Discussion:
Custom Proclib for MVS?
Gary Lee Phillips
2007-06-03 20:49:04 UTC
Permalink
OK, I'm obviously getting rusty in this field. The environment
described here is Volker Bandke's MVS turnkey #3.

I wanted to create an additional SYSx.PROCLIB dataset, in which I
could put PROCs I wanted to test or whatever without adding them to
SYS1.PROCLIB and then making lots of changes that lead to a need to
compress that PDS. I do remember that getting control to compress it
can be a major hassle.

So, I created a new PDS called SYST.PROCLIB, with the same attributes
as SYS1.PROCLIB, and put one simple dummy member into it.

Then I restarted JES2, with P2=SYST so that the new proclib would
appear at the bottom of the PROC00 concatenation, preceded by SYS1...
and SYS2...

No other changes were made, just the substitution of one more PDS
where the default would use the SYS1 twice, but JES2 abends and dumps
to a system dump file. It still starts and works normally when left to
the default values.

After some headscratching to remember how to run AMDPRDMP, I got the
dump printed, but the abend code appears to be 0000. The program
terminated on an SVC 13, which I seem to recall is a deliberate abend,
but normally some sort of information is delivered either by pointing
the PSW at a string or code, or printing a message first.

Obviously, there's something wrong with the way I created the extra
proclib, but I can't figure out what I've forgotten to do. It's
probably something really obvious. Any ideas?
scott
2007-06-03 21:30:02 UTC
Permalink
Post by Gary Lee Phillips
OK, I'm obviously getting rusty in this field. The environment
described here is Volker Bandke's MVS turnkey #3.
I wanted to create an additional SYSx.PROCLIB dataset, in which I
could put PROCs I wanted to test or whatever without adding them to
SYS1.PROCLIB and then making lots of changes that lead to a need to
compress that PDS. I do remember that getting control to compress it
can be a major hassle.
So, I created a new PDS called SYST.PROCLIB, with the same attributes
as SYS1.PROCLIB, and put one simple dummy member into it.
Then I restarted JES2, with P2=SYST so that the new proclib would
appear at the bottom of the PROC00 concatenation, preceded by SYS1...
and SYS2...
No other changes were made, just the substitution of one more PDS
where the default would use the SYS1 twice, but JES2 abends and dumps
to a system dump file. It still starts and works normally when left to
the default values.
After some headscratching to remember how to run AMDPRDMP, I got the
dump printed, but the abend code appears to be 0000. The program
terminated on an SVC 13, which I seem to recall is a deliberate abend,
but normally some sort of information is delivered either by pointing
the PSW at a string or code, or printing a message first.
Obviously, there's something wrong with the way I created the extra
proclib, but I can't figure out what I've forgotten to do. It's
probably something really obvious. Any ideas?
Is the new proclib in the master catalog? If not, it should be. If
you don't want to catalog it in the master catalog, in the JES2 proc,
add a unit= and vol=ser= parameter on it.
Gary Lee Phillips
2007-06-03 21:43:56 UTC
Permalink
Post by scott
Is the new proclib in the master catalog? If not, it should be. If
you don't want to catalog it in the master catalog, in the JES2 proc,
add a unit= and vol=ser= parameter on it.
I knew it would be something just too obvious. Yes, of course that
must be it. The SYST high level qualifier is in a usercat, because
I've been using it for stuff loaded from various distribution tapes.
And I did know that datasets accessed when JES is just starting have
to be in the master catalog. I'll fix that, probably by renaming it to
SYS9 or something. Thanks.

--Gary
scott
2007-06-03 21:51:02 UTC
Permalink
Post by Gary Lee Phillips
Post by scott
Is the new proclib in the master catalog? If not, it should be. If
you don't want to catalog it in the master catalog, in the JES2 proc,
add a unit= and vol=ser= parameter on it.
I knew it would be something just too obvious. Yes, of course that
must be it. The SYST high level qualifier is in a usercat, because
I've been using it for stuff loaded from various distribution tapes.
And I did know that datasets accessed when JES is just starting have
to be in the master catalog. I'll fix that, probably by renaming it to
SYS9 or something. Thanks.
--Gary
No problem. We all slip-up now and then.

Loading...