Course notes and commands from TutsPlus: Virtual Machines with Vagrant and Puppet

vagrant.txt

# https://tutsplus.com/course/virtual-machines-with-vagrant-and-puppet/

# Download and install VirtualBox https://www.virtualbox.org/wiki/Downloads
# Download and install Vagrant http://www.vagrantup.com/downloads.html

# For this lesson, we'll be using VirtualBox as the provider

# add a new box to collection of vagrant boxes
# vagrant box add <name> <location>
vagrant box add precise32 http://files.vagrantup.com/precise32.box

# More boxes available at www.vagrantboxes.es

# list available boxes on local system
vagrant box list

# create folder for virtual machine
mkdir firstVM
cd firstVM

# Initialize a new vagrant virtual machine
# Creates a Vagrantfile in current directory
# This file describes configuration for virtual machine
# vagrant init <box name>
vagrant init precise32

# Remove an existing box from the system
# vagrant box remove <box name> <provider>
vagrant box remove lucid32 virtualbox

# Boot up the virtual machine
# Imports base box and does setup for virtual machine
# After it starts, check VirtualBox Manager - can see machine listed and running
# By default, vagrant vm's are run in headless mode,
# which means there is no window to interact with it directly
# We'll be using command line to work with the VM
vagrant up

# Shut down Virtual Machine - Option 1: Suspend
# Equivalent of putting computer to sleep,
# It's not running but everything on VM's hard drive and RAM
# have been saved to to disk on host machine (your computer)
# Depending on how big the vm is, this can be very resource intensive
vagrant suspend

# Wake up a suspended vm
vagrant resume

# Shut down Virtual Machine - Option 2: Graceful shutdown
# Anything on vm's RAM will not be saved to host computer,
# but vm's hard-drive will be saved to host disk
# This means vm is still using up host hard-drive space, even if vm is not running
vagrant halt

# Bring vm that was halted back up
vagrant up

# Variation: Halt and Restart
# this is required if changes were made to Vagrantfile
vagrant reload

# Shut down Virtual Machine - Option 3: Destroy
# Note still have config saved in Vagrantfile, that is not "destroyed",
# therefore, shouldn't lose any config or provisioning because that's stored separately
vagrant destroy

# Login to vm
# NOTE: Windows 7 must have ssh client on PATH, such as C:\git\Git\bin
vagrant ssh

# We're logged in as 'vagrant' user
whoami

# Check what OS version
cat /proc/version

# 'vagrant' is shared folder linked to host computer dir
# from which we created vm in, vagrant sets up this share by default
cd /vagrant

# create and edit a file in this folder
touch index.html
vi index.html
# edit and add some text, example: <h1> Hi! </h1>

# Boot up a server with python SimpleHTTPServer
# NOTE: sudo is required to choose port 80
sudo python -m SimpleHTTPServer 80

# via config, port 80 is being forwarded to port 8080 on host
config.vm.network "forwarded_port", guest: 80, host: 8080

# Now from host machine, open browser at
http://localhost:8080/
Should see Hi printed on page

# To kill the server
Ctrl+C

# To leave vm
exit

# Vagrantfile
- config for vm, describes setup, what resources should be installed on vm, etc.
- stored in root folder for vm
- should be committed in source control
- is written in ruby
- config object is passed into configure block

Port Forwarding - example
port 8080 on host machine is used to access port 80 on guest machine
config.vm.network "forwarded_port", guest: 80, host: 8080

Folder syncing
Default is not specified in vagrant file, but can also specify additional, example:
config.vm.synced_folder "../data", "/vagrant_data"
sibling data folder on host machine will be synced to /vagrant_data on guest

Provisioning with Shell Scripts
Preparing vm with software and configuration that it needs before it can be used.
For simple requirements only, can provision with shell scripts.

Provisioning with Shell Scripts Method 1 - Inline
config.vm.provision :shell, inline: "echo Hello World!"
vagrant up
If get message about machine already provisioned, run: vagrant provision
don't worry about message: stdin: is not a tty

If make changes to provisioning, don't need to restart vm, just run:
vagrant provision

Provisioning with Shell Scripts Method 2 - Path
config.vm.provision :shell, path: './provision.sh'

Create a file provision.sh in same dir as Vagrant file and edit:
echo "This is our provisioning script"
apt-get clean
apt-get update
apt-get install vim -y

To verify vim was installed
vagrant ssh
which vim

Provisioning with Puppet
Puppet used for actual servers as well as configuring vm's
Don't need to install Enterprise version, use OSS
Note that vagrant already has a version of puppet installed:
vagrant ssh
which puppet
	/opt/vagrant_ruby/bin/puppet
puppet --version
	2.7.19
If you're on a vm that doesn't come with puppet, run:
apt-get install puppet	

Puppet Resources
- describes a single unit of configuration
- eg: file, user, service, etc.
- Puppet has some built in resources, and you can build your own
- resource is NOT a set of instructions, description of state of part of the system
- Puppet scripts use resources to check if system is already in that state, 
if yes, will not re-install packages or have other side effects.

To check state of a resource
puppet resource user
- lists all users in system and prints out their current configuration

Check state of a particular user resource
puppet resource user vagrant

Check state of a package resource
puppet resource package vim

Create a puppet script in same folder as Vagrantfile: vim.pp
exec { "apt-update":
  command => "/usr/bin/apt-get update"
}
Exec["apt-update"] -> Package <| |>
package { 'vim':
  name   => 'vim',
  ensure => 'installed',
}

- puppet script is called: puppet manifest
- if don't explicitly declare name property, first token after package becomes the name
- by convention, line up the arrows, whitespace doesn't matter
- by convention, leave trailing comma after last option, so it's easy to add more stuff

After saving file: vagrant ssh
cd /vagrant

Run the puppet manifest
puppet apply vim.pp
(can also leave off 'apply' option because that's the default action)
- will get err about not being root, fix it with
sudo puppet apply vim.pp

Check status of vim package resource:
puppet resource package vim
package { 'vim':
  ensure => '2:7.3.429-2ubuntu2.1',
}

Other Puppet Resources
- file: manages file
- package: manages packages
- service: manages services running on the node
- notify: logs messages to the termial
- user: manages users
- group: manages user groups
- group: manages user sgroups
- exec: running arbitary command
- cron: managing cron jobs

Puppet Core Types Cheat Sheet
http://docs.puppetlabs.com/puppet_core_types_cheatsheet.pdf

Puppet Docs Resource Types
http://docs.puppetlabs.com/references/latest/type.html

Putting Resources in Order
- puppet is not procedural
- resources may not be installed in the order in which they are declared

create files.pp
file { 'one':
		path 	=> '/vagrant/one',
		content => 'one',
	}
	
file { 'two':
		path 	=> '/vagrant/two',
		content => 'two',
	}
	
In vagrant vm, run
puppet apply files.pp	
- file two may be created before one
- how to ensure one gets created first?

Meta-parameter
- add to one of file resources, example:
file { 'one':
		path 	=> '/vagrant/one',
		content => 'one',
		before	=> File['two'],
	}