diff --git a/nova/tests/functional/regressions/test_bug_1899649.py b/nova/tests/functional/regressions/test_bug_1899649.py
new file mode 100644
index 000000000000..302904658786
--- /dev/null
+++ b/nova/tests/functional/regressions/test_bug_1899649.py
@@ -0,0 +1,110 @@
+# Copyright 2020, Red Hat, Inc. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License"); you may
+# not use this file except in compliance with the License. You may obtain
+# a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
+# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
+# License for the specific language governing permissions and limitations
+# under the License.
+
+from nova.tests import fixtures as nova_fixtures
+from nova.tests.functional.libvirt import base
+from nova.tests.unit.virt.libvirt import fakelibvirt
+
+
+class TestVolAttachmentsAfterFailureToScheduleOrBuild(base.ServersTestBase):
+    """Regression test for bug .
+
+    This regression test aims to ensure a volume attachment remains in place
+    after a failure to either schedule a server or when building a server
+    directly on a compute after skipping the scheduler.
+
+    A volume attachment is required to remain after such failures to ensure the
+    volume itself remains marked as reserved.
+
+    To ensure this is as accurate as possible the tests use the libvirt
+    functional base class to mimic a real world example with NUMA nodes being
+    requested via flavor extra specs. The underlying compute being unable to
+    meet this request ensuring a failure.
+    """
+
+    microversion = 'latest'
+
+    def setUp(self):
+        super().setUp()
+
+        # Launch a single libvirt based compute service with a single NUMA node
+        host_info = fakelibvirt.HostInfo(
+            cpu_nodes=1, cpu_sockets=1, cpu_cores=2, kB_mem=15740000)
+        self.start_compute(host_info=host_info, hostname='compute1')
+
+        # Use a flavor requesting 2 NUMA nodes that we know will always fail
+        self.flavor_id = self._create_flavor(extra_spec={'hw:numa_nodes': '2'})
+
+        # Craft a common bfv server request for use within each test
+        self.volume_id = nova_fixtures.CinderFixture.IMAGE_BACKED_VOL
+        self.server = {
+            'name': 'test',
+            'flavorRef': self.flavor_id,
+            'imageRef': '',
+            'networks': 'none',
+            'block_device_mapping_v2': [{
+                'source_type': 'volume',
+                'destination_type': 'volume',
+                'boot_index': 0,
+                'uuid': self.volume_id}]
+        }
+
+    def _assert_failure_and_volume_attachments(self, server):
+        # Assert that the server is in an ERROR state
+        self._wait_for_state_change(server, 'ERROR')
+
+        # Assert that the volume is in a reserved state. As this isn't modelled
+        # by the CinderFixture we just assert that a single volume attachment
+        # remains after the failure and that it is referenced by the server.
+        attachments = self.cinder.volume_to_attachment.get(self.volume_id)
+        self.assertEqual(1, len(attachments))
+        self.assertIn(
+            self.volume_id, self.cinder.volume_ids_for_instance(server['id']))
+
+    def test_failure_to_schedule(self):
+        # Assert that a volume attachment remains after a failure to schedule
+        server = self.api.post_server({'server': self.server})
+        self._assert_failure_and_volume_attachments(server)
+
+    def test_failure_to_schedule_with_az(self):
+        # Assert that a volume attachment remains after a failure to schedule
+        # with the addition of an availability_zone in the request
+        self.server['availability_zone'] = 'nova'
+        server = self.api.post_server({'server': self.server})
+        self._assert_failure_and_volume_attachments(server)
+
+    def test_failure_to_schedule_with_host(self):
+        # Assert that a volume attachment remains after a failure to schedule
+        # using the optional host parameter introduced in microversion 2.74
+        self.server['host'] = 'compute1'
+        server = self.admin_api.post_server({'server': self.server})
+        self._assert_failure_and_volume_attachments(server)
+
+    def test_failure_to_build_with_az_and_host(self):
+        # Assert that a volume attachments does not remain after a failure to
+        # build and reschedule by providing an availability_zone *and* host,
+        # skipping the scheduler. This is bug #1899649.
+        self.server['availability_zone'] = 'nova:compute1'
+        server = self.admin_api.post_server({'server': self.server})
+
+        # Assert the server ends up in an ERROR state
+        self._wait_for_state_change(server, 'ERROR')
+
+        # FIXME(lyarwood): A single volume attachment should be present for the
+        # instance at this stage as the volume *can* otherwise be marked as
+        # available within Cinder if it isn't multi-attached.
+        attachments = self.cinder.volume_to_attachment.get(self.volume_id)
+        self.assertEqual(0, len(attachments))
+        self.assertNotIn(
+            self.volume_id, self.cinder.volume_ids_for_instance(server['id']))