IBM Mainframe Forum

Hi,

We have one job which sort a big file which is in tape and taking almost 5 CPU mints to complete that.

Do we have any parameters which can improve the performance? In our job we are using below SORT:-

S20Z0270 EXEC PGM=SORT,REGION=0M,
PARM='RELEASE=OFF,VSCORET=12M'
SORTIN DD DSN=#########.#######.DISKOA.D13APR,
DISP=(SHR,KEEP)
SORTOUT DD DSN=##########.######.SORT1.D13APR,
DISP=(NEW,CATLG),
UNIT=TAPE,VOL=(,,,15),
RETPD=&RETPD,
RETPD=30,
DCB=(RECFM=VB,LRECL=28793,BLKSIZE=28797)

SYSIN DD SUBSYS=(CCSS,
' OMIT COND=((17,6,CH,EQ,C''000002''))',
' SORT FIELDS=(5,10,BI,A,17,6,BI,A,27,17,BI,A,',
' 46,1,BI,A,44,2,BI,A,15,2,BI,A)',
' END')

What is the amount of data that you are sorting?

Hello,

Is the entire record needed for the output or only part of it?

If the process can fill up to 15 tape volumes, it will probably take a while.

yes Dick we need entire records in output........

Hello,

As Akatsukami asked - How much data is being sorted?

How about showing the sysout from a completed step?

Hi all,

its qtr end production job so sysout is not available but what I remember it processed 50GB data I believe....

Well, without SYSOUT we cannot advance any further. And why do you not keep the last execution of a production job? That is just criminal (and possiby illegal).