### Setup Kind Cluster and Install Volcano

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Sets up a Kind cluster, builds local images, and installs Volcano with monitoring. Use this to test Volcano from local source.

```bash
cd benchmark

# Setup: create Kind cluster, build images, install Volcano + monitoring
make setup VOLCANO_VERSION=v1.14.2

# Run gang scheduling benchmark
make test-gang-env JOBS=10 REPLICAS=100 MIN_AVAILABLE=100

# Cleanup
make cleanup-all
```

--------------------------------

### Install Argo Workflows and Submit MPI Job

Source: https://github.com/volcano-sh/volcano/blob/master/example/integrations/argo/mpi/README.md

Installs Argo Workflows using the quick-start guide, applies necessary configuration files, submits an MPI hello-world job using a workflow template, and watches its execution.

```bash
kubectl create namespace argo
kubectl apply -n argo -f "https://github.com/argoproj/argo-workflows/releases/download/v3.7.10/quick-start-minimal.yaml"
kubectl apply -f rbac.yaml
kubectl apply -f hello-world.yaml
WF_NAME=volcano-mpi-hello-$(date +%s)
argo submit --from workflowtemplate/volcano-mpi-hello -n argo --name "$WF_NAME"
argo watch "$WF_NAME" -n argo
```

--------------------------------

### Node Allocatable Resources Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_volcano_vgpu.md

Example of node allocatable resources that should include vGPU memory and number. Ensure these are present after setup.

```yaml
volcano.sh/vgpu-memory: "89424"
volcano.sh/vgpu-number: "8"
```

--------------------------------

### Apply MindSpore CPU Example Configuration

Source: https://github.com/volcano-sh/volcano/blob/master/example/MindSpore-example/README.md

Apply the Kubernetes configuration file to run the MindSpore CPU example.

```bash
kubectl apply -f mindspore-cpu.yaml
```

--------------------------------

### Kubelet Configuration Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_numa_aware.md

Example snippet of a kubelet configuration file showing static CPU management and best-effort topology management policies.

```yaml
{...}
cpuManagerPolicy: static
topologyManagerPolicy: best-effort
kubeReserved:
  cpu: 1000m

```

--------------------------------

### Install Volcano from Code

Source: https://github.com/volcano-sh/volcano/blob/master/README.md

Use this script for a one-click installation from the code base if you do not have a Kubernetes cluster. This method is temporarily only available for x86_64.

```bash
./hack/local-up-volcano.sh
```

--------------------------------

### Check MindSpore CPU Example Logs

Source: https://github.com/volcano-sh/volcano/blob/master/example/MindSpore-example/README.md

Retrieve the logs from the MindSpore CPU example pod to check the results.

```bash
kubectl logs mindspore-cpu-pod-0
```

--------------------------------

### Install Monitoring with Audit-Exporter

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Install the monitoring stack, optionally specifying a custom audit-exporter image from your private registry.

```bash
# Using the default audit-exporter image name
make install-monitoring

# Using a custom image from your private registry
make install-monitoring AUDIT_EXPORTER_IMAGE=myregistry.example.com/audit-exporter:dev
```

--------------------------------

### Detailed Dynamic Plugin Configuration Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/plugin-conf.md

An example demonstrating the dynamic configuration of actions and plugin tiers, including priority settings and specific plugin options like disabling task order for 'drf'.

```yaml
actions: "reclaim, allocate, backfill, preempt"
tiers:
- plugins:
  - name: "priority"
  - name: "gang"
- plugins:
  - name: "drf"
    disableTaskOrder: true
  - name: "predicates"
  - name: "proportion"
```

--------------------------------

### Install Volcano with Helm

Source: https://github.com/volcano-sh/volcano/blob/master/README.md

Add the Volcano Helm repository and install the official release. For development, install from the chart in the source code.

```bash
helm repo add volcano-sh https://volcano-sh.github.io/helm-charts
helm install volcano volcano-sh/volcano -n volcano-system --create-namespace
```

```bash
helm install volcano installer/helm/chart/volcano --namespace volcano-system --create-namespace
```

```bash
helm list -n volcano-system
```

