On Wed, 15 Oct 1997, Jacob Suter wrote:

> Well I see mention of a 'miniroot' that is supposed to have edlabel and
> all the other goodies to get a partition going on a vax... Well, I don't
> have edlabel but I got disklabel and its about as user friendly as a
> grizzly bear... 

One that's on a bad acid trip I might add.
 
> I'd love to get these two Conner CP-3100's (they are detected without
> error problem by NetBSD on boot...

It is my understanding that SCSI support on the VS3100's is only
functional on the VS3100/M76.  I haven't had any luck labeling a SCSI disk
on either my M30 or M38.  

> Also I've got a LOT of problems trying to compile ANYTHING...   It wants
> *.h files that are in /usr/include/* but it doesn't seem to detect.  I
> just figure its because I am running 100% under NFS, but...

Perhaps the 1.2G snapshot is a bit flakier than I thought it to be. I
don't have much in the way of problems with compiling, but I also went
through the arduous process of compiling and installing all of 1.2G
myself. Something which proved to be about a 3 week long education and
compilation process.  Much of NetBSD/VAX really didn't behave well for me
until I recompiled it all, but then I think I was using a 1.2G kernel with
a 1.2D snapshot at the time.
 
> I've also got an occasional problem (once while playing
> /usr/games/tetris while the console was running systat) with "le0 device
> timeouts"...  just had it happen twice... is this a known problem with

It's very common.  I'm surprised it's only shown up twice on your system.
:-)  I'd recommend telneting into the system from an xterm to do most
work.  Note, 1.2G uses the X11R6 flavor of xterm terminal description
which has problems when displaying on an X11R5 or lesser xterm session.  I
recommend setting your TERM env variable to xterm-r5 if you're telneting
in from an older xterm program.

> Also, has anyone REALLY stressed the NetBSD/VAX on VS3100's very hard? 
> I've pondered starting a "segment" of an ISP off Intrastar's (I am 1/3
> owner and sysadmin of intrastar) main network and see how well the vaxes
> handle 24x7 moderate usage...

I stress mine pretty hard.  Compiling all the NetBSD sources took about a
solid week of compile time for me.  I do not recommend heavy tasking the
machines when they rely on a Linux 2.0 system for NFS mounted swap.  NFS
is sort of flakey under Linux 2.0.  I guess it'd been known in the Linux
community for a while, but I personally found this out while first trying
to compile large programs on my VAXstation.  Access to the NFS swap file
would tend to time out a bit too long, with fatal results to NetBSD. 
Switching to an OpenBSD server fixed all those problems.  NetBSD,
NeXTSTEP, and IRIX seem to serve equally as well for NFS mounted swap. 
Linux is supposed to have NFS cleaned up by 2.2, and it may already be
fixed in the 2.1 experimental kernels.  I need to check though.

There's still an awful lot of work to be done on the VS3100 port.  It runs
okay, but I wouldn't go and start using it for ISP customer accounts...
Unless A) you enjoy disgruntled users or B) you have users who will be
tolerant of some difficulties but get a kick out of using a VAX.

-brian.
---------------------------------------------------------------------------
Brian D. Chase         Systems Coordinator        brian.chase@carpediem.com
-- Compression, Inc. - 13765 Alton Pkwy, Suite B - Irvine, CA 92618, USA --