WIP on the re-implementation of fact caching and various backends.

lib/ansible/cache/__init__.py

@@ -0,0 +1,59 @@
+# (c) 2014, Michael DeHaan <michael.dehaan@gmail.com>
+#
+# This file is part of Ansible
+#
+# Ansible is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Ansible is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.
+
+from collections import MutableMapping
+
+from ansible import utils
+from ansible import constants as C
+from ansible import errors
+
+
+class FactCache(MutableMapping):
+
+    def __init__(self, *args, **kwargs):
+        self._plugin = utils.plugins.cache_loader.get(C.CACHE_PLUGIN)
+        if self._plugin is None:
+            return
+
+    def __getitem__(self, key):
+        if key not in self:
+            raise KeyError
+        return self._plugin.get(key)
+
+    def __setitem__(self, key, value):
+        self._plugin.set(key, value)
+
+    def __delitem__(self, key):
+        self._plugin.delete(key)
+
+    def __contains__(self, key):
+        return self._plugin.contains(key)
+
+    def __iter__(self):
+        return iter(self._plugin.keys())
+
+    def __len__(self):
+        return len(self._plugin.keys())
+
+    def copy(self):
+        """
+        Return a primitive copy of the keys and values from the cache.
+        """
+        return dict([(k, v) for (k, v) in self.iteritems()])
+
+    def keys(self):
+        return self._plugin.keys()

lib/ansible/cache/base.py

@@ -0,0 +1,15 @@
+class BaseCacheModule(object):
+    def get(self, key):
+        raise NotImplementedError("Subclasses of {} must implement the '{}' method".format(self.__class__.__name__, self.__name__))
+
+    def set(self, key, value):
+        raise NotImplementedError("Subclasses of {} must implement the '{}' method".format(self.__class__.__name__, self.__name__))
+
+    def keys(self):
+        raise NotImplementedError("Subclasses of {} must implement the '{}' method".format(self.__class__.__name__, self.__name__))
+
+    def contains(self, key):
+        raise NotImplementedError("Subclasses of {} must implement the '{}' method".format(self.__class__.__name__, self.__name__))
+
+    def delete(self, key):
+        raise NotImplementedError("Subclasses of {} must implement the '{}' method".format(self.__class__.__name__, self.__name__))

lib/ansible/cache/file.py

@@ -0,0 +1,65 @@
+# (c) 2014, Michael DeHaan <michael.dehaan@gmail.com>
+#
+# This file is part of Ansible
+#
+# Ansible is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Ansible is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.
+from __future__ import absolute_import
+
+import collections
+import json
+import os
+import shutil
+import sys
+import tempfile
+from datetime import datetime
+
+from ansible import constants as C
+from ansible.cache.memory import CacheModule as MemoryCacheModule
+
+
+class CacheModule(MemoryCacheModule):
+    def __init__(self):
+        super(CacheModule, self).__init__()
+        self._timeout = int(C.CACHE_PLUGIN_TIMEOUT)
+        self._filename = '/tmp/ansible_facts.json'
+
+        if os.access(self._filename, os.R_OK):
+            mtime = datetime.fromtimestamp(os.path.getmtime(self._filename))
+            if self._timeout == 0 or (datetime.now() - mtime).total_seconds() < self._timeout:
+                with open(self._filename, 'rb') as f:
+                    # we could make assumptions about the MemoryCacheModule here if we wanted
+                    # to be more efficient, but performance isn't the priority with this module
+                    data = json.load(f)
+                    if isinstance(data, collections.Mapping):
+                        for k, v in data.items():
+                            super(CacheModule, self).set(k, v)
+
+    def set(self, *args, **kwargs):
+        super(CacheModule, self).set(*args, **kwargs)
+        self.flush()
+
+    def delete(self, *args, **kwargs):
+        super(CacheModule, self).delete(*args, **kwargs)
+        self.flush()
+
+    def flush(self):
+        temp = tempfile.TemporaryFile('r+b')
+
+        try:
+            json.dump(self._cache, temp, separators=(',', ':'))
+            temp.seek(0)
+            with open(self._filename, 'w+b') as f:
+                shutil.copyfileobj(temp, f)
+        finally:
+            temp.close()

