Once a dataset is created, the dataclass no longer controls its attributes. The issue is not that the datasets have the wrong dataclass. The issue is they don't have extended addressability.
The datasets did not spring into existence by themselves. They were defined and allocated at some point. At that time, either the wrong dataclass was used or the dataclass did not specify extended addressability. In either case, you cannot fix the problem "in place." If your DB2 guy won't let you redefine them, then he must be willing to take the hit when they fail to act as an extended addressable dataset (such as attempting to exceed 4GB). If you are on the hook for whatever problem occurs, he shouldn't get a vote. An extended addressable VSAM dataset must be SMS managed so playing with non-SMS volumes won't help. Since the datasets are apparently in use now successfully, why do you want to change them? If the need is sufficient, the DB2 guy has to face reality. > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] > On Behalf Of O'Brien, David W. (NIH/CIT) [C] > Sent: Saturday, April 06, 2013 9:09 AM > To: [email protected] > Subject: Change dataclass for existing DB2 datasets for V10 > > Looking at the archives and remembering the thread, dataclass cannot be > changed using DFDSS Copy. > > I've got a situation where some of my **.*db06.** datasets have the > correct dataclass with extended addressibility and some do not. > > Have any of you got any suggestions on how I might force re-allocation > of the problem datasets? > > I was thinking I might Copy them to non-sms volumes but I don't know if > that would remove the incorrect dataclass value. > > I suggested repro/delete/re-define but the DB2 person I'm working with > said No, we don't define these files. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
