关于kubernetes:07kubernetes笔记-Service一-ClusterIPNodePortLoadBalancer

Service简介

Service:能够了解为pod的负债均衡器,规范资源类型,Service Controller 为动静的一组Pod提供一个固定的拜访入口， kubernetes实现SVC工作的是组件是kube-proxy
Endpoint Controller:治理后端端点与svc的绑定,依据标签选择器,筛选适配的pod,监控就绪的pod 并实现svc与pod的绑定
工作流程:Service Controller---->创立雷同标签选择器的 Endpoint Controller依据标签选择器去治理和监听后端Pod状态实现Svc与Pod绑定

Service可能提供负载平衡的能力，然而在应用上有以下限度:

只提供4层负载平衡能力，而没有7层性能，但有时咱们可能须要更多的匹配规定来转发申请，这在4层负载平衡上是不反对的

kube-proxy3种不同的数据调度模式

1.Userspace
Userspace模型:Pod-->Service, iptables拦挡规定,但本人不做调度工作流程: 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(内核实现数据调度)-->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规定
毛病:iptables模式，一个service会生成大量的规定; 如果一个service有50条规定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长处的状况下,同时改良了iptables产生大量规定的毛病,在大规模集群中serice多的状况下劣势更显著,

Service的类型

clusterIP:通过集群外部IP地址裸露服务，但该地址仅在集群外部可见、可达，它无奈被集群内部的客户端拜访;默认类型;倡议由K8S动静指定一个;也反对用户手动明确指定;
NodePort: NodePort是ClusterIP的加强类型，它会于ClusterIP的性能之外，在每个节点上应用一个雷同的端口号将内部流量引入到该Service上来。
LoadBalancer: 是NodePort的加强类型，为各节点上的NodePort提供一个内部负载均衡器;须要私有云反对
ExternalName:内部流程引入到K8S外部,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解析为集群内部的服务的TP地址,实现外部服务与内部服务的数据交互 ExternallP 能够与ClusterIP、NodePort一起应用应用其中一个IP做进口IP

ServicePort

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裸露的话，每个都要独自定义。最终接管申请的是PodIP和ContainerPort;

