kubernetes的存储机制hostpath管理

hostPath

参考资料：

https://kubernetes.io/zh-cn/docs/concepts/storage/volumes/#hostpath

hostPath 可以将宿主机上的目录挂载到 Pod 中作为数据的存储目录

hostPath 一般用在如下场景：

容器应用程序中某些文件需要永久保存
Pod删除，hostPath数据对应在宿主机文件不受影响,即hostPath的生命周期和Pod不同,而和节点相同
宿主机和容器的目录都会自动创建
某些容器应用需要用到容器的自身的内部数据，可将宿主机的/var/lib/[docker|containerd]挂载到Pod中
容器复用宿主机的文件

hostPath 使用注意事项：

不支持数据的漫游，hostpath只有在其所在宿主机可用，其它节点不可用
不同宿主机的目录和文件内容不一定完全相同，所以Pod迁移前后的访问效果不一样
不适合Deployment这种分布式的资源，更适合于DaemonSet
宿主机的目录不属于独立的资源对象的资源，所以对资源设置的资源配额限制对hostPath目录无效

警告：
使用 hostPath 类型的卷存在许多安全风险。如果可以，你应该尽量避免使用 hostPath 卷。例如，你可以改为定义并使用 local PersistentVolume

#配置属性

kubectl explain pod.spec.volumes.hostPath

path  #指定哪个宿主机的目录或文件将被共享给Pod使用
type  #指定路径的类型，一共有7种，默认的类型是没有指定
   空字符串 #默认配置，在关联hostPath存储卷之前不进行任何检查,如果宿主机没有对应的目录，会自
动创建
   DirectoryOrCreate  #宿主机上不存在path，就自动创建0755权限的空目录
   Directory #path必须存在,挂载已存在的目录  
   FileOrCreate #宿主机上不存在挂载文件，就创建0644权限的空文件
   File #必须存在的文件 
 Socket #事先必须存在的Socket文件路径
   CharDevice #事先必须存在的字符设备文件路径
   BlockDevice #事先必须存在的块设备文件路径

#配置格式：

volumes:
 - name: volume_name
  hostPath:
    path: /path/to/host

宿主机磁盘目录是文件定义系统时区东八区

[root@master1 storage]# ll /usr/share/zoneinfo/Asia/Shanghai
-rw-r--r--. 5 root root 561 Nov  6  2025 /usr/share/zoneinfo/Asia/Shanghai
[root@master1 storage]# date
Fri May 22 11:00:30 PM CST 2026

[root@master1 storage]# timedatectl 
               Local time: Fri 2026-05-22 23:03:22 CST
           Universal time: Fri 2026-05-22 15:03:22 UTC
                 RTC time: Fri 2026-05-22 15:03:21
                Time zone: Asia/Shanghai (CST, +0800)
System clock synchronized: yes
              NTP service: active
          RTC in local TZ: no
[root@master1 storage]# ll /etc/localtime 
lrwxrwxrwx. 1 root root 35 May  7 15:28 /etc/localtime -> ../usr/share/zoneinfo/Asia/Shanghai
[root@master1 storage]#

nginx容器不定义时区默认是utc时间第二个容器是东八区时区

[root@master1 storage]# cat  storage-hostpath-timezone.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  name: storage-hostpath-timezone
spec:
  replicas: 1
  selector:
    matchLabels:
      app: storage-hostpath
  template:
    metadata:
      labels:
        app: storage-hostpath
    spec:
      containers:
      - name: c01
        image: registry.cn-beijing.aliyuncs.com/wangxiaochun/nginx:1.20.0
      - name: c02
        image: registry.cn-beijing.aliyuncs.com/wangxiaochun/nginx:1.20.0
        command: ["sh","-c","sleep 3600"]
        volumeMounts:
        - name: localtime
          mountPath: /etc/localtime #容器此为文件是普通文件，挂载节点目录成功
      volumes:
      - name: localtime
        hostPath:
          path: /usr/share/zoneinfo/Asia/Shanghai #配置时区方法2: 此文件用于影响时区，必选
          type: File

kubectl apply -f storage-hostpath-timezone.yaml

[root@master1 storage]# kubectl get pod 
NAME                                         READY   STATUS        RESTARTS        AGE
controller-deployment-demo-5f884bb8-2ppvf    1/1     Running       0               3m22s
controller-deployment-demo-5f884bb8-7svwh    1/1     Terminating   7 (112m ago)    5d23h
controller-deployment-demo-5f884bb8-fscrl    1/1     Running       9 (9m34s ago)   6d3h
controller-deployment-demo-5f884bb8-rfsbv    1/1     Running       8 (9m34s ago)   5d23h
storage-hostpath-timezone-7dd555df9b-pktvv   2/2     Running       0               11s

utc时间

[root@master1 storage]# kubectl exec -it storage-hostpath-timezone-7dd555df9b-pktvv -c c01 -- sh
# 
# date
Fri May 22 15:08:43 UTC 2026

# ls -l /etc/localtime
lrwxrwxrwx 1 root root 27 May 11  2021 /etc/localtime -> /usr/share/zoneinfo/Etc/UTC

[root@master1 storage]# kubectl exec -it storage-hostpath-timezone-7dd555df9b-pktvv -c c02 -- sh 
#  
# date
Fri May 22 23:10:18 CST 2026

