Inventory Guide

Connection methods and details

This section shows you how to expand and refine the connection methods Ansible uses for your inventory.

ControlPersist and paramiko

By default, Ansible uses native OpenSSH, because it supports ControlPersist (a performance feature), Kerberos, and options in ~/.ssh/config such as Jump Host setup. If your control machine uses an older version of OpenSSH that does not support ControlPersist, Ansible will fall back to a Python implementation of OpenSSH called 'paramiko'.

Setting a remote user

By default, Ansible connects to all remote devices with the username you are using on the control node. If that username does not exist on a remote device, you can set a different username for the connection. If you just need to do some tasks as a different user, look at become. You can set the connection user in a playbook:

---
- name: update webservers
  hosts: webservers
  remote_user: admin

  tasks:
  - name: thing to do first in this playbook

  # ...

as a host variable in inventory:

other1.example.com     ansible_connection=ssh        ansible_user=myuser
other2.example.com     ansible_connection=ssh        ansible_user=myotheruser

or as a group variable in inventory:

cloud:
  hosts:
    cloud1: my_backup.cloud.com
    cloud2: my_backup2.cloud.com
  vars:
    ansible_user: admin

ssh_connection
Details on the remote_user keyword and ansible_user variable.

general_precedence_rules
Details on Ansible precedence rules.

Setting up SSH keys

By default, Ansible assumes you are using SSH keys to connect to remote machines. SSH keys are encouraged, but you can use password authentication if needed with the --ask-pass option. If you need to provide a password for privilege escalation <become> (sudo, pbrun, and so on), use --ask-become-pass.

[!NOTE] Ansible does not expose a channel to allow communication between the user and the ssh process to accept a password manually to decrypt an ssh key when using the ssh connection plugin (which is the default). The use of ssh-agent is highly recommended.

To set up SSH agent to avoid retyping passwords, you can do:

$ ssh-agent bash
$ ssh-add ~/.ssh/id_rsa

Depending on your setup, you may wish to use Ansible's --private-key command line option to specify a pem file instead. You can also add the private key file:

$ ssh-agent bash
$ ssh-add ~/.ssh/keypair.pem

Another way to add private key files without using ssh-agent is using ansible_ssh_private_key_file in an inventory file as explained here: intro_inventory.

Running against localhost

You can run commands against the control node by using "localhost" or "127.0.0.1" for the server name:

$ ansible localhost -m ping -e 'ansible_python_interpreter="/usr/bin/env python"'

You can specify localhost explicitly by adding this to your inventory file:

localhost ansible_connection=local ansible_python_interpreter="/usr/bin/env python"

Managing host key checking

Ansible enables host key checking by default. Checking host keys guards against server spoofing and man-in-the-middle attacks, but it does require some maintenance.

If a host is reinstalled and has a different key in 'known_hosts', this will result in an error message until corrected. If a new host is not in 'known_hosts' your control node may prompt for confirmation of the key, which results in an interactive experience if using Ansible, from say, cron. You might not want this.

If you understand the implications and wish to disable this behavior, you can do so by editing /etc/ansible/ansible.cfg or ~/.ansible.cfg:

[defaults]
host_key_checking = False

Alternatively, this can be set by the ANSIBLE_HOST_KEY_CHECKING environment variable:

$ export ANSIBLE_HOST_KEY_CHECKING=False

Also note that host key checking in paramiko mode is reasonably slow, therefore switching to 'ssh' is also recommended when using this feature.

Other connection methods

Ansible can use a variety of connection methods beyond SSH. You can select any connection plugin, including managing things locally and managing chroot, lxc, and jail containers. A mode called 'ansible-pull' can also invert the system and have systems 'phone home' with scheduled Git checkouts to pull configuration directives from a central repository.

Implicit 'localhost'

When you try to reference a localhost and you don't have it defined in inventory, Ansible will create an implicit one for you.:

- hosts: all
  tasks:
    - name: check that i have log file for all hosts on my local machine
      stat: path=/var/log/hosts/{{inventory_hostname}}.log
      delegate_to: localhost

In a case like this (or local_action) when Ansible needs to contact a 'localhost' but you did not supply one, we create one for you. This host is defined with specific connection variables equivalent to this in an inventory:

# ...

hosts:
  localhost:
   vars:
     ansible_connection: local
     ansible_python_interpreter: "{{ansible_playbook_python}}"

This ensures that the proper connection and Python are used to execute your tasks locally. You can override the built-in implicit version by creating a localhost host entry in your inventory. At that point, all implicit behaviors are ignored; the localhost in inventory is treated just like any other host. Group and host vars will apply, including connection vars, which includes the ansible_python_interpreter setting. This will also affect delegate_to: localhost and local_action, the latter being an alias to the former.

[!NOTE]

Building Ansible inventories

[!NOTE] Making Open Source More Inclusive

Red Hat is committed to replacing problematic language in our code, documentation, and web properties. We are beginning with these four terms: master, slave, blacklist, and whitelist. We ask that you open an issue or pull request if you come upon a term that we have missed. For more details, see our CTO Chris Wright's message.

