helpers

package
v0.0.0-...-a846cef Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Jan 26, 2026 License: Apache-2.0 Imports: 24 Imported by: 0

Documentation

Index

Constants

This section is empty.

Variables

This section is empty.

Functions

func CreateSimpleGPUPod 

func CreateSimpleGPUPod(name, namespace string, gpuLimit int) *corev1.Pod

Types

type ChartOptions 

type ChartOptions struct {
	CleanupOnFail bool
	GenerateName  bool
	ReleaseName   string
	Timeout       time.Duration
	Wait          bool
}

type ClusterPolicyClient 

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

func NewClusterPolicyClient 

func NewClusterPolicyClient(client gpuclientset.Interface) *ClusterPolicyClient

func (*ClusterPolicyClient) DisableDCGM 

func (h *ClusterPolicyClient) DisableDCGM(ctx context.Context, name string) error

func (*ClusterPolicyClient) DisableDCGMExporter 

func (h *ClusterPolicyClient) DisableDCGMExporter(ctx context.Context, name string) error

func (*ClusterPolicyClient) DisableGFD 

func (h *ClusterPolicyClient) DisableGFD(ctx context.Context, name string) error

func (*ClusterPolicyClient) EnableDCGM 

func (h *ClusterPolicyClient) EnableDCGM(ctx context.Context, name string) error

func (*ClusterPolicyClient) EnableDCGMExporter 

func (h *ClusterPolicyClient) EnableDCGMExporter(ctx context.Context, name string) error

func (*ClusterPolicyClient) EnableGFD 

func (h *ClusterPolicyClient) EnableGFD(ctx context.Context, name string) error

func (*ClusterPolicyClient) Get 

func (h *ClusterPolicyClient) Get(ctx context.Context, name string) (*nvidiav1.ClusterPolicy, error)

func (*ClusterPolicyClient) SetMIGStrategy 

func (h *ClusterPolicyClient) SetMIGStrategy(ctx context.Context, name, strategy string) error

func (*ClusterPolicyClient) Update 

func (h *ClusterPolicyClient) Update(ctx context.Context, cp *nvidiav1.ClusterPolicy) (*nvidiav1.ClusterPolicy, error)

func (*ClusterPolicyClient) UpdateDriverVersion 

func (h *ClusterPolicyClient) UpdateDriverVersion(ctx context.Context, name, version string) error

func (*ClusterPolicyClient) WaitForReady 

func (h *ClusterPolicyClient) WaitForReady(ctx context.Context, name string, timeout time.Duration) error

type DaemonSetClient 

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

func NewDaemonSetClient 

func NewDaemonSetClient(client kubernetes.Interface) *DaemonSetClient

func (*DaemonSetClient) CheckNoRestarts 

func (h *DaemonSetClient) CheckNoRestarts(ctx context.Context, namespace, name string) error

func (*DaemonSetClient) Get 

func (h *DaemonSetClient) Get(ctx context.Context, namespace, name string) (*appsv1.DaemonSet, error)

func (*DaemonSetClient) GetByLabel 

func (h *DaemonSetClient) GetByLabel(ctx context.Context, namespace, labelKey, labelValue string) (*appsv1.DaemonSet, error)

func (*DaemonSetClient) GetImage 

func (h *DaemonSetClient) GetImage(ctx context.Context, namespace, name string) (string, error)

func (*DaemonSetClient) IsReady 

func (h *DaemonSetClient) IsReady(ctx context.Context, namespace, name string) (bool, error)

func (*DaemonSetClient) WaitForReady 

func (h *DaemonSetClient) WaitForReady(ctx context.Context, namespace, name string, timeout time.Duration) error

type NodeClient 

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

func NewNodeClient 

func NewNodeClient(client kubernetes.Interface) *NodeClient

func (*NodeClient) GetNodesByLabel 

func (h *NodeClient) GetNodesByLabel(ctx context.Context, labelKey, labelValue string) ([]corev1.Node, error)

func (*NodeClient) LabelAllNodes 

func (h *NodeClient) LabelAllNodes(ctx context.Context, key, value string) error

func (*NodeClient) LabelNode 

func (h *NodeClient) LabelNode(ctx context.Context, nodeName, key, value string) error

func (*NodeClient) ListNodes 

func (h *NodeClient) ListNodes(ctx context.Context) ([]corev1.Node, error)

func (*NodeClient) UnlabelAllNodes 

func (h *NodeClient) UnlabelAllNodes(ctx context.Context, key string) error

func (*NodeClient) UnlabelNode 

func (h *NodeClient) UnlabelNode(ctx context.Context, nodeName, key string) error

type NvidiaDriverClient 

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

func NewNvidiaDriverClient 

func NewNvidiaDriverClient(client gpuclientset.Interface, k8sClient kubernetes.Interface) *NvidiaDriverClient

