Version 4.2.6 for Mainframes (Update)
 —  Operations  —

Natural in Batch Mode (All Environments)

This document contains general considerations that apply when running Natural in batch mode.

The following topics are covered:


Adabas Datasets

Adabas datasets must be specified only in single-user mode. They are identical to those required for the execution of any normal application program using Adabas. See the relevant Adabas documentation for detailed information on Adabas datasets.

Top of page

Sort Datasets

Sort datasets must be specified if a Natural program containing a SORT statement is to be executed during the Natural session.

The requirements are identical to those for execution of a normal COBOL or PL/1 application program that invokes the operating system sort program and can vary according to the sort program in use.

Natural does not require the intermediate datasets SORTIN and SORTOUT, but communicates with the sort program via the E15 and E35 user-exit routine interfaces.

Top of page

Subtasking Session Support for Batch Mode Environments

Note:
With Natural for CMS, subtasking is not supported.

Purpose

With subtasking support, you can run multiple Natural batch mode sessions within one address space. This allows parallel processing within one address space, rather than executing subsequent job steps, and can increase throughput dramatically.

Typically, client/server applications and products would take advantage of this functionality, for example, the Natural remote procedure call. Multiple server subtasks can be started to communicate with remote clients.

Prerequisites

If you wish to restart the Natural nucleus, it must be linked as a reentrant module (linkage editor option RENT).

The Adabas link routine (ADALNK) must be generated with reentrancy support.

Functionality

You start a subtask by issuing a CALL statement from a Natural program. The new Natural session ("subtask") is started with an extended front-end parameter list. This list contains up to three parameter sets:

Variable names for standard I/O datasets (for example CMPRINT) and other parameters for the batch mode interface startup can be passed from the starting program in the startup parameter area. Standard I/O datasets can be undefined or dummy datasets; they can be owned by one session or shared by multiple sessions.

Furthermore, a CALL interface is provided for reading the user parameter area with a Natural program.

Starting a Natural Session

Extended Parameter List

The Natural batch mode interface without extended parameter list gets initial control from the operating system using standard linkage call. Register 1 points to an address with high-order bit on as the last address indicator. This address points to a halfword field containing the length of the following parameter area.

The extended parameter list contains up to three parameter addresses. This is indicated by the high-order bit in the last address which can be the first, second or third address. All parameter addresses point to a halfword field containing the parameter length of the following parameter area. Zero length indicates that there is no parameter area.

Startup Parameter Area

When multiple batch mode Natural (sub)tasks are running in the same region, by default these sessions access the very same Natural standard I/O datasets (such as CMPRINT, CMSYNIN, etc), as there are no Natural profile parameters available to set these file names. Also by default the Natural system variables *INIT-ID and *INIT-USER are identical because of their definition for batch mode.

In order to provide unique standard I/O dataset names and unique IDs for Natural subtask sessions the startup parameters in the extented parameter list can be used to overwrite the Natural system defaults. The Startup Parameter area is a table of pairs of 8-character fields:

Keywords and replacement values must be padded with trailing blanks, if necessary.

The following keywords are valid:

CMHCOPY Permanent hardcopy destination
CMSYNIN Command input dataset name
CMOBJIN Object input dataset name
CMPRINT Standard output dataset name
CMPRMIN Dynamic parameter input dataset name
CMPLOG Dynamic parameter output dataset name
CMTRACE Trace output dataset name
INITID Job step name (system variable *INIT-ID)
MSGCLASS Spool class for dynamic allocation of CMPRINT and CMTRACE (z/OS only)
NATRJE Job submission dataset name (z/OS only)
STEPLIB Program load library name (see also profile parameter LIBNAM, Name of Load Library, z/OS only)
SUBPOOL z/OS storage subpool (0 - 127, right justified)
USERID Initial user identification (system variable *INIT-USER)

The usage of these entries is optional and no particular sequence is required. A blank value for a dataset means that this dataset is not available or is empty.

Note concerning z/VSE:

By default, all print output (that is, the one resulting from CMPRINT, CMHCOPY, CMTRACE and CMPLOG) is routed to SYSLST. An overwrite specification for these files starting with SYS is considered a z/VSE system number overwrite. Possible format is SYSnnn where nnn is a three-digit number in the range from 000 to 099; if you specify an invalid number nnn, it is ignored.

User Parameter Area

The format of the user parameter area is free. It can be accessed from any Natural program by a special CALL interface see Accessing the User Parameter Area.

Starting a Subtask

The following call interface is supplied to be used by Natural programs to start a subtask in the same address space.

PGMNAME Natural nucleus name getting control (mandatory). To restart with the same nucleus, an asterisk can be specified as the first character. The actual nucleus name is passed back in this field.
NATPARML Natural dynamic parameter area
STRPARML Startup parameter area
USRPARML User parameter area

All parameter areas must start with the length of the following parameters. The following example illustrates the usage of CMTASK.

Example:

DEFINE DATA LOCAL
01 PGMNAME (A8) INIT <'*'>
01 PARM1
02 NATPARML (I2) INIT <30>
02 NATPARMS (A30) INIT <'INTENS=1,IM=D,STACK=MYPROG'>
01 PARM2
02 STRPARML (I2) INIT <32>
02 STRPARM1 (A16) INIT <'CMPRINT SYSPRINT'>
02 STRPARM2 (A16) INIT <'CMPRMIN MYPARMS'>
01 PARM3
02 USRPARML (I2) INIT <80>
02 USRPARMS (A80) INIT <'special user parameters'>
END-DEFINE
CALL 'CMTASK' PGMNAME NATPARML STRPARML USRPARML
END

A sample program, ASYNBAT, can be found in library SYSEXTP.

Accessing the User Parameter Area

The user parameter area passed during startup can be read from any Natural program with the following CALL statement:

CALL 'CMUPARM' USRPARML USRPARMS

USRPARML is the length (I2) of the USRPARMS area (before the call) and the length of the data returned (after the call). USRPARMS is the parameter data area.

If the length of the data to be returned is greater than the area length, the data is truncated to the area length. The following return codes are possible:

0 Data successfully moved
4 Data moved but truncated
8 No data available
12 Length value not positive
16 Insufficient number of parameters

A sample program, GETUPARM, can be found in library SYSEXTP.

Top of page