View previous topic :: View next topic
|
Author |
Message |
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006 Posts: 6966 Location: porcelain throne
|
|
|
|
Manish,
would you have the jcl for the TMP step, please. |
|
Back to top |
|
|
Phrzby Phil
Senior Member
Joined: 31 Oct 2006 Posts: 1042 Location: Richmond, Virginia
|
|
|
|
Dick's question is the only question that should be answered now.
Please do so. |
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
personally I prefer having one run class unique and only associated with a single initiator. That way serial processing is assured. |
|
Back to top |
|
|
gcicchet
Senior Member
Joined: 28 Jul 2006 Posts: 1702 Location: Australia
|
|
|
|
Hi Dave,
what happens when a job fails, how do you prevent others from running ?
Gerry |
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
Still if u need 3 separate jobs instead of 1,
Another crude approach can be,
having a IKJEFT01 step at the end of the first job with COND=(4,LT) or some other COND to check if the job ran successfully and make it execute the tso command
Code: |
SUB 'YOUR.PDS.WHICH.HAS.THE.NEXTJOB' |
This step will trigger next job, if the triggering job completes successfully.
Hope it helps, |
|
Back to top |
|
|
Manish Kumar Gupta
New User
Joined: 19 Feb 2008 Posts: 6 Location: Bangalore
|
|
|
|
You can code the IKJEFT1B step like this.
Code: |
//STEPXX EXEC PGM=IKJEFT1B,COND=(0,NE)
//SYSPRINT DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SYSTSPRT DD SYSOUT=*
//SYSTSIN DD *
SUBMIT 'YOURPDSNAME(MEMNAME)'
/* |
|
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
Gerry,
First and foremost, Scheduler is best overall choice.
I've seen shops where each pgm calls subpgm prior to successful ending and posts status in a file. Subsequent pgms called common pgm to see if prev pgm was successful. File is usually VSAM and contains running order of jobs/pgms by name -- "the in-house scheduler approach"
My point was serialization, but I overlooked failure. Good catch? |
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1742 Location: Tirupur, India
|
|
|
|
If serialisation is required, then why restrict the initiator to 1? run all jobs at one go with same job name.. they will execute one by one.. |
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
Using same name may be okay in test, but unacceptable in production. |
|
Back to top |
|
|
PeterHolland
Global Moderator
Joined: 27 Oct 2009 Posts: 2481 Location: Netherlands, Amstelveen
|
|
|
|
dneufarth wrote: |
Using same name may be okay in test, but unacceptable in production. |
Not if you have the right agreements with production and/or some
friends there. At least that is my experience. |
|
Back to top |
|
|
dbzTHEdinosauer
Global Moderator
Joined: 20 Oct 2006 Posts: 6966 Location: porcelain throne
|
|
|
|
I have worked at shops (where in the non-prod environment)
where more than 2 submitted jobs under the same name (or user id) will cause the the 3rd and subsequent to fail.
in other words, only 2 jobs max at a time in job queue. |
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
Peter,
I agree, but in all shops I've worked Prod Job names had to be unique. All were under control of a Scheduler or may have triggers based on Job names submitted via CICS region (these were also unique). Those CICS submitted jobs not under Scheduler usually created GDG datasets.
Shop standards obviously vary.
I just happened to be the QA guy and involved with stds.
Everyone can 'boo' now! |
|
Back to top |
|
|
PeterHolland
Global Moderator
Joined: 27 Oct 2009 Posts: 2481 Location: Netherlands, Amstelveen
|
|
|
|
Well Dave im not boo'ing you. QA people like you deserve a place too
I could do those things because of my "support" work. |
|
Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006 Posts: 19244 Location: Inside the Matrix
|
|
|
|
Hello,
I suppose i am really slow this week. . .
So far i have seen no reason that this is more than a single job with multiple steps. . .
From a scheduling perspective, one job is trivial (i.e. without convolutions) to handle as a "single process". Restart/rerun is simplified in a single job. From a quality control or maintenance perspective, maintaining one job should be easier/cleaner than 3. What business reason (that someone "said so" is not a business reason) to inplement 3 jobs?
As i said - i'm slow this week, but i still do not understand . . . |
|
Back to top |
|
|
Marso
REXX Moderator
Joined: 13 Mar 2006 Posts: 1353 Location: Israel
|
|
|
|
madhu2010 wrote: |
First job's output file feeding to Second job, Second Job Output is input to 3rd job.
...
Also please let me know how it works with 1 job and 3 Execs. |
I think Madhu has nothing against 1 job with 3 steps,
maybe he is just lacking some basic knowledge...
Also, using one job allows the use of temporary files. |
|
Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006 Posts: 19244 Location: Inside the Matrix
|
|
|
|
Hi Marso,
Quote: |
I think Madhu has nothing against 1 job with 3 steps,
maybe he is just lacking some basic knowledge... |
Possibly. . .
Madhu,
If you are not familiar with multi-step jobs, you might look at the jcl that is submitted when you compile/link a program.
