David Jousma wrote: >I'm not a DFSORT expert, but have some questions(an ETR has Already been opened with IBM on this). We recently Started adding MOD27's to our >SMS management pool that is used For SORTWKxx allocations. Jobs that worked in the past now fail Because DFSORT cannot switch to BLOCKSET >mode because of some Other issue, which when coding SORTDIAG, DFSORT does tell us. It just Seems to me that DFSORT doesn't handle >MOD27's too well when jobs Coded and run for years that work fine with SORTWKxx allocations to 3390-3 or 3390-9 work volumes work just fine. >...
David, Blockset is far and away DFSORT's most efficient sorting technique and has become even more so over the years. We made a business decision a long time ago NOT to enhance the other much less efficient techniques since they degrade performance significantly. So for all but the simplest, oldest jobs, either Blockset is used or we terminate. For compatability, we let some jobs with "odd" situations still get through to the other techniques, but we strongly recommend that they be changed to use Blockset. You are NOT doing yourself any favors by using these techniques. Even if your jobs using these techniques happen to work, they are working very inefficiently. Blockset provides full MOD27 support for SORTWKs. The other techniques don't. Appropriate error messages are issued when Blockset can't be used with MOD27s. The first line of attack for a job that doesn't use Blockset is to change it to use Blockset if possible. If that's not possible, then you need to do what it takes to run those jobs with the other techniques. Frank Yaeger - DFSORT Team (IBM) Specialties: ICETOOL, IFTHEN, OVERLAY, Symbols, Migration => DFSORT/MVS is on the Web at http://www.ibm.com/storage/dfsort/ ---------------------------------------------------------------------- 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
