This is a text-only version of the following page on https://raymii.org:

This is a text-only version of the following page on https://raymii.org:
---
Title : Automating Openstack with cloud init run a script on VM's first boot
Author : Remy van Elst
Date : 11-03-2015
URL : https://raymii.org/s/tutorials/Automating_Openstack_with_Cloud_init_run_a_script_on_VMs_first_boot.html
Format : Markdown/HTML
---

![openstack][1]

This tutorial will show you how to create a VM in Openstack and execute a script
at the first boot using `cloud-init`'s `user-data` feature. This way you can
eliminate some more manual labor and keep a small base image, instead of
requiring all kinds of specific images for specific tasks.

This tutorial will also give you a few example scripts to use with `cloud-init`
and to create Openstack virtual machines from the command line.

You can see all my [Openstack related articles here][2]. For example, how to
build [an automated High Available website cluster with Ansible and
Openstack][3].

 Recently I removed all Google Ads from this site due to their invasive tracking, as well as Google Analytics. Please, if you found this content useful, consider a small donation using any of the options below: <a href="https://leafnode.nl">I'm developing an open source monitoring app called Leaf Node Monitoring, for windows, linux & android. Go check it out!</a> <a href="https://github.com/sponsors/RaymiiOrg/">Consider sponsoring me on Github. It means the world to me if you show your appreciation and you'll help pay the server costs.</a> <a href="https://www.digitalocean.com/?refcode=7435ae6b8212">You can also sponsor me by getting a Digital Ocean VPS. With this referral link you'll get $100 credit for 60 days. </a> 

For this tutorial I've used the [CloudVPS Openstack platform][5]. It will also
work with any other openstack or cloud provider/software that supports `cloud-
init` and the specific `user_data` feature. If you build your own images then
make sure they have `cloud-init` enabled to support this.

Note that this article is not sponsored nor endorsed by CloudVPS, nor am I
speaking for or as CloudVPS.

### cloud-init

Instead of manually creating a VM, logging in to it and executing a few commands
to set it up, you can automate all these steps. The creation of the VM and the
stuff you do when it is up can all be scripted. You can use a programming
language like Python to do this, but simple Bash scripts also work just as well.

cloud-init is a piece of software created to help with initializing virtual
machines on multiple different cloud software platforms. It is a collection of
Python scripts that run on a VM's first boot.

It understands and talks to different data providers like Amazons or the
Openstack metadata service.

It uses that information to, for example, set a root password, grow the root
filesystem, setup an SSH key, do a callback to an URL when a VM is finished
booting or execute commands at boot. All those things and many more are provided
by so called cloud-init modules. Therefor it can be extended easily.

The metadata provided by the cloud provided can contain things like the VM's
name, its IP addres(es), a root password or an SSH key. You can also provide
your own metadata using the so called `user_data`.

`cloud-init` has a nifty feature that allows us to place a script in the
`user_data` which it will execute at the end of the first boot of the machine.
It can be a bash script, or any other script as long as it starts with `#!`.

This tutorial was tested with `cloud-init` versions 0.7.4 up to 0.7.7. The
`/etc/cloud/cloud.cfg` config file needs the following enabled:

cloud_final_modules:
- scripts-user

### Providing user_data to a new VM in Openstack

To provide the `user_data` script to a new VM you need to place your `user_data`
script in a file, in this example `user_data.file`. See below for an example
script

Make sure you have the [Openstack Command Line Tools][6] installed. For
convinience, also create a `computerc` file which holds your credentials and
`source` it in your shell.

The parameter to supply the user data is `--user-data $filename`. To boot up a
small Ubuntu machine at CloudVPS with our `user_data` file we can use this
command:

nova boot --image "CloudVPS Ubuntu 14.04" --key-name $ssh_key --flavor "Standard 1" --availability-zone NL1 --user-data user_data.file "Example VPS 1"

If you have the console of the machine open (`nova get-vnc-console $UUID novnc`)
then you should see your script executed at the end of the cloud-init run at
boot.

### Example user_data cloud-init script

This is an example bash script you can push via the `user_data`. It gives you a
generic idea of what can be done. You could install and setup your configuration
management framework like Puppet or Chef, or just use plain commands. This
example uses Ansible to deploy the imaginary Example App for your company at
first boot:

#!/bin/bash
# Example script to run at first boot via Openstack
# using the user_data and cloud-init.
# This example installs Ansible and deploys your
# org's example App.

echo "userdata running on hostname: $(uname -n)"
echo "Using pip to install Ansible"
pip2 install --upgrade ansible 2>&1

echo "Cloning repo with example code"
git clone https://gitlab.mycompany.org/ansible/example-app.git /tmp/app

pushd /tmp/app
ansible-playbook ./our-app.yml
popd
exit 0

You can also use Python, Ruby or any of your favorite language. As long as the
`user_data` starts with `#!` cloud-init will see it as a script and not as
specific cloud-init modules. You do need to make sure that your base image has
the interpreter installed (Python, Ruby etc.) or bootstrap that via the script.

