本文主要是介绍k8s学习(三) k8s集群内部署一个docker应用deployment,希望对大家解决编程问题提供一定的参考价值,需要的开发者们随着小编来一起学习吧!
1、获取一个springboot 应用jar包
2、将jar包打成docker镜像
Dockerfile如下:
FROM java:8-alpine
ADD docker-hello-world-0.0.1-SNAPSHOT.jar app.jar
EXPOSE 8080
ENTRYPOINT ["java","-jar", "/app.jar"]
打镜像:
docker build -t hellword:0.0.4 .
3、准备docker 镜像私服
参考https://blog.csdn.net/u011943534/article/details/81331231
4、配置私服地址
编辑/etc/docker/daemon.json,添加nexus仓库地址
{ "exec-opts": ["native.cgroupdriver=systemd"],"insecure-registries":["192.168.100.89:80", "172.16.10.190:8008", "quay.io", "k8s.gcr.io", "gcr.io"]
}
重启docker
systemctl restart docker
5、将镜像推送到远程私服
docker tag helloworld:0.0.4 172.16.10.190:8008/helloworld:0.0.4
登陆nexus
docker login --username=admin 172.16.10.190:8008
输入密码登陆
docker push 172.16.10.190:8008/helloworld:0.0.4
6、配置secret拉取私有镜像仓库内容
kubectl create secret docker-registry myregistrykey --docker-server=172.16.10.190:8008 --docker-username=admin --docker-password=admin123 --docker-email=admin@example.org
获取创建的secret
kubectl get secretsNAME TYPE DATA AGE
cluster-admin-dashboard-sa-token-hld8l kubernetes.io/service-account-token 3 3d19h
default-token-7c9dp kubernetes.io/service-account-token 3 4d
myregistrykey kubernetes.io/dockerconfigjson 1 3d16h
myregistrykey2 kubernetes.io/dockerconfigjson 1 3d16h
myregistrykey3 kubernetes.io/dockerconfigjson 1 3d16h
myregistrykey4 kubernetes.io/dockerconfigjson 1 2d22h
myregistrykey5 kubernetes.io/dockerconfigjson 1 2d22h
myregistrykey6 kubernetes.io/dockerconfigjson 1
7、创建deployment.yaml
imagePullSecrets处配置上一步获得的secret
---apiVersion: extensions/v1beta1kind: Deploymentmetadata:name: kube-nodespec:replicas: 2template:metadata:labels:app: webspec:containers:- name: kube-node-demo-instanceimage: kube-node-demo1ports:- containerPort: 8081imagePullSecrets:- name: myregistrykey
8、部署
kubectl create -f deployment.yaml
9、查看部署状态
如果状态为running部署成功
[root@k8s-node1 k8s]# kubectl get pods
NAME READY STATUS RESTARTS AGE
kube-node-87dcfc9cf-9kj8j 1/1 Running 0 25m
kube-node-87dcfc9cf-l2ppz 1/1 Running 0 25m
10、测试应用
[root@k8s-node1 k8s]# kubectl get pods -o wide
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
kube-node-87dcfc9cf-9kj8j 1/1 Running 0 26m 10.244.1.23 k8s-node3 <none> <none>
kube-node-87dcfc9cf-l2ppz 1/1 Running 0 26m 10.244.2.24 k8s-node2 <none> <none>[root@k8s-node1 k8s]# curl 10.244.2.24:8080/index
hello,world
11、如果要删除应用,可以删除deployment
kubectl delete deployment/kube-node
12、重启deployment
kubectl rollout restart deployment your_deployment_name
13、滚动升级
---apiVersion: extensions/v1beta1kind: Deploymentmetadata:name: kube-nodespec:replicas: 2template:metadata:labels:app: webspec:containers:- name: kube-node-demo-instanceimage: 172.16.10.190:8008/helloworld:0.0.5ports:- containerPort: 8080imagePullSecrets:- name: myregistrykey6minReadySeconds: 10strategy:type: RollingUpdaterollingUpdate:maxSurge: 1maxUnavailable: 1
minReadySeconds:
Kubernetes在等待设置的时间后才进行升级
如果没有设置该值,Kubernetes会假设该容器启动起来后就提供服务了
如果没有设置该值,在某些极端情况下可能会造成服务不正常运行
maxSurge:
升级过程中最多可以比原先设置多出的POD数量
例如:maxSurage=1,replicas=5,则表示Kubernetes会先启动1一个新的Pod后才删掉一个旧的POD,整个升级过程中最多会有5+1个POD。
maxUnavaible:
升级过程中最多有多少个POD处于无法提供服务的状态
当maxSurge不为0时,该值也不能为0
例如:maxUnavaible=1,则表示Kubernetes整个升级过程中最多会有1个POD处于无法服务的状态。
升级
[root@k8s-node1 k8s]# kubectl apply -f deployment.yaml --record=true
deployment.extensions/kube-node configured
查看升级状态
root@k8s-node1 k8s]# kubectl rollout status deployment/kube-node
Waiting for deployment "kube-node" rollout to finish: 1 old replicas are pending termination...
Waiting for deployment "kube-node" rollout to finish: 1 old replicas are pending termination...
deployment "kube-node" successfully rolled out
暂停升级
kubectl rollout pause deployment <deployment>
继续升级
kubectl rollout resume deployment <deployment>
查看RS状态
[root@k8s-node1 k8s]# kubectl get rs
NAME DESIRED CURRENT READY AGE
kube-node-6545cf5596 0 0 0 30h
kube-node-6b9f6d4b59 2 2 2 51m
kube-node-6cd4f7bd56 0 0 0 11m
14、版本回退
查看历史版本:
[root@k8s-node1 k8s]# kubectl rollout history deployment kube-node
deployment.extensions/kube-node
REVISION CHANGE-CAUSE
1 <none>
3 <none>
4 kubectl apply --filename=deployment.yaml --record=true
在升级的时候最好携带参数–record=true,方便查看历史版本信息。如使用 kubectl apply -f deployment.yaml --record=true
在deployment.yaml中最好控制记录历史版本数量,因为每个历史版本都记录在etcd中,多了会造成etcd的负担。
添加一个revisionHistoryLimit: 10参数
spec:revisionHistoryLimit: 10replicas: 2
查看某个历史版本的信息:
[root@k8s-node1 k8s]# kubectl rollout history deployment kube-node --revision=4
deployment.extensions/kube-node with revision #4
Pod Template:Labels: app=webpod-template-hash=6b9f6d4b59Annotations: kubectl.kubernetes.io/restartedAt: 2019-08-29T13:42:33+08:00kubernetes.io/change-cause: kubectl apply --filename=deployment.yaml --record=trueContainers:kube-node-demo-instance:Image: 172.16.10.190:8008/helloworld:0.0.5Port: 8080/TCPHost Port: 0/TCPEnvironment: <none>Mounts: <none>Volumes: <none>
假如现在要直接回退到当前版本的前一个版本:
# kubectl rollout undo deployment kube-node
deployment "kube-node" rolled back
用revision回退到指定的版本:
# kubectl rollout undo deployment kube-node --to-revision=3
deployment "kube-node" rolled back
这篇关于k8s学习(三) k8s集群内部署一个docker应用deployment的文章就介绍到这儿,希望我们推荐的文章对编程师们有所帮助!