SYS$SYSTEM:SYE
SYE displays a series of prompts to which the system manager
with a value or, to take the default, a carriage return:
Prompt
INPUT FILE
-OUTPUT FILE
- OPTIONS
- DEVICE NAME
- AFTER DATE
- BEFORE DATE

responds

Valid Response

[SYSERR]ERRLOG.OLD]
[SYS$0UTPUT]
[ROLL-UP]
[ <CR> ]
[FIRST ENTRY]
[LAST ENTRY]

[

?
?
?
?
?
?

input-file-spec
output-file-spec
report-option
category-type
after-date-spec
before-date-spec

input-file-spec
Input log file;
defaults
to
the
highest
version
of
SYS$SYSDISK: [SYSERR]ERRLOG.OLD;
omitted fields default to those
for VAX-11 FORTRAN unit 1
the
highest
version
of
SYS$SYSDISK:[default-directory]FOR001.DAT
(SYE is written in
FORTRAN)
output-file-spec
Output report file; defaults to SYS$0UTPUT, which is usually the
user's terminal;
omitted fields default to those for VAX-11
FORTRAN
unit
2
the
highest
version
of
SYS$SYSDISK: [default-directory]FOR002.DAT
report-option
R (roll-up), C (cryptic), B (brief), or S
to R

(standard);

defaults

category-type
CP (errors other than for devices), CO (configuration changes),
SY
(system events), or the name of a device
(errors and
configuration changes on a specific device);
defaults to all
category types; a device name can be generic to specify a set of
devices (for example, DB to specify all devices whose names begin
with DB); the colon need not be included on the device name; a
minus sign can precede the category type to indicate all category
types except the specified one (for example, -DBI to specify all
categories except the device whose name is DBI)
after-date-spec
Time on error log
date/time format;

after which reporting begins,
defaults to beginning of log

before-date-spec
Time on error log before which reporting
date/time format; defaults to end of log

ends,

absolute

SYE does not issue error messages on incorrect input, but simply
reprompts.
If an error message appears in the report, the systemmanager should rerun SYE to eliminate the error. Error messages are
reported as VAX-11 FORTRAN messages.
Table 13-2 (Chapter 13) lists and describes the valid device names.
10-2

ERRORS AND OTHER SYSTEM EVENTS
10.1.2

Operations

The error logging facility consists of three parts:
•

A set of executive routines that detects errors and events and
writes relevant information into error log buffers in memory

•

A process called ERRFMT that periodically empties the error
log buffers, transforms the descriptions of the errors into
standard formats, and stores the formatted information in a
file on the system disk

•

The SYE utility

The executive routines and the ERRFMT process operate continuously
without user intervention. The routines fill up the error log buffers
in memory with raw data on every detected error and event.
When one
of the available buffers· becomes full, or when a time allotment
expires, ERRFMT automatically writes the buffers to ERRLOG.SYS.
Sometimes a burst of errors can cause the buffers to fill up before
ERRFMT can empty them. In this case, the system merely assigns a
sequence
number to the errors and events that occur, without
preserving further information. As soon as ERRFMT frees the buffer
space, the executive routines resume preserving error information in
the buffers.
The ERRFMT process displays an error message and deletes itself if it
encounters excessive errors while writing the error log file. The
system manager can restart ERRFMT by invoking the ERFSTART command
procedure in the system manager's directory:
$ SET DEFAULT SYS$SYSDISK: [SYSMGR]
$ @ERFSTART
The command procedure
account (UIC [1,4)).

must

invoked

from

the

system

manager's

10.1.2.1 Using Error Reports - The error reports generated by SYE are
useful tools in two basic ways:
•

Reports aid preventive maintenance by identifying areas within
the system that show potential for failure.

•

Reports speed the diagnosis of a failure by documenting
errors and events that led up to the failure.

the

The detailed contents of the reports are most meaningful to DIGITAL
field service personnel.
However, the system manager can use the
reports as an important indicator of the system's reliability.
For
example, when a report shows that a particular device is producing a
relatively high number of errors, the system manager can consult
DIGITAL field service. By running a diagnostic program to investigate
the device, field service can attempt to isolate the source of the
errors.
Once identified, the source of the errors can possibly be
eliminated and a failure averted.

10-3

ERRORS AND OTHER SYSTEM EVENTS
In the event that a system component does fail, a field service
representative can study error reports of system activity leading up
to and including the failure. For example, if a device fails,
the
system manager can generate error reports immediately after the
failure.
One report might describe in detail all the
errors
associated with the failed device and occurring within the last 24
hours; another report might summarize all types of errors that
occurred within the same time period. The summary report can put the
device errors into a system-wide context.
The
field
service
representative can then run the appropriate diagnostic program for a
thorough analysis of the failed device.
Using the combined error
logging and diagnostic information, the field service representative
can proceed to correct the device.
The information made available by the error logging facility is
essential to efficient maintenance of a VAX/VMS system. Error reports
allow the system manager to track system performance and to anticipate
failures.
In turn, field service personnel rely on the reports as an
aid to both preventive and corrective maintenance. Overall, effective
use of the error logging facility, in conjunction with diagnostic
programs, can significantly reduce the amount of system downtime.
Because the file ERRLOG.SYS can be renamed and the renamed error log
file can then be copied to a removable volume, error reports can be
generated either at the site where the errors occurred or at any other
VAX/VMS installation. For example, a field service representative can
rename and copy the error log file to take back to the field service
off ice, where another VAX/VMS system can be used to generate error
reports. Alternatively, a system manager can rename and copy to a
disk file a version of the error log file that covers a crucial period
of system activity. When a field service representative arrives on
site, he or she can generate one or more reports from the copied file
as well as from the current version of ERRLOG.SYS.

10.1.2.2 Maintaining the Error Log Files - While SYE is accessing
ERRLOG.SYS, ERRFMT cannot write any error information into it.
Therefore, if SYE is accessing the highest version of ERRLOG.SYS when
ERRFMT needs to log an error, ERRFMT creates a new version of the
file. The new version picks up logging errors where the previous
version left off. All the versions of the ERRLOG.SYS file remain on
the system disk until a user explicitly manipulates them in some way.
The fewer the log files, the simpler and more efficient it is to
generate log reports. The system manager or operator can take steps
to minimize or control the number of versions created.
For example, when generating several reports from the current error
log file, a user should first rename the error log file to ERRLOG.OLD
and then use the renamed file as input to SYE. In this way, only one
new error log file is created, and SYE does not prevent ERRFMT from
accessing the new file. In addition, the user ensures that SYE is
accessing the same error log file for each report.

10-4

ERRORS AND OTHER SYSTEM EVENTS
A responsibility of the system operator or system manager is to devise
a plan for maintaining the versions of the error log file. One way to
do this is to rename the highest version of ERRLOG.SYS on a daily
basis.
This action causes a new error log file to be created and
allows the old file (which was renamed) to be copied to a back-up
volume where it can be kept as long as needed. For example, an
operator could rename the current copy of ERRLOG.SYS every morning at
9:00 to ERRLOG.OLD.
To free space on the system disk, the operator
could then back up the renamed version of the error log file on a
different volume and delete the renamed file from the system disk.
Note that caution should be taken to ensure that error log files are
not deleted inadvertently.

10.2

OPERATOR'S LOG FILE

The operator's log file (SYS$SYSDISK: [SYSMGR]OPERATOR.LOG) is a system
management tool that is useful in anticipating and preventing hardware
and software failures.
By regularly examining the operator's log
file, the system manager can often detect tendencies, or trends,
toward failures and can thereby see that corrective action is taken
before these failures occur.
The system operator should, therefore, print out copies of the
operator's log file regularly, and the system manager should retain
these copies for reference. See the VAX/VMS Operator's Guide for
descriptions of the messages that appear in this file and for
instructions on printing copies of it. Figure 10-1 shows some typical
messages of the operator's log file.

DPcom'
0Pcom,
Qpcom,
0Pcom,
Qpcom'
Qpcom,
Qpcom,
0Pcom,

13-AF'R-1978 20:09:43.07, Losfile initialized, oPerator=_Qf'AO:
06: 57: 53. 70, Dev:ic.:~ offline' LF'A-0:
o6:5s:25.70, Device off line' LF'AO:
06:58:57.70, Device off line' LF'AO:
06:59:29.70, Device off line, LF'AO:
07:00:01.70, Device off line, LF'AO:
07:00:33. 70, rievice off line' LF'AO:
07:01:os.70, Device off line, LF'AO:

0Pcom' 11: 30: 47. 70, Device off line, Lf'AO:
Qpcom, 11: 31: 19. 70, Device off line, LF'AO:
Qpcom' 11:31:51.70, Device offlin~' LF'AO:
Qpcom, 14-AF'R-1978 13:59:30.27, Terminal enabled, operator=-TTC3:
0Pcom, 13:59:41.99, ROGERF'
Accnt=VMS
Qpcom, TTC3: , •TEST

Figure 10-1

Operator's Log File

10-5

ERRORS AND OTHER SYSTEM EVENTS
10.3

REPORTING SOFTWARE PROBLEMS

To inform DIGITAL about problems with the VAX/VMS operating system or
about errors in VAX/VMS software documents, the system manager should
use the Software Performance Report (SPR), which is illustrated in
Figure
10-2.
Complete directions for completing the SPR form
accompany the form itself.
A supply of SPR forms is included in the VAX/VMS software distribution
kit; more forms can be obtained from an SPR center. The addresses of
these centers are listed on the backs of the forms.
On a fatal bugcheck, the system manager should (if possible) copy the
file SYS$SYSTEM:SYSDUMP.DMP to tape, and send the tape to DIGITAL with
the SPR.

10-6

--,-------

ERRORS AND OTHER SYSTEM EVENTS
SPR NO.:

SOFTWARE
PERFORMANCE
REPORT

229433

OFTWARE SERVICES NO.:

PAGE - - - - O F - - - -

OPERATING SYSTEM

rERSION

]~YsTEM-PROGR-AMOR DOC~M~NTT:-~J~E-~SION~~~~CU~EN~T~PA~:~~- _ JDA~E -

,_(_S_E_E_E_X_A_M_P_L_E_l_N_l_N_S_T_R_U_,C_T_l_O_N_S_)_ _

DEC OFFICE

DO YOU HAVE SOURCES? -

NAME:

REPORT TYPE/PRIORITY

FIRM:

D
D

ADDRESS:

SUBMITTED BY:

PROBLEM/ERROR
SUGGESTED ENHANCEMENT
OTHER

-~-----------------

PHONE:

---------

CAN THE PROBLEM BE REPRODUCED AT WILL?
YES

-co_u_~--:;=-H-1-s-sP_R_~_v_E_B~-E-N_P_R~~-;;~ E~ e v

ATT ACHM EN T-5
MAG TAPE 0

-y~[J-

FLOPPY DISKS0
LISTING 0
DECTAPE
OTHER - - - - - - -

....C_P_U_T_Y_P_E_ _

BETTER OR MORE DOCUMENTATION?
PLEASE EXPLAIN IN PROVIDED SPACE BELOW .

~J_s_'E_R_l_A_L_N_O__• __ ]~EMORY SIZE- f DISTHIBUTION MEDIU~ -~ rYSTEMDEvlcE

YES

I N~T
DO

NOD

PUBLISH

ALL SUBMISSIONS BECOME THE PROPERTY OF DIGITAL EQUIPMENT CORPORATION
DATE RECEIVED
DATE TO MAINTAINER

-·- - - - - - -

--------to~~AATTEE~ LAONGs--wG--EEDR oSEFNF~-

· ··-DATE RECEIVED FROM MAINTAIN'ER"-"" _____ _,_ -

LOGGED ON

_ _ _ ___,.

·-""""'""

EN-1044G·07 (35C) R0878

ADMINISTRATIVE SERVICES GROUP, SWS

Figure 10-2

Software Performance Report (SPR)

10-7

··-

"'"

-------'

PART III
SYSTEM CONFIGURATION

The system manager is responsible for maintaining an
configuration by:
•

Generating a
workload

system

that

meets

•

Monitoring system activity to determine
levels

•

Adjusting system
performance

parameters

the

and

user's

other

adequate

system

resources

adequate
values

and

performance
to

improve

Part III consists of Chapters 11 through 14:
•

Chapter 11: Tuning Considerations -- Discusses the factors
involved in generating, monitoring, and adjusting the system
for proper performance

•

Chapter
12:
System
Parameters -- Defines
the
system
parameters and the standard system parameter files supplied by
DIGITAL

•

Chapter
13:
System
Generation -- Provides
instructions for the SYSGEN utility

operating

14:
DISPLAY
Utility -- Provides
operating
• Chapter
instructions
for the DISPLAY utility and describes the
displays

CHAPTER 11
TUNING CONSIDERATIONS

Hardware
resources -- mainly
physical
memory
and
secondary
storage -- constitute the primary constraint on system performance.
Adequate hardware resources for the workload generally
provide
adequate performance with little need for tuning, although a certain
amount of tuning will enhance system performance. Inadequate hardware
resources for the workload generally provide inadequate performance
regardless of the tuning effort.
Only in the middle ground of
just-adequate or just-inadequate resources does tuning become a
significant factor. This resource level normally manifests itself in
situations where the user is trying to make do with a small system, or
where the user's workload has been increasing over a period of time
with no corresponding addition of resources. For practical purposes,
tunable general-purpose systems can be considered in two categories:
•

Small systems -- Small in this context means a system with
tight resources.
Normally, this would be a system of lMB or
less, although conceivably a large system with an enormous
workload might fit this category. The primary emphasis in
tuning a small system lies in optimizing the use of physical
memory and achieving a balance between the use of physical
memory and disk I/O.

•

Large systems -- Large in this context means a system with
plentiful resources for the workload. Normally, this would be
a system with physical memory in excess of lMB, although
conceivably a small system with a light workload might fit
this category. The primary emphasis in tuning a large system
is on decreasing disk I/O.

Before undertaking a major tuning effort, the manager of a small
system should weigh the time and effort involved in the venture
against the alternative solution: the purchase of additional physical
memory or faster disks.

11-1

TUNING CONSIDERATIONS

11.l

PRETUNING CONSIDERATIONS

Before starting the tuning effort, the system manager should ensure
that hardware resources are adequate for the workload, that the
workload is distributed as equably as possible, and that frequently
used code is shared.

11.1.1

Hardware Resources

VAX/VMS hardware configurations are designed to handle
workloads:
Workload

Memory

Disk Storage (Capacity)
(Peak Transfer Rate)

1-8 users

512KB

2 RK07 (56MB}
(538KB per second}

9-16 users

512KB

2 RM03 (134MB}
(1200KB per second}

17-32 users

1024KB

2 RP06 (352MB)
(806KB per second)

33-48 users

1536KB

2-4 RP06 (352-704MB}
(806KB per second)

49-64 users

2048KB

2-4 RP06 (352-704MB)
(806KB per second)

the

following

These figures are approximate. The precise number of users for a
memory configuration depends on the size and make-up of the programs
those users are running. If the processing requirements of all users
consist of editing (with an SOS-type editor), compiling, linking, and
running small programs, the system will probably support around the
maximum number of users -- 32 on the lMB system, for example. If, on
the other hand, processing consists mainly of developing very large
programs, sorting, and/or using keypad-type editors, the system will
probably support around the minimum number of users.
For user
production programs, the system manager can lean toward the maximum
figure for programs that contain little data or use VAX-11 RMS to
manipulate data on a per-record basis, and toward the minimum figure
for programs that contain very large data structures. The trend for
native VAX/VMS software products (assemblers, compilers, and so on) is
to require larger amounts of memory, mainly through the use of larger
data structures to reduce I/O activity (resulting in shorter response
times and faster throughput}.

11-2

TUNING CONSIDERATIONS

The system manager can add physical memory up to a total of 8MB of
local memory plus 8MB of multiport memory. Adding memory will most
certainly enhance system performance (as discussed below under "Large
Systems"), but will not necessarily permit more than n4 concurrent
users. Depending on the workload, the system manager should expect to
overload the processor in the range of n4 to 128 users.
Actual disk capacity is a function of user file requirements.
However, the system should ideally contain at least two disk drives
and two disk drive controllers. The system manager should make an
effort to place the disk drives containing the execution files (system
and user image files, paging file, swapping file, and dump file) and
the disk drives containing user data on separate controllers. This
procedure will cut down on the contention between system I/O and user
I/O activities.

11.1.2

Workload Distribution

The system manager is reminded
(although the subject may seem
elementary)
to distribute the workload as evenly as possible over the
time the computer is running.
While scheduling interactive users
evenly is normally not possible due to the weight of convention on
standard working and sleeping hours, either or both of the following
techniques are workable:
•

Run large jobs as batch jobs -- Force the submission of large
jobs on a batch basis;
regulate the number of batch streams
so that batch usage is high when interactive usage is low.
Chapter 9 discusses batch jobs.

•

Restrict system use -- Do not permit more users to log in at
one time than the system can support with an adequate response
time.
The system manager can restrict the
number
of
interactive users with the IJOBLIM system parameter (see
Chapter 12) and the DCL command SET LOGINS
(see the VAX/VMS
Operator's Guide).

11.1.3

Sharing Code

The site-independent start-up command procedure creates permanent
global sections for system programs and subroutines by installing them
as known images with the shared attribute. The system manager should
add to these,
in the site-specific start-up command procedure, user
programs and subroutines that have reached a production status or are
in general use.
Users should, of course, be encouraged to write
shareable code. Chapter 8 of this manual covers start-up command
procedures, while Chapter 7 discusses known images.

11-3

TUNING CONSIDERATIONS
11.2

BACKGROUND DISCUSSION

Figure 11-1 illustrates the physical memory and secondary storage
requirements for supporting process contexts on a VAX/VMS system.

~
....
~
. ...
~
---------------.. ......
~
....

RESIDENT
SYSTEM

PAGE
CACHE

resident
executive
routines

free
page
list

f---------""
non paged
dynamic
memory

1-------modified
page
list

IMAGE FILES

BALANCE SET

user working sets

......

SECTION FI LES

PAGING FILE

system working set

Figure 11-1

SWAPPING FILE

VAX/VMS Memory Configuration

Each active process is associated with a working set that contains the
pages of code and data being used in the execution of that process.
Initially, the working set expands to accommodate clusters of image
and section
(that is, code or data explicitly mapped by a process to
its virtual address space) pages from their disk files as they are
referenced.
The size of the working set (unless altered by certain
system services) equals the cumulative number of pages brought in from
the image and section files, until the working set limit is reached.
At that point, newly referenced pages are accommodated by pushing
current pages out of the working set, on a first-in, first-out (FIFO)
basis.
The system automatically adjusts the working set limit by deducting
pages if the rate of page faults is low, and adding pages if the rate
of page faults is high
(unless the system manager turns off the
automatic working set adjustment feature).
A page fault occurs
whenever a referenced page does not exist in the working set.

11-4

TUNING CONSIDERATIONS
If sufficient space exists, working sets reside in real memory.
If
sufficient space does not exist, one or more of the working sets must
reside on the swapping file. The system determines which working sets
to retain in physical memory and which to outswap to disk according to
each process's state and priority, and the quantum:
•

Current state -- The currently executing
physical memory.

process

resides

•

Compute state -- Processes that can execute immediately if
given processor time should reside in physical memory;
they
are the last candidates for outswapping.

•

Waiting state -- Processes that are waiting for a local event
flag, are hibernating, or are suspended constitute the primary
candidates for outswapping.

•

Priority -- Among processes in the same state, those with
lowest priorities are outswapped.

•

Quantum -- A process cannot be outswapped until it
has
received services for at least one SYSGEN-specified period of
time called a quantum.

the

The system reevaluates physical memory
versus
disk
residency
requirements whenever a process's state changes. For example, if a
local event flag is posted changing the state of a swapped-out process
from wait to compute, the system will
(if possible) place the
process's working set in physical memory, even if this
means
outswapping the working set of another process.
One working set is reserved for those portions of the system executive
that are pageable. In addition, the system requires user working sets
for the system processes ERRFMT, OPCOM, and JOB CONTROL, the Files-11
and magnetic tape ACPs, and the print symbionts, although these
working sets are also swapped.
Pages pushed out of a working set that has reached its limit initially
remain in physical memory by going to the page cache: to the free
page list if they were not written on, or to the modified page list if
they were written on.
The free page list expands to accommodate
additional pages being pushed out of working sets. The free page list
contracts as pages are allocated to processes requesting additional
memory; pages leave the free list on a FIFO basis.
However, the
system cannot reduce the size of the free page list below a
SYSGEN-specified lower limit.
The modified page list expands to accommodate additional pages being
pushed out of working sets, but only to a SYSGEN-specified upper
limit. At that point, pages are written back to their section files
or
(in the case of demand-zero and copy-on-reference pages) to the
paging file, then moved to the free list, until the size of the
modified page list drops below a SYSGEN-specified lower limit. If a
page that was pushed out of a working set is referred to again, a page
fault occurs. The system must return the page to the working set from
either:
•

The page cache -- If the page is still in physical memory
(that is, on the free page list or the modified page list), a
"cheap" page fault occurs as no disk I/O is required.

•

An image, section, or paging file -- If the page has been
overwritten, an "expensive" page fault occurs as the page must
be reread from disk.

11-5

TUNING CONSIDERATIONS

Although pages move through the working sets and page cache on a FIFO
basis, the return, of referred-to-again pages from the page cache to
the working sets produces a net effect of deleting the least
frequently used pages and retaining the most frequently used pages.
When an image terminates, all section pages that were modified are
written back to the section files. All pages associated with the
program in physical memory are released.
(Shared pages, however, are
not released if currently in use by another process.)

TOOLS

11.3

The system manager controls system performance overall with the system
parameters, and on a per-user basis with the entries in the UAF.
Certain DCL commands also aid in improving system performance.
The
effects of tuning can be monitored with the DISPLAY utility.

11.3.1

System Parameters

In lieu of the standard system parameter file appropriate for the
hardware configuration
(as shown in Table 12-7, Chapter 12), the
system manager can create a tailored system parameter file with the
SYSGEN utility. In tailoring a system parameter file for performance,
the system manager should concentrate on the following parameters:
•

WSMAX -- Maximum working set limit for all users

PFRATH,
• AWSTIME

PFRATL,
WSINC,
WSDEC,
Settings for automatic

AWSMIN,
AWSMAX,
and
adjustment of working set

limits
•

QUANTUM -- Minimum amount of time in the balance set a process
will receive before being outswapped; also, maximum amount of
service a process will receive before losing control in a
round-robin situation

SWPRATE

Interval between inswaps of compute-bound

proce~ses

•

FREELIM
page list

Minimum number of pages that must be

the

•

PFCDEFAULT -- Number of pages read into a working set during a
paging operation involving disk I/O

•

MPW HILIM, MPW LOLIMIT, and MPW WRTCLUSTER -- Point at which
(expressed in-number of pages)-pages will be written from the
modified page list to the paging file, point at which the
writing will stop, and the number of pages written in each I/O
operation

•

NPAGEDYN and IRPCOUNT -- Size of nonpaged dynamic pool

•

SYSMWCNT -- Number of pages in system working set

•

RMS_DFMBC -- Blocking factor for VAX-11 RMS I/O operations

•

ACP MAPCACHE,
ACP_HDRCACHE,
ACP DIRCACHE,
ACP FIDCACHE,
ACP-EXTCACHE, ACP QUOCACHE, ACP WRITEBACK, ACP SYSACC, and
ACP-WINDOW
Space used by Files=11 ACP processes for data
str~ctures containing directory and file information

11-6

free

TUNING CONSIDERATIONS
•

ACP MULTIPLE,
ACP SHARE,
ACP SWAPFLAGS,
and
ACP-WORKSET -- Characteristics of ACP- processes:
whether
concurrent I/O operations use the same or separate copies of
the same ACP (ACPs are single threaded); whether concurrent
ACPs share code; whether ACP working sets are locked in
memory or are swapped;
the working set limit for ACP
processes

Chapter 12 explains these and the other system parameters. Chapter 13
explains how to use the SYSGEN utility to modify the parameters.
Although some of the other system parameters affect performance, they
should be considered only in a fine-tuning effort. The system manager
should leave them at the standard settings, or reset them to
reasonable values dictated by other factors.

11.3.2

User Authorization- File

The system manager can affect performance
resetting the following UAF entries:

per-user

basis

•

WSQUOTA -- Maximum working set limit for the user's processes;
cannot exceed the system parameter WSMAX

•

WSDEFAULT -- Working set
cannot exceed WSQUOTA

•

PRIORITY -- Base priority for the user's processes

•

CPUTIME -- CPU time permitted the user's processes

limit

Chapter 2 explains these and the other UAF
entries do not affect system performance.

11.3.3

for

the

user's

entries.

The

processes;

other

UAF

DCL Commands

The following DCL commands affect system performance:
Commands

Function

SET WORKING SET /QUOTA
RUN /MAXIMUM_WORKING_SET

Reset working set
quota for the
user's processes;
cannot exceed UAF
quota

SET WORKING SET /LIMIT
RUN /WORKING_SET

Reset the working set
limit for the user's
processes; cannot
exceed current quota

INITIALIZE QUEUE /BATCH /WSQUOTA*
START QUEUE /BATCH /WSQUOTA*
SUBMIT/WSQUOTA
$JOB /WSQUOTA

Set working set
quota for jobs
submitted to a batch
queue

INITIALIZE QUEUE /BATCH /WSDEFAULT*
START QUEUE /BATCH /WSDEFAULT*
SUBMIT /WSDEFAULT
$JOB /WSDEFAULT

Set working set limit
for jobs submitted to a
batch queue

11-7

TUNING CONSIDERATIONS
Commands

Function

INITIALIZE QUEUE /BATCH /CPUMAXIMUM
/CPUDEFAULT*
START QUEUE /BATCH /CPUMAXIMUM
/CPUDEFAULT*
SUBMIT /CPUTIME
$JOB /CPUTIME

Set the CPU time
permitted batch jobs

SET PROCESS /[NO]SWAPPING

Enables or disables the
swapping in and out of
memory of a process's
working set

SET /PRIORITY
RUN /PRIORITY

Reset the base priority
of a process

SET RMS DEFAULT /BLOCK_COUNT

Resets the blocking
factor for VAX-11 RMS
I/O operations

The VAX/VMS Command Language User's Guide and the VAX/VMS Operator's
Guide (thos·e··-·comma-nds wTth___an ··as.ter-Tsk)·-·-·explain the DCL commands.
--

1L3. 4

DISPLAY Utility

With the DISPLAY utility, the system manager can monitor activities
indicative of system performance.
Useful displays in this regard
include:
•

VAX/VMS PROCESSES (also called the USERS display} -- For .each
current process,
lists the current state and priority of the
process, the physical memory being used in terms of shareable
pages and total pages, the cumulative direct I/O operations,
the cumulative page faults, and the cumulative CPU time

•

I/O SYSTEM RATES -- Provides the sizes of the free and
modified page lists, plus the following information, expressed
in cumulative units per interval, per second during the
preceding interval, and per second since the start of the
display: direct I/O operations, buffered I/O operations, page
faults,
pages read, read I/O operations, pages written, write
I/O operations, and total paqes inswapped

•

PAGE MANAGEMENT STATISTICS -- Substantially the same as I/O
SYSTEM RATES, plus the page faults accumulated by the system
working set

•

TIME IN PROCESSOR MODES -- Percentage of time
processor mode, plus idle time
·

spent

each

•

NONPAGED POOL STATISTICS -- Amount
nonpaged dynamic pool

space

the

•

NUMBER OF PROCESSES IN SCHEDULER STATES -- Provides a snapshot
of the state of the system

unused

Chapter 14 provides operating instructions and detailed
for the DISPLAY utility, and examples of its use.

