Set node "alive" when inspection finished

To make the node fast trackable as soon as
inspection finishes, in addition add a wait for the
agent to callback should it not be available when
fast track is attempted.

Closes-Bug: #2078820

Change-Id: I8a95fc08cf355b7b745a565e3a05c9dc0875a63e

ironic/conductor/cleaning.py

@@ -95,6 +95,12 @@ def do_node_clean(task, clean_steps=None, disable_ramdisk=False):
                      'out-of-band only cleaning has been requested for node '
                      '%s', node.uuid)
             prepare_result = None
+    except exception.AgentConnectionFailed:
+        LOG.info('Agent is not yet running on node %(node)s, waiting for'
+                 ' agent to come up for fast track', {'node': node.uuid})
+        target_state = states.MANAGEABLE if manual_clean else None
+        task.process_event('wait', target_state=target_state)
+        return
     except Exception as e:
         msg = (_('Failed to prepare node %(node)s for cleaning: %(e)s')
                % {'node': node.uuid, 'e': e})

ironic/conductor/deployments.py

@@ -187,6 +187,11 @@ def do_node_deploy(task, conductor_id=None, configdrive=None,
 
     try:
         task.driver.deploy.prepare(task)
+    except exception.AgentConnectionFailed:
+        LOG.info('Agent is not yet running on node %(node)s, waiting for agent'
+                 ' to come up for fast track', {'node': node.uuid})
+        task.process_event('wait')
+        return
     except exception.IronicException as e:
         with excutils.save_and_reraise_exception():
             utils.deploying_error_handler(

ironic/conductor/utils.py

@@ -1150,13 +1150,18 @@ def fast_track_able(task):
 def value_within_timeout(value, timeout):
     """Checks if the time is within the previous timeout seconds from now.
 
-    :param value: a string representing date and time or None.
+    :param value: a datetime or string representing date and time or None.
     :param timeout: timeout in seconds.
     """
     # use native datetime objects for conversion and compare
     # slightly odd because py2 compatibility :(
-    last = datetime.datetime.strptime(value or '1970-01-01T00:00:00.000000',
-                                      "%Y-%m-%dT%H:%M:%S.%f")
+    if isinstance(value, datetime.datetime):
+        # Converts to a offset-naive datetime(as created by timeutils.utcnow())
+        last = value.replace(tzinfo=None)
+    else:
+        defaultdt = '1970-01-01T00:00:00.000000'
+        last = datetime.datetime.strptime(value or defaultdt,
+                                          '%Y-%m-%dT%H:%M:%S.%f')
     # If we found nothing, we assume that the time is essentially epoch.
     time_delta = datetime.timedelta(seconds=timeout)
     last_valid = timeutils.utcnow() - time_delta
@@ -1173,14 +1178,20 @@ def agent_is_alive(node, timeout=None):
     :param node: A node object.
     :param timeout: Heartbeat timeout, defaults to `fast_track_timeout`.
     """
+
+    timeout = timeout or CONF.deploy.fast_track_timeout
+    if node.power_state == states.POWER_ON and \
+       node.inspection_finished_at and \
+       value_within_timeout(node.inspection_finished_at, timeout):
+        return True
+
     # If no agent_url is present then we have powered down since the
     # last agent heartbeat
     if not node.driver_internal_info.get('agent_url'):
         return False
 
     return value_within_timeout(
-        node.driver_internal_info.get('agent_last_heartbeat'),
-        timeout or CONF.deploy.fast_track_timeout)
+        node.driver_internal_info.get('agent_last_heartbeat'), timeout)
 
 
 def is_fast_track(task):

ironic/tests/unit/conductor/test_utils.py

@@ -2238,6 +2238,18 @@ class FastTrackTestCase(db_base.DbTestCase):
                 self.context, self.node.uuid, shared=False) as task:
             self.assertFalse(conductor_utils.is_fast_track(task))
 
+    def test_is_fast_track_inspected_no_heartbeat(self, mock_get_power):
+        mock_get_power.return_value = states.POWER_ON
+        self.node = obj_utils.create_test_node(
+            self.context, driver='fake-hardware',
+            uuid=uuidutils.generate_uuid(),
+            inspection_finished_at=timeutils.utcnow(),
+            power_state=states.POWER_ON
+        )
+        with task_manager.acquire(
+                self.context, self.node.uuid, shared=False) as task:
+            self.assertTrue(conductor_utils.is_fast_track(task))
+
     def test_is_fast_track_powered_after_heartbeat(self, mock_get_power):
         mock_get_power.return_value = states.POWER_ON
         with task_manager.acquire(

releasenotes/notes/set-node-alive-when-inspection-finished-1ec74828852eaeef.yaml

@@ -0,0 +1,7 @@
+---
+fixes:
+  - |
+    Set node "alive" and make it fast trackable
+    as soon as inspection is finished, in addition
+    add a wait for the agent to callback should
+    it not be available when fast track is attempted.