Skip to content

Commit

Permalink
PSUd changes to compute power-budget for Modular chassis (sonic-net#104)
Browse files Browse the repository at this point in the history
PSUd changes to computer power-budget for Modular chassis

HLD: sonic-net/SONiC#646

PSUd will introduce power requirements calculations. Platform APIs are introduced to provide consumers and total consumed power. Number of PSUs will help provide total supplied power

**Output of STATE-DB:**
```
  "CHASSIS_INFO|chassis_power_budget 1": {
    "expireat": 1603182970.639244,
    "ttl": -0.001,
    "type": "hash",
    "value": {
      "SUPERVISOR consumed_power": "80.0",
      "FABRIC-CARD consumed_power": "185.0",
      "FAN consumed_power": "999",
      "LINE-CARD consumed_power": "1000.0",
      "PSU supplied_power": "9000.0"
    }
  },
```
  • Loading branch information
mprabhu-nokia authored Nov 11, 2020
1 parent a14c2bb commit 05c79de
Show file tree
Hide file tree
Showing 10 changed files with 488 additions and 4 deletions.
3 changes: 2 additions & 1 deletion .gitignore
Original file line number Diff line number Diff line change
Expand Up @@ -11,8 +11,9 @@
# Compiled code which doesn't end in '.pyc'
sonic-thermalctld/scripts/thermalctldc
sonic-chassisd/scripts/chassisdc
sonic-psud/scripts/psudc

# Unit test / coverage reports
coverage.xml
.coverage
htmlcov/
htmlcov/
2 changes: 2 additions & 0 deletions sonic-psud/pytest.ini
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
[pytest]
addopts = --cov=scripts --cov-report html --cov-report term --cov-report xml
154 changes: 151 additions & 3 deletions sonic-psud/scripts/psud
Original file line number Diff line number Diff line change
Expand Up @@ -14,11 +14,21 @@ try:
import sys
import threading

from sonic_py_common import daemon_base
from swsscommon import swsscommon
from sonic_py_common import daemon_base, logger

except ImportError as e:
raise ImportError (str(e) + " - required module not found")

try:
from swsscommon import swsscommon
except ImportError as e:
from tests import mock_swsscommon as swsscommon

try:
from sonic_platform_base.device_base import DeviceBase
except ImportError as e:
from tests.mock_device_base import DeviceBase

#
# Constants ====================================================================
#
Expand All @@ -32,6 +42,12 @@ CHASSIS_INFO_TABLE = 'CHASSIS_INFO'
CHASSIS_INFO_KEY_TEMPLATE = 'chassis {}'
CHASSIS_INFO_PSU_NUM_FIELD = 'psu_num'

CHASSIS_INFO_POWER_CONSUMER_FIELD = 'Consumed Power {}'
CHASSIS_INFO_POWER_SUPPLIER_FIELD = 'Supplied Power {}'
CHASSIS_INFO_TOTAL_POWER_CONSUMED_FIELD = 'Total Consumed Power'
CHASSIS_INFO_TOTAL_POWER_SUPPLIED_FIELD = 'Total Supplied Power'
CHASSIS_INFO_POWER_KEY_TEMPLATE = 'chassis_power_budget {}'

PSU_INFO_TABLE = 'PSU_INFO'
PSU_INFO_KEY_TEMPLATE = 'PSU {}'
PSU_INFO_PRESENCE_FIELD = 'presence'
Expand Down Expand Up @@ -133,8 +149,111 @@ def log_on_status_changed(logger, normal_status, normal_log, abnormal_log):
else:
logger.log_warning(abnormal_log)


#
# PSU Chassis Info ==========================================================
#
class PsuChassisInfo(logger.Logger):

def __init__(self, log_identifier, chassis):
"""
Constructor for PsuChassisInfo
:param chassis: Object representing a platform chassis
"""
super(PsuChassisInfo, self).__init__(log_identifier)

self.chassis = chassis
self.master_status_good = True
self.total_consumed_power = 0.0
self.total_supplied_power = 0.0

def run_power_budget(self, chassis_tbl):
self.total_supplied_power = 0.0
self.total_consumed_power = 0.0
total_supplied_power = 0.0
total_fan_consumed_power = 0.0
total_module_consumed_power = 0.0

dict_index = 0
total_entries_len = 2 #For total supplied and consumed
dict_len = self.chassis.get_num_psus() +\
self.chassis.get_num_fan_drawers() +\
self.chassis.get_num_modules() + \
total_entries_len

fvs = swsscommon.FieldValuePairs(dict_len)

for index, psu in enumerate(self.chassis.get_all_psus()):
presence = try_get(psu.get_presence)
if not presence:
continue