11-8

descriptions

TUNING CONSIDERATIONS
The following DCL commands complement the DISPLAY utility:
•

SHOW SYSTEM -- Displays information on the processes currently
executing, including state, priority, elapsed CPU time, number
of page faults, and physical memory occupied

SHOW PROCESS /ACCOUNTING -- Displays information about the
executing process,
including the cumulative number of page
faults, the peak working set size, the peak virtual size,
elapsed total time, and elapsed CPU time

STATUS -- Displays
• SHOW
process,
including the

executing
information about
the
working set limit, the amount of
physical memory being used, the elapsed CPU time, and the
number of page faults

•

SHOW WORKING SET -- Displays a process's
quota, and authorized quota

•

LOGOUT /FULL -- Displays information about the terminating
process, including the number of page faults, the peak working
set size, the peak virtual size, elapsed total time, and
elapsed CPU time

The VAX/VMS Command Language User's Guide provides
of these

commands~~·--~

.. ·--·

--- ·-----·---,.-~

current

full

limit

and

explanations

~~ -~------

SMALL SYSTEMS

11.4

As the background discussion indicates, the
factors
affecting
efficient performance are interrelated and somewhat complex. The
major factors for small systems include working set size, page cache
characteristics, priority, quantum, compute swapping rate, and system
use of the system.

11.4.l

Working Set Size

Decreasing working set limits increases the number of working sets
that can occupy real memory at one time,
reducing the need for
swapping, but i n c re as i n g the need
( i n mo st cases ) f o r pa g i n g •
Normally, excessive paging is preferable to excessive swapping for the
following reasons:
•

$wapping involves the output and input of an entire working
set, while paging usually involves a subset of the working
set.

•

Paging does not necessarily involve disk I/O operations due to
the page cache. In fact, the ratio of disk I/O operations to
page faults may be quite low.

11-9

TUNING CONSIDERATIONS

A process can, then, show larger and larger numbers of page faults
with little effect upon performance. Nevertheless, a point does occur
where the reducti-0n of the working set limit significantly cramps a
process's performance. The optimal working set limit, then, would be
just above the point where performance drops sharply.
A 2-phase
strategy is recommended for setting working sets at their optimal
sizes:
1.

Figure initial working set limits for
processing on a rule-of-thumb basis.

different

types

Adjust working set limits based on observed behavior.

11.4.1.1 Initial Working Set Limits - For processing involving system
components, the following working set limits are suggested:
•

Small (60 to 200 pages) -- For editing, and for compiling
linking small programs ("typical" interactive processing)

and

•

Large (200 to 500 pages) -- For compiling and linking
programs, and for sorting ("typical" batch processing)

large

Working set limits for user programs depend on the code-to-data ratio
of the program and the amount of data in the program. Programs that
are mostly code -- that include a small or moderate amount of data, or
use VAX-11 RMS to process data on a per-record basis -- should require
only a small working set. The amount of code should not matter as
long as it is reasonably linear.
Programs that manipulate large
amounts of data internally require large working sets.
The following guidelines are suggested for initially
sets limits:

setting

working

•

System generation parameters -- Set WSMAX at the highest
number of pages required by any program. At the "typical"
installation where use of the system ranges from smalr
interactive tasks to large batch streams, the initial setting
for WSMAX would be in the range of 200 to 500.
(However,
WSMAX has other effects on the system -- see Section 12.1.8.)

•

UAF options -- For each user, set WSQUOTA at the largest
number of pages required by a program that the user will run
interactively. Set WSDEFAULT at the median number of pages
required by a program that the user will run interactively.

•

Batch queues -- For each batch queue, set WSQUOTA
(using the
INITIALIZE or START command) at the largest number of pages
required by a job that a user will submit. Set WSDEFAULT at
the median number of pages required by a job that a user will
submit.

This arrangement effectively forces users to submit large jobs for
batch processing, as the jobs will not run efficiently interactively.
The system manager can further restrict the user who attempts to run a
large job interactively by imposing CPU time limits in the UAF.

11-10

TUNING CONSIDERATIONS
11.4.1.2 Adjustments - The system automatically adjusts working set
limits based on observed behavior and the values of the PFRATL,
PFRATH, WSINC, WSDEC, AWSMIN, AWSMAX, and AWSTIME system parameters.
The example below demonstrates the resultant working set limits as the
system automatically adjusts a working set with a default limit of 100
pages and system parameter values of PFRATL=2, PFRATH=512, WSINC=21,
and WSDEC=5:
Interval

Page Fault
Rate

Working Set
Limit

Adjustment

100

0
l

2
3
4
5

0
l

0
2

189
692

9
10
11
12

1147
1538
953
492

Deduct
Deduct
Deduct
Deduct
Deduct
None
None
Add 21
Add 21
Add 21
Add 21
None

5 pages
5 pages
5 pages
5 pages
5 pages

90
85
80

75
75
75
96
117

pages
pages
pages
pages

138
159

159

WSINC is a much larger value than WSDEC because quickly increasing the
working set limit may be critical to performance (for example, if a
new module is being referenced), while gradually lowering the limit
from the last increase tends to fine-tune upward adjustments.
The system manager can monitor system performance
parameter values in the following instances:

and

change

system

•

Initial limit -- If the system
typically
increases
or
decreases a working set by a significant amount immediately
after its process starts, the default working set limit for
affected
users
and
batch
streams
should be changed
accordingly. For example, if a user's working set limit is
100, but its processes typically show a working set limit of
50 to 60 for the first few minutes of processing, the user's
default working set value in the UAF should be decreased to
60.

•

Excessive paging I/O activity -- If paging appears to be
generating excessive I/O activity, the system manager might
decrease the value of PFRATH to make the working set limit
more
sensitive
to
paging
activity.
Additionally or
alternatively, the system manager might decrease the value of
PFRATL, increase the value of WSINC, or decrease the value of
WSDEC.

•

Excessive swapping I/O activity -- If swapping I/O operations
appear high, the system manager might increase the value of
PFRATH to make the working set less sensitive to paging
activity.
Working sets will tend to remain at smaller sizes,
hopefully increasing the number of working sets that will fit
in the balance set. Additionally or alternatively, the system
manager might increase the value of PFRATL, decrease the value
of WSINC, or increase the value of WSDEC.

The system manager should ensure that the values for WSDEC and WSINC
are prime in relation to one another. Otherwise, working set limits
will be restricted to a subset of integers. For example, if WSINC is
set to 27 and WSDEC is set to 6, working set limits will always be
multiples of 3 above and below the default working set limit.
11-11

TUNING CONSIDERATIONS

Monitoring aids for adjusting working set limits include:
•

VAX/VMS PROCESSES Display -- The size column
(which displays
shareable/total pages being used by each process) indicates
what portion of its working set each process requires.
The
state column pinpoints processes having scheduling problems:
a repeated state of PAGE FAULT WAIT, or PFW, may mean that the
working set is too small;
repeated states of COMPUTE (OUT OF
BALANCE SET}, or COMO, by several processes usually means that
there are too many users for the available resources. See
Section 14.13.

•

sizes
are
I/O SYSTEM RATES Display -- As working
set
decreased, the swapping rate should decrease. As working set
A
sizes are increased, the swapping rate should increase.
very high swapping rate usually means that there are too many
users for available resources. See Section 14.7.

•

SHOW STATUS DCL Command -- To obtain a per-image page fault
rate, execute the command SHOW STATUS before and after running
the image (preferably as a command procedure}.
Subtract the
before statistics from the after statistics to obtain the net
page faults and CPU time. Divide the page faults by the CPU
time to obtain the page fault rate during execution of the
image.

The system manager can inhibit automatic working set adjustments for a
process by setting the user's working set limit and default equal.
Setting the system parameter WSINC to a value of zero inhibits
automatic working set adjustments for all users.

11.4.2

Page Cache

The page cache, consisting of the free page list and modified page
list, recirculates pages pushed out of the working set when they are
needed again.
Frequently used pages tend to stay in physical
memory -- either in the working sets (unless outswapped) or in the
page cache -- while less frequently used pages tend to be deleted.
Size requirements for the free page list depend mainly on the amount
of code being pushed out of the working set.
Size requirements for
the modified page list depend mainly on the amount of data being
pushed out of the working sets. The cluster size for pages being read
from disk to the working sets also affects the efficiency of the page
cache.

11.4.2.1 Cache Sizes - The FREELIM system generation parameter sets
the minimum size of the free page list. The list may expand as space
permits. The FREELIM values listed in Table 12-7 are normally
adequate.
A high percentage of page faults resulting in read I/O
operations (I/O SYSTEM RATES display}
indicates that the value of
FREELIM might be increased, while a small percentage indicates that
the value of FREELIM is correct or might he smaller. A high swapping
rate
(I/O SYSTEM RATES display) indicates that the value of FREELIM
might be reduced to give more space to the balance set.
(However,
paging and swapping I/O might be better regulated by adjusting working
set limits; see Section 11.4.1.2 above.)

11-12

TUNING CONSIDERATIONS

The following system parameters regulate the size of the modified page
list:
•

MPW HILIM -- Largest size the modified page list is
to reach before pages are written to disk

•

MPW LOLIM -- Point below which writing of pages stops

•

MPW WRTCLUSTER -- Number of pages the system is
write in one I/O operation

permitted

The parameter values in Table 12-7 are normally adequate. The system
manager can monitor the effects of the the size of the modified page
list by watching the write I/O operation and swapping rates on the I/O
SYSTEM RATES display.
Increasing the upper and lower limits on the
modified page list normally lowers the write I/O operation rate, but
may increase the swapping rate.
The system manager can raise the
limits in small increments to determine
whether
the
changes
significantly
improve
the
write
I/O
operation rate without
significantly degrading the swapping rate.
The system manager should not assign arbitrarily high values to
FREELIM and MPW LOLIM, as these values reduce the fluid pages in
physical memory. -The system calculates fluid pages by subtracting the
resident system, the system working set, and the minimum size of the
page cache from available physical memory.
If the value of WSMAX as
specified by the system manager exceeds this value, it is decreased.

11.4.2.2 Cluster Size - The system generation parameter PFCDEFAULT
controls the number of pages being read from disk to the working sets
during one I/O operation. The values suggested in Table 12-7 are
normally adequate. As a rule of thumb, the system manager can specify
a value of 32 where the median working set limit is 100 pages, and a
value of 64 or more where the median working set limit exceeds 200
pages.
A potential problem with a large cluster size is that the pages being
transferred into a working set that has reached its limit push out an
equal number of pages, causing large turn-overs of pages in the
system. However, the actual cluster size during any one I/O operation
is limited by ·other factors, such as the size of the image or section
being transferred.
Only in a few instances do large cluster sizes
adversely affect system performance, but these instances are more than
made up for by the savings in disk I/O that the larger cluster sizes
promote.
For special cases, users can adjust cluster sizes on an individual
basis with the PFC parameter of the CLUSTER option in the linker (for
pages being read from image and section files, not the paging file).

11.4.3

Priority, Quantum, and Compute-Bound Swapping Rate

Certain workloads may give processes a disproportionate share of
service.
The problem can be alleviated somewhat by adjustments to
base priorities, the quantum, and the swapping rate for compute-bound
processes.

11-13

TUNING CONSIDERATIONS

11.4.3.l Very Large Working Sets - Running several working sets that
are very large for the amount of physical memory often causes serious
performance degradation. An example might be two batch streams using
(on the average) 350-page working sets in a 512KB system. Both batch
streams cannot fit in physical memory at the same time, requiring an
inswap and outswap of 350 pages each time one or the other is granted
control.\ The high swapping rate not only slows batch throughput, but
increases interactive response time, as these users are also held up
by the contention on the system disk.
When this situation exists, the system manager normally notices a high
swapping rate (I/O SYSTEM RATES display) and an abundance of processes
in the COMO state (VAX/VMS PROCESSES display).
The obvious solution to the above problem is to run one batch stream.
However, if two batch streams are essential
or perhaps the problem
revolves around two very large production programs that must run
concurrently -- the system manager can improve the situation as
follows:
•

Priority -- Decreasing the base priorities for the batch
streams from 4 to 3 can result in a better interactive
response time, but will slow batch throughput even more.

•

Quantum -- Increasing the quantum in this situation normally
ensures better service all around. The system manager may
find that a rather large quantum -- 1, 2, or even 3 seconds -is in order. The larger quantum means that the large working
sets will be swapped less frequently. While interactive users
must sometimes wait in blocks of several seconds to be
serviced, this may be an improvement over the previous
response time.

11.4.3.2 Compute-Bound Programs - Contention
among
compute-bound
processes can be alleviated by leaving the quantum at its default
setting and increasing the value of the SWPRATE system parameter -- to
as much as 3 to 5 seconds. The compute-bound swapping rate sets the
minimum interval (in real time, not processor time) between inswaps of
compute-bound processes.
(The system defines a compute-bound process
as one whose current priority is the same as the default base
priority.) A high SWPRATE value under such circumstances should reduce
swapping I/O on the system.
Real-time processes are not affected by the value of the SWPRATE
system parameter and interactive processes rarely are. They will be
inswapped and outswapped according to priority and quantum.
If a
compute-bound process must make way for an interactive process, it
will be outswapped even if the time specified by its SWPRATE value has
not passed (provided its quantum has passed).

11.4.4

System Requirements

The system manager can tune system requirements for time and space by
adjusting the nonpaged dynamic pool, system working set, and VAX-11
RMS buffers.

11-14

TUNING CONSIDERATIONS

11.4.4.1 Nonpaged Dynamic Pool - System executive code takes up 150
to 200 pages of physical memory. The nonpaged dynamic pool takes up
another 100 to 180 pages, depending a great deal on the size of the
system.
The NPAGEDYN system parameter determines the size of the
nonpaged dynamic pool. Chapter 12 explains how to calculate the
values of these parameters. Some space can be saved by shaving the
calculations closely, but this is not recommended. The system manager
should use reasonable figures in the calculations and, if anything,
err on the upper side.
During the day-to-day running of the system, the system manager should
monitor the NONPAGED POOL STATISTICS display. If a large amount of
unused space continually exists in the pool, the system manager can
reduce NPAGEDYN.
If very little unused space exists in the pool or
the pool is fragmented, the system manager should increase NPAGEDYN.
The pool should contain 200 to 300 bytes of unused space per user.
Running out of space in the nonpaged dynamic pool degrades service
more than using some extra space to ensure that this condition does
not occur. If the nonpaged dynamic pool is depleted, processes
needing space from the pool will be placed in a miscellaneous resource
wait (MWAIT) state until sufficient memory is returned to the pool by
other processes or by the completion of I/O services. If many
processes enter an MWAIT state, the system manager may be forced to
reboot the system.

11.4.4.2 System Working Set - The system also requires a working set
for the pageable portions of the executive, the paged dynamic pool,
VAX-11 RMS, and the resident portion of the system message file.
The
SYSMWCNT system parameter controls the upper limit (quota) of this
working set. Although the system manager usually would not reduce the
fluid pages available to the system, SYSMWCNT should be set at a
reasonable value. If programs start waiting on the completion of
system services, performance degradation is likely to be sharp •
. During run-time activities, the system automatically decreases the
limit if the page fault rate is low, depending on the values of the
PFRATL, PFRATH, WSINC, WSDEC, AWSMIN, AWSMAX, and AWSTIME system
parameters.
An optimal quota for SYSMWCNT varies depending on how the user
programs are u5ing the system.
Factors dictating a larger size
include high use of VAX-11 RMS
(especially ISAM services), a high
number of logical names, and extensive placement of record locks. As
a rule of thumb, the system manager can start with a value of 120. If
performance is poor, the system manager should immediately increase
SYSMWCNT. An upper limit of 180 or more may be required.
If
performance is good, the system manager can try lowering SYSMWCNT.
The system manager can monitor the system page fault rate by observing
the rate of system page faults on the PAGE MANAGEMENT STATISTICS
display. The system page fault rate should be kept low (one fault per
second or less).

11.4.4.3 VAX-11 RMS Buffers - Where VAX-11 RMS operations are mainly
random accesses of single small records, a process can decrease its
VAX-11 RMS multiblock count to 1 with the SET /RMS DEFAULT command.
This action reduces the size of the VAX-11 RMS ouffer required in
memory. However, the value of the RMS DEFAULT system parameter should
be left at 4.

11-15

TUNING CONSIDERATIONS

11.5

LARGE SYSTEMS

The major strategy in tuning a large system lies in trading more
memory for less I/O. On an individual level, programs should either
maintain larger data structures mapped into memory from section files
(in place of per-record processing by means of VAX-11 RMS), or
increase the multi-buffer
count
and
specify
read-ahead
and
write-behind processing for VAX-11 RMS operations. The system manager
should install as many programs as possible with the shared and
header-resident attributes.
On a system-wide basis, the system
manager should allot as much memory as practical to the Files-11 ACP
data
structures,
primarily
the
header, directory, map, file
identification, extent, and quota caches. The sections that follow
provide guidelines for setting values for the ACP system parameters
(see also Chapter 12).
The system parameters, UAF entries, and DCL commands affecting working
set size, the page cache, priority, the quantum, the compute swapping
rate, the nonpaged dynamic pool, the system working set, and the
VAX-11 RMS buffers should, for the most part, be left at their
standard values. Any changes should favor increasing the amount of
physical
memory
available
to system and user processes.
In
particular, the system manager should provide working set default and
quota values, and PFRATL, PFRATH, WSINC, WSDEC, AWSMIN, AWSMAX, and
AWSTIME system parameter values that minimize paging I/O, although
limit and quota values should not be set arbitrarily or unreasonably
high. On a large system, the system manager should be able to
minimize paging I/O without increasing swapping I/O.
(If swapping I/O
does increase, however, the system manager should adjust working set
limits as described in Section 11.4.1)

11.5.1

Header Cache

Because a file header, once accessed,
is likely to require more
accesses as the file continues to be processed, maintaining header
blocks (one block per page) in memory can save a significant amount of
I/O. The following system parameters control the header cache:
•

ACP WRITEBACK -- This switch should be set to a value of 1 to
enable the deferred writing of file header blocks, which
inhibits the write I/O operation for the file header that
otherwise occurs each time a file is extended. The header is
always written when the file is closed. Headers for relative
and indexed files are always written when the file is
extended. A risk is taken in that the file may be lost if a
header has never been written and the sytem fails. This loss
is usually not serious, since the file is probably in an
inconsistent state.

•

ACP HDRCACHE -- The number of header blocks should equal or
exceed the number of files that are likely to be open
concurrently. The system manager can calculate this value as
the product of the number of concurrent users and the median
number of files each user accesses concurrently.
Reasonable
values for a system configured for 48 to 64 users (64USER
system) fall in the range of 80 to 200.

Even on a small system, the system manager should try to allot extra
pages to the header cache, as the expenditure in memory is well worth
the I/O operations saved.

11-16

TUNING CONSIDERATIONS
11.5.2

Directory Cache

Because a directory block, once accessed, is likely to require more
accesses as files continue to be processed, maintaining directory
blocks (one block per page) in memory can save a significant amount of
I/O.
The ACP DIRCACHE system parameter specifies the size of the directory
cache.
As with ACP HDRCACHE, the system manager should set a value
that equals or exceeds the number of directories that are likely to be
accessed concurrently.
The size of the directory cache stands next to the size of the
cache in importance in reducing Files-11 ACP I/O activity.

11.5.3

Quota Cache

If disk quotas are being enforced, the system manager
quota entry per active user.
The ACP QUOCACHE
specifies the number of entries cached.
One page
entries,
so that, for example, a specification of 64
of storage.

11.5.4

header

should cache one
system parameter
holds 16 quota
requires 4 pages

Map, File Identification, and Extent Caches

The system manager should cache a reasonable number of bit map blocks,
file
identification slots, and extent entries.
Suggested values on a
system configured for 48 to 64 users (64USER system) are as follows:
•

ACP MAPCACHE

20 pages

•

ACP FIDCACHE

8 to 16 entries

•

ACP EXTCACHE

16 to 32 entries

ACP EXTLIMIT can be set to a fairly high value (for example,
200
to
400T,
since free space lost during a system failure is automatically
recovered at volume mount time.

11.5.5

System Directory Cache

The ACP SYSACC parameter specifies the number of directories for which
to save access data on a volume mounted with the /SYSTEM qualifier.
Its value can be overridden with the /ACCESS qualifier
in the DCL
command MOUNT.
It should be set to the number of directories expected
to be in active use on a
system volume at one time.
Each unit
requires 96 bytes of nonpaged pool.
However, too low an ACP SYSACCESS
or /ACCESS value partially negates the benefit of the ACP DIRCACHE
parameter.

11.5.6

Multiple ACPs

Multiple ACPs (that is, duplicate Files-11 ACPs)
are almost never
worthwhile.
The memory spent on the extra ACP would be better spent
on increasing the cache sizes of the first ACP.
The value of
ACP MULTIPLE should be 0 unless all the following conditions exist:
physical memory not less than 2MB;
disks attached on separate
controllers;
high direct I/O activity even with large ACP caches.
11-17

CHAPTER 12
SYSTEM PARAMETERS

The system parameters configure a system in terms of resource
requirements and limits. Values for the parameters exist in several
forms:
•

Current
values -- The
system
(SYS$SYSTEM:SYS.EXE)
contains a set
values.

image
on
disk
of system parameter

•

Active values -- The active system in memory contains a set of
parameter values.

•

Parameter files -- System parameter values can
specially structured files.

stored

The VAX/VMS software distribution kit contains seven standard system
parameter
files,
each
geared
toward
a
particular hardware
configuration:

File Name

Hardware Configuration

MINIMUM

512KB memory
Console terminal
Console floppy diskettes
Bootstrap disk

BUSER

512KB memory
2 RK07 disks
8 DZll lines

16USER

768KB memory
2 RM03 disks
16 DZll lines

32USER

1024-1536KB memory
2 RP06 disks
32 DZll lines

48USER

1536-2048KB memory
2-4 RP06 disks
48 DZ 11 lines

64USER

2048-3072KB memory
2-4 RP06 disks
64 DZll lines

VIRT32MB

1024KB memory
2 RP06 disks
16 DZll lines
Support of 32MB virtual address space

12-1

SYSTEM PARAMETERS

The file specification for each file is SYS$SYSTEM:file-name.PAR.
During the bootstrap process, the operator selects either the current
values
(that is, the values in SYS$SYSTEM:SYS.EXE) or the values in
one of the parameter files, and optionally modifies them.
The
selected values, as modified, become the active values which actually
configure the system, and also become the new current values.
During system operation, the system manager can run the SYSGEN utility
(Chapter 13) to modify the current values or the values of any
parameter file, or create a new parameter file, for use in subsequent
bootstrap operations.
The SYSGEN utility also enables the system
manager to dynamically alter the run-time system configuration by
modifying a subset of the active values.
The system parameters fall into nine general categories:
•

MAJOR -- Parameters most likely to
subset of SYS

require

•

SYS -- Parameters that affect
includes the major parameters

JOB

•

ACP
Parameters associated with Files-11
processes (ACPs)

RMS

•

Parameters associated with process creation limits
PQL
quotas

•

GEN
Parameters that affect the creation and initialization
of data structures at bootstrap time; a subset of SYS, plus
all ACP and PQL parameters

SPECIAL

•

DYNAMIC
Parameters whose active values can be modified;
a
subset of SYS and ACP, plus all JOB, RMS, and PQL parameters

overall

modification;
system

operation;

Job control parameters
ancillary

control

Parameters associated with VAX-11 RMS
and

Special parameters used by DIGITAL

Each parameter has associated with it default, minimum, and maximum
values that define the scope of allowable values. Tables 12-1 through
12-6 briefly describe the parameters and list the default, minimum,
and maximum values.
(The SPECIAL parameters are not documented; they
should be used only by DIGITAL personnel.) Table 12-7 lists the values
of the parameters in the standard system parameter files. The
remaining sections of the chapter describe the parameters in more
detail and provide guidelines for their modification.

12-2

SYSTEM PARAMETERS
Table 12-1
MAJOR Parameters

Parameter Name

Description

Default

Minimum

Maximum3

PFCDEFAULT 2

Default page fault cluster size
in pages

lli

127

GBLSECTIONSl

Number of global section
descriptors

-1

GBLPAGES 1

Number of global page table
entries

2048

512

-1

MAXPROCESSCNTl

Maximum number of processes

251i

SYSMWCNT 1

Maximum size of system working
set in pages

100

lfi384

BALSETCNTl

Maximum number of resident working sets

1024

IRPCOUNT 1

Number of preallocated I/O
request packets

250

327fi8

WSMAXl

Maximum size of any working set
in pages

2511

H384

NPAGEDYN 1

Size of nonpaged dynamic pool in
bytes (but rounded down to an
integral number of pages by the
system)

li4000

lli384

-1

PAGEDYN 1

Size of paged dynamic pool in
bytes (but rounded down to an
integral number of pages by the
system)

24000

819?.

-1

VI RTUALPAG ECNT l

Maximum virtual space per
process in pages

8192

51?.

262144

QUANTUM 2

Maximum time a process can use at
once and minimum service a process must receive before being
outswapped, in lOms units

327f..7

MPW WRTCLUSTERl

Number of pages written per I/O
from the modified page list

127

MPW HILIM

Maximum size of modified page list
in pages

H384

MPW LOLIM 1

Minimum size of modified page list
in pages

H384

1.
2.
3.

Also a GEN parameter
Also a DYNAMIC parameter
The value -1 means no limit.

12-3

SYSTEM PARAMETERS

Table 12-2
SYS Parameters
(excluding MAJOR parameters)
Parameter Name

Description

Minimum3

Default

Maximum3

KFILSTCNT 1

Number of known file list heads

-1

PROCSECTCNTl

Number of process sections

1024

MINWSCNTl

Minimum number of fluid pages in
any working set

-1

INTSTKPAG ES l

Size of interrupt stack in pages

SPTREQ 1

Number of additional system
page table entries

PFRATL 2

-1

700

-1

Page fault rate low limit in faults
per 10 seconds of processor time

-1

PFRATH2

Page fault rate high limit in
faults per 10 seconds of processor
time

512

-1

WSINC 2

Working set increment in pages

-1

WSDEC 2

Working set decrement in pages

-1

AWSMIN 2

Automatic working set minimum in
pages

-1

AWSMAX 3

Automatic workinq set maximum in
pages

41i0

-1

AWSTIME2

Automatic working set time for
collecting sample in !Oms units

-1

EXTRACPU2

Extra CPU time added to process
after CPU time is expired in !Oms
units

1000

-1

MAXSYSGROUP2

Highest system UIC

32768

MAXBUF 2

Maximum number of bytes that can
be transferred in one buffered I/O

1024

512

li4000

DEFMBXBUFQuo2

Default mailbox buffer quota in
bytes

1024

256

li4000

DEFMBXMXMSG2

Default mailbox maximum message
size in bytes

/.511

li4

64000

DEFMBXNUMMSG2

Not implemented

Iii

-1

FREELIM 1

Lower limit of free paqe list in
bytes

-1

XFMAXRATE2

Maximum rate of transfer for DR32
devices

2311

255

LAMAPREGS

Number of map registers allocated to
an LPAll device driver

255
---

(continued on next page)
1.
2.
3.

