Use Storage Operator to deploy and upgrade storage components - Container Service for Kubernetes

Introduction

Storage Operator is used to manage the lifecycle of storage components. Storage Operator runs as a Deployment, which deploys and upgrades storage components based on the default configurations inherited from the image and the custom configurations provided by ConfigMaps. This helps reduce the complexity of container development and maintenance.

Note Storage Operator can only be used in ACK clusters.

Default configurations: Storage Operator provides the default configurations of storage components. The default configurations vary based on the version of Storage Operator.
Custom configurations: ConfigMaps can be used to define custom configurations of storage components, such as version information, and whether to install the component.

Storage Operator preferably uses custom configurations. The default configurations are used only when the custom configurations are not specified.

Note

Each image of Storage Operator contains the default configuration file.
When Storage Operator runs as a Deployment, the Deployment reads configurations from a ConfigMap file that is mounted on the Deployment and contains configurations of storage components.
Storage Operator determines whether to deploy and upgrade a storage component by combining the default and custom configurations.

Step 1: Configure the components

Method 1: Use the default configurations provided by the image

ACK supports the following storage components: storage-snapshot-manager, storage-analyzer, storage-auto-expander, and storage-monitor.

Storage Operator provides the default configurations of storage components. The default configurations vary based on the version of Storage Operator. In this example, V1.18.8.0 is used.

{
  "storage-snapshot-manager": {
    "install":  "true",
    "imageTag": "v1.18.8.0-81508da-aliyun",
    "imageRep": "acs/storage-snapshot-manager",
    "template": "/acs/templates/storage-snapshot-manager/install.yaml"
  },
  "storage-analzyer": {
    "install":  "false",
    "imageTag": "v1.18.8.0-06c5560-aliyun",
    "imageRep": "acs/storage-analyzer",
    "template": "/acs/templates/storage-analyzer/install.yaml"
  },
  "storage-auto-expander": {
    "install":  "false",
    "imageTag": "v1.18.8.0-4852fd4-aliyun",
    "imageRep": "acs/storage-auto-expander",
    "crdTmpl":  "/acs/templates/storage-auto-expander/crd.yaml",
    "template": "/acs/templates/storage-auto-expander/install.yaml"
  },
  "storage-monitor": {
    "install":  "true",
    "imageTag": "v1.18.8.0-c4744b6-aliyun",
    "imageRep": "acs/storage-monitor",
    "template": "/acs/templates/storage-monitor/install.yaml",
    "svcTmpl":  "/acs/templates/storage-monitor/service.yaml"
  }
}

The following table describes the parameters in the default configurations.


Parameter	Description
install	Specifies whether to deploy the component. Default value: `false`.
imageTag	Specifies the default image version if the component is to be deployed.
imageRep	Specifies the default image name if the component is to be deployed.
crdTmpl	Specifies the template path of the Custom Resource Definition (CRD) that you want to deploy. If this parameter is empty, it indicates that no CRD is to be deployed.
svcTmpl	Specifies the template path of the Service that you want to deploy.
template	Specifies the template path of the component that you want to deploy.

Note You can use a ConfigMap to modify the parameters.

Method 2: Use a ConfigMap to customize the configurations

In this example, the following ConfigMap template is provided:

kind: ConfigMap
apiVersion: v1
metadata:
  name: storage-operator
  namespace: kube-system
data:
  storage-snapshot-manager: |
    # deploy config
    install:   false
    imageTag:  v1.16.aaaa
    imageRep:  acs/storage-snapshot-manager
    template:  /acs/templates/storage-snapshot-manager/install.yaml
    # env config
    SNAPSHOT_INTERVAL: 30
  storage-analyzer: |
    # deploy config
    install:   false
    imageTag:  v1.16.bbbb
    imageRep:  acs/storage-analyzer
    template:  /acs/templates/storage-analyzer/install.yaml
    # env config
    LOOP_INTERVAL: 30
  storage-auto-expander: |
    # deploy config
    install:   false
    imageTag:  v1.16.3adsadfs
    imageRep:  acs/storage-auto-expander
    crdTmpl:   /acs/templates/storage-auto-expander/crd.yaml
    template:  /acs/templates/storage-auto-expander/install.yaml
    # env config
    POLLING_INTERVAL_SECONDS: 60

Take note of the following items when you use custom configurations:

The configurations of each component consist of two parts:
- Deployment configurations, which are defined by the parameters in the deploy config section in the preceding example. The parameters specify whether to install the component and the component version that is deployed. Storage Operator deploys and upgrades components based on these configurations.
- Environment configurations, which are defined by the parameters in the env config section in the preceding example. The parameters are used by the storage component. For example, SNAPSHOT_INTERVAL is configured as an environment variable for the storage-snapshot-manager component.
The parameter values that are specified in the custom configurations overwrite the parameter values in the default configurations.
The parameters used in deployment configurations are reserved parameters and cannot be used as environment variables.
You can customize specific parameters based on the features of the component.

Step 2: Deploy Storage Operator

Run the following command to deploy Storage Operator:

kubectl apply -f StorageOperator.yaml

In this example, the following StorageOperator.yaml file is used to deploy Storage Operator:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: storage-operator
  namespace: kube-system
  labels:
    app: storage-operator
spec:
  selector:
    matchLabels:
      app: storage-operator
  template:
    metadata:
      labels:
        app: storage-operator
    spec:
      tolerations:
        - operator: "Exists"
      priorityClassName: system-node-critical
      serviceAccount: storage-operator-admin
      containers:
      - name: storage-operator
        image: registry.cn-hangzhou.aliyuncs.com/acs/storage-operator:v1.18.8.28-18cca7b-aliyun
        imagePullPolicy: Always
        volumeMounts:
        - mountPath: /acs/configmap/
          name: storage-operator
      restartPolicy: Always
      volumes:
      - configMap:
          name: storage-operator
        name: storage-operator

Run the following command to check whether Storage Operator is running:

kubectl get pods -nkubes-system | grep storage-operator

Expected output:

NAME                        READY   STATUS    RESTARTS   AGE         IP                NODE              NOMINATED NODE   READINESS GATES
storage-operator-***        1/1     Running   0          6d20h       192.168.XX.XX     virtual-kubelet   <none>           <none>