Re: Do you need a coupling facility to implement DVIPA?

[Timothy Sipples]

Here's a quote pulled from an older IBM redbook:

"Note: Dynamic VIPA and Sysplex Distributor capabilities do not rely on data 
stored in structures in the Coupling Facility. Therefore, they can be 
implemented using XCF communication without a Coupling Facility (also called 
Basic Sysplex connectivity)."

That's on Page 4 of this redbook:

https://www.redbooks.ibm.com/redbooks/pdfs/sg247800.pdf

But I also found this more modern reference in the IBM WebSphere Application 
Server for z/OS documentation:

https://www.ibm.com/docs/en/was-zos/9.0.5?topic=sysplex-distributor

Funny how we carried forward the z/OS 1.2 (and prior) caveat. That bit of text 
doesn't seem too important for Version 9.0.5 of WAS for z/OS, does it? :-)

—
Timothy Sipples
Senior Architect
Digital Assets, Industry Solutions, and Cybersecurity
IBM zSystems/LinuxONE, Asia-Pacific
sipp...@sg.ibm.com


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Peter Sylvester]

Hi David,

would you mind to explain a bit the input data? A single source that hassome separator to indicate a 
new member or so (similar to iebupdte input)? Or aset of PDSes? Or??


TIA

Peter

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[kekronbekron]

A legendary post from Sri, from 19 years ago is here to help!

https://www.mvsforums.com/helpboards/viewtopic.php?p=8260

Perhaps I was talking about IEBUPDTE - 
https://www.ibm.com/docs/en/zos/2.4.0?topic=examples-example-11-create-new-partitioned-data-set


- KB

--- Original Message ---
On Thursday, April 20th, 2023 at 9:30 AM, kekronbekron 
<02dee3fcae33-dmarc-requ...@listserv.ua.edu> wrote:


> Hi Paul,
> 
> I don't remember the exact thing here but there's a way to flatten PDS files 
> to a PS.
> It'll add things like ./ADDMEMBER or something at the top of each member's 
> content.
> 
> Then, the PS can be used to recreate a PDS. Don't know about PDSEs...
> 
> - KB
> 
> --- Original Message ---
> On Thursday, April 20th, 2023 at 8:02 AM, Paul Gilmartin 
> 042bfe9c879d-dmarc-requ...@listserv.ua.edu wrote:
> 
> 
> 
> > On Thu, 20 Apr 2023 02:13:08 +, kekronbekron wrote:
> > 
> > > How about IEBPTPCH? Write out the members as a huge flatfile, and insert 
> > > member building control cards wherever appropriate.
> > > Then build a PDS from that?
> > 
> > Will IEBPTPCH create a PDS(E)? The description and examples I see use
> > a PDS(E) only as input, not output.
> > 
> > Perhaps I didn't read the right example.
> > https://www.ibm.com/docs/en/zos/2.5.0?topic=program-iebptpch-examples
> > 
> > --
> > gil
> > 
> > --
> > For IBM-MAIN subscribe / signoff / archive access instructions,
> > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
> 
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[kekronbekron]

Hi Paul,

I don't remember the exact thing here but there's a way to flatten PDS files to 
a PS.
It'll add things like ./ADDMEMBER or something at the top of each member's 
content.

Then, the PS can be used to recreate a PDS. Don't know about PDSEs...

- KB

--- Original Message ---
On Thursday, April 20th, 2023 at 8:02 AM, Paul Gilmartin 
<042bfe9c879d-dmarc-requ...@listserv.ua.edu> wrote:


> On Thu, 20 Apr 2023 02:13:08 +, kekronbekron wrote:
> 
> > How about IEBPTPCH? Write out the members as a huge flatfile, and insert 
> > member building control cards wherever appropriate.
> > Then build a PDS from that?
> 
> Will IEBPTPCH create a PDS(E)? The description and examples I see use
> a PDS(E) only as input, not output.
> 
> Perhaps I didn't read the right example.
> https://www.ibm.com/docs/en/zos/2.5.0?topic=program-iebptpch-examples
> 
> 
> --
> gil
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Paul Gilmartin]

On Thu, 20 Apr 2023 02:13:08 +, kekronbekron wrote:

>How about IEBPTPCH? Write out the members as a huge flatfile, and insert 
>member building control cards wherever appropriate.
>Then build a PDS from that?
>
Will IEBPTPCH create a PDS(E)?  The description and examples I see use
a PDS(E) only as input, not output.

Perhaps I didn't read the right example.


-- 
gil

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[kekronbekron]

How about IEBPTPCH? Write out the members as a huge flatfile, and insert member 
building control cards wherever appropriate.
Then build a PDS from that?

