README.Ansible

=============================================================
      reclass — recursive external node classification
=============================================================
reclass is © 2007–2013 martin f. krafft <madduck@madduck.net>
and available under the terms of the Artistic Licence 2.0
'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''

Please make sure to read the generic information in the README file first, or
alongside this document.

Quick start with Ansible
~~~~~~~~~~~~~~~~~~~~~~~~
The following steps should get you up and running quickly. Generally, we will
be working in /etc/ansible. However, if you are using a source-code checkout
of Ansible, you might also want to work inside the ./hacking directory
instead.

Or you can also just look into ./examples/ansible of your reclass checkout,
where the following steps have already been prepared.

/…/reclass refers to the location of your reclass checkout.

  0. Run 'make' in the root of the reclass checkout (see the section
     'Installation' in the README file for the reason).

  1. Symlink /…/reclass/adapters/ansible to /etc/ansible/hosts (or
     ./hacking/hosts)

  2. Copy the two directories 'nodes' and 'classes' from the example
     subdirectory in the reclass checkout to /etc/ansible

     If you prefer to put those directories elsewhere, you can create
     /etc/ansible/reclass-config.yml with contents such as

       storage_type: yaml_fs
       nodes_uri: /srv/reclass/nodes
       classes_uri: /srv/reclass/classes

     Note that yaml_fs is currently the only supported storage_type, and it's
     the default if you don't set it.

  3. Check out your inventory by invoking

       ./hosts --list

     which should return 5 groups in JSON-format, and each group has exactly
     one member 'localhost'.

  4. See the node information for 'localhost':

       ./hosts --host localhost

     This should print a set of keys and values, including a greeting,
     a colour, and a sub-class called '__reclas__'.

  5. Execute some ansible commands, e.g.

       ansible -i hosts \* --list-hosts
       ansible -i hosts \* -m ping
       ansible -i hosts \* -m debug -a 'msg="${greeting}"'
       ansible -i hosts \* -m setup
       ansible-playbook -i hosts test.yml

  6. You can also invoke reclass directly, which gives a slightly different
     view onto the same data, i.e. before it has been adapted for Ansible:

       /…/reclass.py --pretty-print --inventory
       /…/reclass.py --pretty-print --nodeinfo localhost

Integration with Ansible
~~~~~~~~~~~~~~~~~~~~~~~~
The integration between reclass and Ansible is performed through an adapter,
and needs not be of our concern too much.

However, Ansible has no concept of "nodes", "applications", "parameters", and
"classes". Therefore it is necessary to explain how those correspond to
Ansible. Crudely, the following mapping exists:

  nodes         hosts
  classes       groups
  applications  playbooks
  parameters    host_vars

reclass does not provide any group_vars because of its node-centric
perspective. While class definitions include parameters, those are inherited
by the node definitions and hence become node_vars.

reclass also does not provide playbooks, nor does it deal with any of the
related Ansible concepts, i.e. vars_files, vars, tasks, handlers, roles, etc..

  Let it be said at this point that you'll probably want to stop using
  host_vars, group_vars and vars_files altogether, and if only because you
  should no longer need them, but also because the variable precedence rules
  of Ansible are full of surprises, at least to me.

reclass' Ansible adapter massage the reclass output into Ansible-usable data,
namely:

  - Every class in the ancestry of a node becomes a group to Ansible. This is
    mainly useful to be able to target nodes during interactive use of
    Ansible, e.g.

      ansible debiannode@wheezy -m command -a 'apt-get upgrade'
        → upgrade all Debian nodes running wheezy

      ansible ssh.server -m command -a 'invoke-rc.d ssh restart'
        → restart all SSH server processes

      ansible mailserver -m command -a 'tail -n1000 /var/log/mail.err'
        → obtain the last 1,000 lines of all mailserver error log files

    The attentive reader might stumble over the use of singular words, whereas
    it might make more sense to address all 'mailserver*s*' with this tool.
    This is convention and up to you. I prefer to think of my node as
    a (singular) mailserver when I add 'mailserver' to its parent classes.

  - Every entry in the list of a host's applications might well correspond to
    an Ansible playbook. Therefore, reclass creates a (Ansible-)group for
    every application, and adds '_hosts' to the name. This postfix can be
    configured with a CLI option (--applications-postfix) or in the
    configuration file (applications_postfix).

    For instance, the ssh.server class adds the ssh.server application to
    a node's application list. Now the admin might create an Ansible playbook
    like so:

      - name: SSH server management
        hosts: ssh.server_hosts              ← SEE HERE
        tasks:
          - name: install SSH package
            action: …
        …

    There's a bit of redundancy in this, but unfortunately Ansible playbooks
    hardcode the nodes to which a playbook applies.

    It's now trivial to apply this playbook across your infrastructure:

      ansible-playbook ssh.server.yml

    My suggested way to use Ansible site-wide is then to create a 'site'
    playbook that includes all the other playbooks (which shall hopefully be
    based on Ansible roles), and then to invoke Ansible like this:

      ansible-playbook site.yml

    or, if you prefer only to reconfigure a subset of nodes, e.g. all
    webservers:

      ansible-playbook site.yml --limit webserver

    Again, if the singular word 'webserver' puts you off, change the
    convention as you wish.

    And if anyone comes up with a way to directly connect groups in the
    inventory with roles, thereby making it unnecessary to write playbook
    files (containing redundant information), please tell me!

  - Parameters corresponding to a node become host_vars for that host.

It is possible to include Jinja2-style variables like you would in Ansible,
in parameter values. This is especially powerful in combination with the
recursive merging, e.g.

  parameters:
    motd:
      greeting: Welcome to {{ ansible_fqdn }}!
      closing: This system is part of {{ realm }}

Now you just need to specify realm somewhere. The reference can reside in
a parent class, while the variable is defined e.g. in the node.