JMP opens the job pool file (which is in PAM format) and sets up a SAM file (referred to below as a procedure file) for the ENTER commands to be generated.
JMP either determines the file name of the job pool file via a standard link name (link name SJOBPOOL), or the file name is explicitly specified in the JMP statement //OPEN-JOBPOOL-FILE. When transferring the name of the job pool file using a link name, you must execute the command /ADD-FILE-LINK LINK-NAME=SJOBPOOL, FILE-NAME=<filename> prior to calling JMP.
JMP determines the name of the procedure file in the same way, either via the standard link name PJOBPOOL or via explicit specification in the JMP statement //CREATE-PROCEDURE-FILE. The procedure file can be either a newly created file or an existing file which is overwritten. The program is started with /START-JMP.
START-JMP |
VERSION = *STD / <product-version> ,MONJV = *NONE / <filename 1..54 without-gen-vers> ,CPU-LIMIT = *JOB-REST / <integer 1..32767 seconds> |
JMP is controlled by means of control statements read from SYSDTA.
Notes on the job pool file
The job pool file named SYSTEM.JOBPOOL that is current during the session is located in the home pubset. It is opened during a session by the job pool task (system task, TSN JOBP) and can therefore not be accessed by the JMP program.
If the home pubset is imported as a data pubset (after a shutdown), JMP has unlimited access to the job pool file. However, the file is protected by a read password and must therefore be copied to an unprotected file using /COPY-FILE ...,IGNORE-PROTECTION=*SOURCE, before it can be processed by JMP.
The job pool contains information on all of the jobs currently being processed and the jobs yet to be run. These are current or interrupted batch jobs and all batch jobs that have not yet been started (with repeat jobs or calendar jobs, they are the versions that still need to be started). Immediately prior to a planned shutdown, the job pool file should therefore contain only jobs with a start time that is later than the planned session restart. When the system starts up again (STARTUP), a new job pool file is started and set up with job pool information from the previous session. All batch jobs that no longer need to be considered in the new session are removed.