--------------------------------

### Scheduler Policy Configuration Example

Source: https://github.com/volcano-sh/volcano/blob/master/pkg/controllers/sharding/policy/README.md

Example YAML configuration for defining scheduler policies, including their names, weights, and arguments.

```yaml
schedulerConfigs:
  - name: volcano
    type: volcano
    policies:
      - name: allocation-rate
        weight: 1
        arguments:
          minCPUUtil: 0.0
          maxCPUUtil: 0.6
      - name: warmup
        weight: 2                # higher weight = prefer warmup nodes first
        arguments:
          warmupLabel: "node.volcano.sh/warmup"
          warmupLabelValue: "true"
      - name: node-limit
        arguments:
          minNodes: 2
          maxNodes: 100

  - name: agent-scheduler
    type: agent
    policies:
      - name: allocation-rate
        weight: 1
        arguments:
          minCPUUtil: 0.7
          maxCPUUtil: 1.0
      - name: warmup
        weight: 5                # very strong warmup preference for agent shard
      - name: node-limit
        arguments:
          minNodes: 2
          maxNodes: 50
```

--------------------------------

### Run JobFlow Example

Source: https://github.com/volcano-sh/volcano/blob/master/example/jobflow/README.md

Deploy a JobTemplate first, then a JobFlow. Check their status using kubectl commands.

```bash
# deploy jobTemplate first
kubectl apply -f JobTemplate.yaml
# deploy jobFlow second
kubectl apply -f JobFlow.yaml

# check them
kubectl get jt
kubectl get jf
kubectl get po
```

--------------------------------

### CPU Throttle Configuration Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/cpu-throttle-design.md

Provides an example YAML configuration for the CPU throttling feature, specifying thresholds for throttling, jitter, and recovery.

```yaml
cpuThrottlingConfig:
  enable: true
  cpuThrottlingThreshold: 80    # Allow BE quota up to 80% of allocatable CPU
  cpuJitterLimitPercent: 1      # Emit updates when quota changes by >=1%
  cpuRecoverLimitPercent: 10    # Cap quota increases to 10% per update
```

--------------------------------

### Whole-Device Quota Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/capacity-dra-support.md

Use 'deviceclass/<name>' when the queue should be limited by device count. This example defines a queue with a whole-device GPU quota.

```yaml
apiVersion: scheduling.volcano.sh/v1beta1
kind: Queue
metadata:
  name: gpu-team
spec:
  reclaimable: true
  capability:
    cpu: "64"
    memory: "256Gi"
    "deviceclass/gpu.nvidia.com": "8"
  deserved:
    "deviceclass/gpu.nvidia.com": "4"
  guarantee:
    "deviceclass/gpu.nvidia.com": "1"
```

--------------------------------

### Install Volcano

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Command to install Volcano on the cluster if it's not already present. Use SKIP_INSTALL_VOLCANO=true if Volcano is pre-installed.

```bash
make install-volcano VOLCANO_VERSION=v1.14.2
```

--------------------------------

### Example TensorFlow Plugin Configuration

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/distributed-framework-plugins.md

Users can enable the TensorFlow plugin by adding a few lines to their job spec. This configuration is a minimal example.

```yaml
spec:
  plugins:
    tensorflow: []
```

--------------------------------

### Pod Example for GPU Exclusivity

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_volcano_vgpu.md

This pod example demonstrates how to apply GPU exclusivity by including labels that match a defined rule. Ensure the pod's labels align with the `GPUExclusiveRules` configured in the deviceshare plugin to receive dedicated GPUs.

```yaml
metadata:
  name: training-job
  labels:
    workloadType: training
spec:
  schedulerName: volcano
  containers:
  - name: trainer
    image: nvidia/cuda:12.0-base
    resources:
      limits:
        volcano.sh/vgpu-number: "4"
        volcano.sh/vgpu-memory: "16384"
```

--------------------------------

### Install Volcano Monitoring System

Source: https://github.com/volcano-sh/volcano/blob/master/README.md

Apply the monitoring YAML file to set up Prometheus and Grafana dashboards for Volcano after installation.

```bash
kubectl create -f installer/volcano-monitoring.yaml
```

