学习笔记十七：node节点选择器，亲和性-CSDN博客

本文链接：https://blog.csdn.net/weixin_43258559/article/details/132315258

node节点选择器，污点、容忍度、亲和性

node节点选择器
- - nodeName，指定pod节点运行在哪个具体node上
  - nodeSelector：指定pod调度到具有哪些标签的node节点上
亲和性

node节点选择器

我们在创建pod资源的时候，pod会根据schduler进行调度，那么默认会调度到随机的一个工作节点，如果我们想要pod调度到指定节点或者调度到一些具有相同特点的node节点，怎么办呢？
可以使用pod中的nodeName或者nodeSelector字段指定要调度到的node节点

nodeName，指定pod节点运行在哪个具体node上

把tomcat.tar.gz上传到k8snode1和k8snode2，手动解压

ctr -n=k8s.io images import  tomcat.tar.gz

把busybox.tar.gz上传到k8snode1和k8snode2，手动解压

ctr -n=k8s.io images import  busybox.tar.gz

vim pod-node.yaml

apiVersion: v1
kind: Pod
metadata:
  name: demo-pod
  namespace: default
  labels:
    app: myapp
    env: dev
spec:
  nodeName: k8snode1
  containers:
  - name:  tomcat-pod-java
    ports:
    - containerPort: 8080
    image: tomcat:8.5-jre8-alpine
    imagePullPolicy: IfNotPresent
  - name: busybox
    image: busybox:latest
    command:
    - "/bin/sh"
    - "-c"
- "sleep 3600"

kubectl apply -f pod-node.yaml

查看pod调度到哪个节点

kubectl get pods  -o wide

NAME             READY   STATUS    RESTARTS             
demo-pod        1/1     Running   0            k8snode1

nodeSelector：指定pod调度到具有哪些标签的node节点上

同一个yaml文件里定义pod资源，如果同时定义了nodeName和NodeSelector，那么条件必须都满足才可以，有一个不满足都会调度失败

给node节点打标签，打个具有disk=ceph的标签

kubectl label nodes k8snode2 disk=ceph

定义pod的时候指定要调度到具有disk=ceph标签的node上

vim pod-1.yaml

apiVersion: v1
kind: Pod
metadata:
  name: demo-pod-1
  namespace: default
  labels:
    app: myapp
    env: dev
spec:
  nodeSelector:
    disk: ceph
  containers:
  - name:  tomcat-pod-java
    ports:
    - containerPort: 8080
    image: tomcat:8.5-jre8-alpine
    imagePullPolicy: IfNotPresent

kubectl apply -f pod-1.yaml

查看pod调度到哪个节点

kubectl get pods  -o wide

NAME             READY   STATUS    RESTARTS             
demo-pod-1        1/1     Running   0            k8snode2

做完上面实验，需要把default名称空间下的pod全都删除，kubectl delete pods pod名字

kubectl delete pods demo-pod-1

删除node节点打的标签

kubectl label nodes k8snode2 disk-

亲和性

node节点亲和性

node节点亲和性调度：nodeAffinity，查看官方介绍

kubectl explain pods.spec.affinity

KIND:     Pod
VERSION:  v1
RESOURCE: affinity <Object>
DESCRIPTION:
     If specified, the pod's scheduling constraints
    Affinity is a group of affinity scheduling rules.
FIELDS:
   nodeAffinity	<Object> #节点亲和性
   podAffinity	<Object> #pod亲和性
   podAntiAffinity	<Object> #pod反亲和性

kubectl explain  pods.spec.affinity.nodeAffinity

KIND:     Pod
VERSION:  v1
RESOURCE: nodeAffinity <Object>
DESCRIPTION:
     Describes node affinity scheduling rules for the pod.
     Node affinity is a group of node affinity scheduling rules.
FIELDS:
   preferredDuringSchedulingIgnoredDuringExecution	<[]Object>
   requiredDuringSchedulingIgnoredDuringExecution	<Object>
    
prefered表示有节点尽量满足这个位置定义的亲和性，这不是一个必须的条件，软亲和性
require表示必须有节点满足这个位置定义的亲和性，这是个硬性条件，硬亲和性

