Hi, This is one that's got me stumped (I am still fairly new at this,
however).

While using Netscape 4.51 (its the X11R5 version, if that makes any
difference)
under Debian "Slink" I tried to point the browser at one of the *.snm
files in
a new mail directory (I get the impression you're not supposed to do
this, but
how else do you learn?), and got some really disturbing behavior: every
open xterm
window started spouting error messages from "SYSLOGD" (accompanied by
terminal
bells, but I suppose that just adds drama) -- and I mean a lot of them,
they 
looked like HTML code, but I'm not positive as they were going by too
fast and
I don't seem to have a permanent log of them anywhere.

Since then I've had erratic program crashes with (so far) Netscape and
XEmacs 20,
both programs that use a lot of memory, but I don't know of anything
else they
have in common.  The SCARY thing is, this is also happening to OTHER
USERS (well,
there's only one, but still).

The "crash" behavior varies: sometimes I just lose the window (if run
from an
xterm, I get a "Bus Error" message on the terminal afterwards, at least
with
Netscape).  Other times, the X-server apparently goes down dumping me
out onto a text screen (I can recover by restarting xdm).  Or it may
just throw
me back to the xlogin prompt.

That says to me that whatever happened has damaged something in one of
the system
areas (i.e. not just files owned by me).  I've already deleted
Netscape's HTML
caches, and eliminated cookies from websites I was looking at right
before the
incident.  I've also used ... "find * -mtime ..."  to identify files
altered
after the incident (this found that the files I mentioned were altered,
also 
several /var/log/... files, but nothing that seems interesting).  

Now, if this were a Windows machine, I'd say I'd run out of memory, or
swap space
on the hard drive.  However, none of the filesystems is more than 61%
full -- (100s
of megabytes are still available on /, /var, /tmp, /usr, and /home). 
Now, I
don't actually know how to thoroughly check the swap partition, but
"top" shows it
as essentially unused.  One thing, I've considered is that the swap
mechanism
might be broken in some way.  The problem does tend to occur when system
memory
is mostly used up and no swap space has been.

Examining syslog.conf, I find that messages that match "*.emerg" are
sent to all
users logged in (which appears to be the behavior I experienced), but
they are
not logged to disk, so I don't have a copy of what the errors were.  I
have not
tried recreating the circumstances that (may have) caused the problem (I
hesitate
to do this, lest it do more damage).

Its also possible that this was just a coincidence, and that I just
happened to
have some system resource problem at that time.  Maybe I'm missing some
basic
system administration procedure.

Oh yeah, I did of course reboot a couple of times (this apparently had
no effect).
I forced fsck on one of them, apparently it didn't find anything,
though.

Any ideas, references to documentation I ought to read, things I ought
try, etc
would be really handy.  I would prefer to avoid the Windows solution
(i.e. 
re-install EVERYTHING :) ).

Sorry about the length, but I wasn't sure what details would be useful.

Thanks a lot!
-- 
--------------------------------------------
Terry Hancock   [EMAIL PROTECTED]

Reply via email to