Also a GEN parameter
Also a DYNAMIC parameter
The value -1 means no limit.

12-4

SYSTEM PARAMETERS
Table 12-2 (Cont.}
SYS Parameters
(excluding MAJOR parameters}

Minimum3

Maximum3

Parameter Name

Description

REALTIME SPTS

Number of system page table entries
reserved for connect-to-interrupt
processes

-1

CLISYMTBLl 2

Size of command interpreter symbol
table in bytes

128

BUGREBOOT 2

Automatic reboot on fatal bugcheck;
switch

CRDENABLEl

Detection and logging of memorycorrected read errors; switch

DUMPBUG 2

Writing of dump file on fatal
bugcheck; switch

BUGCHECKFATAL2

All bugchecks fatal; switch

POOLPAGING 1

Paging of paged dynamic pool;
switch

SBIERRENABLEl

Internal bus error logging; switch

SETTIME 1

Time-of-day prompt at boot time;
switch

SYSPAGING 1

Paging of executive code; switch

UAFALTERNATEl

Use of alternate UAF; switch

TTYSCANDELTA 1

Terminal dial-up/hang-up scan
interval in increments of lOOns

10000000

100000

-1

TTY SPEEDl

Default speed for terminal; code

TTY BUF'l

Default line width for terminal

65535

TTY DEFCHARl

Default terminal characteristics

40370n528

-1

TTY TYPAHDszl

Size of terminal type-ahead
buffer in bytes

-1

TTY PROT 1

Default protection for all terminals

-1

TTY OWNERl

Owner UIC for TTY PROT specification

-1

1.
2.
3.

Also a GEN parameter
Also a DYNAMIC parameter
The value -1 means no limit.

12-5

Default

SYSTEM PARAMETERS
Table 12-3
JOB Parameters

Parameter Name
---------- -- -

----~-------

Description

Default

Minimum

Maximum

---------

------~-

MAXPRI N'l'SY MB

Maximum number of print
symbionts

DEFPRI

Default priority

IJOBLIM

Maximum number of interactive
jobs

fi4

1024

BJOBLIM

Maximum number of batch jobs

1024

NJOBLIM

Maximum number of network jobs

1024

Default

Minimum

Maximum3

/.55

Table 12-4
ACP Parameters

Parameter Name

Description

ACP MULTIPLE!

Multiple ACPs per volume; switch

ACP SHARE

Sharing of ACP code; switch

ACP MAPCACHE2

Size of bit map cache in pages

-1

ACP HDRCACHE2

Size of file header cache in
pages

-1

ACP DIRCACHE2

Size of directory cache in pages

111

-1

ACP WORKSET 2

Working set default for ACP

-1

ACP FIDCACHEl

Size of file identification
cache in pages

-1

ACP EXTCACHEl

Size of extent cache in pages

-1

ACP EXTLIMIT 1

Maximum amount of free space in
extent cache in tenths of a percent
of available free space

1000

ACP_QUOCACHEl

Number of entries in quota cache

-1

ACP SYSAccl

Size of directory access cache
in pages

-1

ACP MAXREAD2

Maximum directory blocks to read

ACP WINDow2

Default number of window pointers

-1

ACP WRITEBACK 2

Caching of file headers; switch

ACP DATACHECK2

Data verification on ACP I/O
operations; 3-way switch

ACP BASEPRI02

Base priority for ACP processes

ACP SWAPFLGS2

Swapping of ACP working sets;
code

1.
2.
3.

100

Also a DYNAMIC parameter, but effective only when a volume is mounted.
Also a DYNAMIC parameter, but effective only when an ACP is active.
The value -1 means no limit.

12-6

SYSTEM PARAMETERS
Table 12-5
RMS Parameters
.-----

Parameter Name

Description

Default

Minimum

Maximum

RMS DFMBC

Default multiblock count

127

RMS DFMBFSDK

Default multiblock count for
sequential disk _operations

127

RMS DFMBFSMT

Default multiblock count for
magnetic tape operations

127

RMS DFMBFSUR

Default multiblock count for
unit record operations

127

RMS DFMBFREL

Default multiblock count for
relative disk operations

127

RMS DFMBFIDX

Default multiblock oount for
indexed sequential disk
operations

127

RMS DFMBFHSH

Not implemented

127

-----------

Table 12-6
PQL Parameters
- ----·--

Parameter Name

Description
------·-

Default
- - - - - -·--- -

------

Minimuml

Maximuml

PQL_DASTLM

Default number of pending AS Ts

-1

PQL_MASTLM

Minimum number of pending AS Ts

-1

PQL_DBIOLM

Default buffered I/O limit

-1

PQL_MBIOLM

Minimum buffered I/O limit

-1

PQL_DBYTLM

Default buffered I/O byte limit

8192

-1

PQL_MBYTLM

Minimum buffered I/O byte limit

1024

-1

PQL_DCPULM

Default CPU time limit in
increments of lOms

-1

PQL_MCPULM

Minimum CPU time limit in
increments of lOms

-1

PQL_DDIOLM

Default direct I/O limit

-1

PQL_MDIOLM

Minimum direct I/O limit

-1

PQL_DFILLM

Default open file limit

-1

PQL_MFILLM

Minimum open file limit

-1

PQL_DPGFLQUOTA

Default paging file quota

2048

-1

PQL_MPGFLQUOTA

Minimum paging file quota

25fi

-1

PQL_DPRCLM

Default subprocess limit

-1

PQL_MPRCLM

Minimum subprocess limit

-1

PQL_DTQELM

Default time queue entries

-1

PQL_MTQELM

Minimum timer queue entries

-1

PQL_DWSDEFAULT

Default working set size

100

-1

PQL_MWSDEFAULT

Minimum default working set size

-1

PQL_DWSQUOTA

Default working set quota

120

-1

PQL_MWSQUOTA

Minimum working set quota

-1

The value -1 means no limit.

12-7

SYSTEM PARAMETERS
Table 12-7
Standard System Parameter Files
Parameter Name

Parameter File Name
---------~--

PFCDEFAULT
GBLSECTIONS
GBLPAGES
MAXPROCESSCNT
SYSMWCNT
BALSETCNT
IRPCOUNT
WSMAX
NPAGEDYN
PAGEDYN
VI RTUALPAG ECNT
QUANTUM
MPW WRTCLUSTER
MPW-HILIM
MPW-LOLIM
SPTREQ
WSINC
FREELIM
BUG CH ECK FATAL
TTY DEFCHAR
MAXPRINTSYMB
DEFPRI
IJOBLIM
BJOBLIM
NJOBLIM
ACP SHARE
ACP-MAPCACHE
ACP-HDRCACHE
ACP-DIRCACHE
ACP-FIDCACH
ACP-EXTCACHE
ACP-EXTLIMIT
ACP-QUOCACHE
ACP-SWAPFLGS

MINIMUM

BUSER

16
20
1024
12
80
5
0
256
32128
16384
409fi
60
8
10
4
700
0
10
0
268440224
1
4
4
1
16
0
1
4
4
8
8
200
4
15

16
30
2048
20
100
12
200
25fi
51200
24576
4096
fiO
16
24
12
650
21
10
0
2fi8440224
1
4
9
1
16
0
2
6
8
8
8
200
8
15

ll'iUSER
32
32
2048
28
120
20
275
25fi
74240
28672
40%
fiO
32
44
H

700
21
10
0
268440224
1
4
17
1

48USER

fi4
48
3072
48
140
34
530
512
131072
40960
8192
fiO
64
92
32
700
21
lfi
0
268440224
4
4
33
4

127
80
3072
fi8
160
52
fi50
700
159744
49152
8192
60
127
220
100
700
21
lfi
0
268440224
8
4
49

li'i
l

lfi
1
20
80
100

lfi

0
4
12
16
8
lfi
.200
lfi
15

12
40
50
12
16
200
32
14

Note: All other SYS and ACP parameters (not listed in this
take defaults. All RMS and PQL parameters take defaults.

12.1

102~

179.712
fi5'i36
8192
fiO
127
320
200
700
21
lfi
0
2fi8440224
8
4
fi5
8
16

32
200
48
14

127
32
2048
28
120
8
300
1024
921110
327fi8
1;553(-)
fiO
127
220

127
80
3072
84
180
fi8
750

9fi

700
21
10
0
268440224
2
4
17
4
lfi

1
4
12
16

1
20
120
150
16
32
200
64
14

16
200
12
15

table)

MAJOR PARAMETERS

The MAJOR parameters, a subset of the SYS
require adjustment.

12.1.1

VIRT32MB

fi4USER

32USER

parameters,

most

commonly

PFCDEFAULT Parameter

During execution of programs, PFCDEFAULT controls the number of image
pages read from disk per I/O operation when a page fault occurs. The
read I/O operations may take place from an image file or from the
paging file.
The actual size of the cluster may be less than
PFCDEFAULT depending on the size of image sections and the pattern of
page references.
The value of this parameter should not be less than 16 to ensure
adequate I/O performance, nor greater than one-fourth the default size
of the average working set to prevent a single I/O operation from
displacing a major portion of a working set. The average working set
default size can be estimated as the median WSDEFAULT value in the UAF
(see Chapter 2).
The values in the standard parameter files increase with the size of
the system. System managers of 32USER and larger systems especially,
should ensure that the PFCDEFAULT value is not too high for the
average working set size.
PFCDEFAULT can be overridden on an
CLUSTER option of the VAX-11 Linker.

12-8

image-by-image

basis

with

the

SYSTEM PARAMETERS
12.1.2

GBLSECTIONS Parameter

GBLSECTIONS sets the number of global section descriptors allocated in
the system header at bootstrap time. Each global section requires one
descriptor. Each descriptor takes 32 bytes of permanently resident
memory.
The values found in the standard system parameter files are more than
adequate for the images normally installed as shared in the system
start-up command procedures
(see Chapters 7 and 8).
The system
manager may, once the system is running and all global sections are
created, examine the actual requirements with the /GLOBAL option of
the INSTALL utility
(Chapter 7) and reduce the value of GBLSECTIONS
accordingly. However, the value of this parameter should not be cut
too closely -- each descriptor requires only 32 bytes. If the system
manager plans to install many user images as shared or if user
programs are likely to create many global sections, the value of this
parameter must be increased.

12.1.3

GBLPAGES Parameter

GBLPAGES sets the number of global page table entries allocated at
bootstrap time.
Each global section requires one global page table
entry per section page, plus two entries, with the total rounded up to
an even number.
Every 128 global page table entries add 4 bytes to
permanently resident memory in the form of a system page table entry.
The values found in the standard system parameter files are more than
adequate for the images normally installed as shared (see Chapters 7
and 8) in the system start-up command procedures. The system manager
may, once the system is running and all global sections are created,
examine the actual requirements with the /GLOBAL option of the INSTALL
utility (Chapter 7) and reduce the value of GBLPAGES accordingly.
However, the value of this parameter should not be cut too closely
the page table entries are not expensive in terms of permanently
resident memory. If the system manager plans to install many user
images as shared or if user programs are likely to create many global
sections, the value of this parameter must be increased.

12.1.4

MAXPROCESSCNT Parameter

MAXPROCESSCNT sets the number of process entry slots allocated at
bootstrap time.
One slot is required for each concurrent process on
the system. Each slot requires 6 bytes of permanently resident
memory.
Total slot requirements can be figured as follows: 1 per interactive
user; 1 per batch job; 1 per ACP (including the network ACP); l per
print symbiont; l for the input symbiont (where used);
and 3 for
ERRFMT, OPCOM, and JOB CONTROL.
The values in the standard system
parameter files cover these requirements. The system manager should
reduce the value only where interactive use of the system is much
lower than anticipated by the standard system parameter file
for
example, if only 20 users are on a 32USER system.

12-9

SYSTEM PARAMETERS
The system automatically reduces the active value of MAXPROCESSCNT if
that value times the value of WSMAX exceeds the size of the swapping
file in pages. For this reason, the system manager should ensure
reasonable values for MAXPROCESSCNT and WSMAX, and an adequate number
of pages for the swapping file.
(See Chapter 13 for guidelines on
creating the swapping file.) The system automatically increases the
active value of MAXPROCESSCNT when a larger, secondary swapping file
is installed.

12.1.5

SYSMWCNT Parameter

SYSMWCNT sets the quota for the size of the syBtem working set, which
contains the pageable portions of the system, the paged dynamic pool,
VAX-11 RMS, and the resident portion of the system message file.
While a high value takes space away from user working sets, a low
value may seriously impair system performance. Appropriate values
vary deperiding on the level of system use. The system manager should
-- when the system is running at full load -- check the rate of system
faults with the PAGE command in the DISPLAY utility (see Chapter 14).
If the system page fault rate is high, and especially if the system
seems to be slow, the value of SYSMWCNT should be increased.

12.1.6

BALSETCNT Parameter

BALSETCNT sets the number of balance set slots in the system page
table.
Each memory-resident working set requires one balance set
slot. Each balance set slot requires 4 bytes of permanently resident
memory per 128 virtual pages (as specified in the VIRTUALPAGECNT
parameter; see Section 12.1.11). The requirements for the standard
system parameter files are as follows:
MINIMUM
VIRTUALPAGECNT
Bytes per Slot
BALSETCNT
Total Bytes

BUSER

16USER

32USER

48USER

64USER

VIRT32MB

4096
128
12
1536

4096
128
20
2560

8192
256
34
8704

8192
256
52
13312

8192
256
68
17408

65536
2048
8
16384

2048
64
5
320

The system manager can monitor the active system with the DCL command
SHOW SYSTEM or with the USERS command of the DISPLAY utility to
ascertain the actual maximum number of working sets in memory.
If
this number is significantly lower than the value of BALSETCNT, this
parameter value may be lowered. If all balance set slots are being
used, the value of BALSETCNT should be raised.
Lowering the value of BALSETCNT has the effect of increasing the size
of the free page list. The system manager might take this tack if the
page fault rate is excessive. However, the swapping rate is apt to
increase and must be monitored closely.

12.1.7

IRPCOUNT Parameter

IRPCOUNT sets the number of preallocated I/O request packets.
Each
packet requires 96 bytes of permanently resident memory from the
nonpaged dynamic pool. If the amount of space required by this number
exceeds half the space allocated to the nonpaged dynamic pool (see the
NPAGEDYN parameter below), the value of IRPCOUNT is automatically

12-10

SYSTEM PARAMETERS

reduced to half that space. The value of IRPCOUNT is automatically
reduced to 0 for systems configured with the MINIMUM parameter file.
Preallocated I/O request packets (which are also used to fill requests
for nonpaged dynamic memory where the amount of memory requested is
equal to or less than the size of a packet)
reduce system overhead.
When all preallocated packets are in use, I/O packets must be
allocated dynamically from the nonpaged dynamic pool.

12.1.8

WSMAX Parameter

WSMAX sets the maximum number of pages for any one working set.
The
value of WSMAX also affects the allocation of permanently resident
memory for the swapper map and system page table (and for this reason
should not be set at an arbitrarily high figure):
•

Swapper map -- 4 bytes for each page of WSMAX

•

System page table -- 4 bytes for
times BALSETCNT

each

Memory requirements for the standard system
follows:

WSMAX
Swapper Map Bytes
BALSETCNT
Page Table Bytes

128

pages

parameter

WSMAX,

files

are

MINIMUM

BUSER

16USER

32USER

48USER

64USER

VIRT32MB

256
1024
5
40

256
1024

512
2048

1024
4096

160

544

700
2800
52
1096

1024
4096

256
1024
20

68
2176

256

WSMAX also affects the size of the swapping file in that the file must
be large enough to hold WSMAX times MAXPROCESSCNT pages. If the
swapping file is not large enough when the system is initialized, the
value of MAXPROCESSCNT is reduced until WSMAX times MAXPROCESSCNT
pages will fit.
(The number of process slots allocated, however,
is
still based on the original value of MAXPROCESSCNT.) In performing
these calculations, the system determines the space available for the
balance set
(number of pages in the system reduced by the number of
nonpageable pages, and the values of SYSMWCNT and FREELIM) and reduces
WSMAX to this figure if WSMAX exceeds the figure.
Generally, the system manager should use a reasonable value for WSMAX;
that is, whatever the size of the largest working set will be. On the
smaller configurations, the system manager may raise the standard
values
by
as
much as a few hundred pages.
On the larqer
configurations, the system manager may lower the standard values as
much as desired.

12.1.9

NPAGEDYN Parameter

NPAGEDYN sets the size of the nonpaged dynamic pool in
figure is rounded down to an integral number of pages.

bytes.

This

Probably the simplest and best approach to take in setting a value for
this parameter is to initially use the value found in the appropriate
standard system parameter file,
then monitor the amount of space
actually used with the POOL command of the DISPLAY utility. The value
may be decreased if a lot of space is being wasted and should be
increased if little space is unused or if all I/O packets are used.

12-11

SYSTEM PARAMETERS
Insufficient space for the nonpaged dynamic pool
rapidly degrades
system performance,
possibly to the point that the system does not
function.
si~e

To approximate the
following:

of the nonpaged dynamic pool,

first

add

the

* 518

•

MAXPROCESSCNT

WSMAX

* 4

•

IRPCNT

* 96

BALSETCNT

•

MPW WRTCLUSTER

•

KFILSTCNT

•

Size of paging file

•

Size of network device (DMCll)

* 4
* 6

* 68
(in pages) divided by 8
receive buffers

All totals are in bytes.
Next add the space required for
loadable
drivers being used and their data base structures.
The sizes of the
standard drivers and associated data bases are as follows:

Driver
Name

Driver
Size

Controller
Structures

Unit
Structures

CRDRIVER

1328

288

128

DXDRIVER

1008

288

320

DBDRIVER

1776

288

160

DLDRIVER

2128

288

192

DMDRIVER

2288

288

208

DRDRIVER

2080

288

176

DYDRIVER

1776

288

192

LADRIVER

2688

288

400

LPDRIVER

1008

288

128

MBXDRIVER

1776

288

144

TS DRIVER

3232

288

176

TMDRIVER

4256

384

160

NETDRIVER

592

288

144

RTTDRIVER

1136

288

128

DZDRIVER

6544

288

192

XFDRIVER

1504

288

144

XMDRIVER

3232

288

240

12-12

SYSTEM PARAMETERS
All figures are in bytes. Calculate a total for each driver by first
multiplying the next-to-the-last column by the number of controllers
and the last column by the number of units, then totaling all columns.
A DZDRIVER supporting 16 terminals on 2 controllers, for example,
requires 6544 + (288 x 2) + (192 x 16) = 10,192 bytes.
An LPDRIVER
supporting 2 printers requires 1008 + (288 x 2) + (128 x 2) = 1840
bytes (printers require a controller for each unit). In most cases,
the controller structures mean the device data block (DDB), channel
request block (CRB), and interrupt data block (IDB), which require one
I/O request packet each (96 bytes each for a total of 288 bytes). The
controller structures for TMDRIVER, however, require 4 I/O request
packets. The unit structures refer to the unit control blocks (UCBs),
which vary in size.
No matter how many units are in the configuration, only
each driver is required.

12.1.10

one

copy

PAGEDYN Parameter

PAGEDYN sets the size of the paged dynamic pool in bytes.
The
specified value is rounded down to an integral number of pages. Each
512 bytes of paged dynamic pool adds 4 bytes of permanently resident
memory to the system page table;
the paged dynamic pool has no other
direct memory requirements.
The paged dynamic pool is used to allocate storage for system and
group logical names, resident image headers, known file list entries,
and VAX-11 RMS file sharing structures. Substantial amounts of space
for the pool can be overallocated with little effect on system
performance.

12.1.11

VIRTUALPAGECNT Parameter

VIRTUALPAGECNT sets the maximum number of virtual pages that can be
mapped for any one process. Every 128 virtual pages requires 4 bytes
of permanently resident memory in the system page table (as discussed
under the BALSETCNT parameter, Section 12.1.6), 1 page table page
(which is pageable), and 8 bytes in the process header (which is not
pageable, but is swapped). The total number of virtual pages may be
divided between the PO and Pl tables in any proportion except that the
Pl table must be large enough to map 320 pages.
The standard system parameter files provide virtual address spaces of
1 million
(MINIMUM), 2 million
(8USER, and 16USER), and 4 million
(32USER, 48USER, and 64USER) bytes, which should be sufficient for
normal use. The VIRT32MB file provides an address space of 32 million
bytes for a limited number of programs. The system manager should use
the values in the standard parameter system files unless virtual-full
errors occur.
When the System Dump Analyzer is used, the system manager must insure
that the value of VIRTUALPAGECNT is at least the size of the dump file
plus 1024 pages.

12-13

SYSTEM PARAMETERS

12.1.12

QUANTUM Parameter

QUANTUM serves two purposes:
•

Processor time -- Maximum amount of processor time a process
can receive before control passes to another process of equal
priority that is ready to compute

•

Balance set residency -- Minimum amount of service
(includes
processor
time and time waiting for completion of I/O
services) a compute-state process must receive before being
outswapped to secondary storage

The standard system parameter value of 30 (300 milliseconds) should be
adequate for most applications. On a system where the swapping rate
is high (as shown by the !ORATES command of the DISPLAY utility), a
high value
even up to 300
(3 seconds) -- may improve overall
performance by reducing the disk I/O to the swapping file.

12.1.13

Modified Page Writer Parameters

The Modified Page Writer
MPW_HILIM, and MPW LOLIM.

parameters

consist

MPW_WRTCLUSTER,

MPW WRTCLUSTER sets the number of pages to be written from the
modified page list during one I/O operation to the paging file or a
section file.
(The actual size of the cluster may be limited by the
number of pages available for the I/O operation.) Each page in the
cluster requires o bytes of permanently resident memory.
MPW HILIM sets an upper limit for the modified page list.
When the
list accumulates the number of pages specified by this limit, writing
of the list commences.
(The pages that are written are then
·transferred to the free page list.)
MPW LOLIM sets a lower limit for the modified page list. When writing
of the list causes the number of pages on the list to drop to or below
this limit, writing stops.
The system manager may want to adjust the standard system parameter
files
so that the lower limit is 12 to ln pages above the
MPW WRTCLUSTER value. This action ensures that the modified page list
is not depleted by a write I/O operation. Increasing the value of the
lower limit tends to reduce the write I/O operations from the modified
page list, but may increase the swapping rate.
Both write I/O
operationss and the swapping rate can be monitored with the !ORATES
command of the DISPLAY utility.

12.2

SYS PARAMETERS

The remaining (nonmajor) SYS parameters should be checked
reasonableness, as described in the sections that follow.

12-14

for

SYSTEM PARAMETERS
12.2.1

KFILSTCNT Parameter

KFILSCNT sets the number of known file list heads. Each active list
requires 68 bytes of permanently resident memory in the nonpaged
dynamic pool. Each extra, never-used list head requires 4 bytes. The
actual known file entries are allocated from the paged dynamic pool.
One known file head is required for each set of installed images with
a different combination of device name, directory name, and file type.
(See Chapter 7.)

12.2.2

PROCSECTCNT Parameter

PROCSECTCNT sets the number of section descriptors that a process can
contain.
Each section descriptor increases the fixed portion of the
process header by 32 bytes.
The system manager should set a value greater than the maximum number
of image sections in any section to be run, as indicated by the
linkage memory allocation map for the image.

~2.2.3

MINWSCNT Parameter

MINWSCNT sets the minimum number of fluid pages -- that is, pages not
locked in the working set -- required for the execution of a process.
This value plus the size of the process header establishes the minimum
working set size.
The value of MINWSCNT must provide sufficient space to execute any
VAX-11
instruction.
Theoretically,
the worst-case instruction
requires 52 pages; however, all VAX/VMS code can run with 20 fluid
pages.
An insufficient value may inhibit system performance or even
put a process into an infinite loop on some instructions.

12.2.4

INTSTKPAGES Parameter

INTSTKPAGES sets the size of the interrupt stack in pages. Each page
on the interrupt stack requires a page of permanently resident memory.
The default value of 2 should be used unless interrupt-stack-not-valid
exceptions occur.
These may be caused by either an unusually large
number of devices or a driver that requires a very large amount of
stack space.

12.2.5

SPTREQ Parameter

SPTREQ sets the number of system page table (SPT) entries required for
mapping the following components:

Component

SPT Entries
185
126
252
variable
7
24
4

Executive image
VAX-11 RMS image
SYSMSG.MPF file
Multiport memory structures
Each MASSBUSS adapter
Each UNIBUS adapter
Each DR32 adapter
12-15

SYSTEM PARAMETERS
The number of system page table entries required for all other
purposes is automatically computed and added to the value of SPTREQ to
yield the actual size of the system page table.

12.2.6

PFRATL Parameter

PFRATL specifies the page fault rate below which the limit of a
working set will be automatically decreased. The unit of measure is
faults per 10 seconds of processor time. At the default setting, for
example, the system will automatically decrease the limit of a working
set if it is faulting less than 2 pages every 10 seconds.
Increasing the value of this parameter tends to decrease the limits of
the working sets, while decreasing its value tends to increase their
limits.

12.2.7

PFRATH Parameter

PFRATH specifies the page fault rate above which the limit of a
working set will be automatically increased. The unit of measure is
faults per 10 seconds of processor time. At the default setting, for
example, the system will automatically increase the limit of a working
set if it is faulting more than 512 pages per 10 seconds.
Decreasing the value of this parameter tends to increase the limits of
the working sets, while increasing its value tends to decrease their
limits.

12.2.8

WSINC Parameter

WSINC specifies the number of pages by which the limit of a working
set is automatically increased at each adjustment interval (which is
quantum end). At the default setting, for example, the system will
increase the limit of a working set by 21 pages each time an increase
is required.
Decreasing the value of this parameter tends to reduce the speed with
which
working set limits are increased when the need arises.
Normally, this parameter should be kept at a high value because a
rapid increase in limit is often critical to performance.
(The need
for an increase in limit typically occurs when new images or sections
are being introduced to the working set, so that the working set must
suddenly accommodate many more pages than before.) In addition, the
value of this parameter should be prime relative to WSDEC so that
working set limits will vary as they are increased and decreased.
A value of 0 for WSINC disables the automatic adjustment of working
set limits for all processes. Limits stay at their base values. The
automatic adjustment of working set limits can be disabled on a
per-process basis by setting the limit and quota equal.

