Replace shade and os-client-config with openstacksdk.

os-client-config is now just a wrapper around openstacksdk. The shade
code has been imported into openstacksdk. To reduce complexity, just use
openstacksdk directly.

openstacksdk's TaskManager has had to grow some features to deal with
SwiftService. Making nodepool's TaskManager a subclass of openstacksdk's
TaskManager ensures that we get the thread pool set up properly.

Change-Id: I3a01eb18ae31cc3b61509984f3817378db832b47

nodepool/config.py

@@ -230,7 +230,7 @@ def openConfig(path):
 def loadConfig(config_path):
     config = openConfig(config_path)
 
-    # Call driver config reset now to clean global hooks like os_client_config
+    # Call driver config reset now to clean global hooks like openstacksdk
     for driver in Drivers.drivers.values():
         driver.reset()
 

nodepool/driver/fake/__init__.py

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import os_client_config
+from openstack.config import loader
 
 from nodepool.driver import Driver
 from nodepool.driver.fake.config import FakeProviderConfig
@@ -25,7 +25,7 @@ class FakeDriver(Driver):
         self.reset()
 
     def reset(self):
-        self.os_client_config = os_client_config.OpenStackConfig()
+        self.openstack_config = loader.OpenStackConfig()
 
     def getProviderConfig(self, provider):
         return FakeProviderConfig(self, provider)

nodepool/driver/fake/provider.py

@@ -19,7 +19,7 @@ import threading
 import time
 import uuid
 
-import shade
+import openstack
 
 from nodepool import exceptions
 from nodepool.driver.openstack.provider import OpenStackProvider
@@ -39,11 +39,11 @@ class Dummy(object):
             setattr(self, k, v)
         try:
             if self.should_fail:
-                raise shade.OpenStackCloudException('This image has '
+                raise openstack.exceptions.OpenStackCloudException(
-                                                    'SHOULD_FAIL set to True.')
+                    'This image has SHOULD_FAIL set to True.')
             if self.over_quota:
-                raise shade.exc.OpenStackCloudHTTPError(
+                raise openstack.exceptions.HttpException(
-                    'Quota exceeded for something', 403)
+                    message='Quota exceeded for something', http_status=403)
         except AttributeError:
             pass
 

nodepool/driver/openstack/__init__.py

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import os_client_config
+from openstack.config import loader
 
 from nodepool.driver import Driver
 from nodepool.driver.openstack.config import OpenStackProviderConfig
@@ -25,7 +25,7 @@ class OpenStackDriver(Driver):
         self.reset()
 
     def reset(self):
-        self.os_client_config = os_client_config.OpenStackConfig()
+        self.openstack_config = loader.OpenStackConfig()
 
     def getProviderConfig(self, provider):
         return OpenStackProviderConfig(self, provider)

nodepool/driver/openstack/config.py

@@ -195,8 +195,8 @@ class OpenStackProviderConfig(ProviderConfig):
 
     def load(self, config):
         cloud_kwargs = self._cloudKwargs()
-        occ = self.driver_object.os_client_config
+        openstack_config = self.driver_object.openstack_config
-        self.cloud_config = occ.get_one_cloud(**cloud_kwargs)
+        self.cloud_config = openstack_config.get_one(**cloud_kwargs)
 
         self.image_type = self.cloud_config.config['image_format']
         self.region_name = self.provider.get('region-name')

nodepool/driver/openstack/provider.py

@@ -19,7 +19,7 @@ import logging
 import operator
 import time
 
-import shade
+import openstack
 
 from nodepool import exceptions
 from nodepool.driver import Provider
@@ -43,7 +43,7 @@ class OpenStackProvider(Provider):
         self.provider = provider
         self._images = {}
         self._networks = {}
-        self.__flavors = {}
+        self.__flavors = {}  # TODO(gtema): caching
         self.__azs = None
         self._use_taskmanager = use_taskmanager
         self._taskmanager = None
@@ -51,7 +51,7 @@ class OpenStackProvider(Provider):
 
     def start(self, zk_conn):
         if self._use_taskmanager:
-            self._taskmanager = TaskManager(None, self.provider.name,
+            self._taskmanager = TaskManager(self.provider.name,
                                             self.provider.rate)
             self._taskmanager.start()
         self.resetClient()
@@ -67,6 +67,7 @@ class OpenStackProvider(Provider):
     def getRequestHandler(self, poolworker, request):
         return handler.OpenStackNodeRequestHandler(poolworker, request)
 
+    # TODO(gtema): caching
     @property
     def _flavors(self):
         if not self.__flavors:
@@ -78,12 +79,12 @@ class OpenStackProvider(Provider):
             manager = self._taskmanager
         else:
             manager = None
-        return shade.OpenStackCloud(
+        return openstack.connection.Connection(
-            cloud_config=self.provider.cloud_config,
+            config=self.provider.cloud_config,
-            manager=manager,
+            task_manager=manager,
             app_name='nodepool',
-            app_version=version.version_info.version_string(),
+            app_version=version.version_info.version_string()
-            **self.provider.cloud_config.config)
+        )
 
     def quotaNeededByNodeType(self, ntype, pool):
         provider_label = pool.labels[ntype]
@@ -196,19 +197,15 @@ class OpenStackProvider(Provider):
 
     def resetClient(self):
         self._client = self._getClient()
-        if self._use_taskmanager:
-            self._taskmanager.setClient(self._client)
 
     def _getFlavors(self):
         flavors = self.listFlavors()
         flavors.sort(key=operator.itemgetter('ram'))
         return flavors
 