Welcome to the guide to building Ansible inventories. An inventory is a list of managed nodes, or hosts, that Ansible deploys and configures. This guide introduces you to inventories and covers the following topics:

intro_inventory intro_dynamic_inventory intro_patterns implicit_localhost connection_details

Working with dynamic inventory<span id="intro_dynamic_inventory"></span>

If your Ansible inventory fluctuates over time, with hosts spinning up and shutting down in response to business demands, the static inventory solutions described in inventory will not serve your needs. You may need to track hosts from multiple sources: cloud providers, LDAP, Cobbler, and/or enterprise CMDB systems.

Ansible integrates all of these options through a dynamic external inventory system. Ansible supports two ways to connect with external inventory: inventory_plugins and inventory scripts.

Inventory plugins take advantage of the most recent updates to the Ansible Core code. We recommend plugins over scripts for dynamic inventory. You can write your own plugin <developing_inventory> to connect to additional dynamic inventory sources.

You can still use inventory scripts if you choose. When we implemented inventory plugins, we ensured backwards compatibility through the script inventory plugin. The examples below illustrate how to use inventory scripts.

If you prefer a GUI for handling dynamic inventory, the inventory database on AWX or ansible_platform syncs with all your dynamic inventory sources, provides web and REST access to the results, and offers a graphical inventory editor. With a database record of all of your hosts, you can correlate past event history and see which hosts have had failures on their last playbook runs.

Inventory script example: Cobbler

Ansible integrates seamlessly with Cobbler, a Linux installation server originally written by Michael DeHaan and now led by James Cammarata, who works for Ansible.

While primarily used to kickoff OS installations and manage DHCP and DNS, Cobbler has a generic layer that can represent data for multiple configuration management systems (even at the same time) and serve as a 'lightweight CMDB'.

To tie your Ansible inventory to Cobbler, copy this script to /etc/ansible and chmod +x the file. Run cobblerd any time you use Ansible and use the -i command line option (for example, -i /etc/ansible/cobbler.py) to communicate with Cobbler using Cobbler's XMLRPC API.

Add a cobbler.ini file in /etc/ansible so Ansible knows where the Cobbler server is and some cache improvements can be used. For example:

[cobbler]

# Set Cobbler's hostname or IP address
host = http://127.0.0.1/cobbler_api

# API calls to Cobbler can be slow. For this reason, we cache the results of an API
# call. Set this to the path you want cache files to be written to. Two files
# will be written to this directory:
#   - ansible-cobbler.cache
#   - ansible-cobbler.index

cache_path = /tmp

# The number of seconds a cache file is considered valid. After this many
# seconds, a new API call will be made, and the cache file will be updated.

cache_max_age = 900

First test the script by running /etc/ansible/cobbler.py directly. You should see some JSON data output, but it may not have anything in it just yet.

Let's explore what this does. In Cobbler, assume a scenario somewhat like the following:

cobbler profile add --name=webserver --distro=CentOS6-x86_64
cobbler profile edit --name=webserver --mgmt-classes="webserver" --ksmeta="a=2 b=3"
cobbler system edit --name=foo --dns-name="foo.example.com" --mgmt-classes="atlanta" --ksmeta="c=4"
cobbler system edit --name=bar --dns-name="bar.example.com" --mgmt-classes="atlanta" --ksmeta="c=5"

In the example above, the system 'foo.example.com' is addressable by ansible directly, but is also addressable when using the group names 'webserver' or 'atlanta'. Since Ansible uses SSH, it contacts system foo over 'foo.example.com', only, never just 'foo'. Similarly, if you tried "ansible foo", it would not find the system... but "ansible 'foo*'" would do, because the system DNS name starts with 'foo'.

The script provides more than host and group info. In addition, as a bonus, when the 'setup' module is run (which happens automatically when using playbooks), the variables 'a', 'b', and 'c' will all be auto-populated in the templates:

# file: /srv/motd.j2
Welcome, I am templated with a value of a={{ a }}, b={{ b }}, and c={{ c }}

Which could be executed just like this:

ansible webserver -m setup
ansible webserver -m template -a "src=/tmp/motd.j2 dest=/etc/motd"

[!NOTE] The name 'webserver' came from Cobbler, as did the variables for the config file. You can still pass in your own variables like normal in Ansible, but variables from the external inventory script will override any that have the same name.

So, with the template above (motd.j2), this results in the following data being written to /etc/motd for system 'foo':

Welcome, I am templated with a value of a=2, b=3, and c=4

And on system 'bar' (bar.example.com):

Welcome, I am templated with a value of a=2, b=3, and c=5

And technically, though there is no major good reason to do it, this also works:

ansible webserver -m ansible.builtin.shell -a "echo {{ a }}"

So, in other words, you can use those variables in arguments/actions as well.

Other inventory scripts

In Ansible 2.10 and later, inventory scripts moved to their associated collections. Many are now in the ansible-community/contrib-scripts repository. We recommend you use inventory_plugins instead.

Using inventory directories and multiple inventory sources

If the location given to -i in Ansible is a directory (or as so configured in ansible.cfg), Ansible can use multiple inventory sources at the same time. When doing so, it is possible to mix both dynamic and statically managed inventory sources in the same ansible run. Instant hybrid cloud!

In an inventory directory, executable files are treated as dynamic inventory sources and most other files as static sources. Files which end with any of the following are ignored:

~, .orig, .bak, .ini, .cfg, .retry, .pyc, .pyo

You can replace this list with your own selection by configuring an inventory_ignore_extensions list in ansible.cfg, or setting the ANSIBLE_INVENTORY_IGNORE environment variable. The value in either case must be a comma-separated list of patterns, as shown above.

