pod 反亲和测试验证

场景需求

node 节点上不允许有两个一样的pod,即每个节点上每个应用仅仅允许一个pod 运行
  • 测试yaml
[root@k8sNode0 podqinhexing]# cat test.yaml 
apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx
spec:
  selector:
    matchLabels:
      app: nginx
  replicas: 4
  template:
    metadata:
      labels:
        app: nginx
    spec:
      affinity:
        podAntiAffinity:
          requiredDuringSchedulingIgnoredDuringExecution:
          - labelSelector:
              matchExpressions:
              - key: app
                operator: In
                values:
                - nginx
            topologyKey: "kubernetes.io/hostname"
      containers:
      - name: nginx-server
        image: nginx:latest

  • 测试结果(第四个pod 显示pending 状态,即每个节点上落了一个pod)
[root@k8sNode0 podqinhexing]# kubectl  get pod   -o wide |grep nginx
nginx-78dfb97767-747w5                         1/1     Running   0          8s     172.16.4.183   cn-zhangjiakou.172.16.4.39   <none>           <none>
nginx-78dfb97767-bgl2n                         1/1     Running   0          8s     172.16.4.243   cn-zhangjiakou.172.16.4.40   <none>           <none>
nginx-78dfb97767-hrbrn                         1/1     Running   0          8s     172.16.4.16    cn-zhangjiakou.172.16.4.38   <none>           <none>
nginx-78dfb97767-qmjc9                         0/1     Pending   0          8s     <none>         <none>                       <none>           <none>

  • 第四个pod 的报错日志
  Warning  FailedScheduling  15s   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.,


posted @ 2023-08-21 20:44  lixinliang  阅读(88)  评论(0编辑  收藏  举报