On 09/04/2014 12:30 PM, Storr, Lon A CTR USARMY HRC (US) wrote: > Classification: UNCLASSIFIED > Caveats: NONE > > Hello List, > > I'd be interested in feedback regarding allocation in tracks versus cylinders > for certain types of high-usage datasets. > > I believe that there may be certain instances when allocation in units of CYL > is beneficial. One example, I believe, is a PDS that has a multi-track > directory: a single channel program can search up to a CYLinder at a time. > Another example, I believe, is a VSAM dataset allocated in CYLinders will > receive a CA-size of one CYLinder. A benefit similar to the first, if it even > exists, would be achieved by caching the PDS directory in some way (e.g. > BLDL), as I'm sure many system software applications already do (e.g. LLA and > ISPF). > > Are there still pertinent benefits to allocating certain types of datasets in > CYLinder increments? > > Thanks, > Alan > Just as a side comment on VSAM allocation: I believe if the primary and secondary allocation quantities are specified in TRACKS or RECORDS but are both greater than the size of a cylinder, then a CA of 1 cylinder results and allocation will be on cylinder boundaries with primary and secondary quantifies rounded up to multiples of a CA if necessary. It isn't necessary to explicitly specify VSAM space in CYLinders to get Cylinder-boundary allocation and a 1-cylinder CA.
-- Joel C. Ewing, Bentonville, AR [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
