Last modified November 27, 2024

'kubectl gs template nodepool' command reference

The template nodepool command allows to create node pools, which are groups of worker nodes in a cluster sharing common configuration. The command creates a manifest for the custom resources that define a node pool. These are then meant to be applied to the management cluster, e.g. via kubectl apply.

Provider support

This command does not support Cluster API (CAPI) based workload clusters. It only supports AWS and Azure before CAPI. Adding a node pool to a CAPI workload cluster requires modification of the cluster app configuration values.

Resources generated

The resulting resources depend on the provider, set via the --provider flag.

Usage

To create the manifests for a new node pool, use this command:

kubectl gs template nodepool

Here are the supported flags:

• --provider - The infrastructure provider, either aws or azure.
• --availability-zones - list of availability zones to use, instead of setting a number. Use comma to separate values. (e.g. eu-central-1a,eu-central-1b)
• --cluster-name - Unique identifier of the cluster the node pool should be added to.
• --description - User-friendly description of the purpose of the node pool. (default Unnamed node pool)
• --nodes-max - maximum number of worker nodes for the node pool. (default 10)
• --nodes-min - minimum number of worker nodes for the node pool. (default 3)
• --output - Sets a file path to write the output to. If not set, standard output will be used.
• --organization - Name of the organization owning the cluster.
• --release - The workload cluster release version.

AWS specific

• --aws-instance-type - EC2 instance type to use for workers, e.g. m5.2xlarge. (default m5.xlarge)
• --use-alike-instance-types - Enables the use of instance types similar to the one specified via --aws-instance-type (default: false). This can increase the likelihood of getting the required instances, especially when requesting spot instances. See our reference for details.
• --on-demand-percentage-above-base-capacity - To use only on-demand instances, set this to 100. For any other value, the remainder to 100 will be filled with spot instances. For example, 50 will create a node pool that is half spot and half on-demand instances. 0 (zero) will use only spot instances. See our AWS spot instances docs for more information.
• --on-demand-base-capacity - Can be used to set a fixed number of on-demand instances, regardless of the percentage (see above) of spot vs. on-demand to be used otherwise.
• --machine-deployment-subnet: Size of the IPv4 subnet to reserve for the node pool. Must be a number between 20 and 28. For example, 24 stands for a /24 subnet with 256 addresses. Check the alpha.aws.giantswarm.io/aws-subnet-size annotation for details.

Azure specific

• --azure-vm-size - Azure VM size to use for workers (e.g. Standard_D4s_v3).
• --azure-spot-vms - Whether to use spot VMs for this node pool (defaults to false which means not to use spot VMs).
• --azure-spot-vms-max-price - Max hourly price in USD to pay for one spot VM. If the current price for the VM is exceeded, the VM is deleted. If not set, the on-demand price for the same machine size is used as limit.

Examples

kubectl gs template nodepool \
  --provider aws \
  --organization acme \
  --cluster-name a1b2c \
  --description "General purpose" \
  --availability-zones eu-central-1a \
  --aws-instance-type m5.4xlarge \
  --release 17.0.0

Output

The above example command would generate the following output:

apiVersion: cluster.x-k8s.io/v1beta1
kind: MachineDeployment
metadata:
  annotations:
    giantswarm.io/docs: https://docs.giantswarm.io/use-the-api/management-api/crd/machinedeployments.cluster.x-k8s.io/
  creationTimestamp: null
  labels:
    cluster-operator.giantswarm.io/version: 3.13.0
    cluster.x-k8s.io/cluster-name: a1b2c
    giantswarm.io/cluster: a1b2c
    giantswarm.io/machine-deployment: p3a9q
    giantswarm.io/organization: acme
    release.giantswarm.io/version: 17.0.0
  name: p3a9q
  namespace: org-acme
spec:
  clusterName: a1b2c
  selector: {}
  template:
    metadata: {}
    spec:
      bootstrap: {}
      clusterName: a1b2c
      infrastructureRef:
        apiVersion: infrastructure.giantswarm.io/v1alpha3
        kind: AWSMachineDeployment
        name: p3a9q
        namespace: org-acme
status: {}
---
apiVersion: infrastructure.giantswarm.io/v1alpha3
kind: AWSMachineDeployment
metadata:
  annotations:
    giantswarm.io/docs: https://docs.giantswarm.io/use-the-api/management-api/crd/awsmachinedeployments.infrastructure.giantswarm.io/
  creationTimestamp: null
  labels:
    aws-operator.giantswarm.io/version: 10.7.0
    cluster.x-k8s.io/cluster-name: a1b2c
    giantswarm.io/cluster: a1b2c
    giantswarm.io/machine-deployment: p3a9q
    giantswarm.io/organization: acme
    release.giantswarm.io/version: 17.0.0
  name: p3a9q
  namespace: org-acme
spec:
  nodePool:
    description: General purpose
    machine:
      dockerVolumeSizeGB: 100
      kubeletVolumeSizeGB: 100
    scaling:
      max: 10
      min: 3
  provider:
    availabilityZones:
    - eu-central-1a
    instanceDistribution:
      onDemandBaseCapacity: 0
      onDemandPercentageAboveBaseCapacity: 100
    worker:
      instanceType: m5.4xlarge
      useAlikeInstanceTypes: false
status:
  provider:
    worker: {}