k8s部署http服务

1.构建镜像 http.server, 并且把镜像推送到docker hub上, 我推上去的chengfengsunce/httpserver:0.0.1

 

 

2.通过deployment创建pod资源, deployment.yaml如下：

 

 

{
  "apiVersion": "apps/v1",
  "kind": "Deployment",
  "metadata": {
    "labels": {
      "app": "httpserver"
    },
    "name": "httpserver"
  },
  "spec": {
    "progressDeadlineSeconds": 600,
    "replicas": 3,
    "revisionHistoryLimit": 10,
    "selector": {
      "matchLabels": {
        "app": "httpserver"
      }
    },
    "strategy": {
      "rollingUpdate": {
        "maxSurge": "25%",
        "maxUnavailable": "25%"
      },
      "type": "RollingUpdate"
    },
    "template": {
      "metadata": {
        "creationTimestamp": null,
        "labels": {
          "app": "httpserver"
        }
      },
      "spec": {
        "containers": [
          {
            "image": "chengfengsunce/httpserver:0.0.1",
            "imagePullPolicy": "IfNotPresent",
            "livenessProbe": {
              "failureThreshold": 3,
              "httpGet": {
                "path": "/healthz",
                "port": 8080,
                "scheme": "HTTP"
              },
              "initialDelaySeconds": 5,
              "periodSeconds": 10,
              "successThreshold": 1,
              "timeoutSeconds": 1
            },
            "name": "httpserver",
            "readinessProbe": {
              "failureThreshold": 3,
              "httpGet": {
                "path": "/healthz",
                "port": 8080,
                "scheme": "HTTP"
              },
              "initialDelaySeconds": 5,
              "periodSeconds": 10,
              "successThreshold": 1,
              "timeoutSeconds": 1
            },
            "resources": {
              "limits": {
                "cpu": "200m",
                "memory": "100Mi"
              },
              "requests": {
                "cpu": "20m",
                "memory": "20Mi"
              }
            },
            "terminationMessagePath": "/dev/termination-log",
            "terminationMessagePolicy": "File"
          }
        ],
        "dnsPolicy": "ClusterFirst",
        "imagePullSecrets": [
          {
            "name": "cloudnative"
          }
        ],
        "restartPolicy": "Always",
        "schedulerName": "default-scheduler",
        "securityContext": {},
        "terminationGracePeriodSeconds": 30
      }
    }
  }
}

用上面的yaml把文件部署上去，执行命令：kubectl create -f deployment.yaml

执行命令 kubectl create -f 文件名称.yaml，通过查看命令，看看deployment信息， 

root@iZbp12lq02mc4cz0rt0ce9Z:/homework# k get deployment

 

3.部署service代理

 1 apiVersion: v1
 2 kind: Service # 资源类型
 3 metadata:
 4   name: product-server # 名称，这个名称类似dockercompose里面定义的服务名
 5 spec:
 6   ports:
 7   - name: psvc
 8     port: 80 # 服务端口，提供给集群内部访问的端口，外部访问不了
 9     targetPort: 80 #  容器端口
10   - name: grpc
11     port: 81
12     targetPort: 81
13   selector:
14     app: product-server  # 标签，与之对应的是deployment里面的pod标签，它们是多对多的关系
15   type: ClusterIP # 内部网络

执行命令 kubectl create -f 文件名称.yaml，可以查看service信息。

1 root@kubernetes-master:/usr/local/k8s-test01/product# kubectl get service
2 NAME             TYPE           CLUSTER-IP      EXTERNAL-IP   PORT(S)                          AGE
3 kubernetes       ClusterIP      10.96.0.1       <none>        443/TCP                          3d15h
4 mssql            LoadBalancer   10.104.37.195   <pending>     1433:31227/TCP                   26h
5 product-server   ClusterIP      10.107.6.57     <none>        80/TCP,81/TCP                    6h10m
6 rabmq            NodePort       10.102.48.159   <none>        15672:31567/TCP,5672:30567/TCP   13h
7 root@kubernetes-master:/usr/local/k8s-test01/product#

 

deployment和service配置文件可以放在一个yaml文件里面，通过---分开就行，这里分开是为了看起来有层次。

好了，productservice部署完毕了，我们的服务是部署完毕了，但是我们在外部访问不了，需要入口网关，这里我们先使用ingress-nginx-controller。

部署ingress

1.安装ingress-nginx，这里我用的是最新版0.30.0，很是郁闷，下载地址被墙了，最后在github开源代码里面找到安装资源，其实就是一份yaml文件。