- KB

--- Original Message ---
On Thursday, April 20th, 2023 at 6:46 AM, Paul Gilmartin 
<042bfe9c879d-dmarc-requ...@listserv.ua.edu> wrote:


> On Wed, 19 Apr 2023 18:49:14 -0500, Mark Zelden wrote:
> 
> > On Wed, 19 Apr 2023 13:44:39 -0400, David Spiegel wrote:
> > 
> > > Hi,
> > > I am writing a Rexx Exec (Batch TSO) to build 10,000+ members in a PDS
> > > (FB/159/27984).
> > > The average member size is 900 lines.
> > > 
> > > In terms of performance, which method should I use?:
> > > 1) ALLOC ... EXECIO DISKW
> > > 2) LMOPEN ... LMINIT ... LMPUT
> > > 3) Something else
> > > 4) Makes no difference
> > 
> > You didn't explain much else other than the desire to build the PDS.
> > 
> > How important is the actual performance / wall clock time? Is it one off? 
> > Will it run daily?
> > Hourly? Every 15 minutes?
> > 
> > I would just use whatever method you are most comfortable with depending on
> > that answer. But if you are already using REXX, I don't know why you would 
> > then
> > use ISPF services on top of that. When I've done similar things I've used
> > ALLOC / EXECIO.
> 
> ISPF services can avoid 99.99% of the ALLOC overhead.
> 
> Kolusu's suggestion depends on PDSE for concurrency. Is that acceptable,
> or is PDS a requirement.
> 
> What about a UNIX directory as an alternative? No ENQ entanglement.
> 
> --
> gil
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Do you need a coupling facility to implement DVIPA?

[Laurence Chiu]

That is what I thought but I get confused since whenever I  look at DVIPA
discussions like this

https://www.ibm.com/docs/en/zos-basic-skills?topic=addressing-unique-application-instance-dvipa

it always shows a CF.

But in this preso from IBM at Share

https://share.confex.com/share/125/webprogram/Handout/Session17284/Sysplex%20and%20Network%20Technologies%20and%20Considerations%20%28Summer%202015%29.pdf

there is no mention of a CF anywhere in the deck.

Looks like it just uses XCF to communicate the status of the DVIPA stack
between the TCP/IP stacks on each LPAR as you say.

On Thu, Apr 20, 2023 at 1:24 PM Mark Jacobs <
0224d287a4b1-dmarc-requ...@listserv.ua.edu> wrote:

> Not 100% sure, but based on this, likely not.
>
> In order for TCP/IP to communicate DVIPA status among LPARs, TCP/IP uses
> an XCF group called EZBTCPCS. There's no mention of a coupling facility
> structure as a requirement.
>
> Mark Jacobs
>
> Sent from ProtonMail, Swiss-based encrypted email.
>
> GPG Public Key -
> https://api.protonmail.ch/pks/lookup?op=get&search=markjac...@protonmail.com
>
>
> --- Original Message ---
> On Wednesday, April 19th, 2023 at 8:53 PM, Laurence Chiu 
> wrote:
>
>
> > Looking at implementing a basic sysplex where an application will be able
> > to run on two different LPARs but only on one instance at a time.
> >
> > Since it's difficult to change the advertised IP for the application/LPAR
> > we plan to use DVIPA.
> >
> > Is this possible without a coupling facility and only the TCP/IP stake
> and
> > XCF?
> >
> > --
> > For IBM-MAIN subscribe / signoff / archive access instructions,
> > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
>
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
>

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Do you need a coupling facility to implement DVIPA?

[Mark Jacobs]

Not 100% sure, but based on this, likely not.

In order for TCP/IP to communicate DVIPA status among LPARs, TCP/IP uses an XCF 
group called EZBTCPCS. There's no mention of a coupling facility structure as a 
requirement. 

Mark Jacobs 

Sent from ProtonMail, Swiss-based encrypted email.

GPG Public Key - 
https://api.protonmail.ch/pks/lookup?op=get&search=markjac...@protonmail.com


--- Original Message ---
On Wednesday, April 19th, 2023 at 8:53 PM, Laurence Chiu  
wrote:


> Looking at implementing a basic sysplex where an application will be able
> to run on two different LPARs but only on one instance at a time.
> 
> Since it's difficult to change the advertised IP for the application/LPAR
> we plan to use DVIPA.
> 
> Is this possible without a coupling facility and only the TCP/IP stake and
> XCF?
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Paul Gilmartin]

On Wed, 19 Apr 2023 18:49:14 -0500, Mark Zelden wrote:

