After upgrading a kernel on an older CentOS 3 box for a customer the server would no longer boot up. Instead I would get the following at the boot screen:
/etc/rc.d/rc.sysinit: line ###: /dev/null: Read-only file system
The box would then drop into single user maintenance mode. When I tried to fsck the file system it would complain about about /dev/null being read only again. The fix turned out to be fairly simple.
# mount -rw -o remount /
# rm -f /dev/null
# mknod /dev/null c 1 3
# chmod 666 /dev/null[/code]