- re-arrange cluster initialization to a set of variables in a dictionary

which control three things:
    1. should we initialize?
    2. if so, how?
    3. what environment variables and user to use
  The approach taken is very similar to what the Apache formula uses, namely:
  a default dictionary which is over-ridden by:
     os-specific defaults,
     then os codename defaults,
     then os finger defaults,
     and finally user-specified pillar values
- this also adds support for grains['osfinger']

postgres/defaults.yaml

@@ -7,8 +7,6 @@ postgres:
   python: python-psycopg2
   service: postgresql
   conf_dir: /var/lib/pgsql/data
-  data_dir: /var/lib/pgsql/data
-  version: 9.1
   use_upstream_repo: False
   users: {}
   acls: []
@@ -17,10 +15,17 @@ postgres:
   postgresconf_backup: True
   postgresconf: ""
   pg_hba.conf: salt://postgres/pg_hba.conf
-  initdb: True
-  initdb_user: root
-  initdb_args: --data-checksum
-  commands:
-    initdb: service postgresql initdb
   user: postgres
   group: postgres
+  # if prepare_cluster is over-ridden in any of:
+  # - osmap.yaml
+  # - oscodenamemap.yaml
+  # - osfingermap.yaml
+  # you will have to specify a complete dictionary.
+  prepare_cluster:
+    user: root
+    command: service postgresql initdb
+    test: test -f /var/lib/pgsql/data/PG_VERSION
+    env:
+      LC_ALL: C.UTF-8
+