lib/ansible/cache/memcached.py

@@ -0,0 +1,112 @@
+# (c) 2014, Michael DeHaan <michael.dehaan@gmail.com>
+#
+# This file is part of Ansible
+#
+# Ansible is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Ansible is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.
+import collections
+import sys
+import time
+
+from ansible import constants as C
+from ansible.cache.base import BaseCacheModule
+
+try:
+    import memcache
+except ImportError:
+    print 'python-memcached is required for the memcached fact cache'
+    sys.exit(1)
+
+
+class CacheModuleKeys(collections.MutableSet):
+    """
+    A set subclass that keeps track of insertion time and persists
+    the set in memcached.
+    """
+    PREFIX = 'ansible_cache_keys'
+
+    def __init__(self, cache, *args, **kwargs):
+        self._cache = cache
+        self._keyset = dict(*args, **kwargs)
+
+    def __contains__(self, key):
+        return key in self._keyset
+
+    def __iter__(self):
+        return iter(self._keyset)
+
+    def __len__(self):
+        return len(self._keyset)
+
+    def add(self, key):
+        self._keyset[key] = time.time()
+        self._cache.set(self.PREFIX, self._keyset)
+
+    def discard(self, key):
+        del self._keyset[key]
+        self._cache.set(self.PREFIX, self._keyset)
+
+    def remove_by_timerange(self, s_min, s_max):
+        for k in self._keyset.keys():
+            t = self._keyset[k]
+            if s_min < t < s_max:
+                del self._keyset[k]
+        self._cache.set(self.PREFIX, self._keyset)
+
+
+class CacheModule(BaseCacheModule):
+
+    def __init__(self, *args, **kwargs):
+        if C.CACHE_PLUGIN_CONNECTION:
+            connection = C.CACHE_PLUGIN_CONNECTION.split(',')
+        else:
+            connection = ['127.0.0.1:11211']
+
+        self._timeout = C.CACHE_PLUGIN_TIMEOUT
+        self._prefix = C.CACHE_PLUGIN_PREFIX
+        self._cache = memcache.Client(connection, debug=0)
+        self._keys = CacheModuleKeys(self._cache, self._cache.get(CacheModuleKeys.PREFIX) or [])
+
+    def _make_key(self, key):
+        return "{}{}".format(self._prefix, key)
+
+    def _expire_keys(self):
+        if self._timeout > 0:
+            expiry_age = time.time() - self._timeout
+        self._keys.remove_by_timerange(0, expiry_age)
+
+    def get(self, key):
+        value = self._cache.get(self._make_key(key))
+        # guard against the key not being removed from the zset;
+        # this could happen in cases where the timeout value is changed
+        # between invocations
+        if value is None:
+            self.delete(key)
+            raise KeyError
+        return value
+
+    def set(self, key, value):
+        self._cache.set(self._make_key(key), value, time=self._timeout)
+        self._keys.add(key)
+
+    def keys(self):
+        self._expire_keys()
+        return list(iter(self._keys))
+
+    def contains(self, key):
+        self._expire_keys()
+        return key in self._keys
+
+    def delete(self, key):
+        self._cache.delete(self._make_key(key))
+        self._keys.discard(key)

lib/ansible/cache/memory.py

@@ -0,0 +1,37 @@
+# (c) 2014, Michael DeHaan <michael.dehaan@gmail.com>
+#
+# This file is part of Ansible
+#
+# Ansible is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Ansible is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.
+
+
+class CacheModule(object):
+
+    def __init__(self, *args, **kwargs):
+        self._cache = {}
+
+    def get(self, key):
+        return self._cache.get(key)
+
+    def set(self, key, value):
+        self._cache[key] = value
+
+    def keys(self):
+        return self._cache.keys()
+
+    def contains(self, key):
+        return key in self._cache
+
+    def delete(self, key):
+        del self._cache[key]