-    # TODO(mordred): These next three methods duplicate logic that is in
+    # TODO(gtema): These next three methods duplicate logic that is in
-    #                shade, but we can't defer to shade until we're happy
+    #              openstacksdk, caching is not enabled there by default
-    #                with using shade's resource caching facility. We have
+    #              Remove it when caching is default
-    #                not yet proven that to our satisfaction, but if/when
-    #                we do, these should be able to go away.
     def _findFlavorByName(self, flavor_name):
         for f in self._flavors:
             if flavor_name in (f['name'], f['id']):
@@ -226,6 +223,16 @@ class OpenStackProvider(Provider):
         # Note: this will throw an error if the provider is offline
         # but all the callers are in threads (they call in via CreateServer) so
         # the mainloop won't be affected.
+        # TODO(gtema): enable commented block when openstacksdk has caching
+        # enabled by default
+        # if min_ram:
+        #     return self._client.get_flavor_by_ram(
+        #         ram=min_ram,
+        #         include=flavor_name,
+        #         get_extra=False)
+        # else:
+        #     return self._client.get_flavor(flavor_name, get_extra=False)
+
         if min_ram:
             return self._findFlavorByRam(min_ram, flavor_name)
         else:
@@ -314,14 +321,14 @@ class OpenStackProvider(Provider):
 
         try:
             return self._client.create_server(wait=False, **create_args)
-        except shade.OpenStackCloudBadRequest:
+        except openstack.exceptions.BadRequestException:
             # We've gotten a 400 error from nova - which means the request
             # was malformed. The most likely cause of that, unless something
             # became functionally and systemically broken, is stale image
             # or flavor cache. Log a message, invalidate the caches so that
             # next time we get new caches.
             self._images = {}
-            self.__flavors = {}
+            self.__flavors = {}  # TODO(gtema): caching
             self.log.info(
                 "Clearing flavor and image caches due to 400 error from nova")
             raise
@@ -332,7 +339,7 @@ class OpenStackProvider(Provider):
     def getServerConsole(self, server_id):
         try:
             return self._client.get_server_console(server_id)
-        except shade.OpenStackCloudException:
+        except openstack.exceptions.OpenStackCloudException:
             return None
 
     def waitForServer(self, server, timeout=3600, auto_ip=True):

nodepool/logconfig.py

@@ -46,7 +46,7 @@ _DEFAULT_SERVER_LOGGING_CONFIG = {
             'handlers': ['console'],
             'level': 'WARN',
         },
-        'shade': {
+        'openstack': {
             'handlers': ['console'],
             'level': 'WARN',
         },

nodepool/task_manager.py

@@ -21,6 +21,8 @@ import logging
 import queue
 import time
 
+from openstack import task_manager as openstack_task_manager
+
 from nodepool import stats
 
 
@@ -28,24 +30,19 @@ class ManagerStoppedException(Exception):
     pass
 
 
-class TaskManager(object):
+class TaskManager(openstack_task_manager.TaskManager):
     log = logging.getLogger("nodepool.TaskManager")
 
-    def __init__(self, client, name, rate):
+    def __init__(self, name, rate, workers=5):
-        super(TaskManager, self).__init__()
+        super(TaskManager, self).__init__(name=name, workers=workers)
         self.daemon = True
         self.queue = queue.Queue()
         self._running = True
-        self.name = name
         self.rate = float(rate)
-        self._client = None
         self.statsd = stats.get_client()
         self._thread = threading.Thread(name=name, target=self.run)
         self._thread.daemon = True
 
-    def setClient(self, client):
-        self._client = client
-
     def start(self):
         self._thread.start()
 
@@ -70,33 +67,25 @@ class TaskManager(object):
                     if delta >= self.rate:
                         break
                     time.sleep(self.rate - delta)
-                self.log.debug("Manager %s running task %s (queue: %s)" %
+                self.log.debug("Manager %s running task %s (queue %s)" %
-                               (self.name, type(task).__name__,
+                               (self.name, task.name, self.queue.qsize()))
-                                self.queue.qsize()))
+                self.run_task(task)
-                start = time.time()
-                self.runTask(task)
-                last_ts = time.time()
-                dt = last_ts - start
-                self.log.debug("Manager %s ran task %s in %ss" %
-                               (self.name, type(task).__name__, dt))
-                if self.statsd:
-                    # nodepool.task.PROVIDER.subkey
-                    subkey = type(task).__name__
-                    key = 'nodepool.task.%s.%s' % (self.name, subkey)
-                    self.statsd.timing(key, int(dt * 1000))
-                    self.statsd.incr(key)
-
                 self.queue.task_done()
         except Exception:
             self.log.exception("Task manager died.")
             raise
 
-    def submitTask(self, task):
+    def post_run_task(self, elapsed_time, task):
+        super(TaskManager, self).post_run_task(elapsed_time, task)
+        if self.statsd:
+            # nodepool.task.PROVIDER.TASK_NAME
+            key = 'nodepool.task.%s.%s' % (self.name, task.name)
+            self.statsd.timing(key, int(elapsed_time * 1000))
+            self.statsd.incr(key)
+
+    def submit_task(self, task, raw=False):
         if not self._running:
             raise ManagerStoppedException(
                 "Manager %s is no longer running" % self.name)
         self.queue.put(task)
         return task.wait()
-
-    def runTask(self, task):
-        task.run(self._client)

requirements.txt

@@ -7,8 +7,7 @@ extras
 statsd>=3.0
 sqlalchemy>=0.8.2,<1.1.0
 PrettyTable>=0.6,<0.8
-os-client-config>=1.2.0
+openstacksdk>=0.16.0
-shade>=1.21.0
 diskimage-builder>=2.0.0
 voluptuous
 kazoo