>On Wed, 19 Apr 2023 13:44:39 -0400, David Spiegel wrote:
>
>>Hi,
>>I am writing a Rexx Exec (Batch TSO) to build 10,000+ members in a PDS
>>(FB/159/27984).
>>The average member size is 900 lines.
>>
>>In terms of performance, which method should I use?:
>>1) ALLOC ... EXECIO DISKW
>>2) LMOPEN ... LMINIT ... LMPUT
>>3) Something else
>>4) Makes no difference
>
>
>You didn't explain much else other than the desire to build the PDS.   
>
>How important is the actual performance / wall clock time? Is it one off?  
>Will it run daily? 
>Hourly?  Every 15 minutes?
>
>I would just use whatever method you are most comfortable with depending on
>that answer.  But if you are already using REXX, I don't know why you would 
>then 
>use ISPF services on top of that.  When I've done similar things I've used
>ALLOC / EXECIO.   
> 
ISPF services can avoid 99.99% of the ALLOC overhead.

Kolusu's suggestion depends on PDSE for concurrency.  Is that acceptable,
or is PDS a requirement.

What about a UNIX directory as an alternative?  No ENQ entanglement.

-- 
gil

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Do you need a coupling facility to implement DVIPA?

[Laurence Chiu]

Looking at implementing a basic sysplex where an application will be able
to run on two different LPARs but only on one instance at a time.

Since it's difficult to change the advertised IP for the application/LPAR
we plan to use DVIPA.

Is this possible without a coupling facility and only the TCP/IP stake and
XCF?

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Mark Zelden]

On Wed, 19 Apr 2023 13:44:39 -0400, David Spiegel  
wrote:

>Hi,
>I am writing a Rexx Exec (Batch TSO) to build 10,000+ members in a PDS
>(FB/159/27984).
>The average member size is 900 lines.
>
>In terms of performance, which method should I use?:
>1) ALLOC ... EXECIO DISKW
>2) LMOPEN ... LMINIT ... LMPUT
>3) Something else
>4) Makes no difference
>
>Thank you in advance
>
>Regards,
>David
>


You didn't explain much else other than the desire to build the PDS.   

How important is the actual performance / wall clock time? Is it one off?  Will 
it run daily? 
Hourly?  Every 15 minutes?

I would just use whatever method you are most comfortable with depending on
that answer.  But if you are already using REXX, I don't know why you would 
then 
use ISPF services on top of that.  When I've done similar things I've used
ALLOC / EXECIO.   

Regards,

Mark
--
Mark Zelden - Zelden Consulting Services - z/OS, OS/390 and MVS
ITIL v3 Foundation Certified
mailto:m...@mzelden.com
Mark's MVS Utilities: http://www.mzelden.com/mvsutil.html

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[rpinion865]

Please send me your email address.  My email address is 
rpinion...@protonmail.com




Sent with Proton Mail secure email.

--- Original Message ---
On Wednesday, April 19th, 2023 at 2:54 PM, David Spiegel 
<0468385049d1-dmarc-requ...@listserv.ua.edu> wrote:


> Hi Richard,
> I would be interested in obtaining a copy of this.
> 
> Thanks and regards,
> David
> 
> On 2023-04-19 14:41, rpinion865 wrote:
> 
> > I have a very old assembler program. It was written to extract job log 
> > information from a JES external writer output dataset, and create a member 
> > for each job.
> > 
> > GENEDIT CSECT
> > * *
> > ***
> > * *
> > * PROGRAM DESCRIPTION *
> > * *
> > * THIS PROGRAM READS A SEQUENTIAL DATA SET THAT WAS CREATED BY AN *
> > * EXTERNAL WRITER. THE SEQUENTIAL DATA SET CONTAINS CLASS H OUTPUT *
> > * FROM THE PREVIOUS NIGHT'S PRODUCTION JOBS. THE PROGRAM CREATES A *
> > * PDS MEMBER FOR EACH JOB. IF THERE ARE DUPLICATE JOBS, SECONDARY *
> > * MEMBERS ARE CREATED FROM THE FIRST SIX CHARACTERS OF THE JOB NAME *
> > * PLUS A TWO DIGIT COUNT OF THE NUMBER OF TOTAL SECONDARY MEMBER *
> > * NAMES. THE PROGRAM ABENDS WITH A USER ONE, IF A RETURN CODE *
> > * FROM STOW IS GREATER THAN 4. THE PDS CREATED HAS LRECL OF 133, *
> > * BLKSIZE OF 0, AND RECFM OF FB. *
> > * *
> > ***
> > * *
> > 
> > Sent with Proton Mail secure email.
> > 
> > --- Original Message ---
> > On Wednesday, April 19th, 2023 at 2:38 PM, Paul Gilmartin 
> > 042bfe9c879d-dmarc-requ...@listserv.ua.edu wrote:
> > 
> > > On Wed, 19 Apr 2023 14:21:58 -0400, David Spiegel wrote:
> > > 
> > > > If I understand correctly ...
> > > > "ALLOC F(X) DA('MY.PDS') SH REU"
> > > > 
> > > > Do i=1 to n
> > > > "ALLOC F(Y) DA('MY.PDS('member')') SH REU
> > > > "EXECIO * DISKW (FINIS
> > > > End
> > > > 
> > > > Is this correct?
> > > > That's what I remember. But it might be better to remove the ALLOC and 
> > > > EXECIO from
> > > > the loop and do
> > > > LMINIT DDNAME(X) /* outside the loop. /
> > > > then LMPUT, etc. / inside. */
> > > 
> > > LMPUT requires exclusive allocation, not SHR.
> > > It's infuriating that ISPF has no support for REXX compould symbols,
> > > instead providing the bizarre MULTX form.
> > > 
> > > IBM doesn't care.
> > > 
> > > --
> > > gil
> > > 
> > > --
> > > For IBM-MAIN subscribe / signoff / archive access instructions,
> > > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
> > > --
> > > For IBM-MAIN subscribe / signoff / archive access instructions,
> > > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
> 
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[David Spiegel]