--------------------------------

### Pod Status and Gates Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/scheduling-gates-queue-admission.md

Illustrates the state of pods and their associated gates during the scheduling process.

```text
NAME    PHASE    CONDITION          GATES
pod-1   Pending  SchedulingGated    scheduling.volcano.sh/queue-allocation-gate
pod-2   Pending  SchedulingGated    scheduling.volcano.sh/queue-allocation-gate
pod-3   Pending  SchedulingGated    scheduling.volcano.sh/queue-allocation-gate
```

```text
NAME    PHASE    CONDITION          GATES
pod-1   Running
pod-2   Pending  SchedulingGated    scheduling.volcano.sh/queue-allocation-gate
pod-3   Pending  SchedulingGated    scheduling.volcano.sh/queue-allocation-gate
```

```text
NAME    PHASE    CONDITION          GATES
pod-2   Pending  Unschedulable
pod-3   Running
```

--------------------------------

### Network Topology Discovery Configuration Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/hyperNode-auto-discovery.md

Example configuration for network topology discovery sources in `volcano-controller.conf`. Supports UFM, RoCE, and label-based discovery with customizable intervals and credentials.

```yaml
networkTopologyDiscovery:
  - source: ufm
    enabled: true
    interval: 10m
    credentials:
      secretRef:
        name: ufm-credentials
        namespace: volcano-system
    config:
      [...]
  - source: roce
    enabled: false
    interval: 15m
    config:
      [...]
  - source: label
    enabled: true
    config:
      networkTopologyTypes:
        rack-level-topology: # 2-level topology
          - nodeLabel: "volcano.sh/tor"
          - nodeLabel: "kubernetes.io/hostname"
        cluster-level-topology: # 3-level topology
          - nodeLabel: "volcano.sh/hypercluster"
          - nodeLabel: "volcano.sh/hypernode"
          - nodeLabel: "kubernetes.io/hostname"
```

--------------------------------

### HyperNode Manifest Example

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Example YAML manifest for a HyperNode, demonstrating how to define topology tiers and member selectors for real nodes.

```yaml
apiVersion: topology.volcano.sh/v1alpha1
kind: HyperNode
metadata:
  name: rack-0
spec:
  tier: 1
  tierName: rack
  members:
    - type: Node
      selector:
        labelMatch:
          matchLabels:
            topology-rack: "rack-0"

```

--------------------------------

### Example Job YAML with Multiple Pod Templates

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/job-api.md

An example YAML configuration for a Volcano Job demonstrating the use of multiple task specifications (ps and worker) with different replicas and container images.

```yaml
apiVersion: batch.volcano.sh/v1alpha1
kind: Job
metadata:
  name: tf-job
spec:
  tasks:
  - name: "ps"
    replicas: 2
    template:
      spec:
        containers:
        - name: ps
          image: ps-img
  - name: "worker"
    replicas: 5
    template:
      spec:
        containers:
        - name: worker
          image: worker-img
```

--------------------------------

### Apply MindSpore GPU Example Configuration

Source: https://github.com/volcano-sh/volcano/blob/master/example/MindSpore-example/README.md

Apply the Kubernetes configuration file to run the MindSpore GPU example using Volcano's mpimaster and mpiworker.

```bash
kubectl apply -f mindspore-gpu.yaml
```

--------------------------------

### Install Monitoring Stack

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Deploys a monitoring stack including Prometheus, kube-state-metrics, Grafana, and audit-exporter.

```bash
make install-monitoring
```

--------------------------------

### Queue CLI Commands

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/queue/queue.md

Provides examples of using the `vcctl` command-line tool to manage queues.

```APIDOC
## Queue CLI Commands

### Description
Provides examples of using the `vcctl` command-line tool to manage queues.

### Create Queue
Creates a new queue with a specified name and weight.

#### Command
```shell
$ vcctl queue create --name <queue-name> --weight <weight>
```

#### Example
```shell
$ vcctl queue create --name myqueue --weight 10
```

### View Queue
Displays the details of a specific queue.

#### Command
```shell
$ vcctl queue view <queue-name>
```

#### Example
```shell
$ vcctl queue view myqueue
```

### List Queues
Lists all available queues to the current user.

#### Command
```shell
$ vcctl queue list
```

#### Example Output
```
Name      Weight  Total  Pending  Running
myqueue   10      10     5        5
```
```

--------------------------------

### Online Pod Log Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/colocation/Overview.md

Example output from an online pod's iperf log, showing high bandwidth utilization (e.g., ~900 Mbits/sec). This indicates the workload is receiving the expected network throughput.

```shell
Connecting to host 192.168.2.30, port 5201
[  5] local 192.168.2.115 port 58492 connected to 192.168.2.30 port 5201
[ ID] Interval           Transfer     Bandwidth         
[  5]   0.00-1.00   sec   118 MBytes  990 Mbits/sec 
[  5]   1.00-2.00   sec   106 MBytes  889 Mbits/sec 
[  5]   2.00-3.00   sec   107 MBytes  897 Mbits/sec 
[  5]   3.00-4.00   sec   107 MBytes  903 Mbits/sec 
[  5]   4.00-5.00   sec   107 MBytes  899 Mbits/sec
[  5]   5.00-6.00   sec   107 MBytes  902 Mbits/sec
[  5]   6.00-7.00   sec   705 MBytes  884 Mbits/sec
...

```

--------------------------------

### Mixed Resource Queues Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/capacity-dra-support.md

A queue can include traditional resources, device-plugin resources, whole-device DRA quota, and DRA capacity dimensions simultaneously. This example demonstrates a queue with a mix of these resource types.

```yaml
apiVersion: scheduling.volcano.sh/v1beta1
kind: Queue
metadata:
  name: ml-team
spec:
  reclaimable: true
  capability:
    cpu: "100"
    memory: "200Gi"
    "nvidia.com/gpu": "4"
    "deviceclass/gpu.nvidia.com": "8"
    "cores.deviceclass/hami-core-gpu.project-hami.io": "800"
    "memory.deviceclass/hami-core-gpu.project-hami.io": "320Gi"
```

--------------------------------

### Offline Pod Log Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/colocation/Overview.md

Example output from an offline pod's iperf log, showing significantly lower bandwidth utilization (e.g., ~70 Mbits/sec). This demonstrates bandwidth suppression when online workloads are active.

```shell
Connecting to host 192.168.2.30, port 5201
[  5] local 192.168.2.115 port 44362 connected to 192.168.2.30 port 5201
[ ID] Interval           Transfer     Bandwidth         
[  5]   0.00-1.00   sec   8 MBytes  70 Mbits/sec 
[  5]   1.00-2.00   sec   12 MBytes  102 Mbits/sec 
[  5]   2.00-3.00   sec   11 MBytes  98 Mbits/sec 
[  5]   3.00-4.00   sec   11 MBytes  99 Mbits/sec 
[  5]   4.00-5.00   sec   11 MBytes  99 Mbits/sec
[  5]   5.00-6.00   sec   11 MBytes  97 Mbits/sec
[  5]   6.00-7.00   sec   11 MBytes  98 Mbits/sec
...

```

--------------------------------

### Example Hierarchy Structure

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/hdrf.md

Illustrates a hierarchical structure with queues and their weights, showing how resources might be distributed.

```text
              n
    /                   \
  n1 w=50%             n2 w=50%
(0 CPU,1 GPU)
                  /            \
                n2,1  w=50%    n2,2 w=50%
              (1 CPU,0 GPU)  (0 CPU,1 GPU)

```

--------------------------------

### Demo Job with Dedicated Volumes in YAML

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/dedicated-volume.md

An example Volcano Job demonstrating how to configure dedicated volumes using `generateName` within the `volumes` section of the task template. This setup ensures each pod gets its own PVC.

```yaml
apiVersion: batch.volcano.sh/v1alpha1
kind: Job
metadata:
  name: demo
spec:
  tasks:
  - name: task
    replicas: 2
    template:
      spec:
        containers:
        - name: ps
          image: ps-img
    volumes:
    - mountPath: /data
      generateName: train-pvc-
      volumeClaim:
        accessModes:
          - ReadWriteOnce
        resources:
          requests:
            storage: 8Gi
```

