Backup manager: Synchronously call remove_export

Call the volume's rpcapi remove_export method with
a "call" instead of a "cast" when detaching volumes
in the backup manager.

This prevents _detach_volume() from returning before
the remove_export() has completed, which helps prevent
problems where remove_export() races against a
subsequent operation.

Closes-Bug: #1920237

Change-Id: I482c3d5520a7bbb9971942cdb4c4208052106d70
(cherry picked from commit d6895f9d2a)
(cherry picked from commit 5402ac8e2d)

cinder/backup/manager.py

@@ -1072,7 +1072,7 @@ class BackupManager(manager.SchedulerDependentManager):
         if not is_snapshot:
             rpcapi.terminate_connection(ctxt, device, properties,
                                         force=force)
-            rpcapi.remove_export(ctxt, device)
+            rpcapi.remove_export(ctxt, device, sync=True)
         else:
             rpcapi.terminate_connection_snapshot(ctxt, device,
                                                  properties, force=force)

cinder/volume/rpcapi.py

@@ -231,9 +231,12 @@ class VolumeAPI(rpc.RPCAPI):
         return cctxt.call(ctxt, 'terminate_connection', volume_id=volume['id'],
                           connector=connector, force=force)
 
-    def remove_export(self, ctxt, volume):
+    def remove_export(self, ctxt, volume, sync=False):
         cctxt = self._get_cctxt(volume.service_topic_queue)
-        cctxt.cast(ctxt, 'remove_export', volume_id=volume['id'])
+        if sync:
+            cctxt.call(ctxt, 'remove_export', volume_id=volume.id)
+        else:
+            cctxt.cast(ctxt, 'remove_export', volume_id=volume.id)
 
     def publish_service_capabilities(self, ctxt):
         cctxt = self._get_cctxt(fanout=True)

releasenotes/notes/bug-1920237-backup-remove-export-race-941e2ab1f056e54c.yaml

@@ -0,0 +1,8 @@
+---
+fixes:
+  - |
+    `Bug #1920237 <https://bugs.launchpad.net/cinder/+bug/1920237>`_: The
+    backup manager calls volume remove_export() but does not wait for it to
+    complete when detaching a volume after backup.  This caused problems
+    when a subsequent operation started on that volume before it had fully
+    detached.