I've just compiled and installed a 2.4.33.3 kernel; it's installed on hda1, where /dev/ide/hda/model reports hda as a "QUANTUM FIREBALLP KA6.4". I'm using lilo; the applicable lilo.conf entry is:

boot=/dev/hda
map=/boot/map
install=/boot/boot.b
prompt
timeout=50
linear
image=/boot/vmlinuz-2.4.33.3
label=linux-new
read-only
root=/dev/hda1

When I try to boot it, I get a kernel panic and a request for a valid "root=" line. It's clearly finding hda1, because it gives the usual "Uncompressing vmlinuz" line, and the kernel prints the usual startup messages, but when it lists the disks it finds and the partitions it finds on them, I can't see hda there at all, but it goes off the page so quickly that I can't see what the problem is.

Is there a way to slow down the rate of the messages, or perhaps to pass them through a pager? It's not in the boot log - that starts after the disks have been detected.

I'm currently running a 2.4.20 kernel with xfs patches, on a dual-processor 500MHz celeron box. It was originally a redhat 7.2 system, but I've changed almost everything.

Thanks in advance
Wocky

Wocky,

You must either (a) have support for your root filesystem (xfs?) built into your kernel (Y), not built as a module (M), or (b) install and run an initrd image to load that module for the kernel.

This is the most common cause of kernel panic.

Thanks for that; unfortunately it's not the problem (it was the first thing I checked.) /dev/hda1 is an ext2 fs, and support for ext2 is compiled in (Y). As far as I can see, the kernel is not finding the disk at all. In the partition list, hda (which has 4 partitions) doesn't appear at all.

Thanks, though.

Incidentally, it was /proc/ide/hda/model (not /dev/ide...) that reported the disk type (of course!)

Thanks again.

Wocky