The following is a summary of the responses. I hope somebody may be
interested to read it.
Weixing Zhang
**************************************************************************
Please install the Sun Patch 105802-07. This fixes the problem that
you described.
Here is the patch. The problem is that two ttsessions are
being spawned instead of one ttsession and one dtsession.
THIS IS NOT A VNMR BUG, but caused by a bug in an early
Solaris 2.6 patch (105802-06), see also Varian NMR News
1998-10-10!
**************************************************************************
I have a similar problem with my Ultra 1 with Solaris 2.6 and vnmr 6.1B.
But it only locks up if the screen saver is on. We did not have this
problem before we installed vnmr 6.1B and Solaris 2.6. I have contacted one
of Varian's app chemists about this, but have not heard back yet. I'll let
you know if find out anything, although your problem could be from a
completely different source.
**************************************************************************
When you see files leaf.tar or nick in the root directory and/or your root
directory has rw-permissions for root only for . and .. you are definitely
hacked as a few other facilities in the past two months.
You can get some informations also from Rolf Kyburz NMR-news.
**************************************************************************
I wonder if your /var or /tmp partitions have gotten full for some reason
(possibly even root?). Do a df -k with the remote login, and let us know.
There are various possibilities how these could get full, but let's see if
that's
the problem first.
**************************************************************************
I'm not sure about Suns, but on SGIs I would issue the
commands:
/usr/gfx/stopgfx; /usr/gfx/startgfx
Regardless of whether those commands work, it sounds like your
graphics daemon(s) need to be restarted, probably because of too
many (perhaps conflicting) demands.
If this becomes a constant problem you will need to track down the
cause and either allocate more memory or there is a slight chance
that there is a bug. I remember fixing a small bug on one of our
SGIs, but have forgotten what it was.
**************************************************************************
I do solids applications here in palo Alto -- I am not a software person so
my answers are not official but maybe I can help.
You may have tried this already -- but: I thought I'd mention it.
My Ultra will sometimes slow tremendously and seem to lock up. The cause is
usually a process that is not removed from memory and continues to run
slowing everything else down. Text editors left open will do this sometimes
and if you use dg screens with 6.1A they will sometimes lock up the system.
If you have not done this -- when the system locks up - before you reboot -
check the running processes with "ps -ef" in a terminal window. The cuprits
will show a large accumulated runtime. Try killing these processes before
you reboot. (kill "process number").
**************************************************************************
Ensure that you do not use automatic screen blanking when using VNMR under
CDE. It is a known problem. Change the screen blanking settings with the
style manager icon on the applications bar at the bottom of the screen. ...
We
are running 6.1B under Solaris 2.5.1.
Before we learned about not using the screen blanking, I would hunt and peck
with the mouse for the "exit" button on the applications bar to logout.
I hope that this helps.
**************************************************************************
Once I had a problem with an old Sparc station and the problem was diagnosed
as a cpu problem but in the end it was a HAL board lockup. I disconnected
the
spectrometer from the computer and the computer booted up no problem. Have
you tried disconnecting the spectrometer and then booting the computer?
**************************************************************************
I have a 85 Mhz Unity Inova from Varian and we have the same problem on
occassion. We have found two ways around this problem. A) in a terminal
type
ps -e and kill VNMR then go back and restart VNMR or B) in a terminal kill
acqproc then restart it, su acqporc. It appears to be a communication
problem
between the host and acquistion computers that comes up on occassion, but
never
at the right time or often enough to find the cause.
**************************************************************************
In July, many of our SUN computers had been invaded by outsiders. One of
the symptoms is that during login into CDE windows, the process would hung.
The only way to login is to choose Options: failsafe session. This invasion
seems to go through a security weakness in CDE calendar manager and was
documented in VNMR Newsletter (July/99).
Also, after I login with failsafe option and cat the /etc/passwd file, I
found some unauthorized users, such as "doomed", "d", "Gabriel Cortez", and
"lee".
What I did was to reintall Solaris from CDROM using the preserve option to
preserve NMR data partitions. Then I reinstalled Netscape from CDROM, got
the latest Recommended Sun patches from http://sunsolve.sun.com/ and
installed that. I also reinstalled VNMR.
It seems to me that you might have similar problems, although it is almost
1.5 months later. There are also network services in /etc/inetd.conf that
you may wish to turn off. Please check VNMR Newsletter for details.