Specifying Custom Pod Specs¶

As described in the Determined on Kubernetes guide, when tasks (e.g., experiments, notebooks) are started in a Determined cluster running on Kubernetes, the Determined master launches pods to execute these tasks. In this guide, we focus on how users can customize these pods by providing their own pod specs. Common use cases include but are not limited to: assigning pods to specific nodes, specifying additional volume mounts, attaching permissions. Please note that configuring pod specs is not required to use Determined on Kubernetes.

In this topic guide, we will cover:

How Determined uses pod specs.
The different ways to configure custom pod specs.
Supported pod spec fields.
How to configuring default pod specs.
How to configuring per-task pods specs.

How Determined Uses Pod Specs¶

All Determined tasks are launched as pods. Determined pods consists of an initContainer named determined-init-container and a container named determined-container which execute the workload. When users provide a pod spec, Determined inserts the determined-init-container and determined-container into the provided pod spec. As described below users may also configure some of the fields for the determined-container.

Ways to Configure Pod Specs¶

Determined provides two ways to configure pod specs. When Determined is installed, the system administrator can configure pod specs that are used by default for all GPU and CPU tasks. In addition, users can specify a custom pod spec for individual tasks (e.g., for an experiment by specifying environment.pod_spec in the experiment configuration). If a custom pod spec is specified for a task, it overrides the default pod spec (if any).

Supported Pod Spec Fields¶

This section describes which fields users can and cannot configure when specifying custom pod specs.

Determined does not currently support configuring:

Pod Name - Determined automatically assigns a name for every pod that is created.
Pod Namespace - Determined launches all tasks in the Namespace in which the Determined master is running.
Host Networking - This must be configured via the Master Configuration.
Restart Policy - This is always set to Never.

As part of their pod spec, users can specify initContainers and containers. Additionally users can configure the determined-container that executes the task (e.g., training), by setting the container name in the pod spec to determined-container. For the determined-container, Determined currently supports configuring:

Resource requests and limits (except GPU resources).
Volume mounts and volumes.

Configuring Default Pod Specs¶

Defining default pod specs must be done when installing or updating Determined. The default pod specs are configured in values.yaml of the Determined helm chart under taskContainerDefaults.cpuPodSpec and taskContainerDefaults.gpuPodSpec. The gpuPodSpec is applied to all tasks that use GPUs (e.g., experiments, notebooks). cpuPodSpec is applied to all tasks that only use CPUs (e.g., TensorBoards, CPU-only notebooks).

Example of configuring default pod specs:

taskContainerDefaults:
  cpuPodSpec:
    apiVersion: v1
    kind: Pod
    metadata:
      labels:
        customLabel: cpu-label
    spec:
      containers:
        # Will be applied to the container executing the task.
        - name: determined-container
          volumeMounts:
            - name: example-volume
              mountPath: /example-data
        # Custom sidecar container.
        - name: sidecar-container
          image: alpine:latest
      volumes:
        - name: example-volume
          hostPath:
            path: /data
  gpuPodSpec:
    apiVersion: v1
    kind: Pod
    metadata:
      labels:
        customLabel: gpu-label
    spec:
      containers:
        - name: determined-container
          volumeMounts:
            - name: example-volume
              mountPath: /example-data
      volumes:
        - name: example-volume
          hostPath:
            path: /data

Configuring Per-Task Pod Specs¶

In addition to default pod specs, it is also possible to configure custom pod specs for individual tasks. When defining a custom pod spec for a task, it will override the default pod spec if one is defined. Pod specs for individual tasks can be configured under the environment field in the experiment config.

Example of configuring a pod spec for an individual task:

environment:
  pod_spec:
    apiVersion: v1
    kind: Pod
    metadata:
      labels:
        customLabel: task-specific-label
    spec:
      # Specify a pull secret for task container image.
      imagePullSecrets:
        name: regcred
      # Specify a service account that allows writing checkpoints to s3 (for EKS)
      serviceAccountName: <checkpoint-storage-s3-bucket>
      # Specify tolerations for scheduling on tainted nodes
      tolerations:
        - key: "tained-nodegroup-name"
          operator: "Equal"
          value: "true"
          effect: "NoSchedule"