Hi Richard,
I would be interested in obtaining a copy of this.

Thanks and regards,
David

On 2023-04-19 14:41, rpinion865 wrote:

I have a very old assembler program.  It was written to extract job log 
information from a JES external writer output dataset, and create a member for 
each job.

GENEDIT  CSECT
* *
***
* *
*PROGRAM DESCRIPTION  *
* *
*   THIS PROGRAM READS A SEQUENTIAL DATA SET THAT WAS CREATED BY AN   *
*   EXTERNAL WRITER.  THE SEQUENTIAL DATA SET CONTAINS CLASS H OUTPUT *
*   FROM THE PREVIOUS NIGHT'S PRODUCTION JOBS.  THE PROGRAM CREATES A *
*   PDS MEMBER FOR EACH JOB.  IF THERE ARE DUPLICATE JOBS, SECONDARY  *
*   MEMBERS ARE CREATED FROM THE FIRST SIX CHARACTERS OF THE JOB NAME *
*   PLUS A TWO DIGIT COUNT OF THE NUMBER OF TOTAL SECONDARY MEMBER*
*   NAMES.  THE PROGRAM ABENDS WITH A USER ONE, IF A RETURN CODE  *
*   FROM STOW IS GREATER THAN 4.  THE PDS CREATED HAS LRECL OF 133,   *
*   BLKSIZE OF 0, AND RECFM OF FB.*
* *
***
* *




Sent with Proton Mail secure email.

--- Original Message ---
On Wednesday, April 19th, 2023 at 2:38 PM, Paul Gilmartin 
<042bfe9c879d-dmarc-requ...@listserv.ua.edu> wrote:



On Wed, 19 Apr 2023 14:21:58 -0400, David Spiegel wrote:


If I understand correctly ...
"ALLOC F(X) DA('MY.PDS') SH REU"

Do i=1 to n
"ALLOC F(Y) DA('MY.PDS('member')') SH REU
"EXECIO * DISKW (FINIS
End

Is this correct?

That's what I remember. But it might be better to remove the ALLOC and EXECIO 
from
the loop and do
LMINIT DDNAME(X) /* outside the loop. /
then LMPUT, etc. / inside. */

LMPUT requires exclusive allocation, not SHR.
It's infuriating that ISPF has no support for REXX compould symbols,
instead providing the bizarre MULTX form.

IBM doesn't care.

--
gil

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Sri h Kolusu]

>> Can you please supply an example, or, point me to a  website that has one?

David,

Here is a sample that builds 10 members at a time with each member having 900 
lines of data.  You can expand that to 1000 members and then have 10 jobs 
running at the same time to write out 10,000 members to the PDSE

//***
//* Delete the output PDSE if it exist  *
//***
//DELFILES EXEC PGM=IEFBR14
//FILE01   DD DSN=&SYSUID..MYPDSE,
//DISP=(MOD,DELETE,DELETE),
//SPACE=(TRK,(1,0),RLSE)
/*
//***
//* Allocate the PDSE   *
//***
//DELFILES EXEC PGM=IEFBR14
//FILE01   DD DSN=&SYSUID..MYPDSE,
//DISP=(NEW,CATLG,DELETE),
//SPACE=(CYL,(75,75,1),RLSE),
//DCB=(LRECL=159,RECFM=FB),
//DSNTYPE=LIBRARY
/*
//***
//* create the pdse members *
//***
//CREMEM   EXEC PGM=SORT
//SYSOUT   DD SYSOUT=*
//SORTIN   DD *
ABC
//FDD DISP=SHR,DSN=&SYSUID..MYPDSE(M)
//F0001DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0001)
//F0002DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0002)
//F0003DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0003)
//F0004DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0004)
//F0005DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0005)
//F0006DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0006)
//F0007DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0007)
//F0008DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0008)
//F0009DD DISP=SHR,DSN=&SYSUID..MYPDSE(M0009)
//SYSINDD *
  OPTION COPY,ODMAXBF=100K
  OUTFIL REPEAT=900,
  FNAMES=(F,F0001,F0002,F0003,F0004,
  F0005,F0006,F0007,F0008,F0009),
  BUILD=(C'Record number : ',SEQNUM,8,FS,159:X)
/*



