GC28 0667 1_OS_VS2_Planning_Guide_for_Release_2_Nov73 1 OS VS2 Planning Guide For Release 2 Nov73
User Manual: GC28-0667-1_OS_VS2_Planning_Guide_for_Release_2_Nov73
Open the PDF directly: View PDF .
Page Count: 174
Download | |
Open PDF In Browser | View PDF |
GC28-0667-' File No. S370-34 Systems OS/VS2 Planning Guide for Release 2 Program Number 5742-020 VS2 Release 2 Second Edition (November, 1973) This is a major revision of, and obsoletes, GC28-0667-0. See the Summary of Amendments following the Contents. Changes or additions to the text and illustrations are indicated by a vertical line to the left of the change. This edition applies to Release 2 of OS/VS2 and to all subsequent releases until otherwise indicated in new editions or Technical Newsletters. Changes are continually made to the information herein; before using this publication in connection with the operation of IBM systems, consult the latest IBM System/360 and System/370 Bibliography, GA22-6822, for editions that are applicable and current. Requests for copies of IBM publications should be made to your IBM representative or to the IBM branch office serving your locality. A form for readers' comments is provided at the back of this publication. If the form has been removed, comments may be addressed to IBM Corporation, Publications Development, Department D58, Building 706-2, PO Box 390, Poughkeepsie, N.Y. 12602. Comments become the property of IBM. © Copyright International Business Machines Corporation 1973 Preface This publication contains planning information for OS/VS2 Release 2. It is intended for installation managers and system programmers responsible for assessing the effort required to install an OS/VS2 Release 2 system. Readers should have a background in programming and maintaining computer operating systems such as OS/MVT or OS/VS2 Release l. This publication is for planning purposes only. The functions and capabilities described reflect the information that is currently available. This information is subject to change before the availability of OS/VS2 Release 2. Throughout this publication the term MVS is used interchangeably with VS2 Release 2. MVS refers to the mUltiple virtual storage concept of this release. This manual contains the following independent chapters: • Introduction -- The Introduction highlights the major points that should be considered in planning for the installation and implementation of VS2 Release 2. • Defining the System -- This chapter provides preliminary planning information for defining the type of system that an installation needs. For example, the chapter describes the VS2 Release 2: --procedures and macro instructions for system generation. --procedures and parameters for system initialization. --system libraries and data sets. For each topic, a summary of the major changes from both MVT and VS2 Release 1 is provided. • Directing the Use of System Resources new VS2 Release 2 facilities that allow distribution of system resources among information that helps assess the effort facilities. -- This chapter describes some of the an installation to control the system users. It provides planning involved in implementing these • System Integrity -- This chapter discusses integrity in a multi-user computer system. It includes recommendations for maintaining, in any control program extensions or modifications, the same level of system integrity as that of the VS2 Release 2 system. • Conversion Considerations -- This chapter presents preliminary descriptions of some of the most significant differences between VS2 Release 2 and MVT or VS2 Release 1. It provides the information to assess the extent of the modifications required for existing data sets, cataloged procedures, programs, data reduction routines, and operational procedures, to install VS2 Release 2. • Appendix A: Virtual Storage Map -- This appendix shows the layout of virtual storage in VS2 Release 2. • Glossary -- The glossary describes VS2 terms. Preface 3 The following items are described in this publication strictly for advance planning purposes and will not be available with the initial release of VS2 Release 2: • lES] job entry subsystem. • Virtual telecommunications access method (VT AM). Additional information about these items will be available later. Availability dates for the support of the items may be obtained from th{! local IBM branch offices . Prerequisite Publication IBM System/370 Introduction to VS2 Release 2, GC28-0661 This publication describes VS2 Release 2 features and facilities, design coneepts, and basic hardware requirements. Related Publications Introduction to Virtual Storage in System/370, GR20-4260 This publication provides an overview of the advantages and concepts of virtual storage systems. IBM Data Processing Glossary, GC20-1699 This publication defines terms and concepts used in IBM publications. OS/VS Virtual Storage Access Method (V SAM) Planning Guide:, GC26-3799 This publication contains a detailed description of the virtual storage access method (VSAM). IBM System/370 OS/VSl Planning and Use Guide, GC24-S090 This publication describes OS/VSl. It can assist installation personnel in selecting and evaluating VS 1 and can assist system programmers in implementing, modifying, and extending the capabilities of the VS 1 control program. IBM System/370 OS/VS2 Planning and Use Guide, GC28-0600 This publication describes OS/VS2 Release 1. It introduces VS2 concepts and provides planning information for users planning to have VS2 Release 1 installed. Introduction to VTAM, GC27-6987 This publication describes the functions and capabilities that an installation needs to know to plan for installing the virtual telecommunications access method (VTAM). 4 OS/VS2 Planning Guide for Release 2 Contents Summary of Amendments . 9 Introduction . . . . . . VS2 Release 2 Highlights Multiprocessing Job Entry Subsystems. Data Set Handling . . System Resources Management System Activity Measurement Facility Allocation . . . . . . . . . . Multiple Virtual Address Spaces MVS Planning Considerations Functions Not Supported Unsupported Type I Programs 11 12 12 12 12 13 13 13 13 14 15 17 Defining the System . . . . . . System Generation Procedures . The VS2 Release 2 Starter System Minimum Configuration . . . Physical Addresses for Devices Stage I Macro Instructions Job Entry Subsystem Generation JES2 Generation JES3 Generation RMTGEN System Generation Changes Initializing the New System Data Sets Adding Installation-Written Members to System Libraries Unsupported VS2 Release 1 Stage I Macro Instructions System Initialization. . . . . . . . . . . . MVS System Parameters . . . . . . . . . . . . Changes to System Parameters for MVS . . . . Changes to VS2 Release 1 System Parameters Unsupported VS2 Release 1 System Parameters Unsupported MVT System Parameters System Parameter Specification. . . . . System Parameter Library (SYS1.PARMLlB) Changes to SYS I.PARMLIB Lists . IBM-Created Lists. . . . Installation-Created Lists. . . System Libraries and Data Sets. . . Required Libraries and Data Sets Optional Libraries and Data Sets Changes to VS2 Release 1 System Libraries and Data Sets Changes to MVT System Libraries and Data Sets . . . . Adding Paging Space to the System . . . . . . . . . . Creating Temporary Data Sets in External Page Storage Using Virtual I/O 19 20 21 21 21 22 23 24 25 25 26 28 28 29 30 30 36 36 37 38 39 39 40 41 45 47 47 48 49 50 50 51 Directing the Use of System Resources System Resources Management. . . Workload Management Routines Resource-Use Routines . . . . . Control Function. . . . . . . . Workload Management Concepts Transaction. . . . . Performance Variable Service. . . . . Service Rate System Workload . . Installation Performance Specification. Performance Groups. . . . . . . . . Performance Objectives . . . . . . . Associating a Transaction With a Performance Objective The Relationship Between Workload Levels and Service Rates System Resources Manager Parameters IPS Parameters (IEAIPSxx) . . . . . . . . . . . . . . . . 53 54 54 54 56 56 57 57 57 57 58 58 58 59 61 61 65 65 Contents 5 Tuning Parameters (IEAOPTxx) System Activity Measurement Facility (MF /1) MF /1 Operation . . . . . . . . . MF/l Reports and Tracing Records CPU Activity Report Paging Activity Report. . . . . Workload Activity Reports . . . Performance Group Period Report Performance Group Report System Summary Report Channel Activity Report . . I/O Device Activity Reports SMF Recording . Tuning with MF/l . Tuning Example 67 System Integrity . . . . Installation Responsibility Areas of Concern . . . . User-Supplied Addresses for User Storage Areas User-Supplied Addresses for Protected Control Blocks Resource Serialization . . . . . . Resource Identification . . . . . . . . . . SVC Routines Calling SVC Routines . . . . Control Program and User Data Accessibility Control Program Extensions. . . . Installation Procedural Responsibilities 79 80 81 81 81 82 83 84 84 85 86 Convers:ion Considerations . . . . . . Job Entry Subsystem Considerations The JES2 Job Entry Subsystem JES2 Job Flow . . . . . . MVS Operational Changes . JES2/HASP Compatability . The JES3 Job Entry Subsystem JES3 Scheduling Features Job Definition JES3 Job Flow . . . . . Device Scheduling. . . . Job Selection and Scheduling JES Operational Environment The Role of the JES3 Operator The Operator's Console . . Initialization, Control Statements, and Commands JES3/ ASP Compatability . . . . . . . . . . . SMF Considerations. . . . . . . . . . . . . . . . . Differences From VS2 Release t System Management Facilities. Differences From MVT System Management Facilities . SMF Record Differences for VS2 Release 2 New System Resources Manager Data . SMF Recording Replaced by MF /1 New Storage Data. . . . . . . . . Job Entry Subsystem Statistics . . . MUltiprocessing Data . . . . . . . Virtual Storage Access Method Data Time Sharing Record Elimination. . Changes to SMF Exits for VS2 Release 2 Job Control Language . . . . . . . . Changes to VS2 Release 1 JCL Performance Group Specification Assigning Job Selection Priority Specifying Data Sets in Paging Space (Virtual I/O Data Sets) Requesting Data Set Freeing at Data Set Close Time . . . . Specifying the Number of Resources That May Be Held for Reuse Requesting Multiple Copies of an Output Data Set . . Directing the Destination of an Output Data Set . . . Support for Tape and Direct Access Devices . . . . . Format Record Identification for the 3886 OCR Device Output Limit Specification . . . . . . . lncreased Data Set Definition Capability lncreased Volume Sharing Capability . . lnterpretation of the REGION Parameter 6 OS/VS2 Planning Guide for Release 2 68 68 70 70 70 71 71 71 72 72 72 72 73 73 89 90 90 91 93 95 96 98 98 99 lOO 101 102 103 lO3 104 lO4 lO6 lO6 lO6 lO6 109 109 109 110 110 110 llO llO t 12 114 114 114 114 115 115 115 115 116 116 116 116 116 116 Job Step Dispatching Priority Default Value . . . . Cbanges to MVT JCL . . . . . . . . . . . . . . . Requesting Storage for Program Execution. . . . . Specifying the BT AM Online Terminal Test Option JES2 JCL Considerations . . . . JES2 JCL for non-HASP Users . . JES2 Control Cards . . . . . Assigning Job Selection Priority JES2 JCL for HASP Users . . /*JOBPARM Control Card /*OUTPUT Control Card . JES3 JCL Considerations . . . . Communication Between Address Spaces Allocation Considerations . . Allocation Serialization Mount Characteristics . User-Assigned Names . Device Precedence List MVS Allocation Philosophy Reducing Allocation Contention Among Job Steps Operator Command Language . . . Changes to Operator Commands Multiprocessing Support . . Diagnostic Support VTAM Support . . . . . . Time Sharing (TSO) Support System Resources Management Support Miscellaneous Support . . . . . . . Job Entry Subsystem Support. . . . Time Sharing Considerations. . . . . . . System Resources Management Support VSAM Support . . . . . . Allocation Support . . . . . Job Entry Subsystem Support Edit Support. . . . . . . . Accounting Support Time Sharing Integration Support Catalog Support . . . . . . The VSAM Catalog . . . . . . VSAM Order of Search Creating or Updating the VSAM Catalog Updating an OS Catalog From a VSAM Catalog Catalog Utility and Macro Changes OS CVOL Support . . . . . . . . . . . Defining an OS CVOL for MVS Accessing CVOL Catalogs . . . . . . Compatability for OS CVOL Requests. Data Set Conversion . . . . . . . . Reformatting the UADS . . . . . . . . Formatting the Broadcast Data Set Programming Considerations With Virtual=Real Addresses Program Conversion TCAM Conversion. . . . . . . . . . Multiprocessing Considerations . . . . . . Hardware Configuration Considerations Operator Control of Configuration . Recovery Considerations . . . . . Alternate Path Retry . . . . . Dynamic Device Reconfiguration Alternate CPU Recovery . . Manual Switching . . . . . Programmer Considerations . . Loosely-Coupled Configurations 117 117 117 117 117 117 117 118 118 118 118 119 120 121 121 121 121 122 122 123 124 124 127 127 127 128 128 128 128 134 134 134 134 138 138 138 138 139 140 141 142 143 143 144 144 145 145 146 146 146 146 148 148 149 149 149 150 150 150 151 151 151 152 Appendix A: Virtual Storage Map . 153 Glossary 155 Index 161 Contents 7 Figures 8 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 1. 1. 2. 3. 4. S,. 6. 7. 8. 9. HI. 11. 12. 13. Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 14. 15. 16. 17. 18. 19. 20. 2L 22. 22. 23. 24. 25. 26. 27. 28. 29. 29. 29. 30. 3l. 32. 3,. 33. 3,. 34. 35. 36. 37. OS/VS2 Planning Guide for Release 2 Unsupported MVT Functions (Part 1 of 2) 15 Unsupported MVT Functions (Part 2 of 2) 16 Unsupported VS2 Release 1 Functions . . 16 Required Device Physical Addresses for the Starter System 22 Generating a VS2 Release 2 System . . . . . . . . . . 24 Changes in Stage I Macro Instructions . . . . . . . . . . . . . . . . . . 27 System Libraries to Which Installation-Written Members May Be Added 29 System Initialization Parameter Summary 31 Changes to SYS1.PARMLIB Lists . . . . . . . . . . . 40 Performance Objective Definition . . . . . . . . . . . 60 Associating a Transaction With a Performance Objective 61 Performance Objectives. . . . . . . . . . . . . . . . 62 63 System Workload Levels . . . . . . . . . . . . . . . Service Rates for Different Performance Objectives at Various System Workload Levels . . . . . . . . . . . . . 64 Tuning Example . . . . . . . . . . . . 74 Changing a Performance Objective 75 Effects of a Performance Objective Change 76 JES2 Processing Stages . . . . . . . . . 91 JES2 Differences From MVT HASP and VS2 Release 1 HASP 95 Sample JES3/ ASP Multiprocessing Configuration . . . . . 97 JES3 Job Flow . . . . . . . . . . . . . . . . . . . . . 99 JES3 Differences From MVT ASP and VS2 Release 1 ASP 105 Changes to SMF Records for VS2 Release 2 (Part 1 of 2) 107 Changes to SMF Records for VS2 Release 2 (Part 2 of 2) 108 Changes to SMF Exits for VS2 Release 2 . . 1111 JCL Changes for VS2 Release 2 . . . . . . 1] 3 Features Supported With JES3 Control Cards 1119 Communication Between Address Spaces . . 121 Relationship Between User-Assigned and Generic Names 122 Serialization Characteristics of Suggested Allocation Methods 123 Changes to Operator Commands for VS2 Release 2 (Part 1 of 3) 125 Changes to Operator Commands for VS2 Release 2 (Part 2 of 3) 126 Changes to Operator Commands for VS2 Release 2 (Part 3 of 3) 127 JES2 Command Functions 129 JES3 Command Functions . . . . . . . . . . . . . . . . . 130 JES3 Control Command Functions. . . . . . . . . . . . . . 130 Changes to Time Sharing Commands for VS2 Release 2 (Part 1 of 3) 135 Changes to Time Sharing Commands for VS2 Release 2 (Part 2 of 3) 136 Changes to Time Sharing Commands for VS2 Release 2 (Part 3 of 3) 137 Simplified Catalog Structure - OS vs. VSAM 140 Functions of the Access Method Services . 142 Method of Processing OS CVOL Requests 144 Virtual Storage Map . . . . . . . . . . 153 Summary of Amendments for GC28-0667-1 VS2 Release 2 Job Entry Subsystem Catalog Support The JES2 and JES3 job entry subsystems, generally compatible with HASP II and ASP Version 3 respectively, are described with particular reference to compatability to other systems. OS control volumes (CVOLs) are supported under VS2 Release 2. Allocation The system trace function is now available by command (TRACE). Diagnostic Support Allocation has been redesigned in an attempt to reduce contention for devices. V =R Programming Considerations System Resource Management Those considerations necessary for running with virtual addresses equal to real addresses are discussed. The workload management concepts are defined. Error Recovery Under Multiprocessing Virtual I/O I/O may be performed logically rather than physically, in a manner transparent to the user. Manual switching, alternate path retry, dynamic device reconfiguration, and alternate CPU recovery can occur with multiprocessing. Communication Between Address Spaces The interregion communication of previous systems can occur with less movement of data. Summary of Amendments 9 10 OS/VS2 Planning Guide for Release 2 Introduction OS/VS2 Release 2 (MVS) is a virtual storage operating system with multiprogramming, multiprocessing, time sharing (TSO), and job entry subsystems. In addition to providing significant new features, it also enhances existing OS/MVT and OS/VS2 facilities. OS/VS2 Release 2 is generally upward compatible with OS/MFT, OS/MVT, OS/VSl and OS/VS2 Release 1. Introduction 11 VS2 Release 2 Highlights Some of the significant new features provided in MVS include multiprocessing, job entry subsystems, new data set handling facilities, a centralized resource management facility, a system activity measurement facility, a virtual I/O paging mechanism for temporary data sets, and multiple virtual address spaces. These are described in the following sections. For a more detailed description of these features and of the MVS system, see the Introduction to VS2 Release 2. Multiprocessing The im~orporation of multiprocessing into the MVS control program provides for growth to larger and more complex systems. CPU s can be combined in two types of combinations: loosely coupled or tightly coupled. An installation can use combinations of loosely coupled and tightly coupled multiprocessing to tailor configurations to meet individual requirements. In a tightly coupled multiprocessing system, two CPUs share real storage, under a single control program, and communicate with each other by means of an interrupt capability. The system control program treats the CPUs as resources, assigning them to process tasks. Either JES2 or JES3 can be used to provide support for tightly-coupled mUltiprocessing. operatt:~ Loosely coupled multiprocessing systems connect VS2 Release 2 systems by means of channel to channel adapters that pass information between the processors. One processor controls job selection for all of the processors in the system. The JES3 job entry subsystem, a successor of ASP, provides the control for loosely coupled multiprocessing systems. One or more of the JES3 processors can, itself, be a tightly coupled mUltiprocessor. Job Entry Subsystems JES2 and JES3 provide the job processing functions of: • Reading local and remote jobs into the system. • Scheduling jobs. • Maintaining all data submitted with jobs. • Supporting the system management facilities. • Handling all output from batch jobs and time-sharing users. In addition, JES3 provides: • Workload sharing by processors within the JES3 complex. • Network job processing for workload sharing between remote VS2 systems. • Pre··execution fetch and setup of removable I/O volumes. • Scheduling based on job-completion deadline. • Scheduling based on completion deadline. • Scheduling based on completion of one or more jobs within a related set of jobs. Either JES2 or JES3 may be specified during system generation to run as a component of the system control program. 12 OS/VS2 Planning Guide for Releas{! 2 Data Set Handling The virtual storage access method (VSAM) is designed to offer function and flexibility to online and data base environments. In MVS, the key-sequenced VSAM system catalog structure is designed to reduce access time to very large catalogs. It permits the use of multiple private catalogs and the master catalog in one step, job, or session. Virtual I/O (VIO), is a new way of processing temporary data sets. It is accomplished on a page basis in virtual storage. The virtual telecommunications access method (VT AM) is a new direct-control teleprocessing access method with facilities available to applications programs, including those using TeAM. System Resources Management A new facility, called the system resources manager, is a centralized decision-maker that monitors a wide range of data about the condition of the system. It attempts to provide system resources to jobs in a manner specified by the installation. This allows an installation to specify that certain jobs get better response time than other jobs. System Activity Measurement Facility The system activity measurement facility (MF /1) is both a new and standard feature in MVS. It collects information about system activities and produces trace records and reports. MF /1 \s basically system oriented (as opposed to SMF which is basically job oriented). It is designed to aid in analyzing both system performance and system trends and in anticipating future system requirements. Allocation A new allocation design allows an installation to assign names to subsets of devices and to define a precedence list for allocating device types. The intent of the design is to reduce contention for devices (serialization) and allocation time by tailoring the allocation process. Multiple Virtual Address Spaces In VS2, virtual storage is an address range of 16,777,216 bytes. Release 1 of VS2 provided one virtual address range; jobs were assigned regions from this address range, just as jobs were assigned regions in real storage in MVT. MVS provides multiple virtual address spaces; each user (Le., batch job, time-sharing job, and certain system components) receives its own copy of virtual storage, minus the space used for certain system functions (e.g., the nucleus, the link pack area, the system queue area). The new virtual storage design provides the potential for greater multiprogramming, including the concurrent execution of more large jobs and a greater mix of time-sharing and batch jobs. Also, since each job is isolated in a private address space, inter-region virtual storage fragmentation is eliminated and protection features are extended. Introduction 13 MVS Planning Considerations To make use of some of the new or changed MVS facilities, it will be necessary to modify some existing procedures, libraries, data sets, and programs. The following list highlights examples of some of the more significant Release 2 changes that may require modifications to be made by an installation. Neither the list nor the examples are meant to be exhaustive; rather they call attention to areas of change that should be considered when planning to install the MVS system. The actual details of the changes are described in this publication in suqceeding chapters. • While the selection and specification of Stage I system generation options has been simplified, it is now necessary to select and specify job entry subsystem options. • MVS system initialization is designed to enable minimum operator intervention by extending the options that can be specified in SYS I.P ARMLIlB. • The system catalog structure is now a VSAM data set that can handle both VSAM and other data set types. Existing OS system catalogs (SYSCTLGs) are either supported as CVOLS, or can be converted to the new VSAM structure. • The new multi-function access method services are used to convert an OS system catalog to a VSAM master catalog. Access method services also replace most of the catalog functions of system utilities such as IEHMOVE, IEHLIST, IEHPROGM. • Paging space is contained in VSAM data sets in MVS. At least two page data sets with the required space must be defined during system generation. • System resources management routines allow the specification of relative rates for system resource scheduling for various categories of batch and time sharing jobs. The installation identifies the categories by assigning each of them a performance group number to be specified in JCL. The installation specifies the distribution rates and other tuning options in SYS I.P ARMLIB members. • The new system activity measurement facility (MF / I) collects and records information about system activities in either the SMF data set or on formatted and printed reports. Some SMF data, such as CPU or paging activity is now recorded only by MF / I on new SMF records. • System management facilities (SMF) records have been extended to record data provided by new features such as the system resources manager, the virtual storage access method, or job entry subsystems. Data reduction routines may have to be revised to process new records and new or changed record fields. • Job control language parameters have been added to specify items such as performance groups. Job entry subsystem control cards may be used to specify JES2 or JES3 job processing options. • The operator commands have been extended to support multiprocessing and changed for communication with the job entry subsystems. • The installation can name subsets of a generic set of devices in such a manner as to allow simultaneous allocation within the generic set. 14 OS/VS2 Planning Guide for Release 2 Functions Not Supported Figure 1 shows significant MVT functions that are not supported in VS2 Release 2. Figure 2 shows unsupported VS2 Release 1 functions. Unsupported MVT Function Comparable VS2 Release 2 Function Automatic SYSIN batching (ASB) reader Job entry subsystem (JES2 or JES3) BACKSPACE, NOTE, POINT, XDAP, and EXCP addressed to SYSIN/SYSOUT DCBs ---- Conversational remote job entry (CRJE) Time sharing (TSO) Dedicated work files Virtual I/O (VIO) Direct system output (DSO) ---- Graphic job processor (GJP) ---- IEBUPDAT utility program IEHIOSUP utility program IEHMOVE, IEHLIST, IEHPROGM IEBUPDTE utility program ---Access method services (most catalog functions) IMBMDMAP AMBLIST IMCJQDMP service aid Job entry subsystem and scheduler work area (SWA) dump routines IMCOSJQD service aid ---- Main storage hierarchy support ---- OS readers and writers Job entry subsystem (JES2 or JES3) and the external writer Queued telecommunications Telecommunications access access method (QT AM) method (TCAM) QT AM message processing programs TCAM application programs Remote job entry (RJE) JES2 or JES3 remote job entry Rollout/ rollin Paging Satellite graphic job processor (SGJP) ---- Scatter load Paging System environment recording Recovery management routines (SERO and SER 1) OS system catalog (SYSCTLG) support (RMS) VSAM master catalog with CVOL support Figure 1. Unsupported MVT Functions (Part 1 of 2) Introduction 15 Unsupported MVT Function Comparable VS2 Release 2 Function SYSl.ACCT data set System management facilities (SMF) data sets SYSl.SYSJOBQE Job entry subsystem input and output queues and the scheduler work area (SWA) Tape output for SMF ---- TESTRAN program --- - Time slicing System resources manager Transient areas Pageable link pack area TSO driver System resources manager TSO trace Generalized trace fadlity (GTF) -- Figure 1. Unsupported MVT Functions (Part 2 of 2) -Unsupported VS2 Release 1 Function Comparable VS2 Release 2 Function BACKSPACE, NOTE, POINT, XDAP, and EXCP ---- addressed to SYSIN/SYSOUT DCBs Directed swap for TSO - --- -- Access methods services IEHMOVE, IEHLIST, IEHPROGM (most catalog functions) Job entry subsystem and SW A dump routines IMCOSJQD service aid Nucleus-only SYSGEN --- - OS readers and writers Job entry subsystem and the external writer (JES2 or JES3) OS system catalog (SYSCTLG) VSAM master catalog With CVOL support Priority I/O queueing and ordered seek queueing ---- QT AM message processing programs TCAM application programs SYS 1.SYSJOBQE Job entry subsystem input and output queues and the scheduler work area (SWA) Time slicing System resources manager TSO driver System resources manager ---- TSO link pack area Generalized trace facility (GTF) TSO trace Figure 2. Unsupported VS2 Release 1 Functions 16 OS/VS2 Planning Guide for Releas(~ 2 -- Unsupported Type I Programs The following Type I programs which were supported by MVT and VS2 Release t are not supported by MVS (i.e., APARs written against these programs running on MVS will not be accepted). Corresponding program products are supported in their place. (Contact your local IBM branch office for information about these program products). Even though the followlng Type I language compilers are not supported under MVS, generated object cbde for an installation's programs that use any of these higher-level Type I languages, together with their corresponding library modules, will operate on this release of OS/VS2 subject to the constraints outlined in the "Program Conversion" section of this publication. Program Name Program Number COBOL F COBOL F Library Full ANS Cobol V2 Full ANS Cobol V2 Library FORTRAN G FORTRAN H FORTRAN G and H Library PL/t F PL/t F Library PL/ t F Syntax Checker SORT/MERGE 360S-CB-S24 360S-LM-S2S 360S-CB-S4S 360S-LM-S46 360S-FO-S20 360S-FO-SOO 360S-LM-SOt 360S-NL-Stt 360S-LM-St2 360S-PL-SS2 360S-SM-023 Introduction 17 18 OS/VS2 Planning Guide for Release 2 Defining the System During system generation and system initialization, the installation can define or change the operation of a standard or optional feature of the system control program to meet specific needs. This chapter provides planning information for defining the system. It is divided into the following sections: • System Generation Procedures r- • System Initialization • System Libraries and Data Sets Defining the System 19 System Generation Proced.ures System generation is the process of selecting data from IBM-distributed libraries and tailoring it to create an operating system for an installation. An I/O device-only system generation can also be performed. It is performed when only 110 devices and channels are to be added, deleted, or modified. The processor-only generation available with MVT and the nucleus-only generation available with both MVT and VS2 Release 1 are not available with MVS. The MVS system generation process is simplified over the MVT and VS2 Release 1 system generation processes in the following ways: • Many previously optional facilities are now standard. Therefore, fewer macro instructions and parameters need to be specified. • Several user-specified macro instructions have been clarified by consolidating parameters for related options under a single macro instruction. • Additional system initialization lists are available to be pre:formatted and taken from the SYSl.PARMLIB data set for use at IPL time. For MVS, IBM will generate (from installation-provided system generation statements) and install one system control program per uniprocessing or multiprocessing configuration. The IBM installation process also includes performing the installation verification procedure (IVP). There are six major steps in generating a new operating system. Planninn The distribution libraries containing the MVS modules must be ordered from IBM. The options must be selected that, with the standard features, will constitute the desired system. The macro instructions needed to specify the options and standard features of the new system must be selected and coded. Volume and data set initialization The direct access volumes that will contain the distribution libraries (and the starter system) must be initialized. The volumes on which the new system will be generated must also be initialized. The data sets on the system volumes must be allocated (unless they will be specified by SYSGEN macros to be assembled in Stage I for allocation in Stage II). Stage I The macro instructions are assembled and expanded into job control languagt!, utility control statements, assembler language statements, and linkage editor control statements. Stage n (Jobstream Execution) The modules from the distribution libraries are assembled~ link-edited, and copied to the data sets that have been allocated on the new system volume(s). Job entry subsystem generation The job entry subsystem modules can be assembled (JES2 only) and link-edited to reflect the options and limits specified for them. Job entry subsystem generation may be performed concurrently with Stage II. Additionally, 20 OS/VS2 Planning Guide for Release 2 remote station programs may optionally be generated (RMTGEN). For JES2 (most) and for JES3 (all) these options can be specified during initialization. Testing After the new system is generated, the IBM Program Systems Representative uses the installation verification procedure (IVP) to verify that the new system is operational on the hardware configuration. The VS2 Release 2 Starter System The first time that MVS is generated, the IBM starter system is required to drive the system generation processing. The starter system consists of an MVS operating system that uses JES2 as the job entry subsystem. After the first system generation, either the starter system or the new MVS control program can be used to drive subsequent system generations. Minimum Configuration Use of the starter system requires the following minimum configuration: • One IBM System/370 Model 145, 15511, 158, 16511, or 168. (The clock comparator, the CPU timer feature (#2001) and the advanced control program support feature (#1001) are required on the Model 145.) • 768K bytes of real storage. • One multiplexer channel. • One selector or block multiplexer channel. • Three IBM 3330 spindles, or four IBM 2314/2319 Direct Access Storage Facility devices. • One card reader. • One hardcopy console or one hardcopy device with a display console. • One 9-track magnetic tape. • One 1403 or 3211 printer. Physical Addresses for Devices The starter system is configured to support the devices at the addresses specified in Figure 3. Defining the System 21 Function Console Reader Punch Tape Printer Direct Access Storage Device Device Multiplexer Channel Channel 1 Channel 2 158 Console 014,016 3213 015,017 3210/3215 009,01F 3066 040,060 OCO,OOO OEO,OFO 2540 OOC 20C 3505 012 20A 2540 000 200 3525 013 206 Channel 3 Channel 4 -209,21F 140 240 340 440 .. .. 2400 (7-TR-OC) 180, 181 280,281 380,381 480,481 2400 (9-TR) 3400 182, 183 184 282,283 284 3B2,383 482,483 3211 002,004 1403 OOE,OOF 20E 2305-1 1FO 2FO 2305-2 100 200 2314/2319 130, 131 132, 133 134,135 230,231 232,233 234,235 330,331 332,333 3330 150, 151 152, 153 250,251 252,253 350,351 352,353 3330-1 158, 159 15A,156 258,259 25A,256 358,359 35A,356 3340 1CO, 1C1 1C2,1C3 2CO,2C1 2C2,2C3 3CO,3C1 3C2,3C3 Figure 3. Required Device Physical Address~s -- ". -- for the Starter System Stage I Macro Instructions The Stage I macro instructions are used to tailor the new system. They describe the machine configuration, the control program, the data management routines, and the user-written routines. They are also used to specify the program options that are to be included in the system. The macro instructions that can be specified in Stage I of MVS are: AFFINITY specifies program names (e.g., emulators) and the identification of the CPU on which each program is required to execute. CHANNEL specifies the channel characteristics. A CHANNEL macro instruction is required for each channel i~ the installation's computing system. 