Service资源标准

Service名称空间级别的资源不能跨名称空间

apiVersion: v1kind: Servicemetadata:  name: ..  namespace: ...  labels:    key1: value1    key2: value2spec:  type <string>  #Service类型，默认为ClusterIP  selector <map[string]string> #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 <string>#端口名称    protocol <string> #协定，目前仅反对TCP、UDP和SCTP，默认为TCP    port <integer> # Service的端口号    targetPort <string> #后端指标过程的端口号或名称，名称需由Pod标准定义    nodePort <integer> # 节点端口号，仅实用于NodePort和LoadBalancer类型  clusterIP <string> # Service的集群IP，倡议由零碎主动调配  externalTrafficPolicy <string>#内部流量策略解决形式，Local示意由以后节点解决，#Cluster示意向集群范畴调度  loadBalancerIP <string> #内部负载均衡器应用的IP地址，仅实用于LoadBlancer  externalName <string>  # 内部服务名称，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: v1kind: Servicemetadata:  name: demoapp-svc  namespace: defaultspec:   clusterIP: 10.97.72.1   #正式部署不须要指定 会主动生成,手动指定还可能会导致抵触  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 wideNAME          TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)        AGE   SELECTORdemoapp-svc   ClusterIP   10.97.72.1       <none>        80/TCP         11s   app=demoappkubernetes    ClusterIP   10.96.0.1        <none>        443/TCP        30d   <none>my-grafana    NodePort    10.96.4.185      <none>        80:30379/TCP   27d   app.kubernetes.io/instance=my-grafana,app.kubernetes.io/name=grafanamyapp         NodePort    10.106.116.205   <none>        80:31532/TCP   30d   app=myapp,release=stabel[root@k8s-master svc]# curl 10.97.72.1  #通过拜访svc IP拜访到后端节点iKubernetes demoapp v1.0 !! ClientIP: 10.244.0.0, ServerName: demoapp-66db74fcfc-9wkgj, ServerIP: 10.244.2.97![root@k8s-master svc]# curl 10.97.72.1iKubernetes 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-svcName:              demoapp-svcNamespace:         defaultLabels:            <none>Annotations:       <none>Selector:          app=demoappType:              ClusterIPIP:                10.97.72.1Port:              http  80/TCPTargetPort:        80/TCPEndpoints:         10.244.1.98:80,10.244.2.97:80   #后端节点Session Affinity:  NoneEvents:            <none>[root@k8s-master svc]# kubectl get pod -o wide --show-labels   #匹配到前1、2个NAME                          READY   STATUS    RESTARTS   AGE   IP             NODE        NOMINATED NODE   READINESS GATES   LABELSdemoapp-66db74fcfc-9wkgj      1/1     Running   0          39m   10.244.2.97    k8s-node2   <none>           <none>            app=demoapp,pod-template-hash=66db74fcfc,release=stabledemoapp-66db74fcfc-vzb4f      1/1     Running   0          39m   10.244.1.98    k8s-node1   <none>           <none>            app=demoapp,pod-template-hash=66db74fcfc,release=stable,track=dailyliveness-httpget-demo         1/1     Running   3          29m   10.244.1.99    k8s-node1   <none>           <none>            app=livenessliveness-tcpsocket-demo       1/1     Running   3          29m   10.244.1.100   k8s-node1   <none>           <none>            <none>my-grafana-7d788c5479-kpq9q   1/1     Running   4          27d   10.244.1.84    k8s-node1   <none>           <none>            app.kubernetes.io/instance=my-grafana,app.kubernetes.io/name=grafana,pod-template-hash=7d788c5479[root@k8s-master svc]# kubectl get ep  #理论治理后端端点与svc的绑定是EndpointsNAME          ENDPOINTS                       AGEdemoapp-svc   10.244.1.98:80,10.244.2.97:80   2m33skubernetes    192.168.4.170:6443              30dmy-grafana    10.244.1.84:3000                27dmyapp         <none>                          30d[root@k8s-master svc]#  kubectl scale deployment demoapp  --replicas=4  #批改deployment正本数为4deployment.apps/demoapp scaled[root@k8s-master svc]# kubectl get pod --show-labelsNAME                          READY   STATUS    RESTARTS   AGE    LABELSdemoapp-66db74fcfc-9jzs5      1/1     Running   0          18s    app=demoapp,pod-template-hash=66db74fcfc,release=stabledemoapp-66db74fcfc-9wkgj      1/1     Running   0          100m   app=demoapp,pod-template-hash=66db74fcfc,release=stabledemoapp-66db74fcfc-dw9w2      1/1     Running   0          18s    app=demoapp,pod-template-hash=66db74fcfc,release=stabledemoapp-66db74fcfc-vzb4f      1/1     Running   0          100m   app=demoapp,pod-template-hash=66db74fcfc,release=stable,track=dailyliveness-httpget-demo         1/1     Running   3          90m    app=livenessliveness-tcpsocket-demo       1/1     Running   3          90m    <none>my-grafana-7d788c5479-kpq9q   1/1     Running   4          27d    app.kubernetes.io/instance=my-grafana,app.kubernetes.io/name=grafana,pod-template-hash=7d788c5479[root@k8s-master svc]# kubectl get ep  #已实时增加到ep与svc绑定NAME          ENDPOINTS                                                   AGEdemoapp-svc   10.244.1.101:80,10.244.1.98:80,10.244.2.97:80 + 1 more...   63mkubernetes    192.168.4.170:6443                                          30dmy-grafana    10.244.1.84:3000                                            27dmyapp         <none>                                                      30d

示例2: NodePort 演示

[root@k8s-master svc]# cat services-nodeport-demo.yaml apiVersion: v1kind: Servicemetadata:  name: demoapp-nodeport-svc  namespace: defaultspec:   type: NodePort  clusterIP: 10.97.56.1   #正式部署不须要指定 会主动生成手动指定还可能会导致抵触  selector:    app: demoapp  ports:  - name: http    protocol: TCP    port: 80    targetPort: 80    #后端pod端口    nodePort: 31399  #正式部署不须要指定 会主动生成   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 podNAME                          READY   STATUS    RESTARTS   AGEdemoapp-66db74fcfc-9jzs5      1/1     Running   0          8m47sdemoapp-66db74fcfc-9wkgj      1/1     Running   0          109mdemoapp-66db74fcfc-dw9w2      1/1     Running   0          8m47sdemoapp-66db74fcfc-vzb4f      1/1     Running   0          109mliveness-httpget-demo         1/1     Running   3          98mliveness-tcpsocket-demo       1/1     Running   3          98mmy-grafana-7d788c5479-kpq9q   1/1     Running   4          27d[root@k8s-master svc]# kubectl get svcNAME                   TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)        AGEdemoapp-nodeport-svc   NodePort    10.97.56.1       <none>        80:31399/TCP   11s    #能够看到两个prot 其中31399就是nodeport端口demoapp-svc            ClusterIP   10.97.72.1       <none>        80/TCP         72m[root@k8s-master svc]# while true;do curl 192.168.4.171:31399;sleep 1;done  #通过节点IP:prot拜访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!

能够看到下面尽管是通过节点2拜访,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 <string>' #内部流量策略解决形式，
Local示意由以后节点解决
Cluster示意向集群范畴调度

[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正本数为1deployment.apps/demoapp scaled[root@k8s-master ~]# kubectl get pod -o wide  #能够看到惟一的pod运行node2节点上NAME                          READY   STATUS    RESTARTS   AGE    IP             NODE        NOMINATED NODE   READINESS GATESdemoapp-66db74fcfc-9wkgj      1/1     Running   0          123m   10.244.2.97    k8s-node2   <none>           <none>liveness-httpget-demo         1/1     Running   3          112m   10.244.1.99    k8s-node1   <none>           <none>[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2iKubernetes 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: v1kind: Servicemetadata:  name: demoapp-loadbalancer-svc  namespace: defaultspec:   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,无奈创立ELB,所以无奈创立[root@k8s-master svc]# kubectl get svcNAME                       TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)        AGEdemoapp-loadbalancer-svc   LoadBalancer   10.110.155.70    <pending>     80:31619/TCP   31s    #能够看到因为不是Iaas平台上 EXTERNAL-IP始终为pending状态,示意始终在申请资源而挂起,仍然能够通过NodePort的形式拜访demoapp-nodeport-svc       NodePort       10.97.56.1       <none>        80:31399/TCP   30mdemoapp-svc                ClusterIP      10.97.72.1       <none>        80/TCP         102m[root@k8s-master svc]# while true;do curl 192.168.4.171:31399;sleep 1;done  #通过NodePort的形式拜访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 理论利用中还须要在后面加一层负载平衡，以起到对立入口和高可用，而且后端新增的节点也不会主动增加到负载上
externalIPs 在只有1个或多个节点裸露IP的状况下，可通过虚构IP，实现高可用

[root@k8s-master ~]# ip addr add 192.168.100.100/16 dev eth0[root@k8s-master ~]# ip addr show1: lo: <LOOPBACK,UP,LOWER_UP> 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 forever2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> 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.yaml     services-externalip-demo.yaml    services-loadbalancer-demo.yaml  services-nodeport-demo.yaml[root@k8s-master svc]# cat services-externalip-demo.yaml apiVersion: v1kind: Servicemetadata:  name: demoapp-externalip-svc  namespace: defaultspec:   type: ClusterIP  selector:    app: demoapp  ports:  - name: http    protocol: TCP    port: 80    targetPort: 80    #后端pod端口  externalIPs:  - 192.168.100.100  #理论利用中，能够通过过haproxy等实现虚构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 svcNAME                       TYPE           CLUSTER-IP       EXTERNAL-IP       PORT(S)        AGEdemoapp-externalip-svc     ClusterIP      10.110.30.133    192.168.100.100   80/TCP         16sdemoapp-loadbalancer-svc   LoadBalancer   10.110.155.70    <pending>         80:31619/TCP   3h6mdemoapp-nodeport-svc       NodePort       10.97.56.1       <none>            80:31399/TCP   3h36mdemoapp-svc                ClusterIP      10.97.72.1       <none>            80/TCP         4h47m#拜访测试[root@k8s-master svc]# curl 192.168.100.100iKubernetes 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;doneiKubernetes 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!