如何使用storageclass实现动态pv？

如何使用storageclass实现动态pv？针对这个问题，今天小编总结这篇有关storageclass实践的文章，希望帮助更多想解决这个问题的同学找到更加简单易行的办法。

在部署nfs-client-provisioner前，我们需要先准备好nfs存储服务器并在所有node节点上安装

nfs服务器：192.168.248.139

共享存储目录：/data/nfs

nfs-client-provisioner部署文件

vim nfs-client-provisioner.yaml

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nfs-client-provisioner
  namespace: default
spec:
  replicas: 1
  selector:
    matchLabels:
      app: nfs-client-provisioner
  strategy:
    type: Recreate
  template:
    metadata:
      labels:
        app: nfs-client-provisioner
    spec:
      serviceAccountName: nfs-client-provisioner
      containers:
      - name: nfs-client-provisioner
        image: quay.azk8s.cn/external_storage/nfs-client-provisioner:latest
        volumeMounts:
        - name: timezone
          mountPath: /etc/localtime
        - name: nfs-client-root
          mountPath: /persistentvolumes
        env:
        - name: PROVISIONER_NAME
          value: fuseim.pri/ifs
        - name: NFS_SERVER
          value: 192.168.248.139
        - name: NFS_PATH
          value: /data/nfs
      volumes:
      - name: timezone
        hostPath:
          path: /usr/share/zoneinfo/Asia/Shanghai
      - name: nfs-client-root
        nfs:
          server: 192.168.248.139
          path: /data/nfs

Storageclass部署文件

vim nfs-client-class.yaml

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: managed-nfs-storage
  annotations:
    storageclass.kubernetes.io/is-default-class: "true"   #设置其为默认存储后端
provisioner: fuseim.pri/ifs   #or choose another name, must match deployment's env PROVISIONER_NAME'
parameters:
  archiveOnDelete: "false"   #删除pvc后，后端存储上的pv也自动删除

rbac授权文件

vim nfs-client-rbac.yaml

kind: ServiceAccount
apiVersion: v1
metadata:
  name: nfs-client-provisioner
---
kind: ClusterRole
apiVersion: rbac.authorization.k8s.io/v1
metadata:
  name: nfs-client-provisioner-runner
rules:
- apiGroups: [""]
  resources: ["persistentvolumes"]
  verbs: ["get", "list", "watch", "create", "delete"]
- apiGroups: [""]
  resources: ["persistentvolumeclaims"]
  verbs: ["get", "list", "watch", "update"]
- apiGroups: ["storage.k8s.io"]
  resources: ["storageclasses"]
  verbs: ["get", "list", "watch"]
- apiGroups: [""]
  resources: ["events"]
  verbs: ["create", "update", "patch"]
---
kind: ClusterRoleBinding
apiVersion: rbac.authorization.k8s.io/v1
metadata:
  name: run-nfs-client-provisioner
subjects:
- kind: ServiceAccount
  name: nfs-client-provisioner
  namespace: default
roleRef:
  kind: ClusterRole
  name: nfs-client-provisioner-runner
  apiGroup: rbac.authorization.k8s.io
---
kind: Role
apiVersion: rbac.authorization.k8s.io/v1
metadata:
  name: leader-locking-nfs-client-provisioner
rules:
- apiGroups: [""]
  resources: ["endpoints"]
  verbs: ["get", "list", "watch", "create", "update", "patch"]
---
kind: RoleBinding
apiVersion: rbac.authorization.k8s.io/v1
metadata:
  name: leader-locking-nfs-client-provisioner
subjects:
- kind: ServiceAccount
  name: nfs-client-provisioner
  namespace: default
roleRef:
  kind: Role
  name: leader-locking-nfs-client-provisioner
  apiGroup: rbac.authorization.k8s.io

