Subject: Re: system hang
To: None <port-i386@netbsd.org>
From: Germano Cesari <germano.cesari@tesoro.it>
List: port-i386
Date: 10/29/2002 14:52:45
ok ok, but (I said I was a newbie :-)) ok the system may be thrashing,
but shouldnt it just kill the proc responsible for it, instead of hang
dead? how can a 'simple' process crash everything? (really, Id like to
understand)

anyway...

**********WARNING*****************
beware: stupid questions ahead :-)
**********************************

1) I tried mounting mfs on /tmp, but as I did so every application
couldn open display :0 (!!), even if I unmounted the mfs, I had to
reboot... 

2) the /tmp is "in use" it seems, since it is not empty... I tried
making a new dir and mounting mfs on it, but the problem at point 1
persisted

3) heres my /etc/fstab

/dev/wd0a / ffs rw 1 1
/dev/wd0b none swap sw 0 0
/dev/wd0e /usr ffs rw 1 2

but df output is:

/dev/wd0a [blah blah blah...]
/dev/wd0e [blah blah blah...]

mount says:

/dev/wd0a on / type ffs (local)
/dev/wd0e on /usr type ffs (local)

shouldnt df & mount report slice b? if not, IS the system effectively
using slice b as swap space?

thnx :)

Germano


On Fri, 2002-10-25 at 21:35, Juergen Hannken-Illjes wrote:
> ... and be sure the file system containing TMPDIR is not mounted
> with softdeps.
> 
> On Fri, Oct 25, 2002 at 07:23:10PM +0200, Jaromir Dolecek wrote:
> > Your system is likely thrashing. Don't you happen to have MFS /tmp?
> > If yes, try to set TMPDIR=/var/tmp (and export) in your environment
> > before running vmware.
> > 
> > Jaromir 
> > 
> > Germano Cesari wrote:
> > > Hi, just a newbie question:
> > > 
> > > Im using vmware 2.0.4 that comes among the packages of the NetBSD 1.6 6
> > > CD distribution, I actually didnt have any problem installing linux
> > > emulation and running netscape, acrobat reader or even vmware, but
> > > there's something that really puzzles me...
> > > 
> > > Im running NetBSD on a PIII 733 with only 128Mb, and IFF I set the total
> > > memory of a vmware virtual machine at > 32Mb, the ENTIRE NetBSD system
> > > would start slooooowing more and more, ignoring any mouse movements,
> > > till a final total system hang (no mouse, no keyb, clock frozen, no hdd
> > > activity, NO REMOTE CONNETCIONS, and so on, you got the point I
> > > guess...), where a power off is the only solution?
> > > 
> > > just to know, can some1 explain me how is this possible? I mean: that a
> > > process can cause a total system crash? I tought the kernel had control
> > > over every process running, eventually taking care of memory alloc. &
> > > the like, preventing runaways from crashing everything... now, how come
> > > vmware is able to get control over the entire system? has it something
> > > to do with the lkm(s) vmware installs? do they "modify" the kernel in a
> > > way that allows vmware to gain hardware control in a way not usually
> > > possible for other processes? (Ive noted vmware accesses hardware device
> > > independently of NetBSD, "directly" Id say)
> > > 
> > > if so, would it mean that running vmware (or loading his kernel 
> > > modules) can make the entire system unstable? is it possible to write a
> > > kernel module that modifies the kernel making a total system crash
> > > possible (or even achievable)?
> > > 
> > > thnx in advance :)
> > > 
> > > Germano
> > > 
> > 
> > 
> > -- 
> > Jaromir Dolecek <jdolecek@NetBSD.org>            http://www.NetBSD.org/
> > -=- We should be mindful of the potential goal, but as the tantric    -=-
> > -=- Buddhist masters say, ``You may notice during meditation that you -=-
> > -=- sometimes levitate or glow.   Do not let this distract you.''     -=-
> 
> -- 
> Juergen Hannken-Illjes - hannken@eis.cs.tu-bs.de - TU Braunschweig (Germany)