I think that z/OS and its immediate predecessors already make most of
what is needed available.

Only in special circumstances do I see or, if I see them, read
Shmuel's posts; but I found one---It is one of his characteristic
"Nonsense!" posts---in this thread.  It was directed at an earlier
post of mine, and it is therefore a convenient starting point for
discussion.

My comment that elicited his "Nonsense!" response suggested very
briefly that the device of making and marking a load module or program
object reentrant or, better, refreshable could be used to share code
across LPAR boundaries.  His response was that things do not work that
way.  If he had instead written that they do not always work that way,
I should have passed over his post silently, as banal but correct and
certainly inoffensive; but he elected to go for the jugular even
though, as he must know, his knife-fighting skills are now much
diminished.

For many years, since OS/390 V4, I have used a scheme that adds sets
of refreshable modules to the LPA dynamically in order to share them
among several LPARs.  It has always worked well; it continues to work
under z/OS 2.1 as it has worked in the past; and the facilities of
CSVQUERY remain available for ensuring that a dynamic LPA addition has
succeeded.

Multiple resident copies of an excecutable can sometimes be highly
problematic.  From perhaps 20 years ago I recall a situation in which
an intellectually challenged client programmer had replaced an old,
very simple but heavily used assembly-language data-vetting AP used
under CICS with one of his own devising written in RPG, which could
not then be made even quasi-reentrant under CICS, with truly
disastrous results.

Such problems are ordinarily easy to remedy individually when, for
whatever reason, they occur; and when they do occur they cannot be
ignored.

That said, Timothy Sipples is clearly right in general: The use of
significant system-software resources to identify and share
bit-equivalent pages, even big ones, dynamically is not likely to pay
for itself in a z/OS environment.  The major optimizations are already
in place.  The z/OS ASM does not, for example, page out unaltered
pages after doing so once initially; and it is able to make this
distinction efficiently because there is hardware support for
distinguishing altered/stored-into pages from unaltered ones.

John Gilmore, Ashland, MA 01721 - USA

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to