Fix detach notification

Our current `attachment_delete` methods in the volume API and the
manager are using DB methods directly, which makes the OVOs present in
those methods get out of sync with the latest data, which leads to
notifications having the wrong data when we send them on volume detach.

This patch replaces DB method calls with OVO calls and moves the
notification call to the end of the method, where we have the final
status on the volume.

It also adds the missing detach.start notification when deleting an
attachment in the reserved state.

Closes-Bug: #1916980
Closes-Bug: #1935011
Change-Id: Ie48cf55deacd08e7716201dac00ede8d57e6632f

cinder/tests/unit/attachments/test_attachments_manager.py

@@ -166,6 +166,12 @@ class AttachmentManagerTestCase(test.TestCase):
                   mock_db_detached, mock_db_meta_delete, mock_get_attachment):
             mock_elevated.return_value = self.context
             mock_con_term.return_value = False
+            mock_db_detached.return_value = (
+                {'status': 'available',
+                 'attach_status': fields.VolumeAttachStatus.DETACHED},
+                {'attach_status': fields.VolumeAttachStatus.DETACHED,
+                 'deleted': True}
+            )
 
             # test single attachment. This should call
             # detach and remove_export

cinder/volume/api.py

@@ -2259,17 +2259,17 @@ class API(base.Base):
         ctxt.authorize(attachment_policy.DELETE_POLICY,
                        target_obj=attachment)
         volume = attachment.volume
+
         if attachment.attach_status == fields.VolumeAttachStatus.RESERVED:
-            self.db.volume_detached(ctxt.elevated(), attachment.volume_id,
-                                    attachment.get('id'))
-            self.db.volume_admin_metadata_delete(ctxt.elevated(),
-                                                 attachment.volume_id,
-                                                 'attached_mode')
-            volume_utils.notify_about_volume_usage(ctxt, volume, "detach.end")
+            volume_utils.notify_about_volume_usage(ctxt, volume,
+                                                   "detach.start")
+            volume.finish_detach(attachment.id)
+            do_notify = True
         else:
             self.volume_rpcapi.attachment_delete(ctxt,
                                                  attachment.id,
                                                  volume)
+            do_notify = False
         status_updates = {'status': 'available',
                           'attach_status': 'detached'}
         remaining_attachments = AO_LIST.get_all_by_volume_id(ctxt, volume.id)
@@ -2305,6 +2305,8 @@ class API(base.Base):
         volume.attach_status = status_updates['attach_status']
         volume.save()
 
+        if do_notify:
+            volume_utils.notify_about_volume_usage(ctxt, volume, "detach.end")
         return remaining_attachments
 
 

cinder/volume/manager.py

@@ -4954,15 +4954,11 @@ class VolumeManager(manager.CleanableManager,
             # represent a single failed attach out of multiple attachments
 
             # TODO(jdg): object method here
-            self.db.volume_attachment_update(
-                context, attachment.get('id'),
-                {'attach_status': fields.VolumeAttachStatus.ERROR_DETACHING})
+            attachment.attach_status = \
+                fields.VolumeAttachStatus.ERROR_DETACHING
+            attachment.save()
         else:
-            self.db.volume_detached(context.elevated(), vref.id,
-                                    attachment.get('id'))
-            self.db.volume_admin_metadata_delete(context.elevated(),
-                                                 vref.id,
-                                                 'attached_mode')
+            vref.finish_detach(attachment_id)
         self._notify_about_volume_usage(context, vref, "detach.end")
 
     # Replication group API (Tiramisu)

releasenotes/notes/detach-notification-31ae15dafdef36c1.yaml

@@ -0,0 +1,9 @@
+---
+fixes:
+  - |
+    `Bug #1916980 <https://bugs.launchpad.net/cinder/+bug/1916980>`_: Fixed
+    stale volume notification information on volume detach.
+  - |
+    `Bug #1935011 <https://bugs.launchpad.net/cinder/+bug/1935011>`_: Fixed
+    missing detach.start notification when deleting an attachment in reserved
+    state.