cloudinit.d is a python setup package registered with PyPI which means you can easy_install it. From a python virtualenv or an account with access to the system libraries and python in its path run:

$ easy_install cloudinitd

This will download cloudinit.d and all of its dependencies and install them.

If you prefer to fetch the tarball and have a more manual install you can get the latest release here. To install it:

$ tar -zxf cloudinitd-latest.tar.gz
$ cd cloudinitd-<version>
$ python setup.py install

You can now test it out!

$ cloudinitd --help
Usage: [options] <command> [<top level launch plan> | <run name>]
Boot and manage a launch plan
Run with the command 'commands' to see a list of all possible commands

  --version             show program's version number and exit
  -h, --help            show this help message and exit
  -v, --verbose         Print more output
  -x, --validate        Check that boot plan is valid before launching it.
  -y, --dryrun          Check that boot plan is valid before launching it.
  -q, --quiet           Print no output
  -n NAME, --name=NAME  Set the run name, only relevant for boot (by default
                        the system picks)
  -d DATABASE, --database=DATABASE
                        Path to the db directory
  -f LOGDIR, --logdir=LOGDIR
                        Path to the base log directory.
  -l LOGLEVEL, --loglevel=LOGLEVEL
                        Controls the level of detail in the log file : {debug
                        | info | warn | error}
  -c, --noclean         Do not delete the database, only relevant for the
                        terminate command

Create a launch plan

A launch plan is a set of configuration files that tell cloudinit.d what VM images to boot and in what order. It is where the specific cloud to use is set along with the credentials needed to access that cloud. All aspects of contextualization and testing are also set in the launch plan.

Here we will walk through a very basic launch plan. To minimize potential confusion we will only show a bare minimum set of cloudinit.d features.

Simple EC2 example

In this example we show a minimal launch plan needed to start a VM on Amazon’s EC2. In order to use it you will need an account with AWS.


For this example to work you need your default security group to have port 22 and 80.

The launch plan files used in this example can be downloaded:

Top level launch plan

In this very simple example the top level configuration file simply enumerates the run levels. Because we are only launching a single virtual machine there is only one run level. The entire configuration file follows:

level1: helloec2_level1.conf

Run level configuration

In this example all of the interesting information will be held in the run level configuration file. We will need to have the following information in order to run this example.

  • EC2 access key
  • EC2 secret key
  • EC2 ssh key name
  • path to the matching ssh key

These are all very standard things that EC2 users have. If you are not familiar with these items please check out the EC2 tutorial.

Once you have obtained the above information please set the following environment variables in the following way:

$ export CLOUDINITD_IAAS_ACCESS_KEY=<EC2 access key>
$ export CLOUDINITD_IAAS_SECRET_KEY=<EC2 secret key>
$ export CLOUDINITD_IAAS_SSHKEY=<EC2 ssh key name>
$ export CLOUDINITD_IAAS_SSHKEYNAME=<path to the matching ssh key>

Now that we have all of the security information in place we will look at the run level configuration file:


localsshkeypath: env.CLOUDINITD_IAAS_SSHKEY

iaas: us-east-1
image: ami-6fa27506
ssh_username: ubuntu
allocation: t1.micro

bootpgm: ec2bootpgm.py

The name of this configuration section is svc-sampleservice. What this tells cloudinit.d is that it is a service (svc-) named sampleservice. The name must start with svc-.

The first thing to note from this file are the values corresponding to the environment variables we just set. The directive env. tells cloudinit.d to get the value for this field from the environment variable of the following name. The values for these 4 entries are described above.

The remaining entries have the following meanings:

  • iaas This is a string describing the cloud with which we will be interacting. In this case it is us-east-1 on EC2.
  • image This is a handle to an image. The format of this handle will vary with different clouds. In our case we are using EC2 so the format is ami-XXXXXXX. The specific AMI for use with this example is ami-6fa27506. It is a publicly available Ubuntu image. Note that this AMI is managed by the Ubuntu project and could become unavailable in the future. In such case, you can simply pick a new AMI from the AMI locator page.
  • ssh_username This is the username that the VM image has associated with the localsshkeypath defined above.
  • allocation This string tells the cloud details of the resources we want allocated with the image. In our case we pick t1.micro because it is the cheapest.
  • bootpgm This value is described in the next section.

Boot program

This value points to a program that will be run on your virtual machine once to contextualize it. Once you boot your launch program cloudinit.d will wait until its sshd service is running. Once it is this program is uploaded and run as the user defined by ssh_username, in our case ubuntu. Below is the example bootpgm:

#!/usr/bin/env python

import sys
import os

cmd = "sudo sh -c 'env DEBIAN_FRONTEND=noninteractive apt-get -q -y install apache2'"

f = open("hello.html", "w")
f.write("<html><body>Hello cloudinit.d!</body></html>")

cmd = "sudo cp hello.html /var/www/"


This is a very simple program which simply installs the Apache web server and creates a web page. You can also download it here: ec2bootpgm.py.


Now that we have and understand a launch plan, let’s put it into action with cloudinit.d.


You will be charged by AWS for running this command. At the time of this writing the charges are less than 3 cents per hour.

$ cloudinitd -v -v -v boot helloec2.conf
Starting up run 54c03415
    Started IaaS work for sampleservice
Starting the launch plan.
Begin boot level 1...
    Started sampleservice
    SUCCESS service sampleservice boot
        hostname: ec2-50-17-4-154.compute-1.amazonaws.com
        instance: i-2ae3a245
SUCCESS level 1

The above session shows a successful boot of a launch plan. Notice the use of the -v option. This simply increases the amount of output that we will see.

Your VM was successfully started! The output shows the hostname assigned to the boot as ec2-50-17-4-154.compute-1.amazonaws.com. If we point a web browser at http://ec2-50-17-4-154.compute-1.amazonaws.com/hello.html we should see the message we configured with our ec2bootpgm.py program. Note that your URL will be different.


In order to avoid further service charges from AWS we should clean up this launch. This is done by taking the run name from the output of the boot, in our case 54c03415 and giving it as an option to the terminate directive:

$ cloudinitd -v -v -v terminate 54c03415
Terminating 54c03415
Begin terminate level 1...
    Started sampleservice
    SUCCESS service sampleservice terminate
        hostname: None
        instance: None
SUCCESS level 1
deleting the db file /home/bresnaha/.cloudinitd/cloudinitd-54c03415.db