Joshua
I've solved this locally by arranging for the first PROC in my startup (well
before JES2) to run an IEBCOPY which copies all the parms in the appropriate
concatenation into a single PDSE as follows:
//LPARM PROC
//LPARM EXEC PGM=IEBCOPY,TIME=10
//P1 DD DISP=SHR,DSN=CAP.&SYSLIB1..PARMLIB
//P2 DD DISP=SHR,DSN=CAP.&SYSLIB2..PARMLIB
//P3 DD DISP=SHR,DSN=CAP.&SYSLIB3..PARMLIB
//*YSPRINT DD SYSOUT=*,DEST=HOLD
//SYSPRINT DD DISP=SHR,DSN=CAP1.IEBCOPY.LPARM
//POUT DD DISP=OLD,DSN=CAP.&SYSLIB1..PARMOUT
//SYSIN DD DISP=SHR,DSN=CAP.USYS1.PARMLIB(LPARMC1)
// PEND
&SYSLIB1..2 are set up in IEASYMxx members as follows:
For the live system:
SYSDEF SYSNAME(P390)
SYSCLONE(1A)
SYMDEF(&UNIXVER='V1R3M0')
SYMDEF(&SYSR2.='Z3RES2')
SYMDEF(&SYSLIB1='USYS1')
SYMDEF(&SYSLIB2='USYS1')
SYMDEF(&SYSLIB3='USYS1')
For the test system:
SYSDEF SYMDEF(&SYSLIB1='USER')
SYMDEF(&SYSLIB2='USYS1')
SYMDEF(&SYSLIB3='USYS1')
For a sysprog's development IPL system:
SYSDEF SYMDEF(&SYSLIB1='UJULIAN')
SYMDEF(&SYSLIB2='USER')
SYMDEF(&SYSLIB3='USYS1')
With a DISP=OLD on the output PDSE, JES2 cannot start until this relatively
short task has ended.
I then code //parm_ddn DD DISP=SHR,DSN=CAP.&SYSLIB1..PARMOUT for those
products that cannot handle the logical PARMLIB. For dynamic PARMLIB changes
I have to remember to copy changes to the PARMOUT dataset, which is a bit of
an inconvenience.
I control all of the PARMS/PROCs etc using SCLM. If I need to change a
PARM/PROC I just edit it, the changed version is saved into my private
PARM/PROCLIB. I can then test this without effecting test/live environments.
If my change fails I can just use the test or live IPL to correct the
problem and try again. In addition, its very easy to compare changes with
live, to help clarify where I stand with any projects I'm undertaking.
I inherited a system without any source (PARM/PARM) control/procedures, the
above works quite well for us.
I'm glad you asked this question. I've been meaning to ask the list if they
can see any flaws in my scheme. It does seem to work well in practice, but
I'm not the most experienced sysprog and maybe something could come along
and bite me that I haven't considered.
HTH
Julian
----------------------------------------------------------------------
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
