How to recall GDG versions using REXX

NIKHIL SHARMA · New User Joined: 22 Jun 2010 Posts: 12 Location: Bangalore

Hi,

I am writing a rexx program to read all versions of GDG and search for the particular record. However, there are 150 versions present and most of them are migrated to DASD. Need to know how can we recall the migrated versions so that it does not take much time.

Curently, rexx code i am using is taking lots of time to recall these versions if version number is high (around 150 versions are present).

Please suggest something.

Anuj Dhawan · Posted: Mon Aug 09, 2010 1:45 pm

Unfortunately your terminolgy is not correct; using correct words for the things in IT usually increase the chances of getting the correct answers rather fast.

When you talk about GDGs, version and generations both can be correct terms but they have very differnt meanings and in the context of your question what should be used, does matter. There can not be 150 versions for a given GDG - most possibly you menat 150-generations. And why would some migrate them to DASD?

expat · Posted: Mon Aug 09, 2010 2:03 pm

The problem is almost certainly due to your REXX processing each generation one at a time, so when the open for the file is issued HSM will recall the dataset before you can process it.

You will need to recall all of the generations before your job starts. This can be done as shown by the example given by Anuj. The advantage of this is that HSM will only mount a tape once for all datasets to be recalled from that tape, where in your case is migrated to tape the tape may be mounted read and dismounted once for each generation.

Perhaps you should explain the situation tour your friendly storage gurus to see if all generations can be retained on DASD.

Anuj Dhawan · Posted: Mon Aug 09, 2010 3:46 pm

NIKHIL SHARMA · New User Joined: 22 Jun 2010 Posts: 12 Location: Bangalore

Apologies for using wrong words in my questions.

Also i need to hrecall from TAPE (not from DASD)

i need to look for certain record in from the GDG generations and do some processing on that. we can have total of 150 GDG generations.

I am giving here the code i wrote to scan through GDG versions and search and print the desired record.

"ALLOCATE FI(OUTDD) DA("MNRP.OUTPUT.DATA") SHR REUSE"

GETDATA = OUTTRAP("RAWOUT.","*")
"LISTCAT ENTRIES('"ABCD.XYZ.PQRS.DATA"')" /* ABCD.XYZ.PQRS.DATA is GDG base */
STOPGETDATA = OUTTRAP('OFF')
IF WORD(RAWOUT.1,1) <> "GDG" THEN DO
SAY 'NOT GDG BASE'
EXIT
END
GDGLISTINDEX = 0
J = 0
Q = 1
DO LOOP = 1 TO RAWOUT.0 / * looping through total number of GDG generations */
IF WORD(RAWOUT.LOOP,1) <> "NONVSAM" THEN ITERATE
GDGLIST.GDGLISTINDEX = WORD(RAWOUT.LOOP,3)
"ALLOCATE FI(INDD) DA('"GDGLIST.GDGLISTINDEX"')
SHR REUSE"
"EXECIO * DISKR INDD ( STEM FILE."
GDGLISTINDEX = GDGLISTINDEX + 1
DO I = 1 TO FILE.0
J = J + 1
COMP3 = SUBSTR(FILE.I,21,6)
XVAR =C2X(COMP3)
XVAR = SUBSTR(XVAR,2,10)
IF XVAR = ICN THEN
DO
FILEA.Q = FILE.I
Q = Q + 1

END
END
"EXECIO 0 DISKW INDD ( FINIS"
"FREE FI(INDD)"
END

If GDG generations are upto 50, it works pretty fine. But, when we have around 100 generations, it takes time. Today it took 15+ minutes to run

I tried using TSO HRECALL "file.name.*.*" NOWAIT but it did not work. when i gave full gdg name (like "XYZ.ABC.G0157V00), it works fine. but when i try to use 'GDGLIST.GDGLISTINDEX' it does not work. Not sure if i am giving correct syntax here.

Please help

dick scherrer · Posted: Mon Aug 09, 2010 8:06 pm

Hello,

You need to start over. . . Either what you want to do has not been explained properly or you are trying to do something that is not done.

HRECALL retrieves migrated dasd files back to regular dasd.

Also, if you have 150 datasets and they are 20 gig or so, your storage management people will be looking for you with unkindness in their hearts<g>.

expat · Posted: Mon Aug 09, 2010 10:20 pm

dick scherrer · Posted: Mon Aug 09, 2010 10:53 pm