diff --git a/ironic/drivers/modules/irmc/raid.py b/ironic/drivers/modules/irmc/raid.py
index 34d1c3f382..25a856e8ca 100644
--- a/ironic/drivers/modules/irmc/raid.py
+++ b/ironic/drivers/modules/irmc/raid.py
@@ -483,6 +483,7 @@ class IRMCRAID(base.RAIDInterface):
                     if all(fgi_status == 'Idle' for fgi_status in
                            fgi_status_dict.values()):
                         raid_config.update({'fgi_status': RAID_COMPLETED})
+                        raid_common.update_raid_info(node, raid_config)
                         LOG.info('RAID configuration has completed on '
                                  'node %(node)s with fgi_status is %(fgi)s',
                                  {'node': node_uuid, 'fgi': RAID_COMPLETED})
@@ -506,5 +507,4 @@ class IRMCRAID(base.RAIDInterface):
         task.process_event('fail')
 
     def _resume_cleaning(self, task):
-        raid_common.update_raid_info(task.node, task.node.raid_config)
         manager_utils.notify_conductor_resume_clean(task)
diff --git a/ironic/tests/unit/drivers/modules/irmc/test_periodic_task.py b/ironic/tests/unit/drivers/modules/irmc/test_periodic_task.py
index 6bda0fee67..865f589626 100644
--- a/ironic/tests/unit/drivers/modules/irmc/test_periodic_task.py
+++ b/ironic/tests/unit/drivers/modules/irmc/test_periodic_task.py
@@ -308,3 +308,41 @@ class iRMCPeriodicTaskTestCase(test_common.BaseIRMCTest):
         clean_fail_mock.assert_called_once_with(mock.ANY,
                                                 task, fgi_status_dict)
         clean_mock.assert_called_once_with(mock.ANY, task)
+
+    @mock.patch('ironic.drivers.modules.irmc.raid.IRMCRAID._resume_cleaning',
+                autospec=True)
+    @mock.patch('ironic.drivers.modules.irmc.raid.IRMCRAID._set_clean_failed',
+                autospec=True)
+    @mock.patch('ironic.drivers.modules.irmc.raid._get_fgi_status',
+                autospec=True)
+    @mock.patch.object(irmc_common, 'get_irmc_report', autospec=True)
+    @mock.patch.object(task_manager, 'acquire', autospec=True)
+    def test__query_raid_config_fgi_status_avoid_repeatedly_resume_cleaning(
+            self, mock_acquire, report_mock, fgi_mock, clean_fail_mock,
+            clean_mock):
+        mock_manager = mock.Mock()
+        raid_config = self.raid_config
+        fgi_mock.return_value = {'0': 'Idle', '1': 'Idle'}
+        task = mock.Mock(node=self.node, driver=self.driver)
+        mock_acquire.return_value = mock.MagicMock(
+            __enter__=mock.MagicMock(return_value=task))
+        task.node.raid_config = raid_config
+        node_list = [(self.node.uuid, 'irmc', '', raid_config)]
+        mock_manager.iter_nodes.return_value = node_list
+        # Set provision state value
+        task.node.provision_state = 'clean wait'
+        task.node.save()
+        task.driver.raid._query_raid_config_fgi_status(mock_manager,
+                                                       self.context)
+        raid_config = task.node.raid_config
+        node_list = [(self.node.uuid, 'irmc', '', raid_config)]
+        mock_manager.iter_nodes.return_value = node_list
+        task.node.provision_state = 'clean wait'
+        task.node.save()
+        task.driver.raid._query_raid_config_fgi_status(mock_manager,
+                                                       self.context)
+        self.assertEqual(0, clean_fail_mock.call_count)
+        report_mock.assert_called_once_with(task.node)
+        fgi_mock.assert_called_once_with(report_mock.return_value,
+                                         self.node.uuid)
+        clean_mock.assert_called_once_with(mock.ANY, task)
diff --git a/releasenotes/notes/irmc-fix-repeatedly-resuming-clean-020f0dfc2e30d7bc.yaml b/releasenotes/notes/irmc-fix-repeatedly-resuming-clean-020f0dfc2e30d7bc.yaml
new file mode 100644
index 0000000000..02104cc186
--- /dev/null
+++ b/releasenotes/notes/irmc-fix-repeatedly-resuming-clean-020f0dfc2e30d7bc.yaml
@@ -0,0 +1,7 @@
+---
+fixes:
+  - |
+    Fixed the bug of repeated resume cleaning due to the value of
+    `fgi_status` not being updated correctly when obtaining the
+    RAID configuration status of the node managed by the `irmc`
+    hardware type.