准备好以上三个文件后，使用kubectl apply命令应用即可完成nfs-client-provisioner的部署。

[root@k8s-master-01 Dynamic-pv]# kubectl apply -f .
storageclass.storage.k8s.io/managed-nfs-storage created
deployment.apps/nfs-client-provisioner created
serviceaccount/nfs-client-provisioner created
clusterrole.rbac.authorization.k8s.io/nfs-client-provisioner-runner created
clusterrolebinding.rbac.authorization.k8s.io/run-nfs-client-provisioner created
role.rbac.authorization.k8s.io/leader-locking-nfs-client-provisioner created
rolebinding.rbac.authorization.k8s.io/leader-locking-nfs-client-provisioner created

查看pod运行状态和sc

[root@k8s-master-01 Dynamic-pv]# kubectl get pod,sc
NAME                                        READY   STATUS    RESTARTS   AGE
pod/nfs-client-provisioner-c676947d-pfpms   1/1     Running   0          107s

NAME                                                        PROVISIONER      AGE
storageclass.storage.k8s.io/managed-nfs-storage (default)   fuseim.pri/ifs   108s

可以看到nfs-client-provisioner已经正常运行，sc已经创建成功。接下来我们测试创建几个pvc

vim mysql-pvc.yaml

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: mysql-01-pvc
#  annotations:
#    volume.beta.kubernetes.io/storage-class: "managed-nfs-storage"
spec:
  accessModes: ["ReadWriteMany"]
  resources:
    requests:
      storage: 10Gi
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: mysql-02-pvc
spec:
  accessModes: ["ReadWriteMany"]
  resources:
    requests:
      storage: 5Gi
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: mysql-03-pvc
spec:
  accessModes: ["ReadWriteMany"]
  resources:
    requests:
      storage: 3Gi

[root@k8s-master-01 Dynamic-pv]# kubectl apply -f mysql-pvc.yaml 
persistentvolumeclaim/mysql-01-pvc created
persistentvolumeclaim/mysql-02-pvc created
persistentvolumeclaim/mysql-03-pvc created
[root@k8s-master-01 Dynamic-pv]# kubectl get pvc,pv
NAME                                 STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS          AGE
persistentvolumeclaim/mysql-01-pvc   Bound    pvc-eef853e1-f8d8-4ab9-bfd3-05c2a58fd9dc   10Gi       RWX            managed-nfs-storage   2m54s
persistentvolumeclaim/mysql-02-pvc   Bound    pvc-fc0b8228-81c0-4d91-83b0-6bb20ab37cc3   5Gi        RWX            managed-nfs-storage   2m54s
persistentvolumeclaim/mysql-03-pvc   Bound    pvc-c6739d7d-4930-49bd-975f-04bffc05dfd6   3Gi        RWX            managed-nfs-storage   2m54s

NAME                                                        CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS   CLAIM                  STORAGECLASS          REASON   AGE
persistentvolume/pvc-c6739d7d-4930-49bd-975f-04bffc05dfd6   3Gi        RWX            Delete           Bound    default/mysql-03-pvc   managed-nfs-storage            2m54s
persistentvolume/pvc-eef853e1-f8d8-4ab9-bfd3-05c2a58fd9dc   10Gi       RWX            Delete           Bound    default/mysql-01-pvc   managed-nfs-storage            2m54s
persistentvolume/pvc-fc0b8228-81c0-4d91-83b0-6bb20ab37cc3   5Gi        RWX            Delete           Bound    default/mysql-02-pvc   managed-nfs-storage            2m54s

可以看到pvc已经创建成功，并自动创建了一个关联的pv资源对象。我们再查看后端存储目录里面是否生成了对应命名格式的pv

[root@localhost nfs]# pwd
/data/nfs
[root@localhost nfs]# ll
total 12
drwxrwxrwx 2 root root 4096 Feb 20 10:05 default-mysql-01-pvc-pvc-eef853e1-f8d8-4ab9-bfd3-05c2a58fd9dc
drwxrwxrwx 2 root root 4096 Feb 20 10:05 default-mysql-02-pvc-pvc-fc0b8228-81c0-4d91-83b0-6bb20ab37cc3
drwxrwxrwx 2 root root 4096 Feb 20 10:05 default-mysql-03-pvc-pvc-c6739d7d-4930-49bd-975f-04bffc05dfd6

