Guides - Deploy and Manage a Linode Kubernetes Engine (LKE) Cluster with the Linode API

An LKE cluster can be deployed in one of several ways:

• Via the Cloud Manager
• With the Linode API (as presented in this guide)
• With the Linode CLI

These Linode-provided interfaces can be used to create, delete, and update the structural elements of your cluster, including:

• The number of nodes that make up a cluster’s node pools.
• The region where your node pools are deployed.
• The hardware resources for each node in your node pools.
• The Kubernetes version deployed to your cluster’s Master node and worker nodes.

The Kubernetes API and kubectl are the primary ways you interact with your LKE cluster once it’s been created. These tools can be used to configure, deploy, inspect, and secure your Kubernetes workloads, deploy applications, create services, configure storage and networking, and define controllers.

In this Guide

This guide covers how to use the Linode API to:

• Create an LKE cluster
• Connect kubectl to your LKE cluster
• Inspect your LKE cluster
• Modify an existing LKE cluster
• Delete an LKE cluster

Before You Begin

1. Familiarize yourself with the Linode Kubernetes Engine service. This information helps you understand the benefits and limitations of LKE.

2. Create an API Token. You need this to access the LKE service.

3. Install kubectl on your computer. You use kubectl to interact with your cluster once it’s deployed.

4. If you are new to Kubernetes, refer to our A Beginner’s Guide to Kubernetes series to learn about general Kubernetes concepts. This guide assumes a general understanding of core Kubernetes concepts.

Install kubectl

macOS:

Install via Homebrew:

brew install kubectl

If you don’t have Homebrew installed, visit the Homebrew home page for instructions. Alternatively, you can manually install the binary; visit the Kubernetes documentation for instructions.

Linux:

1. Download the latest kubectl release:

   curl -LO "https://dl.k8s.io/release/$(curl -L -s https://dl.k8s.io/release/stable.txt)/bin/linux/amd64/kubectl"

2. Make the downloaded file executable:

   chmod +x ./kubectl

3. Move the command into your PATH:

   sudo mv ./kubectl /usr/local/bin/kubectl

Windows:

Visit the Kubernetes documentation for a link to the most recent Windows release.

Create an LKE Cluster

1. To create an LKE Cluster, send a POST request to the /lke/clusters endpoint. The example below displays all possible request body parameters. Note that tags is an optional parameter.

   curl -H "Content-Type: application/json" \
         -H "Authorization: Bearer $TOKEN" \
         -X POST -d '{
           "label": "cluster12345",
           "region": "us-central",
           "k8s_version": "1.16",
           "tags": ["ecomm", "blogs"],
           "node_pools": [
             { "type": "g6-standard-2", "count": 2},
             { "type": "g6-standard-4", "count": 3}
           ]
         }' https://api.linode.com/v4/lke/clusters

   You receive a response similar to:

   {"k8s_version": "1.16", "updated": "2019-08-02T17:17:49", "region": "us-central", "tags": ["ecomm", "blogs"], "label": "cluster12345", "id": 456, "created": "2019-22-02T17:17:49"}%

2. Make note of your cluster’s ID, as you need it to continue to interact with your cluster in the next sections. In the example above, the cluster’s ID is "id": 456. You can also access your cluster’s ID by listing all LKE Clusters on your account.

   Note

   Each Linode account has a limit to the number of resources they can deploy. This includes services, like Compute Instances, NodeBalancers, Block Storage, etc. If you run into issues deploying the number of nodes you designate for a given cluster’s node pool, you may have run into a limit on the number of resources allowed on your account. Contact Linode Support if you believe this may be the case.

Connect to your LKE Cluster

Now that your LKE cluster is created, you can access and manage your cluster using kubectl on your computer. This gives you the ability to interact with the Kubernetes API, and to create and manage Kubernetes objects in your cluster.

To communicate with your LKE cluster, kubectl requires a copy of your cluster’s kubeconfig. In this section, you access the contents of your kubeconfig using the Linode API and then set up kubectl to communicate with your LKE cluster.