22 OS/VS2 Planning Guide for Release 2 CKPTREST specifies standard system completion codes not eligible for automatic restart, and user completion codes that are eligible for automatic restart. If the CKPTREST macro instruction is not specified, the standard set of codes will be used. CONSOLE specifies the console configuration for the system. CTRLPROG specifies control prograI)} options. If the CTRLPROG macro instruction is not specified, the default values are assumed. DATAMGT specifies optional access methods (BT AM, ISAM, GAM, TCAM and VT AM) to be included in the system. DATASET specifies system data sets and the volumes on which they reside. It is required for each defined data set that is not located on the system residence volume. It is also used to specify user-written macros, routines, and procedures to be added to system libraries. EDIT specifies the physical characteristics and processing attributes of the data sets to be processed by the time sharing EDIT command. GENERATE specifies the data sets, volumes, and I/O devices required for the SYSGEN process, the SYSGEN output options, and the type of generation being performed. IODEVICE specifies characteristics of an I/O device and its operating system requirements. An IODEVICE macro instruction is required for each uniquely addressable I/O device. SCHEDULR specifies job and master scheduler options. SVCTABLE specifies the number, type, APF authorization, and entry status of the user-written SVC routines to be added to the new system. TSO specifies the inclusion of IBM -supplied time sharing command processors. UNITNAME specifies a name for a group of I/O devices and optionally makes the name available for VIO use. A UNITNAME macro instruction is required for each named group of I/O devices in the system, except for unique device types. Job Entry Subsystem Generation Job entry subsystems (JES2 or JES3) are added to the system through the process of job entry subsystem generation, which may be executed concurrently with Stage II. Figure 4 summarizes the steps that are involved in performing a job entry subsystem generation and relates them to the steps that are required to perform a system generation (SYSGEN). Defining the System 23 Step SYSGEN JES2 Generaticm JES3 Generation Planning (selecting options) x x Volume and data set initialization x x x Stage I x Stage \I (jobstream execution) x x x x x x x Stage \I and JES2 generation may be executed concurrently. RMTGEN (optional) Testing x Figure 4., Generating a VS2 Release 2 System JES2 Generation The distribution libraries needed to perform a JES2 generation are distributed by IBM along with the MVS distribution libraries. They contain: • The JES2 source code. • The utilities used by the JES2 jobstream. During JES2 generation, the JES2 control program is assembled to reflect the options and limits that the installation has chosen. The installation takes the following steps to perform a JES2 generation. Planning The JES2 and optional RMTGEN parameters must be chosen and coded. VoIUlD4~ and data set initialization The space required for JES2 data sets must be allocated on the new system volum{!s. Jobstrcam Execution The jobstream on the distribution library must be executed to apply options, assemble, link-edit, and copy the JES2 modules from the distribution library to the data sets that have been allocated on the new system volumes. The jobstream also invokes RMTGEN if any remote terminals were specified by the JES2 generation parameters. JES2 Generation Options: The following characteristics of the JES2 control . program may be specified: • The maximum number of jobs, started tasks, and time sharing users in the JES2 subsystem at anyone time. • The number of card readers, printers, and card punches to be used by the job control manager. • The: maxim-qm number of initiators that may be started at anyone time. • The default limits for a job's printed or punched output and the actions to be taken when the specified limits are exceeded. 24 OS/VS2 Planning Guide for Release 2 • The type of processing for SYSOUT data sets. • The number of direct access storage volumes that may be mounted concurrently as spooling devices. • Many parameters (or their equivalents) formerly specified only during HASPGEN, can now be specified during JES2 initialization. JES3 Generation The initial release of JES3 is distributed as a separately order able independent component. It contains: • A source library containing all JES3 programs and macros. • An object deck library containing assembled JES3 modules suitable for link editing. • A sample program library containing test decks and such procedures as the method for adding the JES3 ICR to the existing MVS system. • A sample JES3 initialization deck. Because JES3 options and limits are selected at initialization time, it is not necessary to specify them at system generation. RMTGEN parameters must be coded. RMTGEN In addition to generating a job entry subsystem, one or more remote station programs may also be generated. The process of generating job entry subsystem multileaving remote terminal programs is called RMTGEN. The parameters used during RMTGEN specify hardware configuration and software options. Remote terminal programs that can be generated by RMTGEN include: • System/360 and System/370 binary synchronous communication (BSC) remote terminal program. • System/360 model 20 remote terminal program. • 1130 remote terminal program. • 1130 remote terminal program load program. • System/3 remote terminal program. Note: Remote terminals can be supported only by 2701 and 2703 emulators on the 3704 and 3705. RMTGEN Options: The following remote terminal program characteristics may be specified: • The type and degree of text compression. • The amount of main storage available to the program. • The number of buffers to be used by the program. • The inclusion or exclusion of unit record devices. • The unit address of the BSC adapter and devices. • The speed of the communication line used. Defining the System 25 • The number of buffer areas to be used for restoring compressed lines to their original form. • The type of assembly listing to be produced during RMTGEN. System Generation Changes Some of the major changes to the system generation processes for MVT and VS2 Release 1 include: • Changes to the process of initializing new system data sets. • The reduction of the number of macro instructions that need to be specified during Stage I (see Figure 5). This reduction is due both to the fact that previously optional facilities are now standard and to the consolidation of related options under single macro instructions in MVS. For example, it is no longer necessary to specify two different SYSGEN macro instructions (RESMODS and LINKLIB) to add installation-written routines to the nucleus and the link library. The parameters on these macro instructions have been consolidated under the DATASET macro instruction which is now used to specify the addition of installation-written routines to any system library. (See the topic "Adding Installation-Written Members to System Libraries.") • The addition of a job entry subsystem generation. • The addition of parameters for specifying authorized appendages. 26 OS/VS2 Planning Guide for Release 2 MVT -- -- AFFINITY * ** *** *** *** ASSEMBLR CENPROCS CHANNEL CENPROCS CHANNEL *** *** -- CHANNEL *** *** CKPTREST DATASET CKPTREST DATASET COBLIB COBOL CTRLPROG *** *** *** *** CTRLPROG CTRLPROG DATAMGT DCMLlB EDIT DATAMGT DATASET EDIT DATAMGT DATASET EDIT EDITOR EMULATOR FORTLIB EDITOR FORTRAN GENERATE GENTSO GJOBCTL GRAPHICS HELP -*** --*** *** *** GENERATE GENERATE --- GRAPHICS DATASET --- DATAMGT DATASET IMAGELIB IOCONTRL IODEVICE -- -- L1NKLIB LOADER MACLIB L1NKLIB LOADER MACLIB DATASET DATASET OUTPUT PARMLIB PL1 TSO PARMLIB TSO DATASET *** *** PL1 LIB PROCLIB PTOP *** *** DATASET DATASET RESMODS RPG SCHEDULR *** VS2 Release 2 ALGLlB ALGOL CHECKER CKPTREST CMDLlB -- VS2 Release 1 DATASET IODEVICE DATASET IODEVICE -- *** *** RESMODS DATASET *** ** * SCHEDULR SCHEDULR SECMODS SECONSLE SORTLIB CENPROCS SECONSLE CONSOLE *** *** SORTMERG SUPRVSOR SVCLlB *** *** DATASET DATASET SVCTABLE SYSUTILS TELCMLlB SVCTABLE SVCTABLE -- --- ** * *** DATASET DATASET TSOPTION UCS UNITNAME TSO UCS UNITNAME TSO -- PAGE DATASET -- UNITNAME no longer specified during system generation the function specified by this macro instruction may be added to the system after system generation Figure 5. Changes in Stage I Macro Instructions Defining the System 27 Initializing the New System Data St~ts New system data sets must be initialized before Stage II is executed. Initialization includes: • allocating space for system data sets on the new system volumes. • building the volume index of the system catalog. • cataloging the data sets in the new system catalog. In MVS, the new system data sets can be initialized either during system generation by using the DATASET macro instruction or at any other time before Stage II by using the access method services. The Stage I DATASET macro instruction did not exist in MFT or MVT. Its use for initializing new system data sets during SYSGEN remains unchanged from VSl Release 1 and VS2 Release 1. A DATASET macro instruction is specifi{~d with the SPACE parameter in the Stage I input deck for every system data set that is to be allocated. The new system data sets are then allocated and cataloged during Stage II execution. In MVS, the IEHPROGM utility program can be used to catalog only non-VSAM data sets without CVOL pointers. In order to preallocate all new system data sets (VSAM and non-VSAM data sets) the access method services should be used. In this case a DATASET macro instruction should be specified without the SPACE parameter for any system data set used during system generation that is not on the system residence volume. This enables the catalog entry to be created during Stage II. Adding Installation-Written Members to System Libraries The DATASET macro instruction is used during SYSGEN to include installation-written members in system libraries. The parameters: PDS=SYS1.name MEMBERS=(name, name ... ) are functionally equivalent to their counterparts on the RESMODS and LINKLIB SYSGEN macro instructions in MVT and VS2 Release 1. A new feature is provided for adding installation-written routines to the nucleus, however. If only the PDS parameter is specified, the members in the data set specified by PDS will be copied into the nucleus as individual members. If both the PDS and MEMBERS parameters are specified, the members in the data set specified by PDS will be copied into the nucleus and link-edited together with the nucleus member IEANUCOl. Figure 6 indicates the system libraries to which installation-written members may be added and the form in which they must be added. 28 OS/VS2 Planning Guide f or Relea~.e 2 System Library Form SYSl.CMDLIB SYS l.IMAGELIB SYS 1.LINKLIB load module load module load module SYS I.LPALIB SYS 1.MACLIB SYSl.NUCLEUS load module system macro load module SYS I.PARMLIB SYS 1. PROCLIB SYS 1. TELCMLIB system parameter system procedure load module SYS 1. VT AMLIB SYS 1. VT AMLST load module , system parameter Figure 6. System Libraries to which InstaUation-Written Members May Be Added Unsupported VS2 Release 1 Stage I Macro Instructions The following VS2 Release 1 macro instructions are not supported in MVS. CENPROCS Model dependent information is no longer needed during system generation processing. EDITOR All functions provided by this macro instruction are available in SYS 1.PROCLIB cataloged procedures. GRAPHICS The DAT AMGT macro instruction now performs the function. LINKLIB The DATASET macro instruction now performs the function. LOADER Options specified by this macro instruction are now available in SYS 1.PROCLIB cataloged procedures. MACLIB The DATASET macro instruction now performs the function. PAGE The DATASET macro instruction now performs the function. PARMLIB The DATASET macro instruction now performs the function. RESMODS The DATASET macro instruction now performs the function. SECONSLE The CONSOLE macro instruction now performs the function. UCS The options specified by this macro instruction are standard in the MVS system. Defining the System 29 System Initialization Before any job can be processed, the control program and its associated control blocks and work areas must be loaded into storage and prepared for operation. Loading these control program modules is the function of the initial program loader (IPL). After the IPL program completes its loading functions, control passes to the nucleus initialization program (NIP), which performs more functions necessary for the operation of the control program. (For example, NIP defines storage areas and initializes certain tables, work areas, and control blocks.) NIP also loads and initializes routines (such as fixed LPA routines) selected by the user. The nucleus initialization program also initializes real storage. It gives the installation control over virtual storage definition by processing installation-supplied system parameters that specify virtual storage options. Master scheduler initialization completes the initialization process by establishing system tasks such as the console communications task, the log task, job entry subsystems, system management facilities, the recovery/termination management task and the missing interruption checker task. In MVS, changes have been made to the system initialization process to provide the installation with greater flexibility in specifying system options as well as the opportunity to simplify the IPL process. These changes include: • The reduction of informational IPL messages. • The elimination of the time-of-day (TOD) clock messages and associate,d replies, unless the installation requests that they be issued or unless there is :an error in TOD clock processing. • The implementation of mUltiple SYS I.P ARMLIB members for SMF, volume attributes, commands, and the link library. • The creation of a SYS l.P ARMLIB member that contains installation-specified operator commands that are issued by the system control program at the completion of system initialization processing. The operator need respond only in case of error conditions. MVS System Parameters The installation uses system parameters to control the system initialization process. System parameters may be specified through the master console or in the IEASYSxx lists in the system parameter library, SYS I.PARMLIB. Each parameter in a SYSl.PARMLIB list is specified in the same format as that required by the "SPECIFY SYSTEM PARAMETERS" message. Figure 7 provides a summary of these parameters. More detailed descriptions of the parameters follow. 30 OS/VS2 Planning Guide for Release 2 Parameter SYS1.PARMLIB List Read Value Specified APF Authorized library name @ APG Automatic priority group for system resources manager • BLDL Pageable directory for SYS1.LlNKLIB BLDLF Nonpageable directory for SYS1.LlNKLIB CLPA New link pack area to be created IEABLDxx • • IEABLDxx IEALODOO IEAPAKOO CMD Command to be issued internally @ CSA Size of the common service area @ CVIO Delete all VIO data sets from paging space @ DUMP Data sets for SYS1.DUMP FIX Reenterable routines for nonpageable LPA HARDCPY Hard copy log IPS I nstallation performance specification @ IEAIPSxx LNK Names of data sets concatenated to SYS1.LlNKLIB @ LNKLSTxx LOGCLS Output class for log data set @ LOGLMT WTL limit for log data set @ MAXUSER Maximum number of virtual address spaces @ MLPA Modifications to pageable LPA NUCMAP New nucleus map for DSS OPI SYS1.PARMLIB operator intervention restrictions • OPT System resources manager tuning parameters @ PAGE Page data set names • • REAL V = R address area size SMF SM F parameters SQA Size of the system queue area SYSP • • COMMNDxx IEAFIXxx IEALPAxx @ IEAOPTxx @ SMFPRMxx System parameter list to be merged with I EASYSOO • IEASYSxx VAL Volume characteristics @ VATLSTxx VRREGN Default region size for a V= R request @ WTOBFRS Number of buffers for WTO routine use @ WTORPLY Number of operator reply elements for WTOR routine use @ • -@ IEAAPFxx Not Included in MVT Systems -- Not Included in VS2 Release 1 Systems and MVT Systems Figure 7. MVS System lnitiaUzation Parameter Summary Defining the System 31 APF specifies the value that is appended to IEAAPF to form the name of the SYS l.P ARMLIB list that contains data set names with their corresponding volume identifications of authorized program libraries. The SYS l.LINKLIB and SYS l.SVCLIB libraries are automatically included as authorized libraries. APG specifies an automatic priority group for use by the system resource management routines. If the APG parameter is not specified, a default value of 7 will be used. BLDL speeifies a value that is appended to IEABLD to form the name of a SYS l.P ARMLIB list that contains the names of modules for which entries are to be built in the pageable BLDL table. The directory created in response to the BLDL parameter will reside in page able storage. If the BLDL and BLDLF parameters are both specified, the BLDL parameter will be ignored. A warning message will be issued to identify this condition. The IEABLDOO list is the default list. BLDLF specifies a value that is appended to IEABLD to form the name of a SYS l.P ARMLIB list that contains the names of modules for which entries are to be built in the resident BLDL table. Thl~ directory created in response to the BLDLF parameter will reside in fixed storage. specified~ the BLDLF parameter will be accepted. A warning message will be issued to identify this condition. If the BLDLF and BLDL parameters are both The IEABLDOO list is the default list. CLPA specifies that a new pageable link pack area is to be created. If a previously created link pack area exists, it is deleted and the space it oceupied is made available for system paging use. If CLP A is specified, CVIO is automatically specified. CMD specifies a value that is appended to COMMND to form the name of the SYS l.PARMLIB list that contains the commands that the installation selects to be issued internally as soon as the system initialization process is complete. If not specified in a system parameter list of by the operator, the COMMNDOO list is used. CSA specifies the common service area. The value specifies the size, in lK byte m~ltiples, of the common service area. If CSA is not specified, the default value of lOOK bytes is used. 32 OS/VS2 Planning Guide for Release 2 CVIO for system restart, specifies that all virtual I/O (VIO) data sets are to be deleted from the paging space. CVIO is automatic if CLP A is specified. DUMP specifies tape volumes and/or the direct access storage devices to be used for the SYS 1.DUMP data sets. FIX specifies a value that is appended to IEAFIX to form the name of the SYS 1.P ARMLIB list that contains the names of reenter able routines from the SYS1.LINKLIB, SYS1.LPALIB, and SYS1.SVCLIB data sets to be made resident as a fixed extension to the link pack area (fixed link pack area). If not specified, a fixed link pack area is not created. HARDCPY specifies that the hard copy log is desired, whether or not the system log is to be used as the hardcopy log, and whether or not messages are to be recorded on the hardcopy log. If the console configuration contains an active graphic console or more than one active console, a hardcopy log is required. In this case, if routing codes 1,2,3,4,7,8,or 10 are desired, they need not be specified since all of these codes are automatically assigned by the system control program. IPS specifies a value that is appended to IEAIPS to form the name of the SYS 1.P ARMLIB member that contains the installation performance specification (IPS). The IEAIPSOO is the default list. LNK specifies a value that is appended to LNKLST to form the name of the SYS 1.PARMLIB list that contains the names of the data sets to be concatenated to SYS 1.LINKLIB. Any number of SYS 1.P ARMLIB lists may be specified in this manner until the number of data sets contained reaches 15. (The maximum number of data set names other than SYS 1.LINKLIB is 15. This allows 16 data sets including SYS1.LINKLIB to be concatenated.) SYS1.LINKLIB is opened first. Then each data set that is specified is opened and concatenated in the order in which it appears (Le., left to right). The LNKLSTOO list is the default list. LOGCLS specifies the JES2 output class for the log data sets. If the LOGCLS parameter is not specified, a default class of A is used. LOGLMT specifies the number of write-to-Iog (WTL) instructions that may be issued before the log data set is scheduled for output processing. If the LOGLMT parameter is not specified, a default value of 500 WTL instructions is used. Defining the System 33 MAXUSER spedfies the maximum number of independent virtual address spaces available for use by the operating system. If the MAXUSER parameter is not specified, a default value of 256 is used. The maximum number that may be specified is 1,536 minus the number of active VIO data sets. MLPA speeifies a value that is appended to IE ALP A to form the name of the SYS 1.P ARMLIB list that contains the reenterable routines from SYSl.LINKLIB, SYSl.LPALIB, and SYSl.SVCLIB data sets to be made resident as a pageable extension to the link pack area (modified link pack area). If the MLP A parameter is not specified, a modified link pack area is not created. NUCMAP specifies the creation of a new map of the nucleus in the SYS 1.DSSVM data set and the deletion of the existing nucleus map. The parameter must be specified if the nucleus for which a map exists is modified and an IPL is performed using the modified nucleus. The NUCMAP parameter need not be specified again until another modification is made to the nucleus. A new nucleus map will be created without the specification of the NUCMAP parameter, if the name of the nucleus load module changes between IPL procedures or if a nucleus map does not exist for the first IPL after system generation. OPI specifies operator intervention restrictions of the system parameters in the SYS 1.P ARMLIB lists. The OPI parameter can only be included in SYS 1.P ARMLIB lists. It may be specified for individual system parameters or for the entire set of system parameters. OPT sp(;:cifies a value that is appended to IEAOPT to form the name of the SYSl.PARMLIB list that contains tuning parameters for the system resources management routines. If not specified in a system parameter list or by the operator, default values provided by the system resources management routines are used. PAGE specifies one or more page data set names to be used for a particular system IPL, in addition to the page data set names in the PAGE parameter specification of P ARM LIB members IEASYSOO or IEASYSxx. The first data set name is used to contain as much of the PLP A as space permits. The values specified include identification of the VSAM data sets for which space was previously allocated and formatted (and, optionally, the unit on which the volume containing the specified data sets should reside). 34 OS/VS2 Planning Guide for Release 2 REAL specifies nonpageable storage (virtual equals real address area). The value specified is the number of IK byte blocks to be reserved for nonpageable (V =R) storage. If the value specified is 0, there will be no V =R storage in the system. If the REAL parameter is not specified, a default of 76K will be used. SMF specifies a value that is appended to SMFPRM to form the name of the SYS I.P ARMLIB list that contains system management facility (SMF) values. If not specified in a system parameter list or by the operator, the SMFPRMOO list is used. SQA specifies the system queue area. The value specified is the number of 64K byte segments to be added to the minimum size of the system queue area. If not specified in a system parameter list or by the operator, a default value of one (64K byte segment) is added to the minimum size of the system queue area. If the specified amount of virtual storage is not available, a warning message will be issued. The SQA parameter may be respecified at that time. The size of the system queue area cannot be changed during a restart that reuses the previously initialized link pack area. If the SQA parameter is specified on the restart and the size is different, the system uses the previous size and issues a warning message. SYSP specifies' the SYS l.P ARMLIB list of system parameters that is to be merged with the default list of system parameters, IEASYSOO. The value specified is appended to IEASYS to form the name of a SYS I.P ARMLIB list that contains the system parameters to be used for the current IPL. Any number of IEASYSxx lists may be specified. If not modified by this parameter, only the IEASYSOO list is used. The IEASYSOO list contains the values selected during system generation for various system parameters unless the contents of the list have been modified by the system programmer at the installation. When multiple IEASYSxx lists are specified by the SYSP parameter, they are treated in ascending priority sequence (Le.,parameters defined in a member specified near the beginning of the list will be replaced by the same parameters defined in a member specified near the end of the list). In the process of merging parameters with the IEASYSOO list, IEASYSOO is always read, and always assumes the lowest priority. Defining the System 35 For I!xample, if the operator responds: R 00, 'SYSP=( 01 ,02 ) , and these two IEASYSxx lists contain the following system parameters: IEASYSOl: MLPA=( 00,01 ) , BLDL=OO IEASYS02: MLPA=02, SQA=1 0 then the effective system parameters will be: MLPA=02, BLDL=OO, and SQA=1 0 VAL specifies a value that is appended to V ATLST to form the name of the SYS l.P ARMLIB list that contains mount and allocation characteristics for any or aU of the direct access volumes used by the installation. If not specified in a system parameter list or by the operator, the V ATLSTOO list is used. VRREGN specifies nonpageable storage (virtual equals real address area) for a specific job. The value specified becomes the default value for the real main storage size (in lK multipltes) to be allocated to a job requiring V =R address space. If the VRREGN parameter is not specified, a default value of 64K is used. The VRREGN system parameter should not be confused with the REAL system parameter. REAL specifies the total amount of nonpageable (V =R) storage available to the system. VRREGN specifies the subset of the total spaee that is available for an individual job. WTOBFRS speeifies WTO buffers. The: value specifies the number of buffers to be used by the write-to-operator (WTO) routines. If the WTOBFRS parameter is not specified, a default value of 20 is used. WTORPLY specifies WTOR elements. The value specifies the number of operator reply elements to be used by the write-to-operator-with-reply (WTOR) routines. If the WTORPLY parameter is not specified, a default value of 5 is used. Changes to System Parameters for MVS Figure 7 shows new system parameters for MVS. Some VS2 Release 1 and some MVT system parameters are not supported. Of these, some have been replaced by new parameters and others have been eliminated, either because their functions are not supported in MVS or because their functions are now performed by the system control program. Chan~:es to VS2 Release 1 System Parameters: The following VS2 Release 1 system parameters have changed for MVS. 36 OS/VS2 Planning Guide for Relea!ie 2 APG The only value that can be specified for the APG parameter in MVS is the priority of the automatic priority group. DUMP Ten SYS I.DUMP data sets can be specified in MVS. PAGE Values specified by the PAGE parameter have changed for MVS because paging space is made up of preformatted VSAM data sets, two of which must be defined and formatted during system generation, (See the topic "Adding Paging Space to the System" in this chapter.) The PAGE parameter also no longer specifies a particular volume to contain the link pack area. REAL In MVS, the value specified by the REAL parameter represents the total amount of V=R storage in lK byte multiples, as opposed to the number of 4K byte blocks added to a 64K byte minimum value in VS2 Release 1. Unsupported VS2 Release 1 System Parameters: The following VS2 Release 1 system parameters are not supported in MVS but are provided for by comparable functions. LSQACEL This parameter is not needed in MVS because quickcells for the local system queue area are controlled by virtual storage management routines. PAL The values specified by this parameter are supplied by the address space supervision routines in MVS. SQACEL This parameter is not needed because quickcells for the system queue area are controlled by virtual storage management routines. TMSL This parameter is not needed because the system resources management routines determine the acceptable time slices. The following VS2 Release 1 system parameters are not supported in MVS and have no comparable MVS functions. AUXLIST This parameter is not needed because in MVS it is not meaningful to distinguish between the auxiliary storage available for background and time-sharing jobs. CPQE The number of channel programs to be used by the address space supervision routines is no longer an installation option. MPA It is not necessary to specify the size of the master scheduler region because in MVS, the master scheduler task has its own independent virtual address space. TSOAUX It is not necessary to specify back-up auxiliary storage space for time sharing jobs in MVS, because each time sharing user is assigned to an individual virtual address space. Defming the System 37 Unsupported MVT System Parameters: The following MVT system parameters are not supported in MVS, but are provided for by comparable MVS functions. MIN This parameter is not needed in VS2 because the initiator modules are contained in the pageable link pack area. MOD The CPU model identification is determined in VS2 by the STIDP instruction. RAM Routines from SYSl.SVCLIB and SYSl.LINKLIB can be: added to the VS2 link pack area by means of the SYS l.LP ALIB data set 01' the MLP A or FIX system parameters. RERP In VS2, error recovery procedure (ERP) routines are always resident in the link pack area. RSVC In VS2, SVC routines are always resident in the link pack area. SQS This parameter is replaced by the SQA system parameter . TMSL This parameter is not needed because the system resource management routines determine the acceptable time slices. The following MVT system parameters are not supported in VS2 Release 2 and have no comparable MVS functions. ALTSYS This parameter is not needed because the system residence devices are t.reated by dynamic device reconfiguration (DDR) as any other swappable device. HRAM Mvr hierarchy support is not needed in VS2. HSVC MVT hierarchy support is not needed in VS2. MPS It is not necessary to specify the size of the master scheduler region because in MVS, the master shceduler task has its own independent virtual address space. QBF This parameter is not needed in MVS because the job queue is replaced by the scheduler work area (SW A) and job entry subsystem data sets. System Parameter Specification An installation can tailor a system through the specification of system parameters, either during system generation (SYSGEN) processing or in the IEASYSOO or IEASYSxx lists of SYS l.P ARMLIB. The system operator specifies 38 OS!VS2 Planning Guide for Releast! 2 system parameters through the master console in response to the "SPECIFY SYSTEM PARAMETERS II message. Proper use of IEASYSOO or IEASYSxx can result in a significant saving of time. The IEASYSOO (system parameter) list is always created during the SYSGEN process, and the system parameters HARDCPY, PAGE and either BLDL or BLDLF are included in it. The installation may also specify that the system parameters CSA, REAL, SQA, VRREGN, APF=OO, and FIX=Ol should be included in the IEASYSOO list during SYSGEN processing. The remainder of the system parameters are specified in two ways: • By using a system utility program to ad,d them to the IEASYSOO or any IEASYSxx list. • By entering them through the operator's console. The IEASYSOO list is read for every IPL. The installation is not only able to add parameters to this list, but may change parameter values that were included in it during SYSGEN processing. Operator interaction will be minimized during system initialization processing if the installation ensures that the IEASYSOO list contains all of the system parameters required to establish the desired operating environment. The system operator will then be able to respond to the I I SPECIFY SYSTEM PARAMETERS" message with end-of-block (EOB), and will not be required to respond further unless error conditions are detected and prompting occurs. The IEASYSOO member should generally contain installation default values, or parameters for which values can be defined that do not change from IPL to IPL. The IEASYSxx members, on the other hand, should be used to contain parameters that are subject to change from IPL to IPL, depending on the user applications. (See the SYSP parameter description for additional information.) System Parameter Library (SYSl.PARMLIB) The SYSl.PARMLIB data set contains both IBM-created and installation-created lists of parameter values. The formats for the lists have been revised from MVT to allow better space utilization within parameter records, and to afford more flexibility in their definition. All MVT and VS2 Release 1 parameter lists that are still valid lists in MVS need not have their formats changed. Defining the System 39 Changes to SYS I.PARMLIB Lists Figure B shows the new or changed MVS SYS l.P ARMLIB lists. List Change from MVT Change from VS2 Release 1 COMMNDxx IEAABDOO IEAAPFxx New New New New New New IEAAPPOO IEABLDxx IEADMPOO New New New New None New IEAFlXxx IEAIGEOO IEAIGGOO IEAIPSxx IEALODOO New No longer supported No longer supported New New None No longer supported No longer supported New None IEALPAxx IEAOPTxx IEAPAKOO IEARSVOO New New New No longer supported None New None No longer supported IEASYSxx IKJPRMOO IQAORDER IRBMFlxx New Now only TIOe parameters Not applicable New None Now only TIOe parameters No longer supported New LNKLSTxx PARMTZ SMFPRMxx No longer a single member New No longer a single member named SMFDEFLT No longer a single member named PRESRES No longer a single member None No longer a single member named SMFDEFLT No longer a single member named PRESRES VATLSTxx Figure 8. MVS Changes to SYS1.PARMLIB Lists 40 OS/VS2 Planning Guide for ReleaSE! 