package
24.9.2+incompatible
Repository: https://github.com/nvidia/gpu-operator.git
Documentation: pkg.go.dev
# Packages
No description provided by the author
# Functions
ClusterRole creates ClusterRole resource.
ClusterRoleBinding creates ClusterRoleBinding resource.
ConfigMaps creates ConfigMap resource(s).
DaemonSet creates Daemonset resource.
Deployment creates Deployment resource.
GetClusterWideProxy returns cluster wide proxy object setup in OCP.
KubernetesVersion fetches the Kubernetes API server version.
OpenshiftVersion fetches OCP version.
PrometheusRule creates PrometheusRule object.
Role creates Role resource.
RoleBinding creates RoleBinding resource.
No description provided by the author
SecurityContextConstraints creates SCC resources.
Service creates Service object.
ServiceAccount creates ServiceAccount resource.
ServiceMonitor creates ServiceMonitor object.
TransformCCManager transforms CC Manager daemonset with required config as per ClusterPolicy.
TransformDCGM transforms dcgm daemonset with required config as per ClusterPolicy.
TransformDCGMExporter transforms dcgm exporter daemonset with required config as per ClusterPolicy.
TransformDevicePlugin transforms k8s-device-plugin daemonset with required config as per ClusterPolicy.
TransformDriver transforms Nvidia driver daemonset with required config as per ClusterPolicy.
TransformGPUDiscoveryPlugin transforms GPU discovery daemonset with required config as per ClusterPolicy.
TransformKataManager transforms Kata Manager daemonset with required config as per ClusterPolicy.
TransformMIGManager transforms MIG Manager daemonset with required config as per ClusterPolicy.
No description provided by the author
TransformNodeStatusExporter transforms the node-status-exporter daemonset with required config as per ClusterPolicy.
TransformSandboxDevicePlugin transforms sandbox-device-plugin daemonset with required config as per ClusterPolicy.
TransformSandboxValidator transforms nvidia-sandbox-validator daemonset with required config as per ClusterPolicy.
TransformToolkit transforms Nvidia container-toolkit daemonset with required config as per ClusterPolicy.
TransformValidator transforms nvidia-operator-validator daemonset with required config as per ClusterPolicy.
TransformValidatorComponent applies changes to given validator component.
TransformValidatorShared applies general transformations to the validator daemonset with required config as per ClusterPolicy.
TransformVFIOManager transforms VFIO-PCI Manager daemonset with required config as per ClusterPolicy.
TransformVGPUDeviceManager transforms VGPU Device Manager daemonset with required config as per ClusterPolicy.
TransformVGPUManager transforms NVIDIA vGPU Manager daemonset with required config as per ClusterPolicy.
# Constants
AppComponentLabelKey indicates the label key of the component.
AppComponentLabelValue indicates the label values of the nvidia-gpu-driver component.
CDIAnnotationPrefixEnvName is the name of the device-plugin envvar for configuring the CDI annotation prefix.
CDIEnabledEnvName is the name of the envvar used to enable CDI in the operands.
CrioConfigModeEnvName is the name of the envvar controlling how the toolkit container updates the cri-o configuration.
DCGMDefaultPort indicates default port bound to DCGM host engine.
DCGMRemoteEngineEnvName indicates env name to specify remote DCGM host engine ip:port.
DefaultCCModeEnvName is the name of the envvar for configuring default CC mode on all compatible GPUs on the node.
DefaultContainerdConfigFile indicates default config file path for containerd.
DefaultContainerdSocketFile indicates default containerd socket file.
DefaultCRIOConfigFile indicates default config file path for cri-o.
DefaultDockerConfigFile indicates default config file path for docker.
DefaultDockerSocketFile indicates default docker socket file.
DefaultDriverInstallDir represents the default path of a driver container installation.
DefaultKataArtifactsDir is the default directory to store kata artifacts on the host.
DefaultMPSRoot is the default MPS root path on the host.
DefaultRuntimeClass represents "nvidia" RuntimeClass.
DefaultRuntimeConfigTargetDir represents target directory where runtime socket dirctory will be mounted.
DefaultRuntimeSocketTargetDir represents target directory where runtime socket dirctory will be mounted.
DefaultToolkitInstallDir is the default toolkit installation directory on the host.
DeviceListStrategyEnvName is the name of the envvar for configuring the device-list-strategy in the device-plugin.
DriverInstallDirCtrPathEnvName is the name of the envvar used by the driver-validator to represent the path of the driver install dir mounted in the container.
