创建maven工程,pom.xml中引入连接k8s的客户端jar包:
<properties>
<maven.compiler.source>8maven.compiler.source>
<maven.compiler.target>8maven.compiler.target>
<fabric.io.version>6.10.0fabric.io.version>
properties>
<dependencies>
<dependency>
<groupId>io.fabric8groupId>
<artifactId>kubernetes-clientartifactId>
<version>${fabric.io.version}version>
dependency>
dependencies>
目标:使用尽可能少的配置项来完成初始化工作,保证后续部署工作量最小
使用kubeconfig文件连接集群
使用场景
环境变量配置
默认配置项:
KUBECONFIG=~/.kube/config
自定义配置:
KUBECONFIG=/opt/file/kubeconfig
示例代码
import io.fabric8.kubernetes.api.model.Namespace;
import io.fabric8.kubernetes.api.model.NamespaceList;
import io.fabric8.kubernetes.client.KubernetesClient;
import io.fabric8.kubernetes.client.KubernetesClientBuilder;
public class K8sClientWithKubeConfig {
public static void main(String[] args) {
KubernetesClient client = new KubernetesClientBuilder().build();
System.out.println(client.getMasterUrl());
NamespaceList myNs = client.namespaces().list();
for (Namespace ns : myNs.getItems()) {
System.out.println(ns.getMetadata().getName());
}
client.close();
}
}
其中master url为kubeconfig文件中的server字段;
使用ServiceAccount连接集群
使用场景
应用部署在k8s集群内,且可以创建Service Account和进行授权等操作(RBAC)
创建SA
apiVersion: v1
kind: ServiceAccount
metadata:
name: fmuser
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
name: fmuser-role
rules:
- apiGroups: [""]
resources: ["pods", "namespaces"]
verbs: ["get", "list", "watch"]
- apiGroups: [""]
resources: ["pods/log"]
verbs: ["get", "list", "watch"]
- apiGroups: [""]
resources: ["pods"]
verbs: ["delete"]
- apiGroups: [""]
resources: ["namespaces"]
verbs: ["get", "list", "watch"]
---
apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRoleBinding
metadata:
name: fmuser-rolebinding
subjects:
- kind: ServiceAccount
name: fmuser
namespace: default
roleRef:
kind: ClusterRole
name: fmuser-role
apiGroup: rbac.authorization.k8s.io
资源授权根据实际使用情况设置即可,Role(当前空间,权限较小)和ClusterRole(整个集群,权限较大)根据需要创建;在代码中使用时,不需要额外配置,客户端会从以下路径自动读取:
/home/app # ls -l /var/run/secrets/kubernetes.io/serviceaccount/
total 0
lrwxrwxrwx 1 root root 13 Jan 22 02:55 ca.crt -> ..data/ca.crt
lrwxrwxrwx 1 root root 16 Jan 22 02:55 namespace -> ..data/namespace
lrwxrwxrwx 1 root root 12 Jan 22 02:55 token -> ..data/token
/home/app #
所需的文件就是ca.crt和token;
如果需要在集群外验证这种认证方式,需要设置如下环境变量:
KUBERNETES_MASTER=https://191.168.7.132:6443;
KUBERNETES_AUTH_SERVICEACCOUNT_TOKEN=/opt/file/serviceaccount-token;
KUBERNETES_CERTS_CA_FILE=/opt/file/serviceaccount-ca.key;
KUBERNETES_AUTH_TRYKUBECONFIG=false
其中serviceaccount-ca.key和serviceaccount-token对应的是上面的ca.crt和token,需要注意的是,使用这种方式,master的url中的端口是6443,而不是默认的443,需要注意;
示例代码
import io.fabric8.kubernetes.api.model.Namespace;
import io.fabric8.kubernetes.api.model.NamespaceList;
import io.fabric8.kubernetes.client.KubernetesClient;
import io.fabric8.kubernetes.client.KubernetesClientBuilder;
public class K8sClientWithServiceAccount {
public static void main(String[] args) {
KubernetesClient client = new KubernetesClientBuilder().build();
System.out.println(client.getConfiguration().getCaCertFile());
System.out.println(client.getConfiguration().getMasterUrl());
System.out.println(client.getConfiguration().getAutoOAuthToken());
NamespaceList myNs = client.namespaces().list();
for (Namespace ns : myNs.getItems()) {
System.out.println(ns.getMetadata().getName());
}
client.close();
}
}
获取sa的ca.crt和token
root@tianshu-ai-platform:~# kubectl get sa
NAME SECRETS AGE
default 1 13d
fmuser 1 7d
test 1 7d
root@tianshu-ai-platform:~# kubectl describe sa fmuser
Name: fmuser
Namespace: default
Labels:
Annotations:
Image pull secrets:
Mountable secrets: fmuser-token-9qb8n
Tokens: fmuser-token-9qb8n
Events:
root@tianshu-ai-platform:~# kubectl describe secret fmuser-token-9qb8n
Name: fmuser-token-9qb8n
Namespace: default
Labels:
Annotations: kubernetes.io/service-account.name: fmuser
kubernetes.io/service-account.uid: 012db65a-e482-4626-ba01-fc136786c5b1
Type: kubernetes.io/service-account-token
Data
====
ca.crt: 1066 bytes
namespace: 7 bytes
token: eyJhbGciOiJSUzI1NiIsImtpZCI6IjJ5SVZq......
root@tianshu-ai-platform:~# kubectl get secret fmuser-token-9qb8n -o jsonpath='{.data.ca\.crt}'
LS0tLS1CRUdJTiBDRVJ....
root@tianshu-ai-platform:~# kubectl get secret fmuser-token-9qb8n -o jsonpath='{.data.token}'
ZXlKaGJHY2lPaUpTVXpJ.....
root@tianshu-ai-platform:~# kubectl get secret fmuser-token-9qb8n -o jsonpath='{.data.ca\.crt}' | base64 --decode
-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
root@tianshu-ai-platform:~# kubectl get secret fmuser-token-9qb8n -o jsonpath='{.data.token}' | base64 --decode
eyJhbGciOiJSUzI1NiIsImtpZCI6IjJ5SVZqUjBOLVdwbUluU2F5M....
在使用中的ca.crt和token都是base64解码过的,存储在secret中的都是经过编码的;