2 IBM-Created Lists The IBM-created lists that are named in column 1 are standard in SYS l.PARMLIB for MVS. Standard List Alternate List Address of areas to be dumped by SYSABEND IEAABDOO IEABLDOO IEABLDxx Names of SYSt.LINKLIB load modules whose directory entries are to made in the BLDL table Addresses of areas to be dumped by SYSUDUMP IEADMPOO IEAIPSOO Contents IEAIPSxx Installation performance specification IEALODOO Names of SYS1.LPALIB modules whose directory control block information is to be maintained in nonpageable storage IEAPAKOO Groups of names of SYS1.LPALIB modules that are to be packed together in the link pack area IEASYSOO IEASYSxx Lists of system parameters LNKLSTOO LNKLSTxx Names of additional data sets that may be concatenated with SYS t .LINKLIB data sets SMFPRMOO SMFPRMxx SMF parameters The following lists are optionally created during the system generation process if specified by the installation. Standard List Alternate List Contents IEAAPFOO IEAAPFxx Names of authorized libraries IEAAPPOO IEAFIXOI Names of installation-written EXCP appendages IEAFIXxx Names of SYS1.SVCLIB, SYS1.LPALIB, and SYSl.LINKLIB modules that are to be made resident as a nonpageable extension to the link pack area Of the IBM-created lists, multiple lists can be created and maintained for those that have alternate names. These lists may therefore be respecified by the installation to continue the system tailoring begun with system generation processing. The desired list may then be selected from the multiple lists by system parameters specified from the operator's console or in the IEASYSOO or IEASYSxx lists. Authorized Appendage List (IEAAPPOO) This list contains the names of installation-written I/O appendage routines and a description of the function of each routine. Any 110 appendage that is to be used by an unauthorized routine must be named in the IEAAPPOO member of SYS l.P ARMLIB. (Authorized routines are able to use any installation-written I/O appendage.) This system integrity requirement prevents unauthorized routines from incorrectly using I/O appendages. To restrict an I/O appendage to use by authorized programs, omit it from this list. It is not necessary to include the names of control program (e.g., system access methods) appendages in an authorized appendage list. Defining the System 41 If the installation chooses to specify the creation of IEAA1PPOO during system generation, appendage names may be entered, even if the appendages have not yet been written. This capability allows the installation to "pIime" the list with the names of appendages that will be added to the system later. See the chapter "System Integrity" for a description of authorized and unauthorized routines. Authoril:ed Program Library List (IEAAPFxx) This list contains the data set names and corresponding volume indentifications for libraries that may contain programs requiring APF authorization. The ability for installation-specified libraries to contain APF-authorized programs extends the VS2 Release 1 APF support that allowed authorized programs to reside only in SYSl.LINKLIB, SYSl.SVCLIB, and SYSl.LPALIB. The specification of additional libraries for authorized programs in the IEAAPFOO or IEAAPFxx list in MVS eliminates the need to concatenate libraries containing programs requiring APF-authorization. It is not necessary to include SYSl.LINKLIB or SYSl.SVCLIB in an authorized program library list. Dump Default Lists (IEAABDOO and IEADMPOO) The dump default lists, IEAABDOO and IEADMPOO, contain the IBM-supplied defaults for the storage areas to be dumped to the data sets defined by SYSABEND and SYSUDUMP DD statements, respectively. The installation may change the defaults in these lists to satisfy individual requirements. Some of the defaults that may be specified include the system queue area, the local system queue area, the scheduler work area for the task, the task's control blocks, the GTF or supervisor trace table, the PSW at entry to ABEND, general register contents at entry to ABEND, the contents of the task's link pack area, the contents of the task's job pack area. Fixed LPA List (IEAFIXxx) The fixed LP A list indicates those modules from SYS 1.SVCLIB, SYS 1.LP ALIB, and/ or SYS 1.LINKLIB that should be included in the nonpageable extension to the link pack area. The list should contain modules that significantly increase system performance when they are fixed rather than paged. The fixed LP A list is created during system generation processing, if so specified by the installation. Installation Performance Specification (IEAIPSxx) The installation performance specification (IPS) consists of control information used by the workload management routines that are part of the system resources management routines. It includes performance group definitions, performance objectives, and service definition coefficients that are used to establish service rates. For a more detailed description of the contents of the IPS, see the chapter: "Directing the Use of System Resources". 42 OS/VS2 Planning Guide for Release 2 Link Library List (LNKLSTxx) The link library list enables concatenating up to 15 data sets, on multiple volumes, with SYS 1.LINKLIB. It is generated during the system generation process but may be modified via the IEBUPDTE utility program. Each data set that is concatenated with SYSl.LINKLIB may have up to 16 extents. (Note: After additions or changes are made that extend data sets concatenated with SYSl.LINKLIB, the system should be re-IPLed to update the description of SYSl.LINKLIB to the system.) In MVS, multiple link library lists may be maintained in SYS 1.PARMLIB. The list to be used is selected at IPL time by the LNK system parameter. LPA Directory Load List (IEALODOO) The link pack area (LP A) directory load list indicates those modules residing on SYS 1.LPALIB or in the pageable LP A that are to have the contents supervision directory control block information initialized by NIP in nonpageable storage. When this list is used, contents supervision routines do not have to search to find the module in the LPA directory. The number of page faults that may occur is therefore reduced. Candidates for this list are modules that are frequently used. LPA Packing List (IEAPAKOO) The link pack area (LP A) packing list contains the names of modules residing on the SYS 1.LPALIB data set that may be grouped together by packing the LP A. Grouping these modules may sharply reduce page faults as well as eliminate wasted space within the LP A. Modules that frequently pass control to one another should be grouped together to reduce page faults. The total size of all modules in a group should not exceed 4K (page size). The proper loading of modules into the LP A is done at NIP time, based upon the data in IEAPAKOO. Modules that are not listed in IEAPAKOO are then loaded into the LP A in the order that they appear in the SYS 1.LP ALIB directory. There are no alternate lists for IEAPAKOO. If IEAPAKOO is not found as a member of SYS 1.PARMLIB, the modules will be loaded in the order that they appear in the SYS 1.LPALIB directory. Resident BLDL List (IEABLDxx) The resident BLDL list specifies the names of modules from SYS 1.LINKLIB, or any data set concatenated to SYS1.LINKLIB, that are to have directory entries built by NIP. This eliminates the directory search required when a load module is requested from SYS 1.LINKLIB. The directory will reside either in the fixed or pageable area of the system, depending upon whether the BLDLF or BLDL reply is specified by the operator in response to the "SPECIFY SYSTEM PARAMETERS" message. The BLDL and BLDLF parameters are mutually exclusive. The member name for the standard list is IEABLDOO. The load module names must be listed in the same order as they appear in the directory; that is, they must appear in ascending collating sequence. (Note: Directory entries in the resident BLDL table are not updated as a result of updating the load module in Defining the System 43 the library. The old version of the load module is used untH an IPL process updates the resident BLDL table.) The IEABLDOO list indicated by BLDL=OO or BLDLF ==00 is created during system generation processing. The list will be null if no entries are specified for it at SYSGEN. If specified during system generation and not modified through operator communication or a list of system parameters (IEASYSxx) the default IEABLDOO is used. SMF Parameter List (SMFPRMxx) The SMF parameter list contains the parameters that control SMF operations. Multiple SMF parameter lists may be maintained in SYS l.P'ARMLIB. The list to be used is selected at IPL time by the SMF system parameter. The single SYS l.P ARMLIB member (SMFDEFLT) supplied by IBM in MVT and VS2 Release 1 is not valid. In MVS, the IBM-supplied SYSl.PARMLIB member that contains the default SMF parameters is called SMFPRMOO. The format of the SMFPRMOO and SMFPRMxx lists is different from the MVT and VS2 Release 1 SMFDEFLT list. The ALT and PRM keywords are not supported because the SMF data sets, SYSl.MANX and SYSl.MANY must be cataloged in MVS. The MDL keyword has also been deleted. The .SID keyword has been expanded to allow the specification of a System/370 model number. System Parameter List (lEASYSxx) The system parameter list may include all system parameters that are valid input to the "SPECIFY SYSTEM PARAMETERS" message. The standard list, IEASYSOO, is created during Stage II of the system generation process. This param,eter list enables an installation to specify all parameters needed during a particular IPL process without requiring that they be specified from an operator's console. Operator intervention for an individual parameter list may be restricted by use of the 'OPI= YES/NO' parameter. 44 OS/VS2 Planning Guide for Release 2 InstaUation-Created Lists The installation is also able to create certain lists, that are not supplied by IBM, to further tailor the system to installation needs. Most of these may also be maintained as mUltiple lists. The following lists may be created by the installation in MVS: Default List Alternate List Contents COMMNDOO COMMNDxx Commands to be issued by the system control program at the completion of system initialization processing. IEALPAOO IEALPAxx Names of SYS1.SVCLIB, SYS1.LPALIB, and SYS1.LINKLIB modules that are to be made resident as a page able extension to the link pack area. IEAOPTOO IEAOPTxx System resource management tuning parameters. IKJPRMOO TIOC buffer parameters IRBMFlOO MF /1 parameters PARMTZ Local time constant VATLSTOO VATLSTxx Volume attributes Command List (COMMNDOO) This list contains commands (e.g., START) that are to be issued by the system control program at the completion of system initialization processing. It may also be used to specify whether or not the time-of-day (TOD) clock initialization messages and associated responses should be issued. The COMMNDxx list is specified by the CMD system parameter. Local Time Constant (PARMTZ) This SYS l.P ARMLIB member may be used to contain the value (in hours, minutes, and seconds) that the local time differs from Greenwich Mean Time (GMT), and the direction (east or west) from the GMT. This constant is used to set the time-of -day clock and overrides the local time value that is specified at system generation for inclusion in the CVT. Modified Link Pack Area List (IEALPAxx) This list may be used to indicate reenterable routines from the SYS l.LINKLIB, SYS1.LPALIB, and SYS1.SVCLIB data sets that are to be contained in the pageable extension to the link pack area. The routines in this area (called the modified link pack area) are resident only for the current IPL. If a routine is searched for, the modified link pack area is searched first, followed by the pageable link pack area. If a system restart is initiated and a previously created link pack area is used, the routines in the modified link pack area must be respecified if desired. The IEALP Axx list is specified by the MLP A system parameter. Defining the System 45 System Resources Management Tuning Parameter List (IEAOPTxx) This parameter list contains the values that are used by the system resources management algorithms to influence decisions about swapping jobs in or out of real storage. The values are described in the chapter: "Directing the Use of System Resources". MF/l Parameter List (IRBMFlxx) This list contains the parameters that control MF /1 operation. The options that may be specified are described in the topic: "MP/1 Operation" in the chapter: "Direeting the Use of System Resources". Volume Attribute List (VATLSTxx) The volume attribute list may be used to define the mount characteristics (permanently resident, reserved) and allocation characteristics (storage, public, private) for any, or all, direct access volumes used by an installation. This list was a single SYS 1.PARMLIB member called PRESRES in MVT and VS2 Release 1. Multiple members of the volume attribute list may be maintained in MVS. The V ATLSTxx list is specified by the VAL system parameter. 46 OS/VS2 Planning Guide for Relea~le 2 System Libraries and Data Sets Some system data sets are basic to MVS and are required for every operating system. Other system data sets are optional and are required only when selected options are included in the system. Required Libraries and Data Sets The following libraries and data sets are required for every MVS operating system and must be on the system residence volume. • SYS I.LOG REC (error recording. data· set) -- The machine error recording data set is used by recovery management routines to record statistical data about machine errors and software errors. • SYS1.NUCLEUS (nucleus library) -- The nucleus library contains the resident portion (nucleus) of the control program and may contain modules selected and link -edited during system generation. • SYS1.SVCLIB (SVC library) -- The SVC library contains some OLTEP modules and some appendage modules. The following libraries and data sets are required for every MVS operating system and must be on direct access volumes. It is not necessary that they be on the system residence volume. • Job Entry Subsystem Data Set(s) -- These data sets contain spooled and checkpointed record used by either JES2 or JES3. Either may be a multivolume data set. • Master Catalog -- The master catalog is a key-sequenced VSAM data set that contains pointers to cataloged data sets and user catalogs. The name can be specified at system generation time on the PDS = parameter of the Stage I DATASET macro instruction. This catalog replaces the MVT and VS2 Release 1 system catalog (SYSCTLG). • Page Data Sets -- The page data sets are known collectively as paging space. Paging space contains the "pagedout" portions of address spaces and the common system area, as well as the data written to virtual 110 data sets. • SYS I.DSSVM -- This data set contains the command language modules used by the dynamic support system (DSS). • SYS1.LINKLIB (link library) -- The link library contains programs and routines that are not in the link pack area and are referred to by ATTACH, LINK, LOAD, or XCTL macro instructions. • SYS1.LPALIB (link pack area library) -- The link pack area library contains all modules that are loaded into the pageable link pack area. In MVS, the pageable link pack area is a read-only area. Therefore, modules contained in it must be refreshable. Modules that are modified during execution (non-refreshable) should be included in the modified link pack area. Because all of the modules in this library are brought into the LP A at IPL time, this library may be placed on a demountable volume that may be removed after IPL. • SYS1.MACLIB (macro library) -- This library contains macro definitions for the system macro instructions used by the assembler-language processor. Defining the System 47 • SYS I.PARMLIB (parameter library) -- The parameter library contains defau.lt parameters for the initialization of system functions. • SYS I.PROCLIB (procedure library) -- The procedure library contains cataloged procedures used for job definition. • SYS I.SAMPLIB -- This library contains independent utility programs, the IPL text, and the SMF exit routines. • SYS I.STGINDEX -- This VSAM data set is used as an internal catalog for the pageable link pack area and virtual 110 data sets that must be saved across job steps and IPLs. • SYSl.BRODCAST (broadcast data set) -- the broadcast data set contains two types of time-sharing messages: notices and mail. This data set must be cataloged but requires space allocation only if time sharing messages are to be written. • SYSl.CMDLIB (command library) -- This library contains time-sharing command processor programs, service routines, and utility programs. This library must be cataloged. It requires space allocation unless the TSO stage 1 system generation macro instruction is coded such that the time sharing command system is not included. • SYSl.MANX/SYSl.MANY (SMF data sets) -- The system management facility (SMF) data sets contain the information collected by the SMF routines. They may also contain measurement information collected by the syst.em activity measurement facility (MF 11). The SMF data sets must be cataloged but require space allocation only if SMF or MF I 1 recording is planned. • SYSl.UADS (user attribute data set) -- This data set contains a list of terminal users who are authorized to use time-sharing and information about the users. This data set must be cataloged but requires space allocation only if the installation plans to initiate terminal sessions. Optional Libraries and Data Sets The following data set is optional and, if selected, must be on the system residence volume. • PASSWORD -- This data set contains records that associate the names of protected non-VSAM data sets with the passwords assigned to the data sets. The following libraries and data sets are optional and, if selected, must be on direct access volumes unless otherwise stated. • SYS I.DCMLIB -- This library contains program function key (PFK) definitions for display consoles and is required if a display console with the optional PFK feature is included in the system. • SYSl.DUMP (dump data set) -- This data set contains core image dumps recorded in the event of abnormal termination of a critical task. The dump data set may reside on tape. • SYS I.HELP -- This data set is required if the time-sharing HELP command is used. Each member of this data set contains information regarding the syntax, subcommands, and functions for each time-sharing command. 48 OS/VS2 Planning Guide for Release 2 • SYS1.1MAGELIB -- This data set contains the 1403 and 3211 universal character set (UCS) and forms control buffers (FCB) image modules. It is required if a 1403 printer with the UCS feature, 3211 printer, 3890 document processor device, or 3886 optical character recognition device is included in the system. • SYS 1.P AGEDUMP -- This data set contains the stand-alone dump program, if this program is on a direct access device. If the stand-alone dump program is on tape, this data set will not exist. • SYS 1. TELCMLIB (telecommunications library) -- This library contains the / telecommunications subroutines in load module format and is required if any of the TP access methods other than the virtual telecommunications access method (VT AM) are selected. • SYS 1. VT AMLST -- This data set contains lists of internal parameters in card image format used during VTAM initialization. • SYS 1. VT AMLIB -- This library contains the VT AM executable code in load module form. • SYSl.VTAMOBJ -- This data set contains the 3705 network control program (NCP) that is loaded by VT AM during VT AM initialization. Changes to VS2 Release 1 System Libraries and Data Sets System data sets and libraries in MVS are the same as their counterparts is VS2 Release 1 with the following exceptions: • The page data set, SYS 1.P AGE, has been replaced by two or more uniquely named VSAM data sets. This paging space must be defined before the first IPL. (See the topic "Adding Paging Space to the System.") • The system catalog, SYSCTLG, has been replaced by a key-sequenced VSAM data set. Conversion aids are discussed in the topic "Catalog Conversion" in the chapter "Conversion Considerations". • The data sets that support job entry subsystems are new. • The format of the SYS 1.BRODCAST data set has been modified to include a change level indicator in the header record and a new field for each user identification in the mail directory records. This field contains the address of the last message in the chain of messages assocated with each user identification. An MVS format for the broadcast data set is obtained by using a new subcommand (SYNC) of the time-sharing ACCOUNT command. (See the topic "Formatting the Broadcast Data Set" in the "Conversion Considerations" chapter.) • The load module size of modules in the SYS 1.DSSVM data set has been increased from 1K bytes to 2K bytes. • The SYS1.SYSJOBQE data set does not exist. It has been replaced by a scheduler work area (SWA) in virtual storage for initiator records and by a job entry subsystem data set for input, queued jobs, and output. • The content of the SYSl.LOGREC data set has been expanded to include records for software error recording, missing interrupts, and dynamic device reconfiguration(DDR) routines. Defining the System 49 • The SMF data sets, SYS I.MANX and SYS I.MANY, may now be used for system activity measurement facility (MF /1) recording. They must be cataloged in MVS. • The log data sets are no longer named SYSl.SYSVLOGX and SYS 1.SYSVLOGY by IBM. They are dynamically allocated by the job entry subsystem. • The SYS I.P ARMLIB library contains new members to control the initialization of the system. See the topic "SYSl.PARMLIB" in the section "Defining the System" for descriptions of these members. • The SYSl.STGINDEX data set is new. • The format of the SYSl.UADS data set has been modified to contain information for new keywords on ACCOUNT subcommands. An existing user attributes data set (UADS) is converted to the new MVS format by using an Account Facility utility program (UADSREFM). (See the topic "Reformatting the UADS" in the "Conversion Considerations" chapter.) • The SYS 1. VTAMLIB, SYS 1. VT AMLST and SYS 1. VT AMOBJ data sets are new. Changes to MVT System Libr.aries and Data Sets System data sets and libraries in MVS are also the same as their counterparts in MVT, with the exceptions previously noted and the following: • The SYS I.ACCT data set no longer exists. The SMF data sets SYS I.MANX and SYSI.MANY should be used in its place. • The SYSl.DSSVM data set is new. • The SYSl.LPALIB data set is new. It contains all of the SVCs and most of the modules that, in MVT were contained in the SYS I.LINKLIB and SYS1.SVCLIB libraries. All of the modules contained in this new data set appear in the fixed or pageable link pack area in VS2. • The page data sets are new. • The SMF data sets, SYSl.MANX and SYSl.MANY, may now be used for VSAM recording. Adding Paging Space to the System Before the first system IPL, the installation must define and preformat paging space. In MVS, paging space contains system pages, the page able link pack area, and virtual I/O data sets. Paging space consists of two or moOre VSAM data sets. The usc of the define statement with this type of VSAM data set causes a paging data set to be formatted with track overflow for maximum performance of paging I/O. At least two such data sets are required for paging space because two copies (primary and secondary) of certain pages are maintained on external storage. These data sets must contain enough space for a minimum number of secondary pages, corresponding primary pages, and additional pages that have no associated secondary copies. The auxiliary storage management initialization routines enforce these minimum values by internally separating paging data sets for exclusive storage of 50 OS/VS2 Planning Guide for Release 2 secondary page copies, from paging data sets used exclusively for primary copies. These routines attempt to select for secondary page copies, enough (entire) data sets on the slowest paging device type to satisfy the minimum number of pages. They classify paging devices by average access time. For best separation, the installation should direct candidate page data sets for secondary page storage onto the slowest device available for paging; preferably directing the other paging data sets to separate devices. The installation can include more paging space than the minimum by specifying additional page data sets during system generation. If it is necessary to add paging space to the syst6"m after system generation, the installation must first use access method services to define and format the data sets, then cause their names to be entered in a PARMLIB member PAGE list, and finally re-IPL the system. The operator can also specify additional names via the PAGE system parameter. No more than 25 page data sets can be specified during system generation. The number of page data sets must never exceed 63. Page data sets can be added to the system in even or odd numbers; there is no need to pair them. Creating Temporary Data Sets in Extemal Page Storage Using Virtual I/O Conventional access methods (BSAM, QSAM, BDAM, and BPAM) and XDAP and EXCP macro instructions, can create and access temporary data sets in external page storage or system paging space by utilizing the virtual input/output (VIa) facility. Specific unitnames can be defined during system generation (UNITNAME macro) as eligible for VIa processing. The VIa unitname defined is coded in a problem program JCL statement which defines a system-named temporary data set without specifying the volume serial number. VIa then interprets and simulates the execution of the channel programs associated with an EXCP issued by an access method or a problem program. The channel programs are not scheduled and translated in the conventional manner by the I/O supervisor. VIa simulates I/O activity using a buffer that is the size of a track for the first device in the list defined by the UNITNAME macro instruction. To the access method or problem program, a VIa device is a real direct access storage device; actually the virtual device is simulated on the volumes assigned to external page storage. VIa processing relies heavily on the services of the paging supervisor. The paging supervisor causes the real I/O that is requested by VIa, and maintains for virtual, auxiliary, and real storage, the page tables that make data accessible in a virtual storage environment. The paging supervisor maintains the contents of the VIa buffer based on requests issued by VIa. VIa offers the following enhancements to conventional processing of temporary data sets: • Channel program translation and page fixing by the I/O supervisor for each EXCP request, are eliminated. • The paging supervisor dynamically allocates page-size physical blocks (auxiliary storage page slots) to a VIa data set as it is being created. DADSM overhead for the allocation of space on a real device is thereby eliminated. Defining the System 51 • Many I/O operations may be satisfied by moving data between a VIO buffer and an access method or user buffer, thus eliminating the channel and device activity associated with normal channel program request processing. The data set is automatically blocked in page size (4K) records, irre:spective of access method or user-buffer size. • There is generally more efficient utilization and easier management of direct access space, since temporary data sets use auxiliary space within the paging data sets. • When a data set page is to be read, the paging supervisor determines if it is still in real storage and available for reclamation (the reclaim function). If it can be reclaimed, the read is accomplished without physical I/O. 52 OS/VS2 Planning Guide for Release 2 Directing the Use of System Resources This chapter describes some of the MVS facilities that allow installations to direct the use of system resources. System resources management routines allow an installation to specify, in measure able terms, the relative rate at which system resources should be provided to various categories of background and terminal jobs. The system activity meas\i:;~ment facility (MF /1) provides reports on selected areas of system activity. The installation uses the information in these reports to determine how to specify performance objectives for the system resources management routines. Directing The Use of System Resources 53 System Resources Management System resources management routines are new in MVS. They provide: • The capability for an installation to specify, in measure able terms, the rate at which system resources should be provided to various use:rs of the system. • A means by which the system control program is able to monitor the use of system resources and attempt to maintain their most efficient use at all system workload levels. These routines monitor the use of system resources to determine when background and time sharing jobs should be swapped in or out to meet the installation-specified objectives. The resources management routines are divided into three primary categories: • Workload management routines. • Resource-use routines. • A control function. Workload Management Routines The function of the workload management routines is to cOllltrol the relative rate at whiGh processing resources are provided to individual jobs or time sharing users so as to conform to installation performance specifications. Their design enables installations to specify different processing rates for a job or time sharing command depending on the system workload level and on th.e age of, or amount of processing done by, a job or command. Workload management routines accomplish this processing rate control by monitoring the processing resources used by each address space and scheduling swaps (to or from main storage) to achieve an intended processing rate. A more detailed discussion of the installation interface with the workload management routines is found under "Workload Management Concept." Resource-Use Routines The resource-use routines consist of a set of algorithms that make scheduling recommendations intended to optimize the utilization of the system CPU, I/O, main storage, and paging resources. These algorithms include: I/O Load Balancing This algorithm attempts to maintain a dispatchable job mix that has a balanced use of logical channels. The algorithm monitors the fraction of I/O requests delayed on each logical channel and ascertains whether this measure of load falls within an acceptable range. When an imbalance -- either an overload or underload situation -- occurs the algorithm seeks out an address space known to be a heavy user of the imbalanced resource, and may then recommend a swap. To reduce the overhead of monitoring the I/O profile of each individual address space, the I/O request counts gathered already on behalf of SMF are analyzed for this purpose. CPU Load Adjustment This algorithm continuously monitors CPU utilization and when this value either goes above or below an acceptable target range, the CPU Load Ad justment algorithm will attempt to remedy the situation by recommelllding the swapping of one or more heavy CPU users into main storage in the case 54 OS/VS2 Planning Guide for Releas(~ 2 of CPU underload, or out of main storage in the case of overload. Because MVS installations are advised to initiate more jobs than will normally fit in main storage, ingredients for a well balanced job mix normally exist on the queue of swapped-out but ready to execute address spaces. A special case occurs when the CPU is 100% utilized. When this occurs the algorithm looks for address spaces near the bottom of the dispatching queue that are not being regularly dispatched. This situation is considered a serious CPU overload because it means that address spaces are tying up real storage, but are either not executing at all or are executing too slowly to warrant the expenditure of rear storage. The solution is to recommend the replacement of heavy CPU users in main storage with 110 users who make better use of the system resources. Main Storage Occupancy This algorithm attempts to ensure that the sum of the working sets of all swapped-in address spaces (those address space pages that are regularly referenced) nearly fills, but does not exceed, the capacity of main storage. The sum is· judged too large if the size of the available frame count drops below a low threshold. When this occurs, an address space is promptly swapped-out to free frames and thereby increase the available frame count. To prevent this corrective action from happening too often, the main storage occupancy algorithm evaluates the consequences of all anticipated swaps in an attempt to keep the available frame count near a target which is above the low threshold. If the low threshold is crossed too frequently, this target is dynamically adjusted away from the low threshold. If it is not crossed at all, the algorithm is too conservative and the target is dynamically adjusted closer to the low threshold. Page Replacement Page stealing makes frames occupied by unreferenced pages, available for reuse by active pages. The page replacement algorithm takes advantage of data maintained on an address space basis, to determine candidates for page stealing. This represents a marked departure from Release 1 which employs a system-wide page stealing algorithm. Perhaps more important however, is the fact that in Release 2 page stealing is not used to p~rform the system-wide control function of ensuring the availability of sufficient deallocated real pages. The problem of thrashing -- excessive demand paging -- is largely eliminated in Release 2 because page stealing is not used to remedy real frame shortages. Instead, when such a shortage occurs, the SRM main storage occupancy algorithm uses swapping to reduce the number of address spaces executing in main storage. To ensure an acceptable percentage of "good steals," the page replacement algorithm monitors the page referencing activity of individual address spaces, and selects a page to be stolen if and only if it has gone un referenced for a predefined amount of that address space's CPU execution time. This criterion prevents the stealing of active working set pages that go unreferenced for periods of elapsed time simply because that address space temporarily did not get dispatched. For pages in the Common Area this approach is not used, since these pages are not associated with any particular address space. Unreferenced elapsed time is used as the prime criterion for stealing these pages. Directing The Use of System Resources 55 Device Allocation This algorithm determines the unit to be used for data sets for which an allocation choice must be made from a list of more than one device candidate. The goal is to allocate from lightly-utilized channels, and to collect allocations for the same address space on the same logical channel without collecting them on the same device, so that swapping that address space will have a predictable effect on the system I/O load. Automatic Priority Group (APG) This algorithm reorders dispatchable address spaces within the automatic priority group based on the degree to which they are I/O bound; within the APG, I/O-bound jobs are given higher dispatching priorities than CPU-bound jobs. To fully realize the benefits of this algorithm, the installation should place all possible users within the APG. In MVS, this is facilitated because the APG priority is used as the default step displl.tching priority. (See the topic "Changes to VS2 Release 1 JCL" in the chapter "Conversion Considerations. ") ENQ/DEQ Algorithm This algorithm determines when a batch or terminal job is in control of a system resource and is delaying other jobs that are enqueued on the same resource. When such a determination is made, the job that is enqueued upon the system