Playbooks¶

Hosts and Users¶

For each play in a playbook, you get to choose which machines in your infrastructure to target and what remote user to complete the steps (called tasks) as.

The hosts line is a list of one or more groups or host patterns, separated by colons, as described in the The Inventory File, Patterns, and Groups documentation. The user is just the name of the user account:

---
- hosts: webservers
  user: root

Support for running things from sudo is also available:

---
- hosts: webservers
  user: yourname
  sudo: True

If you need to specify a password to sudo, run ansible-playbook with --ask-sudo-pass (-K). If you run a sudo playbook and the playbook seems to hang, it’s probably stuck at the sudo prompt. Just Control-C to kill it and run it again with -K.

Vars section¶

The vars section contains a list of variables and values that can be used in the plays, like this:

---
- hosts: webservers
  users: root
  vars:
     http_port: 80
     van_halen_port: 5150
     other: 'magic'

These variables can be used later in the playbook, or on the managed system (in templates), just like this:

{{ varname }}

Within playbooks themselves, but not within templates on the remote machines, it’s also legal to use nicer shorthand like this:

$varname

Further, if there are discovered variables about the system (ansible provides some of these, plus we include ones taken from facter or ohai if installed) these variables bubble up back into the playbook, and can be used on each system just like explicitly set variables.

Facter variables are prefixed with facter_ and Ohai variables are prefixed with ohai_. Ansible variables (0.3 and later) are not surprisingly prefixed with ansible_. So for instance, if I wanted to write the hostname into the /etc/motd file, I could say:

- name: write the motd
  action: template src=/srv/templates/motd.j2 dest=/etc/motd

And in /srv/templates/motd.j2:

You are logged into {{ facter_hostname }}

But we’re getting ahead of ourselves. Let’s talk about tasks.

Tasks list¶

Each play contains a list of tasks. Tasks are executed in order, one at a time, against all machines matched by the host pattern, before moving on to the next task.

Hosts with failed tasks are taken out of the rotation for the entire playbook. If things fail, simply correct the playbook file and rerun.

The goal of each task is to execute a module, with very specific arguments. Variables, as mentioned above, can be used in arguments to modules.

Modules other than command and shell are ‘idempotent’, meaning if you run them again, they will make the changes they are told to make to bring the system to the desired state. This makes it very safe to rerun the same playbook multiple times. They won’t change things unless they have to change things.

The command and shell modules will actually rerun the same command again, which is totally ok if the command is something like ‘chmod’ or ‘setsebool’, etc.

Every task must have a name, which is included in the output from running the playbook. This is output for humans, so it is nice to have reasonably good descriptions of each task step.

Here is what a basic task looks like, as with most modules, the service module takes key=value arguments:

tasks:
  - name: make sure apache is running
    action: service name=httpd state=running

The command and shell modules are the one modules that just takes a list of arguments, and don’t use the key=value form. This makes them work just like you would expect. Simple:

tasks:
  - name: disable selinux
    action: command /sbin/setenforce 0

Variables can be used in action lines. Suppose you defined a variable called ‘vhost’ in the ‘vars’ section, you could do this:

tasks:
  - name: create a virtual host file for $vhost
    action: template src=somefile.j2 dest=/etc/httpd/conf.d/$vhost

Those same variables are usable in templates, which we’ll get to later.

Local Playbooks¶

It may be useful to use a playbook locally, rather than by connecting over SSH. This can be useful for assuring the configuration of a system by putting a playbook on a crontab. This may also be used to run a playbook inside a OS installer, such as an Anaconda kickstart.

To run an entire playbook locally, just set the “hosts:” line to “hosts:127.0.0.1” and then run the playbook like so:

ansible-playbook playbook.yml --connection=local

Alternatively, a local connection can be used in a single playbook play, even if other plays in the playbook use the default remote connection type:

hosts: 127.0.0.1
connection: local

Variables From Other Hosts¶

If your database server wants to check the value of a ‘fact’ from another node, it’s easy to do so within a template or even an action line:

{{ hostvars.get('name_of_host').get('name_of_fact') }}

NOTE: No database or other complex system is required to exchange data between hosts. The hosts that you want to reference data from must be included in either the current play or any previous play.

