diff --git a/ironic/nova/compute/manager.py b/ironic/nova/compute/manager.py
index c2bf47a3f..59782d7bd 100644
--- a/ironic/nova/compute/manager.py
+++ b/ironic/nova/compute/manager.py
@@ -23,16 +23,33 @@ work. The goal here is to generalise the areas where n-c talking to a clustered
 hypervisor has issues, and long term fold them into the main ComputeManager.
 """
 
+from oslo_concurrency import lockutils
+from oslo_log import log as logging
+
 from nova.compute import manager
 import nova.context
-from oslo_concurrency import lockutils
+from nova.i18n import _LW
 
 
+LOG = logging.getLogger(__name__)
+
 CCM_SEMAPHORE = 'clustered_compute_manager'
 
 
+# TODO(jroll) delete this in Ocata
 class ClusteredComputeManager(manager.ComputeManager):
 
+    def __init__(self, *args, **kwargs):
+        LOG.warning(_LW('ClusteredComputeManager is deprecated. As of the '
+                        'Newton release of Nova, the ironic '
+                        'virt driver directly supports using multiple '
+                        'compute hosts without ClusteredComputeManager. '
+                        'Users should unset the `compute_manager` '
+                        'configuration option to use Nova\'s default '
+                        'instead. ClusteredComputeManager will be removed '
+                        'during the Ocata cycle.'))
+        super(ClusteredComputeManager, self).__init__(*args, **kwargs)
+
     def init_host(self):
         """Initialization for a clustered compute service."""
         self.driver.init_host(host=self.host)
diff --git a/releasenotes/notes/deprecate-clustered-compute-manager-3dd68557446bcc5c.yaml b/releasenotes/notes/deprecate-clustered-compute-manager-3dd68557446bcc5c.yaml
new file mode 100644
index 000000000..539f91760
--- /dev/null
+++ b/releasenotes/notes/deprecate-clustered-compute-manager-3dd68557446bcc5c.yaml
@@ -0,0 +1,12 @@
+---
+deprecations:
+  - |
+    In Nova commit 6047d790a32ef5a65d4d6b029f673ce53c3d4141,
+    functionality is added to the ironic virt
+    driver to support multiple compute hosts without using the hack
+    we call ClusteredComputeManager.
+
+    We anticipate this being included in the Newton release of Nova.
+    In advance of that, we are marking this unsupported component
+    as deprecated, and plan to remove it before the end of the Ocata
+    development cycle.