resource is made non-swapp able for an installation-specified time interval. The interval is specified by the ERV parameter in the IEAOPTxx member of SYS 1.P ARMLIB. Control Function The control function gathers the results (recommendation values for potential swap decisions) produced by the workload management routines and the resource-use algorithms. It uses a control algorithm to combine and evaluate the individual recommendation values to produce a composite recommendation value for each potential swap decision. The control function then uses these values to determine the amount of swapping that should be performed in an attempt to satisfy the installation-specified performance objectives andl efficiently use system resources. The control algorithm uses optional installation-specified parameters called resource factor coefficients to weight the recommendation values produced by the I/O and CPU load balancing algorithms. The use of these coefficients (which are contained in the IEAOPTxx member of SYS 1.P ARMLIB) provides one method of system tuning. For example, if the installation increases the CPU coefficient, the recommendation value produced by the CPU load balancing algorithm will be more heavily weighted when it is evaluated by the control algorithm in conjunction with other recommendation values. Workload Management Concepts One of the SRM design objectives is to assure maximum control over user performance by an installation, without requiring the installation to supply its own scheduling algorithm. The solution adopted is to permit an installation to specify in measure able terms, the response and turnaround objectives wanted, and allow the workload management routines to determine how to achieve these objectives. The following concepts support such an approach by providing 56 OS/VS2 Planning Guide for Release 2 measure able terms for the response objectives of individual batch and time sharing users. Transaction The basic entity to which reponse and turnaround objectives are associated is called a transaction. Since both batch and time sharing are supported by the workload management routines, a concept of transaction is needed which supports both environments. A job is the batch entity about which turnaround is usually measured. It is therefore selected as the definition of a batch transaction. Provision also exists to treat individual job-steps as separate transactions. The time sharing entity about which response is measured, normally begins when the user strikes end-of -block and ends when he receives his terminal output. This sequence may be encompassed in a time sharing command or subcommand -- in either case it qualifies as a separate transaction. Performance Variable Having defined what a transaction is, the need now exists for a variable capable of expressing response and turnaround objectives for transactions. Response time itself is not satisfactory because it is too dependent on the individual processing requirements of each transaction. What is needed is a variable that measures the rate at which processing takes place, rather than the amount of processing required. An objective expressed in terms of processing rate can reasonably be used to assure a time sharing user a consistent quality of responsiveness, regardless of the variety of commands he chooses to enter. Service The measure of processing chosen is a composite variable called service -- a combination of an individual address space's CPU, I/O, and main storage usage. The formula for service also includes three installation-supplied coefficients. SERVICE = A* ( SMF Task CPU Time * K) + B* ( SMF I/O Request Count) + C* (Working Set Size * SMF Task CPU Time * K) where: A, B, and C are installation-supplied constants, and K is a model-dependent MIPS adjusting factor. Service Rate Service rate, defined as the service accumulated by a transaction divided by the elapsed time, is the variable used by the installation to express response and turnaround objectives. It is used by the workload management routines to monitor the progress of each executing transaction. The SMF job termination record, the time sharing TIME command, and the MF /1 workload activity report, return to the installation the data necessary to correlate service rate with response and turnaround times. Thus, for example, a user regularly submitting the same batch job should know that if it requires approximately 72K service units, and through arrangement with the installation manager the job is assured 25 sui sec, then it will complete in 4R minutes, while if the arrangement is for 50 su/sec it will complete in 24 minutes. Directing The Use of System Resources 57 There is a practical limit to the service rate a job can absorb. The workload management routines can do no better than provide the job with the highest service rate it can absorb, even if a higher rate is requested. System Workload One important new facility in MVS, is the ability to temporarily swap most batch and time sharing address spaces out of main storage. This facility makes it feasible -- and usually desireable -to start more batch initiators than the real storage capacity can efficiently support. The SRM keeps track of all active address spaces, both batch and time sharing, and schedules swaps in and out so that a number of conditions are satisfied. For the purposes of this discussion, it is sufficient to mention two of these conditions: • The demand for real storage should not cause excessive paging . • The service rate afforded each address space should correspond closely to the installation performance objectives. The combined demand of all address spaces for system resources constitutes the system workload. The more active address spaces that exist -- whether batch or time sharing -- the heavier will be the system workload. MVS is designed to efficiently handle a very wide range of system workloads. An installation should anticipate, however, that under heavy system workload coltlditions, some address spaces must be temporarily swapped out of main storage. This results in lower serviee rates for the delayed address space. Later it will bt~ shown that an installation can specify the service rate reductions that individual batch and time sharing address spaces should experience under different system workload levels. Installation Performance Specific;lltion It is the job of the installation to translate the various performance requirements of its users into one or more installation performance specifications. This control data, which is included in the SYS l.P ARMLIB data set, permits the workload manager to schedule the availability of processing resources among individual users in accordance with the dictates of the installation. An installation performance specification (IPS) is a mapping that associates a particular transaction with an intended service rate. It is important to note that the IPS mappings take into account other considerations, specifically the systt~m workload, the amount of service already accumulated by a transaction, and the age of the transaction. Performance Groups The purpose of performance groups is to group together user transactions that the installation considers to have similar performance requirements. A teleprocessing applications program that has multiple users and executes in a single virtual address space is considered to be a single transaction, and its terminal users are not distinguished from each other by the system resources management routines. The installation can define as many as 255 performance groups, each identified by a distinct performance group number. Each performance group definition includes as many as eight performance group ~riods. By dividing a performance group into periods, an installation can associate different performance objectives with different perioJs in the life of a transaction. The duration of a period can be specified either as a number of real-time seconds or 58 OS/VS2 Planning Guide for Release 2 as a number of accumulated service units. For example, one performance group might be defined for short compile-load-go jobs that should have a very rapid turnaround time. The installation divides the performance group into two periods. The first period is associated with a high service rate and lasts until N number of service units are accumulated, where N has previously been determined to be a sufficient number of service units to complete a short compile-load-go job. If a job in the group is not complete after N service units are accumulated, it enters the second period, for which a lower service rate has been specified, and remains in this period for the duration of the job. Performance Objectives A performance objective specifies service rates; how many service units per second an associated transaction should receive under different system workload conditions. The motive for specifying different performance objectives is to give certain users better service at the expense of other users. However, the relative importance of this motive is dependent on the size of the system workload (i.e., the degree to which the demand for system resources exceeds the supply). The installation specifies the treatment of different subsets of jobs by fixing arbitrary points, called workload levels, at which it defines target service rates for each subset of jobs. Conceptually, increasing workload levels can be thought of as increasing competition or demand for system resources. Thus, in order to preserve a high service rate for high priority jobs as demand (workload level) increases, an installation will reduce the target service rate for lower priority jobs. Typically, an installation supplies an IPS such that if the system workload is moderate, all jobs should receive satisfactory service rates. However, if the system workload is heavy, the installation will want to provide an acceptable service rate to high priority jobs, and lower the service rate for lower priority jobs. Under very heavy workload conditions the installation might completely sacrifice the service rate of lower priority jobs (i.e., assign a service rate of 0) in order for the system to continue to provide acceptable service to the higher priority jobs. By defining workload levels, the installation can express the varying service relationships between groups of users at different system workload levels. For example, an installation defines three performance objectives, numbered 2, 5, and 8. When the system workload is light, each performance objective should specify a better than satisfactory service rate for the transactions associated with it; the installation assigns the number 2 to this workload level and assigns corresponding service rates to each performance objective, as illustrated in the chart at the top of Figure 9. Under very heavy workload conditions, the installation wants to completely sacrifice the service rates for objective 5 and to lower the service rate for objective 8 but still provide acceptable service to jobs associated with objective 2. This workload level is assigned the number 8; the corresponding service rates associated with each objective are illustrated in Figure 9. The two intermediate workload levels shown in Figure 9 are established to reflect the changing relationships between the performance objectives as the workload level increases from 2 to 8. Directing The Use of System Resources S9 Performance Objectives Performance Service Rate Service Rate Service Rate Service Rate Objective for Workload for Workload for Workload for Workload Number Level 2 Level 4 Level 6 Level 8 2 44 40 36 32 5 60 10 0 0 100 20 s 100 Performance Objective 8 80 60 ... CI) 10 a: CI) .;CJ Iii C/) When the system workload is at level 4, transactions associated with performance objective 8 receive service at a rate of 20 40 20 - - - - - . - - - - - - - . _...............__ _ o~-------_~I--------~----------I~--------~I. 2 4 6 8 -w Workload Level Figure 9. Performance Objective Definition 60 OS/VS2 Planning Guide for Release 2 The system resources manager tracks the service rates provided to all users. As the system workload level increases or decreases, it adjusts the service rates in an attempt to maintain the relationships between the performance objectives (and therefore between the transactions associated with the objectives) that the installation has defined. Figure 9 also contains a graphic representation of one of the performance objectives (performance objective 8). This graph illustrates the service rates that the installation has defined for the system workload levels of 2, 4, 6, and 8. The workload management routines attempt· to provide the service rate that applies to the performance objective for the associated job at the current system workload level. In Figure 9, when the current system workload is 4, the workload management routines recominend swapping in order to maintain a service rate of 20 for the jobs that are associated with performance objective 8. Jobs associated with other performance objectives will be targeted at service rates indicated in their performance objectives for a workload level of 4. Associating a Transaction with a Performance Objective Every transaction that enters the system has a performance group number. In Figure 10, transaction A is associated with performance group 9, transaction B is associated with performance group 8. For its first performance group period, transaction B will be associated with performance objective 5. This means that, for this period, the workload management routines will attempt to maintain a service rate for transaction B that equals the service rate specified in performance objective 5 for the workload level that equals the actual workload level of the system at a given time. I nstallation Performance Specification (I PS) ~~~~~--~----~----~------~--Transaction A Performance Group Number = 9 Transaction B Performance Group Number = 8 Workload Level 2 100 4 6 20 18 8 Figure 10. Associating a Transaction With a Perfonnance Objective The Relationship Between Workload Levels and Service Rates To summarize the relationship between workload levels and service rates, consider a hypothetical installation whose users fall into one of three distinct subsets. Each subset can be considered to be a performance group. They are: Directing The Use of System Resources 61 • Performance Group A -- A long running batch job that is run only once a day is the: only job in this class and is extremely important. • Performance Group B' -- Moderate length batch jobs, of intermediate importance make up this class. • Performance Group C -- Relatively short batch jobs, least in importance, make: up this class. Figure 11 illustrates the performance objectives that have been defined for each of the three performance groups. (Three distinct performance groups, each with only one defined period and one associated performanc,~ objective, are assumed for simplicity of illustration.) S 140 J!l III ex: 120 0- 100 0 ~ • 80 CI) u .~ CI) en 0 60 40 20 ~------~--------~------~~----------W 2 4 6 8 o Workload Level Figure 11. Performance Objectives Group A jobs are associated with performance objective 3, group B jobs with performance objective 8, and group C jobs with performance objective 5. Jobs associated with performance objective 3 demand a service rate of approximately 120 at workload levels 2 and 4. At workload level 6, performance objective 3 jobs d~:!mand a service rate of 100 "'hile performance objective 5 jobs have stopped receiving service, and performance objective 8 jobs have reduced their demands to 18 service units per unit time. At workload levt!l 8, even performance objective 8 jobs nearly stop receiving service in deference to A jobs. As long as the A job receives a service rate of approximately 100, the installation knows that the job will complete processing in a satisfactory amount of time. 62 OS/VS2 Planning Guide for Release 2 The workload level at which the system operates is a function of the number of jobs in each performance objective actually initiated at any time. Figure 12 illustrates system workload levels that result from three different possiblilties: • One job from each performance objective. • One job from performance objective 3, four from performance objective 8, and two from performance objective 5. • One job from performance objective 3, seven from performance objective 8, and one from performance objective 5. 1-1-1 S CI.I ....co 1-4-2 S 280 280 280 240 240 240 200 ....CI.Ico a: 160 CI.I .;u i 120 CI) 200 ....IIIco a: 160 a: 160 CI.I 'E Jl 'ECI.I 5 CI) 80 Composite Demanp 40 0 2 4 6 / Performance Objective 3 - 1 job Performance Objective 8 - 1 job Performance Objective 5 - 1 job 120 C . , omposlte Demand 80 8 0 2 4 6 Workload Level System Workload Level Between 2 and 4 Supply i/ Composite ..... ~ 80 Demand ; 40 40 Workload Level i 200 u 120 1-7-1 S Performance Objective 3 - 1 job Performance Objective 8 - 4 jobs Performance Objective 5 - 2 jobs System Workload Level of 4 8 o 2 4 6 8 Workload Level Performance Objective 3 - 1 job Performance Objective 8 - 7 jobs Performance Objective 5 - 1 job System Workload Level Between 6and8 Figure 12. System Workload Levels In these illustrations a simplified assumption is made that the system provides a composite service rate of 220 service units per unit of time to all active jobs; this is the "supply" line. The "demand" line represents the composite service demands from all jobs at each workload level. It is derived by multiplying the number of jobs in each performance group by the service rate specified for that performance objective at the particular workload level. In other words, with a job mix of 1-4-2, (the center graph on Figure 12), the composite service rate demanded by all jobs at workload level 4 is 1x120+4x20+2x10=220. (At workload level 4, a job associated with performance objective 3 demands a service rate of 120, a performance objective 8 job demands 20 and a performance objective 5 job demands 10.) The intersection of the supply and demand lines produces the resulting system workload - that point of equilibrium at which the system can meet the composite demands of all active users. In Figure 13, the different system workload levels that result from the three different job mixes are superimposed over the three performance objectives to illustrate the service rates that apply as the system workload level increases. Directing The Use of System Resources 63 1·1·1 S 1-4·2 ,·7·1 140 ... CD III a: 120 0- 100 ® = = Ii • 80 Q) u .~ ~ ® 60 40 20 ..~------------~W 8 ~------------~----~----~=-------------~ o 2 4 6 Workload Level Figure 13. Service Rates for Different Perfonnance Objectives at Various System Workload Levels As shown in Figure 13, the following conclusions about the installation's performance objective specifications may be drawn from this hypothetical analysis: With 1 job initiated in each performance objective: • The system will be operating at a workload level between 2 and 4. • The performance objective 3 job will receive approximately 121 service units per unit time. • The performance objective 8 job will receive approximately 64 service units per unit time. • The performance objective 5 job will receive approximately 35 service units per unit time. With 1 performance objective 3 job, 4 performance objective 8 jobs, 2 performance objective 5 jobs: • The system will be operating at a workload level of 4. • The performance objective 3 job will receive approximately 120 service units per unit time. • The performance objective 8 jobs will each receive approximately 20 service units per unit time. • The performance objective 5 jobs will each receive approximately 10 service units per unit time. 64 OS/VS2 Planning Guide for Release 2 With 1 performance objective 3 job, 7 performance objective 8 jobs, 1 performance objective 5 job: • The system will be operating at a workload level between 6 and 8. • The performance objective 3 job will receive approximately 100 service units per unit time. • The performance objective 17 jobs will each receive approximately 15 service units per unit time. • The performance objective 5 job will not be receiving service. In each of the preceding examples, the sum of the service rates provided to all users will equal approximately 220 service units per unit time (the composite system "supply"). System Resources Manager Parameters The system resources manager parameters fall into two classes. Those that affect the recommendation values produced by workload management routines are contained in the SYS 1.PARMLIB member IEAIPSOO (or IEAIPSxx); those that otherwise affect swapping decisions are contained in the SYS 1.PARM LIB member IEAOPTOO (or IEAOPTxx). These parameters provide the installation with a great degree of flexibility in stating performance requirements. IPS Parameters (lEAIPSxx) An installation performance specification contains four categories of information: • Service definition coefficients -- The workload management routines use these values as multipliers, in calculating the service that a transaction has received. (See the topic "Service" earlier in this chapter.) • Up to 32 workload level numbers -- Workload level numbers are used to specify the relative service rates that jobs of different priorities are to receive at varying levels of system activity. Each number is related to a specific service rate in each performance objective, and the numbers increase in magnitude as contention for system resources increases. • Up to 64 performance objectives -- Each performance objective is specified in terms of the service rates to be provided at different workload levels. The performance objective number that is specified for each performance group period, identifies the performance objective to be used for that period. • Up to 255 performance group definitions -- A performance group definition is identified by a performance group number. It may be subdivided into as many as eight different periods. Each period has the following parameters associated with it: - period duration - units code - performance objective number - response/throughput bias - interval service value Directing The Use of System Resources 65 These categories of information are specified by the following parameters: Performance Group Number The performance group number is an integer between 1 and 255 that associates a transaction with a performance group definition. All transactions have performance group numbers. 1 is the default for batch jobs and 2 is the default for terminal jobs. Period Duration The period duration is a positive integer that indicates the length of the interval during which a specified performance objective is to be associated with a transaction. Units Code The units code is specified by the letter "s" or the letter "R". "s" indicates that the indicated period duration is to be measured in terms of accumulated service; "R" indicates that it is to be measured in terms of real-time seconds. Performance Objective Number The performance objective number is an integer between 1 and 64, that indicates the performance objective with which the transa(~tion is to be associated for the duration of the indicated period. Response/Throughput Bias The response/throughput bias (RTB) is an integer that indicates, for each performance group period, that deviations from the service. rate specified in the performance objective can be tolerated in favor of higher system throughput. "1" indicates that for jobs in this performance group period, the resource-use algorithms will produce evaluations that will influence swap decisions; "0" indicates that they will not. Interval Service Value The interval service value (lSV) specifies a minimum number of service units that a transaction must receive during each interval of real storage occupancy, before a swap evaluation is made by the workload management routines. Workload Level The workload level is an integer between 1 and 128. Up to 32 workload level numbers can be specified (in increasing magnitude representing heavier workload levels of the system). Workload levels are used to differentiate between light and heavy demands for the use of system resources. The same workload level numbers must be specified for each performance objective in an IPS. Performance Objective Each performance objective is identified by an integer from 1 to 64 followed by a definition of the service rates that are associated with the workload level numbers specified by the workload level parameter. A service rate must be specified for each workload level number defined in the IPS. Service Definition Coefficients The service definition coefficients (CPU, IOC, MSO) are used by the system resources management routines to calculate the service units that are provided to a particular job. Each coefficient is a multiplier on one of the three basic processing units in these calculations. For example, if an installation specifies 66 OS/VS2 Planning Guide for Release 2 an IOC coefficient of 3.0, whenever a job issues an I/O request, the system resources management routines multiply the I/O resource (one processing resource) by the service definition coefficient (3.0), and that job is considered to have used three service units. Each of the service definition coefficients may be specified only once in an IPS. If a service definition coefficient is not specified, an IBM-supplied default value will be used. Service Rates Service rates are integers that define a target rate that system resources are to be supplied to an associated transaction. When grouped in sets called performance objectives, these define an intended service rate for any system workload condition. Tuning Parameters (IEAOPTxx) The parameters in the SYS I.PARMLIB member IEAOPTxx include: Resource Factor Coefficients Resource factor coefficients are values in the range 0.0 to 9.9. They specify values that are used as multipliers in the control algorithm to weight the recommendation values produced by the I/O and CPU load balancing algorithms. The two coefficients are 10C and CPU. Each coefficient has a default value of 1.0. The coefficients are used to make the recommendation values produced by the corresponding algorithms either more or less important than the recommendation values supplied by the workload management routines, for which an implied coefficient of 1.0 applies. Resource Manager Constants This category contains the enqueue residence value (ERV). The ERV is an integer in the range 0-999999. It is used to calculate the amount of time that an address space will not be swapped because the address space is enqueued upon a system resource required by another address space. The time interval is determined by multiplying the ERV by the (model dependent) time required to execute 10,000 machine instructions. (This procedure is intended to keep the execution interval relatively constant on all System/370 models.) Directing The Use of System Resources 67 System Activity Measurement Facility (MF /1) The system activity measurement facility (MF /1) is new in MVS. Installations may use it to monitor selected areas of system activity and obtain feedback in the form of trace records and/or formatted reports. Measurements may be gathered independently on: • CPU activity • Channel activity and channel-CPU overlap activity • I/O device activity and contention - Unit record devices - Graphics devices - Direct access storage devices - Communication equipment - Magnetic tape devices - Character reader equipment • Paging activity • Workloadactivity MF /1 activity is divided into three categories: • SAMG (system activity measurement gathering) obtains the measurements of system activity requested by the installation. • SARG (system activity report generation) produces formatted reports of desired system measurements. • MFC (measurement facility control) controls the collection, tracing and reporting of system activity. SAMG consists of distinct sets of measurement gathering (MG) routines associated with each class of measurement data. Only those sets associated with the measurement data requested by the system operator af(~ loaded and active during MF /1 reporting intervals. SARG generates formatted summary reports of information requested by the system operator, collected by SAMG, and routed to it by MFC. Printing of these reports may be done as they are generated, or may be defe:rred until MF /1 termination. MFC is the system task that controls MF /1 operation. It causes the loading of the MG routines that will be needed to obtain the desired information, passes control to these routines at the desired intervals, routes collected information to SARG as required, and terminates MF /1 execution at the ,end of its specified duration. MF/1 Operation The operation of MF /1 is controlled by a set of parameters that may be contained in: • The P ARM field of the START command that is issued to start MF /1 processing. 68 OS/VS2 Planning Guide for Relea!le 2 • The P ARM field of the EXEC statement in the MF I I cataloged procedure. • The MF I I partitioned data set member (typically a SYS l.P ARMLIB member). The IBM -supplied cataloged procedure for MF I 1 names the MF I 1 partitioned data set member that contains card image MF I 1 control input. Member names are of the form IRBMFlxx, where xx is a decimal digit field. In the absence of any explicit specification, the SYS I.PARMLIB member IRBMF 100 is used. The system operator issues the START command to begin MF I 1 monitoring. The parameters controlling the operation of MF I 1 are gathered from the above sources in the order stated. Explicit specification of an option in the P ARM field of the START command will therefore override any alternative specifications of that option in any other source. If a parameter is not specified in any of the three sources the system default value is used. The MF / I options are summarized below. The underlined values are the system default values, as well as the values that appear in the SYSl.PARMLIB member IRBMFIOO. CPU/NOCPU .PAGING/NOPAGING CHAN/NOCHAN WKLD(PERIOD I GROUP ISYSTEM) INOWKLD DEVICE(device list)/NODEVICE These parameters are used to specify which of the classes of MG routines (CPU activity, paging activity, channel activity, workload activity, 110 device activity) will be loaded and executing during the MF I 1 reporting period. UNITR/NOUNITR TAPE/NOTAPE DASD/NODASD COM'"M/NOCOMM GRAPH/NOGRAPH CHRDR/NOCHRDR These parameters are the elements of the device list above. They specify the device types (unit record, tape, direct access, communications, graphics, character reader) about which information will be collected. REPORT(REALTIME/DEFER)/NOREPORT This parameter specifies whether or not formatted reports of the gathered data will be generated, and when the reports will be generated (i.e., as the data is gathered, or after MF I I processing terminates). SYSOUT(class) This parameter specifies the SYSOUT class to which formatted reports will be directed. Class A is the default. RECORD/NORECORD This parameter specifies whether or not trace records will be written to the SMF data set. INTERVAL(value/valueM) This parameter specifies the interval at which data will be gathered for report formatting and/or trace record writing. The range is from 1 to sixty minutes. The default is 1SM. Directing The Use of System Resources 69 CYCLE(value) This parameter specifies the frequency at which observations are made of sampled data. The range is from 50 to 999 milliseconds. The default is 250 milliseconds. STOP(value/valueM/valueH)/NOSTOP This parameter specifies the desired time interval for MF /1 activity, in minutes or hours (the default is minutes). The operator STOP command will terminate execution at any time, regardless of the value for this parameter. The default value is 15M. MEMBER(nn) The value specified by this parameter is added to IRBMFI to form the name of the partitioned data set member that contains the MF /1 options. The default is 00. OPTIONS/NOOPTIONS This parameter specifies whether or not a list of the keyword options to be used will be printed at the operator's console. If the list is printed, the operator will be able to respond with any desired changes to the option list. MF/1 Reports and Tracing R'ecords Measurement output is produced in two forms: • Printed Reports -- The data is formatted and printed for each class of system activity data collected. • Tra(~e Records -- The data is written as a trace record to the SMF data set. One or more trace records are produced for each class of system activity data collected. The printed reports and trace records contain the same type of information, but in different forms. The installation can specify desired output as: • Printed reports only. • Trace records only. • Both printed reports and trace records. The following sections describe the contents of each type of printed report and trace record. CPU Activity Report This report provides information about the use of the central processing unit(s). The total CPU wait time that has taken place during an MF /1 reporting interval is generated for each CPU. The CPU wait time is also expressed as a percentage of the reporting interval time. Paging Activity Report This report provides detailed information about the demandls made on the systt~m paging facilities during the reporting interval. It also distinguishes between the use of real main storage and external page storage. The following categories of information are included: 70 OS/VS2 Planning Guide for Release 2 • Page reclamations -- Page reclamation rates and percentages are produced, for system pageable areas and the user's private address spaces, for both VIO and non-VIO accesses. • Page-ins -- Page-in rates and percentages are produced for all pages transferred from external page storage to real storage. This information is provided for system pageable areas and the user's private address spaces. • Page-outs -- Page-out rates and percentages are produced for pages transferred from real storage to external page storage. This information is provided for system pageable areas and the user's private address spaces. • External page storage counts -- Total page slots are produced, as well as the number of unused slots, unavailable slots, and slots containing either data set pages or virtual address space pages. • Pageable real storage counts -- The total number of page able real storage frames is produced, as well as breakdowns into counts and percentages of used and unused page frames. • Swap counts -- The total number of swaps are produced, as well as average pages per swap-out, and average pages per swap-in. Workload Activity Reports These reports provide information about the system workload activity and its relationship to the installation-specified performance objective for each performance group period. This information can be used to evaluate and modify the IPS. The three different workload activity reports that are available provide three different levels of detail. The types of reports desired are specified by input parameters indicated below from the most detailed report to the least detailed report. • Performance group period report -- (WKLD(PERIOD)) • Performance group report -- (WKLD( GROUP)) • System summary report -- (WKLD(SYSTEM)) Performance Group Period Report: This report provides, for each performance group period within each performance group: • The total number of service units used by all transactions in that performance group period. • The average service rate for all transactions within that performance group period. • The average workload level of all associated transactions. • The average number of transactions that were concurrently active. • The total number of transactions that terminated. • The average elapsed time of a transaction that terminated. Performance Group Report: This report provides, for each performance group, for one MF /1 reporting interval: • The total number of service units used by all transactions in that performance group. • The average number of associated transactions that were concurrently active. Directing The Use of System Resources 71 • The total number of associated transactions that terminated during the reporting interval. System Summary Report: This report provides, for one MF /1 reporting interval: • Tht: total number of service units used by all transactions. • The mean workload level of the system. • The average number of transactions that were concurrently active. • The total number of transactions that terminated during the reporting interval. Channel Activity Report This report provides information about channel loading for all channels in the system. For each active channel, a "channel activity" count provides the total number of successful SIO instructions (excluding the sense SIO instructions) issued to that channel during the reporting interval. A "percent channel active" indication provides the percentage of the total reporting interval during which each ehannel was busy. A "percent channel busy and CPU waiting" indication provides the percentage of the reporting interval during which each channel was busy while the associated CPU was in a wait state. An installation may use this information, in conjunction with the information contained in the I/O device activity reports, to determine possible bottlenecks associated with channels. I/O Device Activity Reports These reports provide information about I/O device loading for all devices in the chosen device class(es). The classes desired are specified in the input parameter string (UNITR, TAPE, DASD, COMM, GRAPH, CHRDR). For each device report, a "device activity count" indicates the number of successful SIO instructions (excluding the sense SIO instructions) issued to the device during the reporting interval. A "percent busy" indication provides the percentage of the reporting interval during which the device was busy. A "mean queue length" value provides an average number of requests that were enqueued and waiting to be serviced on the device during the reporting interval. An installation may use this information, in conjunction with the information contained in the channel activity reports, to determine possible bottlenecks in the I/O subsystem. SMF Recording If RECORD is specified among the MF /1 input parameters, a record will be written to the SMF data set, at the end of each reporting interval, for each report requested by the parameters. Each record contains information similar to the contents of the corresponding formatted report. Totals., averages, and percentages are not explicitly contained in the record, but may be calculated from the explicit data. The SMF records, and their corresponding printed reports are: 72 OS/VS2 Planning Guide for Release 2 SMF70 -- CPU Activity Report SMF71 -- Paging Activity Report SMF72 -- Workload Activity Report SMF73 -- Channel Activity Report SMF74 -- I/O Device Activity Report Tuning With MF /1 Tuning a system is the process of modifying its hardware or software characteristics to bring performance closer to installation defined objectives. Once an installation's performance objectives are clearly defined, MF /1 can be used to help an installation both specify and meet them. For example, the channel activity report and I/O device activity report can provide information about the efficiency of the system-wide I/O load balance and how it affects system throughput. Analysis of these reports could indicate possible system configuration changes or data set residence modifications. MF /1 can also be used to determine if the IPS reflects the installation's turnaround/ response requirements and, if so, whether the IPS objectives are being met. It can further be used to analyze the specified performance objectives for possible modification. Tuning Example Figure 14 shows the division of the installation's jobs into three performance groups. Assume that users associated with performance group 7 -- the time sharing users -- are experiencing poor response time at their terminals. The installation decides to activate MF /1 for a three hour period during a typical day, to obtain six reports on the workload activity for analysis. To initiate this reporting the operator enters: START MF1.MF1",(NOCPU,NOPAGING,NOCHAN, NODEVICE,INTERVAL(30M),STOP(3H)) (Assume that there are no EXEC statement parameters in the MF /1 procedure, and that the default SYSl.PARMLIB list IRBMFI00 is used.) With the above command and the standard defaults, the following options will control MF /1 execution in this case: NOCPU NOPAGING NOCHAN WKLD(PERIOD) NODEVICE REPORT (DEFER) SYSOUT(A) NOTRACE INTERV AL(30M) CYCLE(250) STOP(3H) OPI Directing The Use of System Resources 73 IPS Values FIRST PERIOD SECOND PERIOD CI> CI> '~ ';:; CI> CI> > > Performance Group Significance of Performance Group to Installation Number c: co .~ c: ::> co Terminal Jobs Interaction 8 High Priority Batch 3000 9 Low Priority Batch Entire Job 100 0 ';:; ~ co a: > I- CI> c.. 5Q S 8 5 System Workload Level " Q) u .~ Q) CI) 40 20 W 2 4 Workload Level Figure 14. Tuning Example 74 OS/VS2 Planning Guide for Release 2 6 E co ....'" C:l :::> :; Rest of Job 60 o co c: ~ CI> c.. 2 80 CI> ..... co c: c: E :; Performance Objectives t,) t,) CI S CI> CI> ';:; Performance Objectives 0 0 0 7 t,) 'B 'B c: -- 8 - 8 I- co a: :::> - - ~ Assuming that the operator does not prematurely terminate MF /1 execution with the STOP command, MF /1 will terminate at the end of three hours, and the six reports that were collected at half-hours intervals will then be printed. The installation can now begin its analysis. Two figures of interest in this case are the average system workload level, and the average active time for all performance group 7 transactions. Suppose that these reports indicate that, during the time that MF /1 monitoring took place, the system was operating approximately at workload level 4 and that the average response time for time sharing users was 30 seconds (considered by the installation to be poor response time). An analysis of the performance objective specification shows that, at workload level 4, performance objective 2 specifies that performance group 7 transactions receive 40 service units per unit time. The installation might reasonably wish to increase the service rate for its time sharing users in order to improve the response time. This increase would raise the performance objective curve for performance objective 2, as shown in Figure 15. S Performance Objectives System Workload Level 100 ® 80 j!! cv a: CII Col .. en 60 'S: ........ CII ~==~ 40 -.-- .... ....... - - .... • 20 o 2 4 6 8 W Workload Level Figure IS. Changing a Perfonnance Objective Directing The Use of System Resources 75 Figure 16 shows that raising the service rate for performance objective 2 raises the system workload level, aU other factors being equal. Transactions associated with performance objectives 5 and 8 would, therefore, be affected. Transactions associated with performance objective 5 would be most severely affected because their demand curve slopes most severely to the right of workload level 4. Performance Objectives S 100 - @ System Workl~ad I Level _- 80 - ... - CIl III a: CIl u .~ 60 - - CIl en -t----_ == § - - ..... ~ 40 t §--t 20 L -______~____--~------~--------~W o 2 4 Workload Level Figure 16. Effects of a Perfonnance Objective Change Before the installation modifies performance objective 2, it is able to perform further analysis to determine the quantitative effects of such a change. Suppose the MF /1 reports showed an average of 20 jobs associated with performance objective 2, 8 jobs with performance objective 5, and 4 jobs with performance objective 8. Then, some measure of the average system rate of service supplied to all jobs can be obtained by multiplying the number of jobs at each performance objective by the service rate for that performance objective at workload level 4 (i.e., 20 x 40 + 8 x 20 + 4 x 10 = 800 + 160 + 40 = 1000 total service units per unit time). In this case, increasing the performance objective 2 curve by 10 service units, results in a total increase in demand of 200 service units (10 times the 20 jobs associated with performance objective 2). Thus, the workload level of the system shifts to the right (increases) to compensate for this increase in demand, while maintaining the system supply of 1000 total service units approximately constant. (See Figure 16.) If the installation considers the service rates provided for all performance objectives at the projected new workload level to be acceptable, it may modify the IPS to contain the new performance objective 2. The installation may then use MF /1 monitoring to verify the accuracy of the projections; repeating the entire procedure, if desired, until satisfactory results are achievt~d. 76 OS/VS2 Planning Guide for Release 2 The preceding example is not meant to reflect actual figures or relationships, but rather to indicate the possibilities for tuning that MF /1 can provide. The illustrated method of improving response time is neither the only one, nor necessarily the best -- other alternatives include varying the interval service value parameter, dividing the time sharing performance group into more than one period, and increasing the number of performance groups provided for time sharing users. Any method chosen to improve this or other system performance characteristics will suggest other uses and variations for MF /1. Directing The Use of System Resources 77 78 OS/VS2 Planning Guide for Release 2 System Integrity A highly desirable property of an operating system is its ability to insure that one program cannot interfere with or modify another program's (system or user) execution unless it is authorized to do so. For reliability and system availability, this is extremely important; for data security, it is essential. MVS provides this capability in the form of system integrity support. System integrity is defined as the ability of the system to protect itself against unauthorized user access to the extent that security controls cannot be compromised. That is, there is no way for an unauthorized* problem program using any system interface to: • Bypass store or fetch protection, Le. read or write from or to another user's areas. • Bypass password checking, Le. access password protected data for which a password has not been supplied. • Obtain control in an authorized state. In MVS all known integrity exposures have been removed. IBM will accept as valid, any AP AR that describes an unauthorized program's use of any system interface (defined or undefined) to bypass store or fetch protection, to bypass password checking, or to obtain control in an authorized state. *An authorized program in MVS is one that executes in a system key (keys 0-7), in supervisor state, or is authorized via the Authorized Program Facility (APF). System Integrity 79 Installation Responsibility To ensure that system integrity is effective and to avoid compromising any of the integrity controls provided in the system, the installation must assume the responsibility for the following items: The installation must be responsible for the physical environment of the computing system. Operations personnel and system programmers have, in effect, uncontrolled access to certain portions of the operating system. Problems of this nature are discussed in The Considerations of Physical Security in a Computer Environment, GS20-2700. The installation must be responsible for the adoption of certain procedures (e.g. the password protection of appropriate system data sets) that are a necessary complement to the integrity support within the operating system itself. The most significant of these procedures are described briefly in the topic "Installation Procedural Responsibilities." The installation must ensure that its own modifications and additions to the control program do not introduce any integrity exposures. That is, all installation-written authorized code (e.g. an installation SVC) must perform the same or an equivalent type of validity checking and control that the MVS control program employs to maintain system integrity. 80 OS/VS2 Planning Guide for Release 2 Areas of Concern Several areas with the potential for integrity exposures have been identified. These areas, and what has been done to eliminate them as potential sources of integrity exposures, are described below. The description is a guideline to aid the installation in preserving system integrity in any modification or addition to the control program. It is also intended to aid in determining whether an impact on existing installation cody might occur, where that code is dependent on the use of non-standard interfaces to the control program. There should be no impact on installation-written routines that use standard interfaces (problem program/system interface described in an SRL), because no standard interface has been removed from the system control program in connection with system integrity support. MVS system integrity support is concerned only with restricting the unauthorized problem program; there is no attempt or intention to prevent the installation manager or any authorized system programmer from modifying the system control program. User-Supplied Addresses for User Storage Areas A potential integrity exposure exists whenever a routine having a system protection key (key 0-7), accepts a user-supplied address of an area to which a store or fetch is to be done. If the system routine does not adequately validate the user-supplied address to ensure that it is the address of an area accessible to the user for storing and fetching data, an integrity violation can occur when the system key program routine: • Stores into (overlays) system code or data (e.g. in the nucleus or the system queue area), or into another user's code or data . • Moves data from a fetch-protected area that is not accessible to the user (e.g. the fetch-protected portion of the common service areas), to an area that is accessible to him. The elimination of this problem requires that system key routines always verify that the entire area to be stored into, or fetched from, is accessible (for storing or fetching) to the user in question. The primary validation technique is the generally established MVS convention that system key routines obtain the protection key of the user before accessing the user-specified area of storage. For example, MVS data management SVC routines (which generally run in key 5) assume the user's key before modifying a data control block (DCB) or an I/O block (lOB). User-Supplied Addresses for Protected Control Blocks A potential integrity exposure exists whenever the control program (system key / privileged mode) accepts the address of a protected system control block from the user. For most system control blocks this situation should not be permitted to exist. However, in certain cases it is necessary to allow the user to provide the address of a system control block that describes his allocation/access to a particular resource (e.g. a data set), in order to identify that resource from a group of similar resources, (e.g. a user may have many data sets allocated). Inadequate validity checking in this situation can create an integrity exposure, since an unauthorized problem program could provide its own (counterfeit) control block in place of the system block and thereby gain the ability to: System Integrity 81 • Access a resource in an uncontrolled manner (since the control block in this case would normally define the restrictions, such as read-only for a data set, on the user's allocation to the resource. • Gain control in a privileged state (because such control blocks may contain the addresses of routines that run in privileged mode or with a system (0-7) key. • Cause various other problems depending on exactly what data the system may be keeping in the control block involved. To avoid this type of exposure, the control program must verify, for every address of this type that is accepted from a problem program, that the address is that of: 1. A protected control block created by the control program. 2. The correct type of control program block (e.g. a TCB versus a DEB, or a QSAM DEB versus an ISAM DEB). 3. A control block created for use in connection with the user (job step) that supplied the address. In MVS, verification is generally accomplished by establishing a chain or table of the particular type of control block to be validated. This chain or table is located via a protected and job step-related control block that is known to be valid (i.e. whose address is not allowed to be supplied by the user, and is located via a chain of protected control blocks that begins with a control block known to be valid or fixed at a known location at IPL time, such as the CVT*). Thus a control block can only be entered in the chain/table by an authorized program, which satisfies (1) above; by definition the chain/table establishes the type of control block, which satisfies (2) above; and also by definition each chain/table is located only through a job step-related control block, satisfying point (3). Resource Serialization Resource serialization is another potential source of integrity problems. Here an integrity exposure can result when serialization controls on sensitive control program resources are nonexistent, or inadequate to the exltent that an unauthorized problem program can directly, or indirectly through a part of the control program (e.g. by issuing an SVC), change the content or status of a system resource (e.g. control block, work area), while another portion of the control program (e.g. another SVC) is using that resource or is in some way dependent on its content or status remaining unchanged for a given period of time. Adequate control of serial resources becomes even more significant in a multiprocessing environment since some uniprocessor serialization techniques (e.g., disablement), are no longer effective because of the possibility of multiple tasks, even mUltiple tasks in the same job step, using the same serial resource and running concurrently on different CPUs. To eliminate this as a potential exposure, a locking technique is used to serialize access to the resources in question. However, a locking technique is not *This does not imply that a system routine must go back to the CVT or similar control block every time it wants to establish a valid chain. Typically, a control block address not too far down on such a chain is available already validated in a register. For example, the first load of an SVC may receive control with a valid TCB address in a lcgister. 82 OS/VS2 Planning Guide for Releas!~ 2 effective unless all routines that have the potential for changing the resource, or that depend on its status remaining unchanged for a given period, make use of the (same) locking mechanism. Effective use of a locking technique therefore requires considerable investigation and effort, to determine and keep track of system resources that must be serialized and routines that access such resources. In MVS, a combination of ENQ/DEQ and a new hierarchical locking structure with multiple typ-es of locks, is the primary method used to synchronize the use of serial resources. The locking mechanism is restricted to key 0 programs, which prevents unauthorized problem programs from interfering with the serialization of system resources. For the same reason, ENQ/DEQ is restricted to authorized programs for all major names of the form SYSZxxxx, and for the following explicit major names: SYSDSN, SYSIEECT, SYSIEFSD, SYSIEAOl, SYSVTOC, SYSPSWRD, SYSCTLG, SYSIGGVl, SYSIGGV2, and SYSVSAM. Another integrity consideration, related to the above problem, but generally applicable to all validity checking for integrity purposes, is the time-of-check-to-time-of-use problem. This refers to the fact that from the time of a validity check until the completion of the operation associated with that validity check, the variables on which the outcome of the validity check is based must not be allowed to change to the extent that the result of the validity check would be changed. While this is a relatively obvious statement, the requirement it imposes has considerably influenced the direction of the integrity support in MVS. In some cases, this requirement can be met "automatically" through use of hardware serialization mechanisms, such as checking the validity of fetch/store operations by assuming the user's protection key. In other cases, techniques such as saving validated data in protected (storage keys 0-7) core and use of the above-mentioned locking mechanisms are required. Resource Identification Resource identification is another area that can be subject to integrity exposures. Exposures can result if the control program does not maintain and use sufficient data to uniquely distinquish one resource from other similar resources. For example, a program must be identified by both name and library to distinquish it from other programs. The consequences of inadequate resource identification are problems such as the ability of an unauthorized problem program to create counterfeit control program code or data, or to cause varying types of integrity problems by intermixing incompatible pieces of control program code and/or data. The general solution can only be stated as the reverse of the problem~ that is, the control program must maintain and use sufficient (protected) data on any control program resource, to distinguish between that resource and other control program or user resources. The following are examples of the controls that MVS employs to comply with the requirement: • In general, authorized program requests to load other authorized programs are satisfied only from authorized system libraries (see the topic "Control Program Extensions"). • MVS takes explicit steps to ensure that routines loaded from authorized system libraries are used only for their intended purpose. This includes expanded validity checking to remove any potential for the unauthorized System Integrity 83 program to specify explicitly which of the authorized library routines are to gain control in any given situation. • Sensitive system control blocks are validated as being the "correct" blocks to be used in any given control program operation. (See the topic "User-Suppli(!d Addresses of Protected Control Blocks"). SVC Routines Calling SVC Routines A potential problem area exists whenever a problem program is allowed to use one SVC routine (routine A) to invoke a second SVC routine (routine B) that the problem program could have invoked directly. An integrity exposure occurs if SVC routine B bypasses some or all validity checking based on the fact that it was called by SVC routine A (an authorized program), and in addition, user-sulPplied data passed to routine B by routine A either is not validity checked by routine A, or is exposed to user modification after it was validated by routine A. This problem will not exist if the user calls SVC routine lB directly, because the validity checking will be performed on the basis of the caller being an unauthorized program. SVC routine A, which is aware that it has been called by an unauthorized program, must ensure that the proper validity checking, etc., is .accomplished. However, it is usually not practical for SVC routine A to do the validity checking itself, because of the potential for user modification of the data prior to or during its use by SVC routine B. The general solution should be for SVC routine A to provide an interface to SVC routine B, informing routine B that the operation is being requested with user-supplied data in behalf of an unauthorized problem program (implying that normal validity checking should be performed). In practice, in MVS, most SVC B type routines that could! be subject to this problem, use the key of their caller as a basis for determining whether or not to perform validity checking. Therefore most SVC A type MVS routines have simply adopted the convention of assuming the key of their caller before calling the SVC B routine. Control Program and User Dtrta Accessibility Important in maintaining system integrity, is the consideration of what system data is sensitive and must be protected from the user, and what data can be exposed to user manipulation. The imp~ications of the exposure of the wrong type of data are obvious. In general, it is necessary to protect the following types of data: • Code, and the location of code, that is to receive control in an authorized state., • Work areas for such code, including areas where it saves the contents of registers. • Control blocks that represent the allocation or use of system resources. S4 OS/VS2 Planning Guide for Release :e MVS maintains items such as the above in system storage, or in a separate region in the case of some APF-authorized (see the following section) programs. It may also be necessary to protect, for a limited period, certain data that is normally under the control of the user (e.g. to prevent its modification during a critical operation). In this case MVS provides fetch protection for such data if: • The data consists of proprietary information (e.g. passwords). • The control program cannot determine the nature of the contents of the data area. Control Program Extensions The final potential problem area involves the fact that there exists a somewhat hazy distinction between the control program and certain types of problem programs. In most installations, there are problem state/user key (keys 8-15) programs that are actually extensions to the control program in that they are allowed (by means of various special SVCs, etc.) to bypass normal system controls over access to system resources. For example, a special utility program that scans all the data on a pack might be able to avoid the normal system extent checking on a direct access volume. The Authorized Program Facility (APF) was introduced in VS2 Release 1 as a means of identifying these control program extensions as "authorized" programs, and restricting the use of the various "special" SVCs used by such programs to bypass normal system controls. This use of APF is continued in MVS with the restriction of SVCs under APF extended to include certain new SVCs introduced in MVS. If an installation has its own control program extensions and special SVCs that allow the bypass of normal system security or integrity checks (e.g., an SVC that returned control in key 0), and if such SVCs are not currently restricted from use by an unauthorized program, then the APF facility should be used to restrict these SVCs and to authorize the control program extensions that use them. A description of how to use the APF facility can be found in appropriate VS2 publications. An important MVS addition is that APF has been extended to recognize and accept authorized programs from "authorized" libraries other than SYSl.SVCLIB, SYSl.LINKLIB, and SYSl.LPALIB. Any library may be designated as an authorized library by inclusion of its name in the SYS 1.PARMLIB member, IEAPFOO or IEAPFxx, prior to IPL. Such libraries are the equivalent of the above system libraries in that nonspecific system requests (library not specified) for routines that will execute with system key, privileged mode, or APF authorization, can be satisfied by a module loaded from any authorized library. APF supports a facility (TESTAUTH) used to control the execution of certain paths through SVCs, such as the path that opens (OPEN) the VTOC. This facility may not be used to control the use of I/O appendages in MVS. Appendages ~ controlled through SYS 1.PARMLIB as described above. Note: The user of APF should also be aware that APF authorization is defined such that any SVC that is restricted by the APF mechanism, can be executed by any system key (0-7) or (in most cases) privileged mode routine. However, this is not true in reverse. An APF authorized program will not automatically be allowed system services that are restricted by system key or privileged mode tests .. System Integrity 85 Installation Procedural Responsibilities As noted previously, the installation is responsible for the implementation of procedures that are necessary in certain instances to complement the MVS contro] program integrity/security support. These responsibilities are discussed below. Note that in some instances, the concern is for basic system integrity, while other items are concerned with a particular type of security. 1. The installation must password protect appropriate system libraries. System integrity obviously cannot be maintained if system code and data is exposed to arbitrary modification by any user on the system. For integrity purposes, it is generally sufficient to protect appropriate libraries from write access; no password is required for read access but a password is required for write access. However, for security purposes, it is necessary to protect certain system data sets (e.g. the PASSWORD data set itself) from read as well as write accesses. To facilitate the protection of such data sets, during IPL and system task initialization password requests are suppressed for data sets being opened by the system. Certain auxiliary storage manager (ASM) initialization procedures make password protection of the SYSl.STGINDEX data set ineffective. However, ASM provides protection via an ENQ on the data set name, effectively preventing another user from allocating the data set. 2. Because the checkpoint data set produced by the Checkpoint/Restart facility contains sensitive system data normally protected from the user, maintaining system integrity requires that such data sets be protected from modification (or from being counterfeited) prior to their use by Restart. MVS implements a facility whereby the installation can adopt a set of special procedures/controls over checkpoint data sets that will eliminate their potential for compromising system integrity. The control mechanism involves a combination of: • System/operator validation of checkpoint data sets. • External labeling procedures for "checkpoint volumes." • Offline control of "checkpoint volumes." • Prohibiting I/O to checkpoint data sets, except through the checkpoint SVC (APF-authorized programs excepted). 3. The installation should be aware of the following password protection requirements, which may necessitate some changes in installation procedure. • Because of the characteristics of magnetic tape devices, once a program has access to one file on a tape, it can not be ensured that the program is not able to access other files on that tape, even though it may not possess passwords for those files. For this reason, MVS has implemented a changt~ in the handling of password-protected tape data sets to attempt to ensure that if a user is allowed access to one data set on a tape volume, he can be presumed also authorized to other data sets on the volume. To this end, whenever a user attempts to create a tape data set whose file sequence number is greater than 1, the protection attributes of the new file will be compared against the protection attributes of the immediately preceding file on the tape (no protection, protected for write, protected for read and write, etc.). If the protection attributes do not match, the new data set will not be created. In addition, if the previous data set is password protected, its password must be supplied before the new data set will be created. 86 OS/VS2 Planning Guide for Release 2 Note that these changes do not eliminate problems caused by an installation library containing already-existing tape files with mixed protection attributes. • The security-conscious installation should ensure that password-protected data sets are created only on tapes which are unused or completely erased. This is necessary because in certain cases residual information left on the tape after the password data set was created, could result in a user being able to open one of the previous data sets on the tape, thus gaining access to the whole tape. 4. In the case of the 2840/2250, the installation may want to use dynamic allocation and JCL validation exits to achieve additional control over device allocation and user identification. Fetch/store protect is not available for the 2840 buffer, which can be shared by up to four 2250 graphic devices. The above-mentioned exits can be used to establish a control mechanism that restricts to users of the same security classification, 2250's that are connected to the same 2840. 5. Although APF -authorized programs bypass the usual system control validity checks, in any user interfaces they must provide validity check mechanisms the same as or equivalent to the system validity checks. The effectiveness of APF is dependent on this. However, it may be impractical to provide these checks for some APF-authorized programs, and in these cases their use should be restricted by placing them in APF-authorized password protected libraries. The passwords for these libraries should then be provided only to individuals authorized to use the programs. Installations might consider using this control mechanism for the IEHINIT, OLTEP, and IEHDASDR routines. 6. If there are two modules with the same name on different authorized libraries and an authorized program attempts to load one of these, it could get the other (either accidentally or deliberately with JOBLIB, STEPLIB, or concatenation). This type of exposure requires the additional restriction that duplicate module names not be permitted across authorized libraries. 7. The optional update password for VSAM catalogs provides protection against unauthorized delete or alter operations on non-VSAM data set entries~ operations on VSAM data set entries are protected under the password for the data set. Catalogs (e.g., the master catalog) containing sensitive non-VSAM data sets should be protected under the update password. 8. OS CVOLs should be password protected (one password for all CVOLs under the name SYSCTLG) to avoid concurrent access between OS catalog routines and an unauthorized user. However, this does not prevent unauthorized delete or alter operations on the catalog. This protection is not offered for OS CVOLs. 