On Mon, 17 Oct 2005 10:51:58 -0400, Bruce Black <[EMAIL PROTECTED]> wrote:
>> >> >>At one time there was a definite problem with tape if the LRECL was less than 18 bytes, as the data was treated as noise errors. >> >The limitation was on blocksize, not LRECL. > >I thought that noise records no longer applied to cartridge tapes, but >the IBM "Using Data Sets" manual still has this text > >"Minimum block size: If you specify a block size other than zero, there is >no minimum requirement for block size except that format-V blocks have a >minimum block size of 8. However, if a data check occurs on a magnetic >tape device, any block shorter than 12 bytes in a read operation, or 18 >bytes in a write operation, is treated as a noise record and lost. No >check for noise is made unless a data check occurs. " > >Data checks are much rarer on cartridge drives, so this rule may rarely >apply, if it really does apply to cartridges. > >-- >Bruce A. Black >Senior Software Developer for FDR >Innovation Data Processing 973-890-7300 >personal: [EMAIL PROTECTED] >sales info: [EMAIL PROTECTED] >tech support: [EMAIL PROTECTED] >web: www.innovationdp.fdr.com > >---------------------------------------------------------------------- >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 Yes, I do not think this applies either, especially since I tested using "virtual tape" and received the same performance. The blocksize on this file is 32760, that was also assigned using SDBS. BUFNO and OPTCD also had no noticable performance increase. ---------------------------------------------------------------------- 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