12-16

SYSTEM PARAMETERS
12.2.9

WSDEC Parameter

WSDEC specifies the number of pages by which the limit of a working
set is automatically decreased at each adjustment interval (which is
quantum end). At the default setting, for example, the system will
decrease the limit of a working set by 5 pages each time a decrease is
required.
Increasing the value of this parameter tends to increase the speed
with which working set limits are decreased when the need arises.
Normally, this parameter should be kept at a low value because a
sudden drop in the working set limit may seriously impair performance.
In addition, the value of this parameter should be prime relative to
WSINC so that working set limits will vary as they are increased and
decreased.

12.2.10

AWSMIN Parameter

AWSMIN establishes the lowest number of pages to which a
limit can be decreased automatically.

12.2.11

working

set

AWSMAX Parameter

AWSMAX establishes the highest number of pages to which a working set
limit can be increased automatically. However, the system will never
increase the limit of a working set above the quota of its process.
A setting equal to the value of the WSMAX parameter permits all
working set limits to increase automatically to the quotas of their
processes.

12.2.12

AWSTIME Parameter

AWSTIME specifies the minimum amount of processor time that must
elapse for the system to collect a significant sample of a working
set's page fault rate.
The time is expressed in units of 10
milliseconds.
The
default value of 20, for example,
is 200
milliseconds. However, the practical value of this parameter never
exceeds the quantum, as page fault rates are sampled at quantum end.

12.2.13

EXTRACPU Parameter

EXTRACPU sets the time, in units of 10 milliseconds, allotted to each
of a process's exit handlers (for each access mode) after the process
times out (that is, reaches its CPU time limit).

12.2.14

MAXSYSGROUP Parameter

MAXSYSGROUP sets the highest value that a group number can have and
still be classified as a system UIC group number. Note that the
specification is not in octal unless preceded by the %0 radix
indicator. This parameter is normally left at 8 (10 octal).

12-17

SYSTEM PARAMETERS
12.2.15

MAXBUF Parameter

MAXBUF sets the maximum size of a buffered I/O transfer (card readers,
console floppy diskettes, line printers, mailboxes, and terminals).
The space for a buffered I/O operation is allocated from the
permanently resident nonpaged dynamic pool. Note that the system adds
approximately 16 bytes to a buffer at allocation time for header
information, so that the default (1024) -- for most devices -- means
that a buffer no larger than 1008 bytes can be allocated.

12.2.16

DEFMBX Parameters

The DEFMBX parameters set defaults for the mailbox buffer quota and
message size when these values are not specified in a Create Mailbox
($CREMBX) system service call.

12.2.17

FREELIM Parameter

FREELIM sets the minimum number of pages that must be on the free page
list.
The system will write pages from the modified page list or
outswap working sets to maintain the minimum count, if necessary.
While the larger free page list generally means less paging I/O, it
also means less space for the balance set, which tends to result in
more swapping I/O. The system manager can monitor the size of the
free page list, the amount of paging, and the amount of swapping with
the !ORATES command of the DISPLAY utility.

12.2.18

XFMAXRATE Parameter

XFMAXRATE limits the data transfer rate that can be set for DR32
devices.
On some hardware configurations (especially those without
interleaved memory), a high DR32 transfer rate could cause a machine
check (CPU timeout).

12.2.19

LAMAPREGS Parameter

LAMAPREGS sets the number of UNIBUS map registers allocated to an
LPAll driver when the driver is loaded, and limits the registers for
the driver to that number. A value of 0 permits dynamic allocation of
an unlimited number of registers.

12.2.20

REALTIME SPTS Parameter

REALTIME SPTS reserves a number of system page table entries for
mapping -connect-to-interrupt processes into system space. This value
should normally remain at the default (0) in an environment that is
not real-time.
Where connect-to-interrupt processes do use the
system, this value should represent the maximum number of pages all
concurrent connect-to-interrupt processes must map into system space.
See the VAX/VMS_ ~~-c:l__~_:_!_!_f!!~ _ _!:!_ser's _<3-u~9~ for details.

12-18

SYSTEM PARAMETERS

12.2.21

CLISYMTBL Parameter

CLISYMTBL sets the size of the command interpreter symbol table.
For
generation of RSX-llS systems, the value of this parameter should be
increased to 35. Otherwise, the default value is adequate.

12.2.22

BUGREBOOT Parameter

BUGREBOOT enables or disables automatic rebooting of the system if a
fatal bugcheck occurs.
This switch should normally be off (0) only
when the executive is being debugged.

12.2.23

CRDENABLE Parameter

CRDENABLE
enables
or
disables
detection
and
logging
of
memory-corrected read data (ECC) errors. This switch should normally
be on ( 1) •

12.2.24

DUMPBUG Parameter

DUMPBUG enables or disables the writing of
memory contents to SYS$SYSTEM:SYSDUMP.DMP
occurs. This switch should be off (0) only
being debugged.

12.2.25

error log buffers and
when a fatal bugcheck
when the executive is

BUGCHECKFATAL Parameter

BUGCHECKFATAL enables or disables the making of all bugchecks fatal
bugchecks.
The system must be rebooted on a fatal bugcheck. A
nonfatal bugcheck just places an entry in the error log. This switch
should normally be on (1).

12.2.26

POOL PAGING Parameter

POOL PAGING enables or disables paging of the pageable dynamic pool.
When- paging is disabled, the pageable dynamic pool is allocated as
permanently resident memory, just like the nonpaged dynamic pool.
This switch should normally be on (1) to conserve physical memory.

12.2.27

SBIERRENABLE Parameter

SBIERRENABLE enables or disables error logging for certain classes of
SBI errors.
This switch should be off only under extraordinary
debugging situations.

12-19

SYSTEM PARAMETERS
12.2.28

SETTIME Parameter

SETTIME enables or disables solicitation of the time of day each time
the system is booted. This switch should normally be off (0), so that
the system sets the time of day at boot time to the value of the
processor time-of-day register. The system manager can reset the time
after the system is up with the SET TIME command (see the VAX/VMS
Opera t ?_!:__'_~ __ (;_~_i~e._) •

12.2.29

SYSPAGING Parameter

SYSPAGING ~nables or disables the paging of code in the system working
set~ ·
This parameter may be off
(0) for a system with sufficient
memory and a stringent real-time response requirement, or in a
debugging situation, but should normally be on (1). Setting this
switch off does not disable the paging of VAX-11 RMS code.

12.2.30

UAFALTERNATE Parameter

UAFALTERNATE enables or disables the assignment of SYSUAF as the
logical name for SYSUAFALT, causing all references to the user
authorization file (SYSUAF) to be translated to SYS$SYSTEM:SYSUAFALT.
Use of the normal user authorization file (SYS$SYSTEM:SYSUAF) can be
restored by deassigning the system logical name SYSUAF.
This switch
should be on (1) only where the system is being used by a restricted
set of users. The system manager must create a user authorization
file named SYSUAFALT prior to its use (see Chapter 2).

12.2.31

RESALLOC Parameter

This parameter is not in use currently.

12.2.32

TTYSCANDELTA Parameter

TTYSCANDELTA sets the interval for polling terminals for dial-up and
hang-up events.
Shorter intervals use more processor time; longer
intervals may result in missing a hang-up event.

12.2.33

TTY SPEED Parameter

TTY SPEED sets the default speed for terminals,
codes:

using

Code

Baud Rate

Code

Baud Rate

1
2
3
4
5

50
75
110
134.5
150
300
600

8
9
10
11
12
13
14
15

1200
1800
2000
2400
3600
4800
7200
9600

12-20

the

following

SYSTEM PARAMETERS
12.2.34

TTY BUF Parameter

TTY BUF sets the default line width for terminals.

12.2.35

TTY DEFCHAR Parameter

TTY DEFCHAR sets the default characteristics for
code derived by summing the following values:

terminals,

Characteristic

Value in Hexadecimal

page length
if SCOPE
if WRAPAROUND
if MECH TAB
if LOWER CASE
if TERMINAL SYNCH

(page length)*l,000,000
1,000
200
100
80
20

Where a condition is false, the value is

Terminal

Common codes include:

Value in Hexadecimal Decimal Equivalent

LA36
VT52
VTlOO

12.2.36

using

80002AO
180012AO
180012AO

134218400
402057952
402057952

TTY TYPAHDSZ Parameter

TTY TYPAHDSZ sets the size of the terminal type-ahead buffer.

12.2.37

TTY PROT Parameter

TTY PROT sets the default protection for all terminals in relation to
the
UIC
specified
for
the
TTY OWNER parameter below.
The
specification must represent the value of a standard lo-bit protection
mask as described in Chapter 3.
However, only read bits are
meaningful. When a read bit is on, that category of user is
prohibited from allocating terminals.
The default (0) provides for no protection on terminals.
The system
manager should normally set TTY PROT to a value of hexadecimal 1110,
which permits only system processes to allocate terminals. The system
manager can change protection on a per-terminal basis with the SET
PROTECTION /TERMINAL command to permit user allocation of remote and
application terminals.

12.2.38

TTY OWNER Parameter

TTY_OWNER specifies the owner UIC against which terminal protection is
set.
The specification must represent the value of a standard 32-bit
UIC -- group number in the high-order word and member number in the
low-order word. The system manager should normally set TTY OWNER to a
value of hexadecimal 10004, which is UIC [1,4].

12-21

SYSTEM PARAMETERS
12.3

JOB PARAMETERS

The JOB parameters should be checked for reasonableness, as
in the sections that follow.

12.3.1

described

MAXPRINTSYMB Parameter

MAXPRINTSYMB sets the maximum number of print symbionts that can be
created.
(However, the maximum number of print symbionts is further
restricted by the value of the PQL DPRCLM parameter
the default
process creation limit.)
-

12.3.2

DEFPRI Parameter

DEFPRI sets the base default priority for processes.
normally be left at 4.

12.3.3

Its value should

JOBLIM Parameters

The JOBLIM parameters set the maximum number of interactive (IJOBLIM),
batch
(BJOBLIM), and network (NJOBLIM) jobs that can be on the system
concurrently. The maximum number of processes (MAXPROCESSCNT) may
actually reduce any or all of the job limits.

12.4

ACP PARAMETERS

Varying circumstances make changes in the ACP
as described in the sections that follow.

12.4.1

parameters

worthwhile,

ACP MULTIPLE Parameter

ACP MULTIPLE enables or disables the default creation of a separate
disk ACP process for each volume mounted on a different device type.
Even where memory is not at a premium, this switch should normally be
off
(0) to queue all requests through the same ACP process.
Performance is better enhanced by increasing cache sizes (see Section
12.4.3) than by adding another ACP. The switch can be overridden on
an individual-volume basis with the DCL command MOUNT.

12.4.2 ACP SHARE Parameter
ACP SHARE enables or disables the creation of a global section for the
first ACP used, so that succeeding ACPs may share its code. This
switch should be set off (0) when ACP MULTIPLE is off.

12-22

SYSTEM PARAMETERS
12.4.3

CACHE Parameters

The CACHE parameters set the number of pages that each Files-11 ACP
can allocate for caching control information rather than going to disk
each time such information is needed or is modified.
Cache sizes
should generally be maximized to reduce ACP I/O activity.
They can be
overridden on an individual volume basis in the DCL command Mount.

12.4.3.1 ACP MAPCACHE Parameter - ACP MAPCACHE
pages for caching bit map blocks.

sets

the

number

12.4.3.2 ACP HDRCACHE Parameter - ACP HDRCACHE
pages for caching file header blocks.

sets

the

number

12.4.3.3 ACP DIRCACHE Parameter - ACP DIRCACHE
pages for caching directory blocks.

sets

the

number

12.4.3.4 ACP FIDCACHE Parameter - ACP FIDCACHE sets the number of
file identification slots cached. A specification of 1 means no cache.

12.4.3.5 ACP EXTCACHE Parameter - ACP EXTCACHE sets the number of
entries in the extent cache.
Each entry points to one contiguous area
of free space on disk. A specification of O means no cache.

12.4.3.~
ACP EXTLIMIT Parameter - ACP EXTLIMIT specifies the maximum
amount of free space to which the extent cache can point, expressed in
thousandths of the currently available free blocks on the disk.
For
example,
if available free space on the disk is 20,000 blocks, a
specification of 10 limits the extent cache to 200 blocks.
The reason
for this parameter is to limit the amount of free space that might be
lost in the event of a system failure.
However, lost free space on a
volume is normally recovered at mount time.

12.4.3.7 ACP QUOCACHE Parameter - ACP QUOCACHE sets the number
quota file entries cached. A specification of o means no cache.

12.4.3.8 ACP WRITEBACK Parameter - ACP WRITEBACK enables the deferred
writing
of - file
headers.
A specification of O causes all
modifications of file headers to be written to disk immediately.

12.4.3.9 ACP SYSACC Parameter - ACP SYSACC sets
the
number
of
directory file control blocks
(FCBs) that will be cached for disks
mounted with the /SYSTEM qualifier.
Each directory FCB contains a
16-byte array containing the first letter of the last entry in each
block of the directory (or group of blocks if the directory exceeds l~
blocks).
Since entries in a directory are alphabetical, the cached
FCB provides quick access to a required directory block.
12-23

SYSTEM PARAMETERS
This parameter value should be roughly equivalent to the number of
directories that will be in use concurrently on each system volume.
It may be overridden on a per-volume basis with the /ACCESSED
qualifier of the MOUNT command.
The value should be kept down in
small systems, as the FCBs require a significant amount of space in
the nonpaged dynamic pool.

ACP WORKSET Parameter

12.4.4

ACP WORKSET sets the default size of a working set for an ACP.
A
value of zero permits the ACP to calculate the size. This value
should be nonzero only on small systems where memory is tight.

ACP MAXREAD Parameter

12.4.5

ACP MAXREAD sets the maximum number of directory blocks read
I/0-operation. This parameter does not affect user file I/O.

one

ACP WINDOW Parameter

12.4.6

ACP WINDOW sets the default number of window pointers to be allocated
in a window for a default file access, for disks mounted with the
/SYSTEM qualifier.

ACP DATACHECK Parameter

12.4.7

ACP DATACHECK enables verification of reading and/or writing of file
structure data
(for example, directories and file headers):
a
specification of 3 means read and write checks; 2 means write check
only;
l means read check only; 0 means no checks. On a read check,
the ACP information is read twice and compared. On a write check, the
ACP information is written, then read and compared.

ACP BASEPRIO Parameter

12.4.8

ACP BASEPRIO sets the base priority for all ACPs.
The SET PROCESS
/PRIORITY command can be used to reset the base priorities of
individual ACPs.

ACP SWAPFLGS Parameter

12.4.9

ACP SWAPFLGS enables or disables swapping for
through the value of a 4-bit number:

Bit
0
1
2
3

Class of ACP
Disks mounted
Disks mounted
Private disks
Magnetic tape

/SYSTEM
/GROUP
ACP

12-24

four

classes

ACPs

SYSTEM PARAMETERS

If the value of the bit is 1, the corresponding class of ACPs can be
swapped.
The default value of hexadecimal F (all bits on) enables
swapping for all classes of ACP.
A value of hexadecimal E would
disable swapping for ACPs for volumes mounted with the /SYSTEM
qualifier. Normally, swapping is enabled for all ACPs except those
for system volumes.

12.5

RMS PARAMETERS

The RMS parameters set default blocking factors for all VAX-11 RMS I/O
operations
(RMS DFMBC) or selected VAX-11 RMS I/O operations (the
remaining RMS parameters). For small systems, the RMS parameters are
normally left at their default values (RMS DFMBC equals 4 and the
remaining RMS parameters equal 0).
For large- systems, the system
manager should increase the value of RMS DFMBC or selected RMS
parameters affecting operations involving many contingent blocks of
data.

12.6

PQL PARAMETERS

The PQL parameters set default limits for processes created by the
Create Process ($CREPRC) system service and the RUN (Process) command.
They are normally left at their default values.

12-25

CHAPTER 13

SYSTEM GENERATION

The system manager performs
configuration of the system:

13.1

various

functions

ensure

proper

•

System parameters -- Creates and modifies the standard system
parameter files for use in subsequent conversational bootstrap
operations; modifies the parameter values in the current
system
image
(SYS$SYSTEM:SYS.EXE) for use in subsequent
bootstrap operations;
dynamically modifies the parameter
values of the active system (applies only to the dynamic
system parameters)

•

Devices and device drivers -- Connects devices and loads their
device drivers (most of this work is automatic)

•

System files -- Creates swapping, paging, and dump files

•

Start-up
command
procedure -- Names
site-independent start-up command procedure

•

Multiport
units

(shared)

memory -- Initializes

the

current

multiport

memory

SYSGEN UTILITY

The system manager performs the system generation functions with the
SYSGEN utility.
Table 13-1 summarizes the SYSGEN commands by format
and function.
Table 13-1
SYSGEN Command Summary
----·-----------.
Format

Function

AUTOCONFIGURE nexus
[/SELECT=(device, ••• )]

Automatically connects devices
physically attached to the system
and loads their drivers

CONNECT device

Connects hardware devices and loads
their drivers (if they are not
already loaded)

/ADAPTER=nexus
[/CSR=cs r-addr]
[/DRIVERNAME=driver]
[/MAXUNITS=max-unit-cnt]
[/NUMVEC=vector-cnt]
[/VECTOR=vector-addr]

___________________ _____________ · - - - - -

..__

_.__

(continued on next page)

13-1

SYSTEM GENERATION

SYSGEN
Table 13-1 (Cont.)
SYSGEN Command Summary

Format

Function

CONNECT device /NOADAPTER
[/DRIVERNAME=driver]

Connects software devices and loads
their drivers (if they are not
already loaded)

CONNECT CONSOLE

Connects the console block storage
device and loads its driver

CREATE file-spec /SIZE=block-count
[/[NO]CONTIGUOUS]

Creates a paging, swapping, or
dump file

DISABLE CHECKS

Disables range checks

ENABLE CHECKS

Enables range checks

EXIT

Terminates SYSGEN

HELP [command

[qualifier]]

Lists the SYSGEN commands

INSTALL file-spec {/PAGEFILE}
/SWAP FILE

ActivatBs a secondary paging or swapping
file

LOAD file-spec

Loads a device driver

RELOAD file-spec

Loads a new version of an existing
device driver

SET parameter-name value

Modifies the value of a system
generation parameter in the SYSGEN
work area

SET /STARTUP file-spec

Names the current site-independent
start-up command procedure

SHARE MPMn mpm-name
/INIT
[/GBLSECTIONS=gbl]
[/MAILBOXES=mail]
[/CEFCLUSTERS=cef]
[/MAXGBLSECTIONS=max-gbl]
[/MAXMAILBOXES=max-mail]
[/MAXCEFCLUSTERS=max-cef]
[/POOLBCNT=block-cnt]
[/POOLBSIZE=block-size]
[/PRQCOUNT=prq-cnt]
[/START=start]

Initializes a multiport memory unit

SHARE MPMn mpm-name
[/GBLSECTIONS=gbl]
[/MAILBOXES=mail]
[/CEFCLUSTERS=cef]

Connects a multiport memory unit

(continued on next page)

13-2

SYSTEM GENERATION

SYSGEN (Cont.)
Table 13-1 (Cont.)
SYSGEN Command Summary
Format

Function
Displays the values of system
parameters in the SYSGEN work area,
plus the default, minimum, and maximum
values of the parameters, and their
units of measure

parameter-name
/ACP
/ALL
/DYNAMIC
/GEN
/JOB
/MAJOR
/NAMES
/PQL
/RMS
/SYS

SHOW

SHOW /DEVICE[=driver-name]

Displays information on connected
devices and loaded drivers

SHOW /STARTUP

Displays the name of the current
site-independent start-up command
procedure

USE

Initializes the SYSGEN work area with
system parameter values from a
parameter file, the current system
image, the active system, or the
default list

file-spec }
CURRENT
{ ACTIVE
DEFAULT

WRITE

Writes the system parameter values
from the SYSGEN work area to a
parameter file, the current system
image, or the active system

file-spec }
CURRENT
{ ACTIVE

' - - - - - - - - - - - - - - - - - - - - - - - - - - - - - L - - - - - - - - - - - - - ·-------

Many of the SYSGEN commands are specific to particular functions:
ENABLE,

DISABLE,

WRITE,

and

•

System parameters -- USE, SET,
SHOW

•

Devices and device drivers -- AUTOCONFIGURE,
RELOAD, and SHOW /DEVICES

System files -- CREATE and INSTALL

Start-up command procedure -- SET /STARTUP and SHOW /STARTUP

•

Multiport memory

CONNECT,

LOAD,

A subset of the SYSGEN commands can be used during
operations;
see the VAX-11 Software Installation Guide.

13-3

bootstrap

SYSTEM GENERATION

SYSGEN (Cont.)
13.1.1

Invoking SYSGEN

The following command invokes the utility:
$ RUN SYS$SYSTEM:SYSGEN

The system responds with the following prompt:
SYSGEN>
The system manager can then enter any of the commands listed in Table
13-1.
These commands follow the standard rules of grammar as
specified in the VAX/Vfv1~_~_om~.~nd ~-~-.~~~.~-2~---~5.~E-~~---s;-~-~.~~ •

13.1.2

AUTOCONFIGURE Command

AUTOCONFIGURE automatically connects devices
attached to the system and loads their drivers.

that are physically
It takes the form:

AUTOCONFIGURE nexus [/SELECT=(device, •.• )]
nexus
Nexus number of a UNIBUS or MASSBUS adapter, or the keyword ALL
device
Standard device name as listed in Table 13-2 (see Section 13.2.2)
and (optionally) a controller designation;
if the controller
designation is omitted, device generically specifies all devices
of the specifiBd type; parentheses can be omitted for a single
device; defaults to all devices on the adapter
The autoconf igure operation assigns standard device names and assumes
standard driver names, as listed in Table 13-2 (see Section 13.2.2).
The following example automatically configures all standard devices:
SYSGEN> AUTOCONFIGURE ALL
The next example automatically configures all terminals, all magnetic
tape units on controller A, all RM03 disks, and all line printers:
SYSGEN> AUTOCONFIGURE ALL /SELECT=(TT,MTA,DR,LP)
Use of the AUTOCONFIGURE
privileges.

command

requires

13-4

the

CMEXEC

and

CMKRNL

SYSTEM GENERATION

SYSGEN {Cont.)
13.1.3

CONNECT (Hardware) Command

CONNECT (hardware) connects a hardware device and loads its driver, if
the driver is not already loaded. It takes the form:
CONNECT device /ADAPTER=nexus
[/CSR=csr-addr]
[/DRIVERNAME=driver]
[/MAXUNITS=max-unit-cnt]
[/NUMVEC=vector-cnt]
[/VECTOR=vector-addr]
device
Name of the device;
should take the form
device
type,
controller, unit -- for example, LPAO for unit 0 on controller A
of device type LP
nexus
Nexus number of the UNIBUS or MASSBUS adapter to which the device
is attached
csr-addr
UNIBUS address of the controller status register for the
must be specified for UNIBUS devices

device;

driver
Name of the driver as recorded in the prologue table;
if the
driver has not been loaded, the system assumes that the driver
name is also the name of an executable (file type of EXE)
image
in SYS$SYSTEM, and loads the driver; DIGITAL standard drivers
are named in Table 13-2 (see Section 13.2.2);
defaults to the
first two characters of the device name plus DRIVER
.max-uni t-cnt
Highest number unit attached to the device controller plus l;
defaults to the number specified in the prologue table of the
driver, or 8 (if not explicitly specified in the prologue table)
vector-cnt
Number of interrupt vectors for the device;
vector-addr
UNIBUS address of the interrupt vector for
lowest vector if there is more than one;
UNIBUS devices

defaults to l
the device, or the
must be specified for

The system manager typically uses this command to connect devices that
are not physically attached or whose drivers are not standard. The
following example connects the device named LPAO to the driver named
LPDRIVER (the default driver name), and loads the driver if it is not
already loaded:
SYSGEN> CONNECT LPAO /ADAPTER=3 /CSR=%0777514/VECTOR=%0200
The next example performs the same operation, but uses
driver named LP2DRIVER:
SYSGEN> CONNECT LPAO /ADAPTER=3 /CSR=%0777514_/DRIVERNAME=LP2DRIVER/VECTOR=%0200

13-5

nonstandard

SYSTEM GENERATION

SVSGEN (Cont.)
The system manager should exercise extreme caution when using the
CONNECT command, as the system does little error checking. An
incorrect vector address or misspelled device name, for example, will
damage the I/O data base and very likely cause the system to fail.
The VAX/VMS Guide to Writing a Device Driver contains more detailed
information on Toading -device drivers and co-nnecting devices.
Use of the CONNECT command requires the CMEXEC and CMKRNL privileges.

13.1.4

CONNECT (Software) Command

CONNECT (software) connects a software device and loads its driver
it is not already loaded. It takes the form:

CONNECT device /NOADAPTER
[/DRIVERNAME=driver]
device
Name of the device
driver
Name of the driver as recorded in the prologue table;
if the
driver has not been loaded, the system assumes that the driver
name is also the name of an executable (file type of EXE)
image
in SYS$SYSTEM, and loads the driver; DIGITAL standard drivers
are named in Table 13-2 (see Section 13.2.2);
defaults to the
first two characters of the device name plus DEVICE
The following example connects the device NET to the driver NETDRIVER,
and loads the driver if it is not already loaded:
SYSGEN>

CONNECT NET /NOADAPTER /DRIVER=NETDRIVER

Use of the CONNECT command requires the CMEXEC and CMKRNL privileges.

13.1.5

CONNECT (Console) Command

CONNECT (console) connects the console block storage device and
its driver. It takes the form:

loads

CONNECT CONSOLE
The name of the console block storage device is CSAl.
Use of the CONNECT command requires the CMEXEC and CMKRNL privileges.

13.1.6

CREATE Command

CREATE creates a file that can be used as a primary paging, swapping,
or dump file, or a secondary paging or swapping file.
It takes the
form:
CREATE file-spec /SIZE=block-count
[/[NO]CONTIGUOUS]
13-6

SYSTEM GENERATION

SYSGEN (Cont.)
file-spec
Name
of
the
file;
primary
files
must
be
SYS$SYSTEM:PAGEFILE.SYS, SWAPFILE.SYS, and SYSDUMP.DMP

named

block-count
Number of blocks to be allocated to the file
Primary files must be
contiguous.
noncontiguous and can span volumes
specification is /CONTIGUOUS.

Secondary
files
can
be
a volume set. The default

The following example creates a primary paging file· of 98,304 blocks:
SYSGEN> CREATE SYS$SYSTEM:PAGEFILE.SYS /SIZE=98304
The next example creates a secondary non-contiguous paging file:
SYSGEN> CREATE DRA5: [SYSTEM]PAGEFILE.SYS /SIZE=98304
/NOCONTIGUOUS
A secondary file
13.1.11)

must

installed

take

effect

(see

Section

The system manager does not have to issue this command directly for
the creation of primary files, but can use the paging, swapping, and
dump files provided with the software distribution kit, or issue the
command procedure SYS$SYSDISK: [SYSUPD]SWAPFILES to create new primary
files.