If there are questions, post them here and someone will be able to clarify. |
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
I was just suplementing one part of OP request for multiple jobs.
Single job may be okay and seems very applicable here unless it runs for hours and hours and involves update of VSAM datasets. Restart is not as simple in that circumstance.
We had std avoiding RESTART and limiting jobs to 45 mins or thereabouts. Few had knowledge of internal reader.
Much came about from a VM/VSE shop moving to MVS in LPARS. PROCS were shunned initially -- compexity with abend diagnosis for those new to expanding procs. Some standards were relaxed over time as knowledge increased.
Just a note: our stds worked very well and a monthly stds meeting for adds/mods/deletes aided that tremendously. It's amazing how quickly the programmers adapted and wanted more leeway, but accepted limitation/conformity stds. Naming conventions rolled right into the as/400 world when we ventured there -- even into the PC/server world.
Programmers moved to various areas were immediately comfortable and just dealt with platform specific. I digress; sorry. |
|
Back to top |
|
|
William Thompson
Global Moderator
Joined: 18 Nov 2006 Posts: 3156 Location: Tucson AZ
|
|
|
|
OK, 37 posts in this topic, only 6 by to OP and he seems to have quit the topic over 2 days ago......Enough? |
|
Back to top |
|
|
dneufarth
Active User
Joined: 27 Apr 2005 Posts: 419 Location: Inside the SPEW (Southwest Ohio, USA)
|
|
|
|
William,
why the cut-off? I thought the point was discussion. Is it because the topic isn't being absolutely followed or something else?
Regards, |
|
Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006 Posts: 19244 Location: Inside the Matrix
|
|
|
|
Hi Bill,
Quote: |
he seems to have quit the topic over 2 days ago...... |
Just watching rather than posting. . .
Personally, i enjoy these discussions when everyone plays nice
And if the view-count can be believed, more than just a few are following the discussion. . .
d |
|
Back to top |
|
|
CICS Guy
Senior Member
Joined: 18 Jul 2007 Posts: 2146 Location: At my coffee table
|
|
|
|
Maybe he's thinking about the old phrase of 'flogging a dead horse'.....grin....
The way I think he sees it is if the OP abandons the thread, we should too.
dneufarth wrote: |
I thought the point was discussion. |
How about we start up topics such as "Discussion: Is one job better than three?" |
|
Back to top |
|
|
toughhou
New User
Joined: 21 Jan 2010 Posts: 5 Location: China
|
|
|
|
Here I have two ways:
1. Using 3 steps
//JOB001 JOB MSGCLASS=A,NOTIFI=&SYSUID
//STEP01 EXEC PGM=PGM01
//FILEIN1 DSN=TOUGH.TEST.FILE.IN1,DISP=SHR
//FILEOUT1 DSN=TOUGH.TEST.FILE.OUT1,DISP=.....
//STEP02 EXEC PGM=PGM02
//FILEIN2 DSN=STEP01.FILEOUT1,DISP=SHR
//FILEOUT2 DSN=TOUGH.TEST.FILE.OUT2,DISP=....
//STEP03 EXEC PGM=PGM03
//FILEIN3 DSN=STEP02.FILEOUT2,DISP=SHR
//FILEOUT3 DSN=TOUGH.TEST.FILE.OUT3,DISP=....
2. Using 3 jobs
JCL1:
//JOB0001 JOB .....
//STEP01 EXEC PGM=PGM01
//FILEIN1 DSN=TOUGH.TEST.FILE.IN1,DISP=SHR
//FILEOUT1 DSN=TOUGH.TEST.FILE.OUT1.DISP=....
//STEP02 EXEC PGM=IKJETF01
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
SUBMIT TOUGH.TEST.JCL(JCL2)
//
JCL2:
//JOB0001 JOB .....
//STEP01 EXEC PGM=PGM02
//FILEIN2 DSN=TOUGH.TEST.FILE.OUT1,DISP=SHR
//FILEOUT2 DSN=TOUGH.TEST.FILE.OUT2.DISP=....
//STEP02 EXEC PGM=IKJETF01
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
SUBMIT TOUGH.TEST.JCL(JCL3)
//
JCL3:
//JOB0001 JOB .....
//STEP01 EXEC PGM=PGM01
//FILEIN3 DSN=TOUGH.TEST.FILE.OUT2,DISP=SHR
//FILEOUT3 DSN=TOUGH.TEST.FILE.OUT3.DISP=....
Thank you for the contribution, but it is best to reply to topics that are currently active.
Also, when postng JCL, code, data etc use the Code tag for alignment and readability. Simply copy/paste from your mainframe screen and do Not try to adjust the spacing. d |
|
Back to top |
|
|
Nic Clouston
Global Moderator
Joined: 10 May 2007 Posts: 2455 Location: Hampshire, UK
|
|
|
|
I think he found a solution over 2 years ago. |
|
Back to top |
|
|
|