If you have access to Omegamon/MVS there are a few commands that are
useful in visualizing your address space region, subpools and TCB
structure. Use major command PEEK (don't forget dash in column 1) and
minor commands STEP, AMAP, TCBS and SUBP ("x" in column 1 for expanded
info. -db
----+----1----+----2----+----3----+-
________________ ZMENU VTM OM/DEX
-PEEK jobname
step
amap
Xtcbs
Xsubp
-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On
Behalf Of Johnny Luo
Sent: Tuesday, July 10, 2007 11:47 AM
To: [email protected]
Subject: JOB STEP TASK and REGION parm
Hi,
I know REGION parm has been discussed many times here and hope I'm not
asking the same one again. (I spent almost two weeks searching archives
and reading Authorized Assembler Service Guide but I'm still not so sure
whether that's enough before asking here.)
Last year I got one good article from NASPA about REGION parm and it
discussed in detail what the REGION parm implies both above and below
the line. However, as I become more accustomed to assembler language i
found there are still something I don't know about this parm.
The most accurate explanation I found is from Mr. Chris Craddock in a
discussion eight years ago. He said that REGION is actually restricting
the amount of 'Private Low' storage owned by tasks which means an
authorized task can get more than that from other subpools.
But to think it in depth will lead to my question:
If I submit a jcl to execute a program, i will get a job step task
attached by initiator. If my program is authorized it can in turn attach
some extra job step tasks. Suppose they all try to GETMAIN from Private
Low subpools, is it true that the total amount mustnot exceed the limit
set by REGION parm? If it attaches normal subtasks, not job step tasks,
will the answer be different?
In other words, does the limit set by REGION parm include all private
low storage GETMAINed by both mother job step task and its subtasks? I
guess the answer is yes or we can attach a subtask to GETMAIN thus
easily break the rule. However, if the subtasks are all job step tasks,
is it still true?
You can see that what is really troubling me is the concept of 'job step
task'. I know it's not common practice to play with job step tasks. But
it's definitely there and there are actually several job tasks in an
address space. (Initiator itself is a job step task, right?)
I cannot find too much information about job step task in manuals. Maybe
it's like Chris Craddock said: There are just too many deep seated
job-steprelated assumptions within z/OS and its ancestors.
>From what I know:
1. To get a job step task, you first need to have an authorized job step
task. It's interesting, like the classic egg-and-chicken problem.
Anyway, in Z/OS we can get our 'first-level' job step task using START
and LOGON command, and a batch job. And they all require a jcl where the
REGION parm can be specified. Don't know whether extra interfaces
exist.
2. A job step task has its own TIOT and JSCB. And its TCBJSTCB points to
itself. This is the biggest difference I know between a job step task
and a normal task. The result is that a job step task can get more
Independence as far as the its relationship with its mother is
concerned.
Thanks for reading my lengthy mail.
Johnny
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send
email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search
the archives at http://bama.ua.edu/archives/ibm-main.html
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
