Path: utzoo!attcan!uunet!ncrlnk!ncr-sd!hp-sdd!hplabs!sm.unisys.com!cs.utexas.edu!rutgers!mailrus!wasatch!cons.utah.edu!kessler
From: kessler%cons.utah.edu@wasatch.UUCP (Robert R. Kessler)
Newsgroups: comp.unix.xenix
Subject: Re: SCO Xenix System Hang
Message-ID: <790@wasatch.UUCP>
Date: 13 Dec 88 14:52:45 GMT
Sender: news@wasatch.UUCP
Reply-To: kessler%cons.utah.edu@wasatch.UUCP (Robert R. Kessler)
Organization: University of Utah CS Dept
Lines: 51


Oops, this should have gone to the whole mailing list:
To: jbayer@ispi.UUCP
Subject: Re: SCO Xenix System Hang
Newsgroups: comp.unix.xenix
In-Reply-To: <347@ispi.UUCP>
References: <766@wasatch.UUCP>
Organization: University of Utah, Computer Science Dept.
Cc: 
Bcc: 

In article <347@ispi.UUCP> you write:
>In article <766@wasatch.UUCP>, kessler%cons.utah.edu@wasatch.UUCP (Robert R. Kessler) writes:
>> We are having problems with SCO 386 Xenix and are looking for some help.
>> 
[Lines deleted]
>> terminals.  When a user runs our software, they all login with the
>> same user id which starts executing our own user interface shell
>   ^^^^^^^^^^^
>This is a very bad idea.  We had a customer who did the same thing.  The
>system would hang if too many people would log on using the same id.

Did you see the same kind of hang that I outlined?  Where it just died
after a while, or did it hang when someone tried to login?

>Make sure they use  different ids.  Otherwise it seems that internal
>system tables  are getting filled up.

As an experiment, that would be easy to try.

>
>Also, how much memory is in the system?  We figure that the base system
>needs about 1 meg, and then we add 512k for each additional terminal.  It
>seems to work.   As I understand RM/Cobol is a memory hog.
>
That particular system has 4 Meg.  The COMPAQ that I was testing the
COBOL program that hung with concurrent access has 5 Meg (and I could
hang it with only 2 users).
>
>Jonathan Bayer
>Intelligent Software Products, Inc.

By the way -- as suggested by Paul De Bra, I installed a script which
slept for 5 minutes and then dumped out a ps -el into a file.  It was
installed and then ran waiting for the system to hang.  It ran for 6
hours without a system hang -- a new record for length of time under
full usage.  They had to reboot the system to leave for the evening,
so it never did hang.  We are rerunning the experiment today.  Boy,
what a strange coincidence.

Bob.