使用requiredDuringSchedulingIgnoredDuringExecution硬亲和性

把myapp-v1.tar.gz上传到k8snode2和k8snode1上，手动解压：

ctr -n=k8s.io images import  myapp-v1.tar.gz

vim pod-nodeaffinity-demo.yaml

apiVersion: v1
kind: Pod
metadata:
  name:  pod-node-affinity-demo
  namespace: default
  labels:
    app: myapp
    tier: frontend
spec:
  affinity:
    nodeAffinity:
     requiredDuringSchedulingIgnoredDuringExecution:
       nodeSelectorTerms:
       - matchExpressions:
         - key: zone
           operator: In
           values: 
           - foo
           - bar
  containers:
  - name: myapp
    image: docker.io/ikubernetes/myapp:v1
imagePullPolicy: IfNotPresent

我们检查当前节点中有任意一个节点拥有zone标签的值是foo或者bar，就可以把pod调度到这个node节点的foo或者bar标签上的节点上

kubectl apply -f pod-nodeaffinity-demo.yaml

kubectl get pods -o wide | grep pod-node

pod-node-affinity-demo             0/1     Pending     0   k8snode1

status的状态是pending，上面说明没有完成调度，因为没有一个拥有zone的标签的值是foo或者bar，而且使用的是硬亲和性，必须满足条件才能完成调度
给这个k8snode1节点打上标签zone=foo，在查看

kubectl label nodes k8snode1 zone=foo

kubectl get pods -o wide

pod-node-affinity-demo             1/1     Running  0   k8snode1

删除pod-nodeaffinity-demo.yaml

kubectl delete -f pod-nodeaffinity-demo.yaml

使用preferredDuringSchedulingIgnoredDuringExecution软亲和性

vim pod-nodeaffinity-demo-2.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-node-affinity-demo-2
  namespace: default
  labels:
    app: myapp
    tier: frontend
spec:
  containers:
  - name: myapp
    image: docker.io/ikubernetes/myapp:v1
    imagePullPolicy: IfNotPresent
  affinity:
    nodeAffinity:
      preferredDuringSchedulingIgnoredDuringExecution:
      - preference:
          matchExpressions: 
          - key: zone1
            operator: In
            values:
            - foo1
            - bar1
        weight: 10
      - preference:
          matchExpressions:
          - key: zone2
            operator: In
            values:
            - foo2
            - bar2
        weight: 20

kubectl apply -f pod-nodeaffinity-demo-2.yaml

kubectl get pods -o wide |grep demo-2

pod-node-affinity-demo-2           1/1     Running     0       k8snode1

上面说明软亲和性是可以运行这个pod的，尽管没有运行这个pod的节点定义的zone1标签
Node节点亲和性针对的是pod和node的关系，Pod调度到node节点的时候匹配的条件

测试完成，删除 pod-nodeaffinity-demo-2.yaml

kubectl delete -f pod-nodeaffinity-demo-2.yaml

weight权重

weight是相对权重，权重越高，pod调度的几率越大
假如给xianchaonode1和xianchaonode2都打上标签

kubectl label nodes k8snode1 zone1=foo1
kubectl label nodes k8snode2 zone2=foo2

vim pod-nodeaffinity-demo-2.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-node-affinity-demo-2
  namespace: default
  labels:
    app: myapp
    tier: frontend
spec:
  containers:
  - name: myapp
    image: docker.io/ikubernetes/myapp:v1
    imagePullPolicy: IfNotPresent
  affinity:
    nodeAffinity:
      preferredDuringSchedulingIgnoredDuringExecution:
      - preference:
          matchExpressions: 
          - key: zone1
            operator: In
            values:
            - foo1
            - bar1
        weight: 10
      - preference:
          matchExpressions:
          - key: zone2
            operator: In
            values:
            - foo2
            - bar2
        weight: 20

kubectl apply -f pod-nodeaffinity-demo-2.yaml

pod在定义node节点亲和性的时候，k8sode1和k8snode2都满足条件，都可以调度pod，且都具有标签，pod在匹配zone2=foo2的权重高，那么pod就会优先调度到k8snode2上

删除对应标签和pod

kubectl delete -f pod-nodeaffinity-demo.yaml
kubectl delete -f pod-nodeaffinity-demo-2.yaml

查看是否还有对应pod未删除

kubectl get pods -owide

Pod节点亲和性

pod自身的亲和性调度有两种表示形式

podaffinity：pod和pod更倾向腻在一起，把相近的pod结合到相近的位置，如同一区域，同一机架，这样的话pod和pod之间更好通信，比方说有两个机房，这两个机房部署的集群有1000台主机，那么我们希望把nginx和tomcat都部署同一个地方的node节点上，可以提高通信效率；
podunaffinity：pod和pod更倾向不腻在一起，如果部署两套程序，那么这两套程序更倾向于反亲和性，这样相互之间不会有影响。
第一个pod随机选则一个节点，做为评判后续的pod能否到达这个pod所在的节点上的运行方式，这就称为pod亲和性；我们怎么判定哪些节点是相同位置的，哪些节点是不同位置的；我们在定义pod亲和性时需要有一个前提，哪些pod在同一个位置，哪些pod不在同一个位置，这个位置是怎么定义的，标准是什么？以节点名称为标准，这个节点名称相同的表示是同一个位置，节点名称不相同的表示不是一个位置。

kubectl explain pods.spec.affinity.podAffinity

KIND:     Pod
VERSION:  v1
RESOURCE: podAffinity <Object>
DESCRIPTION:
     Describes pod affinity scheduling rules (e.g. co-locate this pod in the
     same node, zone, etc. as some other pod(s)).
     Pod affinity is a group of inter pod affinity scheduling rules.
FIELDS:
   preferredDuringSchedulingIgnoredDuringExecution	<[]Object>
   requiredDuringSchedulingIgnoredDuringExecution	<[]Object>
   
requiredDuringSchedulingIgnoredDuringExecution： 硬亲和性
preferredDuringSchedulingIgnoredDuringExecution：软亲和性

kubectl explain pods.spec.affinity.podAffinity.requiredDuringSchedulingIgnoredDuringExecution

KIND:     Pod
VERSION:  v1
RESOURCE: requiredDuringSchedulingIgnoredDuringExecution <[]Object>
DESCRIPTION:
FIELDS:
   labelSelector	<Object>
   namespaces	<[]string>
   topologyKey	<string> -required-

topologyKey：
位置拓扑的键，这个是必须字段
怎么判断是不是同一个位置：
rack=rack1
row=row1
使用rack的键是同一个位置
使用row的键是同一个位置
labelSelector：
我们要判断pod跟别的pod亲和，跟哪个pod亲和，需要靠labelSelector，通过labelSelector选则一组能作为亲和对象的pod资源
namespace：
labelSelector需要选则一组资源，那么这组资源是在哪个名称空间中呢，通过namespace指定，如果不指定namespaces，那么就是当前创建pod的名称空间

在这里插入图片描述

定义两个pod，第一个pod做为基准，第二个pod跟着它走

查看默认名称空间有哪些pod,把看到的pod删除，让默认名称空间没有pod

kubectl get pods

创建的pod必须与拥有app=myapp标签的pod在一个节点上

vim pod-required-affinity-demo-1.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-first
  labels:
    app2: myapp2
    tier: frontend
spec:
    containers:
    - name: myapp
      image: ikubernetes/myapp:v1
      imagePullPolicy: IfNotPresent

kubectl apply -f pod-required-affinity-demo-1.yaml

vim pod-required-affinity-demo-2.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-second
  labels:
    app: backend
    tier: db
spec:
    containers:
    - name: busybox
      image: busybox:latest
      imagePullPolicy: IfNotPresent
      command: ["sh","-c","sleep 3600"]
    affinity:
      podAffinity:
         requiredDuringSchedulingIgnoredDuringExecution:
         - labelSelector:
              matchExpressions:
              - {key: app2, operator: In, values: ["myapp2"]}
           topologyKey: kubernetes.io/hostname

kubectl apply -f pod-required-affinity-demo-2.yaml

第一个pod调度到哪，第二个pod也调度到哪，这就是pod节点亲和性

kubectl get pods -o wide

pod-first              running        k8snode2
pod-second             running        k8snode2

删除测试pod

kubectl delete -f pod-required-affinity-demo-1.yaml
kubectl delete -f pod-required-affinity-demo-2.yaml

pod节点反亲和性

定义两个pod，第一个pod做为基准，第二个pod跟它调度节点相反

vim pod-required-anti-affinity-demo-1.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-first
  labels:
    app1: myapp1
    tier: frontend
spec:
    containers:
    - name: myapp
      image: ikubernetes/myapp:v1
      imagePullPolicy: IfNotPresent

kubectl apply -f  pod-required-anti-affinity-demo-1.yaml

vim pod-required-anti-affinity-demo-2.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-second
  labels:
    app: backend
    tier: db
spec:
    containers:
    - name: busybox
      image: busybox:latest
      imagePullPolicy: IfNotPresent
      command: ["sh","-c","sleep 3600"]
    affinity:
      podAntiAffinity:
         requiredDuringSchedulingIgnoredDuringExecution:
         - labelSelector:
              matchExpressions:
              - {key: app1, operator: In, values: ["myapp1"]}
           topologyKey: kubernetes.io/hostname

kubectl apply -f pod-required-anti-affinity-demo-2.yaml

显示两个pod不在一个node节点上，这就是pod节点反亲和性

kubectl get pods -o wide

pod-first            running        k8snode1
pod-second           running        k8snode2

删除测试用pod

kubectl delete -f pod-required-anti-affinity-demo-1.yaml
kubectl delete -f pod-required-anti-affinity-demo-2.yaml

topologykey 位置拓扑键

kubectl label nodes  k8snode2  zone=foo
kubectl label nodes  k8snode1  zone=foo

vim pod-first-required-anti-affinity-demo-1.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-first
  labels:
    app3: myapp3
    tier: frontend
spec:
    containers:
    - name: myapp
      image: ikubernetes/myapp:v1
      imagePullPolicy: IfNotPresent

kubectl apply -f pod-first-required-anti-affinity-demo-1.yaml

vim pod-second-required-anti-affinity-demo-1.yaml

apiVersion: v1
kind: Pod
metadata:
  name: pod-second
  labels:
    app: backend
    tier: db
spec:
    containers:
    - name: busybox
      image: busybox:latest
      imagePullPolicy: IfNotPresent
      command: ["sh","-c","sleep 3600"]
    affinity:
      podAntiAffinity:
         requiredDuringSchedulingIgnoredDuringExecution:
         - labelSelector:
              matchExpressions:
              - {key: app3 ,operator: In, values: ["myapp3"]}
           topologyKey:  zone

kubectl apply -f pod-second-required-anti-affinity-demo-1.yaml

kubectl get pods -o wide

显示如下

pod-first              running         k8snode1
pod-second            pending         <none>

第二个pod是pending，因为两个节点是同一个位置，现在没有不是同一个位置的了，而且我们要求反亲和性，所以就会处于pending状态，如果在反亲和性这个位置把required改成preferred，那么也会运行。

kubectl delete -f pod-first-required-anti-affinity-demo-1.yaml 
kubectl delete -f pod-second-required-anti-affinity-demo-1.yaml

移除标签

kubectl label nodes  k8snode1   zone-
kubectl label nodes  k8snode2   zone-

podaffinity：pod节点亲和性，pod倾向于哪个pod
poduntiaffinity:pod反亲和性
nodeaffinity：node节点亲和性，pod倾向于哪个node