On Mon, 19 Nov 2012 14:16:59 -0600, McKown, John wrote:

>This though occurred to me when Lindy said that no mainframe shop he knows of 
>has tape drives. I wonder what others would think of the possibility of being 
>able to use both ADRDSSU and AMATERSE in a UNIX environment. Why? Well, image 
>doing a ADRDSSU DUMP of a disk (or even a logical dataset backup) and sending 
>it to a UNIX pipe instead of a data set. This pipe would be the input to 
>AMATERSE. ...
>
The real pity is that IEBCOPY can't write its PDSU (not PDS or PDSE) output to/
read from a UNIX file or other stream.  It employs the block boundaries on
reload, but since the PDSU is RECFM=VBS, they can be reconstructed from
the BDWs.  GIMZIP/GIMUNZIP do this.  Considerable overhead and some
SPACE crises incurred by GIMUNZIP could be circumvented if IEBCOPY read
the GIMZIP UNIX files directly (or via a pipe from uncompress; they're .pax.Z).

Conway's Law:  SMP/E can't impose a requirement for even a trivial
enhancement to IEBCOPY that would reap a much greater saving for
IBM and its customers.

AMATERSE can't compress a PDS(E) directly to tape; it must go via DASD.
The reason appears to be that AMATERSE must POINT back to the first block
to write some sort of summary; perhaps SPACE requirements.  GIMZIP
supplies similar metadata in the GIMFAF files.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to