I dual boot redhat and suse. Redhat LILO is configured to launch SUSE
from the lilo menu.

If I launch SuSe from redhat lilo, I am unable to browse the net. During the
initialization, I see failure messag with dhcp. I also see eho0 as not
being available.
however, if I boot using SuSE boot diskette (with lilo), I am able to
browse the net. However, I still see some dhcp related failure message
during boot process.

1. I am curious to know why the way suse is launched makes difference?
2. How to fix issues related to dchp, eth0 in SuSE?

Thx..
Connie

How do you have the SuSE part of lilo setup? The way that I do it is to copy SuSE's vmlinuz into Redhat's /boot directory and name it something like SuSE.vmlinuz. Then use this in lilo.conf.

And if having the SuSE stuff over in the Redhat directory is causing problems, copy the RH boot stuff over to the SuSE side and run lilo from there.

FWIW, I have Win4Lin, SuSE, Mandrake, & Windows setup on lilo (with no apparent problems)


It's probably the eth0 (eho0?) failure that's hosing you, but I am network ignorant...


John

What does your Red Hat lilo.conf file look like?

Thx rootbuddy & linuxcool for replies.


Linuxcool: Here is how my main lilo looks like (redhat),

boot="/dev/hda6"
timeout="50"
linear
prompt
default=RedHat
read-only
map=/boot/map
install=/boot/boot.b
message=/boot/message

image="/boot/vmlinuz-2.4.2-2"
label="RedHat"
read-only
root="/dev/hda6"
append="hdc=ide-scsi"

image="/boot/vmlinuz"
label="SuSE"
read-only
root="/dev/hda8"
append="hdc=ide-scsi"

image="/boot/vmlinuz"
label="Caldera"
read-only
root="/dev/hda9"
append="hdc=ide-scsi"

other=/dev/hda1
label="Windoz"


You are probably right, rootboy. I must copy the vmlinuz from suse partition and move it into redhat partition. In the lilo.conf above, i am curious to know what the 'image=/boot/vmlinuz for 'SuSE' refer to? Does it refer to the vmlinuz on suse or on the redhat root partition?

Let me try to copy the vmlinuz from both suse and caldera into redhat partitions and see if it brings up the dhcp.

Thx

Yeah, I got that to work.
As rootboy pointed out, I just copied the kernels (vmlinuz) from both suse and caldera partitions and placed them into the redhat's /boot folder under difference names, updated lilo.conf, installed lilo, and lo! it was all set.

Cool

I just recently added SuSE 7.3 to my mix (now up to 4 OS'es) This time around, I left things where they came from and just added the absolute PATH to the lilo.conf statement.


In other words, on my box I have /dev/hdb6 formatted w/reiser and it contains Mandrake. To get to it from SuSE, I mount it as /Hdb6 (on my box, any external partition that is mounted has its first letter capitalized to clue my in to it being an external partition. And I like to keep things simple by naming it after the partition)

So, to add Mandrake to my lilo I:

a) have to be sure that /Hdb6 is mounted
b) include /Hdb6/boot/vmlinuz-Mandrake (or whatever it's called) into the Mandrake portion of my lilo.conf.

This keeps things separate and seems to work well.


John

cool rootboy, great job.

I am curious to know if partitions from /etc/fstab are mounted prior to linux being loaded (rather, I am not sure how to formulate the question)?

Is lilo.conf read everytime linux starts? or is it important only at the LILO installation time?

What exactly happens when you install LILO on the root partition?

Do you have any ide?

Thanks

Quote:

I am curious to know if partitions from /etc/fstab are mounted prior to linux being loaded (rather, I am not sure how to formulate the question)?

If I understand your question correctly, you're asking if these various /boot directories that would be located on a different partitions (at least some of them) need to be mounted first (during the boot).

No, they only need to be mounted whenever you modify your bootloader (when you run lilo). After lilo successfully runs, you can umount these partition(s) unless you need them for some other reason.

And the obvious question would be why not use the absolute partition device (i.e. /dev/hdb6) rather than a mount point. You can, but you would still need to mount the partition, so it doesn't gain you anything (that I can see).


I don't know, but I strongly doubt it. That's why you run lilo in the first place, to convert your "human-readable" parameters into the actual bootloader code.


Ummm, exactly?

Not exactly, but here's a good starting point:

http://www.lysator.liu.se/~forsberg/...with-lilo.html


John

Here's my lilo.conf, I've cleaned it up a bit and added Mandrake 8.1 (which hangs during boot for some reason...)


boot = /dev/hda
vga = 771
read-only
menu-scheme = Wg:kw:Wg:Wg
lba32
prompt
timeout = 80
message = /boot/message

image = /boot/vmlinuz
label = linux7.2
root = /dev/hda6
initrd = /boot/initrd

image = /boot/vmlinuz.suse
label = suse7.2
root = /dev/hda6
initrd = /boot/initrd.suse
optional

image = /Hdb1/vmlinuz
label = MD8.0
root = /dev/hdb6
initrd = /Hdb1/initrd.img


image = /Hdb7/boot/vmlinuz
label = linux7.3
root = /dev/hdb8
initrd = /Hdb7/boot/initrd

image = /Hdb7/boot/vmlinuz.suse
label = suse7.3
root = /dev/hdb7
initrd = /Hdb7/boot/initrd.suse
optional

image = /Hda9/boot/vmlinuz
label = MD8.1
root = /dev/hda9
initrd = /Hda9/boot/initrd.img

other = /dev/hda1
label = windows

image = /boot/memtest.bin
label = memtest86

# Added by NeTraverse for Win4Lin kernel
image = /boot/win4lin
label=win4lin
initrd = /boot/initrd
root=/dev/hda6


Notice the "Hda9" (etc...), that's my mount point for that particular partition. I had to mount these partitions so that lilo could find the vmlinuz's when it wrote the new bootloader.


John

sounds good john,

Were the mount points /Hdb7 or 9 mounted only when you installed Lilo?

Connie

Yup, no need for them to be mounted except for running lilo, or of course, booting into that particular version or distro.


John