power_good = try_get(psu.get_powergood_status)
if not power_good:
continue

name = try_get(psu.get_name, 'PSU {}'.format(index + 1))
supplied_power = try_get(psu.get_maximum_supplied_power, 0.0)
total_supplied_power = total_supplied_power + supplied_power
fvs[dict_index] = (CHASSIS_INFO_POWER_SUPPLIER_FIELD.format(name), str(supplied_power))
dict_index += 1

for index, power_consumer in enumerate(self.chassis.get_all_fan_drawers()):
presence = try_get(power_consumer.get_presence)
if not presence:
continue

name = try_get(power_consumer.get_name, 'FAN-DRAWER {}'.format(index))
fan_drawer_power = try_get(power_consumer.get_maximum_consumed_power, 0.0)
total_fan_consumed_power = total_fan_consumed_power + fan_drawer_power
fvs[dict_index] = (CHASSIS_INFO_POWER_CONSUMER_FIELD.format(name), str(fan_drawer_power))
dict_index += 1

for index, power_consumer in enumerate(self.chassis.get_all_modules()):
presence = try_get(power_consumer.get_presence)
if not presence:
continue

name = try_get(power_consumer.get_name, 'MODULE {}'.format(index))
module_power = try_get(power_consumer.get_maximum_consumed_power, 0.0)
total_module_consumed_power = total_module_consumed_power + module_power
fvs[dict_index] = (CHASSIS_INFO_POWER_CONSUMER_FIELD.format(name), str(module_power))
dict_index += 1

#Record total supplied and consumed power
self.total_supplied_power = total_supplied_power
self.total_consumed_power = total_fan_consumed_power + total_module_consumed_power

#Record in state DB in chassis table
fvs[dict_index] = (CHASSIS_INFO_TOTAL_POWER_SUPPLIED_FIELD, str(self.total_supplied_power))
fvs[dict_index + 1] = (CHASSIS_INFO_TOTAL_POWER_CONSUMED_FIELD, str(self.total_consumed_power))
chassis_tbl.set(CHASSIS_INFO_POWER_KEY_TEMPLATE.format(1), fvs)

def update_master_status(self):
if not self.total_supplied_power or not self.total_consumed_power:
if self.master_status_good is not True:
self.master_status_good = True
return False

master_status_good = (self.total_consumed_power < self.total_supplied_power)
if master_status_good == self.master_status_good:
return False

self.master_status_good = master_status_good

return True

def _set_psu_master_led(self, master_status):
try:
try:
from sonic_platform.psu import Psu
except ImportError as e:
from tests.mock_platform import MockPsu as Psu

color = DeviceBase.STATUS_LED_COLOR_GREEN if master_status else DeviceBase.STATUS_LED_COLOR_RED
Psu.set_status_master_led(color)
except NotImplementedError as e:
pass

# PSU status ===================================================================
#

Expand Down Expand Up @@ -216,6 +335,7 @@ class DaemonPsud(daemon_base.DaemonBase):
self.stop = threading.Event()
self.psu_status_dict = {}
self.fan_tbl = None
self.psu_chassis_info = None

# Signal handler
def signal_handler(self, sig, frame):
Expand Down Expand Up @@ -263,6 +383,7 @@ class DaemonPsud(daemon_base.DaemonBase):
fvs = swsscommon.FieldValuePairs([(CHASSIS_INFO_PSU_NUM_FIELD, str(psu_num))])
chassis_tbl.set(CHASSIS_INFO_KEY_TEMPLATE.format(1), fvs)


# Start main loop
self.log_info("Start daemon main loop")

Expand All @@ -271,13 +392,18 @@ class DaemonPsud(daemon_base.DaemonBase):
self.update_psu_data(psu_tbl)
self._update_led_color(psu_tbl)

if platform_chassis is not None and platform_chassis.is_modular_chassis():
self.update_psu_chassis_info(chassis_tbl)
self.update_master_led_color(chassis_tbl)

self.log_info("Stop daemon main loop")

# Delete all the information from DB and then exit
for psu_index in range(1, psu_num + 1):
psu_tbl._del(PSU_INFO_KEY_TEMPLATE.format(psu_index))

chassis_tbl._del(CHASSIS_INFO_KEY_TEMPLATE.format(1))
chassis_tbl._del(CHASSIS_INFO_POWER_KEY_TEMPLATE.format(1))

self.log_info("Shutting down...")

