PKZIP issue in compressing large file.

ravikumar15 · New User Joined: 27 Mar 2006 Posts: 68

Hi,

I am trying to compress a large file using PKZIP utility. Though I am allocating the compressed output file to tape, the job is failing with SB37 on a temporary file. When I tried to add space for the temporary file, the compressed file is not able to be unzipped. I am using the below commands for compressing the dataset.

-ECHO
-ARCHIVE_OUTFILE(ARCH01)
-ARCHMCLASS(Y28NE)
-ARCHTYPE(FB)
-DATA_TYPE(TEXT)
-ZIPPED_DSN(PRE.X100.PREM.INFILE,prem_file.txt)
PRE.X100.PREM.INFILE
/*
//ARCH01 DD DSN=PRE.X100.PREM.INFILE.ZIP,
// DISP=(NEW,CATLG,DELETE),UNIT=TAPE

The job is giving an error, ZPFM121E. B37 error on the temporary file.

When I googled, I got that we can add additional space to temporary file by giving, -TEMPxxxx or //ARCHTEMP DD statement. I tried with //ARCHTEMP DD UNIT=TEMPDA,SPACE=(CYL,(300,300),RLSE) but still the job failed with B37.

I am not aware of the syntax of -TEMP.

Please help me.

Thanks,
Ravi

PeterHolland · Posted: Mon Aug 11, 2014 12:53 pm

It's all in the PKZIP manual.

ravikumar15 · New User Joined: 27 Mar 2006 Posts: 68

I did refer to the manuals and found the usage of -TEMP commands and I did the below changes to the control card,

-ECHO
-ARCHIVE_OUTFILE(ARCH01)
-ARCHMCLASS(Y28NE)
-ARCHTYPE(FB)
-DATA_TYPE(TEXT)
-TEMP_MGMTCLASS(Y28NE)
-TEMP_VOLUMES((,,,99)
-ZIPPED_DSN(PRE.X100.PREM.INFILE,prem_file.txt)
PRE.X100.PREM.INFILE
/*
//ARCH01 DD DSN=PRE.X100.PREM.INFILE.ZIP,
// DISP=(NEW,CATLG,DELETE),UNIT=TAPE

But still the temporary dataset is allocating in work volumes and the job is failing with B37.

When I give space parameters, the job is completing fine, but it is adding some extra bytes to the zipped file

expat · Posted: Thu Aug 14, 2014 5:12 pm

Just a question, but just how big is this file ?

The objective of tape processing is for very big files. Tapes subsystems also have their own compression routines as does DFSMSsms.

What amount of savings are you looking to achieve, and at what additional CPU cost.

ravikumar15 · New User Joined: 27 Mar 2006 Posts: 68

Hi Expat,

The actual file size is around 100 GB and after compressing it comes to 70GB.

There are other files transferred to the same server, with same PKZIP commands, they are successful except one file.

steve-myers · Posted: Thu Aug 14, 2014 5:53 pm

If it is your intent to send this data set to another z/OS system, you should investigate using the TERSE or AMATERSE utilities. These utilities exist on all z/OS systems. I have never encountered issues with data set size with them. TERSE and AMATERSE are documented in MVS Diagnosis: Tools and Service Aids for your z/OS release. They are slow (by this I mean they use a lot of CPU time), but they always work.

I have encountered references to toy machine utilities that expand files created by TERSE or AMATERSE, but I know nothing about them other than the reference.

I share expat's concern about the size of the source data set.

100 GB to 70 GB is pretty lousy compression unless the source data is already compressed in some fashion. My experience with TERSE is 100 GB should compress to about 30 GB.

ravikumar15 · New User Joined: 27 Mar 2006 Posts: 68

The client requirement is to use PKZIP only.

Please help me in resolving the issue.

PeterHolland · Posted: Sun Aug 17, 2014 10:34 am

Robert Sample · Posted: Mon Aug 18, 2014 3:49 am

Ramsee · New User Joined: 06 Jan 2011 Posts: 53 Location: Chennai

Hi Ravikumar,

You can write up your issue along with the XDC file in an Notepad as an attachment to Kevin Moll [kevin.moll@pkware.com]. He will be able to guide you on this 100 GB to 70 GB Compression.