IBM Mainframe Forum

by **cchapel88** » Fri Sep 17, 2010 8:00 pm

Anybody aware of IBM z/OS JCL trick to create a duplicate of an application program's output file in parallel of the writes to that output? I have a very high-volume weekly process of staging data from our mainframe system applications to MAS "shared" disk for ETL into an analytical database system. We have added a "test" disk area also and the desire is to keep both production and test at the same level...without impacting the performance of the production process (of course).

JCL Code Snippet for the application program output file

Select all

//GB305.OUTFILE DD DSN=GB.#0983864.PPMO.MILS
X/OUTFILE DD DSN=OUTFILE,
X/ DISP=OLD,LABEL=EXPDT=98000,
X/ UNIT=&MAS,VOL=SER=&BFL,
X/ DCB=RECFM=FB,FREE=CLOSE
IEFC653I SUBSTITUTION JCL - DSN=OUTFILE,DISP=OLD,LABEL=EXPDT=98000,UNIT=MASMIL,VOL=SER=BFL140,DCB=RECFM=FB,
FREE=CLOSE

Thanks for any thought and ideas!
Chris.

by **dick scherrer** » Fri Sep 17, 2010 11:10 pm

Hello and welcome to the forum,

without impacting the performance of the production process (of course).

Not likely - writing 100,000,000 records wil always take more resources than writing 50,000,000 records. . .

If you want to avoid extra overhead in the production process, you might consider making the copy after the process has completed. This will require some resources, but should not impact the production process. If you use your system's sort product, a copy can be made very quickly and the file attributes can be copied rather than specified in the jcl.

IBM Mainframe Forum

JCL Trick to create duplicate output file

JCL Trick to create duplicate output file

Re: JCL Trick to create duplicate output file