ansible/docsite/latest/rst/intro_configuration.rst

441 lines
15 KiB
ReStructuredText

The Ansible Configuration File
==============================
.. highlight:: bash
Certain things in Ansible are adjustable. In general, the stock configuration is probably
right for most users, but that doesn't mean you might not want to change them.
The mechanism for doing this is the "ansible.cfg" file, which is looked for in the following locations:
* /etc/ansible/ansible.cfg
* ~/.ansible.cfg
* ansible.cfg (in the playbook directory)
If multiple file locations matching the above exist, the last location on the list is used.
.. contents::
:depth: 2
Getting the latest configuration
````````````````````````````````
If installing ansible from a package manager, the latest ansible.cfg should be present in /etc/ansible, possibly
as a ".rpmnew" file (or other) as appropriate in the case of updates.
If you have installed from pip or from source, however, you may want to create this file in order to override
default settings in Ansible.
You may wish to consult the `ansible.cfg in source control <https://raw.github.com/ansible/ansible/devel/examples/ansible.cfg>_` for
all of the possible latest values.
Environmental configuration
```````````````````````````
Ansible also allows configuration of settings via environment variables. If these environment variables are set, they will
override any setting loaded from the configuration file.
Explanation of values by section
````````````````````````````````
General defaults
----------------
In the [defaults] section of ansible.cfg, the following settings are tunable:
action_plugins
==============
Actions are pieces of code in ansible that enable things like module execution, templating, and so forth.
This is a developer-centric feature that allows low-level extensions around Ansible to be loaded from
different locations::
action_plugins = /usr/share/ansible_plugins/action_plugins
Most users will not need to use this feature. See `developing_plugins` for more details.
ansible_managed
===============
Ansible-managed is a string that can be inserted into files written by Ansible's config templating system, if you use
a string like::
{{ ansible_managed }}
The default configuration shows who modified a file and when::
ansible_managed = Ansible managed: {file} modified on %Y-%m-%d %H:%M:%S by {uid} on {host}
This is useful to tell users that a file has been placed by Ansible and manual changes are likely to be overwritten.
ask_pass
========
This controls whether an Ansible playbook should prompt for a password by default. The default behavior is no::
#ask_pass=True
If using SSH keys for authentication, it's probably not needed to change this setting.
ask_sudo_pass
=============
Similar to ask_pass, this controls whether an Ansible playbook should prompt for a sudo password by default when
sudoing. The default behavior is also no::
#ask_sudo_pass=True
Users on platforms where sudo passwords are enabled should consider changing this setting.
callback_plugins
================
This is a developer-centric feature that allows low-level extensions around Ansible to be loaded from
different locations::
action_plugins = /usr/share/ansible_plugins/action_plugins
Most users will not need to use this feature. See `developing_plugins` for more details
connection_plugins
==================
This is a developer-centric feature that allows low-level extensions around Ansible to be loaded from
different locations::
action_plugins = /usr/share/ansible_plugins/action_plugins
Most users will not need to use this feature. See `developing_plugins` for more details
error_on_undefined_vars
=======================
On by default since Ansible 1.3, this causes ansible to fail steps that reference variable names that are likely
typoed::
#error_on_undefined_vars=True
If set to False, any '{{ template_expression }}' that contains undefined variables will be rendered in a template
or ansible action line exactly as written.
executable
==========
This indicates the command to use to spawn a shell under a sudo environment. Users may need to change this in
rare instances to /bin/bash in rare instances wehn sudo is constrained, but in most cases it may be left as is::
#executable = /bin/bash
forks
=====
This is the default number of parallel processes to spawn when communicating with remote hosts. Since Ansible 1.3,
the fork number is automatically limited to the number of possible hosts, so this is really a limit of how much
network and CPU load you think you can handle. Many users may set this to 50, some set it to 500 or more. If you
have a large number of hosts, higher values will make actions across all of those hosts complete faster. The default
is very very conservative::
forks=5
pattern
=======
This is the default group of hosts to talk to in a playbook if no "hosts:" stanza is supplied. The default is to talk
to all hosts. You may wish to change this to protect yourself from surprises::
hosts=*
Note that /usr/bin/ansible always requires a host pattern and does not use this setting, only /usr/bin/ansible-playbook.
poll_interval
=============
For asynchronous tasks in Ansible (covered in `playbooks_async`), this is how often to check back on the status of those
tasks when an explicit poll interval is not supplied. The default is a reasonably moderate 15 seconds which is a tradeoff
between checking in frequently and providing a quick turnaround when something may have completed::
poll_interval=15
hash_behavior
=============
Ansible by default will override variables in specific precedence orders, as described in `playbooks_variables`. When a variable
of higher precedence wins, it will replace the other value.