lib/ansible/cache/redis.py

@@ -0,0 +1,108 @@
+# (c) 2014, Michael DeHaan <michael.dehaan@gmail.com>
+#
+# This file is part of Ansible
+#
+# Ansible is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Ansible is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.
+from __future__ import absolute_import
+
+import collections
+import pickle
+import sys
+import time
+
+from ansible import constants as C
+from ansible.cache.base import BaseCacheModule
+
+try:
+    from redis import StrictRedis
+except ImportError:
+    print "The 'redis' Python module is required for the redis fact cache"
+    sys.exit(1)
+
+
+class PickledRedis(StrictRedis):
+    """
+    A subclass of StricRedis that uses the pickle module to store and load
+    representations of the provided values.
+    """
+    def get(self, name):
+        pickled_value = super(PickledRedis, self).get(name)
+        if pickled_value is None:
+            return None
+        return pickle.loads(pickled_value)
+
+    def set(self, name, value, *args, **kwargs):
+        return super(PickledRedis, self).set(name, pickle.dumps(value), *args, **kwargs)
+
+    def setex(self, name, time, value):
+        return super(PickledRedis, self).setex(name, time, pickle.dumps(value))
+
+
+class CacheModule(BaseCacheModule):
+    """
+    A caching module backed by redis.
+
+    Keys are maintained in a zset with their score being the timestamp
+    when they are inserted. This allows for the usage of 'zremrangebyscore'
+    to expire keys. This mechanism is used or a pattern matched 'scan' for
+    performance.
+    """
+    def __init__(self, *args, **kwargs):
+        if C.CACHE_PLUGIN_CONNECTION:
+            connection = C.CACHE_PLUGIN_CONNECTION.split(':')
+        else:
+            connection = []
+
+        self._timeout = C.CACHE_PLUGIN_TIMEOUT
+        self._prefix = C.CACHE_PLUGIN_PREFIX
+        self._cache = PickledRedis(*connection)
+        self._keys_set = 'ansible_cache_keys'
+
+    def _make_key(self, key):
+        return "{}{}".format(self._prefix, key)
+
+    def get(self, key):
+        value = self._cache.get(self._make_key(key))
+        # guard against the key not being removed from the zset;
+        # this could happen in cases where the timeout value is changed
+        # between invocations
+        if value is None:
+            self.delete(key)
+            raise KeyError
+        return value
+
+    def set(self, key, value):
+        if self._timeout > 0: # a timeout of 0 is handled as meaning 'never expire'
+            self._cache.setex(self._make_key(key), self._timeout, value)
+        else:
+            self._cache.set(self._make_key(key), value)
+
+        self._cache.zadd(self._keys_set, time.time(), key)
+
+    def _expire_keys(self):
+        if self._timeout > 0:
+            expiry_age = time.time() - self._timeout
+        self._cache.zremrangebyscore(self._keys_set, 0, expiry_age)
+
+    def keys(self):
+        self._expire_keys()
+        return self._cache.zrange(self._keys_set, 0, -1)
+
+    def contains(self, key):
+        self._expire_keys()
+        return (self._cache.zrank(self._keys_set, key) >= 0)
+
+    def delete(self, key):
+        self._cache.delete(self._make_key(key))
+        self._cache.zrem(self._keys_set, key)

lib/ansible/constants.py

@@ -140,6 +140,7 @@ DEFAULT_ASK_SU_PASS       = get_config(p, DEFAULTS, 'ask_su_pass', 'ANSIBLE_ASK_
 DEFAULT_GATHERING         = get_config(p, DEFAULTS, 'gathering', 'ANSIBLE_GATHERING', 'implicit').lower()
 
 DEFAULT_ACTION_PLUGIN_PATH     = get_config(p, DEFAULTS, 'action_plugins',     'ANSIBLE_ACTION_PLUGINS', '/usr/share/ansible_plugins/action_plugins')
+DEFAULT_CACHE_PLUGIN_PATH      = get_config(p, DEFAULTS, 'cache_plugins',      'ANSIBLE_CACHE_PLUGINS', '/usr/share/ansible_plugins/cache_plugins')
 DEFAULT_CALLBACK_PLUGIN_PATH   = get_config(p, DEFAULTS, 'callback_plugins',   'ANSIBLE_CALLBACK_PLUGINS', '/usr/share/ansible_plugins/callback_plugins')
 DEFAULT_CONNECTION_PLUGIN_PATH = get_config(p, DEFAULTS, 'connection_plugins', 'ANSIBLE_CONNECTION_PLUGINS', '/usr/share/ansible_plugins/connection_plugins')
 DEFAULT_LOOKUP_PLUGIN_PATH     = get_config(p, DEFAULTS, 'lookup_plugins',     'ANSIBLE_LOOKUP_PLUGINS', '/usr/share/ansible_plugins/lookup_plugins')
@@ -147,6 +148,11 @@ DEFAULT_VARS_PLUGIN_PATH       = get_config(p, DEFAULTS, 'vars_plugins',       '
 DEFAULT_FILTER_PLUGIN_PATH     = get_config(p, DEFAULTS, 'filter_plugins',     'ANSIBLE_FILTER_PLUGINS', '/usr/share/ansible_plugins/filter_plugins')
 DEFAULT_LOG_PATH               = shell_expand_path(get_config(p, DEFAULTS, 'log_path',           'ANSIBLE_LOG_PATH', ''))
 
+CACHE_PLUGIN                   = get_config(p, DEFAULTS, 'cache_plugin', 'ANSIBLE_CACHE_PLUGIN', 'memory')
+CACHE_PLUGIN_CONNECTION        = get_config(p, DEFAULTS, 'cache_plugin_connection', 'ANSIBLE_CACHE_PLUGIN_CONNECTION', None)
+CACHE_PLUGIN_PREFIX            = get_config(p, DEFAULTS, 'cache_plugin_prefix', 'ANSIBLE_CACHE_PLUGIN_PREFIX', 'ansible_facts')
+CACHE_PLUGIN_TIMEOUT           = get_config(p, DEFAULTS, 'cache_plugin_timeout', 'ANSIBLE_CACHE_PLUGIN_TIMEOUT', (60 * 60 * 24), integer=True)
+
 ANSIBLE_FORCE_COLOR            = get_config(p, DEFAULTS, 'force_color', 'ANSIBLE_FORCE_COLOR', None, boolean=True)
 ANSIBLE_NOCOLOR                = get_config(p, DEFAULTS, 'nocolor', 'ANSIBLE_NOCOLOR', None, boolean=True)
 ANSIBLE_NOCOWS                 = get_config(p, DEFAULTS, 'nocows', 'ANSIBLE_NOCOWS', None, boolean=True)

lib/ansible/playbook/__init__.py

@@ -22,6 +22,7 @@ from ansible.utils.template import template
 from ansible import utils
 from ansible import errors
 import ansible.callbacks
+import ansible.cache
 import os
 import shlex
 import collections
@@ -32,9 +33,10 @@ import pipes
 # the setup cache stores all variables about a host
 # gathered during the setup step, while the vars cache
 # holds all other variables about a host
-SETUP_CACHE = collections.defaultdict(dict)
+SETUP_CACHE = ansible.cache.FactCache()
 VARS_CACHE  = collections.defaultdict(dict)
 
