kernel panic from various initrd/vmlinuz combinations

[Norman Coder]

I recently built a pxeboot/dhcpd server so I can install various Red Hat 
software (RH3, rhel-4, rhel-5, FC-4, FC-6, etc) using PXE boot through 
the network.
It all works fine except for:
when I PXE-boot using (any RH4 or RH5) distribution fails with the same 
error, listed below, and I am using the initrd/vmlinuz files from 
/images/pxeboot/. Fedora Core and RH3 seems to work fine.

EXT2-fs: unable to read superblock
isofs_fill_super: bread failed. dev=md1.
iso_blknum=16. block=32
Kernel panic - not syncing: VFS: Unable to mount root
fs on unknown block(9,1)

If I boot from the first CD the process works fine, but this defeats the 
purpose of PXEbooting.

It appears that I do not have the right combination of initrd/vmlinuz or 
that the options I used on the append line in my 
/tftpboot/pxelinux.cfg/default are incorrect. I have tried several 
combinations of options with no affect other than to display a "kernel 
alive" message and the system hangs when I add "ide=nodma".

kernel is alive
kernel direct mapping tables upt to 1300000....

I can repeat this problem any time I use any RH4 or RH5 distribution.

My basic distribution copy from a set is ISOs is:
for i in 1 2 3 4 5;
do mount -o loop -t iso9660 /iso/redhat/rhel-5-client-x86_64/rhel-5-client-x86_64-disc$i.iso /mnt/dcrom
ls /mnt/dcrom/
if [ $i -eq 1 ]
then
  cp -f /mnt/dcrom/images/pxeboot/initrd.img /tftpboot/initrd-rhel-5-client-x86_64.img
  cp -f /mnt/dcrom/images/pxeboot/vmlinuz /tftpboot/vmlinuz-rhel-5-client-x86_64
  cp -f /mnt/dcrom/RELEASE-NOTES-en* .
fi
echo " "
cp -af /mnt/dcrom/Client .
cp -af /mnt/dcrom/Workstation .
cp -af /mnt/dcrom/images .
umount /mnt/dcrom/
done

My standard append line in /tftpboot/pxelinux.cfg/default is:
# RedHat RHEL 5 Client x86 Interactive (64-bit)
label rh5-client-64
kernel vmlinuz-rhel-5-client-x86_64
append initrd-rhel-5-client-x86_64.img ramdisk_size=9216 load_ramdisk=1 network expert apm=off acpi=off skipddc nofb


(I also included several of the combinations I have tried)
#  append initrd-rhel-5-client-x86_64.img ramdisk_size=92160 load_ramdisk=1 network expert apm=off acpi=off skipddc nofb
#  append initrd-rhel-5-client-x86_64.img ide=nodma ramdisk_size=92160 load_ramdisk=1 network expert apm=off acpi=off skipddcnofb
#  append initrd-rhel-5-client-x86_64.img ide=nodma noapic edd=off apm=off acpi=off skipddc nofbload_ramdisk=1 ramdisk_size=92160
#  append initrd-rhel-5-client-x86_64.img ide=nodma  # it looks like this is the option to avoid kernel panic
#  append initrd-rhel-5-client-x86_64.img ide=nodma noapic edd=off apm=off acpi=off skipddc nofbload_ramdisk=1 ramdisk_size=92160 ks=nfs:192.168.60.15:/pxe_boot/redhat/configs/trdlnxXX_ks.cfg