Add documentation on workflows

docs/source/data_model.rst

@@ -55,6 +55,8 @@ playbooks like purging cluster or removing OSD hosts.
 
 
 
+.. _data-model-server:
+
 Cluster
 +++++++
 
@@ -96,6 +98,8 @@ ignoring `ohai <https://docs.chef.io/ohai.html>`_ and `facter
     discovered by discovery protocol.
 
 
+.. _data-model-server-discovery:
+
 Server Discovery
 ----------------
 
@@ -136,6 +140,8 @@ Server discovery process is done in 6 steps:
    be performed.
 
 
+.. _data-model-playbook-configuration:
+
 Playbook Configuration
 ++++++++++++++++++++++
 
@@ -200,6 +206,9 @@ of `Server`_ models. After that you can modify it as you want.
     machines.
 
 
+
+.. _data-model-execution:
+
 Execution
 +++++++++
 

docs/source/index.rst

@@ -33,6 +33,7 @@ Contents
    installation/index.rst
    deploy/index.rst
    data_model
+   workflows
    cli
    ceph_node
    api/index.rst

docs/source/workflows.rst

@@ -0,0 +1,41 @@
+Workflows
+=========
+
+This chapter describes different workflows for Decapod. These workflows
+should help user to understand how to work with Decapod and perform
+required actions.
+
+
+
+Server discovery
+++++++++++++++++
+
+As mentioned in :ref:`data-model-server-discovery`, it has to be done in
+several steps.
+
+1. Choose OS image with cloud-init support
+2. Generate user-data config with :program:`decapod cloud-config` CLI
+3. Deploy OS with such config
+
+After that, if everything is correct, eventually server will be listed
+in Decapod. Server discovery takes time because of cloud-init execution
+therefore it might be listed in 5 minutes after deployment. If server
+appears in Decapod, it means that tool can use it.
+
+
+
+Cluster Deployment
+++++++++++++++++++
+
+As mentioned in :doc:`data_model` several models are required.
+
+1. First, we need to create empty :ref:`data-model-server` model.
+   This model is a holder for configuration of cluster, also it defines
+   Ceph FSID and the name.
+2. We need to create :ref:`data-model-playbook-configuration` model for
+   ``deploy_cluster`` playbook. This and playbook will allow user to
+   deploy cluster. Also, keep in mind that cluster deployment is
+   idempotent operation and you may execute it several times.
+3. Execute that playbook configuration, creating new
+   :ref:`data-model-execution`. If something went wrong, examine
+   execution steps or log.