func (*NvidiaDriverClient) Create 

func (h *NvidiaDriverClient) Create(ctx context.Context, driver *nvidiav1alpha1.NVIDIADriver) (*nvidiav1alpha1.NVIDIADriver, error)

func (*NvidiaDriverClient) Delete 

func (h *NvidiaDriverClient) Delete(ctx context.Context, name string) error

func (*NvidiaDriverClient) Get 

func (h *NvidiaDriverClient) Get(ctx context.Context, name string) (*nvidiav1alpha1.NVIDIADriver, error)

func (*NvidiaDriverClient) List 

func (h *NvidiaDriverClient) List(ctx context.Context) (*nvidiav1alpha1.NVIDIADriverList, error)

func (*NvidiaDriverClient) Update 

func (h *NvidiaDriverClient) Update(ctx context.Context, driver *nvidiav1alpha1.NVIDIADriver) (*nvidiav1alpha1.NVIDIADriver, error)

func (*NvidiaDriverClient) UpdateDriverVersion 

func (h *NvidiaDriverClient) UpdateDriverVersion(ctx context.Context, name, version string) error

func (*NvidiaDriverClient) WaitForPodsReady 

func (h *NvidiaDriverClient) WaitForPodsReady(ctx context.Context, namespace string, timeout time.Duration) error

WaitForReady waits for the nvidia driver pods to be ready and not terminating. This checks actual pod readiness similar to check_nvidia_driver_pods_ready() in the bash tests.

func (*NvidiaDriverClient) WaitForPodsUpgradeDone 

func (h *NvidiaDriverClient) WaitForPodsUpgradeDone(ctx context.Context, name string, timeout time.Duration) error

WaitForUpgradeDone waits for the driver upgrade to complete on all GPU nodes.

type OperatorClient 

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

func NewOperatorClient 

func NewOperatorClient(opts ...OperatorClientOption) (*OperatorClient, error)

func (*OperatorClient) Install 

func (op *OperatorClient) Install(ctx context.Context, params []string, chartOpts ChartOptions) (string, error)

func (*OperatorClient) Uninstall 

func (op *OperatorClient) Uninstall(releaseName string) error

type OperatorClientOption 

type OperatorClientOption func(client *OperatorClient)

func WithChart 

func WithChart(chart string) OperatorClientOption

func WithKubeConfig 

func WithKubeConfig(kubeconfig string) OperatorClientOption

func WithNamespace 

func WithNamespace(namespace string) OperatorClientOption

type PodClient 

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

func NewPodClient 

func NewPodClient(k8sClient corev1client.CoreV1Interface) *PodClient

func (*PodClient) CreateNamespace 

func (c *PodClient) CreateNamespace(ctx context.Context, namespaceName string, labels map[string]string) (*corev1.Namespace, error)

func (*PodClient) DeleteNamespace 

func (c *PodClient) DeleteNamespace(ctx context.Context, namespaceName string) error

func (*PodClient) EnsureNoPodRestarts 

func (c *PodClient) EnsureNoPodRestarts(ctx context.Context, podName, namespace string) (bool, error)

func (*PodClient) GetPodLogs 

func (c *PodClient) GetPodLogs(ctx context.Context, pod corev1.Pod) (string, error)

func (*PodClient) GetPodsByLabel 

func (c *PodClient) GetPodsByLabel(ctx context.Context, namespace string, labelMap map[string]string) ([]corev1.Pod, error)

func (*PodClient) IsPodReady 

func (c *PodClient) IsPodReady(ctx context.Context, podName, namespace string) (bool, error)

type WorkloadClient 

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

func NewWorkloadClient 

func NewWorkloadClient(client kubernetes.Interface) *WorkloadClient

func (*WorkloadClient) Delete 

func (h *WorkloadClient) Delete(ctx context.Context, namespace, name string) error

func (*WorkloadClient) DeployPod 

func (h *WorkloadClient) DeployPod(ctx context.Context, namespace string, podSpec *corev1.Pod) (*corev1.Pod, error)

func (*WorkloadClient) GetLogs 

func (h *WorkloadClient) GetLogs(ctx context.Context, namespace, name string) (string, error)

func (*WorkloadClient) VerifyGPUAccess 

func (h *WorkloadClient) VerifyGPUAccess(ctx context.Context, namespace, name string) error

VerifyGPUAccess checks pod logs for evidence of GPU access. TODO: Improve this by exec'ing into the container and invoking nvidia-smi directly

func (*WorkloadClient) WaitForCompletion 

func (h *WorkloadClient) WaitForCompletion(ctx context.Context, namespace, name string, timeout time.Duration) error

func (*WorkloadClient) WaitForRunning 

func (h *WorkloadClient) WaitForRunning(ctx context.Context, namespace, name string, timeout time.Duration) error

Source Files

View all Source files

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.