Did you compare the memory usage in the JESJCLMSG sysout of both STC and JOB?


Best Regards

Ituriel do Nascimento Neto
z/OS System Programmer






Em terça-feira, 28 de janeiro de 2025 às 10:46:18 BRT, rpinion865 
<[email protected]> escreveu: 





I did a $DEXIT(*), and all of our exits have ROUTINES=(),STATUS=DESABLED.  So 
no JES2
modifications.  



"Confidentially doc, I am the wabbit."

Bugs Bunny

Sent with Proton Mail secure email.

On Tuesday, January 28th, 2025 at 8:35 AM, David Spiegel 
<[email protected]> wrote:

> Hi Rich,
> Do you have any JES Exits which modify submitted JCL and/or Internal Text?
> 
> Regards,
> David
> 
> On 2025-01-28 07:58, rpinion865 wrote:
> 
> > We have the default IEFUSI, compile date of 05/13/1990, and IBM's 
> > Application Performance Analyzer IEFUSI.
> > I looked at SYS1.SAMPLIB(IEEUSI) and I do not think that code does anything 
> > different for a batch job vs.
> > a STC. The SMFPRMxx parmlib member (and no it has not changed) has 
> > MEMLIMIT(8G).
> > 
> > "Confidentially doc, I am the wabbit."
> > 
> > Bugs Bunny
> > 
> > Sent with Proton Mail secure email.
> > 
> > On Tuesday, January 28th, 2025 at 7:38 AM, Ituriel do Neto 
> > [email protected] wrote:
> > 
> > > The JCL statement REGION overrides the one defined at JOBCLASS.
> > > Do you have an IEFUSI exit? There may be differences related to REGION 
> > > definition because of LSQA settings for JOB and STC.
> > > 
> > > Best Regards
> > > 
> > > Ituriel do Nascimento Neto
> > > z/OS System Programmer
> > > 
> > > Em terça-feira, 28 de janeiro de 2025 às 09:28:46 BRT, rpinion865 
> > > [email protected] escreveu:
> > > 
> > > I forgot to add, the VPS PROC has REGION=0M. The JES2 STC JOBCLASS has 
> > > REGION=4M.
> > > When I ran VPS as a batch job, I had REGION=0M coded on my job statement. 
> > > Does
> > > anyone know if the JES2 STC JOBCLASS REGION would override what is coded 
> > > on the EXEC
> > > statement of a STC?
> > > 
> > > I will dig around and see if I have any programs that report on SMF type 
> > > 30 records.
> > > I would need something that would show region statistics. Perhaps, that 
> > > would show
> > > VPS region usage over the last couple of months.
> > > 
> > > "Confidentially doc, I am the wabbit."
> > > 
> > > Bugs Bunny
> > > 
> > > Sent with Proton Mail secure email.
> > > 
> > > On Tuesday, January 28th, 2025 at 7:22 AM, rpinion865 
> > > [email protected] wrote:
> > > 
> > > > Since I am the z/OS system programmer, I can say with 99% confidence, 
> > > > nothing has changed.
> > > > But I have learned the hard way, never say never. I did not make any 
> > > > parmlib or JES2 parm
> > > > changes. As for the IPL, it was the scheduled monthly IPL (not my 
> > > > decision, that is
> > > > way it has been for many, many years, predating my start date of May 
> > > > 2022).
> > > > 
> > > > LRS had us run a GTF trace and send them the output. After analyzing 
> > > > the GTF trace and
> > > > VPS logs, they suggested we remove printer definitions for printers 
> > > > that were no longer
> > > > present. Again, this company has a bad habit of maintaining 
> > > > pre-historic artifacts.
> > > > After the dead printers were removed, VPS, as a STC, came up and is 
> > > > stable.
> > > > 
> > > > "Confidentially doc, I am the wabbit."
> > > > 
> > > > Bugs Bunny
> > > > 
> > > > Sent with Proton Mail secure email.
> > > > 
> > > > On Tuesday, January 28th, 2025 at 1:55 AM, Brian Westerman 
> > > > [email protected] wrote:
> > > > 
> > > > > Someone could have changed the location of some control blocks in 
> > > > > parmlib and now they are in 64bit areas. Or someone could have added 
> > > > > the TIOT constraint relief stuff or any of several other "minor" parm 
