FC4 Kickstart & Old Megaraid Driver

[Pete Toscano]

Please, can someone help me with this?  I've been working on it for a
while now and just haven't come up with the right combination of words
to Google.

I have an FC2 kickstart CD that gets the ks.cfg file via NFS over eth0,
then gets the rest of the install via NFS too.  This works fine.

When I did my trial install of FC4 on one of my machines, I found that
the newer, probed megaraid driver (megaraid_mbox) doesn't play well with
the MegaRAID Elite 1600 in these boxes.  The manually loaded megaraid
driver seems to work fine though, so I want to go with this.

There seems to be a few paths I can take.

1.  noprobe with ks.cfg over NFS

The problem with this, though, is that the Ethernet driver also isn't
probed.  Is there a way I can "noprobe" only SCSI devices or maybe
specify the Ethernet driver on the append line in the isolinux.cfg file?
 It looks like I can specify the megaraid driver with the "device"
command in the ks.cfg file, so once I get that loaded, I'll be alright
(I hope), but I need the Ethernet to get that loaded.

2. noprobe with ks.cfg on the CD

I think that as long as I can load the ks.cfg file, I'll be okay since
kickstart seems to allow the loading of modules via the "device"
command.  If instead of loading the ks.cfg file via NFS, I just burn it
onto the CD, I should be okay too, right?  This isn't as favored to me
as option one since I do like to tweak the ks.cfg file over time and
don't want to burn a new CD every time, but if it gets me working, I'll
survive.

3.  twiddle with the initrd.img on the boot CD

>From what I've been reading online, I think I could also get things to
work by manipulating the pcitable file in the initrd.img file on the
boot CD.  I gunzip-ed and un-cpio-ed the initrd.img file, but no matter
what I do (even if what I do is nothing), I cannot seem to regenerate
the initrd.img file into something that works.  The kernel loads and the
initrd.img file is loaded, but I always get an error along the lines of
"unable to mount root fs on unknown block (8,3)" and then a kernel dump.
I had read that this might be due to a udev issue, but cannot figure out
how to fix it.

Can anybody help me here?  Any suggestions as to which path is better /
more feasible?

Thanks,
pete