亲和性调度可以分成软策略和硬策略两种方式:
软策略就是如果你没有满足调度要求的节点的话,pod 就会忽略这条规则,继续完成调度过程,说白了就是满足条件最好了,没有的话也无所谓了的策略
硬策略就比较强硬了,如果没有满足条件的节点的话,就不断重试直到满足条件为止,简单说就是你必须满足我的要求,不然我就不干的策略。
1、通过nodeSelector 和label选择节点调度
查看节点的label
[root@k8s-node1 test]# kubectl get nodes --show-labels
NAME STATUS ROLES AGE VERSION LABELS
k8s-node1 Ready master 16d v1.15.2 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node1,kubernetes.io/os=linux,node-role.kubernetes.io/master=
k8s-node2 Ready 16d v1.15.2 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node2,kubernetes.io/os=linux
k8s-node3 Ready 16d v1.15.2 beta.kubernetes.io/arch=amd64,beta.kubernetes.io/os=linux,kubernetes.io/arch=amd64,kubernetes.io/hostname=k8s-node3,kubernetes.io/os=linux
为k8s-node3添加一个标签
[root@k8s-node1 test]# kubectl label nodes k8s-node3 com=zqw
node/k8s-node3 labeled
构建一个pod使用此标签
---
apiVersion: v1
kind: Pod
metadata:
labels:
app: helloworld-nodeselector
name: helloworld-nodeselector
spec:
containers:
- name: helloworld-nodeselector
image: 172.16.10.190:8008/helloworld:0.1.1
imagePullPolicy: Always
imagePullSecrets:
- name: myregistrykey6
nodeSelector:
com: zqw
kubectl apply -f helloworld-nodeSelector.yaml
查看pod
[root@k8s-node1 k8s]# kubectl get pods -owide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
dns-test 1/1 Running 0 14d 10.244.2.26 k8s-node3
helloworld-nodeselector 1/1 Running 0 2m10s 10.244.2.61 k8s-node3
helloworld011-5cf986495c-6zslr 1/1 Running 0 14d 10.244.2.27 k8s-node3
helloworld011-5cf986495c-9xh5d 1/1 Running 0 14d 10.244.1.20 k8s-node2
innocent-boxer-hello-helm-867bd48ccc-k6jmw 1/1 Running 0 19h 10.244.0.33 k8s-node1
kube-hpa-demo-d789d89b-9xgdx 1/1 Running 0 16d 10.244.1.9 k8s-node2
kube-hpa-demo-d789d89b-pw77j 1/1 Running 0 16d 10.244.1.8 k8s-node2
kube-node-6545cf5596-khw4x 1/1 Running 0 16d 10.244.2.3 k8s-node3
kube-node-6545cf5596-q7kd7 1/1 Running 0 16d 10.244.1.5 k8s-node2
liveness-tcpsocket 1/1 Running 0 16d 10.244.1.6 k8s-node2
myrabbitmq-0 1/1 Running 0 18h 10.244.2.60 k8s-node3
mysqldb-6ffc996bdb-lxgsp 0/1 Pending 0 19h
nfs-client-provisioner-84f967cff9-dtxxt 1/1 Running 0 9d 10.244.2.43 k8s-node3
nfs-web-0 1/1 Running 0 9d 10.244.2.44 k8s-node3
nfs-web-1 1/1 Running 0 9d 10.244.1.22 k8s-node2
nginx-ds-7spc2 1/1 Running 0 15d 10.244.1.12 k8s-node2
nginx-ds-ffd9v 1/1 Running 0 15d 10.244.2.13 k8s-node3
nginx-ds-rmqvl 1/1 Running 0 5d23h 10.244.0.7 k8s-node1
static-nginx-k8s-node1 1/1 Running 0 16d 10.244.0.2 k8s-node1
use-pvc-deployment-85997f976b-llg9h 1/1 Running 0 9d 10.244.2.42 k8s-node3
use-pvc-deployment-85997f976b-tm92t 1/1 Running 0 9d 10.244.1.21 k8s-node2
web-0 1/1 Running 0 14d 10.244.2.25 k8s-node3
web-1 1/1 Running 0 14d 10.244.1.19 k8s-node2
发现helloworld-nodeselector在k8s-node3节点上
2、使用节点亲和性策略nodeAffinity
创建helloworld-nodeAffinity.yaml
---
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: helloworld-nodeaffinity
spec:
revisionHistoryLimit: 10
replicas: 3
template:
metadata:
labels:
app: web-nodeaffinity
spec:
containers:
- name: helloworld-nodeaffinity
image: 172.16.10.190:8008/helloworld:0.0.5
ports:
- containerPort: 8080
imagePullSecrets:
- name: myregistrykey6
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution: # 硬策略
nodeSelectorTerms:
- matchExpressions:
- key: kubernetes.io/hostname
operator: NotIn
values:
- k8s-node2
preferredDuringSchedulingIgnoredDuringExecution: # 软策略
- weight: 1
preference:
matchExpressions:
- key: com
operator: In
values:
- zqw
minReadySeconds: 10
strategy:
type: RollingUpdate
rollingUpdate:
maxSurge: 1
maxUnavailable: 1
kubectl apply -f helloworld-nodeAffinity.yaml
查看pod信息
[root@k8s-node1 k8s]# kubectl get pods -owide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
dns-test 1/1 Running 0 15d 10.244.2.26 k8s-node3
helloworld-nodeaffinity-5cfdccb89-9k4vz 1/1 Running 0 13s 10.244.2.63 k8s-node3
helloworld-nodeaffinity-5cfdccb89-dqqlk 1/1 Running 0 13s 10.244.0.37 k8s-node1
helloworld-nodeaffinity-5cfdccb89-k89zq 1/1 Running 0 13s 10.244.2.62 k8s-node3
helloworld-nodeselector 1/1 Running 0 16m 10.244.2.61 k8s-node3
helloworld011-5cf986495c-6zslr 1/1 Running 0 14d 10.244.2.27 k8s-node3
helloworld011-5cf986495c-9xh5d 1/1 Running 0 14d 10.244.1.20 k8s-node2
innocent-boxer-hello-helm-867bd48ccc-k6jmw 1/1 Running 0 19h 10.244.0.33 k8s-node1
kube-hpa-demo-d789d89b-9xgdx 1/1 Running 0 16d 10.244.1.9 k8s-node2
kube-hpa-demo-d789d89b-pw77j 1/1 Running 0 16d 10.244.1.8 k8s-node2
kube-node-6545cf5596-khw4x 1/1 Running 0 16d 10.244.2.3 k8s-node3
kube-node-6545cf5596-q7kd7 1/1 Running 0 16d 10.244.1.5 k8s-node2
liveness-tcpsocket 1/1 Running 0 16d 10.244.1.6 k8s-node2
myrabbitmq-0 1/1 Running 0 18h 10.244.2.60 k8s-node3
mysqldb-6ffc996bdb-lxgsp 0/1 Pending 0 19h
nfs-client-provisioner-84f967cff9-dtxxt 1/1 Running 0 9d 10.244.2.43 k8s-node3
nfs-web-0 1/1 Running 0 9d 10.244.2.44 k8s-node3
nfs-web-1 1/1 Running 0 9d 10.244.1.22 k8s-node2
nginx-ds-7spc2 1/1 Running 0 15d 10.244.1.12 k8s-node2
nginx-ds-ffd9v 1/1 Running 0 15d 10.244.2.13 k8s-node3
nginx-ds-rmqvl 1/1 Running 0 5d23h 10.244.0.7 k8s-node1
static-nginx-k8s-node1 1/1 Running 0 16d 10.244.0.2 k8s-node1
use-pvc-deployment-85997f976b-llg9h 1/1 Running 0 9d 10.244.2.42 k8s-node3
use-pvc-deployment-85997f976b-tm92t 1/1 Running 0 9d 10.244.1.21 k8s-node2
web-0 1/1 Running 0 15d 10.244.2.25 k8s-node3
web-1 1/1 Running 0 15d 10.244.1.19 k8s-node2
发现不在k8s-node2上部署,在k8s-node3上部署
部署规则包括以下几种:
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
3、pod亲和性和反亲和性
PodAffinity和podAntiAffinity参考
https://www.qikqiak.com/k8s-book/docs/51.Kubernetes%E4%BA%B2%E5%92%8C%E6%80%A7%E8%B0%83%E5%BA%A6.html