[rescue] sun3/50 swap

[schaecsn]

Hi!

> On August 12, schaecsn wrote:
> > I do have a SunOS4.1.1U1 question (sun 3/50).
> >
> > My machine crashes under "heavy" load, e.g. 3..5 simultaneous compilier
> > runs.
> > I think the reason is I do have no virtual memory available.

O.k., in the meantime I am pretty sure that I do have virtual memory
enabled (4MB physical and 48MB swap).

I can always reproduce kernel crashes under "all kernels", e.g. GENERIC,
GENERIC_SMALL and a small custom kernel. I have also installed SunOS a
couple of times to make sure that I have a "good" installation.
Furthermore I haven't touched any files. I just installed some random
stuff under /usr/local, but even before installing anything I got these
crashes.

 
>   This is pretty odd.  I must admit I don't have any great ideas right
> off the top of my head...and I've seen just about every possible
> failure mode of SunOS4 on Sun3 hardware.  By "crashing" I assume you
> mean the kernel panics and dumps...if this is the case, are you able
> to capture the panic message on the console?  This would be most
> helpful in diagnosing the problem.  Please look into this if you can
> and I will try to help further.


precisly: when starting several programs (in my case compilers since I
am right now in the process of setting up a working environment) usually
the programs will crash. they die by "random" signals.

this is one choice:
  gcc: Internal compiler error: program cc1 got fatal signal 6

according to /usr/include/signal.h:
  #define SIGIOT  6       /* IOT instruction */
  #define SIGABRT 6       /* used by abort, replace SIGIOT in the futu


sometimes they die by a signal AND crash the kernel, too.

a real killer is copying large files over NFS. just copying
gcc-2.7.2.3.tar (28 MB) is possible. compiling 1 or 2 programms in
parallel crashes the kernel for sure.

When the kernel crashes I get - what is probably - a stack trace of
several screens. so I can't capture it ;-(

it looks something like this:
  called from e0048484 fp=ffff54e4 args = 4 8 0 7
  [etc etc and so on]


I am also pretty sure that I do not run out of memory. top confirms that
the cp command over NFS is just a view 100KB in size and a typical
compiler run is between 1MB and 2MB.

I do understand that is not too much information ;-( it's so mysterious!

--

oh, and finally a probably easier to answer question. I use usually
telnet from my intel-linux box to log into my sun. The terminal
emulation isn't working properly. programs assume a much bigger screen.
when I read manpages or start an editor these programs assume a
screensize of 80x64 but my xterm on the linux site is only of size
80x24.

whose fault is in anyway? linux' or sun's?


below are all environment variables


BASH=/bin/bash
BASH_VERSION=1.14.7(1)
COLUMNS=80
EUID=0
HISTFILE=/root/.bash_history
HISTFILESIZE=500
HISTSIZE=500
HOME=/root
HOSTTYPE=sun3
IFS=    

LINES=65
LOGNAME=root
MAILCHECK=60
OLDPWD=/root
OPTERR=1
OPTIND=1
OSTYPE=SunOS4
PATH=.:/bin:/usr/local/bin:/usr/openwin/bin:/sbin:/usr/sbin:/usr/ucb:/etc/:/usr/
etc:/usr/games
PPID=776
PS1=\h:\w\$ 
PS2=> 
PS4=+ 
PWD=/usr/local/junk
SHELL=/bin/bash
SHLVL=1
TERM=xterm
UID=0
USER=root
_=ls

there is an environment variable LINES. when setting it via "export
LINES=24" to the proper value of 24 programs still assume 80x65. any
help?

--

ah, one more question.


After turning on my computer I would like to boot directly into
sd(0,0,0), but I just get an "invalid device = ''" and I need to boot
"by hand" (b sd(,,)). any suggestions?


thanks
- Stefan

   ___  ___       __         |   Name: Stefan Schckeler
  /  / /  /____  (__)__  ___ | E-mail: schaecsn at gmx.de
 /  /_/  /  _  \/  /\  \/  / |    day: Solaris 2.5.1, HP-UX 10.20
 \___,__/__//__/__/ /__/\__\ |  night: slackware 3.0, netbsd-alpha 1.5