Description

Observing high temp for FPCx SENSOR CPU DIE TEMP that is currently below yellow alarm threshold.

The CPU Die Temp sensor is the sensor which measures the CPU core temperature. The temperature thresholds can be checked via the following command:
> show chassis temperature-thresholds.

As long as the temperature is bellow the alarm threshold, the value can be considered normal. If the threshold is reached and if we observe constant alarms, then it could possibly be due to issue with the sensor or the CPU is overheating.

If the alarms fails to clear, then we should check the environment in which the switch is installed, the load on the switch (maybe the switch needs to process a lot if tasks) and if it turns out that there are no problems with the environment or the load, then the defective part needs to be replaced.

Symptoms

We saw that the CPU die Temp for FPC 1 is at 61, where the normal fan speed thresholds is between 39 to 63 degree C and because of this spike, we observe yellow alarm for the sensor.  If the ambient temperature inside the chassis rises above the threshold temperature, the system shuts down automatically.

root@switch> show chassis environment no-forwarding
Class Item                           Status     Measurement
Temp 
      FPC 1 Sensor TopMiddle Ex2     OK         48 degrees C / 118 degrees F
      FPC 1 Sensor TopRight I        OK         41 degrees C / 105 degrees F
      FPC 1 Sensor TopLeft I         OK         47 degrees C / 116 degrees F
      FPC 1 Sensor TopRight E        OK         53 degrees C / 127 degrees F
      FPC 1 Sensor CPURight I        OK         47 degrees C / 116 degrees F
      FPC 1 Sensor CPULeft I         OK         49 degrees C / 120 degrees F
      FPC 1 Sensor CPU Die Temp      OK         61 degrees C / 141 degrees F

Solution

The fan modules in QFX5100 devices are hot-insertable and hot-removable field-replaceable units (FRUs). These fan modules are designed for one of the two available airflow directions (Airflow In or Airflow Out). Some fan modules are also color coded for the airflow direction as well.
 

Do not mix power supplies with different airflow. If the power supplies are color coded, ensure they are either all azure blue for airflow in models or all gold for airflow out models. If the power supplies are not color coded but have a label, ensure that the chassis is either using all "Airflow In" (AFI) or all "Airflow out" (AFO). Likewise, ensure that all fan modules have the same airflow and match the airflow of the power supplies.
 

Eg: If the fan modules have AIR IN labels, the power supplies must also have AIR IN labels; if the fan modules have AIR OUT labels, the power supplies must also have AIR OUT labels.

Mixing components with different airflows in the same chassis hampers the performance of the cooling system and leads to overheating of the chassis. 
 

You can check the status of fan modules through the command or by looking at the LEDs next to each fan module.
root@switch> show chassis alarms

Modification History

2023-06-12_FD
2023-06-26: Modified visibility

Related Information

https://www.juniper.net/documentation/us/en/hardware/qfx5100/topics/topic-map/qfx5100-cooling-system.html#id-fan-module-led-on-a-qfx5100-device