gtpm6m09Main Supervisor Reference

Initial Program Loading

There are 2 types of IPL, software and hardware.

On a hardware IPL the operator is given the opportunity to modify the current active TPF image.

The devices reset by a hardware IPL are not reset by a software IPL. IPL segments:

Multiple TPF Images

Multiple TPF images allow the TPF user to integrate changes in a TPF environment by:

It is possible to have as many as 8 images. One primary image is designated for use during a hard IPL. Each image has a unique core image restart area (CIMR). Each image also has an IPL area (IPLA and IPLB), E-type program base, and a keypoint staging area (to facilitate the pre-loading of keypoints in NORM state).

Note:
There is one working keypoint area that contains the active keypoints. This area is used no matter what image you IPL on.

If, upon a hardware IPL, you want to modify the current active image and more than one image is enabled, all the enabled images will be displayed. You will then be prompted to select an image for the BSS. If the selected image's IPL area is different from the primary image's IPL area then a software IPL will occur, so that the new IPL area can be read in. The primary image is the image whose IPL area is used during a hardware IPL prior to prompting the operator for image selection.

The ZIMAG commands and ZTPLD command can be used to define and manipulate the TPF images. See TPF Operations for detailed information on these commands.

The ZIMAG Commands

Option
Description

DEFINE
Allows you to define (or redefine) as many as 8 images.

DISPLAY IMAGE
Displays the image name, status, associated IPL and program areas, CTKX version code (if physically loaded), and CIMR component.

ENABLE
Allows you to enable an image for an IPL.

PRIMARY
Defines an enabled image as the primary image. The primary image is used during a hard IPL and is valid only on the basic subsystem (BSS).

DISABLE
Disables an enabled image so that it cannot be IPLed.

CLEAR
Deletes a disabled image.

COPY
Allows you to copy core image restart area (CIMR) components from one image to another by reference (logically) or physically.

UNREF
Deletes the logical references of CIMR components from a disabled image.

MAKEPHYS
Allows you to make all of your CIMR component references physical copies.

DISPLAY PROG
Displays all of the program areas defined in the TPF 4.1 system and identifies which (if any) images they are associated with.

DISPLAY IPL
Displays all the IPL areas that were loaded and identifies which (if any) images they are associated with. This option also displays IPLA and IPLB information for each area.

DISPLAY PROCESSOR
Displays the image associated with each processor in the complex, as well as the status of the processor.

The ZTPLD Command

Message
Description

ZTPLD
Load data to a specified image

Considerations for HPO Users

Under MDBF, the operator specifies the IPL type as one of the following:

You can define a new subsystem configuration when you select the multiple subsystems option. Any existing subsystem configuration is cleared.

The fast IPL option allows you to configure subsystems from the existing configuration. If there is an I/O error, the subsystem configuration is purged. Otherwise, the new subsystem configuration is copied from the existing one, and further operator intervention is not required.

Keypoint record M (CTKM) contains subsystem configuration data. The IPL program uses CTKM each time a new subsystem configuration is generated. Under MDBF, a software IPL can be activated when the system operator issues ZRIPL, or when a catastrophic error occurs. During the IPL, the operator may receive a message indicating there is not enough storage to continue. The operator can use the ZFKPA command to reduce the number of main storage blocks or to cancel the IPL.

MDBF System Definition

The MDBF system is loaded as one of the following:

Generating multiple subsystems is a process that includes:

The IPL program determines the configuration for a basic subsystem with one or more subsystem users. IPL also assigns an identifier or subsystem user ID (SSU ID) to each subsystem user.

In a system with multiple subsystems, the configuration can be determined by either the system operator or the IPL program. For an operator-determined configuration, the operator provides subsystem definitions during loading of the subsystem keypoints. The IPL program configures multiple subsystems when a software IPL is started or fast IPL has been specified.

Both methods require the use of CTKM. The system initialization program (SIP) generates CTKM, but this keypoint record is not completely initialized until the first IPL is completed. System startup, restart, and state change programs perform initialization during the first and subsequent IPLs of the system. The MVS utility program ICKDSF initializes the basic subsystem loader general file with standard IPL records and the IPL segment IPLA. ICKDSF also initializes the volume labels on the online basic subsystem file disk packs. The general file loader online segment (ACPL) writes the IPL1 and IPL2 records to these disk packs.

Processor Identification

Loosely coupled processors require unique identification for a number of reasons, such as identifying processor unique keypoints and communication between processors. The control program records the serial numbers from the processors that are IPLed and places them in keypoint record I (CTKI). CTKI is referred to as the loosely coupled identity table. For each processor it loads, the IPL program establishes a connection between the processor and a central processing unit (CPU) ID. Either the system operator (under IPL program direction) or the IPL program itself (using the information in CTKI) can establish processor identifiers.

Each processor must be assigned an ID for the IPL to continue. Once an ID is assigned to a processor, the process of coupling the processor with the rest of the complex begins.

Lock Management

TPF loosely coupled processors share a common database that resides on shared DASD. Updates to the database are controlled by locking instructions executed in the shared DASD control units. DASD lock facilities-the limited lock facility (LLF) and the concurrency filter lock facility (CFLF)-reside in DASD control units and provide the hardware support for these locking instructions. All DASD used by a TPF loosely coupled complex must be connected to control units with LLF or CFLF.

LLF and CFLF provide locking based on the physical attachment of the processor to the control unit. The point of attachment between a DASD control unit and a channel from a processor is called a control unit interface. In an 8 processor loosely coupled complex each control unit would require 8 interfaces, one for each processor in the complex.

IDs are assigned to the control unit interfaces on each control unit starting with one and as many as the maximum number of interfaces available. A given processor in a loosely coupled complex using LLF must be attached to the same interface ID on all attached DASD control units. This restriction does not apply if CFLF is used.

The IPL program examines all of the control units that a processor has access to and identifies the existing LLF or CFLF relationships. If LLF or CFLF errors are found, the IPL program may limit the system to single processor mode despite the fact that the loosely coupled facility is present.

Only one processor in a loosely coupled complex loads the control program to the online modules. No other processor should be active when the one processor loads the control program.