提示：statefulset控制器主要由pod模版和pvc模版组成；其中pod模版主要定义pod相关属性信息，对应pvc模版主要用来为对应pod提供存储卷，该存储卷可以使用sc资源来动态创建并关联pv，也可以管理员手动创建并关联对应的pv；

3、statefulset控制器的创建和使用

[root@master01 ~]# cat statefulset-demo.yaml
apiVersion: v1
kind: Service
metadata:
name: nginx
labels:
app: nginx
spec:
ports:

• port: 80
  name: web
  clusterIP: None
  selector:
  app: nginx

---

apiVersion: apps/v1
kind: StatefulSet
metadata:
name: web
spec:
selector:
matchLabels:
app: nginx
serviceName: nginx
replicas: 3
template:
metadata:
labels:
app: nginx
spec:
terminationGracePeriodSeconds: 10
containers:
- name: nginx
image: nginx:1.14-alpine
ports:
- containerPort: 80
name: web
volumeMounts:
- name: www
mountPath: /usr/share/nginx/html
volumeClaimTemplates:

• metadata:
  name: www
  spec:
  accessModes: [ “ReadWriteOnce” ]
  resources:
  requests:
  storage: 1Gi
  [root@master01 ~]#
  　　提示：statefulset控制器依赖handless类型service来管理pod的访问；statefulset控制器会根据定义的副本数量和定义的pod模板和pvc模板来启动pod，并给每个pod分配一个固定不变的名称，一般这个名称都是statefulset名称加上一个索引id，如上述清单，它会创建3个pod，这三个pod的名称分别是web-0、web-1、web-2；pod名称会结合handless service为其每个pod分配一个dns子域，访问对应pod就可以直接用这个子域名访问即可；子域名格式为$(po{d}_{n}ame).$(service_name).namespace_name.svc.集群域名（如果在初始化为指定默认集群域名为cluster.local）;上述清单定义了一个handless服务，以及一个statefulset控制器，对应控制器下定义了一个pod模板，和一个pvc模板；其中在pod模板中的terminationGracePeriodSeconds字段用来指定终止容器的宽限期时长，默认不指定为30秒；定义pvc模板需要用到volumeClaimTemplates字段，该字段的值为一个对象列表；其内部我们可以定义pvc模板；如果后端存储支持动态供给pv，还可以在此模板中直接调用对应的sc资源；

在nfs服务器上导出共享目录

[root@docker_registry ~]# cat /etc/exports
/data/v1 192.168.0.0/24(rw,no_root_squash)
/data/v2 192.168.0.0/24(rw,no_root_squash)
/data/v3 192.168.0.0/24(rw,no_root_squash)
[root@docker_registry ~]# ll /data/v*
/data/v1:
total 0

/data/v2:
total 0

/data/v3:
total 0
[root@docker_registry ~]# exportfs -av
exporting 192.168.0.0/24:/data/v3
exporting 192.168.0.0/24:/data/v2
exporting 192.168.0.0/24:/data/v1
[root@docker_registry ~]#
　　手动创建pv

[root@master01 ~]# cat pv-demo.yaml
apiVersion: v1
kind: PersistentVolume
metadata:
name: nfs-pv-v1
labels:
storsystem: nfs
rel: stable
spec:
capacity:
storage: 1Gi
volumeMode: Filesystem
accessModes: [“ReadWriteOnce”,“ReadWriteMany”,“ReadOnlyMany”]
persistentVolumeReclaimPolicy: Retain
mountOptions:

• hard
• nfsvers=4.1
  nfs:
  path: /data/v1
  server: 192.168.0.99

---

apiVersion: v1
kind: PersistentVolume
metadata:
name: nfs-pv-v2
labels:
storsystem: nfs
rel: stable
spec:
capacity:
storage: 1Gi
volumeMode: Filesystem
accessModes: [“ReadWriteOnce”,“ReadWriteMany”,“ReadOnlyMany”]
persistentVolumeReclaimPolicy: Retain
mountOptions:

• hard
• nfsvers=4.1
  nfs:
  path: /data/v2
  server: 192.168.0.99

---

apiVersion: v1
kind: PersistentVolume
metadata:
name: nfs-pv-v3
labels:
storsystem: nfs
rel: stable
spec:
capacity:
storage: 1Gi
volumeMode: Filesystem
accessModes: [“ReadWriteOnce”,“ReadWriteMany”,“ReadOnlyMany”]
persistentVolumeReclaimPolicy: Retain
mountOptions:

• hard
• nfsvers=4.1
  nfs:
  path: /data/v3
  server: 192.168.0.99
  [root@master01 ~]#
  　　提示：手动创建pv需要将其pv回收策略设置为Retain，以免对应pod删除以后，对应pv变成release状态而导致不可用；

应用配置清单

[root@master01 ~]# kubectl apply -f pv-demo.yaml
persistentvolume/nfs-pv-v1 created
persistentvolume/nfs-pv-v2 created
persistentvolume/nfs-pv-v3 created
[root@master01 ~]# kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
nfs-pv-v1 1Gi RWO,ROX,RWX Retain Available 3s
nfs-pv-v2 1Gi RWO,ROX,RWX Retain Available 3s
nfs-pv-v3 1Gi RWO,ROX,RWX Retain Available 3s
[root@master01 ~]#
　　提示：如果后端存储支持动态供给pv，此步骤可以省略；直接创建sc资源对象，然后在statefulset资源清单中的pvc模板中引用对应的sc对象名称就可以实现动态供给pv并绑定对应的pvc；

应用statefulset资源清单

[root@master01 ~]# kubectl apply -f statefulset-demo.yaml
service/nginx created
statefulset.apps/web created
[root@master01 ~]# kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
nfs-pv-v1 1Gi RWO,ROX,RWX Retain Bound default/www-web-0 4m7s
nfs-pv-v2 1Gi RWO,ROX,RWX Retain Bound default/www-web-1 4m7s
nfs-pv-v3 1Gi RWO,ROX,RWX Retain Bound default/www-web-2 4m7s
[root@master01 ~]# kubectl get pvc
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
www-web-0 Bound nfs-pv-v1 1Gi RWO,ROX,RWX 14s
www-web-1 Bound nfs-pv-v2 1Gi RWO,ROX,RWX 12s
www-web-2 Bound nfs-pv-v3 1Gi RWO,ROX,RWX 7s
[root@master01 ~]# kubectl get sts
NAME READY AGE
web 3/3 27s
[root@master01 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 38s
web-1 1/1 Running 0 36s
web-2 1/1 Running 0 31s
[root@master01 ~]# kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.96.0.1 443/TCP 48m
nginx ClusterIP None 80/TCP 41s
[root@master01 ~]#
　　提示：可以看到应用statefulset资源清单以后，对应的pv从available状态变为了bound状态，并且自动创建了3个pvc，对应pod的名称不再是控制器名称加一串随机字符串，而是statefulset控制器名称加一个有序的数字；通常这个数字从0开始，依次向上加，我们把这个数字叫做对应pod的索引；

查看statefulset控制器详细信息

[root@master01 ~]# kubectl describe sts web
Name: web
Namespace: default
CreationTimestamp: Mon, 28 Dec 2020 19:34:11 +0800
Selector: app=nginx
Labels:
Annotations:
Replicas: 3 desired | 3 total
Update Strategy: RollingUpdate
Partition: 0
Pods Status: 3 Running / 0 Waiting / 0 Succeeded / 0 Failed
Pod Template:
Labels: app=nginx
Containers:
nginx:
Image: nginx:1.14-alpine
Port: 80/TCP
Host Port: 0/TCP
Environment:
Mounts:
/usr/share/nginx/html from www (rw)
Volumes:
Volume Claims:
Name: www
StorageClass:
Labels:
Annotations:
Capacity: 1Gi
Access Modes: [ReadWriteOnce]
Events:
Type Reason Age From Message

---

Normal SuccessfulCreate 5m59s statefulset-controller create Claim www-web-0 Pod web-0 in StatefulSet web success
Normal SuccessfulCreate 5m59s statefulset-controller create Pod web-0 in StatefulSet web successful
Normal SuccessfulCreate 5m57s statefulset-controller create Claim www-web-1 Pod web-1 in StatefulSet web success
Normal SuccessfulCreate 5m57s statefulset-controller create Pod web-1 in StatefulSet web successful
Normal SuccessfulCreate 5m52s statefulset-controller create Claim www-web-2 Pod web-2 in StatefulSet web success
Normal SuccessfulCreate 5m52s statefulset-controller create Pod web-2 in StatefulSet web successful
[root@master01 ~]#
　　提示：从上面的详细信息中可以了解到，对应statefulset控制器是创建一个pvc，然后在创建一个pod；只有当第一个pod和pvc都成功并就绪以后，对应才会进行下一个pvc和pod的创建和挂载；简单讲它里面是有序串行进行的；

