k8s持久化存储pv、pvc
1.PV和PVC持久化存储简介
概念：一个应用肯定会跑多个pod，多个pod里面肯定会存放数据，当一个pod死了，数据也会跟着消失，因此k8s就有了pv的概念，pv就是持久化存储，相当于docker的volume，pvc就是分配pv的容量给pod。
在docker的设计实现中，容器中的数据是临时的，即当容器被销毁时，其中的数据将会小时，如果需要持久化数据，需要使用docker数据卷挂载宿主机上的文件或者目录到容器中，在kubernetes中，当pod重建的时候，数据也会丢失，kubernetes也是通过数据卷来挂载pod数据来实现持久化
在编写pvc配置文件时，如果不指定pv是谁，默认随机分配
2.部署基于nfs的持久化存储
2.1.部署nfs
1)在master节点安装nfs
[root@k8s-master ~]# yum -y install nfs-utils
2)准备nfs存储目录
[root@k8s-master ~]# mkdir /data2/kod_pv -p
3)准备配置文件
[root@k8s-master ~]# vim /etc/exports
/data2 192.168.81.0/24(rw,sync,no_root_squash,no_all_squash)
4)启动服务
[root@k8s-master ~]# systemctl restart rpcbind
[root@k8s-master ~]# systemctl enable rpcbind
[root@k8s-master ~]# systemctl restart nfs
[root@k8s-master ~]# systemctl enable nfs
5)查看nfs信息
[root@k8s-master ~]# showmount -e
Export list for k8s-master:
/data2 192.168.81.0/24
6)在node节点安装nfs客户端
[root@k8s-node1 ~]# yum -y install nfs-utils
[root@k8s-node2 ~]# yum -y install nfs-utils
2.2.编写pv yaml文件
1)编写yaml文件
[root@k8s-master pv]# vim k8s_pv.yaml
apiVersion: v1 //api版本
kind: PersistentVolume //资源类型是pv
metadata:
name: kod-pv //pv的名字
labels: //标签，用于pvc指定
type: kod-pv
spec: //pv的属性
capacity:
storage: 10Gi //pv的容量
accessModes: //pv的访问模式
- ReadWriteMany //允许多个客户端读写
persistentVolumeReclaimPolicy: Recycle //pvc的回收策略
nfs: //用nfs提供存储空间
path: "/data2/kod_pv" //nfs存储目录
server: 192.168.81.210 //nfs地址
readOnly: false //关闭只读
2)查看pv
状态为Available
[root@k8s-master pv]# kubectl get pv
NAME CAPACITY ACCESSMODES RECLAIMPOLICY STATUS CLAIM REASON AGE
kod-pv 10Gi RWX Recycle Available 18s
3)查看pv详细信息
[root@k8s-master pv]# kubectl describe pv kod-pv
Name: kod-pv
Labels: type=kod-pv
StorageClass:
Status: Available
Claim:
Reclaim Policy: Recycle
Access Modes: RWX
Capacity: 10Gi
Message:
Source:
Type: NFS (an NFS mount that lasts the lifetime of a pod)
Server: 192.168.81.210
Path: /data2/kod_pv
ReadOnly: false
No events.
2.3.编写pvc yaml文件
1)编写pvc yaml文件
[root@k8s-master pv]# vim k8s_pvc.yaml
apiVersion: v1 //api版本
kind: PersistentVolumeClaim //资源类型是pvc
metadata:
name: kod-pvc //pvc的名称
spec:
accessModes: //pvc的权限
- ReadWriteMany //多个客户端可读可写
resources:
requests:
storage: 5Gi //pvc的资源限制为5G
2)创建pvc
[root@k8s-master pv]# kubectl create -f k8s_pvc.yaml
3)查看pvc的状态
volume对应刚刚创建的kod-pv，在yaml文件中不指定那个pv默认随机分配
[root@k8s-master pv]# kubectl get pvc
NAME STATUS VOLUME CAPACITY ACCESSMODES AGE
kod-pvc Bound kod-pv 10Gi RWX 14s
4)查看pv的状态
现在pv的状态为bound
[root@k8s-master pv]# kubectl get pv
NAME CAPACITY ACCESSMODES RECLAIMPOLICY STATUS CLAIM REASON AGE
kod-pv 10Gi RWX Recycle Bound default/kod-pvc 10m
5)查看pvc的详细
[root@k8s-master pv]# kubectl describe pvc kod-pvc
Name: kod-pvc
Namespace: default
StorageClass:
Status: Bound
Volume: kod-pv //pv的是谁
Labels: <none>
Capacity: 10Gi
Access Modes: RWX
No events.
