def prepare_cleaning(self, task): """Boot into the ramdisk to prepare for cleaning. :param task: a TaskManager object containing the node :raises NodeCleaningFailure: if the previous cleaning ports cannot be removed or if new cleaning ports cannot be created :returns: None or states.CLEANWAIT for async prepare. """ node = task.node conductor_steps.set_node_cleaning_steps(task) if not node.driver_internal_info['clean_steps']: # no clean steps configured, nothing to do. return fast_track = manager_utils.is_fast_track(task) power_state_to_restore = None if not fast_track: power_state_to_restore = manager_utils.power_on_node_if_needed( task) task.driver.network.add_cleaning_network(task) manager_utils.restore_power_state_if_needed(task, power_state_to_restore) boot_opt = deploy_utils.build_agent_options(node) task.driver.boot.prepare_ramdisk(task, boot_opt) if not fast_track: manager_utils.node_power_action(task, states.REBOOT) return states.CLEANWAIT
def test_set_node_cleaning_steps_automated(self, mock_steps, mock_validate_user_steps): mock_steps.return_value = self.clean_steps node = obj_utils.create_test_node( self.context, driver='fake-hardware', provision_state=states.CLEANING, target_provision_state=states.AVAILABLE, last_error=None, clean_step=None) with task_manager.acquire( self.context, node.uuid, shared=False) as task: conductor_steps.set_node_cleaning_steps(task) node.refresh() self.assertEqual(self.clean_steps, node.driver_internal_info['clean_steps']) self.assertEqual({}, node.clean_step) mock_steps.assert_called_once_with(task, enabled=True) self.assertFalse(mock_validate_user_steps.called)
def prepare_cleaning(self, task): """Boot into the ramdisk to prepare for cleaning. :param task: a TaskManager object containing the node :raises NodeCleaningFailure: if the previous cleaning ports cannot be removed or if new cleaning ports cannot be created :returns: None or states.CLEANWAIT for async prepare. """ node = task.node conductor_steps.set_node_cleaning_steps(task) if not node.driver_internal_info['clean_steps']: # no clean steps configured, nothing to do. return power_state_to_restore = manager_utils.power_on_node_if_needed(task) task.driver.network.add_cleaning_network(task) manager_utils.restore_power_state_if_needed( task, power_state_to_restore) boot_opt = deploy_utils.build_agent_options(node) task.driver.boot.prepare_ramdisk(task, boot_opt) manager_utils.node_power_action(task, states.REBOOT) return states.CLEANWAIT
def test_set_node_cleaning_steps_manual(self, mock_steps, mock_validate_user_steps): clean_steps = [self.deploy_raid] mock_steps.return_value = self.clean_steps mock_validate_user_steps.return_value = clean_steps node = obj_utils.create_test_node( self.context, driver='fake-hardware', provision_state=states.CLEANING, target_provision_state=states.MANAGEABLE, last_error=None, clean_step=None, driver_internal_info={'clean_steps': clean_steps}) with task_manager.acquire( self.context, node.uuid, shared=False) as task: conductor_steps.set_node_cleaning_steps(task) node.refresh() self.assertEqual(clean_steps, node.driver_internal_info['clean_steps']) self.assertEqual({}, node.clean_step) self.assertFalse(mock_steps.called) mock_validate_user_steps.assert_called_once_with(task, clean_steps)
def do_node_clean(task, clean_steps=None): """Internal RPC method to perform cleaning of a node. :param task: a TaskManager instance with an exclusive lock on its node :param clean_steps: For a manual clean, the list of clean steps to perform. Is None For automated cleaning (default). For more information, see the clean_steps parameter of :func:`ConductorManager.do_node_clean`. """ node = task.node manual_clean = clean_steps is not None clean_type = 'manual' if manual_clean else 'automated' LOG.debug('Starting %(type)s cleaning for node %(node)s', {'type': clean_type, 'node': node.uuid}) if not manual_clean and utils.skip_automated_cleaning(node): # Skip cleaning, move to AVAILABLE. node.clean_step = None node.save() task.process_event('done') LOG.info('Automated cleaning is disabled, node %s has been ' 'successfully moved to AVAILABLE state.', node.uuid) return # NOTE(dtantsur): this is only reachable during automated cleaning, # for manual cleaning we verify maintenance mode earlier on. if (not CONF.conductor.allow_provisioning_in_maintenance and node.maintenance): msg = _('Cleaning a node in maintenance mode is not allowed') return utils.cleaning_error_handler(task, msg, tear_down_cleaning=False) try: # NOTE(ghe): Valid power and network values are needed to perform # a cleaning. task.driver.power.validate(task) task.driver.network.validate(task) except exception.InvalidParameterValue as e: msg = (_('Validation failed. Cannot clean node %(node)s. ' 'Error: %(msg)s') % {'node': node.uuid, 'msg': e}) return utils.cleaning_error_handler(task, msg) if manual_clean: info = node.driver_internal_info info['clean_steps'] = clean_steps node.driver_internal_info = info node.save() # Do caching of bios settings if supported by driver, # this will be called for both manual and automated cleaning. try: task.driver.bios.cache_bios_settings(task) except