Some users prefer that variables that are hashes (aka 'dictionaries' in Python terms) are merged together. This setting is called 'merge'. This is not the default behavior and it does not affect variables whose values are scalars (integers, strings) or
arrays. We generally recommend not using this setting unless you think you have an absolute need for it, and playbooks in the
official examples repos do not use this setting::
#hash_behavior=replace
The valid values are either 'replace' (the default) or 'merge'.
hostfile
========
This is the default location of the inventory file, script, or directory that Ansible will use to determine what hosts it has available
to talk to::
hostfile = /etc/ansible/hosts
host_key_checking
=================
As described in `intro_gettingstarted`, host key checking is on by default in Ansible 1.3 and later. If you understand the
implications and wish to disable it, you may do so here by setting the value to False::
host_key_checking=True
filter_plugins
==============
This is a developer-centric feature that allows low-level extensions around Ansible to be loaded from
different locations::
action_plugins = /usr/share/ansible_plugins/action_plugins
Most users will not need to use this feature. See `developing_plugins` for more details
jinja2_extensions
=================
This is a developer-specific feature that allows enabling additional Jinja2 extensions::
jinja2_extensions = jinja2.ext.do,jinja2.ext.i18n
If you do not know what these do, you probably don't need to change this setting :)
legacy_playbook_variables
=========================
Ansible prefers to use Jinja2 syntax '{{ like_this }}' to indicate a variable should be substituted in a particular string. However,
older versions of playbooks used a more Perl-style syntax. This syntax was undesirable as it frequently conflicted with bash and
was hard to explain to new users when deferencing complicated variable hierarchies, so we have standardized on the '{{ jinja2 }}' way.
To ensure a string like '$foo' is not inadvertedly replaced in a Perl or Bash script template, the old form of templating (which is
still enabled as of Ansible 1.4) can be disabled like so ::
legacy_playbook_variables = no
library
=======
This is the default location Ansible looks to find modules::
library = /usr/share/ansible
Ansible knows how to look in multiple locations if you feed it a colon seperated path, and it also will look for modules in the
"./library" directory alongside a playbook.
log_path
========
If present and configured in ansible.cfg, Ansible will log information about executions at the designated location. Be sure
the user running Ansible has permissions on the logfile.
log_path=/var/log/ansible.log
This behavior is not on by default.
lookup_plugins
==============
This is a developer-centric feature that allows low-level extensions around Ansible to be loaded from
different locations::
action_plugins = /usr/share/ansible_plugins/action_plugins
Most users will not need to use this feature. See `developing_plugins` for more details
module_name
===========
This is the default module name (-m) value for /usr/bin/ansible. The default is the 'command' module.
Remember the command module doesn't support shell variables, pipes, or quotes, so you might wish to change
it to 'shell'::
module_name = command
nocolor
=======
By default ansible will try to colorize output to give a better indication of failure and status information.
If you dislike this behavior you can turn it off by setting 'nocolor' to 0::
nocolor=0
nocows
======
By default ansible will take advantage of cowsay if installed to make /usr/bin/ansible-playbook runs more exciting.
Why? We believe systems management should be a happy experience. If you do not like the cows, you can disable them
by setting 'nocows' to 1::
nocows=0
private_key_file
================
If you are using a pem file to authenticate with machines rather than SSH agent or passwords, you can set the default
value here to avoid re-specifying --ansible-private-keyfile with every invocation::
private_key_file=/path/to/file.pem
remote_port
===========
This sets the default SSH port on all of your systems, for systems that didn't specify an alternative value in inventroy.
