rjs@okstate.UUCP (03/24/87)
Hello,
Here at Oklahoma State University, we have been using 6 Intel 286/310
systems running Xenix 3.0 & OpenNet (over ethernet) for about 6 months
now and we are having problems. These machines are being used for under-
graduate class work (~150 student accounts) with all 8 of their RS-232
ports running at 9600 baud. The problems seem to arise when we have
3 to 4 people loged on to each machine. The students are given no
idea what machine they are loging onto. Rather, we had expected OpenNet
to handle the load. This however seems to be incorrect.
During the peak load times (day before an assignment is due :-),
one system will randomly lock up and take the entire system down (as
described below). Typing on the console will do nothing. However,
the other machines seem to continue to function as if nothing is wrong.
However, when the next student logs on wanting to access files (over
OpenNet) to the down machine, the machine they are loged onto halts,
locks up and becomes the next one to die. This procedure goes on and
on until all 6 machines are dead as door-nails.
If the first death is caught in time, you can reboot that machine
and then things are fine for a time. However, after about 30-50 minutes
the next machine in the sequence from the name server out will die.
Then the whole process starts over again.
Has anybody else who has these machines heard of something similar?
Is anyone at Intel reading this? If so, we have OpenNet R1.0 (U1.0),
Xenix 286 3.0, 512K of memory, and one iSBC 188/48 per machine.
Help!!!!!!!!!!!!!!!!
Please reply directly and thanks in advance,
Roland J. Stolfa
Department of Computing and Information Sciences
Oklahoma State University
UUCP: {cbosgd, ea, ihnp4, isucs1, mcvax, pesnta, uokvax}!okstate!rjs
Internet: rjs@a.cs.okstate.edu
Disclaimer: You have lost your MIND if you think ANYBODY speaks for
this place!root@sdd.UUCP (03/26/87)
When I used to work for Unisys, we had similar problems using Altos 986 family CPU's and their worknet. The problems that we had were that the machine that one was talking to throught the network would get overloaded, and the network server would get swapped out to disk, packets requesting information would time, out causing worknet to eventually fail. One potential option that you have, that might solve your problem is to add additional memory to your machines or ensuring that your server never gets swapped out by setting the sticky bit on it. Good luck, our final solution was to move to more powerfull machines! Daniel Corbett VP Engineering SDD
mac@uvacs.UUCP (03/30/87)
> During the peak load times (day before an assignment is due :-), > one system will randomly lock up and take the entire system down (as > described below). We have encountered lockups on iSBC552 Commputer Ethernet boards, running iNA961 inder RMX86. Under RMX86 you get a timeout when the board fails to respond.