exception.UnsupportedDriverExtension: LOG.warning('BIOS settings are not supported for node %s, ' 'skipping', task.node.uuid) # TODO(zshi) remove this check when classic drivers are removed except Exception: msg = (_('Caching of bios settings failed on node %(node)s. ' 'Continuing with node cleaning.') % {'node': node.uuid}) LOG.exception(msg) # Allow the deploy driver to set up the ramdisk again (necessary for # IPA cleaning) try: prepare_result = task.driver.deploy.prepare_cleaning(task) except Exception as e: msg = (_('Failed to prepare node %(node)s for cleaning: %(e)s') % {'node': node.uuid, 'e': e}) LOG.exception(msg) return utils.cleaning_error_handler(task, msg) if prepare_result == states.CLEANWAIT: # Prepare is asynchronous, the deploy driver will need to # set node.driver_internal_info['clean_steps'] and # node.clean_step and then make an RPC call to # continue_node_clean to start cleaning. # For manual cleaning, the target provision state is MANAGEABLE, # whereas for automated cleaning, it is AVAILABLE (the default). target_state = states.MANAGEABLE if manual_clean else None task.process_event('wait', target_state=target_state) return try: conductor_steps.set_node_cleaning_steps(task) except (exception.InvalidParameterValue, exception.NodeCleaningFailure) as e: msg = (_('Cannot clean node %(node)s. Error: %(msg)s') % {'node': node.uuid, 'msg': e}) return utils.cleaning_error_handler(task, msg) steps = node.driver_internal_info.get('clean_steps', []) step_index = 0 if steps else None do_next_clean_step(task, step_index)
def continue_cleaning(self, task, **kwargs): """Start the next cleaning step if the previous one is complete. In order to avoid errors and make agent upgrades painless, the agent compares the version of all hardware managers at the start of the cleaning (the agent's get_clean_steps() call) and before executing each clean step. If the version has changed between steps, the agent is unable to tell if an ordering change will cause a cleaning issue so it returns CLEAN_VERSION_MISMATCH. For automated cleaning, we restart the entire cleaning cycle. For manual cleaning, we don't. Additionally, if a clean_step includes the reboot_requested property set to True, this method will coordinate the reboot once the step is completed. """ node = task.node # For manual clean, the target provision state is MANAGEABLE, whereas # for automated cleaning, it is (the default) AVAILABLE. manual_clean = node.target_provision_state == states.MANAGEABLE agent_commands = self._client.get_commands_status(task.node) if not agent_commands: if task.node.driver_internal_info.get('cleaning_reboot'): # Node finished a cleaning step that requested a reboot, and # this is the first heartbeat after booting. Continue cleaning. info = task.node.driver_internal_info info.pop('cleaning_reboot', None) task.node.driver_internal_info = info task.node.save() manager_utils.notify_conductor_resume_clean(task) return else: # Agent has no commands whatsoever return command = _get_completed_cleaning_command(task, agent_commands) LOG.debug('Cleaning command status for node %(node)s on step %(step)s:' ' %(command)s', {'node': node.uuid, 'step': node.clean_step, 'command': command}) if not command: # Agent command in progress return if command.get('command_status') == 'FAILED': msg = (_('Agent returned error for clean step %(step)s on node ' '%(node)s : %(err)s.') % {'node': node.uuid, 'err': command.get('command_error'), 'step': node.clean_step}) LOG.error(msg) return manager_utils.cleaning_error_handler(task, msg) elif command.get('command_status') == 'CLEAN_VERSION_MISMATCH': # Cache the new clean steps (and 'hardware_manager_version') try: self.refresh_clean_steps(task) except exception.NodeCleaningFailure as e: msg = (_('Could not continue cleaning on node ' '%(node)s: %(err)s.') % {'node': node.uuid, 'err': e}) LOG.exception(msg) return manager_utils.cleaning_error_handler(task, msg) if manual_clean: # Don't restart manual cleaning if agent reboots to a new # version. Both are operator actions, unlike automated # cleaning. Manual clean steps are not necessarily idempotent # like automated clean steps and can be even longer running. LOG.info('During manual cleaning, node %(node)s detected ' 'a clean version mismatch. Re-executing and ' 'continuing from current step %(step)s.', {'node': node.uuid, 'step': node.clean_step}) driver_internal_info = node.driver_internal_info driver_internal_info['skip_current_clean_step'] = False node.driver_internal_info = driver_internal_info node.save() else: # Restart cleaning, agent must have rebooted to new version LOG.info('During automated cleaning, node %s detected a ' 'clean version mismatch. Resetting clean steps ' 'and rebooting the node.', node.uuid) try: conductor_steps.set_node_cleaning_steps(task) except exception.NodeCleaningFailure: msg = (_('Could not restart automated cleaning on node ' '%(node)s: %(err)s.') % {'node': node.uuid, 'err': command.get('command_error'), 'step': node.clean_step}) LOG.exception(msg) return manager_utils.cleaning_error_handler(task, msg) manager_utils.notify_conductor_resume_clean(task) elif command.get('command_status') == 'SUCCEEDED': clean_step_hook = _get_post_clean_step_hook(node) if clean_step_hook is not None: LOG.debug('For node %(node)s, executing post clean step ' 'hook %(method)s for clean step %(step)s', {'method': clean_step_hook.__name__, 'node': node.uuid, 'step': node.clean_step}) try: clean_step_hook(task, command) except Exception as e: msg = (_('For node %(node)s, post clean step hook ' '%(method)s failed for clean step %(step)s.' '%(cls)s: %(error)s') % {'method': clean_step_hook.__name__, 'node': node.uuid, 'error': e, 'cls': e.__class__.__name__, 'step': node.clean_step}) LOG.exception(msg) return manager_utils.cleaning_error_handler(task, msg) if task.node.clean_step.get('reboot_requested'): _cleaning_reboot(task) return LOG.info('Agent on node %s returned cleaning command success, ' 'moving to next clean step', node.uuid) manager_utils.notify_conductor_resume_clean(task) else: msg = (_('Agent returned unknown status for clean step %(step)s ' 'on node %(node)s : %(err)s.') % {'node': node.uuid, 'err': command.get('command_status'), 'step': node.clean_step}) LOG.error(msg) return manager_utils.cleaning_error_handler(task, msg)
def heartbeat(self, task, callback_url, agent_version): """Process a heartbeat. :param task: task to work with. :param callback_url: agent HTTP API URL. :param agent_version: The version of the agent that is heartbeating """ # NOTE(pas-ha) immediately skip the rest if nothing to do if (task.node.provision_state not in self.heartbeat_allowed_states and not manager_utils.fast_track_able(task)): LOG.debug('Heartbeat from node %(node)s in unsupported ' 'provision state %(state)s, not taking any action.', {'node': task.node.uuid, 'state': task.node.provision_state}) return try: task.upgrade_lock() except exception.NodeLocked: LOG.warning('Node %s is currently locked, skipping heartbeat ' 'processing (will retry on the next heartbeat)', task.node.uuid) return node = task.node LOG.debug('Heartbeat from node %s', node.uuid) driver_internal_info = node.driver_internal_info driver_internal_info['agent_url'] = callback_url driver_internal_info['agent_version'] = agent_version # Record the last heartbeat event time in UTC, so we can make # decisions about it later. Can be decoded to datetime object with: # datetime.datetime.strptime(var, "%Y-%m-%d %H:%M:%S.%f") driver_internal_info['agent_last_heartbeat'] = str( timeutils.utcnow().isoformat()) node.driver_internal_info = driver_internal_info node.save() if node.provision_state in _HEARTBEAT_RECORD_ONLY: # We shouldn't take any additional action. The agent will # silently continue to heartbeat to ironic until user initiated # state change occurs causing it to match a state below. LOG.debug('Heartbeat from %(node)s recorded to identify the ' 'node as on-line.', {'node': task.node.uuid}) return # Async call backs don't set error state on their own # TODO(jimrollenhagen) improve error messages here msg = _('Failed checking if deploy is done.') try: if node.maintenance: # this shouldn't happen often, but skip the rest if it does. LOG.debug('Heartbeat from node %(node)s in maintenance mode; ' 'not taking any action.', {'node': node.uuid}) return # NOTE(mgoddard): Only handle heartbeats during DEPLOYWAIT if we # are currently in the core deploy.deploy step. Other deploy steps # may cause the agent to boot, but we should not trigger deployment # at that point. elif node.provision_state == states.DEPLOYWAIT: if self.in_core_deploy_step(task): if not self.deploy_has_started(task): msg = _('Node failed to deploy.') self.continue_deploy(task) elif self.deploy_is_done(task): msg = _('Node failed to move to active state.') self.reboot_to_instance(task) else: node.touch_provisioning() else: node.touch_provisioning() elif node.provision_state == states.CLEANWAIT: node.touch_provisioning() if not node.clean_step: LOG.debug('Node %s just booted to start cleaning.', node.uuid) msg = _('Node failed to start the first cleaning step.') # First, cache the clean steps self.refresh_clean_steps(task) # Then set/verify node clean steps and start cleaning conductor_steps.set_node_cleaning_steps(task) # The exceptions from RPC are not possible as we using cast # here manager_utils.notify_conductor_resume_clean(task) else: msg = _('Node failed to check cleaning progress.') self.continue_cleaning(task) elif (node.provision_state == states.RESCUEWAIT): msg = _('Node failed to perform rescue operation.') self._finalize_rescue(task) except Exception as e: err_info = {'msg': msg, 'e': e} last_error = _('Asynchronous exception: %(msg)s ' 'Exception: %(e)s for node') % err_info errmsg = last_error + ' %(node)s' LOG.exception(errmsg, {'node': node.uuid}) if node.provision_state in (states.CLEANING, states.CLEANWAIT): manager_utils.cleaning_error_handler(task, last_error) elif node.provision_state in (states.DEPLOYING, states.DEPLOYWAIT): deploy_utils.set_failed_state( task, last_error, collect_logs=bool(self._client)) elif node.provision_state in (states.RESCUING, states.RESCUEWAIT): manager_utils.rescuing_error_handler(task, last_error)