2.4.指定pvc使用哪个pv
实现思路，可以用标签进行绑定实现
2.4.1.创建多个pv
1)编写yaml文件
[root@k8s-master pv]# vi k8s_pv.yaml
apiVersion: v1
kind: PersistentVolume
metadata:
name: kod-pv01
labels:
pv: kod-pv01 //定义pv标签，用于pvc关联
type: kod-pv01
spec:
capacity:
storage: 10Gi
accessModes:
- ReadWriteMany
persistentVolumeReclaimPolicy: Recycle
nfs:
path: "/data2/kod_pv01"
server: 192.168.81.210
readOnly: false
---
apiVersion: v1
kind: PersistentVolume
metadata:
name: kod-pv02
labels:
pv: kod-pv02 //定义pv标签，用于pvc关联
type: kod-pv02
spec:
capacity:
storage: 10Gi
accessModes:
- ReadWriteMany
persistentVolumeReclaimPolicy: Recycle
nfs:
path: "/data2/kod_pv02"
server: 192.168.81.210
readOnly: false
2)创建
[root@k8s-master pv_kz]# kubectl create -f k8x_pv.yaml
persistentvolume "kod-pv01" created
persistentvolume "kod-pv02" created
3)查看
[root@k8s-master pv_kz]# kubectl get pv
NAME CAPACITY ACCESSMODES RECLAIMPOLICY STATUS CLAIM REASON AGE
kod-pv 10Gi RWX Recycle Bound default/kod-pvc 30m
kod-pv01 10Gi RWX Recycle Available 8s
kod-pv02 10Gi RWX Recycle Available 8s
最主要的地方
2.4.2.创建多个pvc分别指定不同的pv
[root@k8s-master pv_kz]# vim k8x_pvc.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: kod-pvc01
spec:
accessModes:
- ReadWriteMany
resources:
requests:
storage: 5Gi
selector:
matchLabels:
pv: kod-pv01 //定义pv标签，与pv进行关联
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: kod-pvc02
spec:
accessModes:
- ReadWriteMany
resources:
requests:
storage: 5Gi
selector:
matchLabels:
pv: kod-pv02 //定义pv标签，与pv进行关联
2.4.3.验证
会发现pvc01关联pv01，pvc02关联pv02
6.5.删除pv、pvc
pv
[root@k8s-master pv_kz]# kubectl delete pv kod-pv01 kod-pv02
persistentvolume "kod-pv01" deleted
persistentvolume "kod-pv02" deleted
pvc
[root@k8s-master pv_kz]# kubectl delete pvc kod-pvc01
persistentvolumeclaim "kod-pvc01" deleted
6.6.将pvc挂载到应用上
1)挂载pv
[root@k8s-master kod]# vim kod_deployment.yaml
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: kod-deploy
spec:
replicas: 5
template:
metadata:
labels:
app: kodnginx
spec:
containers:
- name: nginxkod
image: 192.168.81.240/k8s/kodexplorer:v4
ports:
- containerPort: 80
volumeMounts: //挂载pv
- name: kod-data //挂载资源的名称，随便起
mountPath: /web/kodexporer/data //挂载目录
volumes:
- name: kod-data //资源名称和volumemounts中的name对应
persistentVolumeClaim:
claimName: kod-pvc01 //pvc的名称
2)更新配置
[root@k8s-master kod]# kubectl apply -f kod_deployment.yaml
deployment "kod-deploy" configured
3)由于挂载问题导致没权限，进入容器修改权限
[root@k8s-master kod]# kubectl exec -it kod-deploy-2316599975-1z3l3 /bin/bash
[root@kod-deploy-2316599975-1z3l3 web]# chmod -R 777 /web/
[root@k8s-master kod]# kubectl exec -it kod-deploy-2316599975-9cxr1 /bin/bash
[root@kod-deploy-2316599975-9cxr1 kodexporer]# chmod -R 777 /web/
[root@k8s-master kod]# kubectl exec -it kod-deploy-2316599975-cfb1d /bin/bash
[root@kod-deploy-2316599975-cfb1d kodexporer]# chmod -R 777 /web/
[root@k8s-master kod]# kubectl exec -it kod-deploy-2316599975-dbhwb /bin/bash
[root@kod-deploy-2316599975-dbhwb kodexporer]# chmod -R 777 /web/
[root@k8s-master kod]# kubectl exec -it kod-deploy-2316599975-hdh2t /bin/bash
[root@kod-deploy-2316599975-hdh2t kodexporer]# chmod -R 777 /web/
不使用pv之前，上传的数据，当一个pod死了之后数据也会消失，使用pv之后pod死了数据也不会丢
