100% network (remote) install -- No direct machine access?

[S. W. Davison]

Many thanks to all who responded to this request.  All the
comments were helpful.  The hands-off install works great.

I started from Robert Citek's notes
(http://www.cwelug.org/cgi-bin/wiki.cgi?FedoraCore3#bootstrap_install)
and adapted them for our situation.  In case anyone is
interested, I have appended a description of the procedure
and the lilo.conf and ks.cfg files.

Tom Diehl's suggestion about using VNC to monitor the
install and using yumgroups.xml to handle platform-to-
platform variance looks interesting.  I don't think we
need to do that -- not many different platform templates --
but it's something to keep in reserve.

There's a lot more to be done before we deploy.  As one
of the responders noted, we have to do a lot of testing.
And there will be all the picky, finicky business of
getting the various servers running properly and the
application software installed.  But all that is
straightfoward, although time-consuming.

Thanks again for your help!

Stowe Davison





<procedure>
1.  Copy the contents of the RedHat directories from the
     four RHEL 3 installation CDs on to an NFS server at the
     site.  That is the "installation tree."

2.  Copy RHEL 3 isolinux versions of vmlinuz and initrd.img
     over to the boot partition of the target machine.  Those
     two files are in the isolinux directory of CD #1 of the
     RHEL 3 installation set.  Same two files for all
     platforms.

3.  Copy a kickstart configuration file, ks.cfg, over to the
     target system.  That file contains the IP addresses of
     the target system and the NFS server from which the RPMs
     will be copied, so it must be customized for each target
     machine.  It also contains the partitioning and
     formatting information and the mount list.  So we would
     probably have a template for each platform type
     and customize them with correct IP addresses for
     each target machine.

4.  Copy a new version of lilo.conf over to the target
     machine, then run lilo to install it.  The new lilo
     configuration boots the isolinux kernel that you copied
     over in step 1, and passes the ks option to the boot
     command line.  Same file for all platforms.

5.  Run "shutdown -r now" on the target system.  When the
     target system finishes rebooting it will be running RHEL
     3 and it will be accessible on the network.

6.  Install application software and configure for operations.
</procedure>






<lilo.conf>
prompt
timeout=50
default=kick_net
boot=/dev/hda
map=/boot/map
install=/boot/boot.b
message=/boot/message
linear

image=/boot/net_ks/vmlinuz
	label=kick_net
	initrd=/boot/net_ks/initrd.img
	read-only
	root=/dev/hda1
	append="ramdisk_size=65536 nofb text ks=hd:hda1/net_ks/ks.cfg"
</lilo.conf>








<ks.cfg>
# Kickstart file automatically generated by anaconda and
# edited for reinstall.

install

# I added the nfs... line.
nfs --server=192.168.0.29 --dir=/home3/rhel3ws_install_tree

lang en_US.UTF-8
langsupport --default en_US.UTF-8 en_US.UTF-8
keyboard us
mouse generic3ps/2 --device psaux
xconfig --card "ATI Mach64" --videoram 8192 --hsync 30-69 --vsync 50-120 
--resolution 1024x768 --depth 24 --startxonboot  --defaultdesktop gnome
network --device eth0 --bootproto static --ip 192.168.0.30 --netmask 
255.255.255.0 --gateway 192.168.0.1 --nameserver 
209.183.205.35,209.183.192.65 --hostname swddeb
rootpw --iscrypted $1$4Ic/R8dr$Q9v5e7qemk1KYsSLYV4Xb0
firewall --disabled
authconfig --enableshadow --enablemd5
timezone America/New_York
bootloader --location=mbr --append hdc=ide-scsi --useLilo
# The following is the partition information you requested
# Note that any partitions you deleted are not expressed
# here so unless you clear all partitions first, this is
# not guaranteed to work
#
# All the partition stuff was commented out in the anaconda-
# generated version.  I uncommented it.
clearpart --linux
part /boot --fstype ext3 --size=1000 --asprimary
part raid.13 --size=5000
part raid.11 --size=5000
part raid.9 --size=5000
part raid.7 --size=5000
part raid.6 --size=5000
part swap --size=1000 --asprimary
# I added quotes around <physical volume (LVM)>
raid pv.22 --fstype "physical volume (LVM)" --level=RAID0 raid.13 
raid.11 raid.9 raid.7 raid.6
volgroup Volume00 pv.22
logvol / --fstype ext3 --name=LogVol00 --vgname=Volume00 --size=3000
logvol /home --fstype ext3 --name=LogVol01 --vgname=Volume00 --size=10000
logvol /tmp --fstype ext3 --name=LogVol02 --vgname=Volume00 --size=3000
logvol /var --fstype ext3 --name=LogVol04 --vgname=Volume00 --size=2000
logvol /usr --fstype ext3 --name=LogVol03 --vgname=Volume00 --size=6000

# That's the end of the section that was commented by anaconda.

# I added reboot
reboot

%packages
@ everything
kernel
lilo

%post

</ks.cfg>