范例: 实现 Redis 数据的持久化

[root@master1 storage]# cat storage-hostpath-redis.yaml  
apiVersion: apps/v1
kind: Deployment
metadata:
  name: hostpath-redis
spec:
  replicas: 1
  selector:
    matchLabels:
      app: hostpath-redis
  template:
    metadata:
      labels:
        app: hostpath-redis
    spec:
      nodeName: node1.org 
      containers:
      - name: hostpath-redis
        image: registry.cn-beijing.aliyuncs.com/wangxiaochun/redis:6.2.5
        volumeMounts:
        - name: redis-backup
          mountPath: /data
      volumes:
      - name: redis-backup
        hostPath:
          path: /backup/redis       #宿主机的目录会自动创建

nodeName: node1.org 只调度在这个节点上

#关键点：spec.containers.volumeMounts的name属性和spec.volumes的那么属性完全一致,因为他们是基于name来关联的。
#注意：redis的镜像将数据保存到了容器的/data 目录下。

目前宿主机上面没有这个目录

[root@node1 ~]# ls /backup/redis 
ls: cannot access '/backup/redis': No such file or directory

创建资源对象

kubectl apply -f storage-hostpath-redis.yaml

[root@master1 storage]# 
[root@master1 storage]# kubectl get node
NAME          STATUS     ROLES           AGE   VERSION
master1.org   Ready      control-plane   12d   v1.35.4
node1.org     Ready      <none>          12d   v1.35.4
node2.org     Ready      <none>          12d   v1.35.4

[root@master1 storage]# kubectl get pod -o wide
NAME                                        READY   STATUS    RESTARTS      AGE    IP             NODE        NOMINATED NODE   READINESS GATES
controller-deployment-demo-5f884bb8-2ppvf   1/1     Running   0             25m    10.244.1.153   node1.org   <none>           <none>
controller-deployment-demo-5f884bb8-fscrl   1/1     Running   9 (31m ago)   6d3h   10.244.1.147   node1.org   <none>           <none>
controller-deployment-demo-5f884bb8-rfsbv   1/1     Running   8 (31m ago)   6d     10.244.1.149   node1.org   <none>           <none>
hostpath-redis-dfcddb5b4-th7ph              1/1     Running   0             92s    10.244.1.155   node1.org   <none>           <none>

kubectl exec -it hostpath-redis-dfcddb5b4-th7ph -- bash

root@hostpath-redis-dfcddb5b4-th7ph:/data# redis-cli 
127.0.0.1:6379> config get dir
1) "dir"
2) "/data"

root@hostpath-redis-dfcddb5b4-th7ph:/data# pwd
/data

生成rdb文件

root@hostpath-redis-dfcddb5b4-th7ph:/data# redis-cli 
127.0.0.1:6379> 
127.0.0.1:6379> set name tom
OK
127.0.0.1:6379> save
OK
127.0.0.1:6379> exit
root@hostpath-redis-dfcddb5b4-th7ph:/data# ls
dump.rdb
root@hostpath-redis-dfcddb5b4-th7ph:/data#

存在宿主机上面

#在对应的worker节点主机查看数据目录自动生成

[root@node1 ~]# ls /backup/redis 
dump.rdb

#删除pod,重新创建

kubectl delete -f storage-hostpath-redis.yaml 
deployment.apps "hostpath-redis" deleted from default namespace

文件还在

[root@node1 ~]# ls /backup/redis 
dump.rdb

#重新创建Pod,验证数据是否还在

kubectl apply -f storage-hostpath-redis.yaml

[root@master1 storage]# kubectl get pod -o wide
NAME                                        READY   STATUS    RESTARTS      AGE    IP             NODE        NOMINATED NODE   READINESS GATES
controller-deployment-demo-5f884bb8-2ppvf   1/1     Running   0             34m    10.244.1.153   node1.org   <none>           <none>
controller-deployment-demo-5f884bb8-fscrl   1/1     Running   9 (40m ago)   6d3h   10.244.1.147   node1.org   <none>           <none>
controller-deployment-demo-5f884bb8-rfsbv   1/1     Running   8 (40m ago)   6d     10.244.1.149   node1.org   <none>           <none>
hostpath-redis-dfcddb5b4-lcqqd              1/1     Running   0             49s    10.244.1.156   node1.org   <none>           <none>

#确认数据还存在

[root@master1 storage]# kubectl exec -it hostpath-redis-dfcddb5b4-lcqqd  -- redis-cli keys '*'     
1) "name"
[root@master1 storage]# 
[root@master1 storage]# kubectl exec -it hostpath-redis-dfcddb5b4-lcqqd  -- redis-cli  get name    
"tom"

hostpath 持久化数据要固定写死pod 节点，不能让他调度别的节点，不然下次数据用不了，如果这个节点挂了，数据就丢了，不能真正实现完全安全

需要弄个备份机器专门做备份，写个脚本rsync同步数据加入定时任务定时备份数据到备份机器

posted @ 2026-05-22 23:56 minger_lcm 阅读(30) 评论(0) 收藏举报

刷新页面返回顶部

kubernetes的存储机制hostpath管理

公告