Skip to content

Latest commit

 

History

History
403 lines (282 loc) · 15.1 KB

README.md

File metadata and controls

403 lines (282 loc) · 15.1 KB
Raw

How to Set Up a DigitalOcean Managed Kubernetes Cluster (DOKS)

Introduction

In this tutorial, you will learn how set up a DigitalOcean managed Kubernetes cluster (DOKS), using the command-line interface. Then, you're going to inspect the cluster state, as well as the available features.

After completing this tutorial, you will be able to:

  • Use the doctl command, to create and manage DOKS clusters.
  • Inspect DOKS clusters state.

Note: As an alternative to this chapter, you can use the DOKS UI to create a cluster. UI is easy to understand and very convenient as well.

Table of contents

Prerequisites

To complete this tutorial, you will need:

  1. A DigitalOcean account, for accessing the DigitalOcean platform.
  2. A DigitalOcean personal access token, for using the DigitalOcean API.
  3. Doctl utility, for managing DigitalOcean resources using the command-line interface.
  4. Kubectl, for Kubernetes cluster interaction.

Step 1 - Doctl CLI Introduction

In this step, you will learn how to use the doctl utility, and explore the available options for the DigitalOcean platform. The way Doctl works, is by using commands and sub-commands, to create and manage DigitalOcean resources. You can get help for each via the --help flag.

Next, you're going to explore the doctl auth command and sub-commands.

Please open a terminal, and type the following command to list all the available options for doctl:

doctl --help

The output looks similar to:

doctl is a command line interface (CLI) for the DigitalOcean API.

Usage:
  doctl [command]

Available Commands:
  1-click         Display commands that pertain to 1-click applications
  account         Display commands that retrieve account details
  apps            Display commands for working with apps
  auth            Display commands for authenticating doctl with an account
  balance         Display commands for retrieving your account balance
  billing-history Display commands for retrieving your billing history
  completion      Modify your shell so doctl commands autocomplete with TAB
  compute         Display commands that manage infrastructure
  databases       Display commands that manage databases
  help            Help about any command
  invoice         Display commands for retrieving invoices for your account
  kubernetes      Displays commands to manage Kubernetes clusters and configurations
  monitoring      [Beta] Display commands to manage monitoring
  projects        Manage projects and assign resources to them
  registry        Display commands for working with container registries
  version         Show the current version
  vpcs            Display commands that manage VPCs
  ...

Next, inspect the auth command help page:

doctl auth --help

The output looks similar to (some parts are hidden for simplicity):

The `doctl auth` commands allow you to authenticate doctl for use with your DigitalOcean account using tokens that you generate in the control panel at https://cloud.digitalocean.com/account/api/tokens.
...

Usage:
  doctl auth [command]

Available Commands:
  init        Initialize doctl to use a specific account
  list        List available authentication contexts
  remove      Remove authentication contexts 
  switch      Switches between authentication contexts
  ...

Then, see what options are available for the list sub-command (part of the doctl auth command):

doctl auth list --help

The output looks similar to:

List named authentication contexts that you created with `doctl auth init`.

To switch between the contexts use `doctl switch <name>`, where `<name>` is one of the contexts listed.

To create new contexts, see the help for `doctl auth init`.

Usage:
  doctl auth list [flags]
...

In the next step, you're going to learn how to authenticate to DigitalOcean API with doctl, to create and manage resources on the DigitalOcean platform.

Step 2 - Authenticating to DigitalOcean API

Doctl needs to authenticate with DigitalOcean API to perform queries, and create resources on your behalf, hence an access token is needed (Step #2 from Prerequisites). For each command or sub-command that you run, doctl performs an API call to DigitalOcean.

To authenticate doctl with DigitalOcean API, you can use the auth command of doctl.

First, list the available options for the auth command:

doctl auth -h

The output looks similar to:

...
Usage:
  doctl auth [command]

Available Commands:
  init        Initialize doctl to use a specific account
  list        List available authentication contexts
  remove      Remove authentication contexts 
  switch      Switches between authentication contexts

Next, you're going to use the init sub-command for doctl auth, to perform the initialization (when asked, please enter the personal access token created in the Prerequisites step):

doctl auth init

The output looks similar to:

Please authenticate doctl for use with your DigitalOcean account. You can generate a token in the control panel at https://cloud.digitalocean.com/account/api/tokens

Enter your access token: <paste_your_personal_token_here>

Validating token... OK

Finally, check that your account is configured for doctl to use:

doctl auth list

The output looks similar to (notice the line containing (current)):

default
doks-starterkit (current)

Next, you're going to learn how to spin up a DOKS cluster, and explore the available options.

Step 3 - Creating the DOKS Cluster

In this step, you will learn how to use the doctl k8s command to create a DOKS cluster.

First, explore the available doctl commands for managing DOKS clusters:

  • Manage a DOKS cluster:

    doctl k8s -h

  • List available options for a DOKS cluster, like: region, size and version:

    doctl k8s options -h

  • List what regions are available to use, when creating a DOKS cluster:

    doctl k8s options regions

  • List machine sizes that can be used in a DOKS cluster:

    doctl k8s options sizes

  • List Kubernetes versions that can be used with DigitalOcean clusters:

    doctl k8s options versions

  • Display commands for managing Kubernetes clusters:

    doctl k8s cluster -h

Next, you're going to focus on the create sub-command of doctl k8s cluster. Inspect the available options via:

doctl k8s cluster create -h

For the Starter Kit tutorial, you will need a DOKS cluster with 3 worker nodes . Use --wait false, if you do not want the command to wait until cluster is ready.

The below example is using 4cpu/8gb AMD nodes ($48/month), 3 default, and auto-scale to 2-4. So, your cluster cost is between $96-$192/month, with hourly billing. To choose a different node type, pick from the following command doctl compute size list.

doctl k8s cluster create starterkit-cluster-2 \
  --auto-upgrade=false \
  --maintenance-window "saturday=21:00" \
  --node-pool "name=basicnp;size=s-4vcpu-8gb-amd;count=3;tag=cluster2;label=type=basic;auto-scale=true;min-nodes=2;max-nodes=4" \
  --region nyc1

The output looks similar to:

Notice: Cluster is provisioning, waiting for cluster to be running
..................................................................
Notice: Cluster created, fetching credentials
Notice: Adding cluster credentials to kubeconfig file found in "/Users/starterkit/.kube/config"
Notice: Setting current-context to starterkit-cluster-2
ID                                      Name                  Region    Version        Auto Upgrade    Status     Node Pools
0922a629-7f2e-4bda-940c-4d42a3f987ad    starterkit-cluster-2  nyc1      1.21.5-do.0    false           running    basicnp

Next, you can verify the cluster details. First, fetch your DOKS cluster ID:

doctl k8s cluster list

The output looks similar to (notice the ID column value):

ID                                      Name                  Region    Version        Auto Upgrade    Status     Node Pools
b4ddaa2e-8c0c-4fd8-b249-cbf99eda0808    starterkit-cluster-2  nyc1      1.21.5-do.0    false           running    basicnp

Now, query the DigitalOcean API, using curl (please make sure to replace the <> placeholders accordingly)

curl -X GET \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <your_do_api_token>" \
  https://api.digitalocean.com/v2/kubernetes/clusters/<cluster_id>

Finally, check if the kubectl context was set to point to your DOKS cluster. The doctl utility does it automatically for you in general, but it's good to know if something goes bad.

kubectl config current-context

The output looks similar to:

starterkit-cluster-2

If the above command output is empty or different, you can use the kubeconfig sub-command of doctl k8s cluster to set kubectl context.

First, list the available DOKS clusters:

doctl k8s cluster list

The output looks similar to:

ID                                      Name                  Region    Version        Auto Upgrade    Status     Node Pools
b4ddaa2e-8c0c-4fd8-b249-cbf99eda0808    starterkit-cluster-2  nyc1      1.21.5-do.0    false           running    basicnp

Next, set kubectl context to point to your cluster:

doctl kubernetes cluster kubeconfig save <your_cluster_name>

The output looks similar to:

Notice: Adding cluster credentials to kubeconfig file found in "/Users/starterkit/.kube/config"
Notice: Setting current-context to starterkit-cluster-2

Finally, list DOKS cluster nodes:

kubectl get nodes

The output looks similar to:

NAME            STATUS   ROLES    AGE     VERSION
basicnp-865x3   Ready    <none>   2m55s   v1.21.5
basicnp-865x8   Ready    <none>   2m21s   v1.21.5
basicnp-865xu   Ready    <none>   2m56s   v1.21.5

If everything was set correctly, you should get a list of all the DOKS cluster worker nodes. The STATUS column should print Ready, if all the nodes are healthy.

Hint: If the worker node(s) STATUS is different from Ready, you can inspect the affected node(s), via (please replace the <> placeholders accordingly):

kubectl describe node <worker_node_name>

After running the above command, please look at the Events section (last line from command output), to check if something went wrong. There are many other useful sections to look at, like Conditions, System Info, Allocated resources, to help you troubleshoot worker nodes issues in the future.

Step 4 [OPTIONAL] - Adding a dedicated node for observability

If you plan to use this cluster to serve in a production environment it is recommended that you also setup, apart from the basic nodes, another fixed size node pool with the purpose of serving the observability stack from Chapter 4 Setup Observability. In general, it is good practice to separate the observability stack from user applications. This way, one cannot interfere with another or get affected by downtime when performing cluster or node pool maintenance, etc. On the other hand, monitoring is a crucial aspect of any modern infrastructure hence high-availability is a must. Later on, you will use Node affinity to schedule observability related pods on the dedicated node pool.

To add another node pool to the cluster created earlier run the following command:

doctl kubernetes cluster node-pool create starterkit-cluster-2 \
    --name "observability" \
    --size "s-4vcpu-8gb-amd" \
    --min-nodes 1 \
    --max-nodes 1 \
    --count 1

The ouput looks similar to:

ID                                      Name             Size                    Count    Tags                                                   
7b5037c8-637c-4a8b-abbe-3296b5aa92fa    observability    s-4vcpu-8gb-amd         1        k8s,k8s:1dcda264-15d6-4bcb-92b1-e64d236f59c1,k8s:worker

Next check the cluster's nodes created:

kubectl get nodes

The output looks similar to (notice the observability prefix in the node name):

NAME                    STATUS   ROLES    AGE     VERSION
basicnp-c4k0f           Ready    <none>   2m34s   v1.22.11
basicnp-c4k0q           Ready    <none>   2m38s   v1.22.11
basicnp-c4k0y           Ready    <none>   2m38s   v1.22.11
observability-cd111     Ready    <none>   2m44s   v1.22.11

Next you will add a label to the new node. This will make it easier to schedule pods onto this node using a distinct label and node affinity.

kubectl label nodes <YOUR_NODE_NAME> preferred=observability

Verify that your node has a preferred=observability label:

kubectl get nodes <YOUR_NODE_NAME> --show-labels

The output looks similar to (notice the preferred=observability label):

NAME                  STATUS   ROLES    AGE     VERSION   LABELS
observability-cd111   Ready    <none>   9m27s   v1.22.8   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/instance-type=s-4vcpu-8gb-amd,beta.kubernetes.io/os=linux,doks.digitalocean.com/node-id=eb199834-a852-40fe-9785-42c361536ec0,doks.digitalocean.com/node-pool-id=92e14637-73d1-4703-a902-11fef09ca4f2,doks.digitalocean.com/node-pool=observability,doks.digitalocean.com/version=1.22.8-do.1,failure-domain.beta.kubernetes.io/region=nyc3,kubernetes.io/arch=amd64,kubernetes.io/hostname=observability-cd111,kubernetes.io/os=linux,node.kubernetes.io/instance-type=s-4vcpu-8gb-amd,preferred=observability,region=nyc3,topology.kubernetes.io/region=nyc3

Conclusion

In this tutorial you learned how to use the doctl utility, inspect the available options, as well as how to get help for a specific command or sub-command. Then, you learned how to create a DOKS cluster, and inspect worker nodes state.

In the next section, you will learn how to use the DigitalOcean Container Registry (DOCR), to easily store and manage private container images for your Kubernetes cluster.

Go to Section 2 - Setting up DOCR.