Expand Down Expand Up @@ -427,6 +553,28 @@ class DaemonPsud(daemon_base.DaemonBase):
])
self.fan_tbl.set(fan_name, fvs)

def update_psu_chassis_info(self, chassis_tbl):
if not platform_chassis:
return

if not self.psu_chassis_info:
self.psu_chassis_info = PsuChassisInfo(SYSLOG_IDENTIFIER, platform_chassis)

self.psu_chassis_info.run_power_budget(chassis_tbl)

def update_master_led_color(self, chassis_tbl):
if not platform_chassis or not self.psu_chassis_info:
return

psu_chassis_info = self.psu_chassis_info
if psu_chassis_info.update_master_status():
log_on_status_changed(self, psu_chassis_info.master_status_good,
'PSU supplied power warning cleared: supplied power is back to normal.',
'PSU supplied power warning: {}W supplied-power less than {}W consumed-power'.format(
psu_chassis_info.total_supplied_power, psu_chassis_info.total_consumed_power)
)
psu_chassis_info._set_psu_master_led(psu_chassis_info.master_status_good)


#
# Main =========================================================================
Expand Down
2 changes: 2 additions & 0 deletions sonic-psud/setup.cfg
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
[aliases]
test=pytest
10 changes: 10 additions & 0 deletions sonic-psud/setup.py
Original file line number Diff line number Diff line change
Expand Up @@ -10,12 +10,21 @@
url='https://github.com/Azure/sonic-platform-daemons',
maintainer='Kevin Wang',
maintainer_email='kevinw@mellanox.com',
packages=[
'tests'
],
scripts=[
'scripts/psud',
],
setup_requires= [
'pytest-runner',
'wheel'
],
tests_require = [
'pytest',
'mock>=2.0.0',
'pytest-cov'
],
classifiers=[
'Development Status :: 4 - Beta',
'Environment :: No Input/Output (Daemon)',
Expand All @@ -29,4 +38,5 @@
'Topic :: System :: Hardware',
],
keywords='sonic SONiC psu PSU daemon psud PSUD',
test_suite='setup.get_test_suite'
)
Empty file added sonic-psud/tests/__init__.py
Empty file.
11 changes: 11 additions & 0 deletions sonic-psud/tests/mock_device_base.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,11 @@
class DeviceBase():
#Device-types
DEVICE_TYPE_PSU = "PSU"
DEVICE_TYPE_FAN = "FAN"
DEVICE_TYPE_FANDRAWER = "FAN-DRAWER"

#LED colors
STATUS_LED_COLOR_GREEN = "green"
STATUS_LED_COLOR_AMBER = "amber"
STATUS_LED_COLOR_RED = "red"
STATUS_LED_COLOR_OFF = "off"
116 changes: 116 additions & 0 deletions sonic-psud/tests/mock_platform.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,116 @@
from .mock_device_base import DeviceBase

class MockDevice:
def __init__(self):
self.name = None
self.presence = True
self.model = 'Module Model'
self.serial = 'Module Serial'

def get_name(self):
return self.name

def set_presence(self, presence):
self.presence = presence

def get_presence(self):
return self.presence

def get_model(self):
return self.model

def get_serial(self):
return self.serial

class MockPsu(MockDevice):

psu_master_led_color = DeviceBase.STATUS_LED_COLOR_OFF

def __init__(self, psu_presence, psu_status, psu_name):
self.name = psu_name
self.presence = True
self.psu_status = psu_status

def get_powergood_status(self):
return self.psu_status

def set_status(self, status):
self.psu_status = status

def set_maximum_supplied_power(self, supplied_power):
self.max_supplied_power = supplied_power

def get_maximum_supplied_power(self):
return self.max_supplied_power

@classmethod
def set_status_master_led(cls, color):
cls.psu_master_led_color = color

@classmethod
def get_status_master_led(cls):
return cls.psu_master_led_color

class MockFanDrawer(MockDevice):
def __init__(self, fan_drawer_presence, fan_drawer_status, fan_drawer_name):
self.name = fan_drawer_name
self.presence = True
self.fan_drawer_status = fan_drawer_status

def get_status(self):
return self.fan_drawer_status

def set_status(self, status):
self.fan_drawer_status = status

def set_maximum_consumed_power(self, consumed_power):
self.max_consumed_power = consumed_power

def get_maximum_consumed_power(self):
return self.max_consumed_power

class MockModule(MockDevice):
def __init__(self, module_presence, module_status, module_name):
self.name = module_name
self.presence = True
self.module_status = module_status

def get_status(self):
return self.module_status

def set_status(self, status):
self.module_status = status

def set_maximum_consumed_power(self, consumed_power):
self.max_consumed_power = consumed_power

def get_maximum_consumed_power(self):
return self.max_consumed_power

class MockChassis:

def __init__(self):
self.psu_list = []
self.fan_drawer_list = []
self.module_list = []

def get_num_psus(self):
return len(self.psu_list)

def get_all_psus(self):
return self.psu_list

def get_psu(self, index):
return self.psu_list[index]

def get_num_fan_drawers(self):
return len(self.fan_drawer_list)

def get_all_fan_drawers(self):
return self.fan_drawer_list

def get_num_modules(self):
return len(self.module_list)

def get_all_modules(self):
return self.module_list
Loading

0 comments on commit 05c79de

Please sign in to comment.