Today, Carol Darnell <[log in to unmask]> wrote:

>We have in fact experienced this situation with master sets - not as a
>result of actually using dynamic expansion directly, but of using 3rd party
>software to perform sampling and capacity adjustments based upon the
>sample.  The Initial (HASH) value remained non-zero, though the increment
>is zero and Dynamic Expansion is turned OFF (via DBUTIL SHOW database
>CAPACITY).
>
>With our very small windows of opportunity to perform maintenance, and very
>large number of IMAGE databases, I use a different method (a compound of
>MPEX expression coding and ADAGER) to manage capacities.  In fact, the
>current release of ADAGER does see this as corruption, as indicated in the
>initial note, at least when attempting to adjust the capacity of the
>master.

Step 1 (relatively short):  Detecting the anomalies.


>I suggest that anyone using the current release of ADAGER to adjust such a
>set.....read the following CAREFULLY lest you have more headaches than you
>want:

Step 2 (a bit longer):  Correcting the anomalies.

My initial approach (i.e., "the current release of Adager") was not too
clever, so Carol kindly warns about potential pitfalls (can pitfalls even
be potential?)


>There is a way (thanks, guys) to re-initialize the Initial value to zero
>(give a *new capacity*,0,0 to ADAGER),

So far, so good...


>but if you supply the ,0,0 to a set
>that has a zero Initial value, your capacity will not be changed.

Mea culpa.  This particular boundary condition managed to sneak by.


>Forcing
>a capacity change on a set with the Initial (HASH) non-zero, without
>that ,0,0 isn't a good idea.  At all.... nope.... don't even think about it.

Sigh...


>We have been working on the fix for this, and today I'm receiving a beta
>version of ADAGER which detects and corrects this situation automatically,
>without resorting to the ,0,0 coding.  Since I've already seen it perform
>as requested, I think I may again be able to sleep through the night from
>time to time!

This is the way that I should have designed this particular section of
Adager.  Live and learn.  Thanks to Carol, I learned quickly.  Correcting
the software was the easy part.  The thinking (and the long hours of
Carol's experimenting) behind the whole "treasure hunt" turned out to be
the hard part.  Thanks, Carol!

Sleep tight,

   _______________
  |               |
  |               |
  |            r  |  Alfredo                     [log in to unmask]
  |          e    |                           http://www.adager.com
  |        g      |  F. Alfredo Rego
  |      a        |  Manager, R & D Labs
  |    d          |  Adager Corporation
  |  A            |  Sun Valley, Idaho 83353-3000            U.S.A.
  |               |
  |_______________|

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *