07.kubernetes笔记 Service(一) ClusterIP、NodePort、LoadBalancer kubernetes容器运维

Service简介

Service:可以理解为pod的负债均衡器,标准资源类型,Service Controller 为动态的一组Pod提供一个固定的访问入口， kubernetes完成SVC工作的是组件是kube-proxy
Endpoint Controller:管理后端端点与svc的绑定,根据标签选择器,筛选适配的pod,监控就绪的pod 并完成svc与pod的绑定
工作流程:Service Controller---->创建相同标签选择器的 Endpoint Controller根据标签选择器去管理和监听后端Pod状态完成Svc与Pod绑定

Service能够提供负载均衡的能力，但是在使用上有以下限制:

只提供4层负载均衡能力，而没有7层功能，但有时我们可能需要更多的匹配规则来转发请求，这在4层负载均衡上是不支持的

kube-proxy3种不同的数据调度模式

1.Userspace
Userspace模型:Pod-->Service, iptables拦截规则,但自己不做调度工作流程: 用户空间-->ptables(内核)-->kube-proxy-->ptables(内核)-->再调度给用户空间效率低
iptables 用户空间-->ptables(内核完成数据调度)-->调度给用户空间效率高
在iptables模型下kube-proxy的作用不在是数据调度转发,而是监听API server所有service中的定义转为本地的iptables规则
缺点:iptables模式，一个service会生成大量的规则; 如果一个service有50条规则那如果有一万个容器,内核的性能就会受到影响
ipvs代理模式: 在继承iptables优点的情况下,同时改进了iptables产生大量规则的缺点,在大规模集群中serice多的情况下优势更明显,

Service的类型

clusterIP:通过集群内部IP地址暴露服务，但该地址仅在集群内部可见、可达，它无法被集群外部的客户端访问; 默认类型; 建议由K8S动态指定一个; 也支持用户手动明确指定;
NodePort: NodePort是ClusterIP的增强类型，它会于ClusterIP的功能之外，在每个节点上使用一个相同的端口号将外部流量引入到该Service上来。
LoadBalancer: 是NodePort的增强类型，为各节点上的NodePort提供一个外部负载均衡器; 需要公有云支持
ExternalName:外部流程引入到K8S内部,借助集群上KubeDNS来实现，服务的名称会被解析为一个CNAME记录，而CNAME名称会被DNS解析为集群外部的服务的TP地址,实现内部服务与外部服务的数据交互 ExternallP 可以与ClusterIP、NodePort一起使用使用其中一个IP做出口IP

ServicePort 【07.kubernetes笔记 Service(一) ClusterIP、NodePort、LoadBalancer】Service:被映射进Pod上的应用程序监听的端口; 而且如果后端Pod有多个端口，并且每个端口都想通过Service暴露的话，每个都要单独定义。最终接收请求的是PodIP和ContainerPort;
Service资源规范
Service名称空间级别的资源不能跨名称空间

apiVersion: v1 kind: Service metadata: name: .. namespace: ... labels: key1: value1 key2: value2 spec: type#Service类型，默认为ClusterIP selector #等值类型的标签选择器，内含“与"逻辑 ports: # Service的端口对象列表 - name #端口名称 protocol#协议，目前仅支持TCP、UDP和SCTP，默认为TCP port # Service的端口号 targetPort#后端目标进程的端口号或名称，名称需由Pod规范定义 nodePort # 节点端口号，仅适用于NodePort和LoadBalancer类型 clusterIP# Service的集群IP，建议由系统自动分配 externalTrafficPolicy #外部流量策略处理方式，Local表示由当前节点处理，#Cluster表示向集群范围调度 loadBalancerIP#外部负载均衡器使用的IP地址，仅适用于LoadBlancer externalName# 外部服务名称，该名称将作为Service的DNS CNAME值

