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

Peformace issues for UNIT=(SYSDA,2) with huge CYL


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

New User


Joined: 01 Aug 2007
Posts: 50
Location: Hosur

PostPosted: Fri Apr 16, 2010 9:48 pm
Reply with quote

Hi,

I have JCL which creates a PS file defined with

UNIT=(SYSDA,2),SPACE=(CYL,(600,100),RLSE),

Now, I have doubt here - since this step executes for hours.

By coding SYSDA,2 for UNIT - does it mean we are searching for CYL,(600,100) within two volumes?

By increasing the UNIT to SYSDA,14 - Will it make any difference?

Thanks!
Back to top
View user's profile Send private message
Robert Sample

Global Moderator


Joined: 06 Jun 2008
Posts: 8700
Location: Dubuque, Iowa, USA

PostPosted: Fri Apr 16, 2010 10:43 pm
Reply with quote

The fact that the step executes for hours is almost certainly not related to the space request. Considering DASD access times are in milliseconds, the space is either acquired or not acquired in a fraction of a second. If not acquired, your job will get an x37 abend.
Back to top
View user's profile Send private message
PeD

Active User


Joined: 26 Nov 2005
Posts: 459
Location: Belgium

PostPosted: Sat Apr 17, 2010 12:19 am
Reply with quote

Quote:
By increasing the UNIT to SYSDA,14 - Will it make any difference?

We can discuss on that, but not significantly.

But .... which language? how many records? blocksize? buffersize?....

Check all that.
Back to top
View user's profile Send private message
dick scherrer

Moderator Emeritus


Joined: 23 Nov 2006
Posts: 19243
Location: Inside the Matrix

PostPosted: Sat Apr 17, 2010 1:02 am
Reply with quote

Hello,

Quote:
Now, I have doubt here - since this step executes for hours.
Allowing the allocation to span more volumes will not significantly alter this. . .

The 2 things to consider are the volume of data and how the code is written. Code that runs sub-second for 100 records may fail when full volume data is processed (which is why it is very risky to promote code that has not gone thru a "full volume" test before promotion to Production).

How many records are to be read? Written? Will any of the records be accessed multiple times? Is the input on tape/cart? If so, how long is the wait between mounts?

Posting the full jcl for this step may help someone help you.
Back to top
View user's profile Send private message
alan_blore

New User


Joined: 01 Aug 2007
Posts: 50
Location: Hosur

PostPosted: Tue May 18, 2010 6:29 pm
Reply with quote

Thanks All for Info...and sorry for late reply
It was the CPU which was full, so making this job to run long. Now it is ok.

But now the other question arises...

Consider this step - a SORT or a program - whatever it is -

What would be ideal value for 'n' in parameter UNIT=(SYSDA,n)?

I believe we can check with storage team for this, but that is tough for me in a day to day scenario, If I need to decide - what should I do..
Back to top
View user's profile Send private message
dick scherrer

Moderator Emeritus


Joined: 23 Nov 2006
Posts: 19243
Location: Inside the Matrix

PostPosted: Tue May 18, 2010 7:12 pm
Reply with quote

Hello,

Just because something is a bit of effort does not mean it should not be done. . . Your storage management people should be your first place to ask. They manage your system and know particulars - we do not.

There is no "ideal value". It is not as though the dataset will be scattered about thus improving performance. . . Suggest you look in the JCL Language Reference (link to "IBM Manuals" at the top of the page) and learn about the UNIT subparameters.
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 Convert Service Unit to CPU Seconds All Other Mainframe Topics 1
No new posts MacKinney Batch to CICS upgrade causi... CICS 7
No new posts Issues Converting From ZD to Signed N... DFSORT/ICETOOL 4
No new posts Issues with VIEW DATASET Command CLIST & REXX 2
No new posts Issues with executing a REXX MACRO th... TSO/ISPF 4
Search our Forums:

Back to Top