验证：在k8s集群上任意节点查看对应nginx服务名称，看看是否能够查到对应服务名称域名下的pod记录

安装dns工具包

1
[root@master01 ~]# yum install -y bind-utils
　　用dig工具查看对应nginx.default.cluster.local域名在coredns上的解析记录

[root@master01 ~]# kubectl get svc
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kubernetes ClusterIP 10.96.0.1 443/TCP 64m
nginx ClusterIP None 80/TCP 20m
[root@master01 ~]# kubectl get svc -n kube-system
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
kube-dns ClusterIP 10.96.0.10 53/UDP,53/TCP,9153/TCP 20d
[root@master01 ~]# dig nginx.default.svc.cluster.local @10.96.0.10

; <<>> DiG 9.11.4-P2-RedHat-9.11.4-26.P2.el7_9.3 <<>> nginx.default.svc.cluster.local @10.96.0.10
;; global options: +cmd
;; Got answer:
;; WARNING: .local is reserved for Multicast DNS
;; You are currently testing what happens when an mDNS query is leaked to DNS
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 61539
;; flags: qr aa rd; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 1
;; WARNING: recursion requested but not available

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;nginx.default.svc.cluster.local. IN A

;; ANSWER SECTION:
nginx.default.svc.cluster.local. 30 IN A 10.244.2.109
nginx.default.svc.cluster.local. 30 IN A 10.244.4.27
nginx.default.svc.cluster.local. 30 IN A 10.244.3.108

;; Query time: 0 msec
;; SERVER: 10.96.0.10#53(10.96.0.10)
;; WHEN: Mon Dec 28 19:54:34 CST 2020
;; MSG SIZE rcvd: 201

[root@master01 ~]# kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
web-0 1/1 Running 0 22m 10.244.4.27 node04.k8s.org
web-1 1/1 Running 0 22m 10.244.2.109 node02.k8s.org
web-2 1/1 Running 0 22m 10.244.3.108 node03.k8s.org
[root@master01 ~]#
　　提示：从上面的查询结果可以看到，对应default名称空间下nginx服务名称，对应在coredns上的记录有3条；并且对应的解析记录就是对应服务后端的podip地址；

验证：查询web-0的记录是否是对应的web-0这个pod的ip地址呢？

[root@master01 ~]# kubectl get pods web-0 -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
web-0 1/1 Running 0 24m 10.244.4.27 node04.k8s.org
[root@master01 ~]# dig web-0.nginx.default.svc.cluster.local @10.96.0.10

; <<>> DiG 9.11.4-P2-RedHat-9.11.4-26.P2.el7_9.3 <<>> web-0.nginx.default.svc.cluster.local @10.96.0.10
;; global options: +cmd
;; Got answer:
;; WARNING: .local is reserved for Multicast DNS
;; You are currently testing what happens when an mDNS query is leaked to DNS
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 13000
;; flags: qr aa rd; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1
;; WARNING: recursion requested but not available

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;web-0.nginx.default.svc.cluster.local. IN A

;; ANSWER SECTION:
web-0.nginx.default.svc.cluster.local. 30 IN A 10.244.4.27

;; Query time: 0 msec
;; SERVER: 10.96.0.10#53(10.96.0.10)
;; WHEN: Mon Dec 28 19:58:58 CST 2020
;; MSG SIZE rcvd: 119

[root@master01 ~]#
　　提示：可以看到在对应查询服务名称域名前加上对应的pod名称，在coredns上能够查询到对应pod的ip记录，这说明，后续访问我们可以直接通过pod名称加服务名称直接访问到对应pod；

验证：把集群节点的dns服务器更改为coredns服务ip，然后使用pod名称加服务名称域名的方式访问pod，看看是否能够访问到pod?

