一、介绍
本案例基于Kubernetes和Docker，其中包括
1、web前端
2、redis master
3、redis slave
其中web前端通过javascript redis api和redis master交互
kubernetes体系架构
二、配置
0、先决条件
Kubernetes 集群
1、启动redis master
使用replication controller确保只有一个pod在运行（当某个节点down了，rc会在另一个健康的node启动redis master），但可能会有数据丢失。
[root@centos1 example]# kubectl create -f redis-master-controller.json replicationcontrollers/redis-master[root@centos1 example]# kubectl get rcCONTROLLER CONTAINER(S) IMAGE(S) SELECTOR REPLICASredis-master master redis name=redis-master 1
验证master运行成功，如下展示了pod运行在centos2/192.168.1.112这台机器上
[root@centos1 example]# kubectl get podsPOD IP CONTAINER(S) IMAGE(S) HOST LABELS STATUS CREATED MESSAGEredis-master-svar7 172.17.0.9 centos2/192.168.1.112 name=redis-master Running 55 seconds master redis Running 55 seconds
SSH到centos2/192.168.1.112查看docker状态
[root@centos2 yum.repos.d]# docker psCONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES91689ce56668 redis:latest "/entrypoint.sh redi 3 minutes ago Up 3 minutes k8s_master.52732b08_redis-master-svar7_default_5b6d5485-1894-11e5-b3ad-000c293c8c19_97e79b7b 38c3180813c3 gcr.io/google_containers/pause:0.8.0 "/pause" 3 minutes ago Up 3 minutes k8s_POD.49eee8c2_redis-master-svar7_default_5b6d5485-1894-11e5-b3ad-000c293c8c19_298e038f
注意：kubectl create执行后，如果镜像不存在，会执行docker pull，根据网络情况，下载中的pods 在kubertnetes UI上会显示pending状态
2、启动master service
一个kubernetes service会对一个或多个container进行负载均衡，这是通过我们上面redis-master中定义的labels元数据实现的，值得注意的是，在redis中只有一个master，但是我们依然为它创建一个service，这是因为这样我们就能使用一个elastic IP来路由到具体某一个master。
kubernetes集群中的service是通过container中的环境变量实现服务发现的，service基于pod label实现container的负载均衡。
在第一步中创建的pod包含了一个label“name=redis-master”，service的selector字段决定了service将流量转发给哪个pod，port和targetPort信息定义了service proxy运行在什么端口。
[root@centos1 example]# kubectl create -f redis-master-service.json services/redis-master[root@centos1 example]# kubectl get servicesNAME LABELS SELECTOR IP(S) PORT(S)redis-master name=redis-master name=redis-master 10.254.154.90 6379/TCP
上面的运行成功后，所有pods都能发现redis master运行在6379端口，从salve到master流量走向会有以下两步：
1、一个redis slave会连接到redis master service的port上
2、流量会从service节点上的port到targetPort
如果targetPort未指定，默认和port一致
3、启动replicated slave pod
虽然redis master是一个单独的pod，redis slaves是一个replicated pod，在Kubernetes中，一个Replication Controller负责管理一个replicated pod的多个实例，RC会自动拉起down掉的replica（可以通过杀死docker 进程方式简单测试）
[root@centos1 example]# kubectl create -f redis-slave-controller.json replicationcontrollers/redis-slave[root@centos1 example]# kubectl get rcCONTROLLER CONTAINER(S) IMAGE(S) SELECTOR REPLICASredis-master master redis name=redis-master 1redis-slave slave kubernetes/redis-slave:v2 name=redis-slave 2[root@centos1 example]# kubectl get podsPOD IP CONTAINER(S) IMAGE(S) HOST LABELS STATUS CREATED MESSAGEredis-master-svar7 172.17.0.9 centos2/192.168.1.112 name=redis-master Running 41 minutes master redis Running 41 minutes redis-slave-31tkb 172.17.0.10 centos2/192.168.1.112 name=redis-slave Running 29 seconds slave kubernetes/redis-slave:v2 Running 28 seconds redis-slave-uk8nu 172.17.0.11 centos2/192.168.1.112 name=redis-slave Running 29 seconds slave kubernetes/redis-slave:v2 Running 28 seconds
可以看到一个master pod和两个slave pod
4、启动slave service
和master一样，我们希望有一个代理服务连接到redis slave，除了服务发现之外，slave service还为web app client提供了透明代理。
这次service 的selector是name=redis-slave，我们可以方便的使用kubectl get services -l "label=value"命令来定位这些服务
