def _inspect_hardware(node):
"""Inspect the node and get hardware information.
:param node: node object.
:raises: HardwareInspectionFailure, if unable to get essential
hardware properties.
:returns: a pair of dictionary and list, the dictionary contains
keys as in IRMCInspect.ESSENTIAL_PROPERTIES and its inspected
values, the list contains mac addresses.
"""
try:
report = irmc_common.get_irmc_report(node)
props = scci.get_essential_properties(report,
IRMCInspect.ESSENTIAL_PROPERTIES)
macs = _get_mac_addresses(node)
except (scci.SCCIInvalidInputError, scci.SCCIClientError,
exception.SNMPFailure) as e:
error = (_("Inspection failed for node %(node_id)s "
"with the following error: %(error)s") % {
'node_id': node.uuid,
'error': e
})
raise exception.HardwareInspectionFailure(error=error)
return (props, macs)
def _inspect_hardware(node, **kwargs):
"""Inspect the node and get hardware information.
:param node: node object.
:param kwargs: the dictionary of additional parameters.
:raises: HardwareInspectionFailure, if unable to get essential
hardware properties.
:returns: a pair of dictionary and list, the dictionary contains
keys as in IRMCInspect.ESSENTIAL_PROPERTIES and its inspected
values, the list contains mac addresses.
"""
capabilities_props = set(CAPABILITIES_PROPERTIES)
# Remove all capabilities item which will be inspected in the existing
# capabilities of node
if 'capabilities' in node.properties:
existing_cap = node.properties['capabilities'].split(',')
for item in capabilities_props:
for prop in existing_cap:
if item == prop.split(':')[0]:
existing_cap.remove(prop)
node.properties['capabilities'] = ",".join(existing_cap)
# get gpu_ids in ironic configuration
values = [gpu_id.lower() for gpu_id in CONF.irmc.gpu_ids]
# if gpu_ids = [], pci_gpu_devices will not be inspected
if len(values) == 0:
capabilities_props.remove('pci_gpu_devices')
try:
report = irmc_common.get_irmc_report(node)
props = scci.get_essential_properties(report,
IRMCInspect.ESSENTIAL_PROPERTIES)
d_info = irmc_common.parse_driver_info(node)
capabilities = scci.get_capabilities_properties(
d_info, capabilities_props, values, **kwargs)
if capabilities:
if capabilities.get('pci_gpu_devices') == 0:
capabilities.pop('pci_gpu_devices')
if capabilities.get('trusted_boot') is False:
capabilities.pop('trusted_boot')
capabilities = utils.get_updated_capabilities(
node.properties.get('capabilities'), capabilities)
if capabilities:
props['capabilities'] = capabilities
macs = _get_mac_addresses(node)
except (scci.SCCIInvalidInputError, scci.SCCIClientError,
exception.SNMPFailure) as e:
error = (_("Inspection failed for node %(node_id)s "
"with the following error: %(error)s") % {
'node_id': node.uuid,
'error': e
})
raise exception.HardwareInspectionFailure(error=error)
return (props, macs)
def _get_sensors_data(task):
"""Get sensors data method.
It gets sensor data from the task's node via SCCI, and convert the data
from XML to the dict format.
:param task: A TaskManager instance.
:raises: FailedToGetSensorData when getting the sensor data fails.
:returns: Returns a consistent formatted dict of sensor data grouped
by sensor type, which can be processed by Ceilometer.
