Hi Ron,
Sure the HURBA will reflect high water mark, but I've worked with journey level folks who look at record count and disk allocation, lots of folks don't pay a lot of attention to HURBA. And our OP does identify himself as new. Of course, most of us are new at something or other. We had some of those heavy tape users - dodgers of the higher excp charges for DASD. We also had some that didn't mind that their tapes danced around the 3420 reel drives at every jobstep. Ran the operators ragged and until the billing folks stepped up with a fairly hefty tape mount charge, then programmers were rushing to code RETAIN. We had a really good performance/capacity specialist working with us for a while, and she always recommended adjusting the billing table to encourage changes that would be beneficial. It really was easier in a lot of ways. Regards, Linda ----- Original Message ----- From: "Ron Hawkins" <[email protected]> To: [email protected] Sent: Saturday, May 7, 2011 8:22:04 AM Subject: Re: Do we need to implement HSM Lisa, I thought about the record deletion, but even when there are only a few records left the HURBA is going to reflect the highest CI used. For VSAM KSDS there's a huge difference between empty and one record, and Ed didn't mention if it was an RRDS or ESDS. Clever programmers and billing - for some people it was as much fun as tax avoidance. Like applications running tape to tape because they were only charged for disk space. Ron > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf > Of > Linda Mooney > Sent: Friday, May 06, 2011 10:52 PM > To: [email protected] > Subject: Re: [IBM-MAIN] Do we need to implement HSM > > Hi Ron, > > > > What you say is true. I did read into the post a bit because I used to get a > lot of questions by folks who would look at the catalog entry and assume that > the candidates were already used. > > > > On the other hand, I have seen nearly empty, and empty, multivolume VSAM > datasets. Turned out that our billing system (at that time), charged for > purges and reorgs (there was a charge to run the job and charges for EXCPs) , > but not for DASD residency or space. So, a clever programmer, who really > paid > attention to the billing table to the benefit of his customer (I am not being > critical of him in the the least), would delete records or move them all to > a > history file, but never reorg the file. I t was a fixed length record, > non- > SMS controlled KSDS . Gen erally, at the beginning of the quarter, that VSAM > file would have at most a couple of hundred records, yet still span 3 > volumes. > > > > SMF records would reveal if Ed's dataset ever had records . IDCAMS or a CA- > DISK VSAM cluster report - would reveal if it is very poorly constructed and > contains maybe 95% free space. People do make some mistakes when defining > datasets. Many shops still use very little SMS. We have a mix. We use it > for some things, not for others. > > > > CA-DISK used to be DMS when Sterling Software had it. > > > > Regards, > > > > Linda > > > > ----- Original Message ----- > From: "Ron Hawkins" <[email protected]> > To: [email protected] > Sent: Friday, May 6, 2011 2:11:36 AM > Subject: Re: Do we need to implement HSM > > Linda, > > Without Guaranteed Space all except the first volume will only be candidate > volumes until the writing records to the first volume causes it to extend to > the second and subsequent volumes. > > The case Ed described is that the dataset is empty, but it had extents across > multiple volumes, which is not the usual behavior for an empty dataset that > does not use guaranteed space. Hence my curiosity. > > I agree completely with your reply, but it does not solve the empty dataset > riddle. > > Ron > > > -----Original Message----- > > From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf > Of > > Linda Mooney > > Sent: Thursday, May 05, 2011 8:22 PM > > To: [email protected] > > Subject: Re: [IBM-MAIN] Do we need to implement HSM > > > > Hi Ron, > > > > > > > > If a VSAM dataset is allocated with a candidate list of volumes > > VOLSER=(VOL001,VOL002,VOL003 ), the datset will allocate to the first > > volser > > in the list, then the next, etc. The catalog entry will show all of the > > volumes, even if there is no VTOC entry there yet . Back a few years ago, > > when we had much smaller volumes, we had some large VSAM datasets (non-SMS) > > that we allocated this way. > > > > > > > > Regards, > > > > > > > > Linda > > > > > > ---------------------------------------------------------------------- > 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 > > ---------------------------------------------------------------------- > 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 ---------------------------------------------------------------------- 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 ---------------------------------------------------------------------- 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