Any group_vars and host_vars subdirectories in an inventory directory are interpreted as expected, making inventory directories a powerful way to organize different sets of configurations. See using_multiple_inventory_sources for more information.

Static groups of dynamic groups

When defining groups of groups in the static inventory file, the child groups must also be defined in the static inventory file, otherwise ansible returns an error. If you want to define a static group of dynamic child groups, define the dynamic groups as empty in the static inventory file. For example:

[tag_Name_staging_foo]

[tag_Name_staging_bar]

[staging:children]
tag_Name_staging_foo
tag_Name_staging_bar

intro_inventory
All about static inventory files

Communication<communication>
Got questions? Need help? Want to share your ideas? Visit the Ansible communication guide

How to build your inventory<span id="intro_inventory"></span>

Ansible automates tasks on managed nodes or "hosts" in your infrastructure by using a list or group of lists known as inventory. Ansible composes its inventory from one or more 'inventory sources'. While one of these sources can be the list of host names you pass at the command line, most Ansible users create inventory files. Your inventory defines the managed nodes you automate and the variables associated with those hosts. You can also specify groups. Groups allow you to reference multiple associated hosts to target for your automation or to define variables in bulk. Once you define your inventory, you use patterns <intro_patterns> to select the hosts or groups you want Ansible to run against.

The simplest inventory is a single file that contains a list of hosts and groups. The default location for this file is /etc/ansible/hosts. You can specify a different inventory source or sources at the command line by using the -i <path or expression> option or by using the configuration system.

Ansible inventory_plugins supports a range of formats and sources, which makes your inventory flexible and customizable. As your inventory expands, you might need more than a single file to organize your hosts and groups. You have the following common options beyond the /etc/ansible/hosts file:

The following YAML snippets include an ellipsis (...) to indicate that the snippets are part of a larger YAML file. You can find out more about YAML syntax at yaml_basics.

Inventory basics: formats, hosts, and groups

You can create your inventory file in one of many formats, depending on the inventory plugins you have. The most common formats are INI and YAML because Ansible includes built-in support for them. This introduction focuses on these two formats, but many other formats and sources are possible.

A basic INI /etc/ansible/hosts might look like this:

mail.example.com

[webservers]
foo.example.com
bar.example.com

[dbservers]
one.example.com
two.example.com
three.example.com

The headings in brackets are group names. You can use group names to classify hosts and to decide which hosts you are controlling at what times and for what purpose. Group names should follow the same guidelines as valid_variable_names.

Here's the same basic inventory file in YAML format:

ungrouped:
  hosts:
    mail.example.com:
webservers:
  hosts:
    foo.example.com:
    bar.example.com:
dbservers:
  hosts:
    one.example.com:
    two.example.com:
    three.example.com:

Default groups

Even if you do not define any groups in your inventory, Ansible creates two default groups: all and ungrouped. The all group contains every host. The ungrouped group contains all hosts that do not belong to any other group. Every host always belongs to at least two groups (all and ungrouped, or all and another group). For example, in the basic inventory above, the host mail.example.com belongs to the all and ungrouped groups. The host two.example.com belongs to the all and dbservers groups. Although all and ungrouped are always present, they can be implicit and might not appear in group listings like group_names.

Hosts in multiple groups

You can put a host in more than one group. For example, you can include a production web server in a data center in Atlanta in the [prod], [atlanta], and [webservers] groups. You can create groups that track the following criteria:

The following example extends the previous YAML inventory to include what, when, and where:

ungrouped:
  hosts:
    mail.example.com:
webservers:
  hosts:
    foo.example.com:
    bar.example.com:
dbservers:
  hosts:
    one.example.com:
    two.example.com:
    three.example.com:
east:
  hosts:
    foo.example.com:
    one.example.com:
    two.example.com:
west:
  hosts:
    bar.example.com:
    three.example.com:
prod:
  hosts:
    foo.example.com:
    one.example.com:
    two.example.com:
test:
  hosts:
    bar.example.com:
    three.example.com:

As the example shows, one.example.com exists in the dbservers, east, and prod groups.

Grouping groups: parent/child group relationships

You can create parent/child relationships among groups. Parent groups are also known as nested groups or groups of groups. For example, if all your production hosts are already in groups such as  atlanta_prod and denver_prod, you can create a production group that includes those smaller groups. This approach reduces maintenance because you add or remove hosts from the parent group by editing the child groups.

To create parent/child relationships for groups, use one of the following methods:

The following example shows the same inventory as above, simplified with parent groups for the prod and test groups:

ungrouped:
  hosts:
    mail.example.com:
webservers:
  hosts:
    foo.example.com:
    bar.example.com:
dbservers:
  hosts:
    one.example.com:
    two.example.com:
    three.example.com:
east:
  hosts:
    foo.example.com:
    one.example.com:
    two.example.com:
west:
  hosts:
    bar.example.com:
    three.example.com:
prod:
  children:
    east:
test:
  children:
    west:

Note the following properties of child groups:

Adding ranges of hosts

Some plugins, like YAML and INI, support adding ranges of hosts. If you have many hosts with a similar pattern, you can add the hosts as a range rather than listing each hostname separately:

In INI:

[webservers]
www[01:50].example.com

In YAML:

# ...
  webservers:
    hosts:
      www[01:50].example.com:

You can specify a stride (increments between sequence numbers) when you define a numeric range of hosts:

In INI:

[webservers]
www[01:50:2].example.com

In YAML:

# ...
  webservers:
    hosts:
      www[01:50:2].example.com:

The example above matches the subdomains www01, www03, www05, ..., www49, but not www00, www02, www50, and so on, because the stride (increment) is 2 units for each step.

