The flex-gpu-device-plugin from wlbf

flex-gpu-device-plugin's People

Contributors

Watchers

flex-gpu-device-plugin's Issues

Identify pod by allocated device IDs

In this case, Device IDs in Allocate rpc request not carry actual gpu index information. We need to identify pod according to device IDs, then extract index value from nvidia.flex.com/index annotaion. Maybe from api-server api or PodResource API.

// - Allocate is expected to be called during pod creation since allocation
//   failures for any container would result in pod startup failure.
// - Allocate allows kubelet to exposes additional artifacts in a pod's
//   environment as directed by the plugin.
// - Allocate allows Device Plugin to run device specific operations on
//   the Devices requested
type AllocateRequest struct {
	ContainerRequests    []*ContainerAllocateRequest `protobuf:"bytes,1,rep,name=container_requests,json=containerRequests,proto3" json:"container_requests,omitempty"`
	XXX_NoUnkeyedLiteral struct{}                    `json:"-"`
	XXX_sizecache        int32                       `json:"-"`
}

type ContainerAllocateRequest struct {
	DevicesIDs           []string `protobuf:"bytes,1,rep,name=devices_ids,json=devicesIds,proto3" json:"devices_ids,omitempty"`
	XXX_NoUnkeyedLiteral struct{} `json:"-"`
	XXX_sizecache        int32    `json:"-"`
}