1. Access your LKE cluster’s kubeconfig file by sending a GET request to the /lke/clusters/{clusterId}/kubeconfig endpoint. Ensure you replace 12345 with your cluster’s ID that you recorded in the previous section:

   curl -H "Authorization: Bearer $TOKEN" \
         https://api.linode.com/v4/lke/clusters/12345/kubeconfig

   The API returns a base64 encoded string (a useful format for automated pipelines) representing your kubeconfig. Your output resembles the following:

   {"kubeconfig": "YXBpVmVyc2lvbjogdjEKY2x1c3RlcnM6Ci0gY2x1c3RlcjoKICAgIGNlcnRpZmljYXRlLWF1dGhvcml0eS1kYXRhOiBMUzB0TFMxQ1JVZEpUaUJEUlZKVVNVWkpRMEZVUlMwdExTMHRDazFKU1VONVJFTkRRV0pEWjBGM1NVSkJaMGxDUVVSQlRrSm5hM0ZvYTJsSE9YY3dRa0ZSYzBaQlJFRldUVkpOZDBWUldVUldVVkZFUlhkd2NtUlhTbXdLWTIwMWJHUkhWbnBOUWpSWVJGUkZOVTFFWjNkTmFrVXpUVlJqTVUxV2IxaEVWRWsx ... 0TFMwdExRbz0K"}%

2. Copy the kubeconfig field’s value from the response body, since you need it in the next step.

   Note

   Make sure you only copy the long string inside the quotes following "kubeconfig": in your output. Do not copy the curly braces or anything outside of them. You receive an error if you use the full output in later steps.

3. Save the base64 kubeconfig to an environment variable:

   KUBE_VAR='YXBpVmVyc2lvbjogdjEK ... 0TFMwdExRbz0K'

4. Navigate to your computer’s ~/.kube directory. This is where kubectl looks for kubeconfig files, by default.

   cd ~/.kube

5. Create a directory called configs within ~/.kube. You can use this directory to store your kubeconfig files.

   mkdir configs
   cd configs

6. Decode the contents of $KUBE_VAR and save it to a new YAML file:

   echo $KUBE_VAR | base64 -D > cluster12345-config.yaml
   Note

   The YAML file that you decode to (cluster12345-config.yaml here) can have any name of your choosing.

7. Add the kubeconfig file to your $KUBECONFIG environment variable.

   export KUBECONFIG=cluster12345-config.yaml

8. Verify that your cluster is selected as kubectl’s current context:

   kubectl config get-contexts

9. View the contents of the configuration:

   kubectl config view
   Note

   You can also access a decoded version of your kubeconfig file in the Cloud Manager.

10. View all nodes in your LKE cluster using kubectl:

    kubectl get nodes

    Your output resembles the following example, but varies depending on your own cluster’s configurations.

    NAME                      STATUS   ROLES  AGE     VERSION
    lke166-193-5d44703cd092   Ready    none   2d22h   v1.14.0
    lke166-194-5d44703cd780   Ready    none   2d22h   v1.14.0
    lke166-195-5d44703cd691   Ready    none   2d22h   v1.14.0
    lke166-196-5d44703cd432   Ready    none   2d22h   v1.14.0
    lke166-197-5d44703cd211   Ready    none   2d22h   v1.14.0

    Now that you are connected to your LKE cluster, you can begin using kubectl to deploy applications, inspect and manage cluster resources, and view logs.

Persist the Kubeconfig Context

If you create a new terminal window, it does not have access to the context that you specified using the previous instructions. This context information can be made persistent between new terminals by setting the KUBECONFIG environment variable in your shell’s configuration file.

These instructions persist the context for users of the Bash terminal. They are similar for users of other terminals:

1. Open up your Bash profile (e.g. ~/.bash_profile) in the text editor of your choice and add your configuration file to the $KUBECONFIG PATH variable.

   If an export KUBECONFIG line is already present in the file, append to the end of this line as follows; if it is not present, add this line to the end of your file:

   export KUBECONFIG=$KUBECONFIG:$HOME/.kube/config:$HOME/.kube/configs/cluster12345-config.yaml
   Note

   Alter the $HOME/.kube/configs/cluster12345-config.yaml path in the above line with the name of the file you decoded to in the previous section.

2. Close your terminal window and open a new window to receive the changes to the $KUBECONFIG variable.

3. Use the config get-contexts command for kubectl to view the available cluster contexts:

   kubectl config get-contexts

   You should see output similar to the following:

   CURRENT  NAME                         CLUSTER     AUTHINFO          NAMESPACE
   *        kubernetes-admin@kubernetes  kubernetes  kubernetes-admin

4. If your context is not already selected, (denoted by an asterisk in the current column), switch to this context using the config use-context command. Supply the full name of the cluster (including the authorized user and the cluster):

   kubectl config use-context kubernetes-admin@kubernetes

   You should see output like the following:

   Switched to context "kubernetes-admin@kubernetes".

5. You are now ready to interact with your cluster using kubectl. You can test the ability to interact with the cluster by retrieving a list of Pods in the kube-system namespace:

   kubectl get pods -n kube-system

Inspect your LKE Cluster

Once you have created an LKE Cluster, you can access information about its structural configuration using the Linode API.

List LKE Clusters

To view a list of all your LKE clusters, send a GET request to the /lke/clusters endpoint.

curl -H "Authorization: Bearer $TOKEN" https://api.linode.com/v4/lke/clusters

The returned response body displays the number of clusters deployed to your account and general details about your LKE clusters:

{"results": 2, "data": [{"updated": "2019-08-02T17:17:49", "region": "us-central", "id": 456, "k8s_version": "1.16", "label": "cluster-12345", "created": "2019-08-02T17:17:49", "tags": ["ecomm", "blogs"]}, {"updated": "2019-08-05T17:00:04", "region": "us-central", "id": 789, "k8s_version": "1.16", "label": "cluster-56789", "created": "2019-08-05T17:00:04", "tags": ["ecomm", "marketing"]}], "pages": 1, "page": 1}%

View an LKE Cluster

You can use the Linode API to access details about an individual LKE cluster. You need your cluster’s ID to access information about this resource. If you don’t know your cluster’s ID, see the List LKE Clusters section.

To view your LKE cluster, send a GET request to the /lke/clusters/{clusterId} endpoint. In this example, ensure you replace 12345 with your cluster’s ID:

curl -H "Authorization: Bearer $TOKEN" https://api.linode.com/v4/lke/clusters/12345

Your output resembles the following:

{"created": "2019-08-02T17:17:49", "updated": "2019-08-02T17:17:49", "k8s_version": "1.16", "tags": ["ecomm", "blogs"], "label": "cluster-12345", "id": 456, "region": "us-central"}%

List a Cluster’s Node Pools

A node pool consists of one or more Compute Instances (worker nodes). Each node in the pool has the same plan type. Your LKE cluster can have several node pools. Each pool is assigned its own plan type and number of nodes. To view a list of an LKE cluster’s node pools, you need your cluster’s ID. If you don’t know your cluster’s ID, see the List LKE Clusters section.

To list your cluster’s node pools, send a GET request to the /lke/clusters/{clusterId}/pools endpoint. In this example, replace 12345 with your cluster’s ID:

curl -H "Authorization: Bearer $TOKEN" https://api.linode.com/v4/lke/clusters/12345/pools

The response body includes information on each node pool’s pool ID, Compute Instance type, and node count; and each node’s individual ID and status.

{"pages": 1, "page": 1, "data": [{"count": 2, "id": 193, "type": "g6-standard-2", "linodes": [{"id": "13841932", "status": "ready "}, {"id": "13841933", "status": "ready"}]}, {"count": 3, "id": 194, "type": "g6-standard-4", "linodes": [{"id": "13841934", "status": "ready"}, {"id": "13841935", "status": "ready"}, {"id": "13841932", "status": "ready"}]}], "results": 2}%

View a Node Pool

You can use the Linode API to access details about a specific node pool in an LKE cluster. You need your cluster’s ID and node pool ID to access information about this resource. To retrieve your cluster’s ID, see the List LKE Clusters section. To find a node pool’s ID, see the List a Cluster’s Node Pools section.

To view a specific node pool, send a GET request to the /lke/clusters/{clusterId}/pools/{poolId} endpoint. In this example, replace 12345 with your cluster’s ID and 456 with the node pool’s ID:

curl -H "Authorization: Bearer $TOKEN" \
    https://api.linode.com/v4/lke/clusters/12345/pools/456

The response body provides information about the number of nodes in the node pool, the node pool’s ID, and type. You also retrieve information about each individual node in the node pool, including the Linode’s ID and status.

{"count": 2, "id": 193, "type": "g6-standard-2", "linodes": [{"id": "13841932", "status": "ready"}, {"id": "13841933", "status": "ready"}]}%

curl -H "Authorization: Bearer $TOKEN" \
    https://api.linode.com/v4/linode/instances/13841932

Modify your LKE Cluster

Once an LKE cluster is created, you can modify the cluster’s label, node pools, and tags. In this section you learn how to modify each of these parts of your cluster.

Update your LKE Cluster Label

To update your LKE cluster’s label, send a PUT request to the /lke/clusters/{clusterId} endpoint. In this example, ensure you replace 12345 with your cluster’s ID:

curl -H "Content-Type: application/json" \
        -H "Authorization: Bearer $TOKEN" \
        -X PUT -d '{
        "label": "updated-cluster-name"
        }' https://api.linode.com/v4/lke/clusters/12345

The response body displays the updated cluster label:

{"created": "2019-08-02T17:17:49", "updated": "2019-08-05T19:11:19", "k8s_version": "1.16", "tags": ["ecomm", "blogs"], "label": "updated-cluster-name", "id": 456, "region": "us-central"}%

Add a Node Pool to your LKE Cluster

A node pool consists of one or more Compute Instances (worker nodes). Each node in the pool has the same plan type and is identical to each other. Your LKE cluster can have several node pools, each pool with its own plan type and number of nodes.

You need your cluster’s ID in order to add a node pool to it. If you don’t know your cluster’s ID, see the List LKE Clusters section.

To add a node pool to an existing LKE cluster, send a POST request to the /lke/clusters/{clusterId}/pools endpoint. The request body must include the type and count parameters. In the URL of this example, ensure you replace 12345 with your own cluster’s ID:

curl -H "Content-Type: application/json" \
        -H "Authorization: Bearer $TOKEN" \
        -X POST -d '{
        "type": "g6-standard-1",
        "count": 5
        }' https://api.linode.com/v4/lke/clusters/12345/pools

The response body resembles the following:

{"count": 5, "id": 196, "type": "g6-standard-1", "linodes": [{"id": "13841945", "status": "ready"}, {"id": "13841946", "status": "ready"}, {"id": "13841947", "status": "ready"}, {"id": "13841948", "status": "ready"}, {"id": "13841949", "status": "ready"}]}%

Add Labels and Taints to your LKE Node Pools

When creating or updating an LKE node pool, you can optionally add custom labels and taints to all nodes using the labels and taints parameters. Defining labels and taints on a per-pool basis through the Linode API has several benefits compared to managing them manually with kubectl, including:

• Custom labels and taints automatically apply to new nodes when a pool is recycled or scaled up (either manually or through autoscaling).
• LKE ensures that nodes have the desired taints in place before they become ready for pod scheduling. This prevents newly created nodes from attracting workloads that don’t have the intended tolerations.

The following cURL command provides an example of using the Linode API to create a new node pool with a custom taint and label. If you are copying this command to run on your own LKE cluster, replace 12345 with the ID of your LKE cluster.

curl -H "Content-Type: application/json" \
        -H "Authorization: Bearer $TOKEN" \
        -X POST -d '{
        "type": "g6-standard-1",
        "count": 3,
        "taints": [
            {
                "key": "myapp.io/app",
                "value": "test",
                "effect": "NoSchedule"
            }
        ],
        "labels": {
            "myapp.io/app": "test"
        }
        }' https://api.linode.com/v4/lke/clusters/12345/pools

In the above command, labels are defined in the labels field as key-value pairs within a single object. Taints are defined as an array of dictionary objects in the taints field.

• Labels: The labels field expects a dictionary object with one or more key-value pairs. These key-value pairs should adhere to the specifications and restrictions outlined in the Kubernetes Labels and Selectors documentation.

  "labels": {
      "myapp.io/app": "test"
  }

  A label’s key and value must begin with a letter or number, and may contain letters, numbers, hyphens, dots, and underscores, up to 63 characters each. Optionally, the key can begin with a valid DNS subdomain prefix and a single slash (/). In this case, the maximum allowed length of the domain prefix is 253 characters. For instance, example.com/my-app is a valid key for a label.

• Taints: The taints field expects an array of one or more dictionary objects, adhering to the guidelines outlined in the Kubernetes Taints and Tolerations documentation. A taint consists of a key, value, and effect:

  "taints": [
      {
          "key": "myapp.io/app",
          "value": "test",
          "effect": "NoSchedule"
      }
  ]
  • Key: The key value must begin with a letter or number, and may contain letters, numbers, hyphens, dots, and underscores, up to 253 characters. Optionally, the key value can begin with a DNS subdomain prefix and a single slash (/), like example.com/my-app. In this case the maximum allowed length of the domain prefix is 253 characters.
  • Value: The value key is optional. If given, it must begin with a letter or number, and may contain letters, numbers, hyphens, dots, and underscores, up to 63 characters.
  • Effect: The effect value must be NoSchedule, PreferNoSchedule, or NoExecute.

You can also add, edit, or remove labels and taints on existing node pools using the Linode API. The example cURL command below demonstrates how to remove taints and update the labels on an existing node pool. If you are copying this command to run on your own LKE cluster, replace 12345 with the ID of your LKE cluster and 196 with the ID of your node pool.

curl -H "Content-Type: application/json" \
    -H "Authorization: Bearer $TOKEN" \
    -X PUT -d '{
        "type": "g6-standard-1",
        "count": 3,
        "taints": [],
        "labels": {
            "myapp.io/app": "prod",
            "example": "foo",
        }
    }' https://api.linode.com/v4/lke/clusters/12345/pools/196

The above command results in the following changes to the node pool, assuming the labels and taints were originally entered as shown in the first create command.

• Removes the “myapp.io/app” taint by specifying an empty array in the taint field.
• Changes the label “myapp.io/app” to have a value of “prod” instead of “test”.
• Adds the new label “example=foo”.

Resize your LKE Node Pool

You can resize an LKE cluster’s node pool to add or decrease its number of nodes. You need your cluster’s ID and the node pool’s ID in order to resize it. If you don’t know your cluster’s ID, see the List LKE Clusters section. If you don’t know your node pool’s ID, see the List a Cluster’s Node Pools section.

To update your node pool’s node count, send a PUT request to the /lke/clusters/{clusterId}/pools/{poolId} endpoint. In the URL of this example, replace 12345 with your cluster’s ID and 196 with your node pool’s ID:

curl -H "Content-Type: application/json" \
    -H "Authorization: Bearer $TOKEN" \
    -X PUT -d '{
        "type": "g6-standard-4",
        "count": 6
    }' https://api.linode.com/v4/lke/clusters/12345/pools/196

Recycle All Nodes Within a Cluster

You can recycle all nodes within an LKE cluster to upgrade the nodes to the most recent patch of the cluster’s Kubernetes version and to otherwise replace the Compute Instances that comprise the cluster. Nodes are recycled on a rolling basis, meaning that only one node is down at a time throughout the recycling process. You need your cluster’s ID in order to recycle it’s nodes. If you don’t know your cluster’s ID, see the List LKE Clusters section.

To recycle all nodes within a cluster, send a POST request to the /lke/clusters/{clusterId}/pools/{poolId}/recycle endpoint. In the URL of this example, replace 12345 with your cluster’s ID:

curl -H "Content-Type: application/json" \
    -H "Authorization: Bearer $TOKEN" \
    -X POST \
    https://api.linode.com/v4/lke/clusters/12345/recycle

Recycle your LKE Node Pool

You can recycle an LKE cluster’s node pool to upgrade its nodes to the most recent patch of the cluster’s Kubernetes version. Nodes are recycled on a rolling basis, meaning that only one node is down at a time throughout the recycling process. You need your cluster’s ID and the node pool’s ID in order to recycle it. If you don’t know your cluster’s ID, see the List LKE Clusters section. If you don’t know your node pool’s ID, see the List a Cluster’s Node Pools section.

To recycle your node pool, send a POST request to the /lke/clusters/{clusterId}/pools/{poolId}/recycle endpoint. In the URL of this example, replace 12345 with your cluster’s ID and 196 with your node pool’s ID:

curl -H "Content-Type: application/json" \
    -H "Authorization: Bearer $TOKEN" \
    -X POST \
    https://api.linode.com/v4/lke/clusters/12345/pools/196/recycle

Recycle a Single Node within a Node Pool

You can recycle an individual node within a LKE Cluster’s Node Pool. You need your cluster’s ID and the node ID in order to recycle it. If you don’t know your cluster’s ID, see the List LKE Clusters section. If you don’t know your node ID, see the List a Cluster’s Node Pools section.

To recycle your node, send a POST request to the /lke/clusters/{clusterId}/nodes/{nodeId}/recycle endpoint. In the URL of this example, replace 12345 with your cluster’s ID and 12345-6aa78910bc with your node ID:

curl -H "Authorization: Bearer $TOKEN" \
    https://api.linode.com/v4/lke/clusters/12345/nodes/12345-6aa78910bc

Upgrade your LKE Cluster to the Next Minor Version

To upgrade your LKE cluster’s version, send a PUT request to the /lke/clusters/{clusterId} endpoint. In this example, ensure you replace 12345 with your cluster’s ID, and 1.17 with whichever Kubernetes version is the next currently available:

curl -H "Content-Type: application/json" \
        -H "Authorization: Bearer $TOKEN" \
        -X PUT -d '{
        "k8s_version": "1.17"
        }' https://api.linode.com/v4/lke/clusters/12345

The response body displays the cluster version that will be applied following a recycle:

{"created": "2019-08-02T17:17:49", "updated": "2019-08-05T19:11:19", "k8s_version": "1.17", "tags": ["ecomm", "blogs"], "label": "updated-cluster-name", "id": 456, "region": "us-central"}%

Add New Tags to your LKE Cluster

Like many Linode resources, you can add tags to your LKE Cluster for organizational purposes. This section shows you how to add new tags to an existing LKE Cluster.

curl -H "Authorization: Bearer $TOKEN" \
    https://api.linode.com/v4/tags

{"data": [{"label": "blogs"}, {"label": "ecomm"}, {"label": "prod"}, {"label": "monitoring"}], "page": 1, "pages": 1, "results": 4}%

1. View the tags currently assigned to your cluster:

   curl -H "Authorization: Bearer $TOKEN" \
     https://api.linode.com/v4/lke/clusters/12345

   The response body contains an array of your cluster’s tags. In the example response, the cluster’s tags are blog, and ecomm.

   {"id": 12345, "status": "ready", "created": "2020-04-13T20:17:22", "updated": "2020-04-13T20:17:22", "label": "cluster-12345", "region": "us-central", "k8s_version": "1.17", "tags": ["blog", "ecomm"]}%

2. To add new tags to your cluster’s existing tags, your request must include a tags array with all previous and new tags. The example request adds the new tags prod and monitoring to the cluster.

   curl -H "Content-Type: application/json" \
         -H "Authorization: Bearer $TOKEN" \
         -X PUT -d '{
           "tags" : ["ecomm", "blog", "prod", "monitoring"]
         }' \
         https://api.linode.com/v4/lke/clusters/12345

   The response displays all of your cluster’s tags. In the example response, the cluster’s tags are now blog, ecomm, prod, and monitoring.

   {"id": 12345, "status": "ready", "created": "2020-04-13T20:17:22", "updated": "2020-04-13T20:17:22", "label": "cluster-12345", "region": "us-central", "k8s_version": "1.17", "tags": ["blog", "ecomm", "monitoring", "prod"]}%

Delete Tags from your LKE Cluster

This section shows you how to delete tags from your LKE Cluster.

1. View the tags currently assigned to your cluster:

   curl -H "Authorization: Bearer $TOKEN" \
     https://api.linode.com/v4/lke/clusters/12345

   The response body contains an array of your cluster’s tags. In the example response, the cluster’s tags are blog, ecomm, prod, and monitoring.

   {"id": 12345, "status": "ready", "created": "2020-04-13T20:17:22", "updated": "2020-04-13T20:17:22", "label": "cluster-12345", "region": "us-central", "k8s_version": "1.17", "tags": [["blog", "ecomm", "monitoring", "prod"]}%

2. To delete a tag from your cluster, issue a request with only the tags you would like to keep assigned to your cluster. In the example request, the tags monitoring and prod are excluded from the tags array and so are deleted from your cluster.

   curl -H "Content-Type: application/json" \
         -H "Authorization: Bearer $TOKEN" \
         -X PUT -d '{
           "tags" : ["ecomm", "blog"]
         }' \
         https://api.linode.com/v4/lke/clusters/12345

   The response displays all of your cluster’s current tags. In the example response, the cluster’s tags are now blog, and ecomm.

   {"id": 12345, "status": "ready", "created": "2020-04-13T20:17:22", "updated": "2020-04-13T20:17:22", "label": "cluster-12345", "region": "us-central", "k8s_version": "1.17", "tags": ["blog", "ecomm"]}%

Delete a Node Pool from an LKE Cluster

When you delete a node pool you also delete the Compute Instances (nodes) and routes to them. The Pods running on those nodes are evicted and rescheduled. If you have assigned Pods to the deleted Nodes, the Pods might remain in an unschedulable condition if no other node in the cluster satisfies the node selector.

To delete a node pool from a LKE cluster, send a DELETE request to the /lke/clusters/{clusterId}/pools/{poolId} end point. In the URL of this example, replace 12345 with your cluster’s ID and 196 with your cluster’s node pool ID:

curl -H "Authorization: Bearer $TOKEN" \
    -X DELETE \
    https://api.linode.com/v4/lke/clusters/12345/pools/196

Delete an LKE Cluster

Deleting an LKE cluster deletes the Master node, all worker nodes, and all NodeBalancers created by the cluster. However, it does not delete any Volumes created by the LKE cluster.

To delete an LKE Cluster, send a DELETE request to the /lke/clusters/{clusterId} endpoint. In the URL of this example, replace 12345 with your cluster’s ID:

curl -H "Authorization: Bearer $TOKEN" \
    -X DELETE \
    https://api.linode.com/v4/lke/clusters/12345

General Network and Firewall Information

In an LKE cluster, some entities and services are only accessible from within that cluster while others are publicly accessible (reachable from the internet).

Private (accessible only within the cluster)

• Pod IPs, which use a per-cluster virtual network in the range 10.2.0.0/16
• ClusterIP Services, which use a per-cluster virtual network in the range 10.128.0.0/16

Public (accessible over the internet)

• NodePort Services, which listen on all Nodes with ports in the range 30000-32768.
• LoadBalancer Services, which automatically deploy and configure a NodeBalancer.
• Any manifest which uses hostNetwork: true and specifies a port.
• Most manifests which use hostPort and specify a port.

Exposing workloads to the public internet through the above methods can be convenient, but this can also carry a security risk. You may wish to manually install firewall rules on your cluster nodes. The following policies are needed to allow communication between the node pools and the control plane and block unwanted traffic:

• Allow kubelet health checks: TCP port 10250 from 192.168.128.0/17 Accept
• Allow Wireguard tunneling for kubectl proxy: UDP port 51820 from 192.168.128.0/17 Accept
• Allow Calico BGP traffic: TCP port 179 from 192.168.128.0/17 Accept
• Allow NodePorts for workload services: TCP/UDP port 30000 - 32767 192.168.128.0/17 Accept
• Block all other TCP traffic: TCP All Ports All IPv4/All IPv6 Drop
• Block all other UDP traffic: UDP All Ports All IPv4/All IPv6 Drop
• Block all ICMP traffic: ICMP All Ports All IPv4/All IPv6 Drop
• IPENCAP for IP ranges 192.168.128.0/17 for internal communication between node pools and control plane.

For additional information, please see this community post. Future LKE release may allow greater flexibility for the network endpoints of these types of workloads.

Please note, at this time, nodes should be removed from the Cloud Firewall configuration before removing/recycling of node pools within the Kubernetes configuration. Also, when adding node pools to the Kubernetes cluster, Cloud Firewall must be updated with the new node pool(s). Failure to add the new nodes creates a security risk.

Where to Go From Here?

Now that you have created an LKE cluster, you can start deploying workloads to it. Review these guides for further help:

• How to Install Apps on Kubernetes with Helm 3
• Create and Deploy a Docker Container Image to a Kubernetes Cluster
• Troubleshooting Kubernetes