> > > > > changes that upset where VPS is able to touch things. It could also 
> > > > > be that the VPS software itself was changed with some maintenance.
> > > > > 
> > > > > Since it's a STC vs JOB thing, then it could be that someone made a 
> > > > > change to the JES2PARMS and moved some of the constraint relief stuff 
> > > > > around. Probably SWA.
> > > > > 
> > > > > The good news is that VPS can run as a batch job without any reall 
> > > > > issues, but you do need to make sure that you code TIME=1440 on the 
> > > > > execute card or jobcard.
> > > > > 
> > > > > The A03 is probably because one of the internal VPS tasks has abended 
> > > > > before it detached it's subtasks. The S0C4 will probably be pointing 
> > > > > to some area above the line or the 64bit line and your program is 
> > > > > likely either 24bit and can't see the 31 bot area or 31bit and cant 
> > > > > see the 64bit area.
> > > > > 
> > > > > So, you can run as a batch job for a while, but you need to run down 
> > > > > the changes to parmlib (could be DIAGxx), or JES2parms to see what 
> > > > > they changed.
> > > > > 
> > > > > Nobody IPLs just for the fun of it so someone cahnged something. If 
> > > > > it's been a while since the last one then lots of stuff could have 
> > > > > changed.
> > > > > 
> > > > > If you can't find the culprit, then restore your res volume from a 
> > > > > previous point in time (like from your previous IPL time) and you 
> > > > > will likely see that things run just fine, then you will need to 
> > > > > compare the parmlibs to see what changed. If it's a VPS change, then 
> > > > > look at the last update date on your VPS loadlibs, if thoe dates are 
> > > > > before the previous IPL, (the one where things still worked) then the 
> > > > > problem is outside VPS and you need to track down the varmint that 
> > > > > changed stuff on the res pack on you.
> > > > > 
> > > > > You might be able to find the culprit just by looking at what has 
> > > > > changed since the previous IPL in parmlib by looking at the dates on 
> > > > > the members.
> > > > > 
> > > > > Remember, if everyone says "nobody changed nothin", then someone is 
> > > > > not telling the truth. Something changed, software doesn't just break 
> > > > > for no reason.
> > > > > 
> > > > > The first thing that should tip you off that something changed was 
> > > > > the fact that you IPLed. There must have been a reason, and maybe 
> > > > > someone thought to slip some other changes in on you.
> > > > > 
> > > > > Look closely, you'll find it.
> > > > > 
> > > > > Brian
> > > > > 
> > > > > ----------------------------------------------------------------------
> > > > > For IBM-MAIN subscribe / signoff / archive access instructions,
> > > > > send email to [email protected] with the message: INFO IBM-MAIN

> > > > > ----------------------------------------------------------------------
> > > > > For IBM-MAIN subscribe / signoff / archive access instructions,
> > > > > send email to [email protected] with the message: INFO IBM-MAIN
> > > 
> > > ----------------------------------------------------------------------
> > > For IBM-MAIN subscribe / signoff / archive access instructions,
> > > send email to [email protected] with the message: INFO IBM-MAIN
> > > 
> > > ----------------------------------------------------------------------
> > > For IBM-MAIN subscribe / signoff / archive access instructions,
> > > send email to [email protected] with the message: INFO IBM-MAIN
> > > ----------------------------------------------------------------------
> > > For IBM-MAIN subscribe / signoff / archive access instructions,
> > > send email to [email protected] with the message: INFO IBM-MAIN
> 
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to