"""
try:
report = irmc_common.get_irmc_report(task.node)
sensor = scci.get_sensor_data(report)
except (exception.InvalidParameterValue,
exception.MissingParameterValue,
scci.SCCIInvalidInputError,
scci.SCCIClientError) as e:
LOG.error(_LE("SCCI get sensor data failed for node %(node_id)s "
"with the following error: %(error)s"),
{'node_id': task.node.uuid, 'error': e})
raise exception.FailedToGetSensorData(
node=task.node.uuid, error=e)
sensors_data = {}
for sdr in sensor:
sensor_type_name = sdr.find('./Data/Decoded/Sensor/TypeName')
sensor_type_number = sdr.find('./Data/Decoded/Sensor/Type')
entity_name = sdr.find('./Data/Decoded/Entity/Name')
entity_id = sdr.find('./Data/Decoded/Entity/ID')
if None in (sensor_type_name, sensor_type_number,
entity_name, entity_id):
continue
sensor_type = ('%s (%s)' %
(sensor_type_name.text, sensor_type_number.text))
sensor_id = ('%s (%s)' %
(entity_name.text, entity_id.text))
reading_value = sdr.find(
'./Data/Decoded/Sensor/Thresholds/*/Normalized')
reading_value_text = "None" if (
reading_value is None) else str(reading_value.text)
reading_units = sdr.find('./Data/Decoded/Sensor/BaseUnitName')
reading_units_text = "None" if (
reading_units is None) else str(reading_units.text)
sensor_reading = '%s %s' % (reading_value_text, reading_units_text)
sensors_data.setdefault(sensor_type, {})[sensor_id] = {
'Sensor Reading': sensor_reading,
'Sensor ID': sensor_id,
'Units': reading_units_text,
}
return sensors_data
def test_get_irmc_report(self, mock_scci):
self.info['irmc_port'] = 80
self.info['irmc_auth_method'] = 'digest'
self.info['irmc_client_timeout'] = 60
mock_scci.get_report.return_value = 'get_report'
returned_mock_scci_get_report = irmc_common.get_irmc_report(self.node)
mock_scci.get_report.assert_called_with(
self.info['irmc_address'],
self.info['irmc_username'],
self.info['irmc_password'],
port=self.info['irmc_port'],
auth_method=self.info['irmc_auth_method'],
client_timeout=self.info['irmc_client_timeout'])
self.assertEqual('get_report', returned_mock_scci_get_report)
def test_get_irmc_report(self, mock_scci):
self.info['irmc_port'] = 80
self.info['irmc_auth_method'] = 'digest'
self.info['irmc_client_timeout'] = 60
mock_scci.get_report.return_value = 'get_report'
returned_mock_scci_get_report = irmc_common.get_irmc_report(self.node)
mock_scci.get_report.assert_called_with(
self.info['irmc_address'],
self.info['irmc_username'],
self.info['irmc_password'],
port=self.info['irmc_port'],
auth_method=self.info['irmc_auth_method'],
client_timeout=self.info['irmc_client_timeout'])
self.assertEqual('get_report', returned_mock_scci_get_report)
def test_get_irmc_report(self, mock_scci):
self.info["irmc_port"] = 80
self.info["irmc_auth_method"] = "digest"
self.info["irmc_client_timeout"] = 60
mock_scci.get_report.return_value = "get_report"
returned_mock_scci_get_report = irmc_common.get_irmc_report(self.node)
mock_scci.get_report.assert_called_with(
self.info["irmc_address"],
self.info["irmc_username"],
self.info["irmc_password"],
port=self.info["irmc_port"],
auth_method=self.info["irmc_auth_method"],
client_timeout=self.info["irmc_client_timeout"],
)
self.assertEqual("get_report", returned_mock_scci_get_report)
def _get_sensors_data(task):
"""Get sensors data method.
It gets sensor data from the task's node via SCCI, and convert the data
from XML to the dict format.
:param task: A TaskManager instance.
:raises: FailedToGetSensorData when getting the sensor data fails.
:returns: Returns a consistent formatted dict of sensor data grouped
by sensor type, which can be processed by Ceilometer.
"""
try:
report = irmc_common.get_irmc_report(task.node)
sensor = scci.get_sensor_data(report)
except Exception as e:
LOG.error(
_LE("SCCI get sensor data failed for node %(node_id)s " "with the following error: %(error)s"),
{"node_id": task.node.uuid, "error": e},
)
raise exception.FailedToGetSensorData(node=task.node.uuid, error=e)
sensors_data = {}
for sdr in sensor:
sensor_type_name = sdr.find("./Data/Decoded/Sensor/TypeName")
sensor_type_number = sdr.find("./Data/Decoded/Sensor/Type")
entity_name = sdr.find("./Data/Decoded/Entity/Name")
entity_id = sdr.find("./Data/Decoded/Entity/ID")
if None in (sensor_type_name, sensor_type_number, entity_name, entity_id):
continue
sensor_type = "%s (%s)" % (sensor_type_name.text, sensor_type_number.text)
sensor_id = "%s (%s)" % (entity_name.text, entity_id.text)
reading_value = sdr.find("./Data/Decoded/Sensor/Thresholds/*/Normalized")
reading_value_text = "None" if (reading_value is None) else str(reading_value.text)
reading_units = sdr.find("./Data/Decoded/Sensor/BaseUnitName")
reading_units_text = "None" if (reading_units is None) else str(reading_units.text)
sensor_reading = "%s %s" % (reading_value_text, reading_units_text)
sensors_data.setdefault(sensor_type, {})[sensor_id] = {
"Sensor Reading": sensor_reading,
"Sensor ID": sensor_id,
"Units": reading_units_text,
}
return sensors_data
def _inspect_hardware(node):
"""Inspect the node and get hardware information.