Thanks,
Kolusu
DFSORT Development
IBM Corporation


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[rpinion865]

I have a very old assembler program.  It was written to extract job log 
information from a JES external writer output dataset, and create a member for 
each job.  

GENEDIT  CSECT   
* *  
***  
* *  
*PROGRAM DESCRIPTION  *  
* *  
*   THIS PROGRAM READS A SEQUENTIAL DATA SET THAT WAS CREATED BY AN   *  
*   EXTERNAL WRITER.  THE SEQUENTIAL DATA SET CONTAINS CLASS H OUTPUT *  
*   FROM THE PREVIOUS NIGHT'S PRODUCTION JOBS.  THE PROGRAM CREATES A *  
*   PDS MEMBER FOR EACH JOB.  IF THERE ARE DUPLICATE JOBS, SECONDARY  *  
*   MEMBERS ARE CREATED FROM THE FIRST SIX CHARACTERS OF THE JOB NAME *  
*   PLUS A TWO DIGIT COUNT OF THE NUMBER OF TOTAL SECONDARY MEMBER*  
*   NAMES.  THE PROGRAM ABENDS WITH A USER ONE, IF A RETURN CODE  *  
*   FROM STOW IS GREATER THAN 4.  THE PDS CREATED HAS LRECL OF 133,   *  
*   BLKSIZE OF 0, AND RECFM OF FB.*  
* *  
***  
* *  




Sent with Proton Mail secure email.

--- Original Message ---
On Wednesday, April 19th, 2023 at 2:38 PM, Paul Gilmartin 
<042bfe9c879d-dmarc-requ...@listserv.ua.edu> wrote:


> On Wed, 19 Apr 2023 14:21:58 -0400, David Spiegel wrote:
> 
> > If I understand correctly ...
> > "ALLOC F(X) DA('MY.PDS') SH REU"
> > 
> > Do i=1 to n
> > "ALLOC F(Y) DA('MY.PDS('member')') SH REU
> > "EXECIO * DISKW (FINIS
> > End
> > 
> > Is this correct?
> 
> That's what I remember. But it might be better to remove the ALLOC and EXECIO 
> from
> the loop and do
> LMINIT DDNAME(X) /* outside the loop. /
> then LMPUT, etc. / inside. */
> 
> LMPUT requires exclusive allocation, not SHR.
> It's infuriating that ISPF has no support for REXX compould symbols,
> instead providing the bizarre MULTX form.
> 
> IBM doesn't care.
> 
> --
> gil
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Paul Gilmartin]

On Wed, 19 Apr 2023 14:21:58 -0400, David Spiegel wrote:
>
>If I understand correctly ...
>"ALLOC F(X) DA('MY.PDS') SH REU"
>
>Do i=1 to n
>    "ALLOC F(Y) DA('MY.PDS('member')') SH REU
>    "EXECIO * DISKW (FINIS
>End
>
>Is this correct?
>
That's what I remember.  But it might be better to remove the ALLOC and EXECIO 
from
the loop and do
LMINIT DDNAME(X)  /* outside the loop.  */
then LMPUT, etc.   /* inside.  */

LMPUT requires exclusive allocation, not SHR.
It's infuriating that ISPF has no support for REXX compould symbols,
instead providing the bizarre MULTX form.

IBM doesn't care.

-- 
gil

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[David Spiegel]

Hi Kolusu,
Can you please supply an example, or, point me to a  website that has one?

Thanks and regards,
David

On 2023-04-19 14:03, Sri h Kolusu wrote:

3) Something else

DFSORT can do simultaneous member write to PDSE (not PDS) and it will haeve 
better performance than rexx.

Thanks,
Kolusu

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[David Spiegel]

Hi Gil,
If I understand correctly ...
"ALLOC F(X) DA('MY.PDS') SH REU"

Do i=1 to n
   "ALLOC F(Y) DA('MY.PDS('member')') SH REU
   "EXECIO * DISKW (FINIS
End

Is this correct?