可以看到下面有名字很长的文件夹，这个文件夹的命名方式是不是和我们上面的规则：${namespace}-${pvcName}-${pvName}是一样的，结果符合我们的预期。

接下来我们部署一个mysql应用，测试下 StorageClass 方式声明的 PVC 对象

cat mysql-config.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: mysql-config
data:
  custom.cnf: |
    [mysqld]
    default_storage_engine=innodb
    skip_external_locking
    skip_host_cache
    skip_name_resolve
    default_authentication_plugin=mysql_native_password

cat mysql-secret.yaml
apiVersion: v1
kind: Secret
metadata:
  name: mysql-user-pwd
data:
  mysql-root-pwd: cGFzc3dvcmQ=

cat mysql-deploy.yaml
apiVersion: v1
kind: Service
metadata:
  name: mysql
spec:
  type: NodePort
  ports:
  - port: 3306
    nodePort: 30006
    protocol: TCP
    targetPort: 3306 
  selector:
    app: mysql

---
apiVersion: apps/v1
kind: Deployment
metadata:
  name: mysql
spec:
  replicas: 1
  selector:
    matchLabels:
      app: mysql
  strategy:
    type: Recreate
  template:
    metadata:
      labels:
        app: mysql
    spec:
      containers:
      - image: mysql
        name: mysql
        imagePullPolicy: IfNotPresent
        env:
        - name: MYSQL_ROOT_PASSWORD
          valueFrom:
            secretKeyRef:
              name: mysql-user-pwd
              key: mysql-root-pwd
        ports:
        - containerPort: 3306
          name: mysql
        volumeMounts:
        - name: mysql-config
          mountPath: /etc/mysql/conf.d/
        - name: mysql-persistent-storage
          mountPath: /var/lib/mysql
        - name: timezone
          mountPath: /etc/localtime
      volumes:
      - name: mysql-config
        configMap:
          name: mysql-config
      - name: timezone
        hostPath:
          path: /usr/share/zoneinfo/Asia/Shanghai
      - name: mysql-persistent-storage
        persistentVolumeClaim:
          claimName: mysql-01-pvc

[root@k8s-master-01 yaml]# kubectl apply -f .
configmap/mysql-config created
service/mysql created
deployment.apps/mysql created
secret/mysql-user-pwd created
[root@k8s-master-01 yaml]# kubectl get pod,svc
NAME                                        READY   STATUS    RESTARTS   AGE
pod/mysql-7c5b5df54c-vrnr8                  1/1     Running   0          83s
pod/nfs-client-provisioner-c676947d-pfpms   1/1     Running   0          30m

NAME                 TYPE        CLUSTER-IP   EXTERNAL-IP   PORT(S)          AGE
service/kubernetes   ClusterIP   10.0.0.1     <none>        443/TCP          93d
service/mysql        NodePort    10.0.0.19    <none>        3306:30006/TCP   83s

可以看到mysql应用已经正常运行，我们通过任意一个node节点的ip和30006端口连接mysql数据库测试

[root@localhost ~]# mysql -uroot -h292.168.248.134 -P30006 -p
Enter password: 
Welcome to the MariaDB monitor.  Commands end with ; or \g.
Your MySQL connection id is 10
Server version: 8.0.19 MySQL Community Server - GPL

Copyright (c) 2000, 2018, Oracle, MariaDB Corporation Ab and others.

Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.

