Hey, everyone,
I'm trying to PXE boot a Red Hat 3 cluster node using a dhcp and tftp server. The dhcp server assigns the node an IP (confirmed with tcpdump on the server) and pxelinux.0 is obtained, as well. The tftp server hosts files at /tftpdboot:
Code:
pxelinux.0
vmlinuz
initrd.img
kickstart/node01-ks.cfg
pxelinux.cfg/default
The default file is obtained successfully by the node, but the Red Hat installation starts instead of the kickstart file.
initrd.img and vmlinuz were obtained from the Red Hat 3 CD under /images/pxeboot. Is this correct? Is this image configured to automatically start the installation instead of using the kickstart file I specify? Should I make my own initrd.img?
Contents of default:
Code:
label linux
kernel vmlinuz
append initrd=initrd.img network ksdevice=eth0 ks=nfs://10.0.0.5/tftpdboot/kickstart/node01-ks.cfg load_ramdisk=1
IPAPPEND 1
10.0.0.5 is my DHCP/TFTP server. 10.0.0.2 is the IP assigned to the node by my DHCP server. If I change the
ks= line to something bogus the node complains about the kickstart file being null instead of starting the Red Hat installation.
Contents of node01-ks.cfg:
Code:
# mount disk from the NFS server which contains support files
mkdir -p /
mount 10.0.0.5:/var/10.0.0.2 -o ro /
Here is my exports file, as well:
Code:
/tftpdboot 10.0.0.2(ro,no_root_squash,anonuid=0)
/var/10.0.0.2 10.0.0.2(ro,no_root_squash,anonuid=0)
How can I make the node use the kickstart file, as opposed to automatically starting the installation process?