Renaming a Tape dataset.

[Pathi Kumaran]

Hi,

Can we rename a tape dataset? I have been told that we can not rename a tape dataset, but why can't we do it?

[Robert Sample]

Tape is a sequential medium. A tape volume contains a volume label. After the volume label, each data set written to that tape (assuming standard labels -- if that weren't true you wouldn't worry about renaming it) contains a data set header record, then the data for the data set, then a tape mark to denote the end of the data set. The data set header record contains the data set name (or rather, the last 17 characters of the data set name). In order to rename that tape data set, the data set header would have to be rewritten. Since physically there may be a slight overlap with the data (no two tape drives are going to align one tape EXACTLY the same every time), that means the data would need to be rewritten as well. And if this is the first data set of ten on the tape, that means all ten would need to be rewritten to ensure they all have valid data. And since the data would have to be read to be rewritten, the entire tape (with all its data sets) would have to be read into memory, just so a data set name could be changed. It makes a lot more sense to just prohibit tape data set renames.

[zprogrammer]

Great post , Thanks Robert

[William Collins]

No, the data would never be required to be re-written (for Standard Labels) to rename a tape dataset. There's an intervening Tape Mark.

I think the sequence would be load the tape, "forward space file", write the new label, unload the tape. The tape has to be writeable as well, of course.

The complicated part of renaming a tape dataset is the Tape Management System. These, along with RACF/ACF2/Top Secret are way clever enough for you not to be able to mess about with "tape" data unless you are authorised to do so. If you are authorised to do it, you'll know how.

If you want to learn a bit more, even though you won't be able to practice any of it, locate a DITTO manual.

[Pathi Kumaran]

Thanks. But I did not get these two statements -

In order to rename that tape data set, the data set header would have to be rewritten. Since physically there may be a slight overlap with the data (no two tape drives are going to align one tape EXACTLY the same every time), that means the data would need to be rewritten as well.

No, the data would never be required to be re-written (for Standard Labels) to rename a tape dataset. There's an intervening Tape Mark.

So what exactly is it?

DITTO was in use at our shop but now it is not, this is what I learn. I'm trying to locate it's manual online now.

[William Collins]

Well, http://publib.boulder.ibm.com/cgi-bin/b ... 0426111816.

I'm now tending to agree with Robert. At a brief read it seems that any editing of tape data is associated with a copy.

The only time I've used DITTO was many years ago, to copy a multi-reel unlabelled tape (yes, I did say that) to a multi-reel labelled tape so that all the data would be used... the VSE VSAM backup allowed multi-reel unlabelled backups, with a little note that it was not possible to *read* then beyond the first volume.

Naturally I had to copy the data, and didn't realise that was an intrinsic part of the process...

I may still be missing something...

[Pathi Kumaran]

Thanks for the DITTO's link. You are very humble.

The only time I've used DITTO was many years ago, to copy a multi-reel unlabelled tape

What is " unlabelled tape"? I have searched for it on google but did not find some good references. Got this one:

Tape labeling with FATS

FATS can label up z/OS magnetic tapes with either standard IBM labels, ANSI V3 labels, or as unlabelled tapes with just a tape mark.

A typical FATS LABEL job to label up 100 tapes with standard IBM labels from XMP000 to XMP099, looks like

Code: Select all

//LABEL EXEC PGM=FATS,REGION=0M 
//SYSPRINT DD SYSOUT=* 
//SYSPRIN1 DD SYSOUT=* 
//SUMMPRT DD SYSOUT=* 
//TAPE1 DD DSN=FATS,UNIT=(3590,,DEFER), 
// LABEL=(,BLP),DISP=(,KEEP) 
//SYSIN DD * 
LABEL(1) VOL=(XMP000),VOLINCR=1,MAXVOLN=100

I thought, all the tapes are labled by default? Is not it so?

[nicc]

Obviously not! Each shop has its own labelling standards regarding volume serial numbers etc. There used to be, and probably still is, a manual - not too big - on tapes.

[Robert Sample]

Tapes arrive from the manufacturer at a site in an unlabeled status. Each site has a process in place to label the tapes with their site-standard volume serial (and usually attach external identifiers as well). An unlabeled tape is one which has not gone through this process. Unlabeled tapes can be handy since some machines (not always computers, either) cannot handle tape labels.

[Gerhard_Adam]

You cannot simply rewrite the header labels [HDR1 and HDR2). Every tape data set also has corresponding trailer labels [EOV or EOF]. If they are EOV, then you'd have to re-write the labels of every tape in the sequence. In any case, you'd still have to advance through the entire tape to modify the trailer labels.

Do not confuse the header labels with the volume record (VOL) which only contains the volume serial number. The VOL record has nothing to do with data set names.

As a result, it doesn't really make sense to do that and not simply copy the tape data set in the first place.

Regarding "no label" [NL] tapes, as mentioned these are tapes that do not have header or trailer labels. They are processing using the LABEL=(,NL) parameter in JCL. Many installations do not allow this mode of processing any longer. The primary advantage was in eliminating the need to specify a data set in JCL for processing and was used as a mechanism for distributing product tapes and maintenance.