9. Installation-written programs (e.g., I/O appendages) that execute in supervisor state or system key (0-7) in the user's address space, must be link edited reentrant to ensure that they are loaded in a protected subpool. System Integrity 87 88 OS/VS2 Planning Guide for Release 2 Conversion Considerations This chapter describes significant differences and incompatibilities between OS/MVS and: • OS/MVT • OS/VS2 Release l. It is divided into the following sections: • Job entry subsystem considerations. • SMF considerations. • Job control language. • Communication between address spaces. • Allocation considerations. • Operator commands. • Time sharing considerations. • Catalog support. • Data set conversion. • V =R programming considerations. • Program conversion. • Multiprocessing considerations. Conversion Considerations 89 Job Entry Subsystem Considerations The job entry subsystems available with MVS are: JES2 -- generally compatible with HASP II. JES3 -- generally compatible with ASP Version 3. With the first release of MVS, only JES2 will be available; therefore JES3 information reflects current design and is for advanced planning purposes only. All jobs, started tasks, and LOGONs processed by the system must be entered through a job entry subsystem, which must therefore be started before any processing is done. The primary job entry subsystem is started automatically at system initialization. Any other is started by the operator. The JES2 Job Entry Subsystel" JES2 functions as a job entry and system output processor. It also starts and stops system initiators, selects jobs to be processed by initiators, and frees resources when the jobs are complete. JES2 controls a job prior to execution and following job termination. 90 OS/VS2 Planning Guide for Release 2 JES2 Job Flow Figure 17 is an overview of the JES2 processing stages. Input, execution, output, and purge services execute in the JES2 address space. Other functions depicted execute in system or user address spaces. Local and Execution Service Input Service • •It! Pre-E xecut ion Control Execution Control Converter Output Service Print/ Punch Service Executing Job or System Task Purge Service Output Processor External Writer Remote or Local Figure 17. JES2 Processing Stages Input Service: Jobs are read into the system from various types of card readers and remote terminals, and through the internal reader interface. An internal reader is an output (SYSOUT) data set allocated to JES2, which is recognized and placed in the input stream. Jobs and system tasks submit jobstreams in this manner. Each job or user is given a unique job identifier; then the JCL and input data are stored on direct access volumes for retrieval during Execution Service. Compatability: Input service functions as a complete replacement for the OS reader; however no catalog procedures are associated with this function. Procedure libraries and interpretation parameters are associated with jobs by JES2. Prior to execution, jobs are maintained on JES2 execution queues rather than on SYS 1.SYSJOBQE. Pre-execution manipulation must occur through the use of JES2 commands rather than system commands. Conversion Considerations 91 Jobs with JCL syntax errors fail during conversion. The job is sent directly to output service, thus the JCL is not interpreted. Support of the DLM parameter on DD compatible with OS support. * or DD DATA JCL statements is JES2 ignores the" / /" (null) statement. The only effect of this statement is a possible higher input record count than MVT. Execution Service: Through this service, jobs are submitted to the system for execution. Conversion takes place immediately after input service. The JCL is converted into internal text, unless syntax errors are discovered. Job selection takes place as soon as an initiator eligible to process the job is available. Jobs are selected in priority sequence within their class. The subsystem interfaces for the following functions are supported through execution service: • Subsystem job selection (including presenting jobs for warmstart). • Allocation/ unallocation of JES2 data sets. • Sysout output interfaces (TSO output TMP). • Job status. • Job cancel. • End of storage. • End of task. • WTO /WTOR/DOM messages. • Command processor. • Verify remote user id. • Job termination. • Job re-enqueue. • OPEN/CLOSE of JES2 data sets. Compeltability: JES2 does not support the Execution Task Monitor of OS/MVT HASP systems. The System Resources Manager of MVS accomplishes the same functions, with no defined interface to JES2. All sysin/ sysout data sets are maintained by JES2. JES2 provides an access method for reading and writing the spool data sets in response to user requests. This method of accessing spool data is different from that provided by OS /MVT HASP. JES2 is responsible for presenting jobs for warmstart or restart. JES2 supports 36 job classes rather than 15. (A-Z, 0-9) EXCP to a JES2 data set is not supported. NOTE/POINT to DCBs processing JES2 data sets is nOit supported. JE82 supports the IEFUSO SMF exit. 92 OS/VS2 Planning Guide for Release 2 Output Service: The print data sets and system messages created during execution are printed; the punch data sets are punched. JES2 operator commands are used to control JES2 printer/punch devices. Operators at remote stations can control only those JES2 devices attached to their particular station. When each group of job-related data sets and each spinoff data set are processed, output service writes one SMF type 6 record to the SMF data set. Compatabi/ity: JES2 does not attach user-written sysout writers. The external writer interface must be used for special sysout processing. Purge Service: When all processing required for a job is completed, the direct access space maintained by JES2 and all JES2 resources associated with the job are made available. The SMF purge exit is scheduled and the SMF purge record written. JES2 Warmstart For a warmstart occurring under JES2, all spool volumes that were up during the last execution must be available, although they need not be mounted on the same drives. Jobs with no job journal or with an irretrievable journal are not presented to the system for warmstart processing. New jobs may be presented to the system and will be processed in priority order -- ahead of lower priority warmstart jobs. MVS Operational Changes JES2 performs the reader function of the MVT Reader/Interpreter, and functions of the MVT Output Writer and Remote Job Entry. Unit record devices used for reading jobs and printing or punching output, and teleprocessing lines used for accessing remote stations are normally assigned to JES2. Thus a deVice, even though idle, may need to be disassociated from JES2 before it can be used by other system jobs or tasks. Converter /Interpreter parameters formerly placed in the Reader Procedure PARM field for each specific reader, are included in the JES2 job class related parameters established during JES2 initialization. All catalogued procedure libraries to be used by jobs, LOGON, or system tasks must be included in the JES2 catalog JCL. JES2 class-related parameters and Job Control Cards associate the procedures with jobs by job class. A card reader can be dedicated to JES2 in such a manner that the operator need only ready the device to cause JES2 to read a jobstream. A catalogued procedure is provided to read a job stream from any QSAM-supported device through the internal reader interface. This method is used to read jobs from local devices other than card readers; it must also be used if input stream positioning is desired. Column binary cards must be read by an installation-written routine and passed to JES2 through the internal reader interface. The OS/360 standard separator page is not supported -- the special JES2 separator page is a combination of those provided by MVT HASP and OS/360. Conversion Considerations 93 An external writer is provided to support user-written output writers and the standard OS/360 separator page. This allows a user to write data to devices not directly supported by JES2 or JES3 (e.g., tape and disk), and to use installation-written job separator routines. Remote stations may be attached with point-to-point based or dial lines. Multi-point support is not provided. The direct sysout writer is not supported. A sysout data set is available for output service processing prior to the end of the job if the data set is unallocated at CLOSE. This facility is called sysout spinoff. JES2 output service attempts to minimize forms, UCS, and FCB loading while processing output data sets. Operator interaction is minimized, and data may be processed in an order different from MVT. A JES2 generation option allows an installation to specify that output in the message class of a job be processed continguously. Sysout data sets and system messages are processed directly by JES2 and written into the JES2 spool data set, rather than into separate direct access data sets. These data sets may be retrieved at a time sharing terminal or by an external writer. Data sets freed by a foreground user are immediately available for output processing by JES2, or for inspection via the OUTPUT command. A time sharing user can use JES2 control cards to specify job-related parameters for his submitted job. JES2 schedules the IEFUSO SMF exit and the SMF purge exit, and writes the SMF output, purge, and subsystem activity records (types 6" 26, and 43-46 respectively) . 94 OS/VS2 Plannjng Guide for Release 2 JES2/HASP Compatibility Figure 18 describes some areas of compatibility between HASP and JES2. Some additional items and further clarification follow: VS2 Release 1/HASP MVT/HASP JES2 Tape input (via HASP START command) Internal reader (via a cataloged procedure) Internal reader (via a cataloged procedure) Execution task monitor Dynamic dispatcher System resources management routines Binary synchronous code adapter and synchronous transmitter receiver communications facilities Binary synchronous code adapter communications facilities only Binary synchronous code adapter communications facilities only Pseudo devices for SYSI N/SYSOUT via QSAM, BSAM, or EXCP Pseudo devices for SYSIN/SYSOUT via QSAM, BSAM, or EXCP Subsystem interfaces for SYSIN/SYSOUT via QSAM and BSAM only Installation-written output writers only via the system output writer Installation-written output writers only via the system output writer Subsystem interface allows installationwritten routines to obtain SYSOUT data sets from JES2 HASP console services Qt multiple console support Multiple console support with HASP interfaces Multiple console support with JES2 interfaces Purge card Purge record written to SMF data set Purge record written to SMF data set System output queued by four types System output queued by forms, printer set-up, and 36 output classes System output queued by forms, printer set-up, and 36 output classes No equivalent function Forms, printer set-up, and routing information may be specified on a HASP OUTPUT control card Forms, printer set-up, and routing information may be specified on a JES2 OUTPUT control card Restricted checkpoint/restart support for jobs using SYSIN/SYSOUT Restricted checkpoint/restart support for jobs using SYSIN/SYSOUT Full checkpoint/restart support No SMF SYSIN record counts No SMF SYSIN record counts No SMF output limiting support No SMF output limiting support No SMF output record SM F output record is written to the SMF data set Numeric forms designation Alphanumeric forms designation Alphanumeric forms designation TSO SYSOUT data sets are written by the system output writer TSO SYSOUT data sets are written by the system output writer Time sharing SYSOUT data sets are processed by JES2 and are retrievable via the OUTPUT command Background jobs scheduled from a terminal must use system facilities rather than HASP TSO users may submit jobs to HASP for scheduling, inquire of their status, and cancel them Time sharing users may submit jobs to JES2 for scheduling, inquire of their status, and cancel them Full SMF support Figure 18. JES2 Differences From MVT/HASP and VS2 Release l/HASP • JES2 is distributed on the MVS Starter System rather than on a separately ordered component tape. Also, JES2 is automatically started by the system at IPL, rather than by the operator. • During initialization, if errors are found in the arrangement of mounted spool volumes the operator is informed and JES2 waits for the situation to be corrected. Conversion Considerations 95 • Pseudo-devices are no loner used for specification of input and output devices and internal readers. System interfaces are provided with data sets identified by JES2-generated internal names, and the Internal Readers identified by the presence of the name· INTRDR as the user-written Writer name in the SYSOUT parameter of the DD statement. • Extended setup control affects the operation of output dlevices. Output service dequeues data for output devices based on output classes associated with each output device, rather than according to the characteristics of the device (e.g. special forms printer). Any of the 36 valid output classes may be specified for a given output device. Default classes specified during job entry subsystem generation can be overridden by the operator. The order of appearance of dat.a sets on an output device may be different from MVT HASP for a given job based on special forms requirements. SYSOUT is sent directly to JES2; otherwise output service processing is compatible between JES2 and VS2 Release 1 HASP. The JES3 Job Entry Subsystem JES3, which includes some scheduling functions not providled by JES2 plus support for loosely-coupled systems, is designed for the uSler with a medium-to-Iarge job-shop environment. As with JES2, JES3 replaces the OS readers and writers, and SYS1.SYSJOBQE. Figure 19 shows a possible JES3 multiprocessing configuration, with MVS and other systems connected via channel-to-channel connectors (CTCs). JES3 executes in all processors. The controlling processor in a JES3 configuration is called a Global processor; under ASP it was called the "Support" and "Local Main" processor. The other JES3 processors are called Local processors; under ASP they wt!re called "Main" or "Real Main" processors. The Global processor controls job input, processes job output, and gives commands to the Local processors through the CTCs. From the global console, the multiple processors of a JES3 complex appear as a single system. The primary enhancement to the ASP concept of loosely-coupled systems is the addition of dynamic system interchange (DSI). In case the Global processor fails, DSI allows the functions of the Global processor to be switched (operator-assisted) to any attached MVS Local processor. This processor then becomes the Global processor for itself and any MVS or MVT processor which is attached to it. 96 OS/VS2 Planning Guide for Release 2 Global Processor Local Processor MVS MVS Up to 3 MVS Local Processors can be connected Local Processor Shared JES3 Job Queue CTC* ASP Main Processor VS2-1 or MVT CTC Maintask -B SYS1.SYSJOBQE Up to 29 MVT or VS2 Release 1 Main Processors can be attached. ASP Main Processor VS2-1 or MVT r r---. ~ CTC Maintask ./ SYS1.SYSJOBQE * A CTC connection between Local processors is recommended (not required) to allow JES3 dynamic Global switching in the event of Global failure. Figure 19. Sample JES3/ ASP Multiprocessing Configuration Conversion Considerations 97 JES3 Scheduling Features Some of the features which enhance loosely-coupled systems and assist uniprocessing environments are: • Pre execution setup is provided for direct access, graphic, unit record, and tape devices which are not assigned to JES3 for job input and output functions. JES3 determines the device and volume requirements for each job. • Selection of a job for a processor is based on the ability of that processor to provide sufficient devices. JES3 switches pooled devices among processors to further control job/processor selection. • A user exit is provided to determine which processor is eligible to execute a given job. Information regarding current workload is available at this exit. • The operator is instructed to premount the volume(s), after the eligible processors are determined and the devices are reserved for the job. • Optionally, preexecution setup need occur for only the largest number of devices used in anyone step of the job. In previous releases (through ASP 3.0), the number of devices setup was the total number of devices used in all steps. For example: Device 2314 2400 # used in step 1 2 3 #used in step 2 4 0 #used in step 3 1 5 ASP total JE to 7 8 • Multiple jobs on different processors can share direct access volumes. Data sets, with disposition specified as shared, can also be shared between processors. • JES3 volumes can be set up dynamically for dynamically allocated data sets. These devices are returned to the system when the data sets are deallocated. • Dependent job control (DJC) allows simple or complex job interdependencies that are present in many commercial data processing installations, to be scheduled through JES3. The success or failure of one job under DJC can cause execution, holding, or cancelling of other dependent jobs. • Devices can be pooled for jobs under DJC, or for job-(~lass groups. This alJows devices to be reserved for jobs with data set dependencies. • Deadline Scheduling increases the probability of a job completing by a specified time. The selection priority of the job is dynamically adjusted. • Network job processing (NJP) allows input and output from compatible JES3 installations. Jobs cannot be transmitted between JES3 and ASP. Job Definition A job in the JES3 system consists of one or more job segments. A job segme:nt (e.g., PRINT or PUNCH) is defined as a logical portion of a job. It is not directly analogous to a job step in MVT, rather it relates the stages of job processing. Except for jobs specifying special JES3 processes (e.g., NJP), job segments are defined by JES3 to control the job's input, control statement conversion, device scheduling, execution, printing and punching, and job purge. AU JES3 support functions are implemented by dynamic support programs (DSPs). A support function is defined as any process performed by the Global 98 OS/VS2 Planning Guide for Release 2 processor as a part of a job in the system. Tasks such as card reading, printing, card punching, and servicing the Local and Global processors are examples of support functions. DSPs can be activated when explicitly requested by the operator or by nonstandard jobs. JES3 Job Flow All jobs are initially read into the Global processor by JES3, and are assigned a unique JES3 job number. Jobs can be submitted via a tape, disk, or card reader attached to the Global processor. In addition, jobs can be submitted to the global processor from remote terminals via remote job processing (RJP), TSO terminals, the internal reader, and from other JES3 systems via network job processing (NJP). The job flow (see figure 20) is as follows: Input Service Converter Service .......J------.L.L.-L.1-_...l'-J ------ Output Service Main Service Executing Job or System Task Output Processor Print! Punch Services Purge Service External Writer Figure 20. JES3 Job Flow 1. The job is read into the system and is initially placed on a spooling disk pack. An entry for the job is placed in the JES3 job queue. The support function that performs this task is called Input Service. Conversion Consl"eratlons 99 2. Immediately after input service, converter service is invoked. If a special catalogued procedure library is specified, it is opened and passed to the converter for JCL conversion. Converter service performs two functions: • Interprets and converts JCL into system control blocks. • Determines job setup requirements from generated system control blocks. During the first phase, JCL conversion and interpretration, syntax and logical errors are detected. This early J CL analysis allows cancellation of jobs with system or JES3 JCL errors before the job enters the setup and execution queues. During the second phase, the system control blocks are sc:anned and the setup requirements for each job are determined. As in the first phase, jobs with errors are scheduled for output service, bypassing execution. 3. Jobs with no JCL errors are passed to the device scheduling function for preexecution setup. Based on device requirements established by converter service, devices are selected. Any necessary volumes that require mounting are requested from the tape and DASD library console operators (See the section called Device Scheduling). 4. Based on device requirements, initialization specifications, job definition, and user exit, the job will be scheduled for processing on either the Global, a Local, or a Main processor. If the selected processor is Local, the Global processor will pass the job to the Main for processing. JES3 then signals the selected processor to begin executing the job. After processing is complete, control is passed to the Global processor for printing and punching. (Refer also to the section called Job Selection and Scheduling.) 5. Output data sets are queued for the external writer, the TSO OUTPUT command, or JES3 printers or punches. 6. Output data sets are printed by the JES3 print service function. Print service informs the operator of any special printer setup necessary. The printer(s) can be dther local or remote as defined by the job. 7. Output data sets are punched by the JES3 punch service function. A local or remote punch may be defined by the job. 8. A support function, purge, releases for use by other jobs, all DASD JES3 queue space associated with the completed job. At this time the SMF purge record is written for the job. In most cases, the standard job sequence described abovl~ will meet the needs of the application programmer; however, through the inclusion of JES3 control statements in the job's JCL, the job function sequence can be altered and special JES3 processing such as NJP can take place. The inclusion of JES3 processing control statements in the J CL causes the job to be classified as nonstandard. Device Scheduling Immediately after the interpretation of the JCL for a job, a table summary of all data set, volume, and device requirements is created. If a JOBCAT or STEPCAT DD statement specifies a special catalog, or if access to the system catalog determines a special catalog is needed, the catalog is dynamically allocated during the building of this table. The table is completed when all ]CL cross reference processing is complete and all volume and unit requirements are determined. The job is then enqueued for selection by the device scheduling function. 100 OS/VS2 Planning Guide for Release 2 The device scheduling function selects jobs for pre execution setup by priority, subject to workload parameters as specified by the installation. JES3 manages all references to tape devices and mountable direct access volumes. Jobs referencing permanently resident volumes will be automatically routed to the Local processor on which the volume resides. Reference to nonspecific direct access volumes are not handled by JES3; they are passed to the system for normal allocation processing. Jobs that require special data set volumes to be mounted are not submitted for execution until the appropriate units are available for mounting. Instructions are then issued to the operator to mount and/ or ready the required volumes. All volumes are assumed to reside in an installation tape and disk library. Messages are sent to these libraries requesting that all volumes required for preexecution setup by JES3 be forwarded to the tape and disk setup areas. This action is performed for each job requiring setup prior to requesting that the volumes be mounted. Jobs which do not require setup devices, or for which setup has already been effected, are executed while the setup process is being carried out for other jobs. JES3 waits for all operators involved to complete their required tasks before allowing the job to execute on a Local processor. If all devices allocated to a job are shared by Local processors, that job will be eligible for execution on any Local processor to which the devices are attached. When setup is complete, the job is released to an execution queue for processing. In addition, device scheduling provides: • Setup and system messages for direct access devices and tapes will be routed to the console located nearest each device. • Multiple jobs on different processors can share direct access volumes and data sets. The data sets can be shared providing the disposition is specified as shared. • Operator commands are available to determine the status of jobs, devices, and volumes in setup control. • JCL references to permanently resident direct access volumes require no operator action. • The number of devices setup can optionally be limited to the number needed at anyone time during the job. During deallocation the device will be returned to the device scheduling function for use by other jobs. Devices used for passed data sets can optionally be made available for other allocations. Job Selection and Scheduling Each processor in the JES3 complex requests work from the Global processor. Jobs are selected for each processing stage in priority order. Priority is initially established via the priority parameter on the JOB statement or with a JES3 control statement. The selection priority of a job can be changed by the operator, by priority aging, or by deadline scheduling. The priority aging feature allows JES3 to increase the priority of a job after it has been passed over by the JES3 an installation-specified number of times, either because of an insufficient number of devices or a low priority level relative Conversion Considerations 101 to that of other jobs in the system. Raising a job's priority will cause the job to be given preferential treatment in the selection of devices. The deadline scheduling feature allows the installation to specify a time of day by which the job should be scheduled. If the job is not scheduled by this time, JES3 will increase the priority of the job at installation-defined intervals until it is scheduled. Certain segments of a job can be designated by a JES3 control statement to be exeeuted in a remote processor. A job can be specified as nonstandard, executed in a remote processor, and its output returned to the submitting processor for printing or punching. The job can be executed in the submitting processor and printed in the remote processor. The user can control NJP processing or allow the operator to control it by making the job eligible for operator-controlled transmission to another processor. Using dependent job control (DJC), an installation can cause JES3 to control job selection based on dependencies among jobs. With JES3 control cards, the user can specify that one set of jobs (predecessor jobs) be eompleted ahead of other jobs (successor jobs). Jobs in a DJC network are held after JCL interpretation until predecessor jobs have completed execution or indicated to JES3 that critical processing is complete. This ensures that interjob catalog and device dependencies are resolved before setting up sucessor jobs. A device pool can be assigned to a set of jobs under DJC, assuring the availability of critical devices when needed. JES3 job scheduling provides flexibility in balancing the installation workload among the JES3 processors. When a system initiator becomes available for job processing, JES3 selects a job by priority from the available classes, and from among jobs requiring no setup volume mounting, or whose setup volumes have been mounted and verified. The installation can also specify that JES3 is to attempt to balance the CPU and I/O load by selecting a job whose CPU-to-I/O ratio complements those jobs currently running in the system. A CPU-to-I/O ratio is specified for each job via a JES3 control statement, or is a default by job class. The installation can also specify the number of initiators that can be started on each processor, and the maximum number of jobs from anyone class which can be simultaneously scheduled. JES3 Operational Environment The operational environment of the loosely coupled JES3 system is considerably different from that of other systems. All JES3 functions for all processors are controlled from consoles attached to the Global processor. The installation can allow system commands to be entered and system messages to be received at these JES3 consoles, reducing or removing the need for an operator to be stationed at system consoles on each separate processor. The implementation of multiple JES3 operator consoles, the separation of JES3 global functions, and the two-way operator communication with JES3 support functions, cause the loosely coupled complex to appear as a single system rather than composed of several separate and independently operated computer systems. JE83 provides an installation with flexibility in the location of machine-room equipment. By using additional operator consoles, JES3 installations can physically separate the operational functions (card input/output, printing, and tape setup), locating them in areas more convenient to the local work flow. The 102 OS/VS2 Planning Guide for Release 2 card reader/punches and the printers can be located in a job dispatching area where programs are submitted for execution and output is returned. The mountable input/output units can be placed in an area that is convenient to the tape and disk library. In addition, an operator console can be placed at the tape and disk librarian's desk to receive library volume requests. The central processing units can then be placed in some other area that is free of the congestion that often surrounds peripheral units. The Role of the JES3 Operator From the Global Processor, the global console operator manages the work flow of the JES3 complex and can modify specific jobs by changing their priorities. Other operators control the devices attached to the Global JES3 system (such as printers, card readers and card punches), mount and demount the Global and Local Processor mountable units, manage the disposition of required special volumes, and monitor the flow of jobs through the system. Each JES3 support function (processing stage) responds to a number of operator commands that permit the operator to cancel job processing, to restart processing, or to resume processing after an operator intervention request. Some functions, such as Print Service (at the print stage), provide additional options. For each printer, processing can be restarted either at the beginning of the data set or job, or at a checkpoint made within several pages of the current position. . the restart can be accomplished on the same printer or on a newly assigned printer. In addition, a request can be made to reload the Universal Character Set buffer of a local printer. Background programs can be invoked from an operator console as well as from the input JCL and JES3 control cards. The JES3 console inquiry function permits the global console operator to determine the status of the JES3 work queue, the status of a given job in the queue, the amount of space left in the JES3 queue, a summary of the workload backlog, and the status of the workload for any JES3 processing stage. This inquiry feature also permits the operator to determine whether the backlogs are adequate or too high, and to obtain an estimate of when a given job will be processed in the current queue. With this information, and with the ability to change job priorities or to place a job in hold status and later release it for processing, the operator can manage the system. He can also inquire into the status of functional components of the JES3 system, such as the number of buffers currently in use, or invoke a support function which will display the entire system status on a printer attached to the JES3 Global system. The Operator's Console: Operational control is maintained at every stage of processing for the following functions: • Deletion of a job. The operator can delete a job from the system by issuing the MODIFY with CANCEL message from the console. • Restart of job processing. The operator can restart applicable job segments on the Global Processor by issuing the RESTART message at a console. • Change of job priority. The operator can change the priority of a particular job. This option is usually used to expedite job processing for a particular job. • Holding and releasing of a job in queue. The operator can withhold a job from processing. Further, a job previously in hold status can be released to be scheduled for additional processing. Jobs may be held and released on the basis of priority or, if received from remote workstations, on an individual Conversion Considerations 103 terminal basis. Jobs in hold status from a particular terminal can be released entirely, or the operator can specify the number of jobs to be released. • Initiation of batch support functions: Using the CALL facility, the operator can request the system to schedule support functions (card-to-tape, tape··to-printer, etc.) to be executed concurrently with the standard preprocessing and postprocessing facilities. • Dynamic reconfiguration of Global processor input/output and terminal devices. The operator can remove, make available, or reroute the output of such devices as printers, card reader/punches, and Local processors. • System status inquiry. The operator can query the system regarding the job queue for functions such as estimated processor running time and estimated number of lines of print. In addition, the operator can request the status of an individual job, as well as a printout of the entire job queue status. Messages can be sent between JES3 consoles. Initialization, Control Statements, and Commands During initialization, JES3 configuration and processing options are defined in much the same manner as with ASP. This process is controlled by JES3 initialization control cards and allows the user flexibility in modifying JES3 operation. Initialization consists of: • Loading the resident portion of the JES3 system. • Loading optional modules. • Formatting the JES3 queue devices if necessary. • Allocating disk space for JES3 data sets. • Defining job scheduling algorithms. .• Defining installation standards. • Defining console message routing. • Defining the number of processors and the operating characteristics of each. • Establishing support processor device allocation tables. • Starting of procedures on selected Main processors. JES3 processes JCL statements in a manner similar to MVT ASP or VS2 Release 1 ASP. Minimal syntax change to existing J CL is required. JES3 control cards communicate special instructions to the system for forms contro] on the printer or card punch, the requirements for execution on a specific processor, and special scheduling requirements for job processing. JES3 / ASP Compatability Unlike ASP which is an optional extension of OS/360, JES3 is a componenet of MVS. 1ES3 functions totally replace SYS 1.SYSJOBQE and the OS readers and writers. Figure 21 relates some of the JES3 compatability items to the systems which support them. 104 OS/VS2 Planning Guide for Release 2 MVT/ASP 3.0 VS2 Release 1/ASP 3.1 VS2 Release 2/JES3 A single system job queue for multiple OS/MVT CPUs. A single system job queue for intermixed MVT and VS2 multiple CPUs. A single system job queue for intermixed MVT, VS2-1 , VS2-2 multiple CPUs. All SYSI N/SYSOUT data is sent across the CTC. All SYSIN/SYSOUT data is sent across the CTC. JES3 systems share the spool queue. The CTC is' used for control information only. MVT and VS2-1 SYSIN/ SYSOUT data is sent across the CTC. I I SYSOUT blocking is required in the Main processor. SYSOUT blocking is required in the Main processor. SYSOUT blocking is transparent to the VS2-2 user. It is required in MVT and VS2-1 processors. ASP 'hot-jobs' remain active in the event of Support Processor or ASP failure. ASP 'hot-jobs' remain active in the event of Support Processor or ASP failure. All jobs running in the failing system are lost but jobs in other systems are not. The Reader/I nterpreter function for all CPUs is executed in the Support Processor. The Reader/Interpreter function for all CPUs is executed in the Support Processor. The Converter and Interpreter functions are executed in the Global Processor for VS2-2, and the R/I function executed in the support processor for MVT and VS2-1. Main storage can be partitioned (fenced) for assignment to specific job groups. Main storage can be partitioned for MVT processors only. It is not applicable to VS. Main storage can be partitioned for MVT processors only. It is not applicable to VS. ASP Dynamic Dispatching. Available for MVT processors only. VS2 provides a dynamic dispatch ing function. Available for MVT processors only. VS2 provides a dynamic dispatching function. ASP job ENQ occurs at the volume level. ASP job ENQ occurs at the volume level. JES3 job ENQ occurs at the data set name level. Pre and post-SYSGEN modification is necessary. Pre and post-SYSGEN modification is necessary. There is a J ES3 generation process. No support. 