[root@centos1 example]# kubectl create -f redis-slave-service.json services/redis-slave[root@centos1 example]# kubectl get servicesNAME LABELS SELECTOR IP(S) PORT(S)redis-master name=redis-master name=redis-master 10.254.154.90 6379/TCPredis-slave name=redis-slave name=redis-slave 10.254.159.145 6379/TCP
5、创建frontend pod
这是一个简单的PHP 服务，用来和master service（写请求）或slave service（读请求）交互
[root@centos1 example]# kubectl create -f frontend-controller.json replicationcontrollers/frontend[root@centos1 example]# kubectl get rcCONTROLLER CONTAINER(S) IMAGE(S) SELECTOR REPLICASfrontend php-redis kubernetes/example-guestbook-php-redis:v2 name=frontend 3redis-master master redis name=redis-master 1redis-slave slave kubernetes/redis-slave:v2 name=redis-slave 2
运行成功后，查看当前pod运行状态
[root@centos1 example]# kubectl get podsPOD IP CONTAINER(S) IMAGE(S) HOST LABELS STATUS CREATED MESSAGEfrontend-fr5z1 172.17.0.13 centos2/192.168.1.112 name=frontend Running 2 minutes php-redis kubernetes/example-guestbook-php-redis:v2 Running 2 minutes frontend-gjx3t 172.17.0.14 centos2/192.168.1.112 name=frontend Running 2 minutes php-redis kubernetes/example-guestbook-php-redis:v2 Running 2 minutes frontend-v608r 172.17.0.12 centos2/192.168.1.112 name=frontend Running 2 minutes php-redis kubernetes/example-guestbook-php-redis:v2 Running 2 minutes redis-master-svar7 172.17.0.9 centos2/192.168.1.112 name=redis-master Running 53 minutes master redis Running 53 minutes redis-slave-31tkb 172.17.0.10 centos2/192.168.1.112 name=redis-slave Running 12 minutes slave kubernetes/redis-slave:v2 Running 12 minutes redis-slave-uk8nu 172.17.0.11 centos2/192.168.1.112 name=redis-slave Running 12 minutes slave kubernetes/redis-slave:v2 Running 12 minutes
可以看到一个redis master，两个redis slave和三个frontend pods
6、创建guestbook service
和其他service一样，你可以创建一个service管理frontend pods
[root@centos1 example]# kubectl create -f frontend-service.json services/frontend[root@centos1 example]# kubectl get servicesNAME LABELS SELECTOR IP(S) PORT(S)frontend name=frontend name=frontend 10.254.154.111 80/TCPredis-master name=redis-master name=redis-master 10.254.154.90 6379/TCPredis-slave name=redis-slave name=redis-slave 10.254.159.145 6379/TCP
我们可以通过frontend service（10.254.154.111）访问pods，但是这个IP明显是无法在外部访问的，下一节讲解如何在外部网络访问guestbook
7、外部网络访问guestbook
kubernetes 支持两种访问暴露一个服务到外部IP地址NodePorts 和LoadBalancers，https://github.com/GoogleCloudPlatform/kubernetes/blob/master/docs/services.md#external-services
另外可以查看防火墙，找到service对应的端口，如下
[root@centos1 example]# kubectl get pods,servicesPOD IP CONTAINER(S) IMAGE(S) HOST LABELS STATUS CREATED MESSAGEfrontend-fr5z1 172.17.0.13 centos2/192.168.1.112 name=frontend Running 22 minutes php-redis kubernetes/example-guestbook-php-redis:v2 Running 22 minutes frontend-gjx3t 172.17.0.14 centos2/192.168.1.112 name=frontend Running 22 minutes php-redis kubernetes/example-guestbook-php-redis:v2 Running 22 minutes frontend-v608r 172.17.0.12 centos2/192.168.1.112 name=frontend Running 22 minutes php-redis kubernetes/example-guestbook-php-redis:v2 Running 22 minutes redis-master-svar7 172.17.0.9 centos2/192.168.1.112 name=redis-master Running About an hour master redis Running About an hour redis-slave-31tkb 172.17.0.10 centos2/192.168.1.112 name=redis-slave Running 32 minutes slave kubernetes/redis-slave:v2 Running 32 minutes redis-slave-uk8nu 172.17.0.11 centos2/192.168.1.112 name=redis-slave Running 32 minutes