 1 # 其他...
 2 apiVersion: apps/v1
 3 kind: Deployment
 4 metadata:
 5 name: nginx-ingress-controller
 6 namespace: ingress-nginx
 7 labels:
 8 app.kubernetes.io/name: ingress-nginx
 9 app.kubernetes.io/part-of: ingress-nginx
10 spec:
11 replicas: 1
12 selector:
13 matchLabels:
14 app.kubernetes.io/name: ingress-nginx
15 app.kubernetes.io/part-of: ingress-nginx
16 template:
17 metadata:
18 labels:
19 app.kubernetes.io/name: ingress-nginx
20 app.kubernetes.io/part-of: ingress-nginx
21 annotations:
22 prometheus.io/port: "10254"
23 prometheus.io/scrape: "true"
24 spec:
25 # wait up to five minutes for the drain of connections
26 terminationGracePeriodSeconds: 300
27 serviceAccountName: nginx-ingress-serviceaccount
28 nodeSelector:
29 kubernetes.io/os: linux
30 containers:
31 - name: nginx-ingress-controller
32 image: quay.io/kubernetes-ingress-controller/nginx-ingress-controller:0.30.0
33 args:
34 - /nginx-ingress-controller
35 - --configmap=$(POD_NAMESPACE)/nginx-configuration
36 - --tcp-services-configmap=$(POD_NAMESPACE)/tcp-services
37 - --udp-services-configmap=$(POD_NAMESPACE)/udp-services
38 - --publish-service=$(POD_NAMESPACE)/ingress-nginx
39 - --annotations-prefix=nginx.ingress.kubernetes.io
40 # ......

安装ingress-nginx ，执行命令 kubectl create -f ingress-nginx.yaml，随后我们通过命令查看。

1 root@kubernetes-master:/usr/local/k8s-test01/product# kubectl get pods -n ingress-nginx
2 NAME                                        READY   STATUS    RESTARTS   AGE
3 nginx-ingress-controller-77db54fc46-tx6pf   1/1     Running   5          40h
4 root@kubernetes-master:/usr/local/k8s-test01/product#

接下来配置并发布ingress-nginx网关服务。

 1 apiVersion: networking.k8s.io/v1beta1
 2 kind: Ingress
 3 metadata:
 4   name: nginx-web
 5   annotations: # 扩展信息，这里可以配置鉴权、认证信息
 6     nginx.ingress.kubernetes.io/rewrite-target: /
 7 spec:
 8   # 路由规则
 9   rules:
10   # 主机名，只能是域名，需要在宿主机配置hosts映射
11   - host: product.com
12     http:
13       paths:
14       - path: /
15         backend:
16           # 后台部署的 Service Name，与上面部署的service对应
17           serviceName: product-server
18           # 后台部署的 Service Port，与上面部署的service对应
19           servicePort: 80
20       - path: /grpc
21         backend:
22           # 后台部署的 Service Name，与上面部署的service对应
23           serviceName: product-server
24           # 后台部署的 Service Port，与上面部署的service对应
25           servicePort: 81

 

执行kubectl create -f productservice-ingress.yaml部署。接下来我们查看网关服务信息。

1 root@kubernetes-master:/usr/local/k8s-test01/product# kubectl get ingress
2 NAME        CLASS    HOSTS         ADDRESS   PORTS   AGE
3 nginx-web   <none>   product.com             80      8h

 

mssql和rabbitmq的部署这里就贴代码了，接下来我们访问product.com这个域名，看看是否部署成功。

 

 

我们测试一个get接口试试，看看能不能通，看图。

 

服务部署就到这，接下来我们简单总结一下

1.这是测试环境所以master没做高可用，ingress也没做高可用，有时间再做顺便补充一下；

2.外部网络请求到ingress-nginx域名，线上环境这个域名肯定是公网地址，ingress做认证鉴权，合法的请求通过path路由到对应的后台service，如果其中一台ingress挂掉了，keepalived会把vip游离到其他slave节点，这样就完成了ingress的高可用；

3.service代理会把请求随机转发到标签匹配的pod里面的容器处理，如果其中一台node挂了或者pod异常退出了（也就是返回值不等于0），deployment会重新启动一个pod，我们下面做个实验，删掉其中一个pod，看看效果怎么样。

1 root@kubernetes-master:/usr/local/k8s-test01/product# kubectl get pods
2 NAME                              READY   STATUS    RESTARTS   AGE
3 mssql-59bd4dc6df-xzxc2            1/1     Running   5          27h
4 product-server-599cfd85cc-2q7zx   1/1     Running   0          6s
5 product-server-599cfd85cc-ppmhx   1/1     Running   0          6h51m
6 rabmq-7c9748f876-9msjg            1/1     Running   0          14h
7 rabmq-7c9748f876-lggh6            1/1     Running   0          14h

 

我们先通过命令显示所有default命名空间下面的所有pod，然后delete一个pod看看，它会不会重新启动。执行删除命令

1 kubectl delete pods product-server-599cfd85cc-ppmhx

接着马上查看pods信息，要快知道吗？

1 root@kubernetes-master:/usr/local/k8s-test01/product# kubectl get pods
2 NAME                              READY   STATUS        RESTARTS   AGE
3 mssql-59bd4dc6df-xzxc2            1/1     Running       5          27h
4 product-server-599cfd85cc-2q7zx   1/1     Running       0          2m8s
5 product-server-599cfd85cc-9s497   1/1     Running       0          13s
6 product-server-599cfd85cc-ppmhx   0/1     Terminating   0          6h53m
7 rabmq-7c9748f876-9msjg            1/1     Running       0          14h
8 rabmq-7c9748f876-lggh6            1/1     Running       0          14h

 

看到没？ppmhx这个pod正在终止，马上就创建了新的pod。