For numeric patterns, you can include or remove leading zeros as desired. Ranges are inclusive. You can also define alphabetic ranges:

[databases]
db-[a:f].example.com

Passing multiple inventory sources

You can target multiple inventory sources (static files, directories, dynamic inventory scripts or anything supported by inventory plugins) at the same time. To do this, specify multiple inventory sources from the command line (see below) or by configuration, either by setting ANSIBLE_INVENTORY or in ansible.cfg (DEFAULT_HOST_LIST). This capability can be useful when you want to target normally separate environments, like staging and production, at the same time for a specific action.

To target two inventory sources from the command line:

ansible-playbook get_logs.yml -i staging -i production

Organizing inventory in a directory

You can consolidate multiple inventory sources in a single directory. The simplest version of this approach is a directory with multiple files instead of a single inventory file. Maintaining a single file becomes difficult when the file gets too long. If you have multiple teams and multiple automation projects, creating one inventory file per team or project lets everyone easily find the hosts and groups that matter to them. You can also still use the files individually or in subsets, depending on how you configure or call Ansible.

These files can use all formats or plugin configurations (for example, YAML or INI). In this case, your directory becomes your 'single' inventory source, and Ansible aggregates the multiple sources it finds in that directory. By default, Ansible ignores some directories and extensions, but you can change this behavior in the configuration (INVENTORY_IGNORE_PATTERNS and INVENTORY_IGNORE_EXTS).

You can also combine multiple inventory source types in an inventory directory. This method can be useful for combining static and dynamic hosts and managing them as one inventory. The following inventory directory combines an inventory plugin source, a dynamic inventory script, and a file with static hosts:

inventory/
  openstack.yml          # configure inventory plugin to get hosts from OpenStack cloud
  dynamic-inventory.py   # add additional hosts with dynamic inventory script
  on-prem                # add static hosts and groups
  parent-groups          # add static hosts and groups

You can target this inventory directory as follows:

ansible-playbook example.yml -i inventory

You can also configure the inventory directory in your ansible.cfg file. See intro_configuration for more details.

Ansible reads and loads files from the top directory down in alphabetically sorted order.

Managing inventory load order

Ansible loads inventory sources in the order you supply them. It defines hosts, groups, and variables as it encounters them in the source files, adding the all and ungrouped groups at the end if needed.

Depending on the inventory plugin or plugins you use, you might need to rearrange the order of sources to ensure that parent/child-defined groups or hosts exist as the plugins expect. Otherwise, you might encounter a parsing error. For example, the YAML and INI inventory plugins discard empty groups (groups with no associated hosts) when they finish processing each source.

If you define a variable multiple times, Ansible overwrites the previous value. The last definition wins.

Adding variables to inventory

You can define variables that relate to a specific host or group in your inventory. A simple way to start is by adding variables directly to the hosts and groups in a YAML or INI inventory source.

This guide documents how to add variables in the inventory source for simplicity. However, you can also use vars_plugins to add variables from many other sources. By default, Ansible ships with the host_group_vars <host_group_vars_vars> plugin, which allows you to define variables in separate host and group variable files. Using separate files is a more robust approach to describing your system policy than defining variables in the inventory source. See splitting_out_vars for guidelines on how to store variable values in individual files in the 'host_vars' and 'group_vars' directories. 

Assigning a variable to one machine: host variables

You can easily assign a variable to a single host and then use that variable later in playbooks. You can do this directly in your inventory file.

In INI:

[atlanta]
host1 http_port=80 maxRequestsPerChild=808
host2 http_port=303 maxRequestsPerChild=909

In YAML:

atlanta:
  hosts:
    host1:
      http_port: 80
      maxRequestsPerChild: 808
    host2:
      http_port: 303
      maxRequestsPerChild: 909

Unique values like non-standard SSH ports work well as host variables. You can add them to your Ansible inventory by adding the port number after the hostname with a colon:

badwolf.example.com:5309

You can use host variables to define 'Connection variables'. Connection variables configure connection, shell, and become plugins to enable task execution on the host. For example:

[targets]

localhost              ansible_connection=local
other1.example.com     ansible_connection=ssh        ansible_user=myuser
other2.example.com     ansible_connection=ssh        ansible_user=myotheruser

Inventory aliases

The inventory_hostname is the unique identifier for a host in Ansible. This identifier can be an IP address or a hostname, but it can also be just an 'alias' or short name for the host.

In INI:

jumper ansible_port=5555 ansible_host=192.0.2.50

In YAML:

# ...
  hosts:
    jumper:
      ansible_port: 5555
      ansible_host: 192.0.2.50

In this example, running Ansible against the host alias "jumper" connects to 192.0.2.50 on port 5555. See behavioral inventory parameters <behavioral_parameters> to further customize the connection to hosts.

