this is obsolete doc -- see http://doc.nethence.com/ instead

Boot recovery 

Quick fix 

In case the machine doesn't restart anymore, but you want it back to production as fast as possible : boot with some liveCD e.g. Gentoo's, Redhat's to launch the daemons in hurry, 

cd /mnt

fsck /dev/sda2

fsck /dev/sda1

mkdir hostexample

mount /dev/sda2 hostexample

mount -t proc none hostexample/proc

mount -o bind /dev hostexample/dev

chroot hostexample /bin/ksh

mount -a

Note change sdaX to your appropriate partition. Here sda2 is /. 

Note use either /bin/ksh, /bin/sh or /bin/bash as default chroot shell. 

Note the last command should mount /boot and swap. 

Check with, 

mount

swapon -s

Then start the needed daemons, 

[...]

service syslog start

service mysqld start

service httpd start

sh /etc/rc.local

service crond start

Note you won't be able to run specific modules like "bonding" and the iptables modules unless you prepare the same kernel version modules into /lib/modules. 

Real fix 

Check and reinstall GRUB or LILO. 

For GRUB, 

grub-install --no-floppy '(hd0)'

grub --no-floppy

   root (hd0,0)

   setup (hd0) 

quit

vi /boot/grub/menu.lst

Note root seems not to be "/" here, it's rather /boot hence the first partition (hd0,0) 

For LILO, 

vi /etc/lilo.conf

lilo

References 

LILO : http://lilo.go.dyndns.org/