Skip to main content

Migrate a Cluster

Prerequisites​

  • Ensure that you are in the correct cluster context for kubectl.
  • If you are importing a cluster, the cluster name must be the same as the primary instance group name.
  • Before doing a workload migration, ensure that there are no unscheduled pods. If there are unscheduled pods, the migration will stop.

Option 1: Hybrid – Migrate Part of an Existing Cluster Managed by AWS Auto Scaling Groups to Ocean​

This procedure requires kops version 1.15 or later.

This section describes how to perform a hybrid migration. This enables you to migrate only some or all of the worker instance groups, while leaving the primary nodes, and any remaining worker node groups to be managed in an AWS ASG.

  1. Enable Spot support by toggling the feature flag on the management station where kops commands are run:

export KOPS_FEATURE_FLAGS="+Spotinst,SpotinstOcean,SpotinstHybrid"

  1. Ensure that Spot environment variables are defined on the kops management server:

SPOTINST_TOKEN and SPOTINST_ACCOUNT

Worker Node Migration​

In the case of multiple worker node Instance Groups:

  1. On your primary Instance Group, set the following labels:

spotinst.io/ocean-default-launchspec: "true"
spotinst.io/hybrid: "true" label

Example:

apiVersion: kops/v1alpha2
kind: InstanceGroup
metadata:
labels:
  kops.k8s.io/cluster: kops.k8s.com
  spotinst.io/hybrid: "true"
  spotinst.io/ocean-default-launchspec: "true"
name: nodes
spec:
  1. Label the rest of the instance groups you want to migrate to Ocean with the label: spotinst.io/hybrid: "true".

Example:

apiVersion: kops/v1alpha2
kind: InstanceGroup
metadata:
  creationTimestamp: null
  labels:
    kops.k8s.io/cluster: kops.ek8s.com
    spotinst.io/hybrid: "true"
  name: nodes
spec:
  1. Run kops update. kops will now identify and create the Ocean Cluster along with the Ocean launch specifications for each labeled instance group.
  2. Workload Migration. Do one of the following:
    • Spot Automated Workload Migration
    • Gradually scale down your existing Auto Scaling Groups to zero. Ocean will automatically provision the required instances to ensure your pods are scheduled.
  3. Clean up (optional):
    1. Delete imported Auto Scaling Groups.
    2. Delete all Launch Configurations.

Option 2: Migrate Primary and Worker Nodes of an Existing Cluster Managed by AWS Auto Scaling Groups to Ocean​

The migration of existing deployment will consist of two separate layers: primary nodes and worker nodes' Instance Group(s). Both will be imported.

  1. Enable Spot support by toggling the feature flag. On the management station where kops commands are run:

export KOPS_FEATURE_FLAGS="+Spotinst,SpotinstOcean"

  1. Ensure that Spot environment variables are defined on the kops management server:
  • SPOTINST_TOKEN
  • SPOTINST_ACCOUNT

Worker Nodes​

  1. In a case of a single Instance Group there is no need to add any labels and you may skip this step. Set the Ocean default launch specification label: spotinst.io/ocean-default-launchspec: true on your primary Instance Group.

Example:

apiVersion: kops/v1alpha2
kind: InstanceGroup
metadata:
labels:
  kops.k8s.io/cluster: kops.k8s.com
  spotinst.io/ocean-default-launchspec: "true"
name: nodes
spec:
  1. Run kops update. kops will now identify and create the Ocean Cluster along with the Ocean launch specifications for each labeled instance group.
  2. Workload Migration. Do one of the following:
    • Spot Automated Workload Migration
    • Gradually scale down your existing Auto Scaling Groups to zero. Ocean will automatically provision the required instances to ensure your pods are scheduled.
  3. Clean up (optional):
    1. Delete imported Auto Scaling Groups.
    2. Delete all Launch Configurations.