13.1.7

DISABLE Command

DISABLE inhibits range checks on parameter
commands. It takes the form:

values

specified

SET

DISABLE CHECKS
Initially, checks are enabled. If the system manager specifies a
value above the maximum, the system sets the parameter to the maximum
and issues an error message. If the system manager specifies a value
below the minimum, the system sets the parameter to the minimum and
issues an error message.
In the following example, the system
manager, foiled in the initial attempt to set WSMAX below the minimum,
disables range checks:
SYSGEN> SET WSMAX 20
%SYSGEN-W-Value set to minimum
SYSGEN> DISABLE CHECKS
SYSGEN> SET WSMAX 20

13.1.8

ENABLE Command

ENABLE ensures that range checks are in effect.

It takes the form:

ENABLE CHECKS
Initially, range checks are enabled, so that this command need be used
only when a DISABLE command occurs earlier in the session.
13-7

SYSTEM GENERATION

SYSGEN (Cont.)

13.1.9

EXIT Command
It

EXIT returns the system manager to DCL command level.
form:

takes

the

EXIT
The system manager can also return to
<CTRL/Z>.

13.1.10

DCL

command

level

typing

HELP Command

HELP lists and explains the SYSGEN commands.

It takes the form:

HELP [command-name [qualifier]]
command-name
Name of a SYSGEN command
qualifier
Name of a SYSGEN qualifier
If no command name is specified, HELP displays general information on
all commands.
Specification of command names and qualifiers obtains
more detailed information.

13.1.11

INSTALL Command

INSTALL activates a secondary paging or swapping file.
form:

takes

the

.
{/PAGEFILE}
INSTALL file-spec
/SWAPFILE
file-spec
Name of a secondary paging or swapping file created with a CREATE
command
Either /PAGEFILE or /SWAPFILE must be specified to indicate
file.

the

type

All processes started after entry of the INSTALL command use the new
paging or swapping file.
Current processes continue using the old
file.
The following example installs a secondary file:
SYSGEN> INSTALL DRA5: [SYSTEM]PAGEFILE.SYS /PAGEFILE
The new paging or swapping file is effective until system shutdown.
Use of the INSTALL command requires the CMKRNL privilege.

13-8

SYSTEM GENERATION

SYSGEN (Cont.)

13.1.12

LOAD Command

LOAD loads an I/O driver.

It takes the form:

LOAD file-spec
file-spec
File specification of the driver image;
EXE

file

type

defaults

If the file specification is the same as a driver already loaded, no
load takes place.
If the file name is the same as a driver already
loaded (but the file speciflcation is different), the specified driver
replaces the existing driver.
DIGITAL supplies drivers for the standard devices, as listed in
13-2 (see Section 13.2.2).

Table

The following example loads the standard driver for a remote terminal:
SYSGEN>

LOAD SYS$SYSTEM:RTTDRIVER

The VAX/VMS Guide to Writing a Device Driver
information on loading a driver.

contains

detailed

Use of the LOAD command requires the CMEXEC and CMKRNL privileges.

13.1.13

RELOAD Command

RELOAD replaces a loaded device driver with a new version.
the form:

takes

RELOAD file-spec
file-spec
File specification of the new driver image;
to EXE

file

type

defaults

The specified image is loaded and replaces any existing driver with
the same file specification. The following example reloads the remote
terminal driver:
SYSGEN>

RELOAD SYS$SYSTEM:RTTDRIVER

Use of the RELOAD command requires the CMEXEC and CMKRNL privileges.

13-9

SYSTEM GENERATION

SYSGEN (Cont.)
13.1.14

SET (Parameter) Command

SET (parameter) assigns a value to a system parameter as it exists
the SYSGEN work area.
The command takes the form:

SET parameter-name value
parameter-name
Name of a system parameter as specified in
12-6 (see Chapter 12)

Tables

12-1

through

value
An integer or the keyword DEFAULT;
integer values must be within
the
minimum and maximum values unless DISABLE CHECKS was
specified;
DEFAULT means the default value for the parameter
This command does not modify parameter files, the current system image
on disk,
or the active system.
See the WRITE command below.
The
following example assigns a value of 20 to the PFCDEFAULT parameter:
SYSGEN>

SET PFCDEFAULT 20

The next example assigns the default value
parameter:
SYSGEN>

13.1.15

(40)

the

GBLSECTIONS

SET GBLSECTIONS DEFAULT

SET (Start-Up Command Procedure) Command

SET (start-up command procedure) names the site-independent start-up
command
procedure to be associated with a parameter file
for
subsequent bootstrap operations.
It takes the form:
SET /STARTUP file-spec
file-spec
File specification of a start-up command procedure on the
disk (maximum of 31 characters)

system

The SET command does not modify parameter files or the current
image on disk.
See the WRITE command below.

system

The following example assigns SYS$SYSTEM:XSTARTUP.COM as
site-independent start-up command procedure:

current

SYSGEN>

the

SET /STARTUP SYS$SYSTEM:XSTARTUP.COM

The
initial
(as
named
in
the
software
distribution
kit)
site-independent start-up command procedure is SYS$SYSTEM:STARTUP.COM.

13-10

SYSTEM GENERATION

SYSGEN (Cont.)

13.1.16

SHARE (Initialize) Command

SHARE (initialize) initializes a multiport memory unit, if it is not
already initialized, and connects it to the processor on which SYSGEN
is running. The command takes the form:
SHARE MPMn mpm-name /INIT
[/GBLSECTIONS=gbl]
[/MAILBOXES=mail]
[/CEFCLUSTERS=cef]
[/MAXGBLSECTIONS=max-gbl]
[/MAXMAILBOXES=max-mail]
[/MAXCEFCLUSTERS=max-cef]
[/POOLBCNT=block-cnt]
[/POOLBSIZE=block-size]
[/PRQCOUNT=prq-cnt]
[/START=start]
n

Number on the front panel of
initialized

the

multiport

memory

unit

being

mpm-name
Name by which the multiport memory unit is to be known to systems
using it;
consists of 1 to 15 alphanumeric characters, dollar
signs, or underlines
gbl
Total global sections permitted in
defaults to 32

the

multiport

memory

unit;

ma i 1
Total mailboxes permitted in the multiport memory unit;
to 32

defaults

Total common event flag clusters
memory unit; defaults to 32

multiport

cef
permitted

max-gbl
Maximum global sections that this processor can
multiport memory unit; defaults to no limit

the

create

the

max-ma i 1
Maximum mailboxes that this processor can create in the multiport
memory unit; defaults to no limit
max-cef
Maximum common event flag clusters that this processor can create
in the multiport memory unit; defaults to no limit
block-cnt
Number of blocks allocated to the multiport memory unit's dynamic
pool; defaults to 128 blocks
block-size
Size of each block in the dynamic pool;

13-11

defaults to 128 bytes

SYSTEM GENERATION

SYSGEN {Cont.)
prq-cnt
Number of interprocess
defaults to 64

request

blocks

(PRQs)

allocated;

start
Relative page number of first page in the multiport
for use by VMS data structures; defaults to 1

memory

unit

The dynamic pool is used for AST control blocks, message buffers, and
other small dynamic structures. The start specification permits space
to be left at the front of the multiport memory unit for the dedicated
use of one processor or for non-VAX/VMS use.
If the specified multiport memory unit is already initialized and
connected to other active processors, the gbl, mail, cef, block-cnt,
block-size, prq-cnt, and start parameter values are ignored, and the
unit is simply connected to the processor.
The following example initializes a multiport memory
defaults on all but the gbl, mail, and cef parameters:

unit

with

SYSGEN> SHARE MPMl SHR MEMl /INIT_IGBLSECTIONS=l28 /MAILBOXES=64 /CEFCLUSTERS=O
The number of the multiport memory unit as it appears on the front
panel is 1. The system manager names the unit SHR MEM 1.
Use of the SHARE command requires the CMKRNL and CMEXEC privileges.

13.1.17

SHARE (Connect) Command

SHARE (connect) connects a processor to a multiport memory unit
already initialized by this or another processor. The command takes
the form:
SHARE MPMn mpm-name [/MAXGBLSECTIONS=max-gbl]
[/MAXMAILBOXES=max-mail]
[/MAXCEFCLUSTERS=max-cef]
n

Number on the front panel of
connected

the

mpm-name
Name of the multiport memory unit
SHARE (initialize) command

multiport

memory

specified

max-gbl
Maximum global sections that this processor can
multiport memory unit; defaults to no limit

unit

create

being

the

13-12

SYSTEM GENERATION

SYSGEN (Cont.)
The number and name of the multiport memory unit must match
an initialized unit or an error condition results.

those

The following example connects a multiport memory unit
on the parameters:

defaults

SYSGEN>

with

SHARE MPMl SHR MEM 1

The unit with a 1 on the front panel must be initialized with the name
SHR MEM 1 for the command to work.
Use of the SHARE command requires the CMKRNL and CMEXEC privileges.

13.1.18

SHOW (Parameter) Command

SHOW (parameter) displays the values of system parameters in the
SYSGEN work area, plus the default, minimum, and maximum values of the
parameters, and their units of measure. It takes the form:

SHOW

parameter-name
/ACP
/ALL
/DYNAMIC
/GEN
/JOB
/MAJOR
/NAMES
/PQL
/RMS
/SYS

parameter-name
Name of a system parameter as specified in
12-6 (see Chapter 12)

Tables

12-1

through

The system manager can specify only one parameter name or one of the
qualifiers.
Specification of a parameter name displays the values of
that parameter. Specification of a qualifier displays the values of
all
ACP parameters
(/ACP), all parameters
(/ALL), all DYNAMIC
parameters (/DYNAMIC), all GEN parameters (/GEN), all JOB parameters
(/JOB), all MAJOR parameters (/MAJOR), all PQL parameters (/PQL), all
RMS parameters (/RMS), all SYS parameters (/SYS), or just the names of
all parameters (/NAMES).
NOTE
When parameter names are abbreviated,
the
first
parameter
matching
the
abbreviation is selected for display.
No
ambiguity checks are made.
For
example, a specification of SHOW GBL
displays the GBLSECTIONS parameter. A
display of
the
GBLPAGES
parameter
requires a specification of (at least)
SHOW GBLP.

13-13

SYSTEM GENERATION

SYSGEN {Cont.)
Figure 13-1 illustrates the SHOW command.
SYSGEN>

SHOW /MAJOR

Parameter Name

-------------PFCDEFAULT

GBLSECTIONS
GBLPAGES
MAXPROCESSCNT
SYSMWCNT
BALSETCNT
IRPCOUNT
WSMAX
NPAGEDYN
PAGEDYN
VIRTUALPAGECNT
QUANTUM
MPW WRTCLUSTER
MPW-HILIM
MPW-LOLIM

Current

-------

16
40
2048
64
48
24
80
250
40448
8192
8192
30
16
24
12

Default

-------

16
40
2048
64
48
24
80
2 5fi
40448
8192
8192
30
16
24
12

Minimum

Maximum

0
20
512
12
20
4
0
60
lfi384
8192
512
2
0
0
0

127
-1
-1
256
16384
1024
32768
10384
-1
-1
65536
32767
127
16384
16384

-------

-·-------··----~-

Figure 13-1

13.1.19

-------

Unit
PAGES
SECTIONS
PAGES
PROCESSES
PAGES
SLOTS
PACKETS
PAGES
BYTES
BYTES
PAGES
lOMS
PAGES
PAGES
PAGES

..----- --·-·······-------·····---·--.. -- ------------------"--

Display Produced by SHOW Command of SYSGEN

SHOW (Device) Command

SHOW (device) displays information on device drivers loaded into the
system,
the devices connected to them, and their I/O data bases.
It
takes the form:
SHOW /DEVICE[=driver-name]
driver-name
Name of the driver;
the system

defaults to all device drivers

loaded

The specific information displayed is as follows:
•

Driver -- Name of the driver

•

Start -- Starting address of the driver

•

End

Ending address of the driver

•

Dev

Name of each device connected to the driver

DDB

Address of the device's device data block

•

CRB

Address of the device's channel request block

IDB

Address of the device's interrupt dispatch block

•

Unit -- Number of each unit on the device

•

UCB -- Address of each unit's unit control block

13-14

into

SYSTEM GENERATION

SYSGEN (Cont.)
All addresses are
in hexadecimal and
illustrates the SHOW /DEVICE command.

are

Figure

virtual.

13-2

Use of the SHOW /DEVICE command requires the CMEXEC privilege •
. . . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -SYSGEN> SHOW /DEVICE=DBDRIVER
Driver
Start
End
Dev
DDS
CRB
IDB

- -- ------------

Unit

UCB

DBDRIVER-----S008239'080082~

DBA 80000848 800988CO 80098920
0 8000087C
1 8008A4FO
2 8008A590
5 8008A630
7 8008MDO

Figure 13-2

13.1.20

Display.Produced by SHOW /DEVICE Command of SYSGEN

SHOW (Start-Up) Command

SHOW (start-up) displays the name of the current
start-up command procedure.
It takes the form:

site-independent

SHOW /STARTUP
Figure 13-3 illustrates the SHOW (start-up) command.
SYSGEN>

SHOW /STARTUP

Startup command file
Figure 13-3

13.1.21

= SYS$SYSTEM:STARTUP.COM

Display Produced by SHOW /STARTUP Command of SYSGEN

USE Command

USE initializes the SYSGEN work area with system parameter values and
the name of the site-independent start-up command procedure, from a
parameter file, the current system image on disk, the active system in
memory, or the default list.
It takes the form:

USE

I~ii:~~~;c l
f

ACTIVE
DEFAULT

file-spec
File specification
defaults to PAR

system

13-15

parameter

file;

file

type

SYSTEM GENERATION

SYSGEN (Cont.)
The system manager specifies one parameter file or one of the
keywords.
The parameter file must be either one of the standard
system parameter files supplied by DIGITAL (see Chapter 12, Table
12-7) or a file created by the system manager with the WRITE command
of SYSGEN. Existing values in the SYSGEN work area are overwritten.
The following example initializes the SYSGEN work area with the
parameter values of the 64USER standard system parameter file:
SYSGEN>

USE SYS$SYSTEM:64USER

The initial values of the SYSGEN work area when the utility is invoked
are the default values.

13.1.22

WRITE Command

WRITE writes the system parameter values and the name of the
site-independent start-up command procedure from the SYSGEN work area
to a parameter file, the current system image on disk, or the active
system in memory.
(However, only the dynamic parameter values are
written to the active system.) The command takes the form:
WRITE { f

~G~~~~;c l

ACTIVE

file-spec
File specification;

file type defaults to PAR

Specification of a file results in the creation
file, as illustrated in the following example:
SYSGEN>

new

parameter

WRITE SYS$SYSTEM:SPECIAL

The next example modifies
(SYS$SYSTEM:SYS.EXE):
SYSGEN>

the

current

system

image

disk

WRITE CURRENT

Use of the WRITE ACTIVE command requires the CMKRNL privilege.

13.1.23

Error Messages

See the VAX/VMS Syste~--~-~-~~9ge~--~r:!~--~~£gy_~~y Procedures__ ~~~~~_! for the
messages issued by the SYSGEN utility.

13-Hi

SYSTEM GENERATION

13.2

OPERATIONS

The system manager uses SYSGEN differently depending on
function being performed.

13.2.l

the

type

System Parameters

The bootstrap process initializes the active system parameter values
in memory from the current system image on disk (that is, the starting
parameter
values
are
those
in
SYS$SYSTEM:SYS.EXE).
In
a
conversational bootstrap operation, the operator can modify these
values by reinitializing the active parameter values from
any
parameter file or the default list, and by setting new parameter
values on an individual basis. The operator performs these activities
with a subset of the SYSGEN commands, as explained in the VAX-11
Software Installation Guide. At the end of the bootstrap operation,
the system image is modified to conform to the active parameter
values.
After the system is booted and running, the system manager can run
SYSGEN to create or modify parameter files, modify the current system
image, and modify the dynamic parameter values of the active system.
The system manager typically uses the following sequence of commands:
•

Invokes SYSGEN -- Invoking SYSGEN initializes a work
the default parameter values.

•

Optionally issues a USE command -- The system manager can
reinitialize the work area to the values of a parameter file,
the current system image, or the active system, if the default
values
do
not provide a suitable base for subsequent
operations.

•

Issues SET commands -- The system manager modifies parameters
on an individual basis. These modifications have no effect
outside the SYSGEN work area.

•

Issues a WRITE command -- The system manager creates
a
parameter file, modifies the current system image on disk, or
modifies the active system on disk.

During these operations, the system manager can use the SHOW
to examine the parameter values in the SYSGEN work area.

area

command

13.2.1.l Create a Parameter File - The creation of a new parameter
file does not immediately affect system performance. At subsequent
conversational bootstrap operations, however, the
operator
can
initialize the active system with the values of the new file.
In the
following example, the system manager creates a new version of the
64USER standard system parameter file with a new value for the
PFCDEFAULT parameter:
$ SET DEFAULT SYS$SYSTEM
$ RUN SYSGEN
SYSGEN> USE 64USER
SYSGEN> SET PFCDEFAULT 52
SYSGEN> WRITE 64USER
SYSGEN> EXIT

13-17

SYSTEM GENERATION

The next example creates a user
using the 64USER file as a base:

file

named

SYS$SYSTEM:52USER.PAR,

$ SET DEFAULT SYS$SYSTEM
$ RUN SYSGEN
SYSGEN> USE 64USER
SYSGEN> SET PFCDEFAULT 52
SYSGEN> WRITE 52USER
SYSGEN> EXIT

13.2.1.2 Modify the System Image - The modification of the current
system image also does not immediately affect system performance. At
subsequent bootstrap operations, however, the active system
is
initialized with the new values. A conversational bootstrap operation
permits the operator to modify these values further, while a nonstop
bootstrap operation makes the new values the values of the active
system. The following example modifies the PFCDEFAULT parameter value
in the system image:
$ RUN SYS$SYSTEM:SYSGEN
SYSGEN> USE CURRENT
SYSGEN> SET PFCDEFAULT 52
SYSGEN> WRITE CURRENT
SYSGEN> EXIT

13.2.1.3 Modify the Active System - Modification of the active system
immediately affects that subset of the system parameters called the
dynamic parameters by changing their values in the active system in
memory.
Chapter 12 lists the dynamic parameters (as does the SHOW
/DYNAMIC command). The other parameters cannot be changed immediately
because they regulate structures that cannot be changed once the
system is running. The following example modifies the active value of
the PFCDEFAULT parameter:
$ RUN SYS$SYSTEM:SYSGEN
SYSGEN> USE ACTIVE
SYSGEN> SET PFCDEFAULT 52
SYSGEN> WRITE ACTIVE
SYSGEN> EXIT
Modification of the active system does not affect the current system
image on disk.
If, for example, the system manager sets new active
parameter values (WRITE ACTIVE) and later wants to use these values
for subsequent bootstrap operations, the values must be explicitly
written to the current system image on disk:
$ RUN SYS$SYSTEM:SYSGEN
SYSGEN> USE ACTIVE
SYSGEN> WRITE CURRENT
SYSGEN> EXIT

13-18

SYSTEM GENERATION

13.2.2

Devices and Device Drivers

Normally, the system manager issues the AUTOCONFIGURE command to
automatically connect all devices physically attached to the system
and load their device drivers, saving a great deal of effort and
reducing the possibility of error. Devices not attached to the system
and devices with nonstandard names can be connected and their device
drivers loaded with explicit CONNECT (or CONNECT and LOAD) commands.
Great care must be exercised
in
issuing
CONNECT
and
LOAD
commands -- see the VAX/VMS Guide to Writing a Device Driver.
Devices not connected automatically by AUTOCONFIGURE include the
network communications logical device and the console block storage
device. Connecting the network communications logical device requires
the following explicit CONNECT command:
CONNECT NET /NOADAPTER /DRIVER=NETDRIVER
Connecting the console block storage
explicit CONNECT command:

device

requires

the

following

CONNECT CONSOLE
The following example autoconfigures the devices physically attached
to the system and explicitly connects the network software device and
the console block storage device:
$ RUN SYS$SYSTEM:SYSGEN
SYSGEN> AUTOCONFIGURE ALL
SYSGEN> CONNECT NET /NOADAPTER /DRIVER=NETDRIVER
SYSGEN> CONNECT CONSOLE
SYSGEN> EXIT
Normally, the SYSGEN commands for connecting devices and loading
device drivers are included in the site-independent start-up command
procedure (See Chapter 8).
Table 13-2 names the standard device types and their device drivers as
supplied
by
DIGITAL.
The
drivers
reside
in
files named
SYS$SYSTEM:file-name.EXE, except for DZDRIVER which
resides
in
SYS$SYSTEM:TTDRIVER.EXE.
Another DIGITAL-supplied driver named CONINTERR (which resides in
SYS$SYSTEM:CONINTERR.EXE) permits real-time processes to connect to
interrupt vectors for quick response to and special handling of
real-time events.
The driver is not associated with any one device
type. See the VAX/VMS Real-Time User's Guide for further information.

13-19

SYSTEM GENERATION
Table 13-2
Standard Device Types and Drivers
Name

Device Type

Driver

Card reader
Console block storage device
RP05, RP06 disk
RL02 cartridge disk
RK06, RK07 cartridge disk
RM03
RX02 floppy diskette
LPAll-K laboratory peripheral
accelerator
Line printer
Shared memory mailbox
TSll magnetic tape
TE16, TU45, TU77 magnetic tape
Network communications logical device
Operator's console
Remote terminal
Interactive terminal
DR32 interface adapter
DUPll synchronous communications line
DMCll synchronous communications line

CRDRIVER
DXDRIVER
DBDRIVER
DLDRIVER
DMDRIVER
DRDRIVER
DYDRIVER
LADRIVER

DB
DL
DM
DR
DY
LA
LP
MB
MS
MT
NET
OP
RT
TT
XF
XJ
XM

LPDRIVER
MBXDRIVER
TSDRIVER
TMDRIVER
NETDRIVER
OPERATOR
RTTDRIVER
DZ DRIVER
XFDRIVER

XMDRIVER

* Available under separate license

13.2.3

System Files

Table 13-3 provides suggested sizes for the paging, swapping, and dump
files for each hardware configuration. The full file specification of
each file is SYS$SYSTEM:file-name.type. Sizes are expressed in pages.
Table 13-3
Suggested Sizes for System Files
System File

MINIMUM 8USER 16USER 32USER 48USER 64USER VIRT32MB

PAG EFI LE • SYS
SWAPFILE.SYS
SYSDUMP.DMP

8192
3072
516
•••

,.•.•

-··-----~--~

16384 32768
7168 24576
2052 3076

8192
5120
1028

......... ,_ _ _ _ _ _ _ ,. ·-

"""~d·-~

...

-~·-

61440
47600
4100

98304
86016
6198

..·--..----~-·,.,---~----------------·---~----

13-20

98304
28672
2052

SYSTEM GENERATION
The VAX/VMS software distribution kit creates system files suitable
for systems configured with the MINIMUM and BUSER parameter files.
For the other configurations, or for variant configurations, the
system manager must specify the file sizes with the CREATE command of
the SYSGEN utility or (to create primary files only) with a DIGITAL
command
procedure
called SYS$SYSDISK: [SYSUPD)SWAPFILES.COM.
The
following example uses the CREATE command to create files suitable for
a system configured with the 16USER parameter file:
$ SET DEFAULT SYS$SYSTEM
$ RUN SYSGEN
SYSGEN> CREATE PAGEFILE.SYS /SIZE=l6384
SYSGEN> CREATE SWAPFILE.SYS /SIZE=7168
SYSGEN> CREATE SYSDUMP.DMP /SIZE=2052
SYSGEN> EXIT
The next example creates the same files using the command procedure:
$ @SYS$SYSDISK: [SYSUPD)SWAPFILES
To leave a file size at its current value type a carriage return
<CR> in response to its file size. Current file sizes are:
DIRECTORY DBO: [SYSEXE]
22-JAN-80 09:59
PAGEFILE.SYS;l
SWAPFILE.SYS;l
SYSDUMP.SYS;l

8192.
3072.
516.

c
c
c

21-JAN-80 09:23
21-JAN-80 09:23
21-JAN-80 09:23

TOTAL OF 11788./11788. BLOCKS IN 3. FILES
Enter new size for paging file: 16384
Enter new size for swapping file: 7168
Enter new size for system dump file: 2052
Use of the command procedure or the CREATE command results in a new
version of the system file.
The old version must be deleted
explicitly (but not until after the next bootstrap operation). In the
case of a primary file (PAGEFILE.SYS, SWAPFILE.SYS, or SYSDUMP.DMP),
the new file does not become effective until the system is shut down
and restarted. In the case of a secondary file, the new file becomes
effective when it is installed. No more than two versions of the
primary paging file (PAGEFILE.SYS) can exist at one time.
The system manager can
following calculations:
•

verify

the

suggested

sizes

through

the

Paging file -- Size of average program (in pages) times the
maximum number of processes (MAXPROCESSCNT system parameter).
Program size statistics can be gathered by looking at the peak
virtual sizes when users log off with the LOGOUT /FULL
command, or by examining user accounting records.
User
program sizes can be limited with the /PGFLQUOTA qualifier of
the ADD and MODIFY commands in the AUTHORIZE utility
(see
Chapter 2).
The suggested file sizes assume an average
program size of 1024 pages. The system manager ·should not
reduce the value of /PGFLQUOTA below 1024, nor reduce the size
of the paging file below the suggested value for
the
configuration, as sufficient space in the paging file is
critical to system performance. Size requirements can vary
depending on the user applications, as the paging file is used
only for copy-on-reference and demand-zero pages.

13-21

SYSTEM GENERATION

•

Swapping file -- Maximum number of processes
(MAXPROCESSCNT
system parameter) times maximum working set size (WSMAX system
parameter). The system manager should not specify a file size
less than this value, as the system will reduce the value of
MAXPROCESSCNT until MAXPROCESSCNT times WSMAX is equal to or
less than the size of the swapping file.

•

Dump file -- Size of physical memory in pages
(to save the
contents of memory if the system fails) plus 4 pages (to
provide continuity of the error log when the system is
shutdown or if the system fails).

At bootstrap time, the system activates the latest versions of
SYS$SYSTEM:PAGEFILE.SYS,
SWAPFILE.SYS,
and
SYSDUMP.DMP.
After
bootstrap, the system manager can replace or augment the primary
paging or swapping file with a secondary file (previously created with
a CREATE command) by issuing an INSTALL command.
The advantages to
using a secondary file are that it does not have to be on the system
disk and it can span volumes in a volume set. Where secondary files
are the norm, the system manager should:
•

Use the MINIMUM sizes for PAGEFILE.SYS and SWAPFILE.SYS

•

Include SYSGEN INSTALL commands for the secondary files in the
site-specific start-up command procedure (see Chapter 8)

The system recalculates the maximum allowable
value
MAXPROCESSCNT system parameter upon installation of a
swapping file.

for
the
secondary

All processes created after installation of the secondary paging file
use that file, while all processes created before its installation
continue to use the primary paging file. Swapping slots are allocated
(at process creation) from whichever swapping file has space, st~rting
with the primary.
Installation of a secondary paging file requires nonpaged dynamic
memory for a bit map, just as the primary paging file does, equal to
(in bytes) the number of pages in the paging file divided by 8.

