Aigh! The 6.1 release (and RPM update) both have this bug. I hate someone.
Please open /etc/rc.d/init.d/halt and look at this logic:

... stuff ...
# Unmount file systems...
... more stuff ...
runcmd "Unmounting proc file system" umount /proc
... eh? ...
# turn off raid
if [ -x /sbin/raidstop -a -f /etc/raidtab ]; then
    # we can not use raidstop -a here because this will only stop
    # devices listed in the default config file which is not always
    # the case. So we look only for the active raid devices
    if [ -f /proc/mdstat ] ; then
        mddevs=$(grep ^md /proc/mdstat | awk '{ print $1 }')
... EH???! ...

This can, of course, be fixed by moving the umount /proc to AFTER examining
and shutting down RAID devices. Works for me.

zbeckman@creativesun.com