REAL storage is the column to use. You can also get better values for this by looking at the completed jobs JESYSMSG DD for that step using the IEF374I STEP/ /STOP line Add the VIRT amount to the EXT amount and that gives you the total (and largest) amount of memory used (Ignore the two SYS fields). The VIRT is the amount of memory used below the 16M line and EXT is the amount used above the 16M line.
Might get better granularity that way. As far as paging goes, I cheat and always use the LSR option that forces the buffers into a no-paging mode (this can only be done on LSR, BLSR and IMS pools, I think). However if the system needs to page out the job when this is done (which is rare on our machine) the job is placed into a wait state then paged out of core entirely, then paged back in when the full resources are available again. -----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Farley, Peter x23353 Sent: Wednesday, May 09, 2007 3:14 PM To: [email protected] Subject: Re: How to measure actual usage of BLSR buffers? Thanks for the idea. A question though -- where exactly in the SDSF panels would I see the pages of core used by the job step -- my SDSF DA panel only has a column for REAL storage (SDSF manual says that column represents real page frames). That would only give me the working-set size, right? Although, since I notice my paging rate is practically always zero, I suppose that could be the virtual size as well. Also, that column is only 4 digits (plus comma), after 9,999 the format looks like " 14T ", which seems to be in thousands. Not very granular, but I guess that would put it within the megabyte-or-so that you suggested. Thanks again for the idea. Peter ---------------------------------------------------------------------- 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