3270 Light Pen support. 3270 Light Pen and Function Key support. Operator control via ASP Console Service. The CTC is the System Master Console. Operator control via ASP Console Service. The CTC is the System Master Console. Operator control is via JES3 Console Service. Messages and commands are optionally routed to JES3 Console Service via the CTC. No Checkpoint/Restart support. No Checkpoint/Restart support. Full Checkpoint/Restart support. ASP accounti ng. ASP accounting. ASP accounting is eliminated. Full SMF support is available. MCS route codes are ignored by ASP. MCS route codes are ignored by ASP. MCS route codes can be mapped onto JES3 route codes for VS2-2 processors (only). POLY ASP is supported. POL YASP is supported. POL Y ASP is not supported. Figure 21. JES3 Differences From MVT ASP and VS2 Release 1 ASP Conversion Considerations 105 SMF Considerations System management facilities (SMF) collect and record system and job information. The information can be used for job accounting and management information reports. Most of the changes to SMF records and exits for MVS have been made to support new system control program design. Because these changes could affect existing data reduction routines, all records and exits should be studied to determine what effect the changes might have. The following topics describe the changes that have been made to SMF for MVS. Differences from VS2 Release 1 System Management Facilities The most significant differences from VS2 Release 1 are: • SMF records contain additional accounting information to record new system environmental characteristics. • New SMF exits from the control program are provided. • SMF supports an output limiting facility. • The SMF data sets, SYS I.MANX and SYS.MANY, must be cataloged. • The SMFDEFLT list in SYS l.P ARMLIB has been changed. Multiple SMF SYS l.P ARM LIB members may be maintained. Their names are of the form SMFI)RMxx. Some of the parameters in these members are also different. See the topic "IBM-Created Lists" in the chapter "Defining the System" for a more detailed description. • The type 20 Job Initiation record is now considered a job record and no longer controlled by the DSV option. • The same SYS l.P ARM LIB member is applicable to background and foreground use. Differences from MVT System Management Facilities In addition to the preceding, the most significant differences from MVT are: • SMF is a standard facility of MVS. • SMF does not support tape for recording SMF data. SMF Record Differences for MVS Figure 22 indicates how the SMF records have changed for ~[VS. The following topics explain why some of the more significant changes were made. 106 OS/VS2 Planning Guide for Release 2 Record Type Differences Between VS2 Release 2 and MVT Differences Between VS2 Release 2 and MFT Virtual storage statistics added Virtual storage statistics added ****** **** .. * 1 --- --- --- --- 2 ****** ** .. * .. * ****** *** ... * .. 3 * ... **** *****'* ****** ****** SYSIN/SYSOUT EXCP counts eliminated Virtual storage statistics added VIO statistics added Paging statistics added CPU time statistics separated into two fields System resource manager statistics added CPU time statistics separated into two fields Virtual storage statistics added VIO statistics added Paging statistics added CPU time statistics separated into two fields SYSIN/SYSOUT EXCP counts eliminated Virtual storage statistics added VIO statistics added Paging statistics added CPU ti me statistics separated into two fields - System resources manager statistics added CPU time statistics separated into two fields System resources manager statistics added CPU time statistics separated into two fields SYSIN/SYSOUT EXCP counts eliminated Virtual storage statistics added VIO statistics added Pagi ng statistics added CPU time statistics separated into two fields System resources manager statistics added CPU time statistics separated into two fields 6 Job entry subsystem change Job entry subsystem change Job entry subsystem change Job entry subsystem change 7 ****** ****** **-5-*** ****** 8 ****** ****** ****** ****** 5 9 10 ****** ****** ****** ****** 11 ****** ****** 12 --- --- --- MP statistics eliminated ****** ****** ******' * ... * ... ** MP statistics eliminated --- --- --- 13 14 VIO statistics added VIO statistics added VIO statistics added VIO statistics added 15 VIO statistics added VIO statistics added VIO statistics added VIO statistics added N/A N/A 17 ****** **** .. * ****** *** .. ** 18 ****** *****'* ****** * .. **** 19 ****** ****** *** ... ** * .. **** 20 ****** ****** ****** 21 ****** * ... **** *** .. ** ****** *** ...... * 22 New record for MP statistics New record for MP statistics New record for MP statistics New record for MP statistics 26 New record for job entry subsystem statistics New record for job entry subsystem statistics New record for job entry subsystem statistics New record for job entry subsystem statistics 30 31 S Differences Between VS2 Release 2 and VS1 0 4 - Differences Between VS2 Release 2 and VS2 Release 1 32 33 --This record is now written when the MODI FY TCAM command with the TS option is issued. N/A New record for time sharing statistics --This record is now written when the MODIFY TCAM command with the TS option is issued --- N/A --- --- N/A --- Figure 22. Changes to SMF Records for VS2 Release 2 (Part 1 of 2) N/A New record for time sharing statistics I I I N/A N/A Record Type 34 35 Differences Between. VS2 Release 2 and VS2 Release 1 Virtual storage statistics added via statistics added Paging statistics added CPU time statistics separated into two fields New record for time sharing statistics CPU time statistics separated into two fields System resources manager statistics added CPU time statistics separated into two fields New record for time sharing statistics CPU time statistics separated into two fields 38 N/A 40 ****** 41 --- N/A 42 --- N/A 43 44 Differences Between VS2 Release 2 and MVT Differences Between VS2 Release 2 and VS 1 Virtual storage -statistics added Paging statistics added CPU time statistics separated into two fields System resources manager statistics added CPU ti me statistics separated into two fields --- New record for dynamic allocation statistics ****** I New record for time sharing statistics CPU time statistics separated into two fields I I --- statistics N/A New record for job entry subsystem statistics --- N/A New record for dynamic a!!ocation N/A --- Job entry subsystem change N/A New record for time sharing statistics CPU time statistics separated into two fields via statistics added N/A New record for job entry subsystem statistics Differences Between VS2 Release 2 and MFT New record for job entry subsystem statistics N/A N/A 45 New record for job entry subsystem statistics New record for job entry subsystem statistics New record for job entry subsystem statistics I New record for job entry subsystem statistics 47 New record for job entry subsystem statistics Job entry subsystem change New record for job entry subsystem statistics I New record for job entry subsystem statistics 48 New record for job entry subsystem statistics Job entry subsystem change New record for job entry subsystem statistics 49 New record for job entry subsystem statistics New record for job entry subsystem statistics New record for job entry subsystem statistics I New record for job entry subsystem statistics New record for job entry subsystem statistics 62 ****** ****** New record for VSAM statistics New record for VSAM statistics 63 ****** New record for VSAM statistics ****** New record for VSAM statistics New record for VSAM statistics 67 ****** ****** ****** ****** New record for VSAM statistics 64 New record for VSAM statistics New record for VSAM statistics 68 ****** ****** New record for VSAM statistics New record for VSAM statistics ****** New record for VSAM statistics New record for VSAM statistics 70 New record for MF 11 New record for MF/1 New record for MF 11 New record for MF 11 71 New record for MF/l New record for MF 11 New record for MF/1 New record for MF 11 72 New record for MF 11 New record for MF /1 New record for MF /1 New record for MF /1 73 New record for MF/1 New record for MF/1 New record for MF/1 New record for MF 11 74 New record for MF 11 New record for MF/1 New record for MF/1 New record for MF /1 69 --- ****** 0 •• ecord nOl+ suppv. led ~r+ ****** ....,. sta"sdCS +i +" Sall.e Figure 22. Changes to SMF Records for VS2 Release 2 (Part 2 of 2) N.I A ,,,. No. appl ....able + New System Resources Manager Data SMF record types 4, 5, 34, and 35 (step termination, job termination, time-sharing step termination and LOGOFF) provide new information obtained from system resources management routines. Fields in record types 4 and 34 are now used to record step termination data such as VIO page-ins, VIO page-outs, total page-ins, and total page-outs. Fields in record types 5 and 35 are used to record job termination data such as: • Total number of service units used by a background job (type 5) or a terminal job (type 35). • The total amount of time that the background job was active (type 5). • The total amount of time that all transactions were active (type 35). • The total number of transactions. This information can be used to approximate TSO terminal response time for a session, and for adjusting IPS values in an attempt to obtain desired response time. (See the chapter "Directing the Use of System Resources" for a description of an installation performance specification (IPS) and service units.) SMF Recording Replaced by MF / t Two records, type 1 (wait time) and type 12 (end-of-day), are no longer supported because the system activity measurement facility (MF /1) now records the data in place of SMF. The information from record types 1 and 12 is now in record types 70 (CPU activity) and 71 (paging activity), although record types 70 and 71 are formatted differently from record types 1 and 12. Furthermore, MF /1 records are only written to the SMF data set when MF /1 is operationa1. Therefore, an installation that wants to use the information that was in record types 1 and 12 should rewrite data reduction routines to recognize the new records and process the new format. MF /1 processing must also be started so that the data will be gathered and recorded. Three other new records (types 72, 73, and 74) have been created to record MF /1 data. See the topic "Tracing" in the chapter "Directingthe Use of System Resources" for a more detailed description of their contents. New Storage Data Three fields in both record types 4 and 34 now provide new step level accounting information such as: • Either the size of the V=R address area specified on an ADDRSPC=REAL request, or the total size of the user's private address space for an ADDRSPC=VIRT request. • The amount of the user's private address space (assigned from the high address) used for the local system queue area (LSQA), the scheduler work area (SWA), and subpools 229 and 230. • The amount of the user's private address space (assigned from the low address) used for the user's programs. Conversion Considerations 109 Job Entry Subsystem Statistics Some SMF records have been modified and others created to record statistics provided by the job entry subsystem. They are: • The type 6 (output) record contains new data such as data set control indicators, JES-assigned job numbers, output route code, and data format error indicators. This record will be produced by the job entry subsystem and external writer which handles all system output processing. • The type 26 (job summary) record is produced when a job is ready to be purged from the system. It contains information about subsystem processing such as the type of job and how it entered the system, p:riorities for job selection and output processing, counts of lines or puched cards generated, and times at which the various phases of processing the job began and ended. • Record types 43, 45, 47, 48, and 49 record information about subsystem activity. Included is data regarding the starting or stopping of a subsystem or a subsystem event (e.g., SIGNON). Record type 49 records the attempt to sign on or start a line with an invalid password. Multiprocessing Data The type 22 record is new in MVS. It indicates the status of resources unique to a multiprocessing configuration. It serves the same function for CPUs, channels, and storage as record types 8, 9, 10, and 11 do for I/O devices. Virtual Storage Access Method Data The virtual storage access method (VSAM) generates new records (types 64, 67, and 68) that contain information similar to existing data management SMF records. VSAM also records statistics when a VSAM data set is opened, allocated, and freed (record types 62, 63, and 69). Note: Record types 63 and 67 are generated for VSAM recovery processing. Time Sharing Record Elimination The following record types are not supported in MVS: • 30 -- Start TS • 32-- Driver • 33 -- Driver Modify • 41-- Modify TS • 42 -- Stop TS These records were associated with starting and stopping TSO and with the TSO driver. These functions have been replaced by the MODIFY TCAM command and the system resources manager in MVS. Changes to SMF Exits for MVS Figure 23 indicates the changes that have been made to SMF exits in other systems for MVS. 110 OS/VS2 Planning Guide for Relea!ie 2 Exit IEFUJV Difference Between VS2 Release 2 and VS2 Release 1 Difference Between VS2 Release 2 and VS1 Difference Between VS2 Rele.~se 2 and MVT Difference Between VS2 Release 2 and MFT This exit is also taken after the internal text is created This exit is also taken after the internal text is created This exit is also taken after the internal text is created This exit is also taken after the internal text is created IEFUIV N/A --- N/A N/A IEFUJI ****** ****** ****** ****** IEFUSI ****** ****** ****** ****** ****** ****** ****** IEFUSO New exit for output limiting facility IEFACTRT Counts are in terms of logical records rather than physical blocks Counts are in terms of logical records rather than physical blocks Counts are in terms of logical records rather than physical blocks Counts are in terms of logical records rather than physical blocks IEFUTL This exit can now be taken from the foreground This exit can now be taken from the foreground This exit can now be taken from the foreground This exit can now be taken from the !oreground There is a new interface to extend the time limit in micro seconds There is a new interface to extend the time limit in micro seconds There is a new interface to extend the time limit in micro seconds There is a new interface to extend the time limit in micro seconds For wait limits there is no longer an extension. A new time limit is set for the remainder of the step. For wait limits there is no longer an extension. A new time limit is set for the remai nder of the step. For wait limits there is no longer an extension. A new time limit is set for the reaminder of the step. For wait limits there is no longer an extension. A new time limit is set for the remainder of the step. New exit for selecting records to be written New exit for selecting records to be written New exit for selecting records to be written Change in exit parameters New exit receives control when job is ready to be purged from the system New exit receives control when job is ready to be purged from the system IEFU83 IEFUJP --****** N/A ****** New exit receives control when job is ready to be purged from the system Exit not supported Unchanged Not applicable Figure 23. Changes to SMF Exits for VS2 Release 2 Conversion Considerations ttt Job Control Language Job control language statements contain information that is needed to control the processing of jobs. This section describes JCL facilities, parameters, and subparameters that are either new or have changed from both VS2 Release 1 to MVS and MVT to MVS. Figure 24 summarizes the more significant JCL changes. 112 OS/VS2 Planning Guide for Release 2 Statement JOB Parameter accounting information field Change C!) ADDRSPC Assigns pageable or nonpageable storage space, and defaults for pageable storage. CLASS (A-Z,0-9). The job entry subsystem enqueues the job on the associated class queue. PERFORM Associates jobs with performance group definitions. 'PRTY This parameter is ignored when the JES2 job entry subsystem is used. Priority values can be specified on a subsystem control card. REGION ROLL Limits the size of variable GETMAIN requests. • This parameter is ignored. • Assigns pageable or nonpageable storage space, and defaults for pageable storage. TYPRUN EXEC ADDRSPC If TYPRUN=SCAN, the JCL is converted but does not execute. DPRTY The default is the value of the APG priority. DYNAMNBR Limits the number of data sets that a job step may hold for reuse. PERFORM Associates job steps with performance group definitions. REGION ROLL DO In addition to standard'OS usage, this field can be used for JES2 parameters. Limits the size of variable GETMAIN requests. • This parameter is ignored; the job is not failed. AFF This parameter is ignored; the job is not failed. AMP Completes information in an access method control block (ACB) for the virtual storage access method (VSAM) data sets. COPIES Specifies multiple copies of an output data set. DCB DEN =4 subparameter indicates 6250 bits-per-inch for 9-track tape. FR 10 subparameter specifies the SYS1.IMAG ELIB member to be used to interpret documents read by a 3886 OCR device. • DEST Allows you to route output to specified destinations. DYNAM Total is added to DYNAMNBR value, if any. FREE Indicates when a data set is to be deallocated. JOBCAT Defines a private user catalog. OUTLIM With SYSOUT=, this parameter will specify the maximum number of logical records (for example, print lines) that may be written into this SYSOUT data set. An SMF exit routine is entered when ~his number is exceeded. RETAIN SEP SPACE SPLIT STEPCAT SUBALLOC SYSOUT=(c,p,f) • o o HIARCHY subparameter is ignored; the job is not failed. C!) (!) (!) (!) (!) (!) ® This parameter is ignored; the job is not failed. This parameter is ignored; the job is not failed. Has a new default value for VIO data sets. Converted to an equivalent SPACE parameter. Defines a private VSAM user catalog. Converted to an equivalent SPACE parameter. c - (A-Z, 0-9) (* is valid in JES2 only.) This output class designation is used to queue the associated output.-l)a-ta-isdequeued in much the same manner as the MVT output writer. p - (p=INTRDR). This data set is sent directly to input service for processing as an input job stream. p - (p +- INTRDR). This data set is enqueued separately for the named writer. f - Data is enqueued separately with this alphameric forms designation. UNIT The SEP subparameter is ignored, the job is not failed. VOLUME The RETAIN subparameter is ignored; the job is not failed. New ,or changed from MVT New or changed from VS2 Release 1 New or changed from MVT and VS2 Release 1 Figure 24. JCL Changes for VS2 Release 2 Conversion Considerations t 13 Changes to VS2 Release 1 JCL The following topics describe the differences from VS2 Release 1 JCL. Perfonnance Group Specification The PERFORM parameter is new for MVS. It is used to assoeiate a job or job step with any of the several performance groups specified by installation-supplied performance group definitions. Each performance group definition specifies a rate for providing system resources to terminal jobs or batch jobs or job steps under varying conditions (see the chapter: "System Resources Management") . • When PERFORM=groupnr is coded on a JOB statement, that job receives system resources at the rate defined by the performance group identified by 'groupnr' (a number between 1 and 255). When PERFORM=groupnr is coded on an EXEC statement, the group number value applies only to the job step specified on the EXEC statement. • When PERFORM.procstep=groupnr is coded on an EXEC statement, the group number value applies to the specified step within the procedure invoked by the EXEC statement. If the PERFORM parameter is coded on the JOB statement, its value overrides any value for PERFORM parameters that may be specified on the EXEC statements for that job. If a PERFORM parameter is specified for a procedure, its value overrides any value for PERFORM parameters that may be specified on EXEC statements within the procedure. If the PERFORM parameter is not coded, the default is 1 for background jobs and 2 for foreground users. Therefore if the previous job control language for a job is not changed, the job will be associated with one of only two possible performance group definitions. Assigning Job Selection Priority The PRTY parameter of the JOB statement is ignored in MVS with the JES2 job entry subsystem. Instead, job selection priority must be specified on a JES2 control card. The PRTY parameter is valid with JES3. Specifying Data Sets in Paging Space (Virtual I/O Data Sets) A new Stage I system generation parameter, VIO, may be specified on the UNITNAME macro instruction to identify unit names that may be used to specify data sets that are to receive virtual I/O (VIO) processing. Any single volume temporary data set defined by a DD statement with a unit name for which VIO= YES was coded at system generation time, is proeessed as a virtual I/O data set. The DD statement that defines the temporary data set must meet the following requirements if the temporary data set is to be allocated to the system's paging space: • The DSNAME parameter must not be specified unless an & name value is supplied for it. (Data sets with non-system generated names are not allocated to paging space). • The data set disposition must be new or passed. 114 OS/VS2 Planning Guide for Release J~ • The UNIT parameter must specify a unit name that indicates system paging space. (VIO= YES must have been specified for that unit name at system generation time.) • Indexed sequential data set organization must not be specified on the DSORG subparameter of the DCB parameter. • No volume serial number may be specified. If the SPACE parameter is not coded for virtual I/O data sets, the default value is 10 primary and 50 secondary b~cks with an average block length of 1000. Requesting Data Set Freeing at Data Set Close Time The FREE parameter of the DD statement is new in MVS. It is used to specify that a data set is to be freed when that data set is closed. If FREE=END is specified, the data set is freed at the end of the job step. If FREE=CLOSE is specified, the data set is freed when the CLOSE macro instruction is issued for it. If the FREE parameter is not coded, the default value is END. Therefore, if the current job control language for a job is not changed, all data sets will be released at the end of the job step. If FREE=CLOSE is specified for a SYSOUT data set, the data set will be sent to the job entry subsystem when the CLOSE macro instruction is issued. The data set is eligible for immediate printout (spinoff). Note: FREE=CLOSE should not be specified for a data set that is opened and closed more than once during a job step. Specifying the Number of Resources That May Be Held for Reuse The DYNAMNBR parameter of the EXEC statement is new in MVS. While it extends (dynamic allocation) and replaces the function of the DD DYNAM statement in MVT and VS2 Release 1, it does not specify a value that limits the number of dynamic allocations allowed to a terminal job. Rather, it specifies a value that is used to control the number of not-in-use data sets that may be allocated to a job step. The DYNAMNBR parameter value is added to the number of DD statements in the job step to determine the number of data sets with the not-in-use attribute that a job step may hold. The value of the DYNAMNBR parameter must be within the range of 0 to 1,635. If the DYNAMNBR parameter is not coded, the default value is O. Requesting Multiple Copies of an Output Data Set The COPIES parameter of the DD statement is new in MVS. It specifies the number of copies of an output data set that are to be produced by the job entry subsystem. The COPIES parameter may be coded only when the SYSOUT parameter is included on the DD statement. Directing the Destination of an Output Data Set The DEST parameter of the DD statement is new in MVS. It specifies that an output data set is to be routed to a remote station or local device by the job entry subsystem. The DEST parameter may be coded only when the SYSOUT parameter is included on the DD statement. Conversion Considerations 115 Support for Tape and Direct Access Devices If INTRDR is not coded for the SYSOUT parameter of the DD statement, the data set is enqueued separately for the named writer. This writer is not invoked by the job entry subsystem -- the operator must start an exte:rnal writer to attach any user-written writer. The main purpose of the external writer is to support output devices not supported by the job entry subsystem, specifically tape and direct access devices. The external writer uses the QSAM access method and any devices supported by QSAM. To separate output destined for an external writer from output processed by the job entry subsystem, it is advisable to place it in separate classes not assigned to job entry subsystem devices. Format Record Identification for thE~ 3886 OCR Device A new subparameter (FRID) of the DCB parameter of the DD statement is used to specify the type of document that is supplied as input to a 3886 optical character recognition (OCR) device. The FRID subparameter indicates the member of the SYS1.lMAGELIB data set that is to be used for interpreting the document to be read by the 3886 OCR device. Output Limit Specification The OUTLIM parameter on the DD statement is used to specify the limit for the number of logical records to be written to an output data set. It is coded the same way as the MVT OUTLIM parameter. Increased Data Set Definition Capability In MVS, as many as 1,635 DD statements may be specified for each job step. In both MVT and VS2 Release 1, the limit was 255 DD statements per step. Increased Volume Sharing Capability The following additional types of volumes may be shared between jobs and job steps in MVS: • Private direct access volumes mounted in response to nonspecific requests . • Direct access volumes for which deferred mounting has be:en requested. Interpretation of the REGION Parameter In MVS, the address space available to an individual job (user private address space) is restricted only by the size of the individual virtual address space (16 megabytes) minus the space required for system control program use. Therefore, for jobs that specify ADDRSPC=VIRT (requesting V=V storage), the REGION paramet.er is not used to specify the size of the address spac(~ available to the individual job. It is, however, used to specify the maximum value for use by the GETMAIN service routines in servicing variable-length GETMAIN requests. When a variable-length GETMAIN is issued that would cause the address space allocated to a user to exceed the value of the REGION parameter, the user is limited to the REGION parameter value. For jobs that specify ADDRSPC=REAL (requesting V=R storage), the value specified is used to limit the size of the address space available to the individual job. 116 OS/VS2 Planning Guide for Release 2 Job Step Dispatching Priority Default Value In MVS, the value of the APG priority will be used as the default dispatching priority for a step when the DPRTY parameter is not specified on the EXEC statement. In MVT and VS2 Release 1, the default dispatching priority for the step was the job's dispatching priority. Changes to MVT JCL In addition to the differences described above, two JCL parameters are new to current MVT users. Requesting Storage for Program Execution The ADDRSPC parameter is new in VS. It is used to assign page able or nonpageable storage space to a job or a job step. It may be specified as either ADDRSPC=VIRT or ADDRSPC=REAL: • If ADDRSPC= VIRT is specified, the job is assigned to pageable (V = V) storage. • If ADDRSPC=REAL is specified, the job is assigned to nonpageable (V =R) storage . • If the ADDRSPC parameter is not coded, the default value is VIRT. Therefore, if the current job control language for a job is not changed, the job will be executed in page able storage. If the job must be executed in nonpageable storage, ADDRSPC=REAL must be specified. If the ADDRSPC parameter is coded on the JOB statement, its value overrides any value in the ADDRSPC parameters that may be specified in the EXEC statements for that job. Specifying the BTAM Online Terminal Test Option In MVS, a new value can be specified for the EROPT subparameter of the DCB parameter of the DD statement. If EROPT=T is coded, it indicates a request for the basic telecommunication access method (BTAM) online terminal test option. JES2 JCL Considerations The JES2 job entry subsystem processes JCL statements in a manner similar to MVT /HASP and VS2 Release l/HASP. The JCL considerations for JES2 therefore fall into two categories, depending upon whether or not an installation currently has a HASP system. In either case, however, no syntax changes to existing J CL will be required. JES2 JCL for Non-HASP Users Two significant points for installations that are unfamiliar with HASP are: • Extended JCL facilities are provided by JES2 control cards. • The cataloged procedure to be used for job conversion can no longer be associated with an input stream but rather with a job class. JES2 Control Cards: The JES2 control cards are identified by a / * in columns 1 and 2 and a JES2 control verb beginning in column 3. They are used to: • Supply information for the JES2 separator page and card. Conversion Considerations 117 • Identify a job selection priority (for queuing the job). • Supply information for calculating a job selection priority if such a priority is not explicitly stated. • Identify volumes required by the job, so that the job is held until the required volumes become available. • Route system output data sets to the local printer or punclh. • Send a message to the central computer operator. • Specify the DDNAME of the cataloged procedure library to be used for conversion of the job's JCL. • Specify extended output processing setup options to be associated with data set groups. Assigning Job Selection Priority: A JES2 control card (priority) is used to assign a selection priority to a job. If this control card is not included, the job entry subsystem assigns a priority determined either by the estimated execution time and estimated number of lines of output (values contained on another JES2 control card) or by the JES2 generation default values. If the PRTY parameter is coded on a JOB statement, it will be ignored. JES2 JCL for HASP Users The two most significant points are: • In addition to the job copies and job routing facilities, it is now possible to specify data set copies and data set routing. • Two new control cards have been created. /*JOBJ>ARM Control Card: The /*JOBPARM control card was new in HASP for VS2 Release 1 and is coded similarly for JES2. It specifit~s job-related parameters that are used by the JES2 job entry subsystem to: • Establish the default job selection priority. • Control the job's processing. • Build the JES2 separator page and card. These parameters extend the information that previously could be specified only on a JOB statement accounting information parameter by adding information such as the name of the cataloged procedure library to be used for JCL conversion. /*OUTPUT Control Card: The /*OUTPUT control card was new in HASP for VS2 Release 1, and is coded similarly for JES2. It specifies c;haracteristics and/or options for groups of output data sets. It is used in conjunction with the form number subparameter in the SYSOUT parameter of the DD statement. Any system output set whose DD statement form subparameter matches the FORMS code specified on as /*OUTPUT control card, will be processed according to the specifications of the /*OUTPUT control card. 118 OS/VS2 Planning Guide for ReleasL~m initialization parameter 33 maximum number specified 36 dump default lists IEAABDxx 42 IEADMPxx 42 overriding defaults with the CHNGDUMP command 128,.125 DUMP system initialization parameter 33,31 difference from VS2 Release 1 36 DYNAM job control language parameter 113,115 dynamic address translation (DAT) definition 156 dynamic allocation validation exit 87 dynamic device reconfiguration (DDR) 150-151 dynamic support program (DSP) 98 dynamic support system (DSS) definition 156 NUCMAP system initialization parameter 34 system data set SYS1.DSSVM 47,49,50 dynamic system interchange (DSI) 96 DYNAMNBR job control language parameter 115,113 EC (extended control) mode definition 156 EDIT command 136 EDIT system generation macro instruction 23,27 EDITOR system generation macro instruction 29,27 emulator 151 specified by AFFINITY macro instruction 22 EMULATOR system generation macro instruction 27 enabled page fault definition 156 END subcommand 136 ENQ macro instruction ensuring serial access for mUltiprocessing 151 major name restrictions for authorized programs 83 ENQ/DEQ algorithm 56 enqueue residence value (ERV) 67,56 EOD operand 125 EROPT JCL subparameter of DCB parameter 117,113 error recording data set (SYS I.LOGREC) 47 change from MVT and VS2 Release 1 49 ERV (enqueue residence value) 67,56 EXCP addressed to SYSIN/SYSOUT DCB 15,16 EXEC command 136 EXEC statement 113 execution service 92 execution task monitor 92 EXPORT verb 142 extended control (EC) mode 156 external page storage definition 156 MF/l counts 71,70 external page table (XPT) definition 156 external writer 94,116 fetch protection 81,87 FIB (foreground initiated background) commands 138 file protection attributes 86 FIX system initialization parameter 33,31 fixed definition 156 fixed BLDL table 153 definition 156 system initialization parameter specification 32 153 fixed link pack area (LPA) definition 156 specifying routines with the FIX parameter 33 fixed LP A list (IEAFIXxx) 42 fixed page 147 definition 156 foreground initiated background (FIB) commands 138 FORTLIB