Add nodes

While working with your cluster, you may need to update your managed node group configuration to add additional nodes to support the needs of your workloads. There are many ways to scale a node group, in our case we will be using eksctl scale nodegroup command.

First let's retrieve the current nodegroup scaling configuration and look at minimum size, maximum size and desired capacity of nodes using eksctl command below:

~$eksctl get nodegroup --name $EKS_DEFAULT_MNG_NAME --cluster $EKS_CLUSTER_NAME

We'll scale the nodegroup in eks-workshop by changing the node count from 3 to 4 for desired capacity using below command:

~$eksctl scale nodegroup --name $EKS_DEFAULT_MNG_NAME --cluster $EKS_CLUSTER_NAME \

--nodes 4 --nodes-min 3 --nodes-max 6

tip

A node group can also be updated using the aws CLI with the following command. See the docs for more info.

aws eks update-nodegroup-config --cluster-name $EKS_CLUSTER_NAME \
  --nodegroup-name $EKS_DEFAULT_MNG_NAME --scaling-config minSize=4,maxSize=6,desiredSize=4

After making changes to the node group it may take up to 2-3 minutes for node provisioning and configuration changes to take effect. Let's retrieve the nodegroup configuration again and look at minimum size, maximum size and desired capacity of nodes using eksctl command below:

~$eksctl get nodegroup --name $EKS_DEFAULT_MNG_NAME --cluster $EKS_CLUSTER_NAME

To wait until the node group update operation is complete you can run this command:

~$aws eks wait nodegroup-active --cluster-name $EKS_CLUSTER_NAME --nodegroup-name $EKS_DEFAULT_MNG_NAME

Once the command above completes we can review the changed worker node count with following command, which lists all nodes in our managed node group by using the label as a filter:

tip

It can take a minute or so for the node to appear in the output below, if the list still shows 3 nodes be patient.

~$kubectl get nodes -l eks.amazonaws.com/nodegroup=$EKS_DEFAULT_MNG_NAME

NAME                                          STATUS     ROLES    AGE     VERSION

ip-10-42-104-151.us-west-2.compute.internal   Ready      <none>   2d23h   v1.29-eks-48e63af

ip-10-42-144-11.us-west-2.compute.internal    Ready      <none>   2d23h   v1.29-eks-48e63af

ip-10-42-146-166.us-west-2.compute.internal   NotReady   <none>   18s     v1.29-eks-48e63af

ip-10-42-182-134.us-west-2.compute.internal   Ready      <none>   2d23h   v1.29-eks-48e63af

Notice that the node shows a status of NotReady, which happens when the new node is still in the process of joining the cluster. We can also use kubectl wait to watch until all the nodes report Ready:

~$kubectl wait --for=condition=Ready nodes --all --timeout=300s