[root@master01 ~]# cat /etc/resolv.conf

Generated by NetworkManager

search k8s.org
nameserver 10.96.0.10
[root@master01 ~]# curl web-0.nginx.default.svc.cluster.local

403 Forbidden

---

nginx/1.14.2 [root@master01 ~]# 　　提示：这里能够响应403，说明pod能够正常访问，只不过对应pod没有主页，所以提示403；

验证：进入对应pod里面，提供主页页面，再次访问，看看是否能够访问到对应的页面内容呢？

[root@master01 ~]# kubectl exec -it web-0 – /bin/sh
/ # cd /usr/share/nginx/html/
/usr/share/nginx/html # ls
/usr/share/nginx/html # echo “this web-0 pod index” > index.html
/usr/share/nginx/html # ls
index.html
/usr/share/nginx/html # cat index.html
this web-0 pod index
/usr/share/nginx/html # exit
[root@master01 ~]# curl web-0.nginx.default.svc.cluster.local
this web-0 pod index
[root@master01 ~]#
　　提示：可以看到对应pod能够被访问到；

删除web-0，看看对应pod是否自动重建？

[root@master01 ~]# kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
web-0 1/1 Running 0 33m 10.244.4.27 node04.k8s.org
web-1 1/1 Running 0 33m 10.244.2.109 node02.k8s.org
web-2 1/1 Running 0 33m 10.244.3.108 node03.k8s.org
[root@master01 ~]# kubectl delete pod web-0
pod “web-0” deleted
[root@master01 ~]# kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
web-0 1/1 Running 0 7s 10.244.4.28 node04.k8s.org
web-1 1/1 Running 0 33m 10.244.2.109 node02.k8s.org
web-2 1/1 Running 0 33m 10.244.3.108 node03.k8s.org
[root@master01 ~]#
　　提示：可以看到手动删除web-0以后，对应控制器会自动根据pod模板，重建一个名称为web-0的pod运行起来；

验证：访问新建后的pod，看看是否能够访问到对应的主页呢？

1
2
3
[root@master01 ~]# curl web-0.nginx.default.svc.cluster.local
this web-0 pod index
[root@master01 ~]#
　　提示：可以看到使用pod名称加服务名称域名，能够正常访问到对应pod的主页，这意味着新建的pod能够自动将之前删除的pod的pvc存储卷挂载到自己对一个的目录下；

扩展pod副本

在nfs服务器上创建共享目录，并导出对应的目录
[root@docker_registry ~]# mkdir -pv /data/v{4,5,6}
mkdir: created directory ‘/data/v4’
mkdir: created directory ‘/data/v5’
mkdir: created directory ‘/data/v6’
[root@docker_registry ~]# echo “/data/v4 192.168.0.0/24(rw,no_root_squash)” >> /etc/exports
[root@docker_registry ~]# echo “/data/v5 192.168.0.0/24(rw,no_root_squash)” >> /etc/exports
[root@docker_registry ~]# echo “/data/v6 192.168.0.0/24(rw,no_root_squash)” >> /etc/exports
[root@docker_registry ~]# cat /etc/exports
/data/v1 192.168.0.0/24(rw,no_root_squash)
/data/v2 192.168.0.0/24(rw,no_root_squash)
/data/v3 192.168.0.0/24(rw,no_root_squash)
/data/v4 192.168.0.0/24(rw,no_root_squash)
/data/v5 192.168.0.0/24(rw,no_root_squash)
/data/v6 192.168.0.0/24(rw,no_root_squash)
[root@docker_registry ~]# exportfs -av
exporting 192.168.0.0/24:/data/v6
exporting 192.168.0.0/24:/data/v5
exporting 192.168.0.0/24:/data/v4
exporting 192.168.0.0/24:/data/v3
exporting 192.168.0.0/24:/data/v2
exporting 192.168.0.0/24:/data/v1
[root@docker_registry ~]#
　　复制创建pv的资源清单，更改为要创建pv对应的配置

[root@master01 ~]# cat pv-demo2.yaml
apiVersion: v1
kind: PersistentVolume
metadata:
name: nfs-pv-v4
labels:
storsystem: nfs
rel: stable
spec:
capacity:
storage: 1Gi
volumeMode: Filesystem
accessModes: [“ReadWriteOnce”,“ReadWriteMany”,“ReadOnlyMany”]
persistentVolumeReclaimPolicy: Retain
mountOptions:

• hard
• nfsvers=4.1
  nfs:
  path: /data/v4
  server: 192.168.0.99

---

apiVersion: v1
kind: PersistentVolume
metadata:
name: nfs-pv-v5
labels:
storsystem: nfs
rel: stable
spec:
capacity:
storage: 1Gi
volumeMode: Filesystem
accessModes: [“ReadWriteOnce”,“ReadWriteMany”,“ReadOnlyMany”]
persistentVolumeReclaimPolicy: Retain
mountOptions:

• hard
• nfsvers=4.1
  nfs:
  path: /data/v5
  server: 192.168.0.99

---

apiVersion: v1
kind: PersistentVolume
metadata:
name: nfs-pv-v6
labels:
storsystem: nfs
rel: stable
spec:
capacity:
storage: 1Gi
volumeMode: Filesystem
accessModes: [“ReadWriteOnce”,“ReadWriteMany”,“ReadOnlyMany”]
persistentVolumeReclaimPolicy: Retain
mountOptions:

• hard
• nfsvers=4.1
  nfs:
  path: /data/v6
  server: 192.168.0.99
  [root@master01 ~]#
  　　应用配置清单，创建pv

[root@master01 ~]# kubectl apply -f pv-demo2.yaml
persistentvolume/nfs-pv-v4 created
persistentvolume/nfs-pv-v5 created
persistentvolume/nfs-pv-v6 created
[root@master01 ~]# kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
nfs-pv-v1 1Gi RWO,ROX,RWX Retain Bound default/www-web-0 55m
nfs-pv-v2 1Gi RWO,ROX,RWX Retain Bound default/www-web-1 55m
nfs-pv-v3 1Gi RWO,ROX,RWX Retain Bound default/www-web-2 55m
nfs-pv-v4 1Gi RWO,ROX,RWX Retain Available 4s
nfs-pv-v5 1Gi RWO,ROX,RWX Retain Available 4s
nfs-pv-v6 1Gi RWO,ROX,RWX Retain Available 4s
[root@master01 ~]#
　　扩展sts副本数为6个

[root@master01 ~]# kubectl get sts
NAME READY AGE
web 3/3 53m
[root@master01 ~]# kubectl scale sts web --replicas=6
statefulset.apps/web scaled
[root@master01 ~]#
　　查看对应的pod扩展过程

[root@master01 ~]# kubectl get pod -w
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 19m
web-1 1/1 Running 0 53m
web-2 1/1 Running 0 53m
web-3 0/1 Pending 0 0s
web-3 0/1 Pending 0 0s
web-3 0/1 Pending 0 0s
web-3 0/1 ContainerCreating 0 0s
web-3 1/1 Running 0 2s
web-4 0/1 Pending 0 0s
web-4 0/1 Pending 0 0s
web-4 0/1 Pending 0 2s
web-4 0/1 ContainerCreating 0 2s
web-4 1/1 Running 0 4s
web-5 0/1 Pending 0 0s
web-5 0/1 Pending 0 0s
web-5 0/1 Pending 0 2s
web-5 0/1 ContainerCreating 0 2s
web-5 1/1 Running 0 4s
　　提示：从上面的扩展过程可以看到对应pod是串行扩展，当web-3就绪running以后，才会进行web-4，依次类推；

查看pv和pvc

[root@master01 ~]# kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
nfs-pv-v1 1Gi RWO,ROX,RWX Retain Bound default/www-web-0 60m
nfs-pv-v2 1Gi RWO,ROX,RWX Retain Bound default/www-web-1 60m
nfs-pv-v3 1Gi RWO,ROX,RWX Retain Bound default/www-web-2 60m
nfs-pv-v4 1Gi RWO,ROX,RWX Retain Bound default/www-web-4 5m6s
nfs-pv-v5 1Gi RWO,ROX,RWX Retain Bound default/www-web-3 5m6s
nfs-pv-v6 1Gi RWO,ROX,RWX Retain Bound default/www-web-5 5m6s
[root@master01 ~]# kubectl get pvc
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
www-web-0 Bound nfs-pv-v1 1Gi RWO,ROX,RWX 57m
www-web-1 Bound nfs-pv-v2 1Gi RWO,ROX,RWX 57m
www-web-2 Bound nfs-pv-v3 1Gi RWO,ROX,RWX 57m
www-web-3 Bound nfs-pv-v5 1Gi RWO,ROX,RWX 3m31s
www-web-4 Bound nfs-pv-v4 1Gi RWO,ROX,RWX 3m29s
www-web-5 Bound nfs-pv-v6 1Gi RWO,ROX,RWX 3m25s
[root@master01 ~]#
　　提示：从上面的结果可以看到pv和pvc都处于bound状态；

