On Sat, 30 Apr 2011 00:31:12 -0400, Robert A. Rosenberg wrote: >At 18:08 -0500 on 04/29/2011, Rick Fochtman wrote about Re: SMPPTS >run out of Space (another approach): > >>I'm sure the SMP/E Team would appreciate ny suggestions you might have. > >They have proven to be unwilling to fix a poor/broken design in the >past so I doubt they would be willing to fix this one (even if a >better design was suggested). > >My favorite broken/bad design is the way RESTORE works. If you >need/want to RESTORE a SYSMOD, SMPE restores that SYSMOD from the >DLIBs BUT runs the APPLY chain until it has restored not only the >elements from that SYSMOD you are restoring but also SYSMODs that are >SUPPED/PRE'ed until you get to a situation where there are no more >APPLIED mods that contain the elements. > IBM's defense of this amounts to "WAD".
>If you are trying to remove SYSMOD1 which contains Elements 1, 2, and >3 than all that is needed is to reapply Elements 1, 2, and 3 from >whatever SYSMODs they resided in if SYSMOD1 was no APPLIED. Right now >you may end up RESTOREing a number of SYSMODs (which then need to be >REAPPLY'ed without SYSMOD1 to do the fallback) since the copy of >Element 1 you are using for the restore is in a SYSMOD along with >Element 4 and the same thing applies to Elements 2 and 3 (and now 4). > There are circumventions. THey are cumbersome and/or irreversible. >IOW: A RESTORE should be preformed as an APPLY of just the replaced >elements that are in the SYSMOD being RESTORED without needing to >restore ANY OTHER Element. > Agreed. The information to do that is available in the SMPPTS. SMP/E simply declines to employ it. -- 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
