I'm running Win/UDPE 7.2 for development.  When running a program
with a few hundred K of variables being tossed around, the
connection running the process will die, sometimes just gasping
out this message:

Size of memory requested (146157568, 35683 local pages) is too
big!

The same code (adjusted for platform-specific nuances) runs
without a hitch on in another MV environment.  It looks like
Unidata is attempting to pre-allocate excessive amounts of memory
as soon as it sees some volume of actual usage.

I've tried tweaking my udtconfig values (restarting Unidata each
time).  I think the program runs a little longer before it dies
but I know I'm not tuning properly and the issue hasn't been
bypassed.

This isn't isolated.  One of our clients running a non-PE (6
user) environment reported the issue and luckily I was able to
reproduce it here.  I suspect none of the memory allocation
values in udtconfig have been changed there either.

I am careful to code without abusing resources, not
pre-allocating massive arrays or leaving large unused buffers
hanging around.  I'll check for sloppy code but I'm hoping
someone here might be able to point to a config issue.

Here are my latest udtconfig values - I'm not cataloging the code
for the test, so global cat SH size shouldn't matter.

SBCS_SHM_SIZE=1048576
SHM_MAX_SIZE=8388608
SHM_ATT_ADD=0
SHM_LBA=4096
SHM_MIN_NATT=4
SHM_GNTBLS=40
SHM_GNPAGES=256
SHM_GPAGESZ=2048
SHM_LPINENTS=10
SHM_LMINENTS=100
SHM_LCINENTS=100
SHM_LPAGESZ=8
SHM_FREEPCT=25
SHM_NFREES=1

Thank you kindly for any guidance - even a link to a Unidata
configuration doc.

Tony Gravagno
Nebula Research and Development
TG@ remove.pleaseNebula-RnD.com

Nebula R&D sells mv.NET and other Pick/MultiValue products
worldwide, and provides related development and training services
Visit the blog, lots of recent updates:
remove.pleaseNebula-RnD.com/blog 
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/

Reply via email to