前言
在实验2.2 – Deployment中，我们将原来基于Pod部署的wordpress+mysql成功改造成基于Deployment部署，基于Deployment部署有很多好处，例如：支持滚动升级（Rolling Update），支持水平扩展。
不过有个问题，就是当我们的Deployment修改了，或者Pod删除重建了，数据也随之丢失了，这是我们不希望看到的，本篇文章我们就来尝试一下基于PV & PVC保存数据状态的有状态应用。
场景
对于单实例的有状态应用，我们可以定义Deployment并且replicas只能为1，用指定PVC的方式关联到一个PV上，使用PV提供的状态存储功能。如果要启动多实例，那么Deployment就无法胜任这个任务，必须使用到我们后面会讲到的StatefulSet+PVC Template的方式创建多实例对多PVC的模型。
本文实验所有的源码保存在：
https://github.com/zrbcool/blog-public/tree/master/k8s-hands-on/lab06
实战
PVC定义
lab06 git:(master) cat 03-wordpress-pvc.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
name: wordpress-pv-claim
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 2Gi
部署后查看状态
lab06 git:(master) kubectl apply -f 03-wordpress-pvc.yaml
persistentvolumeclaim/wordpress-pv-claim created
lab06 git:(master) kubectl get pvc
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
wordpress-mysql-pv-claim Pending 4s
发现PVC处于Pending状态，追查下原因
lab06 git:(master) kubectl describe pvc/wordpress-mysql-pv-claim
Name: wordpress-mysql-pv-claim
Namespace: default
StorageClass:
Status: Pending
Volume:
Labels: <none>
Annotations: kubectl.kubernetes.io/last-applied-configuration:
{"apiVersion":"v1","kind":"PersistentVolumeClaim","metadata":{"annotations":{},"name":"wordpress-mysql-pv-claim","namespace":"default"},"s...
Finalizers: [kubernetes.io/pvc-protection]
Capacity:
Access Modes:
VolumeMode: Filesystem
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal FailedBinding 6s (x4 over 36s) persistentvolume-controller no persistent volumes available for this claim and no storage class is set
Mounted By: <none>
接下来我们创建PV
PV定义
查看定义
lab06 git:(master) cat 04-wordpress-pv.yaml
kind: PersistentVolume
apiVersion: v1
metadata:
name: wordpress-mysql-pv-volume
labels:
type: local
spec:
capacity:
storage: 2Gi
accessModes:
- ReadWriteOnce
hostPath:
path: "/data/pv/wordpress/mysql"
该pv定义是基于hostPath的方式，所以我们要在节点上提前创建好目录，如下：
lab06 git:(master) mkdir -p /data/pv/wordpress/mysql
lab06 git:(master) kubectl create -f 04-wordpress-pv.yaml
persistentvolume/wordpress-mysql-pv-volume created
lab06 git:(master) kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
wordpress-mysql-pv-volume 2Gi RWO Retain Bound default/wordpress-mysql-pv-claim 31s
lab06 git:(master) kubectl get pvc
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
wordpress-mysql-pv-claim Bound wordpress-mysql-pv-volume 2Gi RWO 118s
此时我们之前定义的pvc也已经成功绑定了
使Deployment使用该PVC
查看定义
lab06 git:(master) cat 01-wordpress-mysql-deployment.yaml
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
labels:
app: wordpress
name: wordpress
spec:
replicas: 1
selector:
matchLabels:
app: wordpress
template:
metadata:
labels:
app: wordpress
spec:
containers:
- image: wordpress:latest
imagePullPolicy: IfNotPresent
name: wordpress
env:
- name: WORDPRESS_DB_HOST
value: "127.0.0.1"
- name: WORDPRESS_DB_USER
value: "root"
- name: WORDPRESS_DB_PASSWORD
value: "passw0rd"
- image: mysql:5.7.26
imagePullPolicy: IfNotPresent
name: mysql
env:
- name: MYSQL_ROOT_PASSWORD
value: "passw0rd"
- name: MYSQL_DATABASE
value: "wordpress"
volumeMounts:
- name: mysql-persistent-storage
mountPath: /var/lib/mysql
volumes:
- name: mysql-persistent-storage
persistentVolumeClaim:
claimName: wordpress-mysql-pv-claim
执行更新并查看效果
lab06 git:(master) kubectl apply -f 01-wordpress-mysql-deployment.yaml
deployment.extensions/wordpress configured
lab06 git:(master) kubectl get pod
NAME READY STATUS RESTARTS AGE
wordpress-6cfd879fcd-9mlhb 2/2 Running 0 29s
lab06 git:(master) ls -l /data/pv/wordpress/mysql/
total 188480
-rw-r----- 1 999 999 56 Jun 10 23:42 auto.cnf
-rw------- 1 999 999 1675 Jun 10 23:42 ca-key.pem
-rw-r--r-- 1 999 999 1107 Jun 10 23:42 ca.pem
...
drwxr-x--- 2 999 999 12288 Jun 10 23:42 sys
drwxr-x--- 2 999 999 4096 Jun 10 23:42 wordpress
可见，Pod已经在/data/pv/wordpress/mysql/下产生数据了，接下来我们来试一下配置数据后，删除Pod，使Deployment控制Pod重建，配置数据是否能够保存
通过下面命令删除Pod使其重建
lab06 git:(master) kubectl get pod
NAME READY STATUS RESTARTS AGE
wordpress-6cfd879fcd-9mlhb 2/2 Running 0 7m34s
lab06 git:(master) kubectl delete pod/wordpress-6cfd879fcd-9mlhb
pod "wordpress-6cfd879fcd-9mlhb" deleted
lab06 git:(master) kubectl get pod
NAME READY STATUS RESTARTS AGE
wordpress-6cfd879fcd-29qg9 2/2 Running 1 11s
刷新网页，发布的测试文件还在，测试成功
![]()
新问题
查看官网说明：
![]()
意思是说，对于这个单实例的有状态Deployment不要对其进行扩容，同时当Deployment定义发生更新时也不能使用滚动更新，而应该先删除再创建，也就是strategy: type: Recreate
修改后的文件如下：
lab06 git:(master) cat 01-wordpress-mysql-deployment.yaml
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
labels:
app: wordpress
name: wordpress
spec:
strategy:
type: Recreate
replicas: 1
selector:
matchLabels:
app: wordpress
template:
metadata:
labels:
app: wordpress
spec:
containers:
- image: wordpress:latest
imagePullPolicy: IfNotPresent
name: wordpress
env:
- name: WORDPRESS_DB_HOST
value: "127.0.0.1"
- name: WORDPRESS_DB_USER
value: "root"
- name: WORDPRESS_DB_PASSWORD
value: "passw0rd"
- image: mysql:5.7.26
imagePullPolicy: IfNotPresent
name: mysql
env:
- name: MYSQL_ROOT_PASSWORD
value: "passw0rd"
- name: MYSQL_DATABASE
value: "wordpress"
volumeMounts:
- name: mysql-persistent-storage
mountPath: /var/lib/mysql
volumes:
- name: mysql-persistent-storage
persistentVolumeClaim:
claimName: wordpress-mysql-pv-claim
清除数据
lab06 git:(master) kubectl delete -f .
deployment.extensions "wordpress" deleted
service "wordpress-svc" deleted
persistentvolumeclaim "wordpress-mysql-pv-claim" deleted
persistentvolume "wordpress-mysql-pv-volume" deleted
