Skip site navigation (1)Skip section navigation (2)
Date:      Thu,  4 Nov 1999 09:12:46 -0500 (EST)
From:      Andrew Gallatin <gallatin@cs.duke.edu>
To:        Mike Smith <mike@smith.net.au>
Cc:        Matthew Dillon <dillon@apollo.backplane.com>, current@FreeBSD.ORG
Subject:   Re: Network booting, I'm off to work (was  Re: GENERIC build broken)
Message-ID:  <14369.35974.496839.872957@grasshopper.cs.duke.edu>
In-Reply-To: <199911031909.LAA06730@dingo.cdrom.com>
References:  <199911031843.KAA60534@apollo.backplane.com> <199911031909.LAA06730@dingo.cdrom.com>

next in thread | previous in thread | raw e-mail | index | archive | help


Mike,

I'm not trying to contribute to the FUD, rather I'm just trying to
understand what you're proposing and what effect it will have on my
current environment.

Right now we have an OS research cluster of 30+ Pentium II machines.
These machines are used for both real research & for class projects by
students who have root access to them & who can install their own
hacked kernels.  These machines are essentially crashboxes.  These
machines boot off of local IDE & SCSI disks. They are all rackmounted,
so they run headless & have their console via a serial port.  Most do
not even have graphics cards installed.  Interacting with the BIOS is
a painful process involving walking over to the lab, pulling a box
off the rack, moving it to the workbench, inserting a video card &
attaching a monitor/kbd.

My install process is to boot the machines diskless via a BOOTP kernel
stashed away on the root filesystem of the machine (or on a floppy if
it is a first time install, or if things get really out of hand).
When a machine boots diskless, it mounts a shared volume on the
installation server.  A script is run (I call it mirror) which
disklabels the disk & restores a dump(8)ed backup of a good
'prototype' machine.  The script continues on to massage a few
configuration files (eg, setting the hostname & ipaddr) and builds an
fstab based on what the root disks name is.  All of this takes about
10 minutes max.  Maybe 15 if I have to walk over to the lab & plunk a
floppy in the box.

Because these machines get trashed fairly frequently (students with
root access who are hacking device drivers can trash all sorts of
things in many fun & exciting ways!  You should try it if you're in
need of laughs sometime!) the ability to restore a box to sanity
quickly is very important to me.

My question boils down to:  Will I be able to re-install a machine
using your new i386 netboot just as easily as I can now?  Or will I
have to be physically present at each machine & diddle with the bios
to toggle between disk & netboots?   And what if the NIC doesn't
support PXE?  Am I just SOL?

Again, I'm not trying to contribute to the FUD.  I'm just not sure I
understand how what you are proposing will affect me.

Thanks,

Drew
------------------------------------------------------------------------------
Andrew Gallatin, Sr Systems Programmer	http://www.cs.duke.edu/~gallatin
Duke University				Email: gallatin@cs.duke.edu
Department of Computer Science		Phone: (919) 660-6590





To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?14369.35974.496839.872957>