示例1: ClusterIP 演示

[root@k8s-master svc]# cat services-clusterip-demo.yaml apiVersion: v1 kind: Service metadata: name: demoapp-svc namespace: default spec: clusterIP: 10.97.72.1#正式部署不需要指定会自动生成,手动指定还可能会导致冲突 selector:#定义过滤条件 app: demoapp ports: - name: http protocol: TCP port: 80 targetPort: 80#后端pod端口[root@k8s-master svc]# kubectl apply -f services-clusterip-demo.yaml service/demoapp-svc created[root@k8s-master svc]# kubectl get svc -o wide NAMETYPECLUSTER-IPEXTERNAL-IPPORT(S)AGESELECTOR demoapp-svcClusterIP10.97.72.180/TCP11sapp=demoapp kubernetesClusterIP10.96.0.1443/TCP30d my-grafanaNodePort10.96.4.18580:30379/TCP27dapp.kubernetes.io/instance=my-grafana,app.kubernetes.io/name=grafana myappNodePort10.106.116.20580:31532/TCP30dapp=myapp,release=stabel[root@k8s-master svc]# curl 10.97.72.1#通过访问svc IP访问到后端节点 iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97! [root@k8s-master svc]# curl 10.97.72.1 iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-vzb4f, ServerIP: 10.244.1.98![root@k8s-master svc]# kubectl describe svc demoapp-svc Name:demoapp-svc Namespace:default Labels: Annotations: Selector:app=demoapp Type:ClusterIP IP:10.97.72.1 Port:http80/TCP TargetPort:80/TCP Endpoints:10.244.1.98:80,10.244.2.97:80#后端节点 Session Affinity:None Events:[root@k8s-master svc]# kubectl get pod -o wide --show-labels#匹配到前1、2个 NAMEREADYSTATUSRESTARTSAGEIPNODENOMINATED NODEREADINESS GATESLABELS demoapp-66db74fcfc-9wkgj1/1Running039m10.244.2.97k8s-node2app=demoapp,pod-template-hash=66db74fcfc,release=stable demoapp-66db74fcfc-vzb4f1/1Running039m10.244.1.98k8s-node1app=demoapp,pod-template-hash=66db74fcfc,release=stable,track=daily liveness-httpget-demo1/1Running329m10.244.1.99k8s-node1app=liveness liveness-tcpsocket-demo1/1Running329m10.244.1.100k8s-node1 my-grafana-7d788c5479-kpq9q1/1Running427d10.244.1.84k8s-node1app.kubernetes.io/instance=my-grafana,app.kubernetes.io/name=grafana,pod-template-hash=7d788c5479[root@k8s-master svc]# kubectl get ep#实际管理后端端点与svc的绑定是Endpoints NAMEENDPOINTSAGE demoapp-svc10.244.1.98:80,10.244.2.97:802m33s kubernetes192.168.4.170:644330d my-grafana10.244.1.84:300027d myapp30d[root@k8s-master svc]#kubectl scale deployment demoapp--replicas=4#修改deployment副本数为4 deployment.apps/demoapp scaled[root@k8s-master svc]# kubectl get pod --show-labels NAMEREADYSTATUSRESTARTSAGELABELS demoapp-66db74fcfc-9jzs51/1Running018sapp=demoapp,pod-template-hash=66db74fcfc,release=stable demoapp-66db74fcfc-9wkgj1/1Running0100mapp=demoapp,pod-template-hash=66db74fcfc,release=stable demoapp-66db74fcfc-dw9w21/1Running018sapp=demoapp,pod-template-hash=66db74fcfc,release=stable demoapp-66db74fcfc-vzb4f1/1Running0100mapp=demoapp,pod-template-hash=66db74fcfc,release=stable,track=daily liveness-httpget-demo1/1Running390mapp=liveness liveness-tcpsocket-demo1/1Running390m my-grafana-7d788c5479-kpq9q1/1Running427dapp.kubernetes.io/instance=my-grafana,app.kubernetes.io/name=grafana,pod-template-hash=7d788c5479[root@k8s-master svc]# kubectl get ep#已实时添加到ep与svc绑定 NAMEENDPOINTSAGE demoapp-svc10.244.1.101:80,10.244.1.98:80,10.244.2.97:80 + 1 more...63m kubernetes192.168.4.170:644330d my-grafana10.244.1.84:300027d myapp30d

