On Sun, 11 May 2008 00:27:42 +0000, Linda Mooney wrote: > >I found that I also needed "unpacking" room, the SMPWKDIR, also a HFS file, at >minimum the same size as the SMPNTS, but better to allow more. ... When it >broke on space, just followed the directions and restarted. I did not need to >start over, just restart. > > ... I know that PTFs can be deleted, but it is very convenient to look in > the SMPPTS at the fix rather have having to look online. > Would it save space to delete them from the SMPPTS but retain them in the SMPNTS, since:
o The SMPNTS is compressed with a more efficient formula than the SMPPTS. o IBM recommends using PDSE for the SMPPTS, and multi-volume PDSEs are not supported. SMP/E is clever dealing with relative files. I believe the current scheme un-tars one; IEBCOPY loads it; and deletes the un-tarred copy before proceeding to the next. Explain "restart" as opposed to "start over"? SMP/E is less clever dealing with multiple SMPPTFIN files -- it un-tars all and concatenates them with DYNALLOC into DD SMPPTFIN. This has caused me both to break on space in SMPWKDIR (not an enormous run; I deliberately underallocated SMPWKDIR to test the parameters) and to fail with the limit on DD concatenations. It would be more clever to spawn a process that unpacked each then deleted the workfile before starting the next, while piping its output to a single unconcatenated DD SMPPTFIN. In fact, if the archives were only compressed, not tarred, uncompress could pipe directly into SMPPTFIN with no use of SMPWKDIR. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
