官网文档介绍
通过下图简要介绍

在node节点135上配置的label：key-affinity-label=value-affinity-label

root@server01:~# kubectl label node 192.168.174.135 key-affinity-label=value-affinity-label
node "192.168.174.135" labeled

136服务器k8s master
/root/demo/node-affinity.yaml
在yaml文件中配置label的key；key: key-affinity-label value-affinity-label

apiVersion: v1
kind: Pod
metadata:
  labels:
    run: node-affinity
  name: node-affinity
spec:
  containers:
  - image: nginx
    imagePullPolicy: IfNotPresent
    name: node-affinity
  affinity:
    nodeAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
        nodeSelectorTerms:
        - matchExpressions:
          - key: key-affinity-label
            operator: In
            values:
            - "value-affinity-label"

创建

kubectl create -f node-affinity.yaml
查看node-affinity就创建在了135节点

root@server02:~/demo# kubectl get pod -o wide
NAME                  READY     STATUS    RESTARTS   AGE       IP              NODE
node-affinity         1/1       Running   0          28m       172.20.188.20   192.168.174.135
pod-manual-schedule   1/1       Running   0          3h        172.20.188.19   192.168.174.135
pod-schedule          1/1       Running   0          3h        172.20.40.205   192.168.174.137

刚开始配置的时候，因为key和value与所有3台node的label都不匹配，所以执行命令调度创建的时候，查看创建状态一直都是pending

root@server02:~/demo# kubectl get pod
NAME                  READY     STATUS    RESTARTS   AGE
node-affinity         0/1       Pending   0          13m

查看日志

kubectl describe pod node-affinity

提示default-scheduler 0/3 nodes are available: 3 MatchNodeSelector.

Events:
  Type     Reason            Age                From               Message
  ----     ------            ----               ----               -------
  Warning  FailedScheduling  2m (x72 over 22m)  default-scheduler  0/3 nodes are available: 3 MatchNodeSelector.