This feature is also useful for targeting the same host more than once, but remember that tasks can run in parallel:

In INI:

jumper1 ansible_port=5555 ansible_host=192.0.2.50
jumper2 ansible_port=5555 ansible_host=192.0.2.50

In YAML:

# ...
  hosts:
    jumper1:
      ansible_port: 5555
      ansible_host: 192.0.2.50
    jumper2:
      ansible_port: 5555
      ansible_host: 192.0.2.50

Defining variables in INI format

Ansible interprets values that you pass in the INI format by using the key=value syntax differently depending on where you declare them:

If you need a variable from an INI inventory to have a certain type (for example, a string or a boolean), always specify the type with a filter in your task. Do not rely on types that you set in INI inventories when you consume variables.

Consider using the YAML format for inventory sources to avoid confusion about the actual type of a variable. The YAML inventory plugin processes variable values consistently and correctly.

Assigning a variable to many machines: group variables

If all hosts in a group share a variable value, you can apply that variable to an entire group at once.

In INI:

[atlanta]
host1
host2

[atlanta:vars]
ntp_server=ntp.atlanta.example.com
proxy=proxy.atlanta.example.com

In YAML:

atlanta:
  hosts:
    host1:
    host2:
  vars:
    ntp_server: ntp.atlanta.example.com
    proxy: proxy.atlanta.example.com

Group variables are a convenient way to apply variables to multiple hosts at once. Before executing, however, Ansible always flattens variables, including inventory variables, to the host level. If a host is a member of multiple groups, Ansible reads variable values from all of those groups. If you assign different values to the same variable in different groups, Ansible chooses which value to use based on internal rules for merging <how_we_merge>.

Inheriting variable values: group variables for groups of groups

You can apply variables to parent groups (nested groups or groups of groups) as well as to child groups. The syntax is the same: :vars for INI format and vars: for YAML format:

In INI:

[atlanta]
host1
host2

[raleigh]
host2
host3

[southeast:children]
atlanta
raleigh

[southeast:vars]
some_server=foo.southeast.example.com
halon_system_timeout=30
self_destruct_countdown=60
escape_pods=2

[usa:children]
southeast
northeast
southwest
northwest

In YAML:

usa:
  children:
    southeast:
      children:
        atlanta:
          hosts:
            host1:
            host2:
        raleigh:
          hosts:
            host2:
            host3:
      vars:
        some_server: foo.southeast.example.com
        halon_system_timeout: 30
        self_destruct_countdown: 60
        escape_pods: 2
    northeast:
    northwest:
    southwest:

A child group's variables have higher precedence (they override) than a parent group's variables.

Organizing host and group variables

Although you can define variables in the inventory source, you can also use vars_plugins to define alternate sources for your variables.

The default vars plugin that Ansible ships with, host_group_vars <host_group_vars_vars>, lets you use separate host and group variable files. This method helps you organize your variable values more easily. You can also use lists and hash data in these files, which you cannot do in your main inventory file.

For the host_group_vars plugin, your host and group variable files must use YAML syntax. Valid file extensions are '.yml', '.yaml', '.json', or no file extension. See yaml_syntax if you are new to YAML.

The host_group_vars plugin loads host and group variable files by searching paths relative to the inventory source or the playbook file. If your inventory file at /etc/ansible/hosts contains a host named 'foosball' that belongs to the raleigh and webservers groups, that host will use variables from the YAML files in the following locations:

/etc/ansible/group_vars/raleigh # can optionally end in '.yml', '.yaml', or '.json'
/etc/ansible/group_vars/webservers
/etc/ansible/host_vars/foosball

For example, if you group hosts in your inventory by datacenter, and each datacenter uses its own NTP server and database server, you can create a file named /etc/ansible/group_vars/raleigh to store the variables for the raleigh group:

---
ntp_server: acme.example.org
database_server: storage.example.org

You can also create directories named after your groups or hosts. Ansible reads all the files in these directories in lexicographical order. Here is an example with the 'raleigh' group:

/etc/ansible/group_vars/raleigh/db_settings
/etc/ansible/group_vars/raleigh/cluster_settings

All hosts in the 'raleigh' group have the variables that you define in these files available to them. This method is very useful for keeping your variables organized when a single file gets too big, or when you want to use Ansible Vault <playbooks_vault> on some group variables.

Ansible's host_group_vars <host_group_vars_vars> vars plugin can also add group_vars/ and host_vars/ directories to your playbook directory when you use ansible-playbook. However, not all Ansible commands have a playbook (for example, ansible or ansible-console). For those commands, you can use the --playbook-dir option to provide the directory on the command line. If you have sources for the vars plugins relative to both the playbook directory and the inventory directory, the variables that Ansible sources relative to the playbook override the variables that it sources relative to the inventory source.

To track changes to your inventory and variable definitions, keep your inventory sources and their relative variable directories and files in a Git repository or other version control system.

How variables are merged

[!NOTE] Ansible merges variables from different sources and applies precedence to some variables over others according to a set of rules. For example, variables that occur higher in an inventory can override variables that occur lower in the inventory. See ansible_variable_precedence for more information.