+
 class PlayBook(object):
     '''
     runs an ansible playbook, given as a datastructure or YAML filename.
@@ -98,7 +100,7 @@ class PlayBook(object):
         inventory:        can be specified instead of host_list to use a pre-existing inventory object
         check:            don't change anything, just try to detect some potential changes
         any_errors_fatal: terminate the entire execution immediately when one of the hosts has failed
-        force_handlers:   continue to notify and run handlers even if a task fails 
+        force_handlers:   continue to notify and run handlers even if a task fails
         """
 
         self.SETUP_CACHE = SETUP_CACHE
@@ -187,7 +189,7 @@ class PlayBook(object):
 
     def _get_playbook_vars(self, play_ds, existing_vars):
         '''
-        Gets the vars specified with the play and blends them 
+        Gets the vars specified with the play and blends them
         with any existing vars that have already been read in
         '''
         new_vars = existing_vars.copy()
@@ -470,6 +472,13 @@ class PlayBook(object):
         contacted = results.get('contacted', {})
         self.stats.compute(results, ignore_errors=task.ignore_errors)
 
+        def _register_play_vars(host, result):
+            # when 'register' is used, persist the result in the vars cache
+            # rather than the setup cache - vars should be transient between playbook executions
+            if 'stdout' in result and 'stdout_lines' not in result:
+                result['stdout_lines'] = result['stdout'].splitlines()
+            utils.update_hash(self.VARS_CACHE, host, {task.register: result})
+
         # add facts to the global setup cache
         for host, result in contacted.iteritems():
             if 'results' in result:
@@ -478,22 +487,19 @@ class PlayBook(object):
                 for res in result['results']:
                     if type(res) == dict:
                         facts = res.get('ansible_facts', {})
-                        self.SETUP_CACHE[host].update(facts)
+                        utils.update_hash(self.SETUP_CACHE, host, facts)
             else:
+                # when facts are returned, persist them in the setup cache
                 facts = result.get('ansible_facts', {})
-                self.SETUP_CACHE[host].update(facts)
+                utils.update_hash(self.SETUP_CACHE, host, facts)
             if task.register:
-                if 'stdout' in result and 'stdout_lines' not in result:
+                _register_play_vars(host, result)
-                    result['stdout_lines'] = result['stdout'].splitlines()
-                self.SETUP_CACHE[host][task.register] = result
 
         # also have to register some failed, but ignored, tasks
         if task.ignore_errors and task.register:
             failed = results.get('failed', {})
             for host, result in failed.iteritems():
-                if 'stdout' in result and 'stdout_lines' not in result:
+                _register_play_vars(host, result)
-                    result['stdout_lines'] = result['stdout'].splitlines()
-                self.SETUP_CACHE[host][task.register] = result
 
         # flag which notify handlers need to be run
         if len(task.notify) > 0:
@@ -585,8 +591,8 @@ class PlayBook(object):
         # let runner template out future commands
         setup_ok = setup_results.get('contacted', {})
         for (host, result) in setup_ok.iteritems():
-            self.SETUP_CACHE[host].update({'module_setup': True})
+            utils.update_hash(self.SETUP_CACHE, host, {'module_setup': True})
-            self.SETUP_CACHE[host].update(result.get('ansible_facts', {}))
+            utils.update_hash(self.SETUP_CACHE, host, result.get('ansible_facts', {}))
         return setup_results
 
     # *****************************************************
@@ -620,7 +626,7 @@ class PlayBook(object):
 
     def _run_play(self, play):
         ''' run a list of tasks for a given pattern, in order '''
-        
+
         self.callbacks.on_play_start(play.name)
         # Get the hosts for this play
         play._play_hosts = self.inventory.list_hosts(play.hosts)

lib/ansible/playbook/play.py

@@ -29,6 +29,7 @@ import os
 import sys
 import uuid
 
+
 class Play(object):
 
     __slots__ = [
@@ -85,7 +86,7 @@ class Play(object):
         # now we load the roles into the datastructure
         self.included_roles = []
         ds = self._load_roles(self.roles, ds)
-        
+
         # and finally re-process the vars files as they may have
         # been updated by the included roles
         self.vars_files = ds.get('vars_files', [])
@@ -153,6 +154,7 @@ class Play(object):
         self._tasks      = self._load_tasks(self._ds.get('tasks', []), load_vars)
         self._handlers   = self._load_tasks(self._ds.get('handlers', []), load_vars)
 