:param node: node object.
:raises: HardwareInspectionFailure, if unable to get essential
hardware properties.
:returns: a pair of dictionary and list, the dictionary contains
keys as in IRMCInspect.ESSENTIAL_PROPERTIES and its inspected
values, the list contains mac addresses.
"""
try:
report = irmc_common.get_irmc_report(node)
props = scci.get_essential_properties(
report, IRMCInspect.ESSENTIAL_PROPERTIES)
macs = _get_mac_addresses(node)
except (scci.SCCIInvalidInputError,
scci.SCCIClientError,
exception.SNMPFailure) as e:
error = (_("Inspection failed for node %(node_id)s "
"with the following error: %(error)s") %
{'node_id': node.uuid, 'error': e})
raise exception.HardwareInspectionFailure(error=error)
return (props, macs)
def _inspect_hardware(node, existing_traits=None, **kwargs):
"""Inspect the node and get hardware information.
:param node: node object.
:param existing_traits: existing traits list.
:param kwargs: the dictionary of additional parameters.
:raises: HardwareInspectionFailure, if unable to get essential
hardware properties.
:returns: a pair of dictionary and list, the dictionary contains
keys as in IRMCInspect.ESSENTIAL_PROPERTIES and its inspected
values, the list contains mac addresses.
"""
capabilities_props = set(CAPABILITIES_PROPERTIES)
new_traits = list(existing_traits) if existing_traits else []
# Remove all capabilities item which will be inspected in the existing
# capabilities of node
if 'capabilities' in node.properties:
existing_cap = node.properties['capabilities'].split(',')
for item in capabilities_props:
for prop in existing_cap:
if item == prop.split(':')[0]:
existing_cap.remove(prop)
node.properties['capabilities'] = ",".join(existing_cap)
# get gpu_ids, fpga_ids in ironic configuration
gpu_ids = [gpu_id.lower() for gpu_id in CONF.irmc.gpu_ids]
fpga_ids = [fpga_id.lower() for fpga_id in CONF.irmc.fpga_ids]
# if gpu_ids = [], pci_gpu_devices will not be inspected
if len(gpu_ids) == 0:
capabilities_props.remove('pci_gpu_devices')
# if fpga_ids = [], cpu_fpga will not be inspected
if len(fpga_ids) == 0:
capabilities_props.remove('cpu_fpga')
try:
report = irmc_common.get_irmc_report(node)
props = scci.get_essential_properties(
report, IRMCInspect.ESSENTIAL_PROPERTIES)
d_info = irmc_common.parse_driver_info(node)
capabilities = scci.get_capabilities_properties(
d_info,
capabilities_props,
gpu_ids,
fpga_ids=fpga_ids,
**kwargs)
if capabilities:
if capabilities.get('pci_gpu_devices') == 0:
capabilities.pop('pci_gpu_devices')
cpu_fpga = capabilities.pop('cpu_fpga', 0)
if cpu_fpga == 0 and 'CUSTOM_CPU_FPGA' in new_traits:
new_traits.remove('CUSTOM_CPU_FPGA')
elif cpu_fpga != 0 and 'CUSTOM_CPU_FPGA' not in new_traits:
new_traits.append('CUSTOM_CPU_FPGA')
if capabilities.get('trusted_boot') is False:
capabilities.pop('trusted_boot')
capabilities = utils.get_updated_capabilities(
node.properties.get('capabilities'), capabilities)
if capabilities:
props['capabilities'] = capabilities
macs = _get_mac_addresses(node)
except (scci.SCCIInvalidInputError,
scci.SCCIClientError,
exception.SNMPFailure) as e:
error = (_("Inspection failed for node %(node_id)s "
"with the following error: %(error)s") %
{'node_id': node.uuid, 'error': e})
raise exception.HardwareInspectionFailure(error=error)
return props, macs, new_traits
def _query_raid_config_fgi_status(self, manager, context):
"""Periodic tasks to check the progress of running RAID config."""