示例2: NodePort 演示

[root@k8s-master svc]# cat services-nodeport-demo.yaml apiVersion: v1 kind: Service metadata: name: demoapp-nodeport-svc namespace: default spec: type: NodePort clusterIP: 10.97.56.1#正式部署不需要指定会自动生成手动指定还可能会导致冲突 selector: app: demoapp ports: - name: http protocol: TCP port: 80 targetPort: 80#后端pod端口 nodePort: 31399#正式部署不需要指定会自动生成默认生成端口在30000-32768之间[root@k8s-master svc]# kubectl apply -f services-nodeport-demo.yaml service/demoapp-nodeport-svc created [root@k8s-master svc]# kubectl get pod NAMEREADYSTATUSRESTARTSAGE demoapp-66db74fcfc-9jzs51/1Running08m47s demoapp-66db74fcfc-9wkgj1/1Running0109m demoapp-66db74fcfc-dw9w21/1Running08m47s demoapp-66db74fcfc-vzb4f1/1Running0109m liveness-httpget-demo1/1Running398m liveness-tcpsocket-demo1/1Running398m my-grafana-7d788c5479-kpq9q1/1Running427d [root@k8s-master svc]# kubectl get svc NAMETYPECLUSTER-IPEXTERNAL-IPPORT(S)AGE demoapp-nodeport-svcNodePort10.97.56.180:31399/TCP11s#可以看到两个prot 其中31399就是nodeport端口 demoapp-svcClusterIP10.97.72.180/TCP72m[root@k8s-master svc]# while true; do curl 192.168.4.171:31399; sleep 1; done#通过节点IP:prot访问 iKubernetes demoapp v1.0 !! ClientIP: 10.244.2.1, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97! iKubernetes demoapp v1.0 !! ClientIP: 10.244.2.0, ServerName: demoapp-66db74fcfc-dw9w2, ServerIP: 10.244.1.101! iKubernetes demoapp v1.0 !! ClientIP: 10.244.2.0, ServerName: demoapp-66db74fcfc-vzb4f, ServerIP: 10.244.1.98! iKubernetes demoapp v1.0 !! ClientIP: 10.244.2.1, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97!

可以看到上面虽然是通过节点2访问,但通过IP地址发现还是会轮询到节点1上的pod
这时就要提到 'externalTrafficPolicy ' #外部流量策略处理方式，
Local表示由当前节点处理
Cluster表示向集群范围调度

[root@k8s-master ~]# kubectl edit svc demoapp-nodeport-svc ... spec: clusterIP: 10.97.56.1 externalTrafficPolicy: Local#把默认的Cluster改成Local ...[root@k8s-master svc]# kubectl scale deployment demoapp--replicas=1#调整deployment副本数为1 deployment.apps/demoapp scaled[root@k8s-master ~]# kubectl get pod -o wide#可以看到唯一的pod运行node2节点上 NAMEREADYSTATUSRESTARTSAGEIPNODENOMINATED NODEREADINESS GATES demoapp-66db74fcfc-9wkgj1/1Running0123m10.244.2.97k8s-node2 liveness-httpget-demo1/1Running3112m10.244.1.99k8s-node1[root@k8s-master ~]# curl 192.168.4.171:31399#通过节点1失败^C [root@k8s-master ~]# curl 192.168.4.172:31399#通过节点2 iKubernetes demoapp v1.0 !! ClientIP: 192.168.4.170, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97!