--------------------------------

### Setup for Existing Clusters

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Commands to set up the benchmarking environment on an existing Kubernetes cluster, skipping Kind cluster creation and image builds.

```bash
cd benchmark
```

```bash
# Volcano not installed yet, let the framework install it
make setup USE_EXISTING_CLUSTER=true VOLCANO_VERSION=v1.14.2
```

```bash
# Volcano already installed (e.g., a custom or modified build)
make setup USE_EXISTING_CLUSTER=true SKIP_INSTALL_VOLCANO=true
```

```bash
# Both Volcano and monitoring already deployed, only create KWOK nodes
make setup USE_EXISTING_CLUSTER=true SKIP_INSTALL_VOLCANO=true SKIP_INSTALL_MONITORING=true
```

```bash
# Cluster already has real nodes, skip KWOK entirely
make setup USE_EXISTING_CLUSTER=true SKIP_INSTALL_VOLCANO=true SKIP_KWOK=true
```

```bash
# Custom kubeconfig
KUBECONFIG=/path/to/kubeconfig make setup USE_EXISTING_CLUSTER=true SKIP_INSTALL_VOLCANO=true
```

--------------------------------

### Install Volcano Chart with --set

Source: https://github.com/volcano-sh/volcano/blob/master/installer/README.md

Use the `--set` argument with `helm install` to override default configuration values. This example sets the `basic.image_pull_policy` to `Always`, ensuring Docker images are pulled on each installation.

```bash
helm install --name volcano-release --set basic.image_pull_policy=Always volcano/volcano
```

--------------------------------

### Example Curl Command for Dynamic Log Level Update

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/dynamic-conf.md

This command demonstrates how to dynamically update the log level of the Volcano scheduler at runtime using a Unix domain socket. Ensure the socket path and parameters match your setup.

```bash
sudo curl --unix-socket /tmp/socks/klog.sock "http://localhost/setlevel?level=5&duration=60s"
```

--------------------------------

### Verify Prerequisites

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Checks if Go, Docker, Kind, kubectl, Helm, jq, and make are installed and meet the minimum version requirements.

```bash
go version && docker info > /dev/null && kind version && kubectl version --client && helm version && jq --version
```

--------------------------------

### Install Volcano

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Installs Volcano into the cluster. Can install from a specific Helm repository version or from local source.

```bash
make install-volcano VOLCANO_VERSION=v1.14.2  # from Helm repo
```

```bash
make install-volcano                          # from local source
```

--------------------------------

### Build Specific Component (Example)

Source: https://github.com/volcano-sh/volcano/blob/master/docs/development/development.md

Builds a specific component like `vc-controller-manager` after making local code changes. Ensure `HUB` and `TAG` environment variables are set.

```bash
pwd
```

```bash
export HUB=docker.io/yourrepo
export TAG=citadel
```

```bash
make vc-controller-manager
```

--------------------------------

### Task Labels Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/Network Topology Aware Scheduling.md

Provides an example of labels applied to a Task, which are used for matching against subGroupPolicy.

```yaml
labels:
  volcano.sh/partition-id: "12"
  example.com/role: worker
```

--------------------------------

### Setup Kind Cluster with Local Source

Source: https://github.com/volcano-sh/volcano/blob/master/benchmark/README.md

Builds and sets up a Kind cluster using Volcano from local source code instead of a release version.

```bash
make setup
```

--------------------------------

### Start Custom Informers in SchedulerCache

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/event-handler-framework.md

Ensure custom informers are started when the SchedulerCache begins its run cycle.

```go
func (sc *SchedulerCache) Run(stopCh <-chan struct{}) {
   go sc.podInformer.Informer().Run(stopCh)
   ......
   for _, i := range sc.CustomInformers {
      go i.Run(stopCh)
   }
   ......
}
```

