On 12 Jul 2007 17:13:16 -0700, in bit.listserv.ibm-main you wrote: >On Jul 12, 2007, at 3:09 PM, McKown, John wrote: >-------------- >SNIP-------------------------------------------------------------------- >- > >> >> Same here! Their reason: "I understand what a track and a cylinder >> is!" >> Like they didn't understand allocation by number of records for some >> reason. Or in MiB or GiB. Nope, gotta be in tracks or cylinders >> because >> they understand how big that is. HUH??? >> > > >John, > >I have heard that its easier to figure out trks/cyl than number of >records. I always ask back why? Now I may understand the first run >(maybe) after that you should know exactly (or at least to closest K ># of recs). I am convinced that its just laziness they just put in a >number of cyl(s) and its a guess. I think that they just am afraid >that you are going to hold the number of recs to the exact number and >too bad if it goes over. > >Ed > While this could work well for sequential files where you don't have to worry about CA size, it can bite you with VSAM because you can get a less than optimal CA size. The CA size is the smaller of the primary or secondary allocations when either is less than a cylinder. While this doesn't matter when the secondary is over a cylinder, it can cause unexpected results with primary equating to N cylinders and secondary to only M tracks. And there is no way force optimum CA size. Hence I specified VSAM in track or cylinders.
Clark Morris ---------------------------------------------------------------------- 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