slave kubernetes/redis-slave:v2 Running 32 minutes NAME LABELS SELECTOR IP(S) PORT(S)frontend name=frontend name=frontend 10.254.154.111 80/TCPkubernetes component=apiserver,provider=kubernetes <none> 10.254.0.2 443/TCPkubernetes-ro component=apiserver,provider=kubernetes <none> 10.254.0.1 80/TCPredis-master name=redis-master name=redis-master 10.254.154.90 6379/TCPredis-slave name=redis-slave name=redis-slave 10.254.159.145 6379/TCP[root@centos1 example]# kubectl get servicesNAME LABELS SELECTOR IP(S) PORT(S)frontend name=frontend name=frontend 10.254.154.111 80/TCPkubernetes component=apiserver,provider=kubernetes <none> 10.254.0.2 443/TCPkubernetes-ro component=apiserver,provider=kubernetes <none> 10.254.0.1 80/TCPredis-master name=redis-master name=redis-master 10.254.154.90 6379/TCPredis-slave name=redis-slave name=redis-slave 10.254.159.145 6379/TCP
发现redis-master是在10.254.154.90上，登录到centos2上，执行iptables-save，发现其中有这样一条规则
-A KUBE-PORTALS-HOST -d 10.254.154.90/32 -p tcp -m comment --comment "default/redis-master:" -m tcp --dport 6379 -j DNAT --to-destination 192.168.1.112:49038
说明，centos2本机的49038端口映射到master container内的6379了，当然我们就能执行在本机访问redis了
[root@centos2 yum.repos.d]# redis-cli -p 49038127.0.0.1:49038> set a bOK127.0.0.1:49038> get a"b"
8、使用curl简单测试
提交数据
curl "localhost:8000/index.php?cmd=set&key=messages&value=jay_sais_hi"
查询数据
curl "localhost:8000/index.php?cmd=get&key=messages"
附本案例用到的6个.json文件
1、redis-master-controller.json
{ "kind":"ReplicationController", "apiVersion":"v1beta3", "metadata":{ "name":"redis-master", "labels":{ "name":"redis-master" } }, "spec":{ "replicas":1, "selector":{ "name":"redis-master" }, "template":{ "metadata":{ "labels":{ "name":"redis-master" } }, "spec":{ "containers":[ { "name":"master", "image":"redis", "ports":[ { "containerPort":6379 } ] } ] } } }}
2、redis-master-service.json
{ "kind":"Service", "apiVersion":"v1beta3", "metadata":{ "name":"redis-master", "labels":{ "name":"redis-master" } }, "spec":{ "ports": [ { "port":6379, "targetPort":6379 } ], "selector":{ "name":"redis-master" } }}
3、redis-slave-controller.json
{ "kind":"ReplicationController", "apiVersion":"v1beta3", "metadata":{ "name":"redis-slave", "labels":{ "name":"redis-slave" } }, "spec":{ "replicas":2, "selector":{ "name":"redis-slave" }, "template":{ "metadata":{ "labels":{ "name":"redis-slave" } }, "spec":{ "containers":[ { "name":"slave", "image":"kubernetes/redis-slave:v2", "ports":[ { "containerPort":6379 } ] } ] } } }}
4、redis-slave-service.json
{ "kind":"Service", "apiVersion":"v1beta3", "metadata":{ "name":"redis-slave", "labels":{ "name":"redis-slave" } }, "spec":{ "ports": [ { "port":6379 } ], "selector":{ "name":"redis-slave" } }}
5、frontend-controller.json
{ "kind":"ReplicationController", "apiVersion":"v1beta3", "metadata":{ "name":"frontend", "labels":{ "name":"frontend" } }, "spec":{ "replicas":3, "selector":{ "name":"frontend" }, "template":{ "metadata":{ "labels":{ "name":"frontend" } }, "spec":{ "containers":[ { "name":"php-redis", "image":"kubernetes/example-guestbook-php-redis:v2", "ports":[ { "containerPort":80 } ] } ] } } }}
6、frontend-service.json
{ "kind":"Service", "apiVersion":"v1beta3", "metadata":{ "name":"frontend", "labels":{ "name":"frontend" } }, "spec":{ "ports": [ { "port":80 } ], "selector":{ "name":"frontend" } }}
本文转自 msj0905 51CTO博客，原文链接:http://blog.51cto.com/sky66/1939627