MySQL [(none)]> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| performance_schema |
| sys                |
+--------------------+
4 rows in set (0.01 sec)

MySQL [(none)]>

可以看到mysql数据库连接正常。此时查看nfs存储，mysql数据库数据已经持久化到nfs服务器/data/nfs/default-mysql-01-pvc-pvc-eef853e1-f8d8-4ab9-bfd3-05c2a58fd9dc目录中

[root@localhost nfs]# du -sh *
177M    default-mysql-01-pvc-pvc-eef853e1-f8d8-4ab9-bfd3-05c2a58fd9dc
4.0K    default-mysql-02-pvc-pvc-fc0b8228-81c0-4d91-83b0-6bb20ab37cc3
4.0K    default-mysql-03-pvc-pvc-c6739d7d-4930-49bd-975f-04bffc05dfd6
[root@localhost nfs]# cd default-mysql-01-pvc-pvc-eef853e1-f8d8-4ab9-bfd3-05c2a58fd9dc/
[root@localhost default-mysql-01-pvc-pvc-eef853e1-f8d8-4ab9-bfd3-05c2a58fd9dc]# ls
auto.cnf       binlog.index  client-cert.pem  ibdata1      ibtmp1        mysql.ibd           public_key.pem   sys
binlog.000001  ca-key.pem    client-key.pem   ib_logfile0  #innodb_temp  performance_schema  server-cert.pem  undo_001
binlog.000002  ca.pem        ib_buffer_pool   ib_logfile1  mysql         private_key.pem     server-key.pem   undo_002

另外我们可以看到我们这里是手动创建的一个 PVC 对象，在实际工作中，使用 StorageClass 更多的是 StatefulSet 类型的服务，StatefulSet 类型的服务我们也可以通过一个 volumeClaimTemplates 属性来直接使用 StorageClass，如下

vim web.yaml

apiVersion: v1
kind: Service
metadata:
  name: nginx
  labels:
    app: nginx
spec:
  ports:
  - port: 80
    name: web
  clusterIP: None
  selector:
    app: nginx
---
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: web
spec:
  serviceName: "nginx"
  replicas: 8
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      containers:
      - name: nginx
        image: nginx
        imagePullPolicy: IfNotPresent
        ports:
        - containerPort: 80
          name: web
        volumeMounts:
        - name: www
          mountPath: /usr/share/nginx/html
  volumeClaimTemplates:
  - metadata:
      name: www
    spec:
      accessModes: [ "ReadWriteOnce" ]
      resources:
        requests:
          storage: 10Gi

直接创建上面的对象

[root@k8s-master-01 Dynamic-pv]# kubectl apply -f web.yaml 
service/nginx created
statefulset.apps/web created
[root@k8s-master-01 Dynamic-pv]# kubectl get pod -o wide
NAME                                    READY   STATUS    RESTARTS   AGE   IP             NODE          NOMINATED NODE   READINESS GATES
nfs-client-provisioner-c676947d-wzwhh   1/1     Running   0          41m   10.244.0.176   k8s-node-01   <none>           <none>
web-0                                   1/1     Running   0          32s   10.244.1.167   k8s-node-02   <none>           <none>
web-1                                   1/1     Running   0          31s   10.244.0.188   k8s-node-01   <none>           <none>
web-2                                   1/1     Running   0          29s   10.244.1.168   k8s-node-02   <none>           <none>
web-3                                   1/1     Running   0          27s   10.244.0.189   k8s-node-01   <none>           <none>
web-4                                   1/1     Running   0          24s   10.244.1.169   k8s-node-02   <none>           <none>
web-5                                   1/1     Running   0          22s   10.244.0.190   k8s-node-01   <none>           <none>
web-6                                   1/1     Running   0          21s   10.244.1.170   k8s-node-02   <none>           <none>
web-7                                   1/1     Running   0          19s   10.244.0.191   k8s-node-01   <none>           <none>

查看存储上的数据目录

可以看出可以自动动态的分配nfs存储卷。以上即为k8s持久化存储之storageclass实践。

以上就是使用storageclass实现动态pv的具体步骤，内容较为全面，而且我也相信有相当的一些工具可能是我们日常工作可能会见到或用到的。通过这篇文章，希望你能收获更多。