Skip to content

README.md

Latest commit

 

History

History
51 lines (38 loc) · 2.5 KB

README.md

File metadata and controls

51 lines (38 loc) · 2.5 KB

Quick Start Setup

Scheduling queues

A queue is an object which represents a job queue in the cluster. Queues are an essential scheduling primitive, and can reflect different scheduling guarantees, such as resource quota and priority. Queues are typically assigned to different consumers in the cluster (users, groups, or initiatives). A workload must belong to a queue in order to be scheduled. KAI Scheduler operates with two levels of hierarchical scheduling queue system.

Default Queue on Fresh Install

After installing KAI Scheduler, a two-level queue hierarchy is automatically created:

  • default-parent-queue – Top-level (parent) queue. By default, this queue has no reserved resource quotas, allowing governance of resource distribution for its leaf queues.
  • default-queue – Leaf (child) queue under the default-parent-queue top-level queue. Workloads should reference this queue.

The expected default queues are in default-queues.yaml.

No manual queue setup is required. Both queues will exist immediately after installation, allowing you to start submitting workloads right away. To customize scheduling, you can create additional queues or modify existing ones to set quotas, priorities, and hierarchies.

Creating Additional Queues

To add custom queues, apply your queue configuration:

kubectl apply -f queues.yaml

For detailed configuration options, refer to the Scheduling Queues documentation.

Pods can now be assigned to the new queue and submitted to the cluster for scheduling.

Assigning Pods to Queues

To schedule a pod using KAI Scheduler, ensure the following:

  1. Specify the queue name using the kai.scheduler/queue: default-queue label on the pod/workload.
  2. Set the scheduler name in the pod specification as kai-scheduler This ensures the pod is placed in the correct scheduling queue and managed by KAI Scheduler.

⚠️ Workload namespaces

When submitting workloads, make sure to use a dedicated namespace. Do not use the kai-scheduler namespace for workload submission.

Submitting Example Pods

CPU-Only Pods

To submit a very simple pod that requests CPU and memory resources, use the following command:

kubectl apply -f pods/cpu-only-pod.yaml

GPU Pods

Before you run the below, make sure the NVIDIA GPU-Operator is installed in the cluster.

To submit a pod that requests a GPU resource, use the following command:

kubectl apply -f pods/gpu-pod.yaml