缩减pod数量为4个

[root@master01 ~]# kubectl scale sts web --replicas=4
statefulset.apps/web scaled
[root@master01 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 28m
web-1 1/1 Running 0 61m
web-2 1/1 Running 0 61m
web-3 1/1 Running 0 7m46s
web-4 1/1 Running 0 7m44s
web-5 0/1 Terminating 0 7m40s
[root@master01 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 28m
web-1 1/1 Running 0 62m
web-2 1/1 Running 0 62m
web-3 1/1 Running 0 8m4s
[root@master01 ~]# kubectl get pv
NAME CAPACITY ACCESS MODES RECLAIM POLICY STATUS CLAIM STORAGECLASS REASON AGE
nfs-pv-v1 1Gi RWO,ROX,RWX Retain Bound default/www-web-0 66m
nfs-pv-v2 1Gi RWO,ROX,RWX Retain Bound default/www-web-1 66m
nfs-pv-v3 1Gi RWO,ROX,RWX Retain Bound default/www-web-2 66m
nfs-pv-v4 1Gi RWO,ROX,RWX Retain Bound default/www-web-4 10m
nfs-pv-v5 1Gi RWO,ROX,RWX Retain Bound default/www-web-3 10m
nfs-pv-v6 1Gi RWO,ROX,RWX Retain Bound default/www-web-5 10m
[root@master01 ~]# kubectl get pvc
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
www-web-0 Bound nfs-pv-v1 1Gi RWO,ROX,RWX 62m
www-web-1 Bound nfs-pv-v2 1Gi RWO,ROX,RWX 62m
www-web-2 Bound nfs-pv-v3 1Gi RWO,ROX,RWX 62m
www-web-3 Bound nfs-pv-v5 1Gi RWO,ROX,RWX 8m13s
www-web-4 Bound nfs-pv-v4 1Gi RWO,ROX,RWX 8m11s
www-web-5 Bound nfs-pv-v6 1Gi RWO,ROX,RWX 8m7s
[root@master01 ~]#
　　提示：可以看到缩减pod它会从索引号最大的pod逆序缩减，缩减以后对应pv和pvc的状态依旧是bound状态；扩缩减pod副本过程如下图所示

提示：上图主要描述了sts控制器上的pod缩减–>扩展pod副本的过程；缩减pod副本数量，对应后端的pvc和pv的状态都是不变的，后续再增加pod副本数量，对应pvc能够根据pod名称自动的关联到对应的pod上，使得扩展后对应名称的pod和之前缩减pod的数据保存一致；

滚动更新pod版本

1
2
3
[root@master01 ~]# kubectl set image sts web nginx=nginx:1.16-alpine
statefulset.apps/web image updated
[root@master01 ~]#
　　查看更新过程

[root@master01 ~]# kubectl get pod -w
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 38m
web-1 1/1 Running 0 71m
web-2 1/1 Running 0 71m
web-3 1/1 Running 0 17m
web-3 1/1 Terminating 0 20m
web-3 0/1 Terminating 0 20m
web-3 0/1 Terminating 0 20m
web-3 0/1 Terminating 0 20m
web-3 0/1 Pending 0 0s
web-3 0/1 Pending 0 0s
web-3 0/1 ContainerCreating 0 0s
web-3 1/1 Running 0 1s
web-2 1/1 Terminating 0 74m
web-2 0/1 Terminating 0 74m
web-2 0/1 Terminating 0 74m
web-2 0/1 Terminating 0 74m
web-2 0/1 Pending 0 0s
web-2 0/1 Pending 0 0s
web-2 0/1 ContainerCreating 0 0s
web-2 1/1 Running 0 2s
web-1 1/1 Terminating 0 74m
web-1 0/1 Terminating 0 74m
web-1 0/1 Terminating 0 75m
web-1 0/1 Terminating 0 75m
web-1 0/1 Pending 0 0s
web-1 0/1 Pending 0 0s
web-1 0/1 ContainerCreating 0 0s
web-1 1/1 Running 0 2s
web-0 1/1 Terminating 0 41m
web-0 0/1 Terminating 0 41m
web-0 0/1 Terminating 0 41m
web-0 0/1 Terminating 0 41m
web-0 0/1 Pending 0 0s
web-0 0/1 Pending 0 0s
web-0 0/1 ContainerCreating 0 0s
web-0 1/1 Running 0 1s
　　提示：从上面更新过程来看，statefulset控制器滚动更新是从索引号最大的pod开始更新，并且它一次更新一个pod，只有等到上一个pod更新完毕，并且其状态为running以后，才开始更新第二个，依次类推；

