On Tue, 2 Mar 2010 at 10:10am, Tony Mountifield wrote
> I have a remote CentOS 4 machine on a network where I can't put a DHCP
> or PXE server, and I want to do a complete reinstall. So what I want to
> do is, from the currently-running system, to invoke an installation
> kernel and initrd in just the same way that GRUB would, giving it a boot
> command line that specifies a remote kickstart file, installation tree,
> and other required info.
This is simple. Grab the vmlinuz and initrd.img files from the pxeboot
directory of the repo you want to install from. Put those in /boot on the
server in question. From there, there are a couple of ways you can go.
The easiest is to actually put the ks.cfg on the server itself. Then you
can add a stanza like the following (you'll need to tailor all the hard
drive references to your own setup, of course) to your grub.conf:
title reinstall
root (hd0,0)
kernel /boot/vmlinuz ks=hd:sda1:/ks.cfg ksdevice=eth0
initrd /boot/initrd.img
Make that entry the default, reboot, and your kickstart will start.
Obviously all of your network info needs to be specified in the ks.cfg
file.
If you want to grab the ks.cfg from a remote server, that can be done too,
but you'll need to specify the network config options on the "kernel" line
above. I don't have the exact syntax handy, but it's all documented.
Install the "anaconda" package and look in
/usr/share/doc/anaconda-$VERSION/command-line.txt and you can see all the
options you can pass to the install kernel. On CentOS-5 installs I
always use "noipv6", since it seems to make things go much faster.
For a one-off like this, installing cobbler is a bit (read: a lot) of
overkill.
--
Joshua Baker-LePain
QB3 Shared Cluster Sysadmin
UCSF