[DellEMC] S6000 - Thermal support - Last Reboot Reason (#4097)

- Added support for Thermal event in Last Reboot Reason "show reboot-cause" command.
- Added support for sending log message in case of thermal shutdown.

sonic NOTICE root: Shutting down due to over temperature (40 degree, 30 degree, 34 degree)
This commit is contained in:
Santhosh Kumar T 2020-02-06 18:53:27 +05:30 committed by GitHub
parent 2e0f9ca0f4
commit 74e28ad6ff
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 138 additions and 40 deletions

View File

@ -1,4 +1,5 @@
#!/bin/bash
trap 'cleanup' 0 1 2 3 6 9 10 11 13 15
LEVEL=99
INTERVAL=5
@ -14,11 +15,27 @@ LEVEL3=16000
LEVEL4=19000
LEVEL5=19000
LRR_FILE="/host/reboot-cause/reboot-cause.txt"
I2C_ADAPTER="/sys/class/i2c-adapter/i2c-2/i2c-11"
SENSOR1="$I2C_ADAPTER/11-004c/hwmon/hwmon*/temp1_input"
SENSOR2="$I2C_ADAPTER/11-004d/hwmon/hwmon*/temp1_input"
SENSOR3="$I2C_ADAPTER/11-004e/hwmon/hwmon*/temp1_input"
SENSOR1_MAX="$I2C_ADAPTER/11-004c/hwmon/hwmon*/temp1_max"
SENSOR2_MAX="$I2C_ADAPTER/11-004d/hwmon/hwmon*/temp1_max"
SENSOR3_MAX="$I2C_ADAPTER/11-004e/hwmon/hwmon*/temp1_max"
SENSOR1_MAX_VAL=$(cat $SENSOR1_MAX)
SENSOR2_MAX_VAL=$(cat $SENSOR2_MAX)
SENSOR3_MAX_VAL=$(cat $SENSOR3_MAX)
# Reducing by 63 to differentiate this temperature settings
# from pmon sensors configuration settings
SENSOR1_NEW_MAX=$(expr `echo $SENSOR1_MAX_VAL` + 5000 - 63)
SENSOR2_NEW_MAX=$(expr `echo $SENSOR2_MAX_VAL` + 5000 - 63)
SENSOR3_NEW_MAX=$(expr `echo $SENSOR3_MAX_VAL` + 5000 - 63)
# Three fan trays with each contains two separate fans
# fan1-fan4 fan2-fan5 fan3-fan6
FANTRAY1_FAN1=$I2C_ADAPTER/11-0029/fan1_target
@ -46,6 +63,14 @@ function check_module
fi
}
function cleanup
{
echo $SENSOR1_MAX_VAL > $SENSOR1_MAX
echo $SENSOR2_MAX_VAL > $SENSOR2_MAX
echo $SENSOR3_MAX_VAL > $SENSOR3_MAX
exit 1
}
function check_faulty_fan
{
@ -123,56 +148,95 @@ function update_fan_speed
function monitor_temp_sensors
{
SENSOR1_CUR_MAX_VAL=$(cat $SENSOR1_MAX)
SENSOR2_CUR_MAX_VAL=$(cat $SENSOR2_MAX)
SENSOR3_CUR_MAX_VAL=$(cat $SENSOR3_MAX)
if [ "$SENSOR1_CUR_MAX_VAL" -ne "$SENSOR1_NEW_MAX" ]
then
SENSOR1_NEW_MAX=$(expr `echo $SENSOR1_CUR_MAX_VAL` + 5000 - 63)
SENSOR1_MAX_VAL=$SENSOR1_CUR_MAX_VAL
echo $SENSOR1_NEW_MAX > $SENSOR1_MAX
fi
if [ "$SENSOR2_CUR_MAX_VAL" -ne "$SENSOR2_NEW_MAX" ]
then
SENSOR2_NEW_MAX=$(expr `echo $SENSOR2_CUR_MAX_VAL` + 5000 - 63)
SENSOR2_MAX_VAL=$SENSOR2_CUR_MAX_VAL
echo $SENSOR2_NEW_MAX > $SENSOR2_MAX
fi
if [ "$SENSOR3_CUR_MAX_VAL" -ne "$SENSOR3_NEW_MAX" ]
then
SENSOR3_NEW_MAX=$(expr `echo $SENSOR3_CUR_MAX_VAL` + 5000 - 63)
SENSOR3_MAX_VAL=$SENSOR3_CUR_MAX_VAL
echo $SENSOR3_NEW_MAX > $SENSOR3_MAX
fi
while true # go through all temp sensor outputs
do
sensor1=$(expr `echo $(cat $SENSOR1)` / 1000)
sensor2=$(expr `echo $(cat $SENSOR2)` / 1000)
sensor3=$(expr `echo $(cat $SENSOR3)` / 1000)
sum=$(($sensor1 + $sensor2 + $sensor3))
sensor_temp=$(($sum/3))
# go through all temp sensor outputs
sensor1=$(expr `echo $(cat $SENSOR1)` / 1000)
sensor2=$(expr `echo $(cat $SENSOR2)` / 1000)
sensor3=$(expr `echo $(cat $SENSOR3)` / 1000)
# All sensors output in 1000's
s1=$(cat $SENSOR1)
s2=$(cat $SENSOR2)
s3=$(cat $SENSOR3)
if [ "$sensor_temp" -le "25" ] && [ "$LEVEL" -ne "0" ]
then
# Set Fan Speed to 7000 RPM"
LEVEL=0
update_fan_speed $IDLE
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
if [ "$s1" -ge "$SENSOR1_MAX_VAL" ] || [ "$s2" -ge "$SENSOR2_MAX_VAL" ] || [ "$s3" -ge "$SENSOR3_MAX_VAL" ]
then
# Thermal trip is about to happen
echo "Thermal Overload $sensor1 $sensor2 $sensor3" > $LRR_FILE
logger "Shutting down due to over temperature ($sensor1 degree, $sensor2 degree, $sensor3 degree)"
sync
sleep 1 # Give time to send logger message to server
# Assigning the original max values back in sensors
echo $SENSOR1_MAX_VAL > $SENSOR1_MAX
echo $SENSOR2_MAX_VAL > $SENSOR2_MAX
echo $SENSOR3_MAX_VAL > $SENSOR3_MAX
elif [ "$sensor_temp" -ge "26" ] && [ "$sensor_temp" -le "44" ] && [ "$LEVEL" -ne "1" ]
then
# Set Fan Speed to 10000 RPM"
LEVEL=1
update_fan_speed $LEVEL1
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
### Not Reached ###
# In case if HW fails to shutdown
/sbin/shutdown -P now
fi
sum=$(($sensor1 + $sensor2 + $sensor3))
sensor_temp=$(($sum/3))
elif [ "$sensor_temp" -ge "45" ] && [ "$sensor_temp" -le "59" ] && [ "$LEVEL" -ne "2" ]
then
# Set Fan Speed to 13000 RPM"
LEVEL=2
update_fan_speed $LEVEL2
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
if [ "$sensor_temp" -le "25" ] && [ "$LEVEL" -ne "0" ]
then
# Set Fan Speed to 7000 RPM"
LEVEL=0
update_fan_speed $IDLE
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
elif [ "$sensor_temp" -ge "60" ] && [ "$sensor_temp" -le "79" ] && [ "$LEVEL" -ne "3" ]
then
# Set Fan Speed to 16000 RPM"
LEVEL=3
update_fan_speed $LEVEL3
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
elif [ "$sensor_temp" -ge "26" ] && [ "$sensor_temp" -le "44" ] && [ "$LEVEL" -ne "1" ]
then
# Set Fan Speed to 10000 RPM"
LEVEL=1
update_fan_speed $LEVEL1
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
elif [ "$sensor_temp" -ge "80" ] && [ "$LEVEL" -ne "4" ]
then
# Set Fan Speed to 19000 RPM"
LEVEL=4
update_fan_speed $LEVEL4
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
fi
elif [ "$sensor_temp" -ge "45" ] && [ "$sensor_temp" -le "59" ] && [ "$LEVEL" -ne "2" ]
then
# Set Fan Speed to 13000 RPM"
LEVEL=2
update_fan_speed $LEVEL2
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
elif [ "$sensor_temp" -ge "60" ] && [ "$sensor_temp" -le "79" ] && [ "$LEVEL" -ne "3" ]
then
# Set Fan Speed to 16000 RPM"
LEVEL=3
update_fan_speed $LEVEL3
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
elif [ "$sensor_temp" -ge "80" ] && [ "$LEVEL" -ne "4" ]
then
# Set Fan Speed to 19000 RPM"
LEVEL=4
update_fan_speed $LEVEL4
logger "Adjusted FAN Speed to $IDLE RPM against $sensor_temp Temperature"
fi
# Check for faulty fan
check_faulty_fan
done
}
# Check drivers for sysfs attributes