+
         # apply any missing tags to role tasks
         self._late_merge_role_tags()
 
@@ -167,7 +169,7 @@ class Play(object):
     def _get_role_path(self, role):
         """
         Returns the path on disk to the directory containing
-        the role directories like tasks, templates, etc. Also 
+        the role directories like tasks, templates, etc. Also
         returns any variables that were included with the role
         """
         orig_path = template(self.basedir,role,self.vars)
@@ -242,7 +244,7 @@ class Play(object):
                                 allow_dupes = utils.boolean(meta_data.get('allow_duplicates',''))
 
                         # if any tags were specified as role/dep variables, merge
-                        # them into the current dep_vars so they're passed on to any 
+                        # them into the current dep_vars so they're passed on to any
                         # further dependencies too, and so we only have one place
                         # (dep_vars) to look for tags going forward
                         def __merge_tags(var_obj):
@@ -318,7 +320,7 @@ class Play(object):
                         dep_stack.append([dep,dep_path,dep_vars,dep_defaults_data])
 
             # only add the current role when we're at the top level,
-            # otherwise we'll end up in a recursive loop 
+            # otherwise we'll end up in a recursive loop
             if level == 0:
                 self.included_roles.append(role)
                 dep_stack.append([role,role_path,role_vars,defaults_data])
@@ -505,7 +507,7 @@ class Play(object):
             if not isinstance(x, dict):
                 raise errors.AnsibleError("expecting dict; got: %s, error in %s" % (x, original_file))
 
-            # evaluate sudo vars for current and child tasks 
+            # evaluate sudo vars for current and child tasks
             included_sudo_vars = {}
             for k in ["sudo", "sudo_user"]:
                 if k in x:
@@ -554,7 +556,7 @@ class Play(object):
                 else:
                     default_vars = utils.combine_vars(self.default_vars, default_vars)
 
-                # append the vars defined with the include (from above) 
+                # append the vars defined with the include (from above)
                 # as well as the old-style 'vars' element. The old-style
                 # vars are given higher precedence here (just in case)
                 task_vars = utils.combine_vars(task_vars, include_vars)
@@ -610,8 +612,8 @@ class Play(object):
 
     def _is_valid_tag(self, tag_list):
         """
-        Check to see if the list of tags passed in is in the list of tags 
+        Check to see if the list of tags passed in is in the list of tags
-        we only want (playbook.only_tags), or if it is not in the list of 
+        we only want (playbook.only_tags), or if it is not in the list of
         tags we don't want (playbook.skip_tags).
         """
         matched_skip_tags = set(tag_list) & set(self.playbook.skip_tags)
@@ -774,7 +776,7 @@ class Play(object):
                 inject.update(self.vars)
                 filename4 = template(self.basedir, filename3, inject)
                 filename4 = utils.path_dwim(self.basedir, filename4)
-            else:    
+            else:
                 filename4 = utils.path_dwim(self.basedir, filename3)
             return filename2, filename3, filename4
 
@@ -823,7 +825,7 @@ class Play(object):
             inject.update(self.playbook.SETUP_CACHE.get(host, {}))
             inject.update(self.playbook.VARS_CACHE.get(host, {}))
         else:
-            inject = None            
+            inject = None
 
         for filename in self.vars_files:
             if type(filename) == list:
@@ -854,4 +856,4 @@ class Play(object):
 
         # finally, update the VARS_CACHE for the host, if it is set
         if host is not None:
-            self.playbook.VARS_CACHE[host].update(self.playbook.extra_vars)
+            self.playbook.VARS_CACHE.setdefault(host, {}).update(self.playbook.extra_vars)

lib/ansible/utils/plugins.py

@@ -108,14 +108,14 @@ class PluginLoader(object):
                 if fullpath not in ret:
                     ret.append(fullpath)
 
