Sending GDG file through NDM process

Kamaleshwaran · New User Joined: 10 Apr 2010 Posts: 7 Location: Hyderabad

We are sending the GDG generation file via NDM to the remote system.

Question :
STEP1:
(=1) GDG created
STEP2:
NDM ---> GDG version file.

I need to NDM the latest generation file created in step1.
Do I need to give (0) genaration in NDM step &SRCDSN paramter or (+1) genration???

When I tried with (+1) generation, am getting error while getting the NDM statistics, but the file is transmitted.
I tried to look up the existing NDM JCL's, the JCL's contains only (0) version eventhough a new file is generated in the same JCL.
Am really confused. This is the first time am making changes in NDM process and didnt get much reference. Please help me.

Phrzby Phil · Posted: Thu Feb 17, 2011 8:48 am

Kamaleshwaran -

We extend a personal invitation to you to show (use code tag please) the code you used and also, if it is not too much trouble, the exact error message you received.

This latter information can be found in the select statistics report.

Perhaps then the cause of your distress will be easier to ascertain.

Or, can you think of another way to request assistance?

superk · Posted: Thu Feb 17, 2011 1:58 pm

Kamaleshwaran, I'm trying to think this through logically.

First of all, I think that it's a poor coding design to create the dataset in one step and NDM it in the next. I'd recommend that you make that process two seperate jobs.

Now, the system is going to place an enqueue on the GDG base for your new generation, and if I recall, that enqueue stays throughout the duration of that job. Now, NDM will also place an enqueue the same way. I can't help but believe that, at this step of the job, the dataset that is enqueued would need to be the current (zero) generation, and definitely not a +1 generation, similar to the same situation if you were using an FTP process.

One thing I would do is to allocate the new generation to a DDNAME in the NDM step and have that dataset be the input to the process. As an alternative, I'd also consider the possibility of resolving the new generation to it's explicit name, and passing THAT name to the input of the process.

Kjeld · Posted: Thu Feb 17, 2011 5:49 pm

Generally, when you allocate and use a new GDG generation in a job, you have to refer to the created version as (+1) througout the job, as the dataset is not catalogued before at the end of the job.

gcicchet · Posted: Fri Feb 18, 2011 4:29 am

Hi,

Robert Sample · Posted: Fri Feb 18, 2011 4:42 am

GDG relative references (except for VOL=REF='gen') are resolved at the beginning of the job, hence the need to reference (+1) in all steps of the job to refer to the generation created in the job.

Kjeld · Posted: Fri Feb 18, 2011 3:02 pm

enrico-sorichetti · Posted: Fri Feb 18, 2011 3:32 pm

Anyway I ( personal advice ) usually discourage mixing in the same job
ddname and dsname access for NEW datasets especially GDG' s