一、同一个应用部署至不同宿主机¶

在使用Kubernetes时，一般都会有很多节点运行容器，此时可以使用Pod反亲和力将同一个应用部署到不同的节点上，达到更高的可用率，以免同一个应用部署到相同的宿主机带来的风险。

这里又分为两种情况：

1）同一个应用必须部署到不同宿主机
2）同一个应用尽量部署到不同宿主机

1.1 同一个应用必须部署到不同宿主机¶

1.查看节点污点情况，保证无污点

[root@k8s-master01 ~]# kubectl describe node | grep Taint            
Taints:             <none>
Taints:             <none>
Taints:             <none>
Taints:             <none>

如果存在污点，则参考下面方式删除污点

[root@k8s-master01 ~]# k taint node k8s-master01 node-role.kubernetes.io/control-plane-

2.定义一个名为podAntiAffinity01的yaml文件

[root@k8s-master01 Affinity]# vim podAntiAffinity01.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: must-be-diff-nodes
  name: must-be-diff-nodes
  namespace: kube-public
spec:
  replicas: 3
  selector:
    matchLabels:
      app: must-be-diff-nodes
  template:
    metadata:
      labels:
        app: must-be-diff-nodes
    spec:
      affinity:
        podAntiAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
          - labelSelector:
              matchExpressions:
              - key: app
                operator: In
                values:
                - must-be-diff-nodes
            topologyKey: kubernetes.io/hostname
      containers:
      - image: registry.cn-hangzhou.aliyuncs.com/zq-demo/nginx:1.14.2
        imagePullPolicy: IfNotPresent
        name: must-be-diff-nodes

3.开始部署

[root@k8s-master01 Affinity]# kubectl create -f podAntiAffinity01.yaml

4.查看pod状态，观察到pod被部署到不同节点

[root@k8s-master01 ~]# kubectl get po -n kube-public -owide
NAME                                  READY   STATUS    RESTARTS   AGE     IP               NODE           NOMINATED NODE   READINESS GATES
must-be-diff-nodes-7cb8b45d66-78n96   1/1     Running   0          2m55s   192.168.58.246   k8s-node02     <none>           <none>
must-be-diff-nodes-7cb8b45d66-ltwps   1/1     Running   0          2m55s   192.168.32.129   k8s-master01   <none>           <none>
must-be-diff-nodes-7cb8b45d66-q5mml   1/1     Running   0          2m55s   192.168.85.241   k8s-node01     <none>           <none>

5.扩容副本数为4个

[root@k8s-master01 Affinity]# kubectl scale deployment must-be-diff-nodes --replicas=4 -n kube-public

除了上面扩容副本测试外，也可以尝试重启deployment

[root@k8s-master01 ~]# k rollout restart deployment must-be-diff-nodes  -n kube-public

6.再次查看Pod状态，发现有一个节点处于Pending

[root@k8s-master01 ~]# kubectl get po -n kube-public -owide
NAME                                  READY   STATUS    RESTARTS   AGE     IP               NODE           NOMINATED NODE   READINESS GATES
must-be-diff-nodes-7cb8b45d66-78n96   1/1     Running   0          3m25s   192.168.58.246   k8s-node02     <none>           <none>
must-be-diff-nodes-7cb8b45d66-ltwps   1/1     Running   0          3m25s   192.168.32.129   k8s-master01   <none>           <none>
must-be-diff-nodes-7cb8b45d66-q5mml   1/1     Running   0          3m25s   192.168.85.241   k8s-node01     <none>           <none>
must-be-diff-nodes-7cb8b45d66-xhzp6   0/1     Pending   0          4s      <none>           <none>         <none>           <none>

7.详细查看Pod处于Pending原因，这是因为我们设置的是硬亲和力，必须部署在不同宿主机上，节点只有3个，而副本数设置为4个，所以导致一直处于Pending

[root@k8s-master01 Affinity]# kubectl describe po must-be-diff-nodes-bdbb64998-58ngr -n kube-public
...
...
...
Events:
  Type     Reason            Age   From               Message
  ----     ------            ----  ----               -------
  Warning  FailedScheduling  57s   default-scheduler  0/3 nodes are available: 3 node(s) didn't match pod anti-affinity rules. preemption: 0/3 nodes are available: 3 No preemption victims found for incoming pod.

8.环境复原

[root@k8s-master01 ~]# kubectl delete  -f podAntiAffinity01.yaml

1.2 同一个应用尽量部署到不同宿主机¶

1.查看节点污点情况，保证无污点

[root@k8s-master01 ~]# kubectl describe node | grep Taint            
Taints:             <none>
Taints:             <none>
Taints:             <none>
Taints:             <none>

如果存在污点，则参考下面方式删除污点

[root@k8s-master01 ~]# k taint node k8s-master01 node-role.kubernetes.io/control-plane-

2.定义一个名为podAntiAffinity02的yaml文件

[root@k8s-master01 Affinity]# vim podAntiAffinity02.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: must-be-diff-nodes
  name: must-be-diff-nodes
  namespace: kube-public
spec:
  replicas: 3
  selector:
    matchLabels:
      app: must-be-diff-nodes
  template:
    metadata:
      labels:
        app: must-be-diff-nodes
    spec:
      affinity:
        podAntiAffinity:
          preferredDuringSchedulingIgnoredDuringExecution:
          - weight: 100
            podAffinityTerm:
              labelSelector:
                matchExpressions:
                  - key: app
                    operator: In
                    values:
                      - must-be-diff-nodes
              topologyKey: kubernetes.io/hostname
      containers:
      - image: registry.cn-hangzhou.aliyuncs.com/zq-demo/nginx:1.14.2
        imagePullPolicy: IfNotPresent
        name: must-be-diff-nodes