IBM Mainframe Forum Index
 
Log In
 
IBM Mainframe Forum Index Mainframe: Search IBM Mainframe Forum: FAQ Register
 

JCL and TAPE drives: how to maximize usage of tape drives


IBM Mainframe Forums -> JCL & VSAM
Post new topic   Reply to topic
View previous topic :: View next topic  
Author Message
Sarma Peri

New User


Joined: 02 Oct 2021
Posts: 5
Location: India

PostPosted: Sat Oct 02, 2021 6:42 am
Reply with quote

Hi
We are about to migrate 25000 huge TAPE datasets to AWS . All are independent. So we would like to create many JCLs so that they can run in parallel.
But we are also bound by the actual tape drives defined.
So we are creating many job streams and in each job stream we will have couple of jobs that can run parallel.
In each job we planned to write about 25 or 50 steps to extract copy the TAPE datasets for FTP into AWS.
We would like to know if the tape drives used in the first step will be released once the JCL step is completed. So that those tape drives can be released for other steps in same JCL or other JCL.

I read sometime back that tapes will NOT be released till end of the job, even though we don't need it after a particular step.
But I want to confirm then I will write 25000 JCLs.
And if you suggest a way to release the TAPE locking then we will code few JCLs that have many steps say 100 or so.

Is there any place I can find more information about locking of tape drives?

I searched in ibm manuals and used google search
but no luck.

Any help is highly appreciated

Thank you...
Back to top
View user's profile Send private message
dneufarth

Active User


Joined: 27 Apr 2005
Posts: 336
Location: Inside the SPEW (Cincinnati OH USA)

PostPosted: Sat Oct 02, 2021 6:53 am
Reply with quote

check out DD statement param FREE
Back to top
View user's profile Send private message
Sarma Peri

New User


Joined: 02 Oct 2021
Posts: 5
Location: India

PostPosted: Sat Oct 02, 2021 8:08 am
Reply with quote

Thank you Dave for quick reply. I will check the FREE parameter..
Back to top
View user's profile Send private message
Joerg.Findeisen

Active Member


Joined: 15 Aug 2015
Posts: 623
Location: Bamberg, Germany

PostPosted: Sat Oct 02, 2021 8:11 am
Reply with quote

Code:
** The FREE Parameter                                                       
                                                                             
    The FREE parameter indicates when MVS is to free the resource(s)         
    allocated to this DD statement.  The resource(s) can be device(s),       
    volume(s), or data set(s).                                               
                                                                             
    Note:  Specifying FREE will not release the enqueue on the data set     
           until the last step that requires the data set completes         
           processing.                                                       
                                                                             
                                                                             
    Defaults                                                                 
                                                                             
    If no FREE parameter is specified, the default is END.  Also, if the FREE
    parameter is incorrectly coded, the system substitutes END and issues a 
    warning message.
Back to top
View user's profile Send private message
Sarma Peri

New User


Joined: 02 Oct 2021
Posts: 5
Location: India

PostPosted: Sat Oct 02, 2021 8:37 am
Reply with quote

As we are not planning to use the same dataset in subsequent steps, and the default is FREE=END,

I am thinking that we can get best resource optimization even if we code JCL with MULTIPLE steps which FTPs one dataset each. This helps us to issue minimal number of JCLs.

Thank you for your time Dave & Joerg
🙏
Back to top
View user's profile Send private message
sergeyken
Warnings : 2

Senior Member


Joined: 29 Apr 2008
Posts: 1210

PostPosted: Sat Oct 02, 2021 8:49 am
Reply with quote

Another approach may be as follows.
- use one FTP per job
- use unique (or mostly unique) jobnames, with the same CLASS= parameter
- keep all jobs together in a single DSN, or a single member of library, to submit the whole stream with one SUBMIT command
- provide as many initiators for the selected CLASS= as the desired number of parallel transmissions
Back to top
View user's profile Send private message
Sarma Peri

New User


Joined: 02 Oct 2021
Posts: 5
Location: India

PostPosted: Sat Oct 02, 2021 9:10 am
Reply with quote

Quite interesting suggestion. Unfortunately I am extracting TAPE datasets, compressing into zip file and ftping that zipped file to AWS. Using temporary dataset. So I can't group multiple datasets together, (to minimise space requirements)


And
Is there any way as a non administrator, we can monitor the tape drives usage while executing these number of parallel jobs?
As a programmer, who does not have special access to control M etc?
Any commands etc (sdsf or basic control m commands)

Thank you
Back to top
View user's profile Send private message
sergeyken
Warnings : 2

Senior Member


Joined: 29 Apr 2008
Posts: 1210

PostPosted: Sat Oct 02, 2021 5:16 pm
Reply with quote

Sarma Peri wrote:
Quite interesting suggestion. Unfortunately I am extracting TAPE datasets, compressing into zip file and ftping that zipped file to AWS. Using temporary dataset. So I can't group multiple datasets together, (to minimise space requirements)

If this response is related to my algorithm(?) then those issues have absolutely nothing to do with it. All single-DSN related activities are to be included into a single job. Temporary datasets, compression, grouping together (of what?) - they are all irrelative.
Back to top
View user's profile Send private message
Sarma Peri

New User


Joined: 02 Oct 2021
Posts: 5
Location: India

PostPosted: Sat Oct 02, 2021 5:41 pm
Reply with quote

Thank you Sergeyken. I will take this up for discussion with our team.
Currently we built one job to have all the above ftp activities for one dataset. As we were thinking that the tape drive locks will not be released until the job ends.
As per Dave inputs, the tape drives will be released once the step is completed. So We can now plan to build multiple steps of say 50 datasets into single job so that design will have less number of jobs.

Thank you for your time again.
Regard
Back to top
View user's profile Send private message
dneufarth

Active User


Joined: 27 Apr 2005
Posts: 336
Location: Inside the SPEW (Cincinnati OH USA)

PostPosted: Sat Oct 02, 2021 7:42 pm
Reply with quote

This may also be a consideration UNIT=(TAPE,,DEFER) if multiple input datasets are utilized.
Back to top
View user's profile Send private message
View previous topic :: :: View next topic  
Post new topic   Reply to topic View Bookmarks
All times are GMT + 6 Hours
Forum Index -> JCL & VSAM

 


Similar Topics
Topic Forum Replies
No new posts ADR324E-VOL/DATASET SPECIFIED BY DDNA... JCL & VSAM 6
No new posts Usage of BINSEARCH PL/I PL/I & Assembler 1
No new posts Is there a way to determine Clist usa... CLIST & REXX 10
No new posts Record count of RECFM=U tape file DFSORT/ICETOOL 17
No new posts DB2 group member usage DB2 0
Search our Forums:

Back to Top