Skip to content

Basic operations

This section details the most common actions regarding Kubernetes clusters on Superphenix: deploying, scaling, upgrading and deleting.

Deploying a cluster

Requirements

KaaS cluster nodes are deployed in the same VPC and subnet as any IaaS virtual machines, but their controlplane runs inside a separate VPC which means that the nodes will reach out over the world wide web to access it. Therefore, make sure that your subnet has internet access when creating a cluster. This can be achieved by creating an EIP with the SNAT option enabled and attaching it to your subnet.

To deploy a cluster through the console, follow the steps of the creation wizard.
For a GitOps deployment, you can copy and adjust the example below based on your needs.

  1. Open PaaS → Kubernetes.
  2. Click Create Cluster (top right).
  3. Under General Information set:
    • Name
    • AZ: must match where you want the cluster to be deployed.
    • Cluster Version: Kubernetes version.
  4. Optionally, you can:
  5. Under Cluster Instance Group, create one or more node groups, for each of them set:
    • Number of replicas: How many nodes should be created for this group.
    • Cores: Amount of vCPUs per node.
    • RAM: Memory quantity per node.
    • Storage Class: Storage type to use for the node's disks.
    • Boot disk size: Size of the node's disks.
  6. Under Network Configuration, add one or more subnets to each node group. (Nodes need to talk to each other and should therefore share at least one common subnet between all node groups).
  7. Click Create (bottom right).
my-cluster:
  name: "my-cluster"
  location: "my-az"
  # The list of supported versions for your AZ can be found in the console's creation wizard
  kubeVersion: v1.36.3 
  workers:
    instances:
      node-group-1:
        deployment:
          replicas: 2
        template:
          version: 1
          cores: 2
          memory: 4Gi
          interfaces:
            - subnet: "<my-subnet>"
          bootDisk:
            storage: 20Gi
  kaasEssentials:
    storageClasses:
      default:
        isDefaultClass: true
        tenantClass: default
        infraClass: default
    snapshotClasses:
      default:
        tenantClass: default
        infraClass: default

Full chart values: sfs-kaas.

Accessing your cluster

Once your cluster is ready, you can download its kubeconfig file from the console:

  1. Access your cluster's page under PaaS → Kubernetes.
  2. Click Options → Download kubeconfig (top right).

Scaling a cluster

Horizontal scaling

Steps to add nodes to your cluster:

  1. Access your cluster's page under PaaS → Kubernetes.
  2. Click Options → Edit (top right).
  3. Under Cluster Instance Group you can either:
    • Increase the Number of replicas field of existing node groups.
    • Create new node groups using the Add Group button. Refer to the Deploying a cluster section for details.
  4. If you created new node groups, remember to assign them one or more subnets under the Network Configuration tab.
  5. Navigate to the last tab (Network Configuration) and click Update (bottom right).
  1. Increase <cluster-name>.workers.instances.<node-group>.deployment.replicas.
  2. Commit and push the changes.
  3. Go to the your project's page on the selfservice ArgoCD and synchronize the changes.

You can also create a new worker pool by adding its specification under <cluster-name>.workers.instances like in this example: 

node-group-2:
  deployment:
    replicas: 2
  template:
    version: 1
    cores: 2
    memory: 4Gi
    interfaces:
      - subnet: "<my-subnet>"
    bootDisk:
      storage: 30Gi

Vertical scaling

Steps to increase the size of nodes in an existing node group. This will trigger a rolling update for the given node groups and create new nodes with the updated specification.

  1. Access your cluster's page under PaaS → Kubernetes.
  2. Click Options → Edit (top right).
  3. Under Cluster Instance Group, increase the amount of Cores, RAM or adjust the Boot disk size as needed for each node group.
  4. Navigate to the last tab (Network Configuration) and click Update (bottom right).
  1. Update <cluster-name>.workers.instances.<node-group>.template.cores, .memory or .bootDisk.storage as desired.
  2. Increment <cluster-name>.workers.instances.<node-group>.template.version. This is required for the specification update to take effect.
  3. Commit and push the changes.
  4. Go to the your project's page on the selfservice ArgoCD and synchronize the changes.
Rolling update configuration

You can control how the rolling update process behaves by setting <cluster-name>.workers.instances.<node-group>.template.maxSurge and .maxUnavailable.

Example: 

node-group-1:
  deployment:
    replicas: 2
  template:
    version: 2 # <-- Increment to trigger node rolling update
    cores: 4
    memory: 12Gi
    interfaces:
      - subnet: "<my-subnet>"
    bootDisk:
      storage: 30Gi

Upgrading a cluster

To upgrade your cluster, follow these steps to trigger a controlplane upgrade. Once it is done the worker VMs will undergo a rolling update.

  1. Access your cluster's page under PaaS → Kubernetes.
  2. Click Options → Edit (top right).
  3. Under General information, pick a new Cluster Version from the list.
  4. Navigate to the last tab (Network Configuration) and click Update (bottom right).
  1. Change <cluster-name>.kubeVersion to the desired version.
  2. Commit and push the changes.
  3. Go to the your project's page on the selfservice ArgoCD and synchronize the changes.
Rolling update configuration

You can control how the rolling update process behaves by setting <cluster-name>.workers.instances.<node-group>.template.maxSurge and .maxUnavailable.

Deleting a cluster

To delete a cluster, procceed as described:

  1. Access your cluster's page under PaaS → Kubernetes.
  2. Click Options → Delete (top right).
  3. In the pop-up window, double check the cluster's name and click Confirm.
  1. Remove the cluster specification from your repository.
  2. Access your cluster's application page on the selfservice ArgoCD and click the Delete button in the top row.
  3. In the pop-up window, double check the cluster's name and click OK.