On Tue, 7 Apr 2015 04:29:15 -0500, Norbert Friemel wrote: >On Tue, 7 Apr 2015 03:59:30 -0500, Juergen Kehr wrote: > >> ... I get a RC=40 during UNTERSE of a PDSE (Load) LIBRARY. ... >> > >http://publibfp.dhe.ibm.com/cgi-bin/bookmgr/BOOKS/iea2v1c2/17.4.1 >"Partitioned data sets extended (PDSE) containing program objects are not >supported. " > For Lizette and Scott, what more is needed than a citation of a manual stating that the operation is not supported? (But is the diagnostic message lucid?)
I had expected AMATERSE to use IEBCOPY internally. But that would require a (large) workfile. So the programmer must IEBCOPY unload the PDSE, then AMATERSE the PDSU, so that programmer assumes the onus of the workfile. I believe TSO TRANSMIT uses IEBCOPY. IIRC, there have been reports of TRANSMIT's failing for an underallocated workfile. (I also suspect that IEBCOPY uses Program Management API to deal with Program Objects.) It's a pity that IEBCOPY can't use POSIX pipes for only its PDSU data sets. That would allow AMATERSE and TRANSMIT to use IEBCOPY with only trivial workfiles, piping IEBCOPY PDSU directly into the utility. And SMP/E keeps its SMPNTS in IEBCOPY PDSU copied to UNIX .tar.Z files which it must first unzip to SMPWKDIR, then copy to DSORG=PS, and finally reload with IEBCOPY. Two (large) workfiles. IEBCOPY would do AMATERSE, TRANSMIT and SMP/E a favor by supporting UNIX files (including pipes) as its PDSU. (SMP/E also performs some gyrations to break up large concatenations of SMPPTFIN data sets (I believe it DYNALLOCs directly from SMPWKDIR, not requiring a copy to DSORG=PS.) Even that storage requirement could be reduced by allocating SMPPTFIN to the output of a single POSIX pipe and feeding that with the unzipped .tar.Z files one-by-one, deleting each before unzipping the next. (But would that complicate error reporting?) IIRC, AMATERSE has a restriction that a PDS can not be PACKed directly to a tape, but must first be PACKED to DASD (another (large) workfile), then copied to tape. I suspect this restriction arises from a need to POINT to a prologue block and update it in place at the end of the operation. I know IEBCOPY PDSU also contains a prologue. I wonder how IEBCOPY generates that in a single pass? Does it perform a trial scan of the PDS? -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
