salt.renderers.py¶

Pure python state renderer¶

To use this renderer, the SLS file should contain a function called run which returns highstate data.

The highstate data is a dictionary containing identifiers as keys, and execution dictionaries as values. For example the following state declaration in YAML:

common_packages:
  pkg.installed:
   - pkgs:
      - curl
      - vim

translates to:

{'common_packages': {'pkg.installed': [{'pkgs': ['curl', 'vim']}]}}

In this module, a few objects are defined for you, giving access to Salt's execution functions, grains, pillar, etc. They are:

• __salt__ - Execution functions (i.e. __salt__['test.echo']('foo'))

• __grains__ - Grains (i.e. __grains__['os'])

• __pillar__ - Pillar data (i.e. __pillar__['foo'])

• __opts__ - Minion configuration options

• __env__ - The effective salt fileserver environment (i.e. base). Also referred to as a "saltenv". __env__ should not be modified in a pure python SLS file. To use a different environment, the environment should be set when executing the state. This can be done in a couple different ways:

  • Using the saltenv argument on the salt CLI (i.e. salt '*' state.sls foo.bar.baz saltenv=env_name).

  • By adding a saltenv argument to an individual state within the SLS file. In other words, adding a line like this to the state's data structure: {'saltenv': 'env_name'}

• __sls__ - The SLS path of the file. For example, if the root of the base environment is /srv/salt, and the SLS file is /srv/salt/foo/bar/baz.sls, then __sls__ in that file will be foo.bar.baz.

When used in a scenario where additional user-provided context data is supplied (such as with file.managed), the additional data will typically be injected into the script as one or more global variables:

/etc/http/conf/http.conf:
  file.managed:
    - source: salt://apache/generate_http_conf.py
    - template: py
    - context:
        # Will be injected as the global variable "site_name".
        site_name: {{ site_name }}

When writing a reactor SLS file the global context data (same as context {{ data }} for states written with Jinja + YAML) is available. The following YAML + Jinja state declaration:

{% if data['id'] == 'mysql1' %}
highstate_run:
  local.state.apply:
    - tgt: mysql1
{% endif %}

translates to:

if data['id'] == 'mysql1':
    return {'highstate_run': {'local.state.apply': [{'tgt': 'mysql1'}]}}

 1 #!py
 2
 3 def run():
 4     config = {}
 5
 6     if __grains__['os'] == 'Ubuntu':
 7         user = 'ubuntu'
 8         group = 'ubuntu'
 9         home = '/home/{0}'.format(user)
10     else:
11         user = 'root'
12         group = 'root'
13         home = '/root/'
14
15     config['s3cmd'] = {
16         'pkg': [
17             'installed',
18             {'name': 's3cmd'},
19         ],
20     }
21
22     config[home + '/.s3cfg'] = {
23         'file.managed': [
24             {'source': 'salt://s3cfg/templates/s3cfg'},
25             {'template': 'jinja'},
26             {'user': user},
27             {'group': group},
28             {'mode': 600},
29             {'context': {
30                 'aws_key': __pillar__['AWS_ACCESS_KEY_ID'],
31                 'aws_secret_key': __pillar__['AWS_SECRET_ACCESS_KEY'],
32                 },
33             },
34         ],
35     }
36
37     return config