Before it runs a play, Ansible merges and flattens variables to the specific host. This process keeps Ansible focused on the Host and Task, so groups do not survive outside of inventory and host matching. By default, Ansible overwrites variables, including the ones that you define for a group or host (see DEFAULT_HASH_BEHAVIOUR<DEFAULT_HASH_BEHAVIOUR>). The order/precedence for inventory entities is (from lowest to highest):

The following list shows the order of precedence for inventory entities, from lowest to highest:

By default, Ansible merges groups at the same parent/child level in alphabetical order. Variables from the last group that Ansible loads overwrite variables from the previous groups. For example, Ansible merges an a_group with a b_group, and matching variables from b_group overwrite the variables in a_group.

You can fine-tune this merge behavior by setting the group variable ansible_group_priority. This variable overrides the alphabetical sorting for the merge order for groups of the same level (after Ansible resolves the parent/child order). The larger the number, the later Ansible merges the group, giving it higher priority. This variable defaults to 1 if you do not set it. For example:

a_group:
  vars:
    testvar: a
    ansible_group_priority: 10
b_group:
  vars:
    testvar: b

In this example, if both groups have the same priority, the result would normally be testvar == b. However, because we give a_group a higher priority, the result is testvar == a.

You can set ansible_group_priority only in an inventory source, not in group_vars/. Ansible uses this variable when it loads the group_vars/ directory.

Managing inventory variable load order

This section describes how to control variable precedence by managing the load order of inventory sources. You can pass sources in a specific order at the command line or use prefixes in the filenames of sources within a directory.

When you use multiple inventory sources, remember that Ansible resolves any variable conflicts according to the rules described in how_we_merge and ansible_variable_precedence. You can control the merging order of variables in inventory sources to get the variable value you need.

When you pass multiple inventory sources at the command line, Ansible merges variables in the order you pass those parameters. If the [all:vars] section in the staging inventory defines myvar = 1 and the production inventory defines myvar = 2, then the following outcomes are true:

When you put multiple inventory sources in a directory, Ansible merges the sources in alphabetical order according to their filenames. You can control the load order by adding prefixes to the files:

inventory/
  01-openstack.yml          # configure inventory plugin to get hosts from Openstack cloud
  02-dynamic-inventory.py   # add additional hosts with dynamic inventory script
  03-static-inventory       # add static hosts
  group_vars/
    all.yml                 # assign variables to all hosts

If 01-openstack.yml defines myvar = 1 for the group all, 02-dynamic-inventory.py defines myvar = 2, and 03-static-inventory defines myvar = 3, Ansible runs the playbook with myvar = 3.

For more details on inventory plugins and dynamic inventory scripts see inventory_plugins and intro_dynamic_inventory.

Connecting to hosts: behavioral inventory parameters

As described above, you can set the following variables to control how Ansible interacts with remote hosts.

Host connection:

[!NOTE] Ansible does not expose a channel to allow communication between the user and the ssh process to accept a password manually to decrypt an ssh key when using the ssh connection plugin (which is the default). The use of ssh-agent is highly recommended.

ansible_connection
Specifies the connection type to the host. This can be the name of any Ansible connection plugin. SSH protocol types are ssh or paramiko. The default is ssh.

General for all connections:

ansible_host
Specifies the resolvable name or IP of the host to connect to, if it is different from the alias you wish to give to it. Never set it to depend on inventory_hostname. If you really need something like that, use inventory_hostname_short so it can work with delegation.

ansible_port
The connection port number, if not the default (22 for ssh).

ansible_user
The username to use when connecting (logging in) to the host.

ansible_password
The password to use to authenticate to the host. (Never store this variable in plain text. Always use a vault. See tip_for_variables_and_vaults.)

Specific to the SSH connection plugin:

ansible_ssh_private_key_file
Private key file used by SSH. This is useful if you use multiple keys and you do not want to use SSH agent.

ansible_ssh_common_args
Ansible always appends this setting to the default command line for sftp, scp, and ssh. This is useful for configuring a ProxyCommand for a certain host or group.

ansible_sftp_extra_args
Ansible always appends this setting to the default sftp command line.

ansible_scp_extra_args
Ansible always appends this setting to the default scp command line.

ansible_ssh_extra_args
Ansible always appends this setting to the default ssh command line.

ansible_ssh_pipelining
Specifies whether to use SSH pipelining. This can override the pipelining setting in ansible.cfg.

ansible_ssh_executable (added in version 2.2)
This setting overrides the default behavior to use the system ssh. It can override the ssh_executable setting in the ssh_connection section of ansible.cfg.

Privilege escalation (see Ansible Privilege Escalation<become> for further details):

ansible_become
Equivalent to ansible_sudo or ansible_su; allows you to force privilege escalation.

ansible_become_method
Allows you to set the privilege escalation method to a matching become plugin.

ansible_become_user
Equivalent to ansible_sudo_user or ansible_su_user; allows you to set the user you become through privilege escalation.

ansible_become_password
Equivalent to ansible_sudo_password or ansible_su_password; allows you to set the privilege escalation password. (Never store this variable in plain text. Always use a vault. See tip_for_variables_and_vaults.)

ansible_become_exe
Equivalent to ansible_sudo_exe or ansible_su_exe; allows you to set the executable for the escalation method you selected.

ansible_become_flags
Equivalent to ansible_sudo_flags or ansible_su_flags; allows you to set the flags passed to the selected escalation method. You can also set this globally in ansible.cfg in the become_flags option under privilege_escalation.

