KubernetesLabels represents a set of labels to apply to a Kubernetes resources.

MxJob mxnet job

MxJobList is a list of etcd clusters.

AcceleratorConfig for docker container's volume and enviroment

AcceleratorVolume represents a host path that must be mounted into each container that needs to use GPUs.

ContainerName is an enum for expected containers.

ControllerConfig for docker container with GPU accelerator

EnvironmentVariableConfig for container

Interface offers a common interface for locking on arbitrary resources used in leader election. The Interface is used to hide the details on specific

JobMode mxnet job mode

LeaderCallbacks are callbacks that are triggered during certain lifecycle events of the LeaderElector. These are invoked asynchronously. possible fut

LeaderElectionRecord is the record that is stored in the leader election annotation. This information should be used for observational purposes only a

LeaderElector is a leader election client. possible future methods: * (le *LeaderElector) IsLeader() * (le *LeaderElector) GetLeader()

MXReplicaSet is a set of MX processes all acting as the same role (e.g. worker

MXReplicas is an interface for managing a set of replicas.

MXConfig is a struct representing the MXNET config. This struct is turned into an environment which is used by MXNET processes to configure themselves

MXJobClient defines an interface for working with MxJob CRDs.

MxJobCondition mxnet job condition

MxJobConditionType mxnet job condition type

MxJobCopy for MxJob

There is known issue with TPR in client-go: https://github.com/kubernetes/client-go/issues/8 Workarounds: - We include `Metadata` field in object expl

MxJobPhase mxnet job phase

MxJobRestClient uses the Kubernetes rest interface to talk to the CRD.

TODO(jlewi): Need to define the actual configuration for the MXNet MxJob. MxJobSpec mxnet job specification

MxJobStatus mxnet job status

TODO(jlewi): We probably want to add a name field. This would allow us to have more than 1 type of each worker. This might be useful if you wanted to

MxReplicaStatus mxnet replica status

MxReplicaType determines how a set of Mx processes are handled.

ReplicaState mxnet job replica state

ResourceLockConfig common data that exists across different resource locks

State mxnet job state

TODO(jlewi): We should switch a New pattern and make trainingJob private so we can ensure correctness on creation.

panicTimer panics when it reaches the given duration.