Re: IFS object ownership

On 12-Oct-2010 19:17, James H. H. Lampert wrote:

<<SNIP>>

Tell me, just out of morbid curiosity, does anybody know if
anything bad happens (whether bad as in "may cause a mild
headache" or bad as in "don't cross the streams") if an IFS
object's ownership is transferred to QDFTOWN? Or to QSYS?

<<SNIP>>

The QDFTOWN is a system user profile which exists to implement both ownership recovery [part of RCLSTG] and assignment for create on restore [and probably a few other rare cases by database, for example] for ownership-by-name where that named *USRPRF does not exist, all as OS operations. If the QDFTOWN user profile is used for other [non-OS] purposes, the objects owned by the "default owner" could no longer be inferred to originate only by the OS. Such non-OS use could also effect excessive burden on that operating system UsrPrf for the various authority\owner tracking, and thus affect the [table\object size] limits on that tracking; see the PRTPRFINT noted by Vern. So if not used to excess, and corrections of ownerships by QDFTOWN [i.e. changing owners of all owned objects to a more appropriate owner] are performed periodically, such non-OS use for QDFTOWN would probably not even require an aspirin :-) Regardless, I suggest creating your own user profile to use.

FWiW, using QSYS is probably the worst possible choice, both for potential limits impacts [which any excess could seriously impact functionality of the OS] and for trying to figure out why those objects are owned by QSYS. At least for the QDFTOWN, you know that all of its owned objects are /improperly/ assigned, and thus require some corrective [CHGOWN or CHGOBJOWN] recovery action.

Regards, Chuck