Remote host environment parameters:

ansible_shell_type
Specifies the shell type of the target system. You should not use this setting unless you have set the ansible_shell_executable<ansible_shell_executable> to a non-Bourne (sh) compatible shell.  By default, Ansible formats commands using sh-style syntax.  If you set this to csh or fish, commands that Ansible executes on target systems follow those shell's syntax instead.

ansible_python_interpreter
Specifies the target host Python path. This is useful for systems with more than one Python or for systems where Python is not located at /usr/bin/python, such as *BSD, or where /usr/bin/python is not a 2.X series Python.  We do not use the /usr/bin/env mechanism because that requires the remote user's path to be set correctly and also assumes the python executable is named python, where the executable might be named something like python2.6.

ansible*_interpreter
Works for any language, such as Ruby or Perl, and works just like ansible_python_interpreter<ansible_python_interpreter>. This variable replaces the shebang of modules that will run on that host.

2.1

ansible_shell_executable
This setting sets the shell the Ansible control node will use on the target machine. It overrides executable in ansible.cfg, which defaults to /bin/sh.  You should only change this value if it is not possible to use /bin/sh (in other words, if /bin/sh is not installed on the target machine or cannot be run from sudo.).

Examples from an Ansible-INI host file:

some_host         ansible_port=2222     ansible_user=manager
aws_host          ansible_ssh_private_key_file=/home/example/.ssh/aws.pem
freebsd_host      ansible_python_interpreter=/usr/local/bin/python
ruby_module_host  ansible_ruby_interpreter=/usr/bin/ruby.1.9.3

Non-SSH connection types

As stated in the previous section, Ansible executes playbooks over SSH by default, but it is not limited to this connection type. You can change the connection type with the host-specific parameter ansible_connection=<connection plugin name>. For a full list of available plugins and examples, see connection_plugin_list.

Inventory setup examples

See also sample_setup, which shows inventory along with playbooks and other Ansible artifacts.

Example: One inventory per environment

If you need to manage multiple environments, consider defining only the hosts of a single environment in each inventory. This way, it is harder to, for example, accidentally change the state of nodes inside the "test" environment when you wanted to update some "staging" servers.

For the example mentioned above, you could have an inventory_test file:

[dbservers]
db01.test.example.com
db02.test.example.com

[appservers]
app01.test.example.com
app02.test.example.com
app03.test.example.com

That file only includes hosts that are part of the "test" environment. You can define the "staging" machines in another file called inventory_staging:

[dbservers]
db01.staging.example.com
db02.staging.example.com

[appservers]
app01.staging.example.com
app02.staging.example.com
app03.staging.example.com

To apply a playbook called site.yml to all the app servers in the test environment, use the following command:

ansible-playbook -i inventory_test -l appservers site.yml

Example: Group by function

In the previous section, you already saw an example of using groups to cluster hosts that have the same function. This approach allows you, for example, to define firewall rules inside a playbook or role that affect only database servers:

- hosts: dbservers
  tasks:
  - name: Allow access from 10.0.0.1
    ansible.builtin.iptables:
      chain: INPUT
      jump: ACCEPT
      source: 10.0.0.1

Example: Group by location

Other tasks might focus on where a certain host is located. Let's say that db01.test.example.com and app01.test.example.com are located in DC1, while db02.test.example.com is in DC2:

[dc1]
db01.test.example.com
app01.test.example.com

[dc2]
db02.test.example.com

In practice, you might end up mixing all these setups. For example, you might need to update all nodes in a specific data center on one day, while on another day, you might need to update all the application servers no matter their location.

inventory_plugins
Pulling inventory from dynamic or static sources

intro_dynamic_inventory
Pulling inventory from dynamic sources, such as cloud providers

intro_adhoc
Examples of basic commands

working_with_playbooks
Learning Ansible's configuration, deployment, and orchestration language.

Communication<communication>
Got questions? Need help? Want to share your ideas? Visit the Ansible communication guide

Patterns: targeting hosts and groups

When you execute Ansible through an ad hoc command or by running a playbook, you must choose which managed nodes or groups you want to execute against. Patterns let you run commands and playbooks against specific hosts and/or groups in your inventory. An Ansible pattern can refer to a single host, an IP address, an inventory group, a set of groups, or all hosts in your inventory. Patterns are highly flexible - you can exclude or require subsets of hosts, use wildcards or regular expressions, and more. Ansible executes on all inventory hosts included in the pattern.

Using patterns

You use a pattern almost any time you execute an ad hoc command or a playbook. The pattern is the only element of an ad hoc command<intro_adhoc> that has no flag. It is usually the second element:

ansible <pattern> -m <module_name> -a "<module options>"

For example:

ansible webservers -m service -a "name=httpd state=restarted"

In a playbook, the pattern is the content of the hosts: line for each play:

- name: <play_name>
  hosts: <pattern>

For example:

- name: restart webservers
  hosts: webservers

Since you often want to run a command or playbook against multiple hosts at once, patterns often refer to inventory groups. Both the ad hoc command and the playbook above will execute against all machines in the webservers group.

Common patterns

This table lists common patterns for targeting inventory hosts and groups.