The default is the standard 22::
remote_port = 22
remote_tmp
==========
Ansible works by transferring modules to your remote machines, running them, and then cleaning up after itself. In some
cases, you may not wish to use the default location and would like to change the path. You can do so by altering this
setting::
remote_temp = $HOME/.ansible/tmp
The default is to use a subdirectory of the user's home directory. Ansible will then choose a random directory name
inside this location.
remote_user
===========
This is the default username ansible will connect as for /usr/bin/ansible-playbook. Note that /usr/bin/ansible will
always default to the current user::
remote_user = root
sudo_exe
========
If using an alternative sudo implementation on remote machines, the path to sudo can be replaced here provided
the sudo implementation is matching CLI flags with the standard sudo::
sudo_exe=sudo
sudo_flags
==========
Additional flags to pass to sudo when engaging sudo support. The default is '-H' which preserves the environment
of the original user. In some situations you may wish to add or remote flags, but in general most users
will not need to change this setting::
sudo_flags=-H
sudo_user
=========
This is the default user to sudo to if --sudo-user is not specified or 'sudo_user' is not specified in an Ansible
playbook. The default is the most logical: 'root'::
sudo_user=root
timeout
=======
This is the default SSH timeout to use on connection attempts::
timeout = 10
transport
=========
This is the default transport ot use if "-c <transport_name>" is not specified to /usr/bin/ansible or /usr/bin/ansible-playbook.
The default is 'smart', which will use 'ssh' (OpenSSH based) if the local operating system is new enough to support ControlPersist
technology, and then will otherwise use 'paramiko'. Other transport options include 'local', 'chroot', 'jail', and so on.
Users should usually leave this setting as 'smart' and let their playbooks choose an alternate setting when needed with the
'connection:' play parameter.
vars_plugins
============
This is a developer-centric feature that allows low-level extensions around Ansible to be loaded from
different locations::
action_plugins = /usr/share/ansible_plugins/action_plugins
Most users will not need to use this feature. See `developing_plugins` for more details
Paramiko Specific Settings
--------------------------
Paramiko is the default SSH connection implementation on Enterprise Linux 6 or earlier, and is not used by default on other
platforms. Settings live under the [paramiko] header.
record_host_keys
================
The default setting of yes will record newly discovered and approved (if host key checking is enabled) hosts in the user's hostfile.
This setting may be inefficient for large numbers of hosts, and in those situations, using the ssh transport is definitely recommended
instead. Setting it to False will improve performance and is recommended when host key checking is disabled::
record_host_keys=True
OpenSSH Specific Settings
-------------------------
Under the [ssh] header, the following settings are tunable for SSH connections. OpenSSH is the default connection type for Ansible
on OSes that are new enough to support ControlPersist. (This means basically all operating systems except Enterprise Linux 6 or earlier).
ssh_args
========
If set, this will pass a specific set of options to Ansible rather than Ansible's usual defaults::
ssh_args = -o ControlMaster=auto -o ControlPersist=60s
In particular, users may wish to raise the ControlPersist time to encourage performance. A value of 30 minutes may
be appropriate.
control_path
============
This is the location to save ControlPath sockets. This defaults to::
control_path=%(directory)s/ansible-ssh-%%h-%%p-%%r
On some systems with very long hostnames or very long path names (caused by long user names or
deeply nested home directories) this can exceed the character limit on
file socket names (108 characters for most platforms). In that case, you
may wish to shorten the string to something like the below::
control_path = %(directory)s/%%h-%%r
Ansible 1.4 and later will instruct users to run with "-vvvv" in situations where it hits this problem
and if so it is easy to tell there is too long of a Control Path filename. This may be frequently
encountered on EC2.
scp_if_ssh
==========
Occasionally users may be managing a remote system that doesn't have SFTP enabled. If set to True, we can
cause scp to be used to transfer remote files instead::
scp_if_ssh=False
There's really no reason to change this unless problems are encountered, and then there's also no real drawback
to managing the switch. Most environments support SFTP by default and this doesn't usually need to be changed.
Accelerate Mode Settings
------------------------
Under the [accelerate] header, the following settings are tunable for :doc:`playbooks_acceleration`
accelerate_port
===============
This is the port to use for accelerate mode::
accelerate_port = 5099