(dunno where that surrounding garbage came from, but the readable test is still good) John
-----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of John Compton Sent: 07 February 2012 14:33 To: [email protected] Subject: Physical record size query ??z{S???}???? ?xj???*'???O*^??m??Z?w!j? ?????A question from one of our non-mainframe people arose t'other day: "Is 'half-track blocking' a good thing in these days of RAID arrays?" (or words to that effect). And I don't know enough about RAID architecture to answer. I first learnt about half-track blocking when I was a SysProg on DOS/VSE/AF systems & when I had to deal with LIOCS and PIOCS. Since then, I guess I've sort of become 'married' to it (and, at the risk of getting badly flamed, I'd venture to say that most of us here use the technique 'by default', rather than putting any deep thought into the matter.) Whenever I can exert any control over a file allocation, I do my best to ensure that the physical record size is as close as possible to 27998 bytes. So the question stands... In situations where a RAID array is used for disk storage (as opposed to discrete devices, headed up by a controller), is half-track blocking: (a) worth bothering about; and (b) if it's worth bothering about, is 27998 bytes the best number? B?KKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKKCB????P?KSPRS? ?X????X?H ? ?Y????? ?\??]?HX??\??[?? ?X?[???B??[?[XZ[ ?\? ?\???[XK?XK?Y H ?] HY\??Y?N?S???P?KSPRS?B ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
