Help Center/ Cloud Container Engine/ User Guide/ Scheduling/ Volcano Scheduling/ Scheduling Workloads
Updated on 2024-09-02 GMT+08:00
View PDF
Share

Scheduling Workloads

Volcano is a Kubernetes-based batch processing platform with high-performance general computing capabilities like task scheduling engine, heterogeneous chip management, and task running management. It provides end users with computing frameworks from multiple domains such as AI, big data, gene, and rendering. It also offers job scheduling, job management, and queue management for computing applications.

Kubernetes typically uses its default scheduler to schedule workloads. To use Volcano, specify Volcano for your workloads. For details about the Kubernetes scheduler, see Specify schedulers for pods.

Notes and Constraints

If you schedule a lot of workloads, Volcano will generate a significant number of logs. To prevent the node hosting Volcano from running out of disk space, use Volcano with LTS.

Using Volcano

When using Volcano to schedule workloads, you only need to configure schedulerName in the spec field of the pod and set the parameter to volcano. The following is an example: 
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx
  labels:
    app: nginx
spec:
  replicas: 4
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      annotations:
        # Submit the job to the q1 queue.
        scheduling.volcano.sh/queue-name: "q1"
        volcano.sh/preemptable: "true"
      labels:
        app: nginx
    spec:
      # Specify Volcano as the scheduler.
      schedulerName: volcano
      containers:
      - name: nginx
        image: nginx
        imagePullPolicy: IfNotPresent
        resources:
          limits:
            cpu: 1
            memory: 100Mi
          requests:
            cpu: 1
            memory: 100Mi
        ports:
        - containerPort: 80

Additionally, Volcano supports the workload queues and preemption, which can be implemented through pod annotations. The following table lists the supported annotations.

Table 1 Pod annotations supported by Volcano

Pod Annotations

Description

scheduling.volcano.sh/queue-name: "<queue-name>"

Specifies the queue to which the workload belongs. <queue-name> indicates the queue name.

volcano.sh/preemptable: "true"

Indicates whether a job can be preempted. If this function is enabled, the job can be preempted.

Options:

  • true: Preemption is enabled. This option is enabled by default.
  • false: Preemption is disabled.

You can obtain pod details to check whether the pod is scheduled by Volcano and the allocated queue.

kubectl describe pod <pod_name>

Command output:

Spec:
  Min Member:  1
  Min Resources:
    Cpu:     100m
    Memory:  100Mi
  Queue:     q1
Status:
  Conditions:
    Last Transition Time:  2023-05-30T01:54:43Z
    Reason:                tasks in gang are ready to be scheduled
    Status:                True
    Transition ID:         70be1d7d-3532-41e0-8324-c7644026b38f
    Type:                  Scheduled
  Phase:                   Running
Events:
  Type    Reason     Age              From     Message
  ----    ------     ----             ----     -------
  Normal  Scheduled  0s (x3 over 2s)  volcano  pod group is ready
Parent Topic: Volcano Scheduling