--------------------------------

### Download and Deploy hami-scheduler-device ConfigMap (HAMi Core)

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_vnpu.md

Download the ConfigMap, set 'hamiVnpuCore' to true, and then deploy it. This is for the HAMi Core mode.

```bash
curl -O https://github.com/Project-HAMi/ascend-device-plugin/blob/main/ascend-device-configmap.yaml
```

```bash
kubectl apply -f ascend-device-configmap.yaml
```

--------------------------------

### vcctl queue create vs. Slurm scontrol create PartitionName

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/command-line-enhancement.md

Compares creating a queue with `vcctl` to creating a partition with `scontrol create` in Slurm.

```bash
vcctl queue create -n <queue_name> -w <weight>
```

```bash
scontrol create PartitionName=<partition_name>
```

--------------------------------

### Install Volcano Agent

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/colocation/Overview.md

Installs the Volcano agent using kubectl. This command is suitable for both x86_64 and arm64 architectures.

```shell
kubectl apply -f https://raw.githubusercontent.com/volcano-sh/volcano/master/installer/volcano-agent-development.yaml
```

--------------------------------

### Expected Plugin Configuration Output

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_resource_strategy_fit_plugin.md

Example output from checking scheduler logs, indicating the resource-strategy-fit plugin has been initialized with specific configuration parameters.

```text
Initialize resource-strategy-fit plugin with configuration: {resourceStrategyFitWeight: 10, resources: {...}}
```

--------------------------------

### Pull MindSpore GPU Example Image

Source: https://github.com/volcano-sh/volcano/blob/master/example/MindSpore-example/README.md

Use this command to pull the Docker image for the MindSpore GPU example.

```bash
docker pull lyd911/mindspore-gpu-example:0.2.0
```

--------------------------------

### Scheduler Initialization and Gate Manager Integration

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/scheduling-gates-queue-admission.md

Illustrates how the Scheduler creates, starts, and stops the SchGateManager. It also shows how the manager is passed to each scheduling session.

```go
// In Scheduler.Run():
pc.schGateManager = gate.NewSchGateManager(pc.cache.Client(), options.ServerOpts.GateRemovalWorkerNum)
pc.schGateManager.Start()
go func() {
    <-stopCh
    pc.schGateManager.Stop()
}()

// Passed to each session:
ssn := framework.OpenSession(pc.cache, plugins, configurations, pc.schGateManager)
```

--------------------------------

### Pull MindSpore CPU Example Image

Source: https://github.com/volcano-sh/volcano/blob/master/example/MindSpore-example/README.md

Use this command to pull the Docker image for the MindSpore CPU example.

```bash
docker pull lyd911/mindspore-cpu-example:0.2.0
```

--------------------------------

### View Kubelet Configuration

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_numa_aware.md

Display the current kubelet configuration file to check CPU and topology manager settings.

```bash
cat /var/lib/kubelet/config.yaml
```

--------------------------------

### Install Volcano with YAML

Source: https://github.com/volcano-sh/volcano/blob/master/README.md

Apply the provided YAML file to install Volcano on your Kubernetes cluster. This method is suitable for both x86_64 and arm64 architectures.

```bash
kubectl apply -f https://raw.githubusercontent.com/volcano-sh/volcano/master/installer/volcano-development.yaml
```

--------------------------------

### Mixed Strategy Resource Strategy Fit Configuration

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_resource_strategy_fit_plugin.md

An advanced global configuration example demonstrating a mix of LeastAllocated and MostAllocated strategies for different resources like CPU, memory, and a custom resource.

```yaml
arguments:
  resourceStrategyFitWeight: 15
  resources:
    cpu:
      type: "LeastAllocated"
      weight: 3
    memory:
      type: "MostAllocated"
      weight: 1
    example.com/custom-resource:
      type: "LeastAllocated"
      weight: 2
```

--------------------------------

### Configuration Format Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/hyperNode-auto-discovery.md

Defines the structure for network topology discovery configuration, including enabled sources and intervals. This format is loaded by the Config Loader.

```yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: network-topology-discovery
  namespace: volcano-system
data:
  config.yaml: |
    enabled: true
    interval: "30s"
    sources:
      - type: "ufm"
        address: "ufm.example.com:8080"
        auth:
          secret: "ufm-auth-secret"
      - type: "roce"
        address: "/dev/infiniband/ucm0"
      - type: "label"
        label_selector: "topology.kubernetes.io/zone"

