Walt and Bob, thank for your input. Regarding the Walt's comment:
I agree with you ... My problem is the opposite ... We are already page fixing the buffers for a long time .... And also using LFAREA for 1 MB pages for DB2... I consider it a good thing ... One of the members of the DBA team does not agree on that, and this can prove that even the repetitive issuance of the pagefixing mechanism is wasteful if you do it and undo it zillion times. To the Bob's comment. I totally agree with you regarding the overal picture ... The only comment is that may be a page fixing is trivial compared with a page fault, but I noticed that the PGSER itself with the BRANCH option executed repeatedly accumulates a respectable amount of execution time ....With the SPECIAL option is less expensive but still a number. I will try next week the IARV64 service (In my case also the machine is not paging so the only thing I was thinking about is about the cost of fixing/unfixing) Mauri ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
