unable to limit number of volumes created using pgm=iebgener

gerald williams · New User Joined: 09 Jun 2009 Posts: 7 Location: FLORIDA

When I code VOL=(,,,12) in my iebgener step, it creates more then 12 volumes, does not abend and issues a return code of '0'. In this case it created 20 volumes. Can anyone help me.

Bill Woodger · Posted: Fri Sep 02, 2011 1:01 am

Robert beat me with the answer on the other topic you added to.

You can't use the volume count as a limit to the number of output volumes, it is designed to aid allocation of tape units.

What are you trying to achieve, there is probably some way if we know.

dick scherrer · Posted: Fri Sep 02, 2011 1:01 am

Hello,

Suggest you talk with your tape management people. . .

Just curious, if there is more data volume than will fit on 12 "tapes", why would you want this to stop? The content would basically be useless as there would be no control of where the "file" ended. . .

gerald williams · New User Joined: 09 Jun 2009 Posts: 7 Location: FLORIDA

I have a two-step job. The 1st step is an iebgener to copy the tape files. The second step is a vendor program that calls in the copied file. The 2nd step will only allow up to 16 volumes as input. If there is more than 16 volumes the 2nd step will ABEND after running for several hours. There is a long recovery process to correct the 2nd step ABEND. I would like for the iebgener step to ABEND or issue a return code other than ‘0’ and stop the process there.

dick scherrer · Posted: Fri Sep 02, 2011 1:53 am

Hello,

gerald williams · New User Joined: 09 Jun 2009 Posts: 7 Location: FLORIDA

I am coping Storage Tech 9840 tapes to IBM's virtual tapes. The vendor step (2nd step) has an internal tape limit of 16 and abends when it calls for the 17th tape. Based on the IBM manual I am coding the volume count correctly "VOL=(,,,16)". This should limit the number of volumes createdby the ibgerner step, but it is not. What am i missing? Any help is appreciated.

enrico-sorichetti · Posted: Fri Sep 02, 2011 2:15 am

You have two options ...
1) write a tape MOUNT/EOV exit which does nothing but count and after a certain number of volumes terminates the program.

less intrusive
2) write a program which does its own label processing considering each tape as a <single> file
and after having written a certain number of files/volumes terminates with the proper return code
most probably the easiest way is to do a bit of EXCP programming and handle everything Yourself, VOL,HDR,EOF,EOV and tape marks

You might take as a canvas
the COPYMODS program (CBT Tape File 229),
ot the COPYTAPE (CT) program (CBT Tape File 316)
or search Yourself for something that You can understand better

it should not be a big effort to extract the needed snippets
usually in these program the ugly part is the read part

in this case it would be much simpler ...
write all the header <control>records
read the sequential file, build the buffer
at eot, there is usually the space left to write the trailing <control> recods

dick scherrer · Posted: Fri Sep 02, 2011 2:23 am

Hello,

enrico-sorichetti · Posted: Fri Sep 02, 2011 2:23 am

gerald williams · New User Joined: 09 Jun 2009 Posts: 7 Location: FLORIDA

I will try a volume count of 5 and see what happens. Thanks

Bill Woodger · Posted: Fri Sep 02, 2011 3:22 am

With a volume count of 5, you will get five volumes.

Does the "copy" have to be with IEBGENER?

enrico-sorichetti · Posted: Fri Sep 02, 2011 3:26 am

the TS complained that when specifying a volcount of 12 he gor 20 volumes ( as per manual )
but nothing was said if it was a casual event ( data using only 20 volumes )
or if the volcount was hit
my guess goes to the casual event, he spoke of a 0 return code
so it is quicker to test with 5 volumes ( sure hit ) than hope to have data taking more than 20 volumes

Bill Woodger · Posted: Fri Sep 02, 2011 3:56 am

The zero is indeed interesting. No hint I can see from the manual so far of a non-zero, no mention of rc. Specify six to 20, you get up to 20, you want 21, tough, but what actually happens? Yes, better to test with five and see :-)

20 vs 20 is too big a co-incidence, but then there is that zero. Zero usually means OK :-)

My bold prediction is five gets five, with identical input as original 20, and gets... silly to predict that, when the data is on the way soon...

Bill Woodger · Posted: Fri Sep 02, 2011 4:05 am

837-08.

Coincidence with 20th volume, causal. Spot-on enrico.

Still a predication, of course...

dick scherrer · Posted: Fri Sep 02, 2011 9:29 am

Hello,

IIRC the "default" max volumes is site specific. Many (most) use 5 as this is the value in the install but long ago places that typically created many volumes of tapes as output set their limit higher.

I suspect that if this copy was run with no volume count specified, it would get the s837 after filling 5 volumes (regardless of device type) so long as the max volume count was 5 . . .

gerald williams · New User Joined: 09 Jun 2009 Posts: 7 Location: FLORIDA

I used vol=(,,,5) and the step abended with an 837-08 when it tried to create the six volume. I guess I will allow the iebgener step to create as many volumes as it wants and then add a step to check the number of volumes created and ABEND the new step if it finds more than 16. Thanks for everyone help.

Bill Woodger · Posted: Fri Sep 02, 2011 9:02 pm

That sounds like an amount of wasted time.

Does it have to be IEBGENER copying the data? What do you do if there is too much data for the 2nd step?

enrico-sorichetti · Posted: Fri Sep 02, 2011 9:07 pm

Just curious...

one of the rules of IT is data consistency...
if Your data needs 18 volumes, for overall consistency You need to process all of it it means 18 volumes

what happens to Your data consistency when some application can process only 16 volumes of data ???

are You just going to throw away the residual 2

gerald williams · New User Joined: 09 Jun 2009 Posts: 7 Location: FLORIDA

I am coping Storage Tech 9840 tapes to IBM's virtual tapes. The MVS guys have put a limit on the amount of data each virtual tape can hold. I have to load the copied tapes into a Vendor Software that has a limit of 16 volumes per dataset. So my temporary resolution is to develop a work around for the Storage Tech tapes that cannot be copied. This work around will be in place until the MVS guys increased the amount of data the virtual tapes can hold or the Vendor increase the number of volumes they will accept per dataset. I have no control over the MVS guys or the Vendor. I am only playing the hand I was dealt. Thanks for all of your help.

dick scherrer · Posted: Fri Sep 02, 2011 11:14 pm

Hello,

gerald williams · New User Joined: 09 Jun 2009 Posts: 7 Location: FLORIDA

There is no vendor "record limit" there is a vendor 16 volume limit. We are not going to drop any data or corrupt the file. We will just not import the virtual tapes into the vendor software if more than 16 are created from the Storage Tech tapes. In that case we will continue to use those Storage Tech tapes until the MVS guys increased the amount of data the virtual tapes can hold, or the Vendor increase the number of volumes they will accept per file. Thanks for your help until next time.