-        # look in any configured plugin paths, allow one level deep for subcategories 
+        # look in any configured plugin paths, allow one level deep for subcategories
         configured_paths = self.config.split(os.pathsep)
         for path in configured_paths:
             path = os.path.realpath(os.path.expanduser(path))
             contents = glob.glob("%s/*" % path)
             for c in contents:
                 if os.path.isdir(c) and c not in ret:
-                    ret.append(c)       
+                    ret.append(c)
             if path not in ret:
                 ret.append(path)
 
@@ -181,7 +181,7 @@ class PluginLoader(object):
         return getattr(self._module_cache[path], self.class_name)(*args, **kwargs)
 
     def all(self, *args, **kwargs):
-        ''' instantiates all plugins with the same arguments '''       
+        ''' instantiates all plugins with the same arguments '''
 
         for i in self._get_paths():
             matches = glob.glob(os.path.join(i, "*.py"))
@@ -195,24 +195,31 @@ class PluginLoader(object):
                 yield getattr(self._module_cache[path], self.class_name)(*args, **kwargs)
 
 action_loader = PluginLoader(
-    'ActionModule',   
+    'ActionModule',
     'ansible.runner.action_plugins',
     C.DEFAULT_ACTION_PLUGIN_PATH,
     'action_plugins'
 )
 
+cache_loader = PluginLoader(
+    'CacheModule',
+    'ansible.cache',
+    C.DEFAULT_CACHE_PLUGIN_PATH,
+    'cache_plugins'
+)
+
 callback_loader = PluginLoader(
-    'CallbackModule', 
+    'CallbackModule',
-    'ansible.callback_plugins', 
+    'ansible.callback_plugins',
-    C.DEFAULT_CALLBACK_PLUGIN_PATH, 
+    C.DEFAULT_CALLBACK_PLUGIN_PATH,
     'callback_plugins'
 )
 
 connection_loader = PluginLoader(
-    'Connection', 
+    'Connection',
-    'ansible.runner.connection_plugins', 
+    'ansible.runner.connection_plugins',
-    C.DEFAULT_CONNECTION_PLUGIN_PATH, 
+    C.DEFAULT_CONNECTION_PLUGIN_PATH,
-    'connection_plugins', 
+    'connection_plugins',
     aliases={'paramiko': 'paramiko_ssh'}
 )
 
@@ -224,30 +231,30 @@ shell_loader = PluginLoader(
 )
 
 module_finder = PluginLoader(
-    '', 
+    '',
-    '', 
+    '',
-    C.DEFAULT_MODULE_PATH, 
+    C.DEFAULT_MODULE_PATH,
     'library'
 )
 
 lookup_loader = PluginLoader(
-    'LookupModule',   
+    'LookupModule',
-    'ansible.runner.lookup_plugins', 
+    'ansible.runner.lookup_plugins',
-    C.DEFAULT_LOOKUP_PLUGIN_PATH, 
+    C.DEFAULT_LOOKUP_PLUGIN_PATH,
     'lookup_plugins'
 )
 
 vars_loader = PluginLoader(
-    'VarsModule', 
+    'VarsModule',
-    'ansible.inventory.vars_plugins', 
+    'ansible.inventory.vars_plugins',
-    C.DEFAULT_VARS_PLUGIN_PATH, 
+    C.DEFAULT_VARS_PLUGIN_PATH,
     'vars_plugins'
 )
 
 filter_loader = PluginLoader(
-    'FilterModule', 
+    'FilterModule',
-    'ansible.runner.filter_plugins', 
+    'ansible.runner.filter_plugins',
-    C.DEFAULT_FILTER_PLUGIN_PATH, 
+    C.DEFAULT_FILTER_PLUGIN_PATH,
     'filter_plugins'
 )
 

setup.py

@@ -38,6 +38,7 @@ setup(name='ansible',
       package_dir={ 'ansible': 'lib/ansible' },
       packages=[
          'ansible',
+         'ansible.cache',
          'ansible.utils',
          'ansible.utils.module_docs_fragments',
          'ansible.inventory',