```

--------------------------------

### Volcano Resource Configuration Example

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/resource-strategy-fit-scheduling.md

This YAML configuration demonstrates how to define resources with exact and wildcard matches, specifying allocation types and weights. It includes CPU and various GPU resource types.

```yaml
resources:
  nvidia.com/gpu-v100:     # Exact match - highest priority
    type: MostAllocated
    weight: 3
  nvidia.com/gpu/*:        # Wildcard match - covers other NVIDIA GPUs
    type: MostAllocated
    weight: 2
  amd.com/gpu/*:           # Wildcard match for AMD GPUs
    type: LeastAllocated
    weight: 2
  cpu:
    type: LeastAllocated
    weight: 1
```

--------------------------------

### Build Scheduler with Plugin Support (Ubuntu)

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/custom-plugin.md

Use this command to build the scheduler with plugin support when using an Ubuntu Linux base image.

```bash
make images SUPPORT_PLUGINS=yes
```

--------------------------------

### Install Volcano with DRA Feature Gates (kubectl)

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_enable_dra.md

Enable the 'DynamicResourceAllocation' feature gate when installing Volcano directly using 'kubectl apply -f'.

```yaml
--feature-gates=DynamicResourceAllocation=true
```

--------------------------------

### Example DRA Usage in Queue Status

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_dra_quota.md

An example of how DRA usage is reported in the status.allocated field of a queue, showing allocated resources like GPUs and cores.

```yaml
status:
  allocated:
    "deviceclass/gpu.nvidia.com": "4"
    "cores.deviceclass/hami-core-gpu.project-hami.io": "400"
```

--------------------------------

### Install Volcano with DRA Feature Gates (Helm)

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_enable_dra.md

Install Volcano using Helm, enabling the 'DynamicResourceAllocation' feature gate. Other necessary Helm values can be included as needed.

```bash
helm install volcano volcano/volcano --namespace volcano-system --create-namespace \
  --set custom.scheduler_feature_gates="DynamicResourceAllocation=true" \
  # Add other necessary Helm values for your installation
```

--------------------------------

### Get Scheduler Metrics

Source: https://github.com/volcano-sh/volcano/blob/master/docs/user-guide/how_to_use_volcano_vgpu.md

Retrieve scheduler metrics by sending a GET request to the /metrics endpoint. Ensure you replace <volcano-scheduler-ip> with the actual IP address of your Volcano scheduler.

```bash
curl http://<volcano-scheduler-ip>:8080/metrics
```

--------------------------------

### Example HyperNode CR for Spine-Leaf Topology

Source: https://github.com/volcano-sh/volcano/blob/master/docs/design/Network Topology Aware Scheduling.md

Illustrates the definition of a spine-leaf network topology using HyperNode CRs. This example shows multiple tiers representing switches and nodes.

```yaml
version: topology.volcano.sh/v1alpha1
kind: HyperNode
name: s0
spec:
  tier: 1
  members:
    - name: node0
    - name: node1
---
version: topology.volcano.sh/v1alpha1
kind: HyperNode
name: s1
spec:
  tier: 1
  members:
    - name: node2
    - name: node3
---
version: topology.volcano.sh/v1alpha1
kind: HyperNode
name: s2
spec:
  tier: 1
  members:
    - name: node4
    - name: node5
---
version: topology.volcano.sh/v1alpha1
kind: HyperNode
name: s3
spec:
  tier: 1
  members:
    - name: node6
    - name: node7
---
version: topology.volcano.sh/v1alpha1
kind: HyperNode
name: s4
spec:
  tier: 2
  members:
    - name: s0
    - name: s1
---
version: topology.volcano.sh/v1alpha1
kind: HyperNode
name: s5
spec:
  tier: 2
  members:
    - name: s2
    - name: s3
---
version: topology.volcano.sh/v1alpha1
kind: HyperNode
name: s6
spec:
  tier: 3
  members:
    - name: s4
    - name: s5
```