9

minikube metrics-server HPA 自动扩缩容错误

 3 years ago
source link: https://www.cnblogs.com/wangiqngpei557/p/15676674.html
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
neoserver,ios ssh client

minikube metrics-server pod 错误

启动 minikube addons enable metrics-server 之后查看 metrics-server pod 会有如下错误

Failed to pull image "k8s.gcr.io/metrics-server-amd64:v0.5.2": rpc error: code = Unknown desc = Error response from daemon: Get https://k8s.gcr.io/v2/: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)

这个错误是k8s.gcr.io已经废弃无法访问，需要替换成在国内的镜像，可以使用阿里云的。

registry.cn-hangzhou.aliyuncs.com/google_containers

完整镜像地址。

registry.cn-hangzhou.aliyuncs.com/google_containers/metrics-server-amd64:v0.5.2

进入 minikube docker 手动pull镜像。

minikube ssh
docker pull registry.cn-hangzhou.aliyuncs.com/google_containers/metrics-server-amd64:v0.5.2

为了让metrics-server-deployment能工作需要手动打个tag并且让pod拉取镜像的规则设置成IfNotPresent。

docker tag registry.cn-hangzhou.aliyuncs.com/google_containers/metrics-server:v0.5.2 k8s.gcr.io/metrics-server/metrics-server:v0.5.2

如果不做这一步，也可以手动修改deployment image地址，但是这样比较麻烦。
修改镜像拉取规则

 imagePullPolicy: IfNotPresent

手动安装 metrics-server

如果你本地有其他错误，我建议直接关掉minikube metrics-server addons。自行安装metrics-server。

minikube addons disable metrics-server

拉取metrics-server manifast

git://github.com/kubernetes-incubator/metrics-server.git

manifast文件metrics-server/manifests/base/deployment.yaml
（注意，手动安装前还需要像第一节介绍的那样，手动拉取镜像、修改imagePullPolicy模式。）
可能会出现类似 heapster 之类错误。

Error from server (NotFound): the server could not find the requested resource (get services http:heapster:)

（注意：heapster模式的测量机制在k8s 1.6之后已经不再使用。需要手动打开控制器管理器hpa-rest拉取测量指标。）
但是minikube controller-manager需要在启动的时候设置。

kube-system   kube-controller-manager-minikube   1/1     Running   5          2d13h

启动metrics-server测量。

minikube start --extra-config 'controller-manager.horizontal-pod-autoscaler-use-rest-clients=true'

hpa 错误

创建hpa

k autoscale deployment kubia --cpu-percent=30 --min=1 --max=5

查看hpa状态 k get hpa

kubia   Deployment/kubia   <unknown>/30%          1         5         5          2d12h

查看详情 k describe hpa kubia

the HPA was unable to compute the replica count: failed to get cpu utilization: unable to get metrics for resource cpu: no metrics returned from resource metrics API

其实这里的日志是说已经能正常拉取pod的测量，虽然没有任何返回，但是不影响使用，我们做个测试。
创建一个应用程序。

apiVersion: apps/v1
kind: Deployment
metadata:
  name: kubia
  labels:
    app: kubia
spec:
  selector:
    matchLabels:
      app: kubia
  replicas: 3
  template:
    metadata:
      labels:
        app: kubia
    spec:
      containers:
      - image: luksa/kubia:v1
        name: kubia
        resources:
          requests:
            cpu: 100m

映射一个请求端口。

k expose deployment kubia --port=80 --target-port=8080

监控HPA
watch kubectl get hpa

NAME    REFERENCE          TARGETS         MINPODS   MAXPODS   REPLICAS   AGE
kubia   Deployment/kubia   <unknown>/30%   1         5         3          2m34s

kubia 程序默认是3个副本。
在另外一个pod中循环调用应用程序。

k run -it --rm --restart=Never loadgenerator --image=busybox -- sh -c "while true; do wget -O -q http://kubia.default; done"

NAME    REFERENCE          TARGETS         MINPODS   MAXPODS   REPLICAS   AGE
kubia   Deployment/kubia   51%/30%         1         5         3          3m53s

hpa生效了，我们看下describe。

NAME    REFERENCE          TARGETS         MINPODS   MAXPODS   REPLICAS   AGE
kubia   Deployment/kubia   30%/30%         1         5         5          5m12

HPA-Autoscaler 识别到CPU测量值，已经完成弹性扩容。SuccessfulRescale。

minikube version: v1.16.0

Client Version: version.Info{Major:"1", Minor:"14", GitVersion:"v1.14.9"}
Server Version: version.Info{Major:"1", Minor:"20", GitVersion:"v1.20.0"}

report

Recommend

About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK