Kubernetes-Affinity(node亲和、pod亲和)
目录
一、节点亲和性
（1）节点硬亲和
（2）节点软亲和
（3）软硬亲和共同使用
二、Pod亲和
1.Pod亲和性
2.Pod反亲和性
三、亲和性/反亲和性调度策略比较
一、节点亲和性
pod.spec.nodeAffinity
preferredDuringSchedulingIgnoredDuringExecution：软策略
requiredDuringSchedulingIgnoredDuringExecution：硬策略
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
（1）节点硬亲和
requiredDuringSchedulingIgnoredDuringExecution（硬亲和）
apiVersion: v1
kind: Pod
metadata:
name: affinity
labels:
app: node-affinity-pod
spec:
containers:
- name: node-affinity
image: centos:7.9.2009
affinity: #亲和性
nodeAffinity: #节点亲和性
requiredDuringSchedulingIgnoredDuringExecution: #硬性限制
nodeSelectorTerms: #节点的选择器限制
- matchExpressions: #匹配运算符
- key: kubernetes.io/hostname
operator: NotIn #不在列表
values:
- k8s-node02 #不匹配node02节点（不将此pod调度到k8s-node02节点）
节点硬亲和实验：
vim node_required_affinity.yaml
apiVersion: v1
kind: Pod
metadata:
name: affinity
labels:
app: node-affinity-pod
spec:
containers:
- name: node-affinity
image: nginx:latest
imagePullPolicy: IfNotPresent
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: kubernetes.io/hostname
operator: NotIn
values:
- k8s-node02
kubectl create -f node_required_affinity.yaml
#创建pod
kubectl get pod -o wide
#查看新建的pod调度到哪个节上
vim node_required_affinity.yaml
#修改operator:NotIn改为In
#values改为k8s-node03
apiVersion: v1
kind: Pod
metadata:
name: affinity
labels:
app: node-affinity-pod
spec:
containers:
- name: node-affinity
image: nginx:latest
imagePullPolicy: IfNotPresent
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: kubernetes.io/hostname
operator: In
values:
- k8s-node03
kubectl delete -f node_required_affinity.yaml
#删除刚刚创建的pod
kubectl apply -f node_required_affinity.yaml
#根据更改后的yaml文件，再次创建pod
kubectl get pod -o wide
#没有node03节点，硬亲和匹配不到，所以pod一直pending。
（2）节点软亲和
preferredDuringSchedulingIgnoredDuringExecution（软亲和）
apiVersion: v1
kind: Pod
metadata:
name: affinity
labels:
app: node-affinity-pod
spec:
containers:
- name: node-affinity
image: nginx:latest
affinity: #亲和性
nodeAffinity: #节点亲和性
preferredDuringSchedulingIgnoredDuringExecution: #软性亲和
- weight: 1 #权重
preference:
matchExpressions:
- key: kubernetes.io/hostname
operator: In #因为没有三节点，所以软亲和会在从节点1或2上选
values:
- k8s-node02
节点软亲和实验：
vim node_preferred_affinity.yaml
apiVersion: v1
kind: Pod
metadata:
name: affinity
labels:
app: node-affinity-pod
spec:
containers:
- name: with-node-affinity
image: nginx:latest
imagePullPolicy: IfNotPresent
affinity:
nodeAffinity:
preferredDuringSchedulingIgnoredDuringExecution:
- weight: 1
preference:
matchExpressions:
- key: kubernetes.io/hostname
operator: In #匹配规则IN
values:
- k8s-node02 #优先匹配k8s-node02节点
kubectl create -f node_preferred_affinity.yaml
kubectl get pod -o wide
#查看pod调度的节点
kubectl delete -f node_preferred_affinity.yaml
vim node_preferred_affinity.yaml
#修改k8s-node02为k8s-node03
apiVersion: v1
kind: Pod
metadata:
name: affinity
labels:
app: node-affinity-pod
spec:
containers:
- name: with-node-affinity
image: nginx:latest
imagePullPolicy: IfNotPresent
affinity:
nodeAffinity:
preferredDuringSchedulingIgnoredDuringExecution:
- weight: 1
preference:
matchExpressions:
- key: kubernetes.io/hostname
operator: In #匹配规则IN
values:
- k8s-node03 #优先匹配k8s-node03节点，如果没有node03节点，会在node01和02之间随机选择。
kubectl create -f node_preferred_affinity.yaml
kubectl get pod -o wide
#查看pod调度的节点（3）软硬亲和共同使用
apiVersion: v1
kind: Pod
metadata:
name: affinity
labels:
app: node-affinity-pod
spec:
containers:
- name: node-affinity
image: nginx:latest
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution: #硬亲和
nodeSelectorTerms: #节点选择器
- matchExpressions: #匹配运算符
- key: kubernetes.io/hostname
operator: NotIn #除了node02节点，其他都可以运行
values:
- k8s-node02
preferredDuringSchedulingIgnoredDuringExecution: #软亲和
- weight: 1
preference:
matchExpressions:
- key: source
operator: In
values:
- aaa #除了node02节点不匹配，优先匹配带有aaa标签的节点。
二、Pod亲和
1.Pod亲和性
pod.spec.affinity.podAffinity/podAntiAffinity
preferredDuringSchedulingIgnoredDuringExecution：软策略
requiredDuringSchedulingIgnoredDuringExecution：硬策略
apiVersion: v1
kind: Pod
metadata:
name: pod-3
labels:
app: pod-3
spec:
containers:
- name: pod-3
image: nginx:latest
affinity:
podAffinity: #pod亲和性
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app
operator: In #在列表
values:
- pod-1 #pod-1在哪个节点，就匹配到哪个节点。
topologyKey: kubernetes.io/hostname
#是否在一起，根据这个topologyKey决定，他的values只要相同，那么就认为是在一起的
podAntiAffinity: #反亲和性
preferredDuringSchedulingIgnoredDuringExecution:
- weight: 1
podAffinityTerm:
labelSelector:
matchExpressions:
- key: app
operator: In
values:
- pod-2
topologyKey: kubernetes.io/hostname
pod亲和性实验：
vim pod_affinity.yaml
apiVersion: v1
kind: Pod
metadata:
name: pod-affinity
labels:
app: pod-affinity
spec:
containers:
- name: pod-affinity
image: nginx:latest
affinity:
podAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app
operator: In
values:
- pod-1
topologyKey: kubernetes.io/hostname
kubectl create -f pod_affinity.yaml
kubectl get pod -o wide
#必须得有一个具备此标签pod才可以，现在没有，所以是Pending状态
vim pod-1.yaml
#创建一个带有pod-1标签的pod
apiVersion: v1
kind: Pod
metadata:
name: pod
labels:
app: pod-1
spec:
containers:
- name: pod-1
image: nginx:latest
imagePullPolicy: IfNotPresent
kubectl create -f pod-1.yaml
#创建一个带有pod-1标签的pod
kubectl get pod -o wide
#pod-affinity找到了匹配的标签，达到running状态。
kubectl get pod --show-labels
#查看pod的标签2.Pod反亲和性
apiVersion: v1
kind: Pod
metadata:
name: pod-antiaffinity
labels:
app: pod-antiaffinity
spec:
containers:
- name: pod-antiaffinity
image: nginx:latest
affinity:
podAntiAffinity: #反亲和性
preferredDuringSchedulingIgnoredDuringExecution: #软限制策略
- weight: 1
podAffinityTerm:
labelSelector:
matchExpressions:
- key: app
operator: In #在列表
values:
- pod-2 #匹配pod-2
topologyKey: kubernetes.io/hostname
#由于是反亲和，根据上述配置，软限制，优先选择pod-2不在的节点运行。
pod反亲和实验：
vim pod_antiaffinity.yaml
apiVersion: v1
kind: Pod
metadata:
name: pod-antiaffinity
labels:
app: pod-antiaffinity
spec:
containers:
- name: pod-antiaffinity
image: nginx:latest
imagePullPolicy: IfNotPresent
affinity:
podAntiAffinity:
preferredDuringSchedulingIgnoredDuringExecution:
- weight: 1
podAffinityTerm:
labelSelector:
matchExpressions:
- key: app
operator: In #在列表
values:
- pod-1
topologyKey: kubernetes.io/hostname
#最终pod优先不与pod-1标签的pod在一个节点
kubectl create -f pod_antiaffinity.yaml
kubectl get pod -o wide
#查看pod的运行节点三、亲和性/反亲和性调度策略比较
| 调度策略 |
匹配标签 |
操作符 |
拓扑域支持 |
调度目标 |
| nodeAffinity |
主机 |
In, NotIn, Exists, DoesNotExist, Gt, Lt |
否 |
指定主机 |
| podAffinity |
POD |
In, NotIn, Exists, DoesNotExist |
是 |
POD与指定POD同一拓扑域 |
| podAnitAffinity |
POD |
In, NotIn, Exists, DoesNotExist |
是 |
POD与指定POD不在同一拓扑域 |