全部文档
当前文档
最近更新时间：2024-08-13 17:02:40
您可以在金山云容器服务Kubernetes集群中使用金山云KFS存储卷。
目前，金山云提供两种kubernetes挂载方式：
静态存储卷
可以通过以下两种方式使用KFS文件存储静态存储卷：
动态存储卷
挂载文件系统（KFS）的前提是您有创建好的文件系统。如果您还未创建文件系统，您需要先创建文件系统。有关如何创建文件系统的详细信息，参见创建文件系统及挂载点
金山云KFS为共享存储，可以同时为多个 Pod 提供共享存储服务，即一个PVC可以同时被多个Pod 使用。
在没有卸载文件系统前，务必不要删除文件系统的挂载点，否则会造成操作系统Hang。
Flexvolume模式下使用KFS动态存储卷能力，若集群创建时间在2021-2-24及之后的集群可以直接使用，创建时间在2021-2-24之前，需要更新集群中Kube-system命名空间下disk-provisioner组件，YAML见附录
如上图所示：
server：10.0.1.xx
挂载路径：/cfs-eHhkjGxxxx
直接通过Volume使用
apiVersion: apps/v1
kind: Deployment
metadata:
name: kfs
spec:
replicas: 2
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx
volumeMounts:
- name: "kfs"
mountPath: "/data"
volumes:
- name: "kfs"
nfs:
server: "10.0.1.xx"
path: "/cfs-eHhkjGxxxx"通过PV/PVC使用
创建PV：
apiVersion: v1
kind: PersistentVolume
metadata:
name: kfs-pv
spec:
storageClassName: "kfs"
capacity:
storage: 100Mi
accessModes:
- ReadWriteMany
mountOptions:
- nfsvers=3
nfs:
server: "10.0.1.xx"
path: "/cfs-eHhkjGxxxx"创建PVC：
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: kfs-pvc
spec:
storageClassName: "kfs"
accessModes:
- ReadWriteMany
resources:
requests:
storage: 100Mi创建Deployment：
apiVersion: apps/v1
kind: Deployment
metadata:
name: kfs-deploy
spec:
replicas: 2
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx
volumeMounts:
- name: kfs
mountPath: "/data"
volumes:
- name: kfs
persistentVolumeClaim:
claimName: kfs-pvc参数 | 说明 |
|---|---|
server | 指定kfs挂载点地址（必填） |
path | 指定kfs挂载目录 (必填，子目录不存在时可自动创建) |
archiveOnDelete | 表示删除PVC、PV时候，处理kfs子目录的方式：如果reclaimPolicy为Delete，且archiveOnDelete为false：会直接删除远端目录和数据，请谨慎使用。如果reclaimPolicy为Delete，且archiveOnDelete为true：会将远端的目录更新为其他名字备份。如果reclaimPolicy为Retain，远端的目录不作处理。（选填，默认为false） |
storageType | ksc/kfs，指定使用kfs存储（必填） |
创建Storageclass：
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: ksc-kfs
parameters:
server: 10.0.1.xx
path: /cfs-eHhkjGxxxx/test-path
archiveOnDelete: "false"
storageType: ksc/kfs
provisioner: ksc/storage创建StatefulSet：
apiVersion: apps/v1
kind: StatefulSet
metadata:
name: web
spec:
serviceName: test
selector:
matchLabels:
app: nginx
replicas: 2
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx
volumeMounts:
- name: www
mountPath: "/data"
volumeClaimTemplates:
- metadata:
name: www
spec:
accessModes:
- ReadWriteOnce
storageClassName: ksc-kfs
resources:
requests:
storage: 10Gi说明：仅csi-driver 2.0.0及以上版本支持KFS，如您的组件版本过低，请提交工单联系后台进行组件版本升级。参数 | 说明 |
|---|---|
server | 指定kfs挂载点地址 |
share | 指定kfs挂载目录，同“前提->查看文件系统”中的挂载路径 |
reclaimPolicy | 持久卷删除时的回收策略，目前NFS仅支持Retain策略：当 pvc对象被删除时，pv卷仍然存在，对应的数据卷变为released状态，用户需要手动回收资源。 |
volumeBindingMode | 绑定持久卷的属性 |
mountOptions | 挂载kfs的options参数 |
创建StorageClass：
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: kfsplugin
provisioner: com.ksc.csi.nfsplugin
allowVolumeExpansion: false
parameters:
server: 10.0.1.xx
share: /cfs-eHhkjGxxxx
reclaimPolicy: Retain
volumeBindingMode: Immediate
mountOptions:
- vers=3
- nolock
- proto=tcp
- noresvport创建PVC：
apiVersion: "v1"
kind: "PersistentVolumeClaim"
metadata:
name: "pvc-kfs"
namespace: "default"
spec:
accessModes:
- "ReadWriteMany"
resources:
requests:
storage: "100Mi"
storageClassName: "kfsplugin"创建deployment：
apiVersion: apps/v1
kind: Deployment
metadata:
name: web-server-deployment-kfs
spec:
replicas: 2
selector:
matchLabels:
app: nginx
template:
metadata:
labels:
app: nginx
spec:
containers:
- name: nginx
image: nginx
volumeMounts:
- name: mypvc
mountPath: /usr/share/nginx/html
volumes:
- name: mypvc
persistentVolumeClaim:
claimName: pvc-kfs
readOnly: false验证
##使用命令查看pod、pvc状态，当pod状态为Running，pvc状态为Bound时绑定成功
kubectl get pods
kubectl get pvcdisk-provisioneryaml文件如下：
apiVersion: apps/v1
kind: Deployment
metadata:
name: disk-provisioner
namespace: kube-system
spec:
selector:
matchLabels:
app: disk-provisioner
replicas: 1
revisionHistoryLimit: 2
template:
metadata:
labels:
app: disk-provisioner
spec:
dnsPolicy: Default
tolerations:
# this taint is set by all kubelets running `--cloud-provider=external`
- key: "node.cloudprovider.kubernetes.io/uninitialized"
value: "true"
effect: "NoSchedule"
containers:
- image: hub.kce.ksyun.com/ksyun/disk-provisioner:latest
name: ebs-provisioner
env:
- name: OPENAPI_ENDPOINT
value: "internal.api.ksyun.com"
- name: OPENAPI_PREFIX
value: "http"
volumeMounts:
- name: kubeconfig
mountPath: /root/.kube/config
- name: clusterinfo
mountPath: /opt/app-agent/arrangement/clusterinfo
- image: hub.kce.ksyun.com/ksyun/disk-provisioner:latest
name: ksc-storage-provisioner
securityContext:
privileged: true # do mount
args:
- --provisioner=ksc/storage
env:
- name: OPENAPI_ENDPOINT
value: "internal.api.ksyun.com"
- name: OPENAPI_PREFIX
value: "http"
volumeMounts:
- name: kubeconfig
mountPath: /root/.kube/config
- name: clusterinfo
mountPath: /opt/app-agent/arrangement/clusterinfo
volumes:
- name: kubeconfig
hostPath:
path: /root/.kube/config
- name: clusterinfo
hostPath:
path: /opt/app-agent/arrangement/clusterinfo纯净模式