On Tue, 25 Feb 2003, John wrote:
> On Mon, 24 Feb 2003, Jim Wildman wrote:
> > 3) Configure PXE boot so if the box is rebooted and a particular link is
> > in place it reinstalls, else boot from the local drive.
>
> This has possibilities, but needs some care and attention to work. AFAIK
> it only works if your NIC actually supports PXE.
naw -- no worries if your host cannot PXE, lacks a floppy, and
lacks a CD drive, so long as you have a hard drive, or the
capability to NFS boot no-root-squash (I cannot visualize
how the last would be able to NFS boot, and yet not PXE --
maybe on an OpenBoot prom with rarpd -- it should be
possible on a Sparc port. I'll ask Tom Callaway ...).
A. If it does not PXE and has a floppy, use etherboot
B. If it does not PXE and you don't want to use the floppy,
use a CD based boot -- see the RELEASE-NOTES for RHL 8.0 to build
the ISO; or the use the prebuilt in Phoebe-X
C. If it does not PXE and you don't want to use the floppy and
you don't want to use the CD, you can:
1/ Copy the ./images/pxeboot/* content into, say,
a newly created /boot/pxeboot/
2/ Update /boot/grun/grub.conf to have a stanza pointing to
the kernel, the initrd, and passing along a proper ks.cfg
source
------------- snip ----------------
# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this file
# NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.
# root (hd0,0)
# kernel /vmlinuz-version ro root=/dev/hda2
# initrd /initrd-version.img
#boot=/dev/hda
default=1
timeout=10
splashimage=(hd0,0)/grub/splash.xpm.gz
title Red Hat Linux (2.4.20-2.48)
root (hd0,0)
kernel /vmlinuz-2.4.20-2.48 ro root=LABEL=/ hdb=ide-scsi
initrd /initrd-2.4.20-2.48.img
#
# RPH edits for John on the West Coast of .au
#
title Red Hat Linux Upgrade
root (hd0,0)
kernel /pxestuff/vmlinuz ro root=LABEL=/ hdb=ide-scsi
initrd /pxestuff/initrd.img
append ksdevice=eth0 load_ramdisk=1 prompt_ramdisk=0 \
ramdisk_size=32768 network \
ks=http://10.16.33.105/pub/kickstart/ks.cfg
#
# note that the foregoing line uses backslash
# continuation only for clarity -- don't use them
# in production -- string the darn thing out ...
#
-------------- end snip -------------
I have not seen method C formally documented -- that is,
simply booting from hard drive into a kickstart based upgrade
-- (an install would work just as well).
Clearly method C is not a Red Hat supported model -- heck,
while PXE file support is present, it is for advanced
installers. The issues of having a working PXE clients, DNS,
DHCP, TFTP, FTP and HTTP server make it really tricky to debug
if one is not systematic. Red Hat support could not
economically meet their commitment to purchasers to get
supported hardware installed in finite time, because there is
so much to go wrong. <g>
BTW: This is not theory. I am away from my office notes, and
did this from memory. I just followed exactly the stated
instructions, and used the install information here on a
machine on the bench across the room.
This is a variation on my notes at:
http://www.owlriver.com/tips/pxe-install/ which I
developed with editorial review and 'bouncing ideas off' Jim
Wildman. Thanks, Jim.
- Russ Herrold
--
end
======================================+
.-- -... ---.. ... -.- -.-- |
Copyright (C) 2003 R P Herrold | Owl River Company
herrold@xxxxxxxxxxxx NIC: RPH5 (US) | "The World is Open to Linux (tm)"
My words are not deathless prose, | Open Source LINUX solutions ...
but they are mine. | info@xxxxxxxxxxxx -- Columbus, OH
gpg --keyserver pgp.mit.edu --recv-key 0x7BFB98B9
gpg --list-keys 2> /dev/null | grep 7BFB98B9
