Skip to content

Conversation

@mahendrapaipuri
Copy link
Collaborator

  • When GPUs are configured to use in passthrough mode, there is no gurantee that nvidia-smi command will return GPUs. In that case we should not block exporter from starting. In any case, we cannot fetch metrics of passed through GPUs so it should not block exporter.

  • Simulate the case in unit test

Ref: #402 (reply in thread)

* When GPUs are configured to use in passthrough mode, there is no gurantee that `nvidia-smi` command will return GPUs. In that case we should not block exporter from starting. In any case, we cannot fetch metrics of passed through GPUs so it should not block exporter.

* Simulate the case in unit test

Signed-off-by: Mahendra Paipuri <mahendra.paipuri@gmail.com>
@mahendrapaipuri mahendrapaipuri added the bug Something isn't working label Sep 5, 2025
@mahendrapaipuri mahendrapaipuri merged commit 3e375c5 into main Sep 5, 2025
16 checks passed
@mahendrapaipuri mahendrapaipuri deleted the fix/exporter/libvirt/allow_gpu_fetch_to_fail branch September 5, 2025 14:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant