|
[Date Index]
[Thread Index]
[Author Index]
.math22prefs.mb file getting corrupted
- To: mathgroup at christensen.cybernetics.net
- Subject: [mg640] .math22prefs.mb file getting corrupted
- From: Chris Williams <chris at hpcpbla.bri.hp.com>
- Date: Tue, 4 Apr 1995 09:38:57 +0100
Dear MathGroupers,
I am having a problem with corruption of the preferences file when running
Mathematica on a HP700 Unix system. The work I do involves running several
long term background (low priority) Math jobs, and one or two foreground jobs.
Here is a brief description.
System:
HP735/125, hpux 9.01,
Mathematic v2.2, single processor, unlimited number of processes type
license,
Conditions:
Several math sessions are open at once, from different windows
(possibly on different machines where my home directory is nfs
mounted). The sessions are all started asynchronously ie
mathematica &
Observations:
The .math22prefs.mb file seems to get corrupted, as once it has
happened I can't start any new mathematica sessions (I get an
error message saying can't find an important resource or I have a
bad or modified copy of the program). I can fix the problem
by deleting the .math22prefs.mb file. Of course I lose my kernel
definitions when I do this.
The corruption can result in a mathematica session core dumping, giving
the message Memory Fault. Or it can take out the parent window as
well (and of course any other processes spawned from it).
Once the .math22prefs.mb file has been deleted to fix the above,
existing jobs can produce messages like
"File /users/chris/.math22prefs.mb appears to be broken for TEXT.2701.0"
although I'm not sure of the consequences of this.
For the work I do, I really want the individual Math sessions to be robust,
and totally independent of each other. It seems that just starting them from a
separate window is not enough.
Does anyone know what causes the problem?
Does anyone know if it's possible to make Math sessions for a given user
more independent, say by forcing the use of a different .math22prefs.mb file
for each?
Regards,
Chris Williams
(Hewlett Packard CPB, R+D)
Prev by Date:
Mathematica 2.2.3 and WinNT 3.5
Next by Date:
Summary:Ways to get Odd Columns&Rows of Matrix
Previous by thread:
Re: Mathematica 2.2.3 and WinNT 3.5
Next by thread:
Summary:Ways to get Odd Columns&Rows of Matrix
|