system generation macro instruction 27 FORTRAN G 17 FORTRAN H 17 FORTRAN system generation macro instruction 27 fragmentation definition 156 virtual storage 13 FREE command 136 job entry subsystem support 138 FREE job control language parameter 115,113 freeing data sets 115 FRID job control language subparameter of DCB parameter 116,113 Full ANS COBOL V2 17 functions not supported MVT 15 VS2 Release 1 16 GAM (graphics access method) definition 156 system generation specification 23 GOG (generation data group) 142 127 generalized trace facility (GTF) replacement for trace and TSO trace 16 TRACE system initialization parameter 37 GENERATE system generation macro instruction 23,27 142 generation data group (GDG) GENTSO system generation macro instruction 27 GETMAIN, restriction for variable-length 116 GJOBCTL system generation macro instruction 27 GJP (graphic job processor) 15 global definition 156 global lock definition 156 global processor 96 global services definition 156 GRAPH MF/l parameter 69 graphic job processor (GJP) 15 graphics access method (GAM) definition 156 system generation specification 23 GRAPHICS system generation macro instruction 29,27 GTF (generalized trace facility) 127 replacement for trace and TSO trace 16 TRACE system initialization parameter 37 HALT command 125,127 hard copy log 33 HARDCPY system initialization parameter 33,27 hardware error recovery management system definition 156 HASP 156 (see also JES2) HELP system generation macro instruction 27 HIARCHY job control language subparameter of DCB parameter 113 hierarchy support unsupported MVT function 15,38 highlights of VS2 Release 2 12 HOLD command 125 HOLD operand 126 HRAMsystem initialization parameter 38 HSVC system initialization parameter 38 IBM-created SYS1.PARMLIB lists 41-44 IEAABDxx (dump default list) description 42 overriding defaults with the CHNGDUMP command 128,125 IEAAPFxx (authorized program library list) 42 IEAAPPOO (authorized appendage list) 41 IEABLDxx (resident BLDL list) 43 IEADMPxx (dump default list) description 42 overriding defaults with the CHNGDUMP command 128,125 IEAFIXxx (fixed LP A list) 42 IEAIPSxx (see installation performance specification) IEALODOO (LP A directory load list) 43 IEALP Axx (modified link pack area list) 45 IEAOPTxx (system resources management tuning parameter list) 67,45-46 Index 163 lEAP AKOO (LP A packing list) 43 IEASYSxx (system parameter list) description 44 recommendations for values 39 IEBUPDAT utility program 15 IEBUPDTE utility program 15 IEFACTRT SMF exit 111 IEFUIV SMF exit 111 IEFUJI SMF exit 111 IEFUJP SMF exit 111 IEFUJV SMF exit 111 IEFUSI SMF exit 111 IEFUSO SMF exit 111 IEFUTL SMF exit 111 IEFU83 SMF exit 111 IEHIOSUP utility program 15 IEHLIST utility program 143,15,16 IEHMOVE utility program 143,15,16 IEHPROGM utility program 143,28,15,16 IEHUCAT utility program 143 IMAGELIB system generation macro instruction 27 IMBMDMAP program 15 IMCOSJQD service aid 16 IMPORT verb 142 IN operand 125 indexed sequential access method (ISAM) 23 INIT operand DISPLAY command 125 TRACK command 127 initial program loader (IPL) 30 initialization, master scheduler 30 initialization, new system data sets 28 initialization, system 30-46 parameters 30-39 input service 91 installation-created SYS l.P ARM LIB lists 45-46 installation performance specification (IPS) 58 definition 156 IEAIPSxx list 65-67,42 parameters 65-67 tuning example 73-77 109 using SMF data to adjust installation verification procedure (IVP) 20,21 integrity, system 79-87 areas of concern 81-85 definition 79 installation responsibility 80 procedural responsibilities 86-87 interaction 156 internal reader 91 definition 156 internal writer definition 156 interprocessor communication (IPC) definition 156 120 interregion communication INTERV AL MF /1 parameter 69 interval service value (ISV) 66 definition 156 10C resource factor coefficient 67 10C service definition coefficient 67,57 IOCONTRL system generation macro instruction 27 IODEVICE system generation macro instruction 23,27 I/O appendage integrity 87 I/O device activity measurement 68 I/O device activity report 72 tuning example 73 I/O load balancing algorithm 54 weighted by resource factor coefficients 67,56 I/O measure 57 I/O, Virtual 51-52 IPC (interprocessor communication) definition 156 IPL (initial program loader) 30 IPS (see installation performance specific:ation) 58 IPS operand 126 IPS system initialization parameter 33,31 IQAORDER list 40 IRBMFlxx (MF/1 parameter list) 43 ISAM (indexed sequential access method) 23 ISV (interval service value) 66 164 OS/VS2 Planning Guide for Release 2 IVP (installation verification procedure) 20,21 JCL (job cont~ollanguage) 112-118 (see also parameters, JCL) changes to MVT 117,113 changes to VS2 Release 1 114-117 for JES2 117-118 summary of changes 113 JCL validation exit 87 JES2 definition 156 differences from MVT /HASP and VS2 Release I/HASP 93-95 generation 24-25 highlights 12 input service 91 JCL 117-118 job flow 91 operator commands 124-131 SMF record support no time sharing commands 138 JES2 control cards 117-118 JES2 generation 24-25 JES3 96-105 commands 132-133,130 comparison with ASP 104-105 definition 156 generation 25,24 highlights 12 JCL 118 job flow 99 mUltiprocessing 152 job class, JES2 25 job control language (JCL) 112-118 changes to MVT 117,113 changes to VS2 Release 1 114-116 for JES2 117-118 summary of changes 113 job entry subsystem (see JES2, JES3) data set 47,49 generation 20-21,23-25 highlights 12 job initiation record 106 job journal 93 job purge SMF exit 111 job queue data set (SYSl.SYSJOBQE) 49 unsupported MVT function 16 unsupported VS2 Release 1 function 16 job segment 98 job selection priority 117 JOB statement 113 job step dispatching priority default value 117 job summary record (SMF) 110 JOBCAT catalog 141 JOB CLASS operand 125 jobname operand 125-126 JOBPARM control card 118 JOBS operand DISPLAY command 125 TRACK command 127 journal, job 93 link library (SYS I.LINKLIB) 47 adding installation-written members 29 link library list (LNKLSTxx) 43 link pack area (LP A) defined 157 fixed LPA 153 modified LPA 153 pageable LPA 153 link pack area library (SYSl.LPALIB) 47 adding installation-written members 29 LINKLIB system generation macro instruction LISTCAT verb 142 LNK system initialization parameter 33,31 LNKLSTxx (link library list) 43 load balancing 54 LOADER system generation macro instruction local lock definition 157 29,27 29,27 local main processor 96 local processor 96 local service definition 157 local system queue area (LSQA) 153 definition 157 local time constant (P ARMTZ) 45 locking technique 83,149 log data sets 50 specification of output class with LOGCLS 33 specification of WTL limit with LOGLMT 33 LOGCLS system initialization parameter 33,27 LOGLMT system initialization parameter 33,27 LOGOFF command 136 LOGON command 136 system resources manager support 134 152,12 loosely coupled mUltiprocessing (MP) LP A Oink pack area) definition 157 fixed LPA 153 modified LPA 153 pageable LPA 153 LPA directory load list (IEALODOO) 43 LPA packing list (IEAPAKOO) 43 LSQA (local system queue area) 147 definition 157 LSQA operand MOUNT command 128,126 START command 128,126 LSQACEL system initialization parameter 37 M operand 125 MACLIB system generation macro instruction macro instructions, system generation AFFINITY 22,27 ALGLIB 27 ALGOL 27 ASSEMBLER 27 CENPROCS 29,27 CHANNEL 22,27 CHECKER 27 CKPTREST 23,27 COBLIB 27 COBOL 27 CONSOLE 23,27 CTRLPROG 23,27 DAT AMGT 23,27 DATASET 23,27 DCMLIB 27 EDIT 23,27 EDITOR 29,27 EMULATOR 27 FORTLIB 27 FORTRAN 27 GENERATE 23,27 GENTSO 27 GJOBCTL 27 GRAPHICS 29,27 HELP 27 IMAGELIB 27 IOCONTRL 27 IODEVICE 23,27 LINKLIB 29,27 LOADER 29,27 MACLIB 29,27 OUTPUT 27 PAGE 29,27 PARMLIB 29,27 PLI 27 PLlLIB 27 PTOP 27 RESMODS 29,27 RPG 27 SECONSLE 29,27 SCHEDULR 23,27 SORTLIB 27 SORTMERG 27 SUPRVSOR 27 SVCTABLE 23,27 29,27 SYSUTILS 27 TELCMLIB 27 TSO 23,27 TSOPTION 27 UCS 29,27 UNITNAME 23,27 macro library (SYS1.MACLIB) 47 adding installation-written members 29 main processor 96 main storage (see real storage) main storage hierarchy support 15 main storage occupancy algorithm 55 manual switching 151 master address space 37,38 definition 157 master scheduler initialization 30 MATRIX operand of DISPLAY command 127,125 MAXUSER system initialization parameter 34,31 MCP (message control program) definition 157 TCAM 148 measurement facilities MF/l 68-77 SMF 106-111 measurement facility control (MFC) 68 MEMBER MF /1 parameter 70 MEMBERS system generation parameter 28 message control program (MCP) definition 157 TCAM 148 MFC (measurement facility control) 68 MF/l (system activity measurement facility) 68-77 definition 157 highlights 13 operation 68-69 parameters 69-70 reports 70-72 SYSl.PARMLIB member 43 tracing 72 tuning example 73-77 MF/l parameter list (IRBMFlxx) 43 MF /1 reports 70-72 channel activity 72 CPU activity 70 I/O device activity 72 paging activity 70-71 workload activity 71-72 MIC (missing interruption checker) 157 definition initialization 30 MIN system initialization parameter 37 minimum configuration for multiprocessing 149 for starter system 21 missing interruption checker (MIC) definition 157 initialization 30 MLPA system initialization parameter 34,31 MN operand 126 MOD system initialization parameter 38 modified link pack area 153 system initialization parameter specification 33 modified link pack area list (IEALP Axx) 45 MODIFY command 125 MODIFY subcommand 136 MONITOR command 128,126 MONITOR subcommand 136 MOUNT command 126,128 MP (see multiprocessing) MPA system initialization parameter 37 MPS system initialization parameter 38 MSGRT command 126 multiple address spaces 13 definition 157 specifying maximum number with MAXUSER parameter 34 multiple locks 149 multiple virtual address spaces 13 definition 157 specifying maximum number with MAXUSER parameter 34 Index 165 multiprocessing (MP), tightly coupled considerations for VS2 Release 2 149-151 definition 157 highlights 12 operator command support 127 SMF record support 110 system integrity considerations 82 multiprocessing (MP), loosely coupled 12 mUltiprogramming 13 multitasking considerations for a multiprocessing environment 82,151 N operand 125 NET operand 127 DISPLAY command 125 HALT command 127 MODIFY command 126 START command 126 VARY command 127 Network job processing (NJP) 98,102 NIP (nucleus initialization program) 30 NJP 98,102 NOCHAN MF /1 parameter 69 NOCHRDR MF/l parameter 69 NOCOMM MF/l parameter 69 NOCPU MF/l parameter 69 NODASD MF/l parameter 69 NODEVICE MF/l parameter 69 NOGRAPH MF /1 parameter 69 nonpageable dynamic area 146-147 nonstandard job 100 non-standard control program interface 81 not-in-use data sets 115 NOTE addressed to SYSIN/SYSOUT DCB 15,16 NOOPTIONS MF/l parameter 70 NOPAGING MF/l parameter 69 NORECORD MF/l parameter 69 NOREPORT MF/l parameter 69 NOT APE MF/l parameter 69 NOTRACE MF/l parameter 69 NOUNITR MF/l parameter 69 NOWKLD MF/l parameter 69 nucleus 153 nucleus initialization program (NIP) 30 nucleus library (SYS1.NUCLEUS) 47 adding installation-written members 29 nucleus-only generation 20,15 NUCMAP system initialization parameter 34,31 o PER operand 126 operator commands 124-131 console, JES3 103 control of multiprocessing configurations 149-151 reducing interaction during system initialization 39 OPERATOR command 136 operator commands 124-131 job entry subsystem support 128-131 multiprocessing support 127 summary of changes 125-127 system resources management support 128 time sharing support 128 VT AM support 127 operator intervention restriction 34,31 OPI system initialization parameter 34,31 OPT system initialization parameter 34,31 OPTIONS MF/l parameter 70 ordered seek queueing 16 OUT operand CANCEL command 125 RESET command 126 OUTCLASS operand 125 OUTUM job control language paramet(~r 116,113 OUTPUT command 137 job entry subsystem support 138 OUTPUT control card 118 output limit specification 116 output record (SMF) 110 166 OS/VS2 Planning Guide for Release 2 output service 93 OUTPUT system generation macro )mstruction 27 page definition 157 page data sets 47 adding paging space to the system 50 change from VS2 Release 1 49 definition 157 14 highlights naming (PAGE parameter) 34 page fault definition 157 page fixing definition 157 page frame definition 157 MF/l counts 71 page frame table entries 153 page-in counts in SMF records 106 157 definition rates and percentages in MF /1 n:ports 71 page-out counts in SMF records 106 definition 157 rates and percentages in MF /1 reports 71 page reclamation definition 157 rates and percentages in MF / 1 n~ports 71 page replacement algorithm 55 page stealing 55 definition 157 PAGE system generation macro instruction 29,27 PAGE system initialization parameter 34,31 adding paging space to the system 50 difference from VS2 Release 1 36 page table (PGT) definition 157 page translation exception definition 157 page able BLDL table 153 system initialization parameter specification 32 page able link pack area 153 creation using the CLP A parameter 32 SYS1. LPAUB 47 transient area replacement 16 pageable real storage counts 71 paging 157 definition replacement for rollout/rollin 15 replacement for scatter load 1:5 space 50 paging activity measurement 68,70-71 paging activity report 70-71 paging device definition 157 PAGING MF/l parameter 69 paging rate MF/l counts 71 definition 158 paging space (page data sets) 47,50 adding paging space to the system 50 change from VS2 release 1 49 definition 158 highlights 14 naming (PAGE parameter) 34 PAL system initialization parameter 37 parameters, JCL ADDRSPC 116-117,113 AFF 113 COPIES 115,113 DCB 113 DEST 115,113 DPRTY 117,113 DYNAM 113,115 DYNAMNBR 115,113 FREE 115,113 HIARCHY 113 OUTUM 116,113 PERFORM 113,114 PRTY 117-118,114,113 REGION 116,113 RETAIN 113 ROLL 113 SEP 113 SPACE 113 SPLIT 113 SUBALLOC 113 SYSABEND 42 SYSUDUMP 42 UNIT 113 parameters, MF /1 69-70 example 73-75 parameters, system initialization ALTSYS 38 APF 32,31 APG 32,31 AUXLIST 37 BLDL 32,31 BLDLF 32,31 CLPA 32,31 CMD 32,31 CPQE 37 CSA 32,31 CVIO 33,31 DUMP 33,31 FIX 33,31 HARDCPY 33,31 HRAM 38 HSVC 38 IPS 33,31 LNK 33,31 LOGCLS 33,31 LOGLMT 33,31 LSQACEL 37 MAXUSER 34,31 MIN 37 MLPA 34,31 MOD 38 MPA 37 MPS 38 NUCMAP 34,31 OPI 34,31 OPT 34,31 PAGE 34,31 PAL 37 QBF 38 RAM 38 REAL 34,31 RERP 38 RSVC 38 SMF 35,31 SQA 35,31 SQACEL 37 SQS 38 SYSP 35,31 TMSL 37,38 TRACE 37 TSOAUX 37 VAL 36,31 VRREGN 36,31 WTOBFRS 36,31 WTORPLY 36,31 PARMLIB system generation macro instruction 29,27 PARMTZ (local time constant) 45 password protection integrity 87 PASSWORD data set 48 system data sets 86,80 system libraries 86 tape data sets 86-87 PATH operand 127 PAUSE operand 125 PDS system generation parameter 28 percent channel active value 72 percent channel busy and CPU waiting value 72 PERFORM job control language parameter 113,114 performance groups 158 performance group number 66 JCL specification 114 . time sharing LOGON command 134 performance group periods 158 definition 158 tuning example 73-77 performance group period report 71 performance group report 71-72 performance objective number 67,59-65 performance objectives 59-65 definition 158 effect of changing 76 motive for specifying 59 tuning example 73-77 performance specifications 58 performance variable 57 period duration parameter 66 PGT (page table) definition 158 physical addresses for devices 21 physical environment, computing system 80 PL 1 system generation macro instruction 27 PL/l F compiler 17 library 17 syntax checker 17 PL 1LIB system generation macro instruction 27 POINT addressed to SYSIN/SYSOUT DCB 15,16 preexecution setup 98,100 pre-fixed storage area (PSA) 153 PRES RES list 46,40 PRINT verb 142 priority aging 101 priority I/O queueing 16 priority, job selection 117 priority, job step dispatching 117 private address space definition 158 size recorded by SMF 109 PROC operand 126 procedure library (SYS1.PROCLIB) 48 adding installation-written members 29 processor-only generation 20 procname operand 125 PROFILE subcommand 136 ED IT command 138 program conversion 148 programming conventions 148 V=R 146-147 multiprocessing 151-152 protected system control block 81-82 PRTY job control language parameter 116,114,113 PRTYoperand 126 PSA (pre-fixed storage area) 153 pseudo device 96 PTOP system generation macro instruction 27 Q operand HOLD command 125 RELEASE command 126 SET command 126 QBF system initialization parameter 38 queued telecommunications access method (QT AM) QUIESCE command 127,126 QT AM (queued telecommunications access method) 15 15 R operand 125 RAM system initialization parameter 38 reader 91,156,15,16 internal 156 OS 15,16 real address 158 definition range set for multiprocessing 149-151 real main processor 96 real storage ADDRSPC=REAL 148 definition 158 multiprocessing considerations 149-151 MVT 13 paging activity report of use 70-71 shared 12 Index 167 real storage management (RSM) definition 158 real storage occupancy resource-use algorithm 54-55 service definition coefficient MSO 67 service received during interval of 67 REAL system initialization parameter '34,31 difference from VS2 Release 1 37 Reclaim (of data set) 52 reconfiguration 149-151 definition 158 JES3 104 RECORD MF /1 parameter 69 recovery definition 158 mUltiprocessing 150 recovery management routines 153 recovery termination manager definition 158 initia.lization 30 refreshable definition 158 modules in the page able link pack an:a 47 REGION job control language parameter 116,113 REGSIZE operand 126 RELEASE command 126 relocate hardware (dynamic address translation) definition 155 remote job entry (RJE) 15 remote terminal access method (RT AM) definition 158 remote terminal programs 25 REPORT MF/l parameter 69 REPRO verb 142 RERP system initialization parameter 38 RESET command 126 resident BLDL list (IEABLDxx) 43,44 RESMODS system generation macro instruction 29,27 resource factor coefficients 67 use of 56 resource identification 83 resource manager constants 67 resource-use routines 54-55 response/throughput bias (RTB) 66 definition 158 RET AIN job control language parameter 113 RJE (remote job entry) 15 RMTGEN 25 definition 158 options 25 ROLL job control language parameter 113 rollout/rollin 15 RPG system generation macro instruction 27 RSM (real storage management) definition 158 RSVC system initialization parameter 38 RTAM (remote terminal access method) definition 158 RTB (response/throughput bias) 66 definition 158 RUN subcommand 136 SAMG (system activity measurement gathering) 68 SARG (system activity report generation) 68 satellite graphic job processor (SGJP) 15 SAVE operand 126 15 scatter load scheduler work area (SW A) 153 definition 158 SYS1.SYSJOBQE replacement 16,49 SCHEDULR system generation macro instruction 23,27 secondary storage (auxiliary storage) 155 definition paging activity report of use 70-71 SECONSLE system generation macro instruction 29,27 security, system 86-87 segment definition 158 segment table (SGT) definition 158 168 OS/VS2 Planning Guide for Release 2 SEND operator command 128,126 SEND subcommand EDIT command 138,136 OPERATOR command 137 SEND time sharing command 137 SEP job control language parameter 113 separator page 93 serialization, allocation 121,13 serialization techniques 82-83 disablement 82 ENQ/DEQ 83,151 locking 83 WAIT/POST 151 service 57 service definition coefficients 67,65 use of 57-59 service rate 59-65 definition 158 effect of change for a performance objective 76 IPS specification 67 tuning example 73-77 workload activity reports 71-72 service units 57-65 definition 158 MF/l workload activity reports 71-72 SMF record counts 106 tuning example 73-77 SERO routine 15 SER 1 routine 15 SET command used to identify the CPU with an acceptable clock value 127 used to restore time zone constant to IPL value 127 used to specify an IPS 126 SET operand 125 setup JES2 96 JES3 98,100 SGJP (satellite graphic job processor) 15 SGT (segment table) definition 158 shared DASD, recovery of 150 SIO instruction counts 72 slot definition 158 MF/l counts 71 SMF (system management facilities) 106-111 differences from MVT 106 differences from VS2 Release 1 106 exits 111 records 107-108 system initialization parameter specification 35 SMF data sets (SYS1.MANX/SYSl.MANY) 48 change from MVT 50 change from VS2 Release 1 49 SMF operand 126 SMF parameter list (SMFPRMxx) 44 SMF system initialization paramete:r 35,31 SMFDEFLT parameter list 44,40 SMFPRMxx (SMF parameter list) 44 software recording facility definition 158 SORT/MERGE 17 SORTLIB system generation macro instruction 27 SORTMERG system generation macro instruction 27 SPACE job control language parameter 113 SPACE system generation macro parameter 28 spin lock definition 159 spinoff 93,94 SPLIT job control language parameter 113 spool 99 SQSA (system queue area) 153 definition 159 system initialization parameter specification 35 SQA operand 125,147 SQS system initialization paramet~:r 35,31 SQACEL system initialization parameter 37 SQA system initialization parameter 38 SSM instruction 148 stage I 20 stage II 20 standard control program interface 81 START command 128,126 START operand 127 starter system 21 minimum configuration 21 physical addresses for devices 21-22 STEPCAT catalog 141 STIDP instruction 38 STOP command 128 STOP MF/l parameter 70 STOPMN subcommand 136 STOPTR command 128,127 STORAGE operand 127 storage map, virtual 153 store protection 87 SUBALLOC job control language parameter 113 SUBMIT command 137 job entry subsystem support 138 SUBMIT operand 126 support processor 96 suspend lock definition 159 SUPRVSOR system generation macro instruction 27 SVC library (SYSl.SVCLIB) 47 adding installation-written members 29 SVC routines calling SVC routines 84 SVC routines, installation-written integrity considerations 85 SYSGEN specification with SVCT ABLE macro 23 SVCTABLE system generation macro instruction 23,27 SW A (scheduler work area) 153 definition 159 replacement for SYSl.SYSJOBQE 16,49 swap counts 71 swapping definition 159 ER V effect on 67 ISV effect on 66 use by the system resources manager 57 switching, manual 151 symmetric processors definition 159 symmetric storage configuration definition 159 symmetricalI/O 149 SYNC subcommand 146,135 SYSABEND job control language parameter 42 SYSCTLG data set 49 conversion to VSAM format 139-145 SYSDSN major name 83 SYSGEN (see system generation) SYSIEAOI major name 83 SYSIEECT major name 83 SYSIEFSD major name 83 SYSOUT MF/l parameter 69 SYSP system initialization parameter 35,31 SYSPSWRD major name 83 system activity measurement facility (MF/l) 68-77 definition 159 13 highlights operation 68-69 parameters 69-70 reports 70-72 SYSl.PARMLIB member 43 tracing 72 tuning example 73-77 system activity measurement gathering (SAMG) 68 system activity report generation (SARG) 68 system area 153 system catalog (SYSCTLG) 49 139-145 conversion to VSAM format highlights 12 system data sets 46-50 changes to MVT 50 changes to VS2 Release 1 49-50 optional 48,49 password protection 86 required 47-48 system environment recording routines (SERO and SERl) 15 system generation changes 25-29 JES2 generation 24-25 job entry subsystem generation 20,24-25 macro instructions (see macro instructions, system generation) procedures 20-29 RMTGEN 25 starter system 21-22 unsupported VS2 Release 1 macro instructions 29 system initialization 30-46 parameters 30-39 (see also parameters, system initialization) system parameter library (SYS 1.PARMLIB) 39-46 system integrity 79-87 areas of concern 81-85 definition 79 installation responsibility 80 procedural responsibilities 86-87 system key 81 definition 159 system parameter library (SYS1.PARMLIB) 39-46 IBM-created lists 41 installation-created lists 45 required system data set 47 summary of changes 40 106-111 system management facilities (SMF) differences from MVT 106 differences from VS2 Release 1 106 111 exits records 107-108 system initialization parameter specification 35 system parameter list (IEASYSxx) 44 system parameter specification 37 system protection key 81 definition 159 system queue area (SQA) 153 definition 159 system initialization parameter specification 35 system resources management 54-67,73-77 definition 159 highlights 13 JCL support 114 operator command support 128 parameters 65-67 SMF records 107-108 time sharing command support 134 . tuning example 73-77 system resources management tuning parameter list (IEAOPTxx) 67,45-46 system restart 33 system security 86-87 system summary report 74 system workload 58 SYSUDUMP job control language parameter 42 SYSUTILS system generation macro instruction 27 SYSVTOC major name 83 SYS l.ACCT data set 50 unsupported MVT function 16 SYS l.BRODCAST data set 48 formatting 146 mounted for system initialization 138 new format description 49 SYS1.CMDLIB data set 48 adding installation-written members 29 SYS1.DCMLIB data set 48 SYSl.DSSVM data set 47 change from MVT 50 change from VS2 Release 1 49 specifying creation of a new nucleus map with the NUCMAP parameter 34 SYS 1. DUMP data set 48 device specification by a system initialization parameter 33 maximum number specified 36 SYS 1. HELP data set 48 SYS1.IMAGELIB data set 48 adding installation-written members 29 116 use for 3886 OCR devices SYS1.LINKLIB data set 47 adding installation-written members 29 Index 169 SYS I.LOGREC error recording data set 47 change from MVT and VS2 Release I 49 SYS1.LPALIB data set 47 adding installation-written members 29 SYS I.MACLIB data set 47 adding installation-written members 29 SYS1.MANX data set 48 change from MVT 50 change from VS2 Release 1 49 SYS I.MANY data set 48 change from MVT 50 change from VS2 Release 1 49 SYS1.NUCLEUS data set 47 adding installation-written members 29 SYS1.PAGE data set 49 SYS1.PAGEDUMP 49 SYS 1. P ARMLIB (system parameter library)' 39-46 IBM-created lists 41 installation-created lists 45 required system data set 47 summary of changes 40 SYS1.PARMLIB lists changes to 40 IEAABDxx 42 IEAAPFxx 42 IEAAPPOO 41 IEABLDxx 43-44 IEADMPxx 42 IEAFIXxx 42 IEAIPSxx 42 IEALODOO 43 IEALPAxx 45 IEAOPTxx 45-46 IEAPAKOO 43 IEASYSxx 44 IQAORDER 40 IRBMFlxx 43 LNKLSTxx 43 PARMTZ 45 PRES RES 46,40 SMFDEFLT 44,40 SMFPRMxx 44 VATLSTxx 46 SYS1.PROCLIB data set 48 adding installation-written members 29 SYSl.SAMPLIB data set 48 SYSl . STGINDEX data set 48 SYS l.SVCLIB data set 47 adding installation-written members 29 SYS 1.SYSJOBQE data set 49 JES2 91 JES3 96 unsupported MVT function 16 unsupported VS2 Release 1 function 16 SYS I.SYSVLOGX (see log data sets) SYS I.SYSVLOGY (see log data sets) SYS 1. TELCMLIB data set 48-49 adding installation-written members 29 SYS1.UADS (user attribute data set) 48 formatting 146 mounted for system initialization 138 new format description 50 SYS 1. VT AMLIB data set 49 adding installation-written members 29 SYS 1. VT AMLST data set 49 adding installation-written members 29 SYS1.VTAMOBJ data set 49 T operand 125 tape data set password protection 86-87 TAPE MF /1 parameter 69 TCAM (telecommunications access method) conversion considerations 148 system generation specification 23 TELCMLIB system generation macro instruction 27 telecommunications access method (TCAM) conversion considerations 148 system generation specification 23 telecommunications library (SYS 1.TELCMLIB) 48-49 adding installation-written members 29 170 OS/VS2 Planning Guide for Releas·e 2 temporary data set 51 TEST AUTH facility 51 TESTRAN program 16 tightly coupled mUltiprocessing (MP') considerations for VS2 Release 2 149-151 definition 157 highlights 12 operator command support 127 SM.F record support 110 TIME command 137 time dependency 148 83 time-of-check-to-time-of-use problem time sharing accounting support 138 allocation support 134 command changes 135-137 edit support 138 job entry subsystem support 138 SMF record elimination 1 10 system resources manager support 134 time sharing integration support 138 VSAM support 134 time slicing 16 TMSL system initialization parameter 37,38 TOO clock messages 30 TP operand 125 TRACE command 127 TRACE MF/l parameter 69 trace, TSO 16 tracing, MF /1 72 SMF records 106 TRACK command 128,127 transaction 159 transient areas 16 translation tables 159 TS operand DISPLAY command 125 TRACK command 127 TSO (see time sharing) TSO driver 16 TSO system generation macro instruction 23,27 TSO trace 16 TSOAUX system initialization parameter 37 TSOPTION system generation macro instruction turnaround time 73 tuning parameter list (IEAOPTxx) 67,45-46 tuning with MF/1 73-77 type I programs 17 u operand 125 UADS (see SYS1.UADS) UADSREFM utility program 146 UCS system generation macro instruction 29,27' unallocating data sets (see freeing data sets) uniprocessing 159 UNIT job control language parameter 1 13 UNITNAME system generation macro instruction 121-122,23,27 114,51 specifying VIO processing UNITR MF /1 parameter 69 units code parameter 66 unsupported MVT functions job control language summary 113 operator command summary 125-127 SMF record summary 107 -108 summary 15-16 system data sets 50 system initialization parameters 37-38 time sharing command summary 135-137 unsupported Type I programs 17 unsupported VS2 Release 1 functions job control language summary 113 operator command summary 125-127 SMF record summary 107 -108 summary 16 system data sets 49-50 system generation macro instructions 29 system initialization parameters 37 time sh",ring command summary 135-137 27 user address space (see private address space) user area 153 user-assigned names 121 user attribute data set (SYS1.UADS) 48 formatting 146 mounted for system initialization 138 new format description 50 user key 85 USER operand 125 user protection key 85 USERID operand 125 user-supplied addresses 81 USERS operand 126 utility programs IEHIOSUP 15 IEHLlST 143,15,16 IEHMOVE 143,15,16 IEHPROGM 143,28,15,16 IEHUCAT 143 v= R (virtual equals real) address area ADDRSPC JCL parameter specification 116-117 candidate programs 148 definition 159 programming considerations 146-147 REAL system initialization specification 34 VRREGN system initialization specification 36 V=R programming considerations 146-147 VAL system initialization parameter 36,31 validity checking 80 fetch/store operations 83 SVC routines calling SVC routines 84 unauthorized program requests for authorized library routines 84 user-supplied addresses 81 VARY command 149,127 VATLSTxx (volume attribute list) 46 VIO (virtual I/O) catalog for data sets (STGINDEX) 48 data set allocation 122-123 definition 159 12 highlights JCL requirements 114-115 MF /1 report data 71 SMF record support 106 SYSGEN specification with the UNITNAME macro 23 system initialization parameter specification 33 virtual address definition 159 virtual equals real address area (V =R) ADDRSPC JCL parameter specification 117 candidate programs 146-148 definition 159 REAL system initialization specification 34 VRREGN system initialization specification 36 virtual I/O (VIO) catalog for data sets (STGINDEX) 50 definition 159 12 highlights JCL requirements 114-115 MF /1 report data 71 SMF record support 106 SYSGEN specification with the UNITNAME macro 2.3 system initialization parameter specification 33 virtual memory (virtual storage) definition 159 highlights 13 virtual storage definition 159 13 highlights virtual storage access method (VSAM) catalog 139-145 definition 159 12 highlights SMF record support 110 time sharing support 134 virtual storage management (VSM) definition 159 virtual storage map 153 virtual telecommunications access method (VT AM) command support 125-127 highlights 12 system data sets 49 volume attribute list (V ATLSTxx) 46 volume sharing 116 139 volume table of contents (VTOC) VRREGN system initialization parameter 36,31 VSAM (virtual storage access method) catalog (see VSAM catalog) definition 159 highlights 12 SMF record support 110 time sharing support 134 VSAM catalog 139-145 creating 142-143 master catalog 47 search strategy 141 structure 139-140 updating 142-143 142-143 use of access method services VSM (virtual storage management) 159 VTAM (virtual telecommunications access method) command support 125-127 13 highlights system data sets 49 VT AM data sets SYS 1. VT AMLIB 49 SYS 1. VT AMLST 49 SYS1.VTAMOBJ 49 VTOC (volume table of contents) 139 WKLD MF /1 parameter 69 workload activity measurement 68 workload activity reports 71 performance group period report 71 performance group report 71-72 system summary report 72 workload level 59-66 function of initiated jobs 63 MF /1 data in reports 71-72 tuning example 73-77 used to express service relationships between groups of users 59 workload management 54,56,58 definition 159 WRITELOG command 127 writer internal 156 OS 15,16 WTOBFRS system initialization parameter 36,31 WTORPLY system initialization parameter 36,31 XDAP addressed to SYSIN/SYSOUT DCB XPT (external page table) definition 156 2-channel switches 3886 OCR device 15,16 149 116 Index 171 172 OS/VS2 Planning Guide for Release 2 OS/VS2 Planning Guide for Release 2 G C28-0667-1 Your views about this publication may help improve its usefulness; this form will be sent to the author's department for appropriate action. Using this form to request system assistance or additional publications will delay response, however. For more direct handling of such requests, please contact your IBM representative or the IBM Branch Office serving your locality. Possible topics for comment are: Clarity Accuracy Completeness Organization Index Figures Examples Legibility What is your occupation? ----------------------------_____________ _ Number of latest Technical Newsletter (if any) concerning this publication: ____________ _ Please indicate in the space below if you wish a reply. Thank you for your cooperation. No postage stamp necessary if mailed in the U.S.A. Elsewhere, an IBM office or representative will be happy to forward your comments. READER'S COMMENT FORM GC28-0667-1 Your comments, please ... This manual is part of a library that serves as a reference source for system analysts, programmers, and operators of IBM systems. Your comments on the other side of this form will be carefully reviewed by the persons responsible for writing and publishing this material. All comments and suggestions become the property of IBM. Fold I Fold - - - - - - - - - - - --~ First Class Permit 81 Poughkeepsie New York I I I I o (/) < (/) I\.) ""0 Qj" ::J ~. Business Beply Mail ::J co G) No postage stamp necessary if mailed in the U.S.A. 0.: CD I ..,0' I I I Postage will be paid by: I nternational Business Machines Corporation Department D58, Building 706-2 PO Box 390 Poughkeepsie, New York 12602 I I I I ----------------~ Fold Fold I I I I I I I I l!J]3llir (!) International Business Machines Corporation Data Processing Division 1133 Westchester Avenue, White Plains, New York 10604 (U.S.A. only) IBM World Trade Corporation 821 United Nations Plaza, New York, New York 10017 (International) c I I I I I I I I
Source Exif Data:
File Type : PDF File Type Extension : pdf MIME Type : application/pdf PDF Version : 1.3 Linearized : No XMP Toolkit : Adobe XMP Core 4.2.1-c043 52.372728, 2009/01/18-15:56:37 Create Date : 2012:06:08 18:40:35-08:00 Modify Date : 2012:06:08 23:16:27-07:00 Metadata Date : 2012:06:08 23:16:27-07:00 Producer : Adobe Acrobat 9.51 Paper Capture Plug-in Format : application/pdf Document ID : uuid:6890bc4d-3432-4aae-9eea-cdb5932650bc Instance ID : uuid:67282d0f-20a7-413c-b705-cfd896039632 Page Layout : SinglePage Page Mode : UseOutlines Page Count : 174EXIF Metadata provided by EXIF.tools