Description Pattern(s) Targets
All hosts all (or *)
One host host1
Multiple hosts host1:host2 (or host1,host2)
One group webservers
Multiple groups webservers:dbservers all hosts in webservers plus all hosts in dbservers
Excluding groups webservers:!atlanta all hosts in webservers except those in atlanta
Intersection of groups webservers:&staging any hosts in webservers that are also in staging

[!NOTE] You can use either a comma (,) or a colon (:) to separate a list of hosts. The comma is preferred when dealing with ranges and IPv6 addresses.

Once you know the basic patterns, you can combine them. This example:

webservers:dbservers:&staging:!phoenix

targets all machines in the groups 'webservers' and 'dbservers' that are also in the group 'staging', except for any machines in the group 'phoenix'.

You can use wildcard patterns with FQDNs or IP addresses, as long as the hosts are named in your inventory by FQDN or IP address:

192.0.*
*.example.com
*.com

You can mix wildcard patterns and groups at the same time:

one*.com:dbservers

Limitations of patterns

Patterns depend on inventory. If a host or group is not listed in your inventory, you cannot use a pattern to target it. If your pattern includes an IP address or hostname that does not appear in your inventory, you will see an error like this:

[WARNING]: No inventory was parsed, only implicit localhost is available
[WARNING]: Could not match supplied host pattern, ignoring: *.not_in_inventory.com

Your pattern must match your inventory syntax. If you define a host as an alias<inventory_aliases>:

atlanta:
  hosts:
    host1:
      http_port: 80
      maxRequestsPerChild: 808
      ansible_host: 127.0.0.2

you must use the alias in your pattern. In the example above, you must use host1 in your pattern. If you use the IP address, you will once again get the error:

[WARNING]: Could not match supplied host pattern, ignoring: 127.0.0.2

Pattern processing order

The processing is a bit special and happens in the following order:

  1. : and ,
  2. &
  3. !

This positioning only accounts for processing order inside each operation: a:b:&c:!d:!e == &c:a:!d:b:!e == !d:a:!e:&c:b

All of these result in the following:

Host in/is (a or b) AND host in/is all(c) AND host NOT in/is all(d, e).

Now a:b:!e:!d:&c is a slight change as the !e gets processed before the !d, though this doesn't make much of a difference:

Host in/is (a or b) AND host in/is all(c) AND host NOT in/is all(e, d).

Advanced pattern options

The common patterns described above will meet most of your needs, but Ansible offers several other ways to define the hosts and groups you want to target.

Using variables in patterns

You can use variables to enable passing group specifiers with the -e argument to ansible-playbook:

webservers:!{{ excluded }}:&{{ required }}

Using group position in patterns

You can define a host or subset of hosts by its position in a group. For example, given the following group:

[webservers]
cobweb
webbing
weber

you can use subscripts to select individual hosts or ranges within the webservers group.

Slicing at specific items

If i is negative, the index is relative to the end of sequence s: len(s) + i is substituted. However -0 is 0.

webservers[0]       # == cobweb
webservers[-1]      # == weber

Slicing with start and end points

The slice of s from i to j is defined as the sequence of items with index k such that i <= k <= j. If i is omitted, use 0. If j is omitted, use len(s). The slice omitting both i and j, results in an invalid host pattern. If i is greater than j, the slice is empty. If i is equal to j, the s[i] is substituted.

webservers[0:2]     # == webservers[0],webservers[1],webservers[2]
                    # == cobweb,webbing,weber
webservers[1:2]     # == webservers[1],webservers[2]
                    # == webbing,weber
webservers[1:]      # == webbing,weber
webservers[:3]      # == cobweb,webbing,weber

Using regexes in patterns

You can specify a pattern as a regular expression by starting the pattern with ~:

~(web|db).*\.example\.com

Patterns and ad-hoc commands

You can change the behavior of the patterns defined in ad-hoc commands using command-line options. You can also limit the hosts you target on a particular run with the --limit flag.

$ ansible all -m <module> -a "<module options>" --limit "host1"

$ ansible all -m <module> -a "<module options>" --limit "host1,host2"

$ ansible all -m <module> -a "<module options>" --limit 'all:!host1'

$ ansible all -m <module> -a "<module options>" --limit 'group1'

Patterns and ansible-playbook flags

You can change the behavior of the patterns defined in playbooks using command-line options. For example, you can run a playbook that defines hosts: all on a single host by specifying -i 127.0.0.2, (note the trailing comma). This works even if the host you target is not defined in your inventory, but this method will NOT read your inventory for variables tied to this host and any variables required by the playbook will need to be specified manually at the command line. You can also limit the hosts you target on a particular run with the --limit flag, which will reference your inventory:

ansible-playbook site.yml --limit datacenter2

Finally, you can use --limit to read the list of hosts from a file by prefixing the file name with @:

ansible-playbook site.yml --limit @retry_hosts.txt

If RETRY_FILES_ENABLED is set to True, a .retry file will be created after the ansible-playbook run containing a list of failed hosts from all plays. This file is overwritten each time ansible-playbook finishes running.

ansible-playbook site.yml --limit @site.retry

To apply your knowledge of patterns with Ansible commands and playbooks, read intro_adhoc and playbooks_intro.

intro_adhoc
Examples of basic commands

working_with_playbooks
Learning the Ansible configuration management language

Communication<communication>
Got questions? Need help? Want to share your ideas? Visit the Ansible communication guide