验证：查看对应sts信息，看看对应版本是否更新为我们指定的镜像版本？

[root@master01 ~]# kubectl get sts -o wide
NAME READY AGE CONTAINERS IMAGES
web 4/4 79m nginx nginx:1.16-alpine
[root@master01 ~]#
　　回滚pod版本为上一个版本

[root@master01 ~]# kubectl get sts -o wide
NAME READY AGE CONTAINERS IMAGES
web 4/4 80m nginx nginx:1.16-alpine
[root@master01 ~]# kubectl rollout undo sts/web
statefulset.apps/web rolled back
[root@master01 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 6m6s
web-1 1/1 Running 0 6m13s
web-2 0/1 ContainerCreating 0 1s
web-3 1/1 Running 0 12s
[root@master01 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 6m14s
web-1 0/1 ContainerCreating 0 1s
web-2 1/1 Running 0 9s
web-3 1/1 Running 0 20s
[root@master01 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 1s
web-1 1/1 Running 0 8s
web-2 1/1 Running 0 16s
web-3 1/1 Running 0 27s
[root@master01 ~]# kubectl get sts -o wide
NAME READY AGE CONTAINERS IMAGES
web 4/4 81m nginx nginx:1.14-alpine
[root@master01 ~]#
　　使用partition字段控制更新pod数量，实现金丝雀效果更新pod版本

sts控制器默认的更新策略是依次从索引最大的pod开始逆序更新，先删除一个pod等待对应pod更新完毕以后，状态处于running以后，接着更新第二个依次更新完所有的pod，要想实现金丝雀更新pod版本的效果，我们需要告诉sts更新在那个位置；在deploy控制器中我们使用的是kubectl rollout pause命令来暂停更新，从而实现金丝雀更新pod版本的效果，当然在sts中也可以；除此之外，sts还支持通过sts.spec.updateStrategy.rollingUpdate.partition字段的值来控制器更新数量；默认partition的值为0，表示更新到索引大于0的pod位置，即全部更新；如下图所示

提示：在sts控制器中更新pod模板的镜像版本，可以使用partition这个字段来控制更新到那个位置，partition=3表示更新索引大于等于3的pod，小于3的pod就不更新；partition=0表示全部更新；

示例：在线更改sts控制器的partition字段的值为3

提示：在线修改sts的配置，使用kubectl edit命令指定类型和对应控制器实例，就可以进入编辑对应配置文件的界面，找到updateStrategy字段下的rollingUpdate字段下的partition字段，把原有的0更改为3，保存退出即可生效；当然我们也可以直接更改配置清单，然后再重新应用一下也行；如果配置清单中没有定义，可以加上对应的字段即可；

再次更新pod版本，看看它是否只更新索引于等3的pod呢？

9
[root@master01 ~]# kubectl get sts -o wide
NAME READY AGE CONTAINERS IMAGES
web 4/4 146m nginx nginx:1.14-alpine
[root@master01 ~]# kubectl set image sts web nginx=nginx:1.16-alpine
statefulset.apps/web image updated
[root@master01 ~]# kubectl get sts -o wide
NAME READY AGE CONTAINERS IMAGES
web 3/4 146m nginx nginx:1.16-alpine
[root@master01 ~]#
　　查看更新过程

[root@master01 ~]# kubectl get pods -w
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 64m
web-1 1/1 Running 0 64m
web-2 1/1 Running 0 64m
web-3 1/1 Running 0 64m
web-3 1/1 Terminating 0 51s
web-3 0/1 Terminating 0 51s
web-3 0/1 Terminating 0 60s
web-3 0/1 Terminating 0 60s
web-3 0/1 Pending 0 0s
web-3 0/1 Pending 0 0s
web-3 0/1 ContainerCreating 0 0s
web-3 1/1 Running 0 1s
^C[root@master01 ~]# kubectl get pods
NAME READY STATUS RESTARTS AGE
web-0 1/1 Running 0 65m
web-1 1/1 Running 0 65m
web-2 1/1 Running 0 65m
web-3 1/1 Running 0 50s
[root@master01 ~]#
　　提示：从上面的更新过程可以看到，对应sts控制器此时更新只是更新了web-3，其余索引小于3的pod并没有发生更新操作；

亚马逊测评 www.yisuping.com