Here is another script that installs Wordpress on CentOS, including nginx, php-
fpm and mysql:

#!/bin/bash
# Example script to run at first boot via Openstack using the user_data and cloud-init. This example installs Wordpress, nginx, MySQL and PHP-FPM.
# Author: Remy van Elst, https://raymii.org; License: GNU GPLv3

printf "\033c" #clear screen
VERSION="$(grep -Eo "[0-9]\.[0-9]" /etc/redhat-release | cut -d . -f 1)"

echo "Installing EPEL"
rpm -Uvh http://cdn.duplicity.so/utils/epel-release-${VERSION}.noarch.rpm 2>&1

echo "Installing Ansible and Git"
yum -y install ansible git gmp 2>&1

echo "Cloning repo with Wordpress Playbook"
git clone https://github.com/RaymiiOrg/ansible-examples.git /tmp/app 2>&1

echo "Creating Ansible inventory file"
echo -e "[wordpress-server]\n127.0.0.1" > /tmp/app/wordpress-nginx/inventory

echo "Starting playbook"
cd /tmp/app/wordpress-nginx
ansible-playbook -i inventory ./site.yml 2>&1

exit 0

The repository was forked from Ansible's example repo and changed so that the
`site.yml` playbook includes the `connection: local` line. That way we don't use
SSH to run the playbook. It also randomly generates the database password
instead of using a variable.

### Re-execute or debugging

The script only runs at first boot of the machine via `cloud-init`. If you
execute the `cloud-init` command again it will not execute the script because it
already did it. Testing and debugging the script can be quite intensive if you
need to boot up a machine every time.

We can however fool `cloud-init` by letting it think the machine did a fresh
first boot. We need to remove the following two files:

/var/lib/cloud/instances/$UUID/boot-finished
/var/lib/cloud/instances/$UUID/sem/config_scripts_user

Replace `$UUID` by your instance's UUID.

Execute the following command to run the cloud-init final module again:

cloud-init modules --mode final

The `final` module will execute our `user_data` script again. Before every new
test run you need to remove the two files listed above.

Keep in mind as well that if you for example touch a file and run the script
again, the file will still be there. Changes are persistent, build your code
idempotent so that it handles that.

If you've by accident deleted to much cloud-init data you can re-initialize it
with the following command:

cloud-init init

### Command Line script to create VM's

Here is an example script you can use to create an amount of VM's using the
command line. It will wait until the VM is active before creating the next one,
and it passes through a `user_data` file. You can use this, for example, to
easily start up 20 servers and set them up as Apache webservers to scale up when
your site gets a lot of traffic and needs to scale up.

You do need to place a credentials file named `computerc` in your home folder.

#!/bin/bash
KEY="SSH Key Name"
BOOTIMG="IMAGE UUID"
ZONE="NL1"
FLAVOR="Standard 1"

source ~/computerc

for RUN in {1..20}; do
echo "Creating VM ${RUN}""
VMUUID=$(nova boot \
--image "${BOOTIMG}" \
--flavor "${FLAVOR}" \
--availability-zone "${ZONE}" \
--nic net-id=00000000-0000-0000-0000-000000000000 \
--key-name "${KEY}" \
--user-data user_data.file \
"VPS-${RUN}-${ZONE}" | awk '/id/ {print $4}' | head -n 1);

until [[ "$(nova show ${VMUUID} | awk '/status/ {print $4}')" == "ACTIVE" ]]; do
:
done

echo "VM ${RUN} (${VMUUID}) is active."

done

[1]: https://raymii.org/s/inc/img/openstack-software-diagram.png
[2]: https://raymii.org/s/tags/openstack.html
[3]: https://raymii.org/s/articles/Building_HA_Clusters_With_Ansible_and_Openstack.html
[4]: https://www.digitalocean.com/?refcode=7435ae6b8212
[5]: https://cloudvps.com
[6]: http://docs.openstack.org/user-guide/content/install_clients.html

---

License:
All the text on this website is free as in freedom unless stated otherwise.
This means you can use it in any way you want, you can copy it, change it
the way you like and republish it, as long as you release the (modified)
content under the same license to give others the same freedoms you've got
and place my name and a link to this site with the article as source.

This site uses Google Analytics for statistics and Google Adwords for
advertisements. You are tracked and Google knows everything about you.
Use an adblocker like ublock-origin if you don't want it.

All the code on this website is licensed under the GNU GPL v3 license
unless already licensed under a license which does not allows this form
of licensing or if another license is stated on that page / in that software:

This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.

Just to be clear, the information on this website is for meant for educational
purposes and you use it at your own risk. I do not take responsibility if you
screw something up. Use common sense, do not 'rm -rf /' as root for example.
If you have any questions then do not hesitate to contact me.

See https://raymii.org/s/static/About.html for details.