External Variables and Prompted or Sensitive Data¶

It’s a great idea to keep your playbooks under source control, but you may wish to make the playbook source public while keeping certain important variables private. Similarly, sometimes you may just want to keep certain information in different files, away from the main playbook.

You can do this by using an external variables file, or files, just like this:

---
- hosts: all
  user: root
  vars:
    favcolor: blue
  vars_files:
    - /vars/external_vars.yml
  tasks:
  - name: this is just a placeholder
    action: command /bin/echo foo

This removes the risk of sharing sensitive data with others when sharing your playbook source with them.

The contents of each variables file is a simple YAML dictionary, like this:

---
# in the above example, this would be vars/external_vars.yml
somevar: somevalue
password: magic

Alternatively, you may wish to prompt the user for certain input, and can do so with the similarly named ‘vars_prompt’ section. This has uses beyond security, for instance, you may use the same playbook for all software releases and would prompt for a particular release version in a push-script:

---
- hosts: all
  user: root
  vars:
    from: "camelot"
  vars_prompt:
    name: "what is your name?"
    quest: "what is your quest?"
    favcolor: "what is your favorite color?"

There are full examples of both of these items in the github examples/playbooks directory.

Conditional Execution¶

Sometimes you will want to skip a particular step on a particular host. This could be something as simple as not installing a certain package if the operating system is a particular version, or it could be something like performing some cleanup steps if a filesystem is getting full.

This is easy to do in Ansible, with the only_if clause. This clause can be applied to any task, and allows usage of variables from anywhere in ansible, either denoted with $dollar_sign_syntax or {{ braces_syntax }} and then evaluates them with a Python expression. Don’t panic – it’s actually pretty simple:

vars:
  favcolor: blue
  is_favcolor_blue: "'$favcolor' == 'blue'"
  is_centos: "'$facter_operatingsystem' == 'CentOS'"
tasks:
  - name: "shutdown if my favorite color is blue"
    action: command /sbin/shutdown -t now
    only_if: '$is_favcolor_blue'

Variables from tools like facter and ohai can be used here, if installed, or you can use variables that bubble up from ansible (0.3 and later). As a reminder, these variables are prefixed, so it’s $facter_operatingsystem, not $operatingsystem. Ansible’s built in variables are prefixed with ansible_. The only_if expression is actually a tiny small bit of Python, so be sure to quote variables and make something that evaluates to True or False. It is a good idea to use ‘vars_files’ instead of ‘vars’ to define all of your conditional expressions in a way that makes them very easy to reuse between plays and playbooks.

Conditional Imports¶

Sometimes you will want to do certain things differently in a playbook based on certain criteria. Having one playbook that works on multiple platforms and OS versions is a good example.

As an example, the name of the Apache package may be different between CentOS and Debian, but it is easily handled with a minimum of syntax in an Ansible Playbook:

---
- hosts: all
  user: root
  vars_files:
    - "vars/common.yml"
    - [ "vars/$facter_operatingsystem.yml", "vars/os_defaults.yml" ]
  tasks:
  - name: make sure apache is running
    action: service name=$apache state=running

Note that a variable ($facter_operatingsystem) is being interpolated into the list of filenames being defined for vars_files.

As a reminder, the various YAML files contain just keys and values:

---
# for vars/CentOS.yml
apache: httpd
somethingelse: 42

How does this work? If the operating system was ‘CentOS’, the first file Ansible would try to import would be ‘vars/CentOS.yml’, followed up by ‘/vars/os_defaults.yml’ if that file did not exist. If no files in the list were found, an error would be raised. On Debian, it would instead first look towards ‘vars/Debian.yml’ instead of ‘vars/CentOS.yml’, before falling back on ‘vars/os_defaults.yml’. Pretty simple.

To use this conditional import feature, you’ll need facter or ohai installed prior to running the playbook, but you can of course push this out with Ansible if you like:

# for facter
ansible -m yum -a "pkg=facter ensure=installed"
ansible -m yum -a "pkg=ruby-json ensure=installed"

# for ohai
ansible -m yum -a "pkg=ohai ensure=installed"

Ansible’s approach to configuration – seperating variables from tasks, keeps your playbooks from turning into arbitrary code with ugly nested ifs, conditionals, and so on - and results in more streamlined & auditable configuration rules – especially because there are a minimum of decision points to track.

