Re: Cross-library Logicals and RSTLIB/CRTDUPOBJ

[Mark Waterbury <mark.s.waterbury@xxxxxxxxxxxxx>]

 Hi, Joe,

Clearly, there are still plenty of way to "fool" the OS ... :-o

Renaming libraries during a restore seems to be one of them.

Did you try the (related) RSTDFROBJ command?

Also, I seem to recall that RSTOBJ was somewhat sensitive to the library list, as far as finding dependent objects ... perhaps you could create the new named library first, then set it as your *CURLIB or add it to the top of your *LIBL, prior to these restore attempts?

Otherwise, it might be worth opening a "case" (PMR) with IBM Support to see if they can shed any additional light on the situation.

Or, when all else fails, you can always just recreate those (few) join LFs from DDS source after the restore(s) are completed?


All the best,

Mark S. Waterbury

    On Wednesday, January 8, 2020, 10:16:59 AM EST, Joe Pluta <joepluta@xxxxxxxxxxxxxxxxx> wrote:  
 
 Well, I stopped using them 30 years ago, so I'm happy that something has 
changed.

But it doesn't look like that parameter works when you change the 
library names on the restore.

It just keeps telling me that DFRID REFRESH has one object remaining.


On 1/D8/2020 8:45 AM, Mark Waterbury wrote:
>   Hi, Joe,
>
> This was fixed by IBM around V5R4, IIRC ...
>
> Prompt the RSTLIB or RSTOBJ commands, and press F10=Additional parameters, then look for the (then) "new" parameter for "Defer ID . . . "  DFRID ...
>
> So, not a problem for 30 years, but only perhaps 20 years?  ;-)
>
> Happy New Year!   "Welcome to the future!"
>
> All the best,
>
> Mark S. Waterbury
>
>       On Wednesday, January 8, 2020, 9:21:10 AM EST, Joe Pluta <joepluta@xxxxxxxxxxxxxxxxx> wrote:
>   
>   Has anyone ever identified a process to handle either RSTLIB or
> CRTDUPOBJ for cross-library logicals: those pesky logical files that
> join physical files in two different libraries (they can also be created
> as SQL VIEWs these days).
>
> The problem is simple.  I can create a logical ABC1/LF over ABC1/PF and
> ABC2/PF.  But if I save and restore those libraries to a different name
> (say XYZ1 and XYZ2), I end up with a "broken" logical.  XYZ1/LF ends up
> being created over XYZ1/PF and ABC2/PF.   The same happens if I try to
> do a CRTDUPOBJ of ABC1/LF to XYZ1/LF.
>
> I have found no way short of recreating the logical file with CRTLF (or
> RUNSQLSTM for a VIEW) to link the logical correctly to both physicals.
>
> This has been a problem for 30 years, which is why I've since avoided
> them like the plague.  But I have a few of them I'm dealing with right
> now and they're quite annoying, and I was hoping something had changed
> over the years.  :)