Without having cross-checked everything you are doing, I'll say that
most of what you have outlined should work in principle. Run APPLY and
ACCEPT against clones of your existing target and DLIB zones and
libraries. Of course, always take a backup first, too, because "stuff
happens."
See below for some thoughts on what ACCEPT level you want for the DVD's
DLIB zone and libraries before running BUILDMCS.
You need to run BUILDMCS for all the missing FMIDs for this feature that
are listed in the z/OS Program Directory, and then APPLY them to your
*existing* target zone (they are dependent functions and cannot be
installed without their corresponding base functions), using your
*existing* libraries, some of which might need to be expanded before
APPLY runs if they don't have enough free space.
The PD will show the original space deltas between the two in the space
tables either in or pointed to by Chapter 5. If you used a reasonable
amount of free space when you installed your z/OS ServerPac, you might
get lucky and skate! If you have to reallocate something, you can use
the ALLOCDS ServerPac job to find the allocation requirements (JCL,
IDCAMS commands, etc.) for each library. If you ran GENSKEL, look in
the order's SCPPBENU data set.
You *must* use the existing libraries unless you want to have a
discussion I'd rather not get into about how to tell whether it's
actually necessary; simplicity wins, and the odds are that you need to
do it anyway. This will result in an APPLY level for these FMIDs in
your target zone equal to the ACCEPT level you got on the DVD. (You
might have to take care of some requisite service along the way, which
could make them somewhat inequal, but that's OK.) Dead-current HOLDDATA
and GEXT are your friends. There is no need for post-APPLY manual
copies, as you will be installing the parts directly into the libraries
where things need to wind up.
Then, ACCEPT the FMIDs. Next, determine whether the overall APPLY level
matches either (a) what you got on the first ServerPac or (b) the level
you want. Both are unlikely, as the archive order was created at a
point in time not matching the original ServerPac order, and both were
build with different HOLDDATA and different available PTFs. So, some
APPLY processing for, e.g., RSU PTFs (with GEXT and including SEC/INT,
as I *know* that anyone as rational as you are must be signed up for the
Security Portal ;-) will likely follow. I'd shoot for current RSU,
HIPER fix, PE fix, and SEC/INT with GEXT, myself, but of course what to
choose is up to you.
BUILDMCS is what I recommend to SMP/E-literate sysprogs who need to
rebuild corrupted or missing target libraries, too. In that case, it's
entirely true that one must pre-ACCEPT at least all applicable service
(and "all service" is simpler!), but not in this case. This is not
regenerating a library with the required matching levels of all its
parts that have to work with other parts; it's more like installing an
additional product. Specifying GEXT on APPLY will make SMP/E figure out
what other PTFs are needed during APPLY and pull them in automagically
unless a HOLD stands in the way. After resolving any ERROR HOLDs (and
of course taking care of any SYSTEM HOLDs), you will have the matching
part levels you need to work and play well with the rest of the system's
parts.
It's a bit counterintuitive, but I think that in this case you really
want the *lowest* possible pre-BUILDMCS ACCEPT level. That can avoid
potential unremovable PEs, make it easier to get the whole resulting
target zone at more or less the same APPLY level, and leave open the
ability to RESTORE off (some) newer PTFs later on if need be (for
example, if they are PEd).
HTH...
Gibney, David Allen wrote:
By the silence, I guess everyone thinks this will work, or are too polite to
tell me I'm crazy and aiming at my foot.
<snip>
--
John Eells
IBM Poughkeepsie
ee...@us.ibm.com
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
