healthcheck

Documentation

Index

• type GPUDevice
• type GPUHealthChecker
  • func NewGPUHealthChecker(devices map[string]pluginapi.Device, health chan pluginapi.Device, codes []int) *GPUHealthChecker
  • func (hc *GPUHealthChecker) Start() error
  • func (hc *GPUHealthChecker) Stop()

Types

type GPUDevice

type GPUDevice struct{}

type GPUHealthChecker

type GPUHealthChecker struct {
	// contains filtered or unexported fields
}

GPUHealthChecker checks the health of nvidia GPUs. Note that with the current device naming pattern in device manager, GPUHealthChecker will not work with MIG devices.

func NewGPUHealthChecker

func NewGPUHealthChecker(devices map[string]pluginapi.Device, health chan pluginapi.Device, codes []int) *GPUHealthChecker

NewGPUHealthChecker returns a GPUHealthChecker object for a given device name

func (*GPUHealthChecker) Start

func (hc *GPUHealthChecker) Start() error

Start registers NVML events and starts listening to them

func (*GPUHealthChecker) Stop

func (hc *GPUHealthChecker) Stop()

Stop deletes the NVML events and stops the listening go routine