filters = {
'reserved': False,
'provision_state': states.CLEANWAIT,
'maintenance': False
}
fields = ['raid_config']
node_list = manager.iter_nodes(fields=fields, filters=filters)
for (node_uuid, driver, conductor_group, raid_config) in node_list:
try:
lock_purpose = 'checking async RAID configuration tasks'
with task_manager.acquire(context,
node_uuid,
purpose=lock_purpose,
shared=True) as task:
node = task.node
node_uuid = task.node.uuid
if not isinstance(task.driver.raid, IRMCRAID):
continue
if task.node.target_raid_config is None:
continue
if not raid_config or raid_config.get('fgi_status'):
continue
task.upgrade_lock()
if node.provision_state != states.CLEANWAIT:
continue
# Avoid hitting clean_callback_timeout expiration
node.touch_provisioning()
try:
report = irmc_common.get_irmc_report(node)
except client.scci.SCCIInvalidInputError:
raid_config.update({'fgi_status': RAID_FAILED})
raid_common.update_raid_info(node, raid_config)
self._set_clean_failed(task, RAID_FAILED)
continue
except client.scci.SCCIClientError:
raid_config.update({'fgi_status': RAID_FAILED})
raid_common.update_raid_info(node, raid_config)
self._set_clean_failed(task, RAID_FAILED)
continue
fgi_status_dict = _get_fgi_status(report, node_uuid)
# Note(trungnv): Allow to check until RAID mechanism to be
# completed with RAID information in report.
if fgi_status_dict == 'completing':
continue
if not fgi_status_dict:
raid_config.update({'fgi_status': RAID_FAILED})
raid_common.update_raid_info(node, raid_config)
self._set_clean_failed(task, fgi_status_dict)
continue
if all(fgi_status == 'Idle'
for fgi_status in fgi_status_dict.values()):
raid_config.update({'fgi_status': RAID_COMPLETED})
LOG.info(
'RAID configuration has completed on '
'node %(node)s with fgi_status is %(fgi)s', {
'node': node_uuid,
'fgi': RAID_COMPLETED
})
self._resume_cleaning(task)
except exception.NodeNotFound:
LOG.info(
'During query_raid_config_job_status, node '
'%(node)s was not found raid_config and presumed '
'deleted by another process.', {'node': node_uuid})
except exception.NodeLocked:
LOG.info(
'During query_raid_config_job_status, node '
'%(node)s was already locked by another process. '
'Skip.', {'node': node_uuid})
def _query_raid_config_fgi_status(self, manager, context):
"""Periodic tasks to check the progress of running RAID config."""
filters = {'reserved': False, 'provision_state': states.CLEANWAIT,
'maintenance': False}
fields = ['raid_config']
node_list = manager.iter_nodes(fields=fields, filters=filters)
for (node_uuid, driver, conductor_group, raid_config) in node_list:
try:
lock_purpose = 'checking async RAID configuration tasks'
with task_manager.acquire(context, node_uuid,
purpose=lock_purpose,
shared=True) as task:
node = task.node
node_uuid = task.node.uuid
if not isinstance(task.driver.raid, IRMCRAID):
continue
if task.node.target_raid_config is None:
continue
if not raid_config or raid_config.get('fgi_status'):
continue
task.upgrade_lock()
if node.provision_state != states.CLEANWAIT:
continue
# Avoid hitting clean_callback_timeout expiration
node.touch_provisioning()
try:
report = irmc_common.get_irmc_report(node)
except client.scci.SCCIInvalidInputError:
raid_config.update({'fgi_status': RAID_FAILED})
raid_common.update_raid_info(node, raid_config)
self._set_clean_failed(task, RAID_FAILED)
continue
except client.scci.SCCIClientError:
raid_config.update({'fgi_status': RAID_FAILED})
raid_common.update_raid_info(node, raid_config)
self._set_clean_failed(task, RAID_FAILED)
continue
fgi_status_dict = _get_fgi_status(report, node_uuid)
# Note(trungnv): Allow to check until RAID mechanism to be
# completed with RAID information in report.
if fgi_status_dict == 'completing':
continue
if not fgi_status_dict:
raid_config.update({'fgi_status': RAID_FAILED})
raid_common.update_raid_info(node, raid_config)
self._set_clean_failed(task, fgi_status_dict)
continue
if all(fgi_status == 'Idle' for fgi_status in
fgi_status_dict.values()):
raid_config.update({'fgi_status': RAID_COMPLETED})
LOG.info('RAID configuration has completed on '
'node %(node)s with fgi_status is %(fgi)s',
{'node': node_uuid, 'fgi': RAID_COMPLETED})
self._resume_cleaning(task)
except exception.NodeNotFound:
LOG.info('During query_raid_config_job_status, node '
'%(node)s was not found raid_config and presumed '
'deleted by another process.', {'node': node_uuid})
except exception.NodeLocked:
LOG.info('During query_raid_config_job_status, node '
'%(node)s was already locked by another process. '
'Skip.', {'node': node_uuid})
