More on Crash Exit
14 messages in this thread
G – (G for Gary?)
I have solved the crash problem, but only by removing my Novell 3.11 IPX and
NETX drivers from the system. Are there any known problems with this version
Netware or is it a conflict with my IPX/SPX v3.02 or shell v3.26 versions. My
IPX is loaded at interrupt 5, base 360 and no DMA or RAM.
????????
JD.
(yup, Gary) We've got a large number of sites using Novell 3.11
(including our internal lab at Autodesk) and it works perfectly. I'm
not a Novell expert but Paul Lind, Jamie Clay and a number of other
folks here on the forum are. Perhaps they can help with your
questions.
Thanks, since there are not any known problems I will start
reconfiguring my IPX, etc. right away and see just what configuration
is going to work. I appreciate the help and information! I am
really enjoying R3, by the way when is R4 expected <G>.
You'll get R4 right after you've gotten R3 configured correctly! <g>
No known problems that I'm aware of. There is always a possibility
of memory conflicts that need to be weeded out, and you should be
able to get your setup working again with the drivers if you add them
experimentally.
Nik,
Thanks, I am starting to reconfigure the whole system as we speak.
JD
There are no bugs or problems that we've run into using Novell 3.11
or 4.0. The IPX and NETX drivers work fine. Make sure your not
running into memory problems. We've run R3 across 53 machines
simultaneously under 3.11 and we will be running it simultaneously
across over a hundred within the next 45 days. With the load, I can
assure you that R3 runs flawlessly. As always network integrity is an
extremely high priority.
Thanks for your response. We are currently trying different
configurations of our memory allocations and I am certain we will
work out this "little" problem now that we know that it's the memory
setup and not hardware, Novell or R3 problems.
>>… and we will be running it simultaneously across over a hundred
>>within the next 45 days.
YES! You win the rendering speed award, Mr. L! Can I send you a
little 4,000,000 polygon, 10,000 frame animation to render?
We just got a new warehouse facility to move exverything into, so we
should have a fairly phenomenal rendering complex when it's all done.
It sure is nice not having to buy 100 copies of 3D Studio.
Even though it hasn't been brought up yet, I just want to offer my
deepest heartfelt thanks that the Yost Group and Adesk decided to
include such a great network renderer in R3. It is truly a work of
art unto itself, and could have easily been sold for several thousand
as a seperate package. Or on a sliding scale for the number of nodes
supported. The fact that it was directly included in the package
coupled with the unbelievable new features of R3 proves without a
doubt that both the Yost Group and Adesk are committed to bringing
workstation level animation/rendering to the masses.
Everyone should easily be able to go head to head with workstation
based firms and know that they're playing on a pretty level field.
Thaanks Again! Paul
Are you _sure_ you're not having any technical problems running all
those machines?
Problems, we had alot of problems …… Power problems, heat
problems, cabling problems, etc. We managed to get it sorted out by
dividing the units into different subsytems across the net and
putting them into different rooms. I don't currently have a method of
running them all continuously, that's why we're moving to larger
space. We are running 8 continuously for the proposal animation and
haven't had a single problem running days on end.
Sure ?????? Well I guess as sure as I can be at this point. The
original test was just to see if it could be done, and wasn't meant
to check all of _your_ network stability. We were choking the net
with the file transfers but we've devised a workaround for that. It's
a fairly large undertaking to bring them all up. I sincerely hope I
don't run into any problems with it when we reconnect and bring it
up. I'm going through "ALOT" of hassle and expense bringing them all
together under one roof. I need it to work right for a large contract
I'm bidding. Is there a specific reason, I should recheck everything
before I bid this? At the moment there are only two ways to do this.
Invest in the rest of the machines and run R3 across a large net, or
put the money into several VGXTs. The problems with the latter is
that I have to get alot of copies of an SGI package.
We have decided that if it became a problem with the file transfers,
we'd break the system into five distinct nets of ~20 machines each.
Lock them to a central hub and then send tasks to the individual
lans. Other than that I don't forsee any problems. ALthough I've had
sleepless nights thinking about it, "it's sorta untested", I am
fairly confident that it can be done. At least I was until the chief
architect of the program started asking about problems.
Maybe I should approach this project on a tried and tested platform?
Lets put it this way, I have three of the "best" Novell techs in the
world, if we encounter any problems, I'll be sure to voice them
extremely loudly here on CIS.
P.S. – I certainly hope we don't run into problems, Adesk has been
advertising 9,999 nodes with one copy. You know I'm fairly neurotic,
now I won't sleep for a while.
.. 4,000,000 polys x 10,000 frames…
time to RISC it!
100 machines sure will help on that 10,000 frames but you will
still need upwards of 120mb of RAM on each of those machines to
handle the 4,000,000 polygons…