I've a ConnectX6 MCX653106A-HDA_Ax in server and when I plugged in an active cable, it gets temperature warning message: 'High temperature on sensors with bit set 0 1' . I have updated the ConnectX6 to the latest firmware 20.29.2002 but it doesn't help.

Infrastructure & Networking Adapters and Cables
1

driver: mlx5_core

version: 4.7-3.2.9

firmware-version: 20.29.2002 (MT_0000000225)

expansion-rom-version:

bus-info: 0000:03:00.0

Feb 23 14:35:58 localhost kernel: mlx5_core 0000:03:00.0: temp_warn:170:(pid 0): High temperature on sensors with bit set 0 1

Feb 23 14:35:58 localhost kernel: mlx5_core 0000:03:00.1: temp_warn:170:(pid 0): High temperature on sensors with bit set 0 1

Feb 23 14:39:08 localhost kernel: mlx5_core 0000:03:00.1: print_health_info:421:(pid 0): synd 0x10: High temperature

Feb 23 14:39:08 localhost kernel: mlx5_core 0000:03:00.0: print_health_info:421:(pid 0): synd 0x10: High temperature

2

Hello Patrick,

Thank you for posting your inquiry on the NVIDIA Networking Community.

Make sure when using an active cable, you use one of the validated and supported cables based on your installed f/w RN → https://docs.mellanox.com/display/ConnectX6Firmwarev20292002/Firmware+Compatible+Products

Also please make sure, the adapter and transceiver is getting adequate cooling by increasing the fan-speed of the system.

If you are using a validated/ supported cable and after increasing the fan speed of the system, you still are experiencing this issue, please open a support case (valid support contract needed), by sending an email to networking-support@nvidia.com

Thank you and regards,

~NVIDIA Networking Technical Support