Description

Device is showing inconsistent fan metrics (with "show chassis fan"), such as "absent", "OK" and "spinning at normal speed", with fan metrics changing within seconds. This is seemingly affecting all fans inconsistently.

Symptoms

Please check below symptoms:

Step 1: Fan Tray was showing not present.

root@switch> show chassis alarms no-forwarding

 

5 alarms currently active

Alarm time        Class Description

2023-03-10 17:13:44 UTC Major FPC0: Fan Tray 4 is not present

2023-03-10 17:13:44 UTC Major FPC0: Fan Tray 3 is not present

2023-03-10 17:13:34 UTC Major FPC0: Fan Tray 0 is not present

2023-03-10 17:13:34 UTC Major FPC0: Fan Tray 2 is not present

2023-03-10 17:13:34 UTC Major FPC0: Fan Tray 1 is not present

Step 2: Not showing fan in below in show chassis hardware.

root@switch> show chassis hardware detail no-forwarding

 

Hardware inventory:

Item       Version Part number Serial number   Description

Chassis                WHXXXXXXXXXX   QFX5200-32C-32Q

Pseudo CB 0

Routing Engine 0     BUILTIN   BUILTIN      RE-QFX5200-32C-32Q

 ada0  9316 MB QEMU         QM00001      Virtio Block Disk

 ada1  4096 MB QEMU         QM00002      Virtio Block Disk

 ada2  512 MB QEMU         QM00003      Virtio Block Disk

 ada3  1024 MB QEMU         QM00004      Virtio Block Disk

 usb0 (addr 0.1) UHCI root HUB 0   Intel       uhub0

 usb1 (addr 1.1) EHCI root HUB 0   Intel       uhub1

 usb2 (addr 1.2) product 0x0020 32  vendor 0x8087   uhub2

FPC 0      REV 34  650-059719  WHXXXXXXXXXX   QFX5200-32C-32Q

 CPU           BUILTIN   BUILTIN      FPC CPU

 PIC 0          BUILTIN   BUILTIN      32X40G/32X100G-QSFP

Power Supply 0  REV 05  740-053352  1GXXXXXXXXX    JPSU-850W-AC-AFO

Power Supply 1  REV 05  740-053352  1GXXXXXXXXX    JPSU-850W-AC-AFO

Step 3: The output of "show chassis environment is showing also FAN Tray 0,1 and 2 is not present"

root@switch> show chassis environment no-forwarding

 

Class Item              Status   Measurement

Power FPC 0 Power Supply 0      OK     28 degrees C / 82 degrees F

   FPC 0 Power Supply 1      OK     28 degrees C / 82 degrees F

Temp FPC 0 Sensor TopLeft I     OK     31 degrees C / 87 degrees F

   FPC 0 Sensor TopRight E    OK     30 degrees C / 86 degrees F

   FPC 0 Sensor TopCenter I    OK     32 degrees C / 89 degrees F

   FPC 0 Sensor TopLeft E     OK     37 degrees C / 98 degrees F

   FPC 0 Sensor CPULeft I     OK     36 degrees C / 96 degrees F

   FPC 0 Sensor CPURight I    OK     34 degrees C / 93 degrees F

   FPC 0 Sensor CPU Die Temp   OK     43 degrees C / 109 degrees F

   FPC 0 Sensor TopCenter E    OK     27 degrees C / 80 degrees F

   FPC 0 Sensor BottomRight E   OK     37 degrees C / 98 degrees F

   FPC 0 Sensor TopRight I    OK     27 degrees C / 80 degrees F

   FPC 0 Sensor BottomLeft E   OK     43 degrees C / 109 degrees F

Fans FPC 0 Fan Tray 0        Absent

   FPC 0 Fan Tray 1        Absent

   FPC 0 Fan Tray 2        Absent

   FPC 0 Fan Tray 3        OK     Spinning at normal speed

   FPC 0 Fan Tray 4        OK     Spinning at normal speed

 

Solution

  1. Please check , the dust on Fan Tray the very first thing and clean it.
  2. Perform a swap test on the fan.
  3.  Remove the problem fan tray and put it into another bay. If the fan tray does not work in the other bay, it is probably faulty and must be replaced. If the fan tray works in another bay, there is probably a problem with the power connectivity from the midplane.
  4. Check the fan power connector.
  5. If this issue is observed on multiple fan trays of the same chassis, as optional workaround (only if business can afford to), you may reboot the switch to restart all processes and to clear any potential misbehavior on the routing engine.

Conclusion: If there is no dust, a swap test on the fan didn’t help, fan power connector check didn’t help then this is hardware issue and need to contact JTAC for hardware replacement.

Modification History

05-17-2023 Initial Draft
04/11/2024 - Added reboot as optional workaround.