Include Files And Reuse¶

Suppose you want to reuse lists of tasks between plays or playbooks. You can use include files to do this.

An include file simply contains a flat list of tasks, like so:

---
# possibly saved as tasks/foo.yml
- name: placeholder foo
  action: command /bin/foo
- name: placeholder bar
  action: command /bin/bar

Include directives look like this:

- tasks:
   - include: tasks/foo.yml

You can also pass variables into includes directly. We might call this a ‘parameterized include’.

For instance, if deploying multiple wordpress instances, I could contain all of my wordpress tasks in a single wordpress.yml file, and use it like so:

- tasks:
  - include: wordpress.yml user=timmy
  - include: wordpress.yml user=alice
  - include: wordpress.yml user=bob

Variables passed in can be used in the included files. Using jinja2 syntax, in the included file, you can reference them like this:

{{ user }}

or, more simply, using Ansible’s simplified variable syntax:

$user

In addition to the explicitly passed in parameters, all variables from the vars section are also available for use here as well.

Includes can also be used in the ‘handlers’ section, for instance, if you want to define how to restart apache, you only have to do that once for all of your playbooks. You might make a handlers.yml that looks like:

----
# this might be in a file like handlers/handlers.yml
- name: restart apache
  action: service name=apache state=restarted

And in your main playbook file, just include it like so, at the bottom of a play:

handlers:
  - include: handlers/handlers.yml

You can mix in includes along with your regular non-included tasks and handlers.

Note that you can not conditionally path the location to an include file, like you can with ‘vars_files’. If you find yourself needing to do this, consider how you can restructure your playbook to be more class/role oriented.

Using Includes To Assign Classes of Systems¶

Include files are really powerful when used to reuse logic between playbooks. You could imagine a playbook describing your entire infrastructure like this, in a list of just a few plays:

---
- hosts: atlanta-webservers
  vars:
    datacenter: atlanta
  tasks:
  - include: tasks/base.yml
  - include: tasks/webservers.yml database=db.atlanta.com
  handlers:
    - include: handlers/common.yml
- hosts: atlanta-dbservers
  vars:
    datacenter: atlanta
  tasks:
  - include: tasks/base.yml
  - include: tasks/dbservers.yml
  handlers:
    - include: handlers/common.yml

There is one (or more) play defined for each group of systems, and each play maps each group to several includes. These includes represent ‘class definitions’, telling the systems what they are supposed to do or be. In the above example, all hosts get the base configuration first and further customize it depending on what class or nature of machines they are.

Loop Shorthand¶

To save some typing, repeated tasks can be written in short-hand like so:

- name: add user $item
  action: user name=$item state=present groups=wheel
  with_items:
     - testuser1
     - testuser2

The above would be the equivalent of:

- name: add user testuser1
  action: user name=testuser1 state=present groups=wheel
- name: add user testuser2
  action: user name=testuser2 state=present groups=wheel

Asynchronous Actions and Polling¶

By default tasks in playbooks block, meaning the connections stay open until the task is done on each node. If executing playbooks with a small parallelism value (aka --forks), you may wish that long running operations can go faster. The easiest way to do this is to kick them off all at once and then poll until they are done.

You will also want to use asynchronous mode on very long running operations that might be subject to timeout.

To launch a task asynchronously, specify its maximum runtime and how frequently you would like to poll for status. The default poll value is 10 seconds if you do not specify a value for poll:

---
- hosts: all
  user: root
  tasks:
  - name: simulate long running op (15 sec), wait for up to 45, poll every 5
    action: command /bin/sleep 15
    async: 45
    poll: 5

Alternatively, if you do not need to wait on the task to complete, you may “fire and forget” by specifying a poll value of 0:

---
- hosts: all
  user: root
  tasks:
  - name: simulate long running op, allow to run for 45, fire and forget
    action: command /bin/sleep 15
    async: 45
    poll: 0