View File

@ -11,6 +11,7 @@ try:
import os
import time
import datetime
import struct
import subprocess
from sonic_platform_base.chassis_base import ChassisBase
from sonic_platform.sfp import Sfp
@ -41,6 +42,7 @@ class Chassis(ChassisBase):
reset_reason_dict = {}
reset_reason_dict[0xe] = ChassisBase.REBOOT_CAUSE_NON_HARDWARE
reset_reason_dict[0x6] = ChassisBase.REBOOT_CAUSE_NON_HARDWARE
reset_reason_dict[0x7] = ChassisBase.REBOOT_CAUSE_THERMAL_OVERLOAD_OTHER
def __init__(self):
ChassisBase.__init__(self)
@ -100,6 +102,36 @@ class Chassis(ChassisBase):
rv = rv.lstrip(" ")
return rv
def _nvram_write(self, offset, val):
resource = "/dev/nvram"
fd = os.open(resource, os.O_RDWR)
if (fd < 0):
print('File open failed ',resource)
return
if (os.lseek(fd, offset, os.SEEK_SET) != offset):
print('lseek failed on ',resource)
return
ret = os.write(fd, struct.pack('B', val))
if ret != 1:
print('Write failed ',str(ret))
return
os.close(fd)
def _get_thermal_reset(self):
reset_file = "/host/reboot-cause/reboot-cause.txt"
if (not os.path.isfile(reset_file)):
return False
try:
with open(reset_file, 'r') as fd:
rv = fd.read()
except Exception as error:
return False
if "Thermal Overload" in rv:
return True
return False
def get_name(self):
"""
Retrieves the name of the chassis
@ -181,6 +213,8 @@ class Chassis(ChassisBase):
# NVRAM. Only Warmboot and Coldboot reason are supported here.
# Since it does not support any hardware reason, we return
# non_hardware as default
if self._get_thermal_reset() == True:
self._nvram_write(0x49, 0x7)
lrr = self._get_cpld_register('last_reboot_reason')
if (lrr != 'ERR'):