13.2.4

Start-Up Command Procedure

Chapter 8 describes the site-independent start-up command procedure
supplied by DIGITAL -- SYS$SYSTEM:STARTUP.COM. The system manager can
create alternate site-independent start-up command procedures;
for
example, by copying STARTUP.COM to other files of type COM and editing
those files.

13-22

SYSTEM GENERATION
Following a bootstrap operation, the system executes the current
site-independent start-up command procedure. Initially (that is, as
supplied
in
the
software
distribution
kit),
the
current
site-independent start-up command procedure is STARTUP.COM.
The
system manager can name an alternate site-independent start-up command
procedure to be used during subsequent bootstrap operations with the
SET /STARTUP command in SYSGEN. The SHOW /STARTUP command displays
the
current
site-independent
start-up command procedure.
The
following example displays the current site-independent start-up
command procedure, then specifies an alternate:
$ RUN SYS$SYSTEM:SYSGEN
SYSGEN> USE CURRENT
SYSGEN> SHOW /STARTUP
Startup command file
SYS$SYSTEM:STARTUP.COM
SYSGEN> SET /STARTUP SYS$SYSTEM:XSTARTUP.COM
SYSGEN> WRITE CURRENT
SYSGEN> EXIT

13.2.5

Multiport (Shared) Memory

A single processor can attach up to four multiport memory units, each
of which may vary in size from 256KB to 2MB. The front panel of each
multiport memory unit displays the number of .that. unit.
When the
system manager issues a SHARE (initialize) command, SYSGEN polls the
processors with ports on the specified multiport memory unit.
If no
other processors are using the unit, the unit is initialized. If
another processor is using the unit, the unit is connected.
A SHARE
(connect) command for an uninitialized multiport memory unit results
in an error condition.
The system manager should power down a multiport memory unit only
after first shutting down and rebooting all systems connected to the
unit.
A multiport memory unit managed by VAX/VMS contains
structures with space requirements as indicated:

following

page

for

•

Pointer pages -- A common data page plus
active port

•

Global section description
(GSD)
table -- Total
global
sections, as specified in the SHARE command, times 96 bytes,
rounded up to the next full page

•

Mailbox table -- Total mailboxes, as specified in the SHARE
command, times 48 bytes, rounded up to the next full page

•

CEF table -- Total common event flag clusters, as specified in
the SHARE command, times 80 bytes, rounded up to the next full
page

•

PRQ pool -- Total interprocess or request
messages,
as
specified in the SHARE command, times 64 bytes, rounded up to
the next full page

13-23

one

the

each

SYSTEM GENERATION
•

Dynamic pool -- Number of blocks allocated to the pool, as
specified in the SHARE command, times the size of each block,
as specified in the SHARE command

•

Non-VAX/VMS structures -- Relative number of first page in
multiport memory for use by shared structures, as specified in
the SHARE command, minus 1

•

Global section bit map -- One page

•

Global sections -- Size of multiport memory unit minus the sum
of the above preallocated structures (that is, the remaining
space)

Figure 13-4 illustrates the appearance of a multiport memory unit
where the SHARE command specifies the default structure values,
assuming a 256KB unit with four active ports.
SYSGEN> SHARE MPM1 SHR_MEM __ 1 /INIT/GBLSECTIONS 32
/MAILBOXES 32/CEFCLUSTERS 32 /POOLBCNT 128/POOLBSIZE 128
/PRQCOUNT 64
POINTER PAGES

1 t 4

pages

DYNAMIC
POOL
_____________

128 x 128

pages

GSD TABLE

32 x 96

page

PRO POOL

---- ----··-·-- - .... ---·· ·---·---- --·
,,_

-~--·-~-~~~--~-~~.

- --~e-~•-·=~~=·e'-~"'"'~'

MAILBOX TABLE

32 x 48

page

CEF TABLE

32 x 8

page

BIT MAP
GLOBAL SECTIONS

Figure 13-4

1 page

512 - 49

463 pages

relative
- - - page 1

Example of Multiport Memory Structures

The following guidelines are suggested in selecting values for the
SHARE
qualifiers
that regulate the sizes of the preallocated
structures:
e

/CEFCLUSTERS, /GBLSECTIONS, and
/MAILBOXES -- The
system
manager should simply specify the maximum number of each type
of structure required by all processors at any one time.
The
same structure being used by many processes on one or more
processors counts as just one structure.

13-24

SYSTEM GENERATION
•

/POOLBCNT and /POOLBSIZE -- The primary use of the dynamic
pool is to buffer mailbox messages. The size of a message is
28 bytes plus the data in the message. Since space from the
pool is always allocated in whole blocks, the recommended
block size is the median message size plus 28. A block size
that is too small for a message requires extra system overhead
to concatenate the message blocks into the user buffer and
segment them out of the user buffer. The number of blocks
should be sufficient to satisfy all messages that might be
outstanding at once. If a mailbox request cannot be satisfied
due to insufficient pool space, the requesting process enters
a resource wait state or the request fails (if resource wait
mode is not enabled), just as if the nonpaged dynamic pool
were depleted.
For this reason, the system manager should
tend to overestimate space requirements in the dynamic pool.

•

/PRQCOUNT -- The system uses interprocessor request blocks
internally to transfer requests among the VAX/VMS executive
routines and mailbox drivers on the different processors.
PRQs are allocated and deallocated rapidly, so that a large
number should not be needed.
The default value normally
suffices.
If an executive or driver request cannot be
satisfied due to depletion of the PRQs, the requesting routine
waits until a PRQ becomes available.

The system manager should calculate the space remaining for the global
sections and be satisfied that it is sufficient. If the space is
insufficient, the system manager might reduce the size of the dynamic
pool.
However, this condition really suggests the need for a larger
or additional multiport memory unit.
Where a multiport memory unit is a normal part of the system
configuration, the system manager should include the SYSGEN commands
to initialize and connect it in the site-specific start-up command
procedure (see Section 8.2.9).

13-25

CHAPTER 14
DISPLAY UTILITY

The system manager runs the DISPLAY utility to display information on
system performance. Table 14-1 summarizes the DISPLAY commands. Use
of the utility is restricted to VTS2, VTSS, and VTlOO terminals.
(VTlOO terminals automatically operate in VTS2 emulation mode.) Print
files (for example, assigning a file to SYS$0UTPUT) are not permitted.
Table 14-1
DISPLAY Command Summary
Format

Function

!-------------~-~--~ ---------~- --·~--·-------!

Returns the user to DISPLAY command level

CYCLE

Produces all displays cyclically, except for
PRIMITIVE STATISTICS display

EXIT

Returns the user to DCL command level

FCP

Produces the FILE PRIMITIVE STATISTICS display

HELP

Lists the DISPLAY commands

I ORATES

Produces the I/O SYSTEM RATES display

Produces the TIME IN PROCESSOR MODES display

Produces the TIME IN PROCESSOR MODES display as
graph (VTSS terminals only)

PAGE

Produces the PAGE MANAGEMENT STATISTICS display

POOL

Produces the NONPAGED POOL STATISTICS display

Produces the NUMBER OF PROCESSES
display

SCHEDULER

STATES

Produces the NUMBER OF PROCESSES IN SCHEDULER
display as a bar graph (VTSS terminals only)

STATES

TOPUSERS

Produces the TOP CPU TIME USERS display

USERS

Produces the VAX/VMS PROCESSES display

14-1

the

FILE

bar

DISPLAY UTILITY

DISPLAY
14.1

INVOKING DISPLAY

The following command invokes the utility:
RUN SYS$SYSTEM:DISPLAY

The system responds with the following prompt:
DISPLAY NAME?
The system manager can then enter any of the commands listed in Table
14-1.
The command names can be abbreviated to their first two
characters (or their first characters if the abbreviation is unique).
If the command is a request for a display (all commands except
<CTRL/C>, EXIT, and HELP), the system responds with the following
prompt:
UPDATE INTERVAL IN SECONDS?
The system manager responds with a positive integer to represent the
number of seconds over which statistics for the display are gathered.
At the end of each interval, the system posts the results to the
display.
For example, if the system manager specifies an interval of
10 seconds, the system:
1.

Displays the headings of the display

Collects statistics for 10 seconds

Posts the results to the display

Collects statistics for another 10 seconds

Posts the new results to the display

And so on until the system manager types <CTRL/C>

A carriage return results in the default interval of 3 seconds.
CYCLE command is an exception; see Section 14.1.3).

(The

The top of each display contains
the
time
in
the
hours:minutes:seconds that the last statistics were posted.

format

14.2

CTRL/C COMMAND

Typing <CTRL/C> (that is, pressing the CTRL key and, while holding it
down, typing C) terminates the current display and returns the user to
DISPLAY command level. The screen is cleared and the prompt DISPLAY
NAME? appears.

14-2

DISPLAY UTILITY

DISPLAY (Cont.)

14.3

CYCLE COMMAND

CYCLE cyclically
indicated:

produces

the

following

displays

the

order

TIME IN PROCESSOR MODES (VT55 terminals only)

VAX/VMS PROCESSES

TOP CPU TIME USERS

PAGE MANAGEMENT STATISTICS

NUMBER OF PROCESSES IN SCHEDULER STATES (VT55 terminals only)

I/O SYSTEM RATES

TIME IN PROCESSOR MODES

NONPAGED POOL STATISTICS

NUMBER OF PROCESSES IN SCHEDULER STATES

The command takes the form:
CYCLE
In the case of the CYCLE command, the response to the interval prompt
means the number of seconds between different displays, and defaults
to 30. The update interval is fixed at 3 seconds.
The following example produces the
between each display:
DISPLAY NAME?

CYCLE

with

seconds

CYCLE

INTERVAL BETWEEN DISPLAYS IN SECONDS?

14.4

displays

EXIT COMMAND

EXIT returns the system manager to DCL command level.
form:

takes

the

takes

the

EXIT

14.5

FCP COMMAND

FCP produces the FILE PRIMITIVE
form:

STATISTICS

FCP

14-3

display.

DISPLAY UTILITY

DISPLAY (Cont.)
The FILE PRIMITIVE STATISTICS display
measures
activities for all Files-11 ACPs on the system:

the

following

•

FCP CALLS -- QIO requests received by the system

•

ALLOCATIONS -- Calls that caused allocation of disk space

•

CREATES -- New files created

•

DISK READS -- Read I/O operations from disk by the file system

•

DISK WRITES -- Write I/O operations to disk by the file system

•

CACHE HITS -- Requested blocks
cache

•

CPU TICS -- CPU time used by the file system in 10 millisecond
tics

•

WINDOW TURNS -- File mapping window misses

FILE LOOKUPS -- File
• directories

located

the

file

system

file

during

the

•

RATE/SEC -- Average number of times the activity occurred
second during the preceding interval

per

•

AVG RATE -- Average number of times the activity occurred
second since the display began

per

•

name

look-up

operations

FILE OPENS -- Files that were opened

The display provides three measurements for each activity:
-- Number of times
• VALUE
preceding interval

The following example produces a
with an interval of 20 seconds:
DISPLAY NAME?

the

FILE

activity

PRIMITIVE

occurred

STATISTICS

display

FCP

UPDATE INTERVAL IN SECONDS?

Normally, a lengthy interval provides better results, because the
sample is larger. If a short interval is used, the system manager can
allow several intervals to pass and watch the AVG RATE statistics.
Figure 14-1 illustrates a FILE PRIMITIVE STATISTICS display.
From
this display, the system manager might, for example, learn that over
the past interval the system performed 4 read operations from disk for
an average rate of .97 reads per second during the interval and .85
reads per second since the display started.

14-4

DISPLAY UTILITY

DISPLAY (Cont.)

FILE PRIMITil,IE STAT I ST I.CS
1£1:27:27

l,IALUE

RATE
/SEC

Al.JG
RATE

CACHE HITS

1. 9ll

2.87

0. Oll

CPU TICS

2. ll3

2.85

o.oo

0.01

WINDOW TURNS

0.2ll

1. 01

0.97

0.85

FILE LOOI\ UPS

1. 70

2. Gll

o.oo

0. 15

FILE OPENS

0. 2ll

0.97

l,IALUE

RATE
/SEC

Al.JG
RATE

NAME

FCP CALLS

1. 70

3.92

ALLOCATIONS

o.oo

CREATES

DISi\ READS
DIS!\ WRITES

NAME

Figure 14-1

14.6

FILE PRIMITIVE STATISTICS Display

HELP COMMAND

HELP lists and explains the DISPLAY commands.

It takes the form:

HELP

14.7

!ORATES COMMAND

!ORATES produces the I/O SYSTEM RATES display.

It takes the form:

!ORATES
The I/O SYSTEM RATES display measures the following activities:
•

DIRECT I/Os -- Direct
operations

I/O

(for

•

BUFFERED I/Os -- Buffered
printer) I/O operations

•

MAILBOX WRITES -- Write-to-mailbox requests
system

•

WINDOW TURNS -- File mapping window misses

•

LOGNAME TRANS -- Logical name translations

(for

14-5

example,
example,

disk

tape)

and

terminal
received

and
by

line
the

DISPLAY UTILITY

DISPLAY (Cont.)
•

FILE OPENS -- Files opened

•

PAGE FAULTS -- Page faults for all working sets

•

PAGES READ -- Pages read from disk as a result of page faults

•

READ I/Os -- Read I/O operations from disk as a result of page
faults

•

PAGES WRITTEN -- Pages written to the paging file

•

WRITE I/Os -- Write I/O operations to the paging file

•

TOTAL INSWAPS -- Working
swapping file

sets

read

into

memory

from

the

The display provides three measurements for each activity:
•

VALUE -- Number of times
preceding interval

the

activity

occurred

during

the

•

RATE/SEC -- Average number of times the activity occurred
second during the preceding interval

per

•

AVG RATE -- Average number of times the activity occurred
second since the display began

per

The top corners of the display contain the number of pages on the free
page and modified page lists at post time.
The following example produces an I/O SYSTEM
interval of 20 seconds:
DISPLAY NAME?

RATES

display

with

!ORATES

UPDATE INTERVAL IN SECONDS?

Normally, a lengthy interval provides better results, because the
sample is larger. If a short interval is used, the system mqnager can
allow several intervals to pass and watch the AVG RATE statistics.
Figure 14-2 illustrates an I/O SYSTEM RATES display.
From this
display, the system manager might, for example, learn that over the
past interval 65 page faults occurred, resulting in 4 pages being read
from disk in 2 I/O operations.
The system manager can calculate further statistics from this display:
•

Percentage of page faults resulting in a disk I/O operation
READ I/Os divided by PAGE FAULTS

•

Average cluster size for reading pages into
PAGES READ divided by READ I/Os

•

Average cluster size for writing pages from modified page list
-- PAGES WRITTEN divided by WRITE I/Os

•

Total swapping I/O operations -- TOTAL INSWAPS times 2

•

Disk I/O rate -- DIRECT I/Os plus READ I/Os plus WRITE I/Os
plus total swapping I/O operations (TOTAL INSWAPS times 2)
allowing for any DIRECT I/Os to tape
14-6

working

sets

DISPLAY UTILITY

DISPLAY {Cont.)

FREE LIST: 201 G

I/O SYSTEM RATES

MODIFY LIST: 35

lG: 17: 09

1,JALUE

RATE
/SEC

Al.JG
RATE

NAME

DIRE CT I/Os

7.30

1.50

PAGE FAULTS

BUFFERED I/Os

G+G2

3+24

MA I LBO>( WRITES

o.oo

WINDOW TURNS

LOGNAME TRANS
FILE OPENS

NAME

Al,JG

RATE
/SEC

RATE

14.84

1. 83

PAGES READ

0.91

0 +11

o.oo

READ I/Os

(I+ 45

0.07

O.G8

0 +14

PAGES WR ITT EN

o.oo

0. 00

8+90

0 +98

WRITE I/Os

o.oo

o. 00

O+G8

0.07

TOTAL INSWAPS

o.oo

1.JALUE

Figure 14-2 I/O SYSTEM RATES Display

14.8

M2 AND MS COMMANDS

M2 and MS produce the TIME IN PROCESSOR MODES display.
The MS command
produces the display as a bar graph;
its use is restricted to VTSS
terminals.
The M2 and MS commands take the forms, respectively:
M2
MS
The display measures the following activities as
total CPU time used by the system:
STACK -- Time

interrupt

the

TIME ON INTERRUPT
(INTER time)

•

TIME IN KERNEL MODE -- Time spent in kernel mode, but
interrupt stack (KERNEL time)

•

TIME IN EXEC MODE -- Time spent in executive mode

{EXEC time)

•

TIME IN SUPER MODE -- Time spent
time)

mode

•

TIME
IN USER MODE -- Time spent
executing compatibility-mode images

14-7

spent

percentages

supervisor

in user mode,
(USER time)

stack
not

(SUPER
but

not

DISPLAY UTILITY

DISPLAY (Cont.)
•

TIME
IN
COMPATIBILITY
MODE -- Time
compatibility-mode images (COMPAT time)

•

IDLE TIME -- Time spent executing the NULL process

spent

executing

The system manager can calculate the approximate time spent executing
system
programs by adding the INTER, KERNEL, EXEC, and SUPER
percentages; and the approximate time spent executing user programs
by adding the USER and COMPAT percentages. User programs include the
DIGITAL editors, compilers, linker, and utilities.
The following example produces a TIME IN PROCESSOR MODES display
an interval of 20 seconds:
DISPLAY NAME?

with

UPDATE INTERVAL IN SECONDS?
Normally, a lengthy interval
sample is larger.

provides

better

results,

because

the

Figure 14-3 illustrates a TIME IN PROCESSOR MODES display in standard
format.
Figure 14-4 illustrates a TIME IN PROCESSOR MODES display as
a bar graph (VT55 terminals only). From the standard display, the
system manager might, for example, see that system programs used about
38 percent of the system's CPU time and user programs used about 25
percent.

TIME IN PROCESSOR MODESC%1
14:30:35
0

100

+ -

- +

+ -

TIME ON INTERRUPT STACK
TIME IN KERNEL MODE
TIME IN EXEC MODE
TIME IN SUPER MODE
TIME IN USER MODE
TIME IN COMPATIBILITY MODE
IDLE TIME

Figure 14-3 TIME IN PROCESSOR MODES Display (M2 Command)

14-8

DISPLAY UTILITY

DISPLAY (Cont.)

TIME

PROCESSOR

MODES ( 'X,)

14:22:23
100
90

80
70
GO

50
40
30

-20
10

INTER

KERNEL

E}(EC

SUPER

USER

COM PAT

IDLE

Figure 14-4 TIME IN PROCESSOR MODES Display (M5 Command)

PAGE COMMAND

14.9

PAGE produces the PAGE MANAGEMENT STATISTICS display.
form:

takes

the

following

PAGE
The PAGE MANAGEMENT
activities:

STATISTICS

display

measures

•

PAGE FAULTS -- Page faults for all working sets

•

PAGES READ -- Pages read from disk as a result of page faults

•

READ I/Os -- Read I/O operations from disk as a result of page
faults

14-9

DISPLAY UTILITY

DISPLAY (Cont.)
•

PAGES WRITTEN -- Pages written to the paging file

•

WRITE I/Os -- Write I/O operations to the paging file

•

FREE LIST -- Pages read from the free page list as a result of
page faults

•

MODIFIED LIST -- Pages read from the modified page list
result of page faults

•

DEMAND ZERO -- Zero-filled pages allocated as a result of page
faults

•

WRITE IN PROG -- Pages read that were in the process of
written back to disk, when faulted

•

SYSTEM FAULTS -- Page faults for pages in system space

being

The display provides three measurements for each activity:
-- Number of times
• VALUE
preceding interval

during

the

RATE/SEC -- Average number of times the activity occurred
second during the preceding interval

per

RATE -- Average number of times the activity occurred
• AVG
second since the display began

per

•

the

activity

occurred

The top corners of the display contain the number of pages on the free
page and modified page lists at post time.
The following example produces a PAGE
with an interval of 20 seconds:
DISPLAY NAME?

MANAGEMENT

STATISTICS

display

PAGE

UPDATE INTERVAL IN SECONDS?

Normally, a lengthy interval provides better results, because the
sample is larger. If a short interval is used, the system manager can
allow several intervals to pass and watch the AVG RATE statistics.
Figure 14-5 illustrates a PAGE MANAGEMENT STATISTICS display.
From
this display, the system manager might, for example, learn that over
the past interval 50 page faults occurred, resulting in 9 pages being
read from disk in 3 I/O operations.

14-10

DISPLAY UTILITY

DISPLAY (Cont.)

FREE LIST: 181

PAGE MANAGEMENT STATISTICS

MODIFY LIST: 19

14: 25: 19

NAME

l.IALUE

RATE
/SEC

Al.IG
RATE

NAME

l,IALUE

Al,IG

RATE
/SEC

RATE

0.23

0. 10

PAGE FAULTS

11. 96

8. 04

FREE LIST

PAGES READ

2. 15

2.08

MODIFIED LIST

0.71

1. 55

READ I/Os

o. 71

0. 46

DEMAND ZERO

1. 91

2. 44

PAGES WRITTEN

o.oo

0. 42

WRITE IN PROG

o. 00

o.oo

WRITE I/Os

o. 00

(l t 02

SYSTEM FAULTS

o.oo

Figure 14-5 PAGE MANAGEMENT STATISTICS Display

14.10

POOL COMMAND

POOL produces the NONPAGED POOL
form:

STATISTICS

display.

takes

the

POOL
The NONPAGED POOL STATISTICS display measures space allocations in the
nonpaged dynamic pool:
•

I/O REQUEST PACKETS LEFT -- I/O request packets not in use

•

NUMBER OF HOLES IN POOL -- Unused blocks of contiguous
in the dynamically allocated portion of the pool

•

TOTAL SPACE LEFT -- Total unused
allocated portion of the pool

•

LARGEST BLOCK -- Size in bytes of the largest block of
space in the dynamically allocated portion of the pool

•

SMALLEST BLOCK -- Size in bytes of the smallest block of
unused space in the dynamically allocated portion of the pool

•

NUMBER OF BLOCKS LEQ 32 BYTES -- Blocks less than or equal to
32 bytes in size in the dynamically allocated portion of the
pool

14-11

bytes

the

space

dynamically
unused

DISPLAY UTILITY

DISPLAY (Cont.)
The display provides two measurements for each space allocation:
•

VALUE -- Amount of space at the end of the preceding interval

•

AVG -- Average amount of space since the display began

The following example produces a NONPAGED POOL STATISTICS display with
an interval of 8 seconds:
DISPLAY NAME?

POOL

UPDATE INTERVAL IN SECONDS?

The system manager should set the update interval as short as possible
without impeding readability to obtain many pictures of the pool at
different times.
Figure 14-6 illustrates a NONPAGED POOL STATISTICS display. From this
display, the system manager might, for example, learn that the dynamic
portion of the pool has 12,048 bytes of unused space in 27 blocks, but
that most of the space is in one contiguous 11,136-byte block.

NONPAGED POOL STATISTICS
lll: 28: 26

t.JALUE

At.JG

I/O REQUEST PACKETS LEFT

BL.I

GL.1.00

NUMBER OF HOLES IN POOL

27.00

TOTAL SPACE LEFT

12048

12048.00

LARGEST BLOCK

11136

11136.00

SMALLEST BLOCK

16.00

NUMBER OF BLOCKS LEQ 32 BYTES

18.00

--------------·

'~~-~-·-,·-··-"'·'"

Figure 14-6 NONPAGED POOL STATISTICS Display

14.11

S2 AND SS COMMANDS

S2 and SS produce the NUMBER OF PROCESSES IN SCHEDULER STATES display.
The SS command produces the display as a bar graph;
its use is
restricted to VTSS terminals. The S2 and SS commands take the forms,
respectively:
S2
SS
14-12

DISPLAY UTILITY

DISPLAY (Cont.)
The display shows the number of processes in each of the 14
states:
•

COLLIDED PAGE WAIT (CPG) -- Waiting
transition

for

scheduler

faulted

page

AND MISC •
RESOURCE WAIT
(MWT) -- Waiting for
the
• MUTEX
availability of a mutual exclusion
(mutex) semaphore or a
dynamic resource
e

COMMON EVENT FLAG WAIT (CEF) -- Waiting for
some· combination
of event flags to be set in a common event block

PAGE FAULT WAIT (PFW) -- Waiting for a page to be
result of a page fault;
resident processes

LOCAL EVENT FLAG WAIT (LEF) -- Waiting for some combination of
local event flags to be posted;
resident processes

•

LOCAL EVENT FLAG (OUT OF
some
combination
of
outswapped processes

•

HIBERNATE WAIT
(HIB) -- Waiting
resident processes

•

HIBERNATE WAIT (OUT OF BALANCE SET)
(HBO) -- Waiting
hibernate request;
outswapped processes

SUSPENDED WAIT
(SSP) -- Waiting
resident processes

•

SUSPENDED WAIT (OUT OF BALANCE SET)
(SPO) -- Wating
suspend request;
outswapped processes

•

FREE PAGE WAIT (FPG) -- Waiting for a free page of memory

COMPUTE
(COM}
processes

COMPUTE
(OUT OF BALANCE SET)
processor;
outswapped processes

CURRENT PROCESS (CUR) -- Using the processor

Ready

BALANCE SET)
local event

use

for

the

(LFO)
flags
a

read

request;

suspend

for

request;
for

resident

processor;

(CMO) -- Ready

Waiting for
be posted;

hibernate

use

the

The CUR process
is the one currently executing.
LFO processes
normally belong to interactive users wpo are thinking, although they
might be processes waiting for disk I/O on a crowded system.
LEF
processes are normally interactive users who are thinking or processes
waiting on a disk I/O. A state of CMO for any process indicates a
very crowded system and one on which performance (interactive response
time and batch throughput, for example) is poor.
The following example produces a NUMBER OF PROCESSES
STATES display with an interval of 8 seconds:
DISPLAY NAME?

UPDATE INTERVAL IN SECONDS?

14-13

SCHEDULER

DISPLAY UTILITY

DISPLAY (Cont.)
The system manager should set the update interval as short as possible
without impeding readability to obtain many pictures of the processes
at different times.
Figure 14-7 illustrates a NUMBER OF PROCESSES IN SCHEDULER STATES
display in standard format.
Figure 14-8 illustrates a NUMBER OF
PROCESSES IN SCHEDULER STATES display as a bar graph
(VT55 terminals
only).
From the standard display, the system manager might, for
example, learn that 8 resident processes are ready to use the
processor.

