Ron, With Sequential pre-fetch in all cache controllers nowadays I don't think you will see much of a difference between BLKSIZE=7000 and BLKSIZE=21000, unless this is a very large file. If you want almost equivalent performance then for SDB you can triple your buffers - DCB=BUFNO=15. Then you get better space usage with the same performance.
If you add DSORG=PS to your two SEQ files SDB will have enough information to know what to do and you will get the correct Block sizes (I just tested it). Coding BLKSIZE=0 on your sample JCL does not give SDB enough to work with either (I tested that as well). One of my soap boxes - best practice for SDB is not to specify BLKSIZE=0, it is to specify the DCB and omit the BLKSIZE, or specify the DCB and code BLKSIZE=0. Ron > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On > Behalf Of Ron MacRae > Sent: Wednesday, 15 February 2006 6:31 PM > To: [email protected] > Subject: Re: DASD allocation guidelines > > Guys, > That explains most of my issues. We have some fairly large LRECL FB > sequential work files and were finding them unblocked, which increased our > I/O quite a bit. So SDB always blocks for maximum capacity, regardless of > performance implications. I guess we have to manually block these files > via JCL. Or is there a way to tell SDB that you want to favor performance > over capacity for a particular dataset? > > I have another issue. Here is an example of some allocation JCL. > > //STEP1 EXEC PGM=IEFBR14,REGION=4M > //SYSPRINT DD SYSOUT=* > //DD1 DD DSN=OSBSYS.RM.TEST.SEQVB, > // SPACE=(TRK,(5,1)), > // DCB=(RECFM=VB,LRECL=80), > // DISP=(,CATLG,DELETE),UNIT=3390 > //DD2 DD DSN=OSBSYS.RM.TEST.PDSVB, > // SPACE=(TRK,(5,1,1)), > // DCB=(RECFM=VB,LRECL=80), > // DISP=(,CATLG,DELETE),UNIT=3390 > //DD3 DD DSN=OSBSYS.RM.TEST.SEQFB, > // SPACE=(TRK,(5,1)), > // DCB=(RECFM=FB,LRECL=80), > // DISP=(,CATLG,DELETE),UNIT=3390 > //DD4 DD DSN=OSBSYS.RM.TEST.PDSFB, > // SPACE=(TRK,(5,1,1)), > // DCB=(RECFM=FB,LRECL=80), > // DISP=(,CATLG,DELETE),UNIT=3390 > > The 2 PDSs end up half track blocked but both sequential files end up > blocksize=0 and so the files cannot be edited. Allocating the same > sequential files via ISPF 3.2 creates them half track blocked. > What other information does SDB need to block the sequential files > correctly in batch? > ---------------------------------------------------------------------- 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
