Page 2 of 2

Re: compress by batch

PostPosted: Thu Mar 25, 2010 7:23 pm
by expat
One method I use is HMIG the PDS and the HRECALL it, as HSM uses IEBCOPY to migrate / recall PDS datasets.

Re: compress by batch

PostPosted: Fri Mar 26, 2010 12:42 am
by dick scherrer
Hello,

Something else to look into is whether TWS has a built-in compress feature. . . The TWS documentation might mention how this needs to be handled even if there is nothing built-in.

Is there any possibility to use a PDS/E rather than just a PDS?

Re: compress by batch

PostPosted: Fri Apr 09, 2010 7:26 pm
by stefkrastev
samb01 wrote:OK thanks. ;)
Hello sambo01 and Robert
To share one simple idea using ZOS standard tool with ISPF/TSO :
- p.3.4 – BCA1MGP.STEF.JCL
- Z BCA1MGP.STEF.JCL or with ‘/’ (and choose 'compress' from Data Set List Actions)
- Result with S BCA1MGP.STEF.JCL is:
Used cylinders – decreased by 1
Used extents - decreased by 1
It is very easy and fast!
Stefkrastev

Re: compress by batch

PostPosted: Fri Apr 09, 2010 7:30 pm
by Robert Sample
Stefkrastev, have you tried this with a data set in use by a started task? What happens when you do so?

Re: compress by batch

PostPosted: Mon Apr 12, 2010 2:30 pm
by stefkrastev
Robert Sample wrote:Stefkrastev, have you tried this with a data set in use by a started task? What happens when you do so?
Robert,
Yes, the PDS /library/ file should not be used at this time by another task – ZOS will issue ‘Data set in use’ . But this is still usable for private or system libraries requiring immediate action because of waiting users missing space and after informed them to stop the access for 5 minutes.
Another approach for compress a PDS in batch with expanding the PDS at the same time is given in an example for IEBCOPY at OS/390 Dataset Utilities (http://www.lascon.co.uk/d017003.htm) where the existing PDS is copied in a larger with a new name and after checking the bigger file is working OK to delete the old and alter the new name to original one.
Hope this clear the case!
Stefan

Re: compress by batch

PostPosted: Tue Apr 13, 2010 12:36 am
by dick scherrer
Hello,

But this is still usable for private or system libraries requiring immediate action because of waiting users missing space and after informed them to stop the access for 5 minutes.
Informing people they should not use the library does not guarantee the dataset is not in use. . .

IEBCOPY at OS/390 Dataset Utilities (http://www.lascon.co.uk/d017003.htm) where the existing PDS is copied in a larger with a new name and after checking the bigger file is working OK to delete the old and alter the new name to original one.
This also will work only when the dataset is not already in use. . .

Hope this clear the case!
If the "case" is to compress the dataset while it is in use, no this does not "clear the case".

Re: compress by batch

PostPosted: Tue Apr 13, 2010 12:41 pm
by stefkrastev
OK Dick, agree - you are right!
I accept the explanation of Robert for DISP=OLD and I will try to compress the dataset while it is in use!
Thanks!

Re: compress by batch

PostPosted: Tue Apr 13, 2010 11:28 pm
by dick scherrer
Hello,

Suggest you make a "throw away" copy of the library to experiment with. . .

Even if a compress runs, it may corrupt the library and if this is not detected immediately, the problem may be quite difficult to correct.

Re: compress by batch

PostPosted: Fri Apr 16, 2010 1:49 pm
by stefkrastev
Hello Dick,
I've tested batch compress with disp=old for a library where a member was editing by IBMUSER and it doesn't work,
waiting the data set to be available:
//CONDENS JOB (9999),
// CLASS=B,
// REGION=0K,TIME=1440,
// MSGCLASS=X,MSGLEVEL=(1,1)
//S15 EXEC PGM=IEBCOPY
//SYSPRINT DD SYSOUT=*
//SYSUT3 DD UNIT=SYSDA,SPACE=(CYL,(4,2))
//SYSUT4 DD UNIT=SYSDA,SPACE=(CYL,(4,2))
//PDS DD DISP=OLD,DSN=BCA1MGP.STEF.JCL1
//SYSIN DD *
C INDD=PDS,OUTDD=PDS

09.10.21 JOB00195 $HASP373 CONDENS STARTED - INIT 1 - CLASS B - SYS BC2A
09.10.21 JOB00195 IEF861I FOLLOWING RESERVED DATA SET NAMES UNAVAILABLE TO CONDENS
09.10.21 JOB00195 IEF863I DSN = BCA1MGP.STEF.JCL1 CONDENS
09.10.21 JOB00195 *IEF099I JOB CONDENS WAITING FOR DATA SETS
Then I asked who was using the library:
/D GRS,RES=(*,BCA1MGP.STEF.JCL1)
BC2A IBMUSER 0058 008FFBF8 SHARE OWN
BC2A CONDENS 0025 008F17E0 EXCLUSIVE WAIT
I canceled IBMUSER (/c u=IBMUSER) and job CONDENS finished OK:
---------------------------
IEB152I RESBIO28 COMPRESSED - WAS ALREADY IN PLACE AND NOT MOVED
IEB152I DELBIO5 COMPRESSED - WAS ALREADY IN PLACE AND NOT MOVED
IEB1097I FOLLOWING MEMBER(S) MOVED IN DATA SET REFERENCED BY PDS
IEB154I ADDSLTAP HAS BEEN SUCCESSFULLY MOVED
IEB154I ADSLTAP2 HAS BEEN SUCCESSFULLY MOVED
IEB154I BACKFAST HAS BEEN SUCCESSFULLY MOVED
------------------------
IEB1098I 104 OF 120 MEMBERS MOVED IN DATA SET REFERENCED BY PDS
IEB144I THERE ARE 157 UNUSED TRACKS IN OUTPUT DATA SET REFERENCED BY PDS
IEB149I THERE ARE 0 UNUSED DIRECTORY BLOCKS IN OUTPUT DIRECTORY
IEB147I END OF JOB - 0 WAS HIGHEST SEVERITY CODE
During the compress batch is not possible any task to have access for editting/processing this PDS (the library)
I think this clear the case!
Stefan

Re: compress by batch

PostPosted: Sat Apr 17, 2010 2:29 am
by dick scherrer
Good to hear you have a resolution - thank you for the followup :)

d