Hi
My previous SystemImager experience is with a cluster in which
I used autoinstall boot floppies.
Now, because of new equipment that does not have floppy or CD
drives, and because of a desire to avoid trying to squeeze
custom kernels onto autoinstall floppies, I'm trying to use
SystemImager's PXE functionality to autoinstall systems on a
new cluster.
I'm having difficulty with the setup of SystemImager's PXE mode.
Hardware:
---------
Cluster of Shuttle XPC SB61G2 boxes.
One SB61G2 box is the designated PBS frontend, DHCP server,
boot server, and imager server. It has a CD drive
Another SB61G2 is the golden client. It has a CD drive
Other SB61G2s are compute nodes to be autoinstalled with the
golden client's cloned system via SystemImager.
No floppy or CD drives exist on the compute nodes.
Basic system software:
----------------------
Fedora Core 1 linux (very similar to rh 9.0)
All packages were installed from the FC1 3-CD distro onto image server
and golden client, including the standard syslinux-2.06-1 package that
contains pxelinux.0.
Basic FC1 linux system is running fine on imager server and golden
client.
The /tftpboot directory contains the following structure:
linux-install/
pxelinux.0
pxelinux.cfg/
msgs/
boot.msg
expert.msg
general.msg
param.msg
rescue.msg
snake.msg
Network:
--------
All boxes are isolated from the Internet on a private 192.168.5.0/24
private switched 100 Mbit/s LAN.
Using the native RealTek RTL8100B ethernet interface on each SB61G2 box,
which successfully uses the linux 8139too ethernet driver.
SIS components and dependencies installed on image server:
----------------------------------------------------------
perl-XML-Simple-1.08-1.noarch.rpm
perl-AppConfig-1.52-4.noarch.rpm
systemimager-i386boot-standard-3.2.0-4.noarch.rpm
systemimager-common-3.2.0-4.noarch.rpm
systemimager-server-3.2.0-4.noarch.rpm
systemconfigurator- 2.0.9-1.noarch.rpm
These SystemImager-related RPMs installed ok on the image server.
mkdhcpserver set up a preliminary /etc/dhcpd.conf ok.
(Had to edit /etc/dhcpd.conf to remove gateway and DNS options
since this cluster is isolated on a private LAN that will use
flat /etc/hosts file for hostname/IP lookups.)
mkbootserver detected tftp server ok.
Problem:
--------
mkbootserver aborted after it failed to find a pxe daemon.
It's not clear what the difference between the pxe boot
loader /tftpboot/linux-install/pxelinux.0 and the
pxe daemon is. Apparently they may be two different things.
But some documentation implies that the pxe daemon is
not necessary for newer pxe client firmware. SystemImager's
mkbootserver program requires the presence of the pxe
daemon, but this daemon is apparently not a part of the
standard fedora-1 distribution.
pxe-1.4.1.tar.gz was obtained from www.kano.org.uk, compiled,
and installed ok on the image server.
But the pxe-1.4.1.tar.gz distribution contains no formal
/etc/rc.d/init.d/pxe start/top script (as mkbootserver
wants), and no guidance on how to start this daemon.
Manually running this pxe daemon using the command "/usr/sbin/pxe"
results in no messages on stdout, stderr, or /var/log/messages, and
the pxe process quickly exits with status 0.
Advice on how to proceed to allow mkbootserver to set up the
fedora-1 image server for pxe autoinstall would be much appreciated.
Thanks,
Bill McKie
NASA Ames Research Center
[EMAIL PROTECTED]
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
Sisuite-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/sisuite-users
