mac 上学习k8s系列（26）ReplicationController、ReplicaSet、StatefulSet

golang算法架构leetcode技术php 2021-12-05

171

如果说svc、pod、pv等资源是k8s的数据流，那么ReplicationController、ReplicaSet、StatefulSet便是k8s的控制流，可以控制k8s 中pod的副本数和动态伸缩。我们基于以前制作的apple:5678镜像来进行学习。

1，ReplicationController

apiVersion: v1
kind: ReplicationController
metadata:
  name: apple-rctl
  labels:
    name: apple-rctl
spec:
  replicas: 2
  selector:
    app: apple
  template:
    metadata:
     labels:
       app: apple
    spec:
     containers:
     - name: apple
       image: apple:5678
       env :
       - name : GET_HOSTS_FROM
         value : env
       ports:
       - containerPort: 5678

RC只支持基于等式的selector（app!= apple或app!=apple）

 % kubectl apply -f deployment/ReplicationController.yaml
replicationcontroller/apple-rc created

% kubectl get rc  
NAME       DESIRED   CURRENT   READY   AGE
apple-rc   3         3         3       32s

查看下pod，发现有两个pod已经创建成功了

% kubectl get pod |grep apple-rctl
apple-rctl-2hdpt                            1/1     Running     0          29s
apple-rctl-wrmwp                            1/1     Running     0          15s

2，ReplicaSet

Replica Set还支持新的、基于集合的selector（version in (v1.0, v2.0)或env notin (dev, qa)）

apiVersion: apps/v1
kind: ReplicaSet 
metadata:
  name: apple-rs 
spec:
  replicas: 3 #有3个副本
  selector: #标签选择器
    matchLabels:
      app: apple 
  template: #模板
    metadata:
      labels:
        app: apple 
    spec:
      containers:
      - name: apple-app 
        image: apple:5678
        env:
        - name: GET_HOSTS_FROM 
          value: dns 
        ports:
        - containerPort: 5678

中间遇到的问题

% kubectl apply -f deployment/ReplicaSet.yaml
error: unable to recognize "deployment/ReplicaSet.yaml": no matches for kind "ReplicaSet" in version "extensions/v1beta1"

需要查看下当前k8s支持的版本

%  kubectl api-versions
admissionregistration.k8s.io/v1
admissionregistration.k8s.io/v1beta1
apiextensions.k8s.io/v1
apiextensions.k8s.io/v1beta1
apiregistration.k8s.io/v1
apiregistration.k8s.io/v1beta1
apps/v1
authentication.k8s.io/v1
authentication.k8s.io/v1beta1
authorization.k8s.io/v1
authorization.k8s.io/v1beta1
autoscaling/v1
autoscaling/v2beta1
autoscaling/v2beta2
batch/v1
batch/v1beta1
catalog.cattle.io/v1
certificates.k8s.io/v1
certificates.k8s.io/v1beta1
coordination.k8s.io/v1
coordination.k8s.io/v1beta1
discovery.k8s.io/v1
discovery.k8s.io/v1beta1
events.k8s.io/v1
events.k8s.io/v1beta1
extensions/v1beta1
flowcontrol.apiserver.k8s.io/v1beta1
management.cattle.io/v3
networking.k8s.io/v1
networking.k8s.io/v1beta1
node.k8s.io/v1
node.k8s.io/v1beta1
policy/v1
policy/v1beta1
rbac.authorization.k8s.io/v1
rbac.authorization.k8s.io/v1beta1
samplecontroller.k8s.io/v1alpha1
scheduling.k8s.io/v1
scheduling.k8s.io/v1beta1
storage.k8s.io/v1
storage.k8s.io/v1beta1
ui.cattle.io/v1
v1

然后修改成对应的版本即可。

 % kubectl apply -f deployment/ReplicaSet.yaml
replicaset.apps/apple-rc create

3，StatefulSet

使用statefulset需要使用对应的StorageClass，先声明一个

apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: apple-nfs-storage
provisioner: docker.io/hostpath # or choose another name, must match deployment's env PROVISIONER_NAME' fuseim.pri/ifs  改为docker.io/hostpath 
parameters:
  archiveOnDelete: "false" # # When set to "false" your PVs will not be archived
                           # by the provisioner upon deletion of the PVC.

statefulset是有状态的，有状态的副本集都会用到持久存储，所以各个节点不能使用同一存储卷，每个节点有自已的专用存储，因此需要一个pv的模板StorageClass。

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: apple-ss
spec:
  selector:
    matchLabels:
      app: apple #必须匹配 .spec.template.metadata.labels
  serviceName: "apple-headless"  #声明它属于哪个Headless Service.
  replicas: 3 #副本数
  template:
    metadata:
      labels:
        app: apple # 必须配置 .spec.selector.matchLabels
    spec:
      terminationGracePeriodSeconds: 1
      containers:
      - name: apple
        image: apple:5678
        ports:
        - containerPort: 5678
          name: apple
        volumeMounts:
        - name: apple-pvc
          mountPath: /usr/share/apple/html


  volumeClaimTemplates:   #可看作pvc的模板
  - metadata:
      name: apple-pvc
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "apple-nfs-storage"  #存储类名，改为集群中已存在的 apple-nfs-storage 改为docker.io/hostpath  改为 hostpath
      resources:
        requests:
          storage: 1Ki

在StatefulSet中与之对应的headless service，headless service，即无头服务，因为是有状态的服务，希望我们每一次连接到同一个pod，关联到同一份pv存储，但是每一次重启后后pod的ip都会被重新分配，所以需要headless service，它和普通服务的区别是创建服务的时候制定 clusterIp：none，这样kubeproxy就不会做负载均衡，方便应用内部自己控制，同时给每一个pod都提供了域名，我们可以保证pod的域名重启后不变，保证了每次连接到同一个pod。

% kubectl apply -f deployment/StorageClass.yaml
storageclass.storage.k8s.io/apple-nfs-storage created

% kubectl apply -f deployment/statefulset.yaml
statefulset.apps/apple-ss created

启动后我们发现状态一直没有ready

% kubectl get statefulset
NAME       READY   AGE
apple-ss   0/3     22s

看下pod的描述信息

% kubectl describe pod apple-ss-0        
Name:           apple-ss-0
Namespace:      default
Priority:       0
Node:           <none>
Labels:         app=apple
                controller-revision-hash=apple-ss-6bb7949cbd
                statefulset.kubernetes.io/pod-name=apple-ss-0
Annotations:    <none>
Status:         Pending
IP:             
IPs:            <none>
Controlled By:  StatefulSet/apple-ss
Containers:
  apple:
    Image:        apple:5678
    Port:         5678/TCP
    Host Port:    0/TCP
    Environment:  <none>
    Mounts:
      usr/share/apple/html from apple-pvc (rw)
      var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-qzvxw (ro)
Conditions:
  Type           Status
  PodScheduled   False 
Volumes:
  apple-pvc:
    Type:       PersistentVolumeClaim (a reference to a PersistentVolumeClaim in the same namespace)
    ClaimName:  apple-pvc-apple-ss-0
    ReadOnly:   false
  kube-api-access-qzvxw:
    Type:                    Projected (a volume that contains injected data from multiple sources)
    TokenExpirationSeconds:  3607
    ConfigMapName:           kube-root-ca.crt
    ConfigMapOptional:       <nil>
    DownwardAPI:             true
QoS Class:                   BestEffort
Node-Selectors:              <none>
Tolerations:                 node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
                             node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:
  Type     Reason            Age   From               Message
  ----     ------            ----  ----               -------
  Warning  FailedScheduling  39s   default-scheduler  0/1 nodes are available: 1 pod has unbound immediate PersistentVolumeClaims.
  Warning  FailedScheduling  38s   default-scheduler  0/1 nodes are available: 1 pod has unbound immediate PersistentVolumeClaims.

发现pod绑定pv出现了问题

1 pod has unbound immediate PersistentVolumeClaims.

仔细研究发现storageClass里声明的parameters：fuseim.pri/ifs 在docker for mac 上不支持，改为docker.io/hostpath就好了

修改后我们看下StorageClass资源

% kubectl get storageclass                     
NAME                 PROVISIONER          RECLAIMPOLICY   VOLUMEBINDINGMODE   ALLOWVOLUMEEXPANSION   AGE
apple-nfs-storage    docker.io/hostpath   Delete          Immediate           false                  4s
hostpath (default)   docker.io/hostpath   Delete          Immediate           false                  104d

查看我们的pod和statefullset，发现已经起来了

% kubectl get pod  |grep apple-ss
apple-ss-0                                  1/1     Running     0          64s
apple-ss-1                                  1/1     Running     0          64s
apple-ss-2                                  1/1     Running     0          62s
 % kubectl get statefulset
NAME       READY   AGE
apple-ss   3/3     83s

文章转载自golang算法架构leetcode技术php，如果涉嫌侵权，请发送邮件至：contact@modb.pro进行举报，并提供相关证据，一经查实，墨天轮将立刻删除相关内容。

mac 上学习k8s系列（26）ReplicationController、ReplicaSet、StatefulSet

评论