Parsing of section nvidia_smi failed

Troubleshooting
,
1

CMK version: 2.4.0p10
OS version: Docker

Monitoring Windows VM on Proxmox with GPU Passthrough (NVIDIA GeForce RTX 4060 Ti)

Error message:

Parsing of section nvidia_smi failed - please submit a crash report!

Agent output for nvidia smi section: nvidia_smi_plugin_out.txt (69.3 KB)

Crash logs when checkmk tries to parse it: crash.info.txt (122.0 KB)

Error from crash info:

1 validation error for PowerReadings\npower_state\n  Input should be 'P0', 'P1', 'P2', 'P3', 'P4', 'P5', 'P6', 'P7', 'P8', 'P9', 'P10', 'P11', 'P12', 'P13', 'P14' or 'P15' [type=literal_error, input_value='Requested functionality has been deprecated', input_type=str]\n    For further information visit https://errors.pydantic.dev/2.10/v/literal_error

Thanks for your help!

2

I’ve created a PR to avoid that crash. It just adds

 "Requested functionality has been deprecated",

as a valid value for PowerState in cmk/plugins/collection/agent_based/nvidia_smi.py. We will see if this will be accepted.

1 Like
3

can confirm the same for the linux based check

4

For Linux a simple workaround could be replacing the problematic output in the agent plugin with some dummy value:

nvidia-smi -q -x | sed ‘s|<power_state>Requested functionality has been deprecated</power_state>|<power_state>P0</power_state>|’

5

For Windows the agent could be modified to something like this:

nvidia-smi.exe -q -x | foreach {$_.replace("<power_state>Requested functionality has been deprecated</power_state>","<power_state>P0</power_state>")}
1 Like
6

Confirm this workaround works on my side

image
image1352×451 51.8 KB