IBM Mainframe Forum

Hi,

I need to take back up of around 100 Ps files.I have all PS file name in another PS.How do I take back up of all datasets using JCL.

Use the JCL to execute a program to do the backups, of course. Since JCL does nothing but execute programs anyway, just what kind of answer did you expect?

Talk to your site support group and find out which products are installed at your site to do backups. DF/DSS (ADRDSSU) is the IBM utility for doing backups, although others can be used. There are also independent software vendor products that can do backups. Which are installed at your site? Only your site people will know for sure.

I regret for mentioning back up..

I meant I need to copy 1 PS to another,But there are hundred such PS.I used IEBGENER to copy 1 by one.Can IEBGENER be used to copy 100 PS file at once.

IEBGENER allows for one SYSUT1 (input) DD statement and one SYSUT2 (output) statement. You could do this by using IDCAMS REPRO (requiring one input and one output DD statement per file to be copied -- along with the REPRO utility statement for each), or by making a backup using DF/DSS of the high level qualifier (assuming there's a common set of data set name nodes for the 100 data sets) and restoring the data sets to the new high level qualifier, or by coding up 100 IEBGENER steps to execute..

The best choice would be to contact your site support group, explain what you're trying to do, and see what they recommend. There may be an installed package that would do this for you with very little effort -- but if you don't ask, you won't know.

Thanks..

I am copying with the exact names..but HLQ being different..DOinf it thru IEBGENER copying it 100 times,may not be a feasible option..

Using Repro also..Can I include a PS which has all PS names which needs to be copied?

If your HLQ is different between datasets then you will have to hand code one-by-one. Generics can be used by IDCAMS and DF/Dss but not on the HLQ.

Hello,

DOinf it thru IEBGENER copying it 100 times,may not be a feasible option..

I hesitate to mention this, but often we actually have to do some work. . .

If you put all of the dataset names in a file, you could read this file and generate the copy jcl for all 100 files.

If you create one set of copy jcl it can be cloned 100 times in a very short amount of time (10-15 minutes maybe).

If you create an inline PROCedure of a copy step with the dataset name as a symbolic parameter, you could clone 100 EXEC statements in a few minutes also. You could also read the file of dataset names to generate the EXEC statements.

if the dataset have some common pattern the fasset way would be a DFDSS copy
otherwise You will have to specify the one by one,
but even in this case DFDSS is useful
the advantage is that You do not have to allocate them by hand, a DFDSS copy will allocate
with the same space characteristics of the source

spmething along the lines of



read the dfdss manuals
and keep executing it with PARM='TYPRYN=NORUN'
when the filtering is what You want , take away the PARM for the final go