Stopping the cluster

Stop the cluster

First, list node scaling groups:

$ kops get ig
NAME                ROLE    MACHINETYPE MIN    MAX    ZONES
master-us-east-1a   Master  t3.medium   1   1   us-east-1a
nodes               Node    t3.medium   0   0   us-east-1a

Save, name for the master group:

$ MASTER_NAME=master-us-east-1a

Edit the group for Kubernetes worker nodes:

$ kops edit ig nodes

apiVersion: kops.k8s.io/v1alpha2
kind: InstanceGroup
metadata:
  ...
  name: nodes
spec:
  image: kope.io/k8s-1.16-debian-stretch-amd64-hvm-ebs-2020-01-17
  machineType: t3.medium
  maxSize: 0
  minSize: 0
  ...

Edit the group for Kubernetes master nodes:

$ kops edit ig $MASTER_NAME

apiVersion: kops.k8s.io/v1alpha2
kind: InstanceGroup
metadata:
  ...
  name: master-us-east-1a
spec:
  image: kope.io/k8s-1.16-debian-stretch-amd64-hvm-ebs-2020-01-17
  machineType: t3.medium
  maxSize: 0
  minSize: 0
  ...

Update the cluster:

$ kops update cluster --yes

Using cluster from kubectl context: istio-workshop.k8s.local

I0607 16:18:35.680522   18665 apply_cluster.go:556] Gossip DNS: skipping DNS validation
W0607 16:18:35.827462   18665 firewall.go:250] Opening etcd port on masters for access from the nodes, for calico.  This is unsafe in untrusted environments.
I0607 16:18:38.189527   18665 executor.go:103] Tasks: 0 done / 95 total; 46 can run
I0607 16:18:39.544345   18665 executor.go:103] Tasks: 46 done / 95 total; 24 can run
I0607 16:18:40.482241   18665 executor.go:103] Tasks: 70 done / 95 total; 21 can run
I0607 16:18:41.809540   18665 executor.go:103] Tasks: 91 done / 95 total; 3 can run
I0607 16:18:42.444500   18665 executor.go:103] Tasks: 94 done / 95 total; 1 can run
I0607 16:18:42.651568   18665 executor.go:103] Tasks: 95 done / 95 total; 0 can run
I0607 16:18:43.972348   18665 update_cluster.go:305] Exporting kubecfg for cluster
kops has set your kubectl context to istio-workshop.k8s.local

Cluster changes have been applied to the cloud.


Changes may require instances to restart: kops rolling-update cluster

Run the rolling update of the cluster:

$ kops rolling-update cluster --cloudonly

Now, open the AWS console, go to the EC2 dashboard and ensure that all instances are in terminated state.

Restart the cluster

In order to restart the stopped cluster, run the exact same steps but provide different values for minSize and maxSize parameters in the instance group settings.

For Kubernetes worker nodes:

maxSize: 2
minSize: 2

For Kubernetes master nodes:

maxSize: 1
minSize: 1

Wait 5-10 minutes until the cluster is ready and all pods are recreated. Inspect default and istio-system namespaces to ensure that are pods are up and running.