DriverInstallDirEnvName is the name of the envvar used by the driver-validator to represent the driver install dir.
DriverInstallPathVolName represents volume name for driver install path provided to toolkit.
DriverLabelKey indicates pod label key of the driver.
DriverLabelValue indicates pod label value of the driver.
GDSEnabledEnvName is the env name to enable GDS support with device-plugin.
GPUDirectRDMAEnabledEnvName indicates if GPU direct RDMA is enabled through GPU operator.
HostRootEnvName is the name of the envvar representing the root path of the underlying host.
KataManagerAnnotationHashKey is the annotation indicating the hash of the kata-manager configuration.
Liveness probe.
MetricsConfigFileName indicates custom dcgm metrics file name.
MetricsConfigMountPath indicates mount path for custom dcgm metrics file.
MigDefaultGPUClientsConfigMapName indicates name of ConfigMap containing default gpu-clients.
MigPartedDefaultConfigMapName indicates name of ConfigMap containing default mig-parted config.
MigStrategyEnvName indicates env name for passing MIG strategy.
MOFEDEnabledEnvName is the env name to enable MOFED devices injection with device-plugin.
MPSRootEnvName is the name of the envvar for configuring the MPS root.
NLSClientTokenFileName is the NLS client config token filename.
NLSClientTokenMountPath inidicates the target mount path for NLS client config token file (.tok).
NvidiaAnnotationHashKey indicates annotation name for last applied hash by gpu-operator.
NvidiaCDIHookPathEnvName is the name of the envvar specifying the path to the 'nvidia-cdi-hook' binary.
NvidiaCtrRuntimeCDIPrefixesEnvName is the name of toolkit container env for configuring the CDI annotation prefixes.
NvidiaCtrRuntimeModeEnvName is the name of the toolkit container env for configuring the NVIDIA Container Runtime mode.
NvidiaDisableRequireEnvName is the env name to disable default cuda constraints.
OpenKernelModulesEnabledEnvName is the name of the driver-container envvar for enabling open GPU kernel module support.
PodControllerRevisionHashLabelKey is the annotation key for pod controller revision hash value.
Readiness probe.
ServiceMonitorCRDName is the name of the CRD defining the ServiceMonitor kind.
Startup probe.
ToolkitInstallDirEnvName is the name of the toolkit container env for configuring where NVIDIA Container Toolkit is installed.
TrustedCABundleFileName indicates custom user ca certificate filename.
TrustedCABundleMountDir indicates target mount directory of user ca bundle.
TrustedCACertificate indicates injected CA certificate name.
TrustedCAConfigMapName indicates configmap with custom user CA injected.
UpgradeSkipDrainLabelSelector indicates the pod selector label to skip with drain.
UseHostMOFEDEnvName indicates if MOFED driver is pre-installed on the host.
ValidatorImageEnvName indicates env name for validator image passed.
ValidatorImagePullPolicyEnvName indicates env name for validator image pull policy passed.
ValidatorImagePullSecretsEnvName indicates env name for validator image pull secrets passed.
ValidatorRuntimeClassEnvName indicates env name of runtime class to be applied to validator pods.
VgpuDMDefaultConfigMapName indicates name of ConfigMap containing default vGPU devices configuration.
VgpuDMDefaultConfigName indicates name of default configuration in the vGPU devices config file.
VGPULicensingConfigMountPath indicates target mount path for vGPU licensing configuration file.
VGPULicensingFileName is the vGPU licensing configuration filename.
VGPUTopologyConfigFileName is the vGPU topology daemon configuration filename.
VGPUTopologyConfigMountPath indicates target mount path for vGPU topology daemon configuration file.
# Variables
CertConfigPathMap indicates standard OS specific paths for ssl keys/certificates.
RepoConfigPathMap indicates standard OS specific paths for repository configuration files.
SubscriptionPathMap contains information on OS-specific paths that provide entitlements/subscription details on the host.
# Structs
ClusterPolicyController represents clusterpolicy controller spec for GPU operator.
ClusterPolicyReconciler reconciles a ClusterPolicy object.
NVIDIADriverReconciler reconciles a NVIDIADriver object.
OpenShiftDriverToolkit contains the values required to deploy OpenShift DriverToolkit DaemonSet.
OperatorMetrics defines the Prometheus metrics exposed for the operator status.
Resources indicates resources managed by GPU operator.
UpgradeReconciler reconciles Driver Daemon Sets for upgrade.
# Type aliases
ContainerProbe defines container probe types.
MountPathToVolumeSource maps a container mount path to a VolumeSource.