Thanks and regards,
David

On 2023-04-19 14:10, Paul Gilmartin wrote:

On Wed, 19 Apr 2023 13:44:39 -0400, David Spiegel wrote:

I am writing a Rexx Exec (Batch TSO) to build 10,000+ members in a PDS
(FB/159/27984).
The average member size is 900 lines.

In terms of performance, which method should I use?:
1) ALLOC ... EXECIO DISKW


Performance is likely to be dominated by allocation overhead.
It was once reported here that allocating another DDNAME to the same data set
*and*leaving*it*allocated* throughout the process improves performance by
shortcutting ENQ overhead.


2) LMOPEN ... LMINIT ... LMPUT


Allocate only once; LM OPEN  by DDNAME, not DSN.


3) Something else
4) Makes no difference


What's the source of your data?  Isn't there a utility that will
create multiple PDS members from a single PS input?



--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Charles Mills]

Not the answer to the question you asked but ...

"Performance" has at least two definitions: least elapsed wall clock time and 
least CPU utilization.

Another option would be BPXWDYN ... EXECIO DISKW.

Charles

On Wed, 19 Apr 2023 13:44:39 -0400, David Spiegel  
wrote:

>Hi,
>I am writing a Rexx Exec (Batch TSO) to build 10,000+ members in a PDS
>(FB/159/27984).
>The average member size is 900 lines.
>
>In terms of performance, which method should I use?:
>1) ALLOC ... EXECIO DISKW
>2) LMOPEN ... LMINIT ... LMPUT
>3) Something else
>4) Makes no difference

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Paul Gilmartin]

On Wed, 19 Apr 2023 13:44:39 -0400, David Spiegel wrote:
>
>I am writing a Rexx Exec (Batch TSO) to build 10,000+ members in a PDS
>(FB/159/27984).
>The average member size is 900 lines.
>
>In terms of performance, which method should I use?:
>1) ALLOC ... EXECIO DISKW
>
Performance is likely to be dominated by allocation overhead.
It was once reported here that allocating another DDNAME to the same data set
*and*leaving*it*allocated* throughout the process improves performance by
shortcutting ENQ overhead.

>2) LMOPEN ... LMINIT ... LMPUT
>
Allocate only once; LM OPEN  by DDNAME, not DSN.

>3) Something else
>4) Makes no difference
>
What's the source of your data?  Isn't there a utility that will
create multiple PDS members from a single PS input?

-- 
gil

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[Sri h Kolusu]

>>3) Something else

DFSORT can do simultaneous member write to PDSE (not PDS) and it will haeve 
better performance than rexx.

Thanks,
Kolusu

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Rexx Exec to Build 10,000 PDS Members: ALLOC vs ISPF Services vs ?

[David Spiegel]

Hi,
I am writing a Rexx Exec (Batch TSO) to build 10,000+ members in a PDS 
(FB/159/27984).

The average member size is 900 lines.

In terms of performance, which method should I use?:
1) ALLOC ... EXECIO DISKW
2) LMOPEN ... LMINIT ... LMPUT
3) Something else
4) Makes no difference

Thank you in advance

Regards,
David

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: JESLOG Parameter on JOB Card, executing SUB=MSTR

[Mark Jacobs]

I opened an idea on the IBM Idea Portal for this and it was accepted as a 
"Future Consideration." For anyone who wants to vote for it, it's ZOS-I-3641.

Mark Jacobs 

Sent from ProtonMail, Swiss-based encrypted email.

GPG Public Key - 
https://api.protonmail.ch/pks/lookup?op=get&search=markjac...@protonmail.com


--- Original Message ---
On Saturday, April 8th, 2023 at 8:09 AM, Steve Horein  
wrote:


