我實驗了一下kubenetes的dynamic volume，基於ceph RDB. 以下是使用步驟：

1. 準備storageclass, 配置ceph rdb的資訊， ceph mon的ip, 使用者，密碼和ceph pool， 之前secret要生成。

storageclass:

apiVersion: storage.k8s.io/v1beta1

kind: StorageClass

metadata:

 name: kubepool

 annotations:

   storageclass.beta.kubernetes.io/is-default-class: 'true'

provisioner: kubernetes.io/rbd

parameters:

   monitors: 10.0.200.11:6789

   adminId: kube

   adminSecretName: ceph-secret

   adminSecretNamespace: default

   pool: kube

   userId: kube

   userSecretName: ceph-secret

 2. 準備 persistent volume claim,

kind: PersistentVolumeClaim

apiVersion: v1

metadata:

 name: ceph-claim-sc

 annotations:

   volume.beta.kubernetes.io/storage-class: 'kubepool'

spec:

 accessModes:

    -ReadWriteOnce

 resources:

   requests:

     storage: 20Gi

 3. 檢查kube resource 和 rbd的image. 發現image已經在pool裡被建立，並且格式化。預設是ext4

[[email protected] kube]# kubectl getstorageclass

NAME                 TYPE

kubepool (default)   kubernetes.io/rbd  

[[email protected] kube]# kubectl get pvc

NAME            STATUS    VOLUME                                    CAPACITY   ACCESSMODES   AGE

ceph-claim      Bound    ceph-pv                                    50Gi       RWO           3h

ceph-claim-sc   Bound    pvc-ac668f99-3b8b-11e7-8af9-fa163e01317b   20Gi      RWO           7m

[[email protected] kube]# kubectl get pv

NAME                                      CAPACITY   ACCESSMODES   RECLAIMPOLICY   STATUS   CLAIM                  REASON    AGE

ceph-pv                                    50Gi       RWO           Recycle         Bound     default/ceph-claim                3h

pvc-ac668f99-3b8b-11e7-8af9-fa163e01317b   20Gi      RWO           Delete          Bound     default/ceph-claim-sc             7m

 note: delete reclaimpolicy， 預設的policy, 意思是pvc被刪除，相應的ceph rbd image也將被刪除

 [[email protected] kube]# rbd -p kube -nclient.kube ls

kubernetes-dynamic-pvc-ac6b857a-3b8b-11e7-bdfc-fa163e01317b

vol1

vol2

vol50

4.  建立rc， 驗證是否能夠使用volume

apiVersion: v1
kind: ReplicationController
metadata:
  name: frontendpvcsc
  labels:
    name: frontendpvcsc
spec:
  replicas: 1
  selector:
    name: frontendpvcsc
  template:
    metadata:
      labels:
        name: frontendpvcsc
    spec:
      containers:
      - name: frontendpvcsc
        image: kubeguide/guestbook-php-frontend
        env:
        - name: GET_HOSTS_FROM
          value: env
        ports:
        - containerPort: 80
        volumeMounts:
        - mountPath: /mnt/rbd
          name: ceph-vol
      volumes:
      - name: ceph-vol
        persistentVolumeClaim:
          claimName: ceph-claim-sc

[[email protected] kube]# kubectl exec frontendpvcsc-xzz15 -it bash
[email protected]:/var/www/html# df -h
Filesystem                                                                                        Size  Used Avail Use% Mounted on
/dev/mapper/docker-253:1-528252-1915f387c1f17925e19bbcaa4324e401cc7c1abb5e86a11ee6bddda38f0db1da   10G  609M  9.4G   6% /
tmpfs                                                                                             3.9G     0  3.9G   0% /dev
tmpfs                                                                                             3.9G     0  3.9G   0% /sys/fs/cgroup
/dev/vda1                                                                                          19G  4.9G   14G  27% /etc/hosts
/dev/rbd0                                                                                          20G   45M   19G   1% /mnt/rbd
shm                                                                                                64M     0   64M   0% /dev/shm

能夠看到volume已經被mount上來了。

總結：

Dynamic volume能夠很方便的被使用者使用，像一個儲存資源池，能動態的分配儲存資源。