Container Service for Kubernetes:Create an ACK Serverless cluster by using Alibaba Cloud CLI

Install Alibaba Cloud CLI

If you use Cloud Shell, Alibaba Cloud CLI is preinstalled and pre-configured — skip this section.

Otherwise, install Alibaba Cloud CLI for your platform:

After installing, run aliyun configure to set up your credentials:

aliyun configure

Enter your Access Key ID, Access Key Secret, default region, and preferred output format when prompted:

Configuring profile 'default' in 'AK' authenticate mode...
Access Key Id []: ************
Access Key Secret []: ************
Default Region Id []: cn-beijing
Default Output Format [json]: json (Only support json)
Default Language [zh|en] en:
Saving profile[default] ...Done.

Verify the installation

Run the following command to confirm the CLI is installed and configured correctly:

aliyun version

Step 1: Set configuration variables

Set environment variables for the cluster parameters you'll reference throughout this guide. Replace the values with your own region and zone:

export CLUSTER_NAME="test-serverless-k8s"
export REGION_ID="cn-hangzhou"
export ZONE_ID="cn-hangzhou-h"
export KUBECONFIG="$HOME/.kube/ask-config"

Step 2: Create the cluster configuration file

Create a file named create.json with the following content. The key fields are explained in the table below:

{
    "cluster_type": "ManagedKubernetes",
    "profile": "Serverless",
    "name": "test-serverless-k8s",
    "region_id": "cn-hangzhou",
    "zoneid": "cn-hangzhou-h",
    "nat_gateway": true,
    "private_zone": false,
    "tags": [
        {"key": "env", "value": "test"}
    ]
}

For a full list of configuration parameters, see Create an ACK Serverless cluster.

Step 3: Submit the cluster creation request

aliyun cs POST /clusters --header "Content-Type=application/json" --body "$(cat create.json)"

The response includes your cluster ID:

{
    "cluster_id": "************************",
    "instanceId": "************************",
    "request_id": "**********-****-****-****-************",
    "task_id": "*-************"
}

Save the cluster_id value. You'll need it in the next steps. Set it as an environment variable:

export CLUSTER_ID="<your-cluster-id>"

Step 4: Verify the cluster is running

Cluster creation takes several minutes. Run the following command to check the status:

aliyun cs GET /clusters/$CLUSTER_ID

The cluster is ready when "state": "running" appears in the response:

{
    "cluster_id": "************************",
    "cluster_spec": "ack.standard",
    "cluster_type": "ManagedKubernetes",
    "created": "2024-05-06T14:48:40+08:00",
    "current_version": "1.28.3-aliyun.1",
    "deletion_protection": false,
    "external_loadbalancer_id": "lb-*********",
    "init_version": "1.28.3-aliyun.1",
    "name": "test-serverless-k8s",
    "network_mode": "vpc",
    "profile": "Serverless",
    "region_id": "cn-hangzhou",
    "resource_group_id": "rg-*********",
    "security_group_id": "sg-*********",
    "service_domain_name": "",
    "size": 0,
    "state": "running",
    "tags": [
        {"key": "env", "value": "test"},
        {"key": "ack.aliyun.com", "value": "cc98dd6edd4ff4c*****************"}
    ],
    "updated": "2024-05-06T14:52:44+08:00",
    "vpc_id": "vpc-*********",
    "vswitch_id": "vsw-*********",
    "zone_id": "cn-hangzhou-*"
}

Step 5: Configure kubectl access

Download the kubeconfig and confirm kubectl can reach the cluster:

aliyun cs GET /k8s/$CLUSTER_ID/user_config | jq -r '.config' > $KUBECONFIG
kubectl get ns

The cluster is accessible when you see the four system namespaces:

NAME              STATUS   AGE
default           Active   7m43s
kube-node-lease   Active   7m45s
kube-public       Active   7m45s
kube-system       Active   7m45s

Step 6: Deploy a test application

Deploy an NGINX application to verify the cluster accepts workloads:

kubectl run nginx --image=registry-vpc.cn-shenzhen.aliyuncs.com/acs-sample/nginx:latest

Expected output:

deployment.apps/nginx created

Step 7: Verify the deployment is ready

kubectl get deploy nginx

The deployment is ready when READY shows 1/1:

NAME    READY   UP-TO-DATE   AVAILABLE   AGE
nginx   1/1     1            1           58s