> FWIW, I will likely create some "Extended Status" pseudo messages UP_JES3
> (NetView is UP, JES3 just reached UP status) and ISUP_JES3 (NetView
> restarted, JES3 was already UP) to issue the desired command once per life
> of NetView opposed to the "every night at 22:00" approach currently in
> place.
> 
> https://www.ibm.com/docs/en/z-system-automation/4.3.0?topic=messages-pseudo-invoke-extended-status-commands
> 
> 
> "Don't change what ain't broke... is a great way to stand still."
> 
> On Fri, Apr 7, 2023 at 5:54 PM Mark Jacobs <
> 0224d287a4b1-dmarc-requ...@listserv.ua.edu> wrote:
> 
> > I'll try all that on our sandbox system. Thanks all.
> > 
> > Mark Jacobs
> > 
> > Sent from ProtonMail, Swiss-based encrypted email.
> > 
> > GPG Public Key -
> > https://api.protonmail.ch/pks/lookup?op=get&search=markjac...@protonmail.com
> > 
> > --- Original Message ---
> > On Friday, April 7th, 2023 at 5:21 PM, Steve Horein <
> > steve.hor...@gmail.com> wrote:
> > 
> > > I meant to include that I didn't find the doc very clear relating to:
> > > $T REQJOBID,JESLOG=SUPPRESS
> > > 
> > > Is "Jobid" the JES2 number, or can it be a name, or is it literal and
> > > touch
> > > "all" jobs?
> > > 
> > > On Fri, Apr 7, 2023 at 4:18 PM Steve Horein steve.hor...@gmail.com
> > > wrote:
> > > 
> > > > Okay, I logged onto to work to see what was happening.
> > > > 
> > > > JES3 shop. Every night at 22:00, the following command is issued:
> > > > *F J=NETVIEW,NOLOG,N=ALL
> > > > 
> > > > I don't know if this is an equivalent command for JES2 (
> > 
> > https://www.ibm.com/docs/en/zos/2.4.0?topic=section-t-reqjobid-set-jobid-attributes
> > 
> > > > ):
> > > > $T REQJOBID,JESLOG=SUPPRESS
> > > > or maybe (
> > 
> > https://www.ibm.com/docs/en/zos/2.4.0?topic=crs-t-job-change-jobs-class-scheduling-priority-affinity
> > 
> > > > ):
> > > > $T J,NETVIEW,SPIN
> > > > 
> > > > On Fri, Apr 7, 2023 at 1:21 PM Mark Jacobs <
> > > > 0224d287a4b1-dmarc-requ...@listserv.ua.edu> wrote:
> > > > 
> > > > > I just tried that. No difference in what I'm observing in JESMSGLG.
> > > > > 
> > > > > Mark Jacobs
> > > > > 
> > > > > Sent from ProtonMail, Swiss-based encrypted email.
> > > > > 
> > > > > GPG Public Key -
> > 
> > https://api.protonmail.ch/pks/lookup?op=get&search=markjac...@protonmail.com
> > 
> > > > > --- Original Message ---
> > > > > On Friday, April 7th, 2023 at 2:13 PM, Steve Horein <
> > > > > steve.hor...@gmail.com> wrote:
> > > > > 
> > > > > > We do START NETVIEW,SUB=MSTR,MSGLEVEL=(0,0) for NetView.
> > > > > > I have not tried omitting the MSGLEVEL parameter to see if it
> > > > > > really
> > > > > > makes
> > > > > > a difference or not.
> > > > > > 
> > > > > > On Fri, Apr 7, 2023 at 1:01 PM Mark Jacobs <
> > > > > > 0224d287a4b1-dmarc-requ...@listserv.ua.edu> wrote:
> > > > > > 
> > > > > > > The problem is with our Netview STCs. They execute under MSTR
> > > > > > > and on
> > > > > > > our
> > > > > > > GDPS systems can stay up for months and months at a time. Their
> > > > > > > JESMSGLG
> > > > > > > can grow to several million lines which consumes a
> > > > > > > non-insignificant
> > > > > > > amount
> > > > > > > of spool space. I was hoping to add JESLOG=SUPPRESS to its job
> > > > > > > card to
> > > > > > > suppress this unneeded data. It didn't work. I opened an idea on
> > > > > > > IBM's idea
> > > > > > > portal for future support.
> > > > > > > 
> > > > > > > Mark Jacobs
> > > > > > > 
> > > > > > > Sent from ProtonMail, Swiss-based encrypted email.
> > > > > > > 
> > > > > > > GPG Public Key -
> > 
> > https://api.protonmail.ch/pks/lookup?op=get&search=markjac...@protonmail.com
> > 
> > > > > > > --- Original Message ---
> > > > > > > On Friday, April 7th, 2023 at 1:50 PM, Steve Thompson
> > > > > > > ste...@wkyr.net
> > > > > > > wrote:
> > > > > > > 
> > > > > > > > I think it depends on the started task first issuing a Request
> > > > > > > > JOB ID CALL to JES. Then you get minimal JES support. At that
> > > > > > > > point the JOBLOG= may possibly have meaning. Take this with
> > > > > > > > some
> > > > > > > > salt, because I haven't done this level of programming for many
> > > > > > > > years and do not currently have the ability to try this on the
> > > > > > > > system I am using.
> > > > > > > > 
> > > > > > > > Otherwise I think you have to then issue the equivalent call to
> > > > > > > > JES for this.
> > > > > > > > 
> > > > > > > > Steve Thompson
> > > > > > > > 
> > > > > > > > On 4/7/2023 1:18 PM, Mark Jacobs wrote:
> > > > > > > > 
> > > >

Re: PKCS #7

[Pierre Fichaud]

I wrote a certificate management package some years ago on z/OS. I wrote the 
package in C using the gsk functions.
IBM stated that the functions couldn't be called directly in assembler.
Regards, Pierre.

