I'm not entirely sure I understand your question but I'll try to explain how DFSORT uses Hiperspace and maybe that will help.
As part of DFSORT initialization, the characteristics of the sort are evaluated along with available resources and a decision is made about what type and how much storage to use. If a decision was made to use Hiperspace storage than a Hiperspace is created. At this point the Hiperspace is empty which may be why you see it as "swapped out". As the DB2 utility passes input records to DFSORT, they are written to the Hiperspace. I'm guessing this filling of the Hiperspace might be what is looking like page ins to you. Once all input records have been processed, DFSORT then begins to read the records back from the Hiperspace and pass them back to the DB2 Utility in sequence. Normally we use RELEASE=YES so that the pages are released as they are read. At completion, DFSORT then deletes the Hiperspace. Have a nice day, Dave Betten DFSORT Development, Performance Lead IBM Corporation email: [email protected] 1-301-556-0943 DFSORT/MVSontheweb at http://www.ibm.com/storage/dfsort/ IBM Mainframe Discussion List <[email protected]> wrote on 01/23/2009 01:32:12 PM: > We are trying to understand how hiperspace is being used. z/os 1.9. Using > DSNUTILB and it's invocation of df/sort, it appears that when the hiperspace > is initially created, it is immediately swapped out, then demand paged back > in. When I look at type30 records, or in MON3 we see pagins, but never any > page outs, nor any swap outs during the step duration. in MON3, we see > sometimes 1K's worth of pages in DIV status, but again, we never see any > page/swap outs. What am I missing or where should I be looking ? > > ---------------------------------------------------------------------- > 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 ---------------------------------------------------------------------- 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
