HPA自动伸缩

安装metric-server

启用HPA功能首先得安装metric-server(kubernetes 1.11 之后已经完全启用 heapster)

附件metrics-server.zip

exec

1 2 3 4 5

unzip metrics-server.zip cd metrics-server kubectl apply -f . kubectl top node

使用kubectl top node 能查看到节点资源使用情况是说明安装成功.

 

编写编排文件

这里以云中心为例

cloud-center-web.yaml

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104

apiVersion: apps/v1 kind: Deployment metadata:   namespace: zxy   name: cloud-web-server spec:   replicas: 1   selector:     matchLabels:       app: cloud-web-server   strategy:     type: "RollingUpdate"     rollingUpdate:       maxSurge: 1       maxUnavailable: 0   template:     metadata:       labels:         app: cloud-web-server     spec:  #     nodeName: 192.168.1.120       containers:         - name: cloud-web-server           resources:             requests:               memory: "1024Mi"               cpu: "100m"             limits:               cpu: "3000m"               memory: "1024Mi"           image: docker.zhixueyun.com:5000/cloud-center-web-server:v201810261009-master           volumeMounts:           - mountPath: /log             name: cloud-center-web-log           - mountPath: /etc/hosts             name: hosts           - mountPath: /tmp             name: cloud-web-datatmp           ports:           - containerPort: 8080           env:           - name: SPRING_APPLICATION_JSON             value: '{"spring.rabbitmq.virtual-host": "/","spring.redis.cluster": "false","spring.redis.cluster.nodes": "redis.zxy.svc.cluster.local:6379","spring.fastdfs.tracker-servers": "192.168.1.120:22122","spring.fastdfs.tracker.http-port": "10402","server.context-path": "/api/v1/cloud-center"}'           - name: dubbo_registry_address             value: 'zookeeper://zookeeper.zxy.svc.cluster.local:2181'           - name: dubbo_application_version             value: '1'           - name: dubbo_protocol_port             value: '8080'           - name: spring_rabbitmq_host             value: 'rabbitmq.zxy.svc.cluster.local'           - name: spring_rabbitmq_port             value: '5672'           - name: spring_rabbitmq_username             value: 'guest'           - name: spring_rabbitmq_password             value: 'guest'           - name: graphite_server             value: 'graphite.zxy.svc.cluster.local'           - name: graphite_port             value: '2003'           - name: server_port             value: '8080'           - name: logging_level_org_jooq             value: 'DEBUG'       volumes:         - hostPath:             path: /app/contaniner_data/v9_5_8/cloud-center-web           name: cloud-center-web-log         - hostPath:             path: /app/contaniner_data/datatmp-cloud-web           name: cloud-web-datatmp         - hostPath:             path: /etc/hosts           name: hosts   --- apiVersion: v1 kind: Service metadata:   namespace: zxy   name: cloud-web-server spec:   type: NodePort   ports:     - port: 8080       nodePort: 31254       name: cloud-web-server   selector:     app: cloud-web-server --- apiVersion: autoscaling/v1 kind: HorizontalPodAutoscaler metadata:   namespace: zxy   name: cloud-web-server-hpa spec:   maxReplicas: 3   minReplicas: 1   scaleTargetRef:     apiVersion: "apps/v1"     kind: Deployment     name: cloud-web-server   targetCPUUtilizationPercentage: 80

说明:

在原先的基础上改变了Controller,并添加滚动更新策略，限制资源使用大小.上述的例子是以CPU为标准来进行伸缩，也可以以mem来进行，只需要根据改变api-versions来做出相应的变动。

example

1 2	kubectl apply -f cloud-center-web-server.yaml kubectl get hpa -n zxy #查看hpa状态