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: How to remove an empty USS file shipped with z/OS

[Dave Jousma]

On Mon, 17 Apr 2023 09:01:21 -0700, Lizette Koehler  
wrote:

>List -
>
>
>
>I am being asked to take one of two actions on a USS file that is part of
>Server Pac
>
>the /etc/security seems to be part of Server pac
>
>
>

So, I don’t know about you, but the /etc that comes from serverpac is not 
mounted, and I treat as guidelines on what needs to be changed/updated in my 
/etc/ that is actually mounted and in use.   So, unless you are actually using 
Serverpac provided /etc you should be free to do with as you see fit.


--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: If You Had Invested $10,000 In IBM In 2013, This Is How Much You Would Have Today

[zMan]

Kirk, the article talks about dividends at length; in fact, the subhed is:
Dividends have played a significant role in the stock's returns.

On Wed, Apr 19, 2023 at 9:02 AM Kirk Wolf  wrote:

> Bill is correct.  I would add that the OP also forgets this minor detail:
> https://www.marketbeat.com/stocks/NYSE/IBM/dividend/
>
> Kirk Wolf
> Dovetailed Technologies
> https://coztoolkit.com
>
> On Mon, Apr 17, 2023, at 9:58 AM, Bill Johnson wrote:
> > If you had invested 10,000 in Tesla 2 years ago, you’d now have 5,000.
> If you bought bitcoin in 2021, you lost half. If you purchased GM stock in
> 2006, you lost it all. If you bought a real bank like JPM in 2013, you
> tripled your money. Most of the “high growth” stocks that Silicon Valley
> has pushed on America as disruptors, like Uber, Chewy, don’t even make any
> money at all & their stocks are below the first trade price. Uber, who I
> drove for as a possible retirement hobby, (3 months lost $600) has lost
> over 10 billion dollars & the stock is significantly below the first trade
> price. IBM hasn’t been a great investment over the last 10, but you can
> always pick a time frame for every stock where it doesn’t perform. MSFT
> stock did nothing from 1999 til 2016.
> >
> >
> > Sent from Yahoo Mail for iPhone
> >
> >
> > On Monday, April 17, 2023, 10:20 AM, zMan 
> wrote:
> >
> >
> https://www.fool.com/investing/2023/04/17/if-you-had-invested-1-in-ibm-in-2013-this-is-h/
> >
> > TL;DR: Less than $10,000.
> > --
> > zMan -- "I've got a mainframe and I'm not afraid to use it"
> >
> > --
> > For IBM-MAIN subscribe / signoff / archive access instructions,
> > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
> >
> >
> >
> >
> > --
> > For IBM-MAIN subscribe / signoff / archive access instructions,
> > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
> >
>
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
>


-- 
zMan -- "I've got a mainframe and I'm not afraid to use it"

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Re: If You Had Invested $10,000 In IBM In 2013, This Is How Much You Would Have Today

[Kirk Wolf]

Bill is correct.  I would add that the OP also forgets this minor detail:
https://www.marketbeat.com/stocks/NYSE/IBM/dividend/

Kirk Wolf
Dovetailed Technologies
https://coztoolkit.com

On Mon, Apr 17, 2023, at 9:58 AM, Bill Johnson wrote:
> If you had invested 10,000 in Tesla 2 years ago, you’d now have 5,000. If you 
> bought bitcoin in 2021, you lost half. If you purchased GM stock in 2006, you 
> lost it all. If you bought a real bank like JPM in 2013, you tripled your 
> money. Most of the “high growth” stocks that Silicon Valley has pushed on 
> America as disruptors, like Uber, Chewy, don’t even make any money at all & 
> their stocks are below the first trade price. Uber, who I drove for as a 
> possible retirement hobby, (3 months lost $600) has lost over 10 billion 
> dollars & the stock is significantly below the first trade price. IBM hasn’t 
> been a great investment over the last 10, but you can always pick a time 
> frame for every stock where it doesn’t perform. MSFT stock did nothing from 
> 1999 til 2016.
> 
> 
> Sent from Yahoo Mail for iPhone
> 
> 
> On Monday, April 17, 2023, 10:20 AM, zMan  wrote:
> 
> https://www.fool.com/investing/2023/04/17/if-you-had-invested-1-in-ibm-in-2013-this-is-h/
> 
> TL;DR: Less than $10,000.
> -- 
> zMan -- "I've got a mainframe and I'm not afraid to use it"
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
> 
> 
> 
> 
> --
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
> 

--
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN