Container Service for Kubernetes:Schedule pods to elastic container instances that are deployed as virtual nodes

How it works

Elastic Container Instance is a serverless compute service that is provided by Alibaba Cloud for containerization. You can use elastic container instances to efficiently set up an O&M-free and isolated runtime environment for your containers. Elastic container instances allow you to focus on containerized applications without the need to purchase or manage Elastic Compute Service (ECS) instances. This eliminates the need for infrastructure maintenance. You can create elastic container instances to meet your business requirements. You are charged for resource usage on a per second basis.

In most cases, an ACK cluster has at least one ECS node pool. After you create a pod, ACK schedules the pod to an ECS instance in the node pool. This scheduling mode is suitable for applications that receive a stable volume of traffic. However, this scheduling mode cannot handle traffic spikes, even though ACK can scale out ECS instances. This is because the creation and startup of ECS instances is time-consuming. To resolve this problem, we recommend that you schedule pods to elastic container instances that are deployed as virtual nodes. This eliminates the need to add new nodes, prevents resource waste, and reduces computing costs.

Step 1: Grant RAM permissions to ack-virtual-node

Step 2: Install ack-virtual-node

onectl addon install ack-virtual-node

Addon ack-virtual-node, version **** installed.

Step 3: Schedule pods to elastic container instances

After ack-virtual-node is deployed in your cluster, you can schedule pods to elastic container instances that are deployed as virtual nodes in your cluster. You can use one of the following methods to schedule pods to elastic container instances in a registered cluster: Make sure that the virtual nodes are in the Ready state before you schedule the pods.

1. Run the following command to query the status of the virtual node:

   kubectl  get no |grep virtual-kubelet

   Expected output:

   virtual-kubelet-cn-hangzhou-b   Ready    agent                  18d   v1.20.11-aliyun.1

   The output indicates that the virtual node is in the Ready state.

2. You can use one of the following three methods to schedule pods to elastic container instances.

   Method 1: Add labels to pods (version of the cluster is later than 1.16)

   Add the alibabacloud.com/eci=true label to the pod that you want to create. Then, an Elastic Container Instance-based pod is created and scheduled to the virtual node. Example:

   1. Run the following command to add a label to the pod:

   kubectl run nginx --image nginx -l alibabacloud.com/eci=true

   2. Run the following command to query the pod:

   kubectl get pod -o wide|grep virtual-kubelet

   3. Expected output:

   nginx-7fc9f746b6-r4xgx     0/1     ContainerCreating   0          20s   192.168.XX.XX   virtual-kubelet        <none>           <none>

   Method 2: Add labels to namespaces

   Add the alibabacloud.com/eci=true label to the namespace to which the pod belongs. Then, an Elastic Container Instance-based pod is created and scheduled to the virtual node. Example:

   1. Run the following command to create a virtual node:

   kubectl create ns vk

   2. Run the following command to add a label to the namespace to which the pod belongs:

   kubectl label namespace vk alibabacloud.com/eci=true

   3. Run the following command to schedule the pod to the virtual node:

   kubectl -n vk run nginx --image nginx

   4. Run the following command to query the pod:

   kubectl -n vk get pod -o wide|grep virtual-kubelet

   5. Expected output:

   nginx-6f489b847d-vgj4d      1/1     Running             0          1m   192.168.XX.XX   virtual-kubelet        <none>           <none>

   Method 3: Specify node names

   Schedule only specified pods to virtual nodes. Add the nodeName: virtual-kubelet-cn-shanghai-k label to the the pod. Then, an Elastic Container Instance-based pod is created and scheduled to a virtual node. Example:

   1. Create a file named nginx-deployment.yaml and copy the following content to the file:

      apiVersion: apps/v1 # for versions before 1.8.0 use apps/v1beta1
      kind: Deployment
      metadata:
        name: nginx-deployment-basic
        labels:
          app: nginx
      spec:
        replicas: 2
        selector:
          matchLabels:
            app: nginx
        template:
          metadata:
            labels:
              app: nginx
          spec:
            nodeName: virtual-kubelet-cn-shanghai-k             # nodeName of the specified virtual kubelet
            containers:
            - name: nginx
              image: nginx:1.7.9 # replace it with your exactly <image_name:tags>
              ports:
              - containerPort: 80
              resources:
                limits:
                  cpu: "500m"

   2. Run the following command to deploy the application:

      kubectl apply -f nginx-deployment.yaml

   3. Run the following command to query the pod:

      kubectl  get pod -o wide|grep virtual-kubelet

   4. Expected output:

      nginx-6f489b847d-XXX      1/1     Running             0          1m   192.168.XX.XX   virtual-kubelet        <none>           <none>
      nginx-6f489b847d-XXX      1/1     Running             0          1m   192.168.XX.XX   virtual-kubelet        <none>           <none>

What to do next

Modify the configurations of the vSwitch to which an elastic container instance is connected

You can modify the ECI_VSWITCH environment variable to change the vSwitch of the pods that are scheduled to the virtual node. We recommend that you configure multiple vSwitches that are deployed in different zones to ensure high availability. When elastic container instances in the current zone are out of stock, the virtual node controller creates pods in another zone.

Run the following command to modify the configurations of the vSwitch to which an elastic container instance is connected:

kubectl edit configmap eci-profile -n kube-system

Configuration example:

data:
  enableClusterIp: "true"
  enableHybridMode: "false"
  enablePrivateZone: "false"
  securityGroupId: sg-2zeeyaaxlkq9sppl****
  selectors: ""
  vSwitchIds: vsw-2ze23nqzig8inprou****,vsw-2ze94pjtfuj9vaymf****     # the vSwitch to which an elastic container instance pod is connected. Separate multiple vSwitch IDs with commas (,).
  vpcId: vpc-2zeghwzptn5zii0w7****

For more information about how to configure an eci-profile, see Configure an eci-profile.

Delete a virtual node

1. Uninstall ack-virtual-node from the registered cluster.

   Use onectl

   Run the following command to uninstall ack-virtual-node:

   onectl addon uninstall ack-virtual-node

   Expected output:

   Addon ack-virtual-node uninstalled.

   Use the console

   Go to the Add-ons page in the ACK console and uninstall ack-virtual-node.

2. Run the kubectl delete node <node name> to delete the virtual nodes that you do not need.

   Note

   After you uninstall ack-virtual-node, existing elastic container instances in your cluster are not deleted.