karmada安装和基本使用
karmada是k8s的一个多集群实现方案
安装
wget https://github.com/karmada-io/karmada/releases/download/v1.9.0/karmadactl-linux-amd64.tgz- 也可以安装 kubectl的插件
kubectl krew install karmada- 安装karmadactl到集群中
karmadactl init - 安装完成之后默认会在目录
/etc/karmada/添加这个虚拟集群的相关文件
多集群成员
push模式添加成员
- 添加集群push模式，–kubeconfig决定要加入的集群
karmadactl --kubeconfig /etc/karmada/karmada-apiserver.config join cluster1 --cluster-kubeconfig=$HOME/.kube/config查看集群
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config get clusterspull模式添加成员
- 获取注册命令和凭据
karmadactl token create --print-register-command --kubeconfig /etc/karmada/karmada-apiserver.config
# karmadactl register 10.7.143.254:32443 --cluster-name cluster2 --token po8les.a05eqne2hqwy8gly --discovery-token-ca-cert-hash sha256:3bfb29c846092b61af5bb51a47a88bd52ed834d8468158a7fb341180d5a3bc74删除集群
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config delete cluster cluster2应用分发
- 首先查看下有那些集群
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config get clusters
# NAME VERSION MODE READY AGE
# cluster1 v1.24.12 Push True 10d
# kubernetes v1.24.12 Pull True 9d- 部署分发策略
# propagationpolicy.yaml
apiVersion: policy.karmada.io/v1alpha1
kind: PropagationPolicy
metadata:
name: example-policy # The default namespace is `default`.
spec:
resourceSelectors:
- apiVersion: apps/v1
kind: Deployment
name: nginx # If no namespace is specified, the namespace is inherited from the parent object scope.
placement:
clusterAffinity:
clusterNames:
- cluster1
- kubernetes- 部署上面的分发策略
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config apply -f propagationpolicy.yaml- 其中一个集群中部署一个测试的nginx
kubectl create deployment nginx --image nginx- 在karmada中也创建一个测试的
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config create deployment nginx --image nginx- 在第二个集群中可以看到同步过来了，此时同步完成
kubectl --kubeconfig=./2.yaml get po
# NAME READY STATUS RESTARTS AGE
# nginx-8f458dc5b-w526c 1/1 Running 0 5m43s- 当我们删除karmada中的nginx
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config delete deploy
# deployment.apps "nginx" deleted
# 集群1的资源还在
kubectl get po
# NAME READY STATUS RESTARTS AGE
# nginx-8f458dc5b-t6z4k 1/1 Running 0 7m35s
# 集群2的资源不在了
kubectl --kubeconfig=./2.yaml get po
# No resources found in default namespace.- 将集群1的测试资源也删除
kubectl delete deploy nginx
# deployment.apps "nginx" deleted- 我们只在karmada中部署nginx测试
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config create deployment nginx --image nginx
#deployment.apps/nginx created
# 可以看到 ready是2/1 因为2个集群都同步了
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config get deploy
# NAME READY UP-TO-DATE AVAILABLE AGE
# nginx 2/1 2 2 15s
# 使用karmadactl查看
karmadactl --kubeconfig=/etc/karmada/karmada-apiserver.config get po
# NAME CLUSTER READY STATUS RESTARTS AGE
# nginx-74f96fcc5c-zphtd cluster1 1/1 Running 0 6m6s
# nginx-8f458dc5b-mkpx9 kubernetes 1/1 Running 0 11m- 从上面的现象可以看到同步的时候karmada集群中必须有需要同步资源,当成员集群已经有了该资源则在删除的时候不会删除
覆盖应用
覆盖主要用于重写一些属性
- 保存下面的覆盖策略为overridepolicy.yaml
# overridepolicy.yaml
apiVersion: policy.karmada.io/v1alpha1
kind: OverridePolicy
metadata:
name: example
spec:
resourceSelectors:
- apiVersion: apps/v1
kind: Deployment
name: nginx
labelSelector:
matchLabels:
app: nginx
overrideRules:
- targetCluster:
clusterNames:
- cluster1
overriders:
imageOverrider:
- component: Tag
operator: replace
value: '1.20'- apply这个策略到karmada集群
kubectl --kubeconfig /etc/karmada/karmada-apiserver.config apply -f overridepolicy.yaml- 查看集群1的nginx已经被改为1.20版本了
kubectl get deploy -o wide
# NAME READY UP-TO-DATE AVAILABLE AGE CONTAINERS IMAGES SELECTOR
# nginx 1/1 1 1 6m12s nginx nginx:1.20 app=nginx