On Tue, 11 Jan 2000, Mauricio Tavares wrote:

> 	This last week I have been experiencing a rather interesting problem.
> When I come to work in the morning, I try to access my dec 5000/20 (1.4.1)
> and nothing.  So, I run ttermpro and select the serial line, so i can go to
> console, and this is what I am getting:
> 
> 
> Bad character
> ?
> db>
> 
> That to me makes me feel the machine got into a core dump or something
> during the night.  As soon as I reboot the machine, all goes well
> throughout the day.  Then, in the next day I find the very same problem.
> What could be causing such behavior?  I had samba (not really configured),
> apache, and netatalk (configured to fit my needs).  Right now, I commented
> out the lines in rc.local that run samba just to see if it may be th
> esource of problems, but it really does not make sense.  


try entering 

  bt

at the db> prompt to get a backtrace of what was happening when the
machine crashed.  

A long time ago (NetBSD-1.1) I used to sometimes have problems like that
from the 'daily' script (or was that the 'weekly').  I never did quite
track down why and I haven't had the problem since then.  If the 'bt'
doesn't tell you anything interesting, you could turn off those scripts.
('crontab -e ' and then comment out the daily,weekly lines).

-Dan