I have a question re: Anaconda that I was hoping that someone might be
able to help me with. I apologize if this is the wrong list for this
discussion.
I'm trying to kickstart a new RedHat Enterprise 4 system that has 2 x
Intel e1000 gigabit on the motherboard (Intel S5000V). I'm having a big
problem that has been causing me much grief all day today.
Unfortunately, the PCI ID for our e1000 (0x1096) isn't built in to the
standard e1000 module that comes with RedHat Enterprise 4, so kickstart
doesn't recognize the network, and prompts me for a driver disk/manual
install. For numerous reasons, I'd rather not prepare a driver disk. I
would prefer that our installs remain hands off. It's much easier for
me to recompile a later version of the Linux kernel which includes all
the drivers built-in to the kernel (instead of as modules), and then
rebuild the install disk using that kernel so that I have a totally
hands off install. Anyway, after compiling a later Linux kernel
(2.6.17.7) and rebuilding the custom install CD, now when I boot with
"ks", I can see that my e1000 is now recognized fine by the kernel.
However, my kickstart doesn't start, and instead I get taken to the
"language selection" screen, after which it gives me the menu that lets
me choose "Install from CD", "Hard Disk", "NFS", etc. If I choose any
of the network operations like "NFS"/"FTP", it will say that I don't
have the required network device. On the other hand, if I say to
install from "Local CD", the installer installs just fine and recognizes
that both eth0 and eth1 are there. As far as I know, that part is
running Anaconda as well .. so my question is this -- how in the world
can Anaconda recognize the network for a manual install, but not
recognize it for kickstart ???? (I did try the manual install method,
and went to the shell prompt during install where I was able to ifconfig
the interface, and use it, so I know the driver is working...)
I would be very greatful for any help that anyone can provide...
Jason.
