K8S调度亲和性和反亲和性

K8s官网介绍:

​​https://kubernetes.io/zh/docs/tasks/configure-pod-container/assign-pods-nodes-using-node-affinity/​​

节点亲和性:

pod.spec.nodeAffinity

requiredDuringSchedulingIgnoredDuringExecution(硬策略)表示pod必须部署到满足条件的节点上，如果没有满足条件的节点，就不停重试。其中IgnoreDuringExecution表示pod部署成功之后运行的时候，如果节点标签发生了变化，不再满足pod指定的条件，pod也会继续运行。

 

preferredDuringSchedulingIgnoredDuringExecution(软策略)表示优先部署到满足条件的节点上，如果没有满足条件的节点，就忽略这些条件，按照正常逻辑部署。

键值运算关系:

In：label的值在某个列表中

NotIn:label的值不在某个列表中

Gt:label的值大于某个值

Lt:label的值小于某个值

Exists：某个label存在

DoesNotExist:某个label不存在

 

查看node标签属性:

[root@master ~]# kubectl get nodes --show-labels
NAME     STATUS   ROLES    AGE   VERSION   LABELS
master   Ready    master   40d   v1.17.5   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=master,kubernetes.io/os=linux,node-role.kubernetes.io/master=
node1    Ready    <none>   40d   v1.17.5   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=node1,kubernetes.io/os=linux
node2    Ready    <none>   40d   v1.17.5   beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=node2,kubernetes.io/os=linux

硬策略测试:

硬策略是强制性的要求，必须满足，不满足pod只能处于pending状态。

[root@master afff]# cat nodeaf.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: aff
  labels:
    app: test-node
spec:
  containers:
  - name: test-pod
    image: nginx
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: kubernetes.io/hostname
            operator: NotIn
            values:
            - node2
[root@master afff]# kubectl apply -f nodeaf.yaml 
pod/aff created
[root@master afff]# kubectl get pod -o wide

软策略测试:

软策略定义是尽量满足，如果满足不了也可以选择其他node运行。

[root@master afff]# cat nodeaf.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: aff
  labels:
    app: test-node
spec:
  containers:
  - name: test-pod
    image: nginx
  affinity:
    nodeAffinity:
      preferredDuringSchedulingIgnoredDuringExecution:
      - weight: 1
        preference:
          matchExpressions:
          - key: kubernetes.io/hostname
            operator: In
            values:
            - node3
[root@master afff]# kubectl apply -f nodeaf.yaml 
pod/aff created
[root@master afff]# kubectl get pod -o wide

Pod亲和性:

pod亲和性(podAffinity)：和某些pod调度到同一节点

pod反亲和性(podAntiAffinity)：和某些pod调度在不同节点

亲和性调度可以分成两种策略:

硬策略(requiredDuringSchedulingIgnoredDuringExecution):

指定了将 pod 调度到一个节点上必须满足的规则,不满足则会处于pending状态，一直进行重试，直到满足为止。

软策略(preferredDuringSchedulingIgnoredDuringExecution):

优先满足指定的规则，如果不满足则会调度到其他节点上。如果有多个软策略还可以设置优先级(weight)

首先创建一个pod，打上自定标签:

创建一个nginx pod:

[root@master afff]# cat pod.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: nginx-example
  labels:
    app: nginx
spec:
  containers:
  - name: test-nginx
    image: nginx

查看pod标签:

[root@master afff]# kubectl get pod --show-labels
NAME            READY   STATUS    RESTARTS   AGE     LABELS
nginx-example   1/1     Running   0          4m40s   app=nginx

Pod亲和性:指定pod亲和性标签是nginx:

[root@master afff]# cat pod2.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: pod2
  labels:
    app: test-node
spec:
  containers:
  - name: test-pod
    image: nginx
  affinity:
    podAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
      - topologyKey: kubernetes.io/hostname
        labelSelector:
          matchExpressions:
          - key: app
            operator: In
            values:
            - nginx
[root@master afff]# kubectl get pod -o wide
NAME            READY   STATUS              RESTARTS   AGE   IP            NODE    NOMINATED NODE   READINESS GATES
nginx-example   1/1     Running             0          21m   10.244.2.55   node2   <none>           <none>
pod2            0/1     ContainerCreating   0          23s   <none>        node2   <none>           <none>

Pod反亲和性:

[root@master afff]# cat pod2.yaml 
apiVersion: v1
kind: Pod
metadata:
  name: pod2
  labels:
    app: test-node
spec:
  containers:
  - name: test-pod
    image: nginx
  affinity:
    podAntiAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
      - topologyKey: kubernetes.io/hostname
        labelSelector:
          matchExpressions:
          - key: app
            operator: In
            values:
            - nginx
[root@master afff]# kubectl apply -f pod2.yaml 
pod/pod2 created
[root@master afff]# kubectl get pod -o wide
NAME            READY   STATUS              RESTARTS   AGE   IP            NODE    NOMINATED NODE   READINESS GATES
nginx-example   1/1     Running             0          26m   10.244.2.55   node2   <none>           <none>
pod2            0/1     ContainerCreating   0          7s    <none>        node1   <none>           <none>