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

reducing exception count and improving performance


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

Active User


Joined: 25 Apr 2007
Posts: 206
Location: Bangalore

PostPosted: Tue Apr 14, 2009 8:50 pm
Reply with quote

Hi,

We are using a KSDS dataset but it is giving us a very high exception count leading to longer runtime of the job step. The job step is executing a cobol program opening the file in I-O mode and doing only 'WRITE'.

We changed the freespace parameter to 5,5 from 1,1 and increased space allocation also. But still the job is taking longer time.

Is there any way to monitor the CI-CA split activity? Also how can we reduce the exception counts? I think reducing excepiton count will help improve run time. Will increasing BUFSPC in define cluster or AMP parameter help?

Thanks,
Back to top
View user's profile Send private message
dbzTHEdinosauer

Global Moderator


Joined: 20 Oct 2006
Posts: 6966
Location: porcelain throne

PostPosted: Tue Apr 14, 2009 8:59 pm
Reply with quote

I realize you are very busy and do not have any time to search the forum for possible threads that are like yours.

this link:
ibmmainframes.com/viewtopic.php?t=39682&sid=0d0e014e32c1a1685173120e4acac888
is from today. there are some good comments and links to reference material that you should read.
Back to top
View user's profile Send private message
Robert Sample

Global Moderator


Joined: 06 Jun 2008
Posts: 8697
Location: Dubuque, Iowa, USA

PostPosted: Tue Apr 14, 2009 9:18 pm
Reply with quote

Splits don't matter enough to batch jobs to make a difference -- although they can have an impact when doing online processing. Changing BUFSP can help but may have side effects if the file is used online -- it's less effective as you have less control over how the buffers are allocated between data and index components.

Your best bet is adding AMP=('BUFND=??,BUFNI=??') to your batch job. Is your file being accessed sequentially or randomly? For sequential access, use enough data (BUFND) buffers to read a cylinder -- if you're not sure how many, run a LISTCAT and get the value from there. For random access, you need to have at least enough index (BUFNI) to read in the index tree (see the INDEX LEVELS on the LISTCAT output); adding a few won't hurt and could help.

The other thing to think about is that flat files benefit from additional buffers as well. If your COBOL program reads a flat file record then writes it to VSAM you definitely want to look at buffering for both files.
Back to top
View user's profile Send private message
Frank Yaeger

DFSORT Developer


Joined: 15 Feb 2005
Posts: 7129
Location: San Jose, CA

PostPosted: Tue Apr 14, 2009 9:21 pm
Reply with quote

Quote:
a very high exception count


Just out of curiosity, do you mean a very high EXCP count? EXCP stands for Execute Channel Program - not exception.
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 To get the count of rows for every 1 ... DB2 3
No new posts To find whether record count are true... DFSORT/ICETOOL 6
No new posts Check data with Exception Table DB2 0
No new posts Validating record count of a file is ... DFSORT/ICETOOL 13
No new posts exploiting Z16 performance PL/I & Assembler 2
Search our Forums:

Back to Top