What's the difference between /etc/rc.local and /etc/init.d/rc.local?

/etc/init.d is maintained on ubuntu for backward compatibility with sysvinit stuff. If you actually look at /etc/init.d/rc.local you'll see (also from a 12.04 LTS Server):

#! /bin/sh
### BEGIN INIT INFORMATION
# Provides:          rc.local
# Required-Start:    $remote_fs $syslog $all
# Required-Stop:
# Default-Start:     2 3 4 5
# Default-Stop:
# Short-Description: Run /etc/rc.local if it exist
### END INIT INFO

And "Run /etc/rc.local" is exactly what it does. The entirety of /etc/rc.local is:

#!/bin/sh -e
#
# rc.local
#
# This script is executed at the end of each multiuser runlevel.
# Make sure that the script will "exit 0" on success or any other
# value on error.
#
# In order to enable or disable this script just change the execution
# bits.
#
# By default this script does nothing.

exit 0

I would guess the purpose in doing this is to provide a dead simple place to put shell commands you want run at boot, without having to deal with the stop|start service stuff, which is in /etc/init.d/rc.local.

So it is in fact a service, and can be run as such. I added a echo line to /etc/rc.local and:

»service rc.local start
hello world

However, I do not believe it is referenced by anything in upstart's /etc/init (not init.d!) directory:

»initctl start rc.local
initctl: Unknown job: rc.local

There are a few "rc" services in upstart:

»initctl list | grep rc
rc stop/waiting
rcS stop/waiting
rc-sysinit stop/waiting

But none of those seem to have anything to do with rc.local.

This is more of a distribution specific thing. (like, you will not find different rc.local in CentOS).

Now coming to your actual question, I think adding anything inside /etc/init.d/rc.local makes it to start as a "service" whereas, anything inside /etc/rc.local would simply launch that script at boot time.

I am not really sure on why Ubuntu still maintains both of them? (Perhaps someone else might shed some light over this part !!)