Add vGPU support #52
Add vGPU support #52
Conversation
✱ Stainless preview buildsThis PR will update the Edit this comment to update it. It will appear in the SDK's changelogs. ✅ hypeman-typescript studio · code · diff
✅ hypeman-go studio · code · diff
⏳ These are partial results; builds are still running. This comment is auto-generated by GitHub Actions and is automatically kept up to date as you push. |
![tembo[bot]](https://avatars.githubusercontent.com/in/1223475?s=60&v=4){kind=small}
| if err != nil { | ||
| continue | ||
| } | ||
| instances, err := strconv.Atoi(strings.TrimSpace(string(data))) |
There was a problem hiding this comment.
The error from DestroyMdev is being silently ignored during cleanup. Consider logging this error to aid debugging when orphaned mdevs fail to clean up.
| instances, err := strconv.Atoi(strings.TrimSpace(string(data))) | |
| if err := DestroyMdev(mdev.UUID); err != nil { | |
| // Log but continue - best effort cleanup | |
| fmt.Fprintf(os.Stderr, "failed to destroy orphaned mdev %s: %v\n", mdev.UUID, err) | |
| continue | |
| } |
| for i, p := range parts { | ||
| if strings.HasPrefix(p, "0000:") && i+1 < len(parts) && parts[i+1] == uuid { | ||
| vfAddress = p | ||
| break |
There was a problem hiding this comment.
Consider wrapping this error with additional context about which VF was being targeted - this will help debugging when mdev creation fails in production.
| break | |
| if err := os.WriteFile(createPath, []byte(mdevUUID), 0200); err != nil { | |
| return nil, fmt.Errorf("create mdev on VF %s: %w", targetVF, err) | |
| } |
| } | ||
|
|
||
| // getProfileNameFromType resolves internal type (nvidia-556) to profile name (L40S-1Q) | ||
| func getProfileNameFromType(profileType, vfAddress string) string { |
There was a problem hiding this comment.
The error from mdevctl undefine is silently discarded. While this is "best effort", if it fails unexpectedly (e.g., mdevctl not installed), subsequent sysfs removal might also fail. Consider logging the error at debug level if mdevctl is available but fails.
| log.ErrorContext(ctx, "failed to create mdev", "profile", req.GPU.Profile, "error", err) | ||
| return nil, fmt.Errorf("create vGPU mdev for profile %s: %w", req.GPU.Profile, err) | ||
| } | ||
| gpuProfile = req.GPU.Profile |
There was a problem hiding this comment.
There's a potential race condition here: if multiple instances request the same profile concurrently, both could succeed at CreateMdev targeting the same available VF before either completes. Consider adding a mutex or using file-based locking around mdev creation to prevent this.
| }) | ||
| } | ||
|
|
||
| return vfs, nil |
There was a problem hiding this comment.
This calls ListMdevDevices() for every VF during discovery, which could result in O(n*m) operations where n is VFs and m is mdevs. Consider listing mdevs once and building a lookup map to improve performance on hosts with many VFs.
There was a problem hiding this comment.
there is some performance issue when I try to call the resources endpoint so this is probably why, needs investigating
| } | ||
|
|
||
| switch mode { | ||
| case devices.GPUModeVGPU: |
There was a problem hiding this comment.
DetectHostGPUMode() internally calls DiscoverAvailableDevices() which does filesystem I/O. Then for vGPU mode, getVGPUStatus() calls DiscoverVFs() doing more I/O. Consider caching the mode detection result or combining detection with status retrieval to reduce redundant syscalls on every /resources API call.
No description provided.