NUMBER OF PROCESSES IN SCHEDULER STATES
1ll:29: l!O
0

COLLIDED PAGE WAIT
MUTEX AND MISC. RESOURCE WAIT
COMMON EVENT FLAG WAIT
PAGE FAULT WA IT
LOCAL EVENT FLAG WAIT
LOCAL El.JENT FLAG <OUT OF BALANCE SET>
HIBERNATE WAIT
HIBERNATE WAIT <OUT OF BALANCE SET)
SUSPENDED WAIT
SUSPENDED WAIT <OUT OF BALANCE SET>
FREE PAGE WAIT
COMPUTE
COMPUTE <OUT OF BALANCE SET>
CURRENT PROCESS

l!O

+ -

- +

••••••111111111111111111
11111111111111111111111111

Figure 14-7

20
-

NUMBER OF PROCESSES IN SCHEDULER STATES Display (S2 Command)

14-14

DISPLAY UTILITY

DISPLAY (Cont.)

NUMBER OF PROCESSES IN SCHEDULER STATES
14:23:43
20
18

14
12
10
8

- - - - - - - - --- --- - - - - - - - - - - - - - - - - - - - - - - · · - CPG

MWT

CEF

PFW

LEF

LFO

HIB

HBO

SSP

SPO

FPG

COM

CMO

CUR

Figure 14-8 NUMBER OF PROCESSES IN SCHEDULER STATES Display (S5 Command)

14.12

TOPUSERS COMMAND

TOPUSERS produces the TOP CPU TIME USERS display.

It takes the form:

TOPUSERS
The TOP CPU TIME USERS display measures the percentage of the total
CPU time taken by each user during the preceding interval. Users are
identifed by UIC and name.
Processes not in memory during the
interval do not participate in the display.
The following example produces a TOP CPU TIME USERS
interval of 8 seconds:
DISPLAY NAME?

display

with

TOPUSERS

UPDATE INTERVAL IN SECONDS?

The system manager should set the updat~ interval as short as possible
without impeding readability to obtain many pictures of the processes
at different times.

14-15

DISPLAY UTILITY

DISPLAY (Cont.)
Figure 14-9 illustrates a TOP CPU TIME USERS display.
From this
display, the system manager might, for example, learn that user QAl
used over 25 percent of the total CPU time during the preceding
interval.

TOP CPU TIME USERS!%)
1G:15: 41

[320r001]

QAl

[3201001]

_JOB1255

[1101010]

ROGER

[0011002]

DBA3ACP

[2221020]

_JOB125G

[0011002]

DBA2ACP

[0071022]

LI NDAP

[2651010]

DISPLAY

Figure 14-9

14.13

+ -

100
-

TOP CPU TIME USERS Display

USERS COMMAND

USERS produces the VAX/VMS PROCESSES display.

It takes the form:

USERS
For each process on the system during the preceding interval,
VAX/VMS PROCESSES display provides the following information:
•

PID -- Process identification as assigned by
hexadecimal

•

UIC -- User identification code, in octal

•

STATE -- Process's scheduler state (some abbreviations are
longer than those used in the NUMBER OF PROCESSES IN SCHEDULER
STATES display; COMO instead of CMO and LEFO instead of LFO,
for example)

14-16

the

system,

the
in

DISPLAY UTILITY

DISPLAY (Cont.)
•

PRI -- Current (as opposed to base) priority of the process

NAME

•

Number of shareable pages and total number of pages in
SIZE
use by the process

•

DIOCNT -- Cumulative direct I/O operations performed by the
process since its start; not displayed if process is swapped
out

•

FAULTS -- Cumulative page faults since the
not displayed if process is swapped out

•

CPU
TIME -- Cumulative
CPU
time
in
the
format
hours:minutes:seconds used by the process since its start;
not displayed if process is swapped out

Process name

process

started;

The top corners of the display contain the number of processes on the
system and the time in days, hours, and minutes since the system was
last booted. Processes that are swapped out are so noted.
If more processes are on the system than can be displayed on the
terminal screen, the system segments the list of processes and
displays each segment at the user-specified interval.
The following example produces a VAX/VMS
interval of 10 seconds:
DISPLAY NAME?

PROCESSES

display

with

USERS

UPDATE INTERVAL IN SECONDS?

The system manager should set an update interval as short as possible
without impeding readability to obtain many pictures of the system at
different times.
Figure 14-10 illustrates a VAX/VMS PROCESSES display.
From this
display, the system manager might, for example, learn that user
DAVIDP, waiting on a local event flag, has been running for quite a
while and doing a fair amount of disk I/O.

14-17

DISPLAY UTILITY

DISPLAY (Cont.)

t.JA){/t.JMS PROCESSES

PROCESSES: 37

UPTIME:

300:48

14:20: 12
PIO

UIC

STATE

PRI

NAME

SIZE

001D002C
0017002D

[1011010]

LEF

[2011011]

LEFO

0009002E

[1051020]

COM

0031002F

[0111010]

LEF

00130030

[2511020]

00080031

[010 1040]

LEF
LEFO

0 IDCNT

FAULTS

CPU TIME

ANDREW

RMSTST

Gil 100
14/43

1824

3370

00:00:35.10

RICHARDP

29/150

LI NDAP

14 I 100

8378

4778

00:01:57.99

18/128

12128

8254

00:03:57.25

JAMESP
_TTCG:

SWAPPED OUT
4498
3589 00:02:39.29

0/32

00170032

[0071007]

LEFO

HANKP

20/48

00010033

[0011002]

HIB

DBA2ACP

31171

00350034

[3801015]

LEFO

JONES

15/44

00220035

[2211020]

LEF

KATHYP

31/102

00010038

[0071377]

LEFO

3/35

SWAPPED OUT
SWAPPED OUT
47

99114

00:23:04.82

SWAPPED OUT
9044

9734

00:03:38.99

00150037

[2411010]

LEF

SYSTEM
WILLI AM_B

SWAPPED OUT

38/170

20771

8904

00:08:15.74

00010038

[0011002]

LEF

OBAlACP

31171

44381

00:05:24.83

00020039

[0011002]

HIB

PRTSYMB2

8/31

20470

00:11:40.82

OOOB003A

[010 1010]

LEF

OAt.JIOP

37/114

452875

212280

14:51:45.25

00010035

[ 001 100G J

HIB

ERRFMT

0/2G

900

00:00:08.25

0001003C

[0021001]

LEFO

OPERATOR

0/30

SWAPPED OUT

Figure 14-10 VAX/VMS PROCESSES Display

14.14

ERROR MESSAGES

Table 14-2 lists the error messages issued by the DISPLAY utility
alphabetical order).

(in

Table 14-2
DISPLAY Error Messages
.---------·---·············-·-·-·-········· ··----· ·-··--·-··--··········---············· -· ···-- ·-········-··- ---··· ... ····-------.-------- --··-..··-- - --····--------····-··-·-· ..-·-· - - - - - - - - - -

Message
1---------------·····--·-----·-·-··--···-·---·

Remedy
-v,..,, __ ._ ·~-··-----·-

-+---·----- - - - - -

COMMAND NOT UNIQUE

Spell out the command name more.

INVALID DISPLAY NAME, TYPE "HELP"
FOR HELP

Enter a valid command.

THIS DISPLAY FOR VTSS's ONLY -TERMINAL IS NOT A VTSS

Informational message
respond.

-- do not

Where correction of an entered value is indicated, the entire
must be reentered.

14-18

command

INDEX

A
$ACCDEF module in SYS$LIBRARY:LIB.
MLB library, 4-15
Access,
delete, execute, read, and write,
3-3
file, 3-2 (figure)
group, 4-9
logical and physical I/O, 3-3
physical page, 4-11
/ACCESSED, 5-6
/ACCOUNT, 2-3
Accounting, 4-15
disable, 4-7
log file, 4-15
Account name, 2-3, 4-15
ACCOUNTNG.DAT, 4-15
ACNT privilege, 4-7
ACP, see Ancillary control process
ACP system parameters, 12-6 (table)
12-22
Active system parameter values,
12-1, 13-17
modify, 13-18
use, 13-15
write, 13-16
/ADAPTER, 13-5
ADD command,
in AUTHORIZE, 2-7
in DISKQUOTA, ~-3
Add entry to disk quota file, 6-3
Add entry to user authorization
file, 2-7
Adjustment of working set limit,
automatic, 11-11, 12-16
/ALL, 13-13
Allocate,
card reader, 3-5, 3-8
devices, 4-7
terminal, 3-5, 3-8
ALLSPOOL privilege, 4-7
Alter base priority, 4-7
Alternate user authorization file,
.2-18, 12-20
ALTPRI privilege, 4-7
Analyzer, system dump, 8-6
Ancillary control process, 5-8
cluster size, 12-24
multiple, 11-17, 12-22
priority, 12-24
read verification, 12-24
shared, 12-22
swapping, 12-24
window pointers, 12-24
working set size, 12-24
write verification, 12-24

Assign,
logical names in site-independent
start-up command procedure,
8-2
logical names in site-specific
start-up command procedure,
8-5
logical print queue, 9-10
/ASTLM, 2-3
AST queue limit, 2-3, 4-2
Asynchronous write errors, 10-1
Authorization file, see User
authorization file
AUTHORIZE utility, 2-2
ADD command, 2-7
commands, 2-2 (table)
DEFAULT command, 2-9
EXIT command, 2-10
HELP command, 2-10
invoking, 2-7
LIST command, 2-10
messages, 2-14 (table)
MODIFY command, 2-11
operations, 2-17
qualifiers 2-3 (table)
REMOVE command, 2-12
SHOW command, 2-12
SHOW display, 2-13 (figure)
AUTOCONFIGURE command of SYSGEN,
13-4
Automatic,
adjust working set limit, 11-11,
12-16
connect devices and load drivers,
13-4
reboot on fatal bugcheck,
12-19
rebuild disk quota file, 5-10
AWSMAX system parameter, 11-11,
12-17
AWSMIN system parameter, 11-11,
12-17
AWSTIME system parameter, 11-11,
12-17

B
Back-up,
home block 5-2
index file header, 5-3
log file (BACKUP.SYS), 5-4
of public volumes, 5-10
BACKUP.SYS (back-up log file), 5-4
BADBLK.SYS (bad block file), 5-3
BADLOG.SYS (pending bad block log
file), 5-4

Index-1

INDEX

Bad block,
f i 1 e (BAD BL K • SYS ) , 5- 3
log file, pending (BADLOG.SYS),
5-4
Balance set, 4-12
slots, 12-10
BALSETCNT system parameter, 12-10
Base priority, 4-4
alter, 4-7
Batch job, 9-1, 11-3
command procedure as, 9-4
versus interactive job, 9-6
Batch queue,
create, 9-5
delete, 9-6
guidelines, 9-6
initialize, 9-5
start, 9-5
stop, 9-6
/BIND, 5-7
/BIOLM, 2-3
Bit map,
cache, 11-17, 12-23
index file bit map, 5-3
storage bit map file
(BITMAP.SYS), 5-3
BITMAP.SYS (storage bit map file),
5-3
BJOBLIM system parameter, 12-22
Block storage device
console, 13-6
Blocking factor, record management
services, 12-25
Bootstrap,
block, 5-2
operation, 13-17
Brief,
known image display, 7-6 (figure)
SYE report, 10-1
/BRIEF qualifier to SHOW, 2-12
Buffer,
quota for mailboxes, 12-18
record management services,
11-15
terminal type-ahead, 12-21
Buffered,
I/O byte limit, 2-3, 4-2
I/O count limit, 2-3, 4-2
I/O transfer limit, 12-18
Bugcheck,
all bugchecks fatal, 12-19
automatic reboot on fatal,
12-19
dump file on fatal, 12-19
BUGCHECKFATAL system parameter,
12-19
BUGCHK privilege, 4-8
BUGREBOOT system parameter, 12-19
BYPASS privilege, 4-8

Bypass protection, 4-8
/BYTLM, 2-3

c
Caches, Files-11, 5-8, 11-ln,
12-23
bit map, 11-17, 12-23
directory, 11-17, 12-23
disk quota, 11-17, 12-23
extent, 11-17, 12-23
file control block, 12-23
file header, 12-23
file identification, 11-17,
12-23
header, 11-16
map, 11-17
page, 11-5, 11-12
system directory, 11-17
Capacity, disk, 11-2
Card reader, 9-3
allocation and protection, 3-5,
3-8
/CEFCLUSTERS, 13-11, 13-24
/CHARACTERISTICS, 9-12
Characteristics,
printer, 9-12
terminal, 12-21
CHARTYPE.DAT, 9-12
Checks, range, 13-7
/CL!, 2-3
CLISYMTBL system parameter, 12-19
/CLUSTER, 5-6
Cluster,
common event flag, 4-11
size, 11-13, 12-8, 12-14
size, ancillary control process,
12-24
CMFEXEC privilege, 4-8
CMKRNL privilege, 4-8
COBOL logical names, 8-2
Command interpreter, 2-4
default, 2-1, 2-3
symbol table, 12-19
Command procedure,
as batch job, 9-4
login, 2-1, 2-9
to create system files, 13-21
see also Site-independent
start-up command procedure
see also Site-specific start-up
command procedure
Commands, DCL, 11-7, 11-9
Common event flag cluster, 4-11
protection, 3-6
Communication, operator, 4-10
Components of the VAX/VMS operating system, 1-2
Compute-bound programs, 11-14

Index-2

INDEX

Compute state, 11-5
Configuration changes, 10-1
Configure devices automatically,
13-4
CONNECT (console) command of
SYSGEN, 13-6
CONNECT (hardware) command of
SYSGEN, 13-5
CONNECT (software) command of
SYSGEN, 13-6
Connect,
console block storage device,
13-6
devices and load drivers, 13-5,
13-9
automatically, 13-4
in site-independent start-up
command procedure, 8-4
multiport memory, 13-12, 13-23
software devices and load
drivers, 13-6
_Connect-to-interrupt processes,
12-18
Console block storage device,
connect, 13-6
CONTIN.SYS (continuation file),
5-4
Continuation file (CONTIN.SYS),
5-4
Core image file (CORIMG.SYS), 5-3
CORIMG.SYS (core image file), 5-3
CPU time limit, 2-3, 4-3
/CPUTIME, 2-3
CRDENABLE system parameter, 12-19
CREATE command,
in DISKQUOTA, 6-3
in SYSGEN, 13-6
Create,
batch queue, 9-5
disk quota file, 6-3
dump file, 13-6
first-level user directory, 2-9
primary paging file, 13-6
primary swapping file, 13-6
print queue, 9-9
secondary paging file, 13-6
secondary swapping file, 13-6
system files, command procedure
to, 13-21
system parameter file, 13-17
Creation limit, subprocess, 2-5,
4-3
Cryptic SYE report, 10-1
/CSR, 13-5
<CTRL/C> command of DISPLAY, 14-2
<CTRL/Y>, 2-1, 2-4
Current state, 11-5
Current system parameter values,
12-1, 13-17
mod i f y , 1 3-18

Current system parameter
values, (Cont.)
use, 13-15
write, 13-16
CYCLE command of DISPLAY, 14-3

D
Data structures and devices, protection of, 3-2
Day, time of, 12-20
DCL commands, 11-7, 11-9
Deassign logical print queue, 9-11
Debugger logical names, symbolic,
8-2
DECnet, 4-10
Decrement for automatic working set
adjustment, 12-17
Deductible limits, 4-1
Default,
account name, 2-3
command interpreter, 2-1, 2-3
device, 2-3
directory, 2-4
file specification, 2-1
priority, 12-22
system parameter values, 12-2,
13-15
working set default, 2-61 4-4,
11-10
DEFAULT command in AUTHORIZE, 2-9
DEFAULT record, 2-1
DEFCLI, 2-4
Deferred writing of file header,
12-23
DEFMBX system parameters, 12-18
DEFPRI system parameter, 12-22
/DELETE, 7-5
Delete,
access, 3-3
batch queue, 9-6
known image, 7-5, 7-11
print queue, 9-10
DETACH privilege, 4-9
Detached process, 4-9
Device,
allocate, 4-7
automatically configure, 13-4
connect, 13-5, 13-19
in site-independent start-up
command procedure, 8-4
default device, 2-3
errors, 10-1
physical device queue, 9-9
protection, 3-2
software, 13-6
spooled, 4-10
standard device types and
drivers, 13-20 (table)
/DEVICE, 2-3

Index-3

INDEX

/DEVICES, 13-14
DIAGNOSE privilege, 4-9
/DIOLM, 2-3
Direct I/O count limit, 2-3, 4-3
Directory,
cache, 11-17, 12-23
create first-level user, 2-9
default, 2-4
master file, 5-3
system directory cache, 11-17
/DIRECTORY, 2-4
Disable,
accounting, 4-7
disk quotas, 6-4
range checks, 13-7
DISABLE command,
in DISKQUOTA, 6-4
in SYSGEN, 13-7
DISCTLY, 2-4
Disk,
capacity, 11-2
Files-11 file structure, 5-2
integrity, 5-8
public volumes, 4-10
initialize, 5-5
mount, 5-7, 8-5
transfer rate, 11-2
Disk quota,
add entry to disk quota file,
6-3
cache, 11-17, 12-23
create disk quota file, n-3
disable, 6-4
enable, 6-4
exceed, 4-9, 6-9
modify entry in disk quota file,
6-4
overdraft , 6-1
rebuild disk quota file, 6-5
automatically, 6-10
remove entry from disk quota
file, 6-5
show disk quota entry, 6-6
suspend, 6-10
DISKQUOTA utility, 6-1
ADD command, 6-3
commands, 6-2 (table)
CREATE command, 6-3
DISABLE command, 6-4
ENABLE command, 6-4
EXIT command, 6-4
HELP command, 6-4
invoking, 6-3
messages (table), 6-7
MODIFY command, 6-4
operations, 6-9
REBUILD command, 6-5
REMOVE command, 6-5
SHOW command, 6-6
USE command, 6-6

Dismount volume, 7-11, 10-1
Display,
FILE PRIMITIVE STATISTICS, 14-3,
14-5 (figure)
I/O SYSTEM RATES, 11-12, 14-5,
14-7 (figure)
NONPAGED POOL STATISTICS, 11-15,
14-11, 14-12 (figure)
NUMBER OF PROCESSES IN SCHEDULER
STATES, 14-12, 14-14 (figure)
14-15 (figure)
PAGE MANAGEMENT STATISTICS,
11-15, 14-9, 14-11 (figure)
TIME IN PROCESSOR MODES, 14-7,
14-8 (figure), 14-9 (figure)
TOP CPU TIME USERS, 14-15, 14-16
(figure)
VAX/VMS PROCESSES, 11-12, 14-16,
14-18 (figure)
DISPLAY utility, 11-8, 14-1
commands, 14-1 (table)
<CTRL/C> command, 14-2
CYCLE comand, 14-3
EXIT command, 14-3
FCP command, 14-3
HELP command, 14-5
invoking, 14-2
!ORATES command, 14-5
M2 command, 14-7
MS command, 14-7
messages, 14-18 (table)
PAGE command, 14-9
POOL command, 14-11
S2 command, 14-12
SS command, 14-12
terminate a display, 14-2
TOPUSERS command, 14-15
USERS command, 14-16
Distribution, workload, 11-3
Driver,
load, 13-5, 13-9, 13-19
automatically, 13-4
in site-independent start-up
command procedure, 8-4
reload, 13-9
show devices, drivers, and
structures, 13-14, 13-15
(figure)
sizes, 12-12
standard device types and
d r i ve rs , 1 3- 2 0 (tab 1 e )
/DRIVERNAME, 13-5, 13-6
DR32 transfer rate, 12-18
Dump,
Analyzer, System, 8-6
file, 13-20
create, 13-6
on fatal bugcheck, 12-19
DUMPBUG system parameter, 12-19
/DYNAMIC, 13-13

Index-4

INDEX

Dynamic pool,
see Nonpaged dynamic pool
see Paged dynamic pool

E
ECC errors, 10-1, 12-19
Enable,
disk quotas, n-4
range checks, 13-7
ENABLE command,
in DISKQUOTA, n-4
in SYSGEN, 13-7
/ENQLM, 2-4
ERFSTART command procedure, 10-3
ERRFMT process, 10-3
ERRLOG.SYS, 10-1
Error log, 10-1
Event flag cluster, see Common
event flag cluster
Events, system, 10-1
Exceed disk quota, 4-9, n-9
Executable image, 7-2
Execute access, 3-3
Executive mode, 4-8
EXIT command,
in AUTHORIZE, 2-10
in DISKQUOTA, 6-4
in DISPLAY, 14-3
in SYSGEN, 13-8
Exit handler, process, 12-17
EXQUOTA privilege, 4-9
/EXTENSION, 5-6
Extent cache, 11-17, 12-23
Extent limit, 12-23
EXTRACPU system parameter, 12-17

F
Fatal bugcheck,
all bugchecks fatal, 12-19
automatic reboot on, 12-19
dump file on, 12-19
Fault, see Page fault
FCB, see File control block
FCP command of DISPLAY, 14-3
FIELD record, 2-2, 2-6 (table)
File,
access, 3-2 (figure)
control block cache, 12-33
default file specification, 2-1
header, 5-3
cache, 12-23
deferred writing of, 12-23
identification cache, 11-17,
12-23
limit, open, 2-4, 4-3
protection, 3-4 (figure)

File, (Cont.)
public, 5-1
RMS file sharing, 5-8
in site-independent start-up
command procedure, 8-4
see also System parameter file
Fi les-11
caches, 5-8, 11-16, 12-23
disk file structure, 5-2
protection, 3-4
Structure Levels 1 and 2, 5-4
First-level user directory, 2-9
/FLAGS, 2-4
Flag,
login, 2-1, 2-4
see also Common event flag
cluster
Fluid pages in working set, 12-15
Forms control, 9-11
/FORMS TYPE, 9-11
FORMSTYPE.DAT, 9-11
FORTRAN logical names, 8-2
FREELIM system parameter, 11-12,
12-18
Free page list, 11-5, 11-12, 12-18
/FULL, 2-11, 7-6
Full known image display, 7-7
(figure)

G
GBL$, 7-12
GBLPAGES sytem parameter, 12-9
/GBLSECTIONS, 13-11, 13-24
GBLSECTIONS sytem parameter, 12-9
/GEN, 13-13
Generic queue, 9-9
/GLOBAL, 7-6
Global page table entries, 12-9
Global section, 4-13, 7-2, 11-3
descriptors, 12-9
display, 7-8 (figure)
permanent, 4-12
protect ion, 3-6
system, 4-13
Group,
access, 4-9
logical name, 4-10
protection, 3-6
users, 3-2, 3-6
GROUP privilege, 4-9
GRPNAM privilege, 4-10

H
Hard ECC errors, 10-1
Hardware configurations, 11-2,
12-1

Index-5

INDEX

header, file, 5-3
cache, 11-ln, 12-23
deferred writing of, 12-23
header resident known image, 7-1
/HEADER RESIDENT, 7-4
/HEADERS, 5-6
HELP command,
in AUTHORIZE, 2-10
in DISKQUOTA, n-4
in DISPLAY, 14-5
in SYSGEN, 13-8
Home block, 5-2
back-up, 5-2

IJOBLIM system parameter, 12-22
Image,
executable, 7-2
shareable, 7-2, 7-12
see also Known image
Inaccessible user authorization
file, 2-17
Increase in working set limit,
12-In

/INDEX, 5-6
Index file (INDEXF.SYS), 5-2
bit map, 5-3
header, back-up, 5-3
INDEXF.SYS (index file), 5-2
/!NIT, 13-11
Initialize,
batch queue, 9-5
multiport memory, 13-11, 13-23
print queue, 9-9
public disk volumes, 5-5
queues in site-specific start-up
command procedure, 8-6
Input symbiont, 9-3
Install,
known image, 7-1, 7-4
in site-independent start-up
command procedure, 8-3
in site-specific start-up
command procedure, 8-5
secondary paging file, 13-8
secondary swapping file, 13-8
INSTALL command of SYSGEN, 13-8
INSTALL utility, 7-3
brief display, 7-6 (figure)
commands, 7-3 (table)
delete, 7-5
full display, 7-7 (figure)
global section display, 7-8
(figure)
header resident, 7-4
invoking, 7-3
list, 7-6
messages, 7-10

INSTALL utility, (Cont.)
operations, 7-10
permanently open, 7-4
privileged, 7-4
protected, 7-4
replace, 7-5
shared, 7-4
terminate, 7-9
writeable, 7-4
Integrity, disk volume, 5-8
Interrupt stack size, 12-15
INTSTKPAGES system parameter, 12-15
Invoking,
AUTHORIZE, 2-7
DI SK QUOTA, n-3
DISPLAY, 14-2
INSTALL, 7-3
SYSGEN, 13-4
I/O,
logical access, 3-3
physical access, 3-3
byte limit, buffered, 2-3, 4-2
count limit, buffered, 2-3, 4-2
count limit, direct, 2-3, 4-3
operations
logical, 4-10
physical, 4-11
request packets, preallocated,
12-10

transfer limit, buffered, 12-18
I/O SYSTEM RATES display, 11-12,
14-5, 14-7 (figure)
!ORATES command of DISPLAY, 14-5
IRPCOUNT system parameter, 12-10

J
/JOB, 13-13
Job,
limits, 12-22
see also Batch job
see also Print job
JOB system parameters, 12-6, 12-22

K
Kernel mode, 4-8
KFILSTCNT system parameter, 12-15
Known file list, 7-2
head, 12-15
Known image
delete, 7-5, 7-11
header resident, 7-1
install, 7-1, 7-4
in site-independent start-up
command procedure, 8-3
in site-specific start-up
command procedure, 8-6

Index-6

INDEX

Known image, (Cont.)
list, 7-6
order of installation, 7-11
permanently open, 7-1
privileged, 7-1, 7-11
protected, 7-1
replace, 7-5
shared, 7-1, 7-12
writeable, 7-1
display, 7-5 (figure), 7-7
(figure)

Logical name,
COBOL, 8-2
FORTRAN, 8-2
group, 4-10
PASCAL, 8-2
protection, 3-6
RSX-llM, 8-2
site-independent start-up command
procedure, 8-2
site-specific start-up command
procedure, 8-5
Symbolic Debugger, 8-2
system, 4-13
L
Logical print queue, 9-9
assign, 9-10
LAMAPREGS system parameter, 12-18
deassign, 9-11
Large systems, 11-1, 11-lh
Login,
Large working sets, very, 11-14
command procedure, 2-1, 2-4, 2-9
Levels 1 and 2, Files-11 Structures,
flags, 2-1, 2-4
/LGICMD, 2-4
restrictions, 11-3
limits on resources, 2-1, 4-1,
LOG IO privilege, 4-10
12-25
LPAll map registers, 12-18
AST queue, 2-3, 4-2
buffered I/O byte, 2-3, 4-2
buffered I/O count, 2-3, 4-2
buffered I/O transfer, 12-18
CPU time, 4-3
M2 command of DISPLAY, 14-7
deductible, 4-1
MS command of DISPLAY, 14-7
default working set, 2-6
Machine checks, 10-1
direct I/O count, 2-3, 4-3
Mailbox, 4-13
extent, 12-23
buffer quota, 12-18
job, 12-22
message size, 12-18
nondeductible, 4-1
permanent, 4-12
open file, 2-4, 4-3
protection, 3-5
pooled, 4-1
temporary, 4-14
paging file, 2-5, 4-3
/MAILBOXES, 13-11, 13-24
subprocess creation, 2-5, 4-3
/MAJOR, 13-13
timer queue entry, 2-5, 4-4
MAJOR system parameters, 12-3, 12-8
working set, 2-6, 4-4, 11-9
Mask, protection, 3-3
Line width, terminal, 12-21
Master file directory (MFD), 5-3
/LIST, 7-6
MAXBUF system parameter, 12-18
LIST command in AUTHORIZE, 2-10
/MAXCEFCLUSTERS, 13-11, 13-12
LOAD command in SYSGEN, 13-9
/MAXGBLSECTIONS, 13-11, 13-12
Load driver, 13-5, 13-9, 13-19
/MAXIMUM, 5-6
automatically, 13-4
Maximum,
in site-independent start-up
CPU time, 2-3
command procedure, 8-4
print symbionts, 12-22
Lock password, 2-4
system parameter values, 12-2
LOCKPWD, 2-4
virtual pages, 12-13
Log file,
/MAXMAILBOXES, 13-11, 13-12
accounting, 4-15
MAXPRINTSYMB system parameter,
back-up (BACKUP.SYS), 5-4
12-22
error, 10-1
MAXPROCESSCNT system parameter,
operator's, 10-5
12-9
pending bad block (BADLOG.SYS),
MAXSYSGROUP system parameter, 12-17
5-4
/MAXUNITS, 13-5
Logical I/O,
Memory,
access, 3-3
physical, 11-2, 11-4 (figure)
operations, 4-10
see also Multiport memory

Index-7

INDEX

Messages,
AUTHORIZE, 2-14
DI SKQUOTA, o-7
DISPLAY, 14-18
INSTALL, 7-10
SYSGEN, 13-16
MFD, see Master file directory
Minimum,
fluid pages in working set,
12-15
system parameter values, 12-2
MINIMUM.PAR, 12-1
MINWSCNT system parameter, 12-15
Modified page list, 11-5, 11-12,
12-14
Modify,
active system, 13-18
current system image, 13-18
entry in disk quota file, o-4
entry in user authorization file,
2-11
MOUNT privilege, 4-10
Mounting public disk volumes, 5-7,
10-1
MPW HILIM system parameter, 12-14
MPW-LOLIM system parameter, 12-14
MPW-WRTCLUSTER system parameter,
-12-14
Multiple ancillary control
processes, 11-17, 12-22
Multiport memory, 4-13, 7-12
connect, 13-12, 13-23
initialize, 13-11, 13-23
structures, 13-23, 13-24
MWAIT state, 11-15

N
name,
see Logical name
see User name
/NAMES, 13-13
NETMBX privilege, 4-10
NJOBLIM system parameter, 12-22
/NOADAPTER, 13-6
/NOCONTIGUOUS, 13-6
Nondeductible limits, 4-1
Nonpaged dynamic pool, 11-15,
12-11
NONPAGED POOL STATISTICS display,
-11-15, 14-11, 14-12 (figure)
NPAGEDYN system parameter, 11-15,
12-11
NUMBER OF PROCESSES IN SCHEDULER
STATES display, 14-12, 14-14
(figure), 14-15 (figure)
/NUMVEC, 13-5

0
/OPEN, 7-4
Open,
file limit, 2-4, 4-3
permanently open known image, 7-1
system, 2-17
OPER privilege, 4-10
Operating system, components of
the VAX/VMS, 1-2
Operations,
AUTHORIZE, 2-17
DI SK QUOTA, 6-9
INSTALL, 7-10
SYE, 10-3
SYSGEN, 13-1 7
Operator communications, 4-10
OPERATOR.LOG, 10-5
Operator's log file, 10-5
Order to install known images,
7-11
Output symbionts, 9-3
/OVERDRAFT, n-3, n-4
Overdraft of disk quota, 6-1
/OWNER, 2-4
Owner,
access, 3-2
name, 2-4

p
Packets, preallocated I/O request,
12-10
Page access, physical, 4-11
cache, 11-5, 11-12
fault, 11-5, 12-8
rate, 12-16
size, vertical, 9-11
table entries,
global, 12-9
system, 12-15
PAGE command of DISPLAY, 14-9
Paged dynamic pool, 5-8, 12-13,
12-19
PAGEDYN system parameter, 12-13
/PAG EFI LE, 13-8
PAGEFILE.SYS, 3-7, 13-20
PAGE MANAGEMENT STATISTICS display,
11-15, 14-9, 14-11 (table)
Pages,
in working set, minimum fluid,
12-15
maximum virtual, 12-13
Paging, 11-9
paged dynamic pool, 12-19
system working set, 12-20

Index-8

INDEX

Paging file,
limit, 2-5, 4-3
primary, 13-20
create, 13-6
secondary, 13-22
create, 13-6
install, 13-8
Parameter, see System parameter
PASCAL logical names, 8-2
/PASSWORD, 2-5
Password, 2-1, 2-5, 3-7
lock, 2-4
/PBYTLM, 2-5
Pending bad block log file
(BADLOG.SYS), 5-4
Permanent
global section, 4-12
ma i 1 box , 4 -1 2
Permanently open known image, 7-1
/PERMQUOTA, 6-3, 6-4
PFCDEFAULT system parameter, 11-13,
12-8
PFNMAP privilege, 4-11
PFRATH system parameter, 11-11,
12-Hi
PFRATL system parameter, 11-11,
12-H
PHY IO privilege, 4-11
Physical
device queue, 9-9
I/O access, 3-3
I/O operations, 4-11
memory configuration, 11-2, 11-4
(figure)
page access, 4-11
Polling terminals, 12-20
Pool,
see Nonpaged dynamic pool
see Paged dynamic pool
POOL command of DISPLAY, 14-11
/POOLBCNT, 13-11, 13-25
/POOLBSIZE, 13-11, 13-25
Pooled limits, 4-1
POOL PAGING system parameter,
12-19
/PQL, 13-13
PQL system parameters, 12-7, 12-25
/PRCLM, 2-5
Preallocated I/O request packets,
12-10
Primary,
paging file, 13-20
create, 13-6
swapping file, 13-20
create, 13-6
Printer,
characteristics, 9-12
in site-specific start-up
command procedure, 8-5

Printer, (Cont.)
set printer spooled, 9-3
in site-specific start-up
command procedure, 8-5
Print job, 9-1
spool, 9-8
Print queue,
assign logical, 9-10
create, 9-9
deassign logical, 9-11
delete, 9-10
guidelines, 9-12
initialize, 9-9
start, 9-10
stop, 9-10
Print symbionts, maximum, 12-22
/PRIO, 2-5
Priority, 2-1, 2-5, 11-5, 11-14
alter base, 4-7
ancillary control process, 12-24
base, 4-4
default, 12-22
Privilege, 2-1, 3-7, 4-5 (table),
4-13, 7-9
ACNT, 4-7
ALLSPOOL, 4-7
ALTPRI, 4-7
bits, 7-9 (table)
BUGCHK, 4-8
BYPASS, 4-8
CMEXEC, 4-8
CMKRNL, 4-8
DETACH, 4-9
DIAGNOSE, 4-9
EXQUOTA, 4-9
GROUP, 4-9
GRPNAM, 4-10
LOG IO, 4-10
MOUNT, 4-10
names, 7-9 (table)
NETMBX, 4-10
OPER, 4-10
PFNMAP, 4-11
PHY IO, 4-11
PRMCEB, 4-11
PRMGBL, 4-12
PRMMBX, 4-12
PSWAPM, 4-12
SETPRV, 4-13
SHMEM, 4-13
SY SG BL , 4 -13
SYSNAM, 4-13
SYSPRV, 4-14
TMPMBX, 4-14
VOLPRO, 4-14
WORLD, 4-14
/PRIVILEGED, 7-4
Privileged known image, 7-1, 7-11
/PRIVILEGES, 2-5

Index-9

INDEX

Procedure,
see Command procedure
Process,
connect-to-interrupt, 12-18
detached, 4-9
entry slots, 12-9
exit handler, 12-17
limits, 12-25
section descriptors, 12-15
state, 11-5
/PROCESSOR, 5-7
PROCSECTCNT system parameter, 12-15
/PROTECTED, 7-4
Protected known image, 7-1
Protection, 3-2
bypass, 4-8
card reader, 3-5, 3-8
common event flag, 3-6
Files-11, 3-4, 3-4 (figure)
global section, 3-6
group, 3-6
logical name, 3-6
mailbox, 3-5
mask, 3-3
sensitive system files, 3-7
terminal, 3-5, 3-8, 12-21
volume, 4-14
/PRQCOUNT, 13-11, 13-25
PSWAPM privilege, 4-12
Public disk volumes, 4-10, 5-1
back-up, 5-10
initializing, 5-5
mount i n g , 5 - 7
in site-specific start-up
command procedure, 8-5

Q
Quantum, 11-5, 11-14, 12-14
QUANTUM system parameter, 11-14,
12-14
Queue,
generic, 9-9
logical, 9-9
physical device, 9-9
see also Batch queue
see also Print queue
Quota,
mailbox buffer, 12-18
system working set, 12-10
working set, 2-6, 4-4, 11-10
see also Disk quota
QUOTA.SYS (quota file), 6-1

R
Range checks,
disable, 13-7
enable, 13-7

Read access, 3-3
Read verification, ancillary
control process, 12-24
REALTIME SPTS system parameter,
12-18
Reboot on fatal bugcheck, automatic, 12-19
REBUILD command in DISKQUOTA, 6-5
Rebuild disk quota file, 6-5
automatically, 6-10
Record Management Services (RMS)
blocking factor, 12-25
buffers, 11-15
file sharing, 5-8
in site-independent start-up
command procedure, 8-4
RELOAD command in SYSGEN, 13-9
Reload driver, 13-9
REMOVE command,
in AUTHORIZE, 2-12
in DISKQUOTA, 6-5
Remove entry,
from disk quota file, 6-5
from user authorization file,
2-12
/REPLACE, 7-5
Replace known image, 7-5
Request packets, preallocated I/O,
12-10
RESALLOC system parameter, 12-20
Resources,
accounting for system, 4-15
hardware, 11-2
limits on system, 2-1, 4-1
/RMS, 13-13
RMS, see Record Management Services
RMSSHARE utility, 5-8
RMS system parameters, 12-7, 12-25
RMTNODE.DAT, 3-7
Roll-up SYE report, 10-1
RSX-llM logical names, 8-2

s
S2 command in DISPLAY, 14-12
SS command in DISPLAY, 14-12
SB! errors, 10-1, 12-19
Secondary paging file, 13-22
create, 13-6
install, 13-8
Secondary swapping file, 13-22
create, 13-6
install, 13-8
Section, see Global section
/SELECT, 13-4
Sensitive system files, protection
of, 3-7
SET (parameter) command in SYSGEN,
13-10

Index-10

INDEX

SET (start-up command procedure) in
SYSGEN, 13-10
SETPRV privilege, 4-13
SETTIME system parameter, 12-20
SHARE (connect) command of SYSGEN,
13-12
SHARE (initialize) command of
SYSGEN, 13-11
Shareable image, 7-2, 7-12
/SHARED, 7-4
Shared,
ancillary control processes,
12-22
known image, 7-1, 7-12
SHMEM privilege, 4-13
Show,
devices, drivers, and structures,
13-14, 13-15 (figure)
disk quota entry, 6-6
system parameter values, 13-13,
13-14 (figure)
user authorization file entry,
2-12, 2-13 (figure)
SHOW (device) command in SYSGEN,
13-14
SHOW (parameter) command in SYSGEN,
13-13
SHOW (start-up) command in SYSGEN,
13-15
SHOW command,
in AUTHORIZE, 2-12
in DISKQUOTA, n-6
/SHRFILLM, 2-5
Site-independent start-up command
procedure, 8-1
connect devices in, 8-4
create new, 13-22
install known images in, 8-3
load drivers in, 8-4
Record Management Services file
sharing in, 8-4
set name, 13-10
show name, 13-15
Site-specific start-up command
procedure, 8-5
assign logical names in, 8-5
initialize queues in, 8-6
install known images in, 8-6
mount public disk volumes in, 8-5
set printer characteristics in,
8-5
set printer spooled in, 8-5
set terminal speed in, 8-5
start queues in, 8-6
/SIZE, 13-6
Slots
balance set, 12-10
process entry, 12-9
Small systems, 11-1, 11-9

Soft ECC errors, 10-1
Software devices, 13-6
Software Performance Report (SPR),
10-6
Speed, terminal, 12-20
Spooled,
devices, 4-10
printers, 9-3, 9-12
in site-specific start-up
command procedure, 8-5
S po o 1 i n g , 9- 2
pr int jobs, 9-8
SPR, see Software Performance
Report
SPT, see System page table
SPTREQ system parameter, 12-15
Standard,
device types and drivers, 13-20
(table)
SYE report, 10-1
/START, 13-11
Start,
batch queue, 9-5
print queue, 9-10
in site-specific start-up
command procedure, 8-5
/STARTUP, 13-15
STARTUP.COM, 8-1
Start-up command procedure
see Site-independent start-up
command procedure
see Site-specific start-up
command procedure
State, process, 11-5
Stop,
batch queue, 9-6
print queue, 9-10
Storage bit map file (BITMAP.SYS)'
5-3
Structure,
driver structure sizes, 12-12
Levels 1 and 2, Files-11, 5-4
in multiport memory, 13-23, 13-24
(figure)
protection, 3-2
Submit batch job, 9-4
Subprocess creation limit, 2-5, 4-3
Suspend disk quota, 6-10
/SWAP FI LE, 13-8
SWAPFILE.SYS, 3-7, 13-20
SWAPFILES.COM, 13-21
Swapping, 4-12, 11-5, 11-9
ancillary control process, 12-24
see also Primary swapping file
see also Secondary swapping file
SWPRATE system parameter, 11-14
SYE utility, 10-1
operations, 10-3
reports, 10-1

Index-11

INDEX

Symbionts,
input and output, 9-3
maximum print, 12-22
Symbol table, command interpreter,
12-19
Symbol Debugger logical names, 8-2
/SYS, 13-13
SYS, system parameters, 12-4, 12-14
SYS$BATCH, 9-3
SYSDUMP.DMP, 3-7, 13-20
[SYSERR], 1-2
[SYSEXE], 1-3
SYS.EXE, 13-17
SYSGBL privilege, 4-13
SYSGEN utility, 13-1
AUTOCONFIGURE command, 13-4
commands, 13-1 (table)
CONNECT (console) command, 13-n
CONNECT (hardware) command, 13-5
CONNECT (software) command, 13-6
CREATE command, 13-6
DISABLE command, 13-7
ENABLE command, 13-7
EXIT command, 13-8
HELP command, 13-8
INSTALL command, 13-8
invoking, 13-4
LOAD command, 13-9
messages, 13-16
operations, 13-17
RELOAD command, 13-9
SET (parameter) command, 13-10
SET (start-up command procedure),
13-10
SHARE (connect) command, 13-12
SHARE (initialize) command,
13-11
SHOW (device) command, 13-14
SHOW (parameter) command,
13-13
SHOW (start-up) command, 13-15
USE command, 13-16
WRITE command, 13-16
SYS$HELP, 8-3
[SYSHLP], 1-2
SYS$LIBRARY:, 8-3
LIB.MLB, 2-18, 4-16
[SYSLIB], 1-2
SYS$LP LINES, 9-11
[SYSMAINT], 1-3
SYS$MESSAGE, 8-3
[SYSMGR], 1-2
[ SYSMSG] , 1-2
SYSMWCNT system parameter, 12-10
SYSNAM privilege, 4-13
SYSPAGING system parameter, 12-20
SYSPRV privilege, 4-14
SYS$SHARE, 7-12
SYS$SYSDISK: [SYSERRJERRLOG.SYS,
10-1

SYS$SYSDISK: [SYSMGR],
ACCOUNTING.DAT, 4-15
CHARTYPE.DAT, 9-12
FORMSTYPE.DAT, 9-11
OPERATOR.LOG, 10-5
SWAPFILES.COM, 13-21
SYS$SYSTEM:,
BUSER.PAR, 12-1
16USER.PAR, 12-1
32USER.PAR, 12-1
48USER.PAR, 12-1
64USER.PAR, 12-1
ERFSTART.COM, 10-3
MINIMUM.PAR, 12-1
PAGEFILE.SYS, 3-7, 13-20
RMTNODE.DAT, 3-7
STARTUP.COM, 8-1
SWAPFILE.SYS, 3-7, 13-20
SYSDUMP.DMP, 3-7, 13-20
SYS.EXE, 13-17
SYSUAFALT.DAT, 2-18
SYSUAF.DAT, 2-17, 3-7
VIRT32MB.PAR, 12-1
System directory cache, 11-17
System Dump Analyzer, 8-6
System events, 10-1
System files,
command procedure to create,
13-21
protection of sensitive, 3-7
System global section, 4-13
System image, see Current system
image
System logical name, 4-13
System page table entries, 12-15
System parameter, 12-6 (tables)
AC P , 1 2- 6 ( tab 1 e ) , 1 2- 2 2
ACP BASEPRIO, 12-24
ACP-DATACHECK, 12-24
ACP-DIRCACHE, 11-17, 12-23
ACP-EXTCACHE, 11-17, 12-23
ACP-EXTLIMIT, 11-17, 12-23
ACP-FIDCACHE, 11-17, 12-23
ACP-HDRCACHE, 11-16, 12-23
ACP-MAPCACHE, 11-17, 12-23
ACP-MAXREAD, 12-24
ACP-MULTIPLE, 11-17, 12-22
ACP-QUOCACHE, 11-17, 12-23
ACP-SHARE, 12-22
ACP-SWAPFLGS, 12-24
ACP-SYSACC, 11-17, 12-23
ACP-WINDOW, 12-24
ACP-WORKSET, 12-24
ACP-WRITEBACK, 11-16, 12-23
AWSMAX, 11-11, 12-17
AWSMIN, 11-11, 12-17
AWSTIME, 11-11, 12-17
BALSETCNT, 12-10
BJOBJIM, 12-22
BUGCHECKFATAL, 12-19

Index-12

INDEX

System parameter, (Cont.)
BUGREBOOT, 12-19
CLISYMTBL, 12-19
CRDENABLE, 12-19
DEFMBX, 12-18
DEFPRI, 12-22
DUMPBUG, 12-19
EXTRACPU, 12-17
FREELIM, 11-12, 12-18
GBLPAGES, 12-9
GBLSECTIONS, 12-9
IJOBLIM, 12-22
INTSTKPAGES, 12-15
IRPCOUNT, 12-10
JOB, 12-6 (table), 12-22
KFILSTCNT, 12-15
LAMAPREGS, 12-18
MAJOR, 12-3 (table), 12-8
MAXBUF, 12-18
MAXPRINTSYMB, 12-22
MAXPROCESSCNT, 12-9
MAXSYSGROUP, 12-17
MINWSCNT, 12-15
MPW HILIM, 12-14
MPW-LOLIM, 12-14
MPW-WRTCLUSTER, 12-14
NJOBLIM, 12-22
NPAGEDYN, 11-15, 12-11
PAGEDYN, 12-13
PFCDEFAULT, 11-13, 12-8
PFRATH, 11-11, 12-16
PRFATL, 11-11, 12-16
POOL PAGING, 12-19
PQL,-12-7 (table), 12-25
PROCSECTCNT, 12-15
QUANTUM, 11-14, 12-14
REALTIME SPTS, 12-18
RESALLoc-; 12-20
RMS' 12-7 (table) ' 12-2 5
SBIERRENABLE, 12-19
SETTIME, 12-20
SPTREQ, 12-15
SWPRATE, 11-14
SYS, 12-4 (table), 12-14
SYSMWCNT, 12-10
SY SP AG ING, 12-20
TTY BUF, 12-21
TTY-DEFCHAR, 12-21
TTY-OWNER, 12-21
TTY-PROT, 12-21
TTYSCANDELTA, 12-20
TTY SPEED, 12-20
TTY-TYPAHDSZ, 12-21
UAFALTERNATE, 12-20
VIRTUALPAGECNT, 12-13
WSDEC, 11-11, 12-17
WSINC, 11-11, 12-16
WSMAX, 11-10, 12-11
XFMAXRATE, 12-18

System parameter file, 12-1,
12-8 (table)
8 USER • PAR , 1 2-1 , 1 2- 8 ( tab 1 e )
16 us ER • p AR ' 1 2-1 ' 1 2-8 ( tab 1 e )
3 2 US ER • PAR , 1 2 -1 , 1 2 - 8 ( t a b 1 e )
48USER.PAR, 12-1, 12-8 (table)
6 4 us ER • p AR ' 1 2-1 ' 1 2-8 ( tab 1 e )
create, 13-17
MI NIM uM. p AR ' 1 2-1 ' 1 2-8 ( tab 1 e )
use, 13-15
VIRT32MB.PAR, 12-1, 12-8 (table)
write, 13-16
System parameter value,
active, 12-1
use, 13-15
write, 13-16
current, 12-1
use, 13-15
write, 13-16
default, 12-2, 13-15
maximum, 12-2
minimum, 12-2
set, 13-10
show, 13-13, 13-14 (table)
SYSTEM record, 2-2, 2-6 (table)
System resources,
accounting for, 4-15
limits on, 4-1
System user, 3-2, 4-14, 12-17
System working set, 11-15
paging of, 12-20
quota, 12-10
SYSTEST record, 2-2, 2-6 (table)
[SYSTEST], 1-2
SYSUAFALT.DAT, 2-18
SYSUAF.DAT, 2-17, 3-7
[SYSUPD], 1-3
[SYSUPD.EXAMPLES], 1-3

T
Temporary mailbox, 4-14
Terminal,
allocation and protection, 3-5,
3-8
characteristics, 12-21
line width, 12-21
po 11 i ng, 12-2 0
protection, 12-21
queues, 9-19
speed, 12-20
in site-specific start-up
command procedure, 8-5
type-ahead buffer, 12-21
TIME IN PROCESSOR MODES display,
14-7
for M2 command, 14-8 (figure)
for MS command, 14-9 (figure)

Index-13

INDEX

Time limit, CPU, 4-3
Time of day, 12-20
Timer queue entry limit, 2-5, 4-4
TMPMBX privilege, 4-14
TOP CPU TIME USERS display, 14-15,
14-16 {figure}
TOPUSERS command of DISPLAY, 14-15
/TQELM, 2-5
Transfer rate,
disk, 11-2
DR32, 12-18
TTY BUF system parameter, 12-21
TTY-DEFCHAR system parameter,
-12-21
TTY OWNER system parameter, 12-21
TTY-PROT system parameter, 12-21
TTYSCANDELTA systeme parameter,
12-20
TTY SPEED system parameter, 12-20
TTY-TYPAHDSZ system parameter,
-12-21
Tun in g , 11-1
Type-ahead buffer, terminal, 12-21

v
validation, user, 2-17
VAX-11 RMS, see Record Management
Services
VAX/VMS PROCESSES display, 11-12,
14-16, 14-18 (table}
/VECTOR, 13-5
Vertical page size, 9-11
VIRT32MB.PAR, 12-1
VIRTUALPAGECNT system parameter,
12-13
Virtual pages, maximum, 12-13
VOLPRO privilege, 4-14
VOLSET.SYS {volume set list file}
5-4
Volume, 4-10, 5-1
back-up of public, 5-10
dismount, 7-11
initialize public, 5-10
integrity, 5-8
mount public, 5-7
protection, 4-14
quota, see Disk quota
set list file (VOLSET.SYS}, 5-4

UAF, see User authorization file
UAFALTERNATE system parameter,
12-20
$UAFDEF module in
SYS$LIBRARY:LIB.MLB, 2-18
/UIC, 2-5
UIC, see User identification code
Update interval, 14-2
USE command,
in DISKQUOTA, 6-6
in SYSGEN, 13-15
User authorization file, 2-1, 11-7
add entry to, 2-7
alternate, 2-18, 12-20
inaccessible, 2-17
list entries in, 2-10
modify DEFAULT record, 2-9
modify entry in, 2-11
remove entry from, 2-12
show entries in, 2-12
user data areas in the, 2-18
version 2.0 upgrade of, 2-19
User data areas in the user
authorization file, 2-18
User directory, create first-level,
2-9
User identification code, 2-5, 3-1
User name, 2-1, 4-15
User validation, 2-17
USERS command of DISPLAY, 14-16

Waiting state, 11-5
Width, terminal line, 12-21
/WINDOW, 5-7
Window pointers, ancillary contro
process, 12-24
Working set, 11-4
adjustment, automatic, 11-11,
12-16
ancillary control process, 12-2
default, 2-6, 4-4, 11-9
minimum fluid pages in, 12-15
quota, 2-6, 4-4, 11-10, 12-10
system, see System working set
very large, 11-14
Workload,
configurations, 11-2, 12-1
distribution, 11-3
World access, 3-2, 4-14
WORLD privilege, 4-14
Write,
access, 3-3
errors, 10-1
verification, ancillary control
process, 12-24
/WRITEABLE, 7-4
Writeable known image, 7-1
WRITE command of SYSGEN, 13-16
WSDEC system parameter, 11-11,
12-17
/WSDEFAULT, 2-6, 11-10

Index-14

INDEX

WSINC system parameter, 11-11,
12-Hi
WSMAX system parameter, 11-10,
12-11
/WSQUOTA, 2-6, 11-10

x
XFMAXRATE system parameter, 12-18

Index-15

VAX/VMS
System Manager's Guide
AA-D027B-TE
READER'S COMMENTS

NOTE:

This form is for document comments only. DIGITAL will
use comments submitted on this form at the company's
discretion.
If you require a written reply and are
eligible to receive one under Software Performance
Report (SPR) service, submit your comments on an SPR
form.

Did you find this manual understandable, usable, and well-organized?
Please make suggestions for improvement •

0
0

Did you find errors in this manual?
page number.

If so, specify the error and the

Please indicate the type of reader that you most nearly represent.
[] Assembly language programmer
[] Higher-level language programmer
[] Occasional programmer (experienced)
[] User with little programming experience
[] Student programmer
[] Other (please specify>~~~~~~~~~~~~~~~~~~~

CitY~~~~~~~~~~~~~~State~~~~~~-Zip Code~~~~~~

or
Country

Do Not Tear - Fold Here and Tape
No Postage
Necessary
if Mai led in the
United States

'"]

··-------·····--·--·····---------- .................................

SI NESS REPLY MAIL
T CLASS PERMIT N0.33 MAYNARD MASS.
·-·~--·
------....-..·-------~----~-~·-··-"-"""·'•'·"'"" ___....--~--·..----..~--·-...······-_._..._._,,,_.

POSTAGE WILL BE PAID BY ADDRESSEE

BSSG PUBLICATIONS TW/A 14
DIGITAL EQUIPMENT CORPORATION
1925 ANDOVER STREET
TEWKSBURY, MASSACHUSETTS

Do Not Tear - Fold Here

01876