Mark,

Ummm.....24MB is a bit thin but if all is well and the environment
doesn't change it would seem that your ok as is.

From the Trudeau files of "Does certification cover this?".

For a couple of years we were running our production 987 (22Gb
disc, 9+Gb of Image databases, 30-40 concurrent users, misc
developers doing Lord knows what) with 192 Mb of memory.  I
knew this was just a tad low but "all was well".  Enter the
desktop people with new machines and software on a bunch of
desks and suddenly the 987 was getting more hung sessions
than you could shake a stiff cat at.

Never mind that it in retrospect looks like those machines had some
flaky config parms.  From my perspective nothing had changed as
no would told me this activity was occuring.

So what happend?  It is well (?) known that these RISC boxes are
much more susceptible to the infamous "knuckle" in the performance
curve.  That is, you add just a bit more load to the system somewhere
and response time goes "yaaaaaaaaaaa" right over the cliff.  This is
what we experianced.  We jacked the memory up to 384 Mb and all
has been well.  My only point being that this sort of degradation can
happen in a figurative heart beat.

Moral: You may get away for a long time with a short system resource
or two in a "controlled" development or whatever environment, but if
you push it on a "growing" system, particularly if the growth is in
spurts, be prepared for strange and unpleasant suprises.

Jim (hmmm, might rain today) Trudeau
Sr. muck about

> -----Original Message-----
> From: Mark Wilkinson [SMTP:[log in to unmask]]
> Sent: Friday, August 21, 1998 5:32 AM
> To:   [log in to unmask]
> Subject:      Amazed with our 927LX..
>
> I just ran SYSINFO and discovered that our development machine, a 927LX,
> has only
> got 48 meg of memory. It runs just fine and at times there are four of us
> happily compiling
> COBOL progs and Adagering databases, all at the same time.
>
> Stunned..
>
> Mark Wilkinson.
> Sony Pictures Entertainment.