示例3: LoadBalancer 演示

[root@k8s-master svc]# cat services-loadbalancer-demo.yaml apiVersion: v1 kind: Service metadata: name: demoapp-loadbalancer-svc namespace: default spec: type: LoadBalancer selector: app: demoapp ports: - name: http protocol: TCP port: 80 targetPort: 80#后端pod端口 #loadBalancerIP: 1.2.3.4#这里应该不是在Iaas平台上,无法创建ELB,所以无法创建[root@k8s-master svc]# kubectl get svc NAMETYPECLUSTER-IPEXTERNAL-IPPORT(S)AGE demoapp-loadbalancer-svcLoadBalancer10.110.155.7080:31619/TCP31s#可以看到因为不是Iaas平台上 EXTERNAL-IP一直为pending状态,表示一直在申请资源而挂起,依然可以通过NodePort的方式访问 demoapp-nodeport-svcNodePort10.97.56.180:31399/TCP30m demoapp-svcClusterIP10.97.72.180/TCP102m[root@k8s-master svc]# while true; do curl 192.168.4.171:31399; sleep 1; done#通过NodePort的方式访问 iKubernetes demoapp v1.0 !! ClientIP: 10.244.1.0, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97! iKubernetes demoapp v1.0 !! ClientIP: 10.244.1.1, ServerName: demoapp-66db74fcfc-2jf49, ServerIP: 10.244.1.103! iKubernetes demoapp v1.0 !! ClientIP: 10.244.1.0, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97! iKubernetes demoapp v1.0 !! ClientIP: 10.244.1.1, ServerName: demoapp-66db74fcfc-5dp5n, ServerIP: 10.244.1.102!

示例4: externalIPs 演示
NodePort 实际应用中还需要在前面加一层负载均衡，以起到统一入口和高可用，而且后端新增的节点也不会自动添加到负载上
externalIPs 在只有1个或多个节点暴露IP的情况下，可通过虚拟IP，实现高可用

[root@k8s-master ~]# ip addr add 192.168.100.100/16 dev eth0 [root@k8s-master ~]# ip addr show 1: lo: mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host valid_lft forever preferred_lft forever 2: eth0: mtu 1500 qdisc pfifo_fast state UP group default qlen 1000 link/ether 52:54:00:44:16:16 brd ff:ff:ff:ff:ff:ff inet 192.168.4.170/24 brd 192.168.4.255 scope global noprefixroute eth0 valid_lft forever preferred_lft forever inet 192.168.100.100/16 scope global eth0 valid_lft forever preferred_lft forever[root@k8s-master svc]# cat services- services-clusterip-demo.yamlservices-externalip-demo.yamlservices-loadbalancer-demo.yamlservices-nodeport-demo.yaml [root@k8s-master svc]# cat services-externalip-demo.yaml apiVersion: v1 kind: Service metadata: name: demoapp-externalip-svc namespace: default spec: type: ClusterIP selector: app: demoapp ports: - name: http protocol: TCP port: 80 targetPort: 80#后端pod端口 externalIPs: - 192.168.100.100#实际应用中，可以通过过haproxy等实现虚拟IP 达到高可用[root@k8s-master svc]# kubectl apply -f services-externalip-demo.yaml service/demoapp-externalip-svc created [root@k8s-master svc]# kubectl get svc NAMETYPECLUSTER-IPEXTERNAL-IPPORT(S)AGE demoapp-externalip-svcClusterIP10.110.30.133192.168.100.10080/TCP16s demoapp-loadbalancer-svcLoadBalancer10.110.155.7080:31619/TCP3h6m demoapp-nodeport-svcNodePort10.97.56.180:31399/TCP3h36m demoapp-svcClusterIP10.97.72.180/TCP4h47m#访问测试 [root@k8s-master svc]# curl 192.168.100.100 iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97! [root@k8s-master svc]# while true; do curl 192.168.100.100; sleep 1; done iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-z682r, ServerIP: 10.244.2.99! iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-5dp5n, ServerIP: 10.244.1.102! iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-5dp5n, ServerIP: 10.244.1.102! iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97! iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-5dp5n, ServerIP: 10.244.1.102!