k8s-Metrics Serverv0.5.2安装以及报错解决

在查看k8s 环境中，使用top命令查看Pod的CPU、内存使用过程中，遇到以下问题：

$ kubectl top node

W0818 03:22:46.090578   26207 top_pod.go:140] Using json format to get metrics.e-protocol-buffers flag
error: Metrics API not available

如上看到ERROR信息“Metrics API not available”，这是由于该Kuernetes环境没有安装metric-server组件导致的。

安装metric-server组件可以参考Github上的安装参考资料：https://github.com/kubernetes-sigs/metrics-server，如下所示：

[root@sdw88 glj_tmp]# kubectl create -f components.yaml
serviceaccount/metrics-server created
clusterrole.rbac.authorization.k8s.io/system:aggregated-metrics-reader created
clusterrole.rbac.authorization.k8s.io/system:metrics-server created
rolebinding.rbac.authorization.k8s.io/metrics-server-auth-reader created
clusterrolebinding.rbac.authorization.k8s.io/metrics-server:system:auth-delegator created
clusterrolebinding.rbac.authorization.k8s.io/system:metrics-server created
service/metrics-server created
deployment.apps/metrics-server created
apiservice.apiregistration.k8s.io/v1beta1.metrics.k8s.io created

在正常可以科学上网的情况下，可以拉取到镜像k8s.gcr.io/metrics-server/metrics-server:v0.5.2，即可安装完成。

但是呢公司服务器就是不行!只能通过:

#1,先到阿里云下载
docker pull registry.aliyuncs.com/google_containers/metrics-server:v0.5.2
#2.更换标签命:
docker tag registry.aliyuncs.com/google_containers/metrics-server:v0.5.2 k8s.gcr.io/metrics-server/metrics-server:v0.5.2

然后查看Pod是否OK,等待一段时间查看哈:

[root@sdw88 glj_tmp]# kubectl get pod -n kube-system 
NAME                              READY   STATUS    RESTARTS   AGE
coredns-7f6cbbb7b8-44nfm          1/1     Running   0          7d19h
coredns-7f6cbbb7b8-pjdt7          1/1     Running   0          7d19h
etcd-sdw88                        1/1     Running   0          7d19h
kube-apiserver-sdw88              1/1     Running   0          7d19h
kube-controller-manager-sdw88     1/1     Running   0          7d19h
kube-flannel-ds-bndj6             1/1     Running   0          7d18h
kube-flannel-ds-cnxxc             1/1     Running   0          7d18h
kube-flannel-ds-l28g9             1/1     Running   0          7d18h
kube-proxy-8lppt                  1/1     Running   0          28h
kube-proxy-ns4tv                  1/1     Running   0          28h
kube-proxy-pnmm8                  1/1     Running   0          28h
kube-scheduler-sdw88              1/1     Running   0          7d19h
metrics-server-5b6dd75459-7wzvr   0/1     Running   0          12s

该metric-server Pod尚未准备运行完成，describe查看其详细信息：

[root@sdw88 glj_tmp]# kubectl describe pod metrics-server-5b6dd75459-7wzvr -n kube-system 


Events:
  Type    Reason     Age        From               Message
  ----    ------     ----       ----               -------
  Normal  Scheduled  31s        default-scheduler  Successfully assigned kube-system/metrics-server-5b6dd75459-7wzvr to sdw34
  Normal  Pulled     <invalid>  kubelet            Container image "k8s.gcr.io/metrics-server/metrics-server:v0.5.2" already present on machine
  Normal  Created    <invalid>  kubelet            Created container metrics-server
  Normal  Started    <invalid>  kubelet            Started container metrics-server
  Warning  Unhealthy  9s (x5 over 49s)  kubelet            Readiness probe failed: HTTP probe failed with statuscode: 500

可以看到在描述中的事件信息中直接:Readiness probe failed: HTTP probe failed with statuscode: 500

然后日志全是这种错:

E1216 10:54:51.417065       1 scraper.go:139] "Failed to scrape node" err="Get \"https://192.168.1.33:10250/stats/summary?only_cpu_and_memory=true\": x509: cannot validate certificate for 192.168.1.33 because it doesn't contain any IP SANs" node="sdw33"
I1216 10:54:53.739687       1 server.go:188] "Failed probe" probe="metric-storage-ready" err="not metrics to serve"
I1216 10:55:03.739164       1 server.go:188] "Failed probe" probe="metric-storage-ready" err="not metrics to serve"
E1216 10:55:06.408232       1 scraper.go:139] "Failed to scrape node" err="Get \"https://192.168.1.33:10250/stats/summary?only_cpu_and_memory=true\": x509: cannot validate certificate for 192.168.1.33 because it doesn't contain any IP SANs" node="sdw33"
E1216 10:55:06.409199       1 scraper.go:139] "Failed to scrape node" err="Get \"https://192.168.1.34:10250/stats/summary?only_cpu_and_memory=true\": x509: cannot validate certificate for 192.168.1.34 because it doesn't contain any IP SANs" node="sdw34"
E1216 10:55:06.411028       1 scraper.go:139] "Failed to scrape node" err="Get \"https://192.168.1.88:10250/stats/summary?only_cpu_and_memory=true\": x509: cannot validate certificate for 192.168.1.88 because it doesn't contain any IP SANs" node="sdw88"

针对这个报错，仔细查阅Github上的安装参考资料：https://github.com/kubernetes-sigs/metrics-server (全英文-日了狗)

1.安装要求，如下红框中说明，Kubelet证书需要由群集证书颁发机构签名（或可以禁用证书验证，通过对Metrics Server配置参数–Kubelet-insecure-tls不安全）

2.配置，如下红框中说明，添加了--Kubelet-insecure-tls这个配置，就不会去验证Kubelets提供的服务证书的CA。只能在测试环境哦-不安全

修改之前的apply的components.yaml文件，添加--Kubelet-insecure-tls参数，

–kubelet-insecure-tls：kubelet 的10250端口使用的是https协议，连接需要验证tls证书。–kubelet-insecure-tls不验证客户端证书

如下所示

重新apply components.yaml文件,以及查看Pod

[root@sdw88 glj_tmp]# kubectl get pod -n kube-system 
NAME                              READY   STATUS    RESTARTS   AGE

metrics-server-5b6dd75459-7wzvr   1/1     Running   0          33s

验证top：

[root@sdw88 glj_tmp]# kubectl top pod -n kube-system
NAME                              CPU(cores)   MEMORY(bytes)   
coredns-7f6cbbb7b8-44nfm          3m           23Mi            
coredns-7f6cbbb7b8-pjdt7          3m           21Mi            
etcd-sdw88                        34m          56Mi            
kube-apiserver-sdw88              102m         290Mi           
kube-controller-manager-sdw88     30m          58Mi            
kube-flannel-ds-bndj6             3m           23Mi            
kube-flannel-ds-cnxxc             3m           21Mi            
kube-flannel-ds-l28g9             4m           26Mi            
kube-proxy-8lppt                  11m          20Mi            
kube-proxy-ns4tv                  1m           19Mi            
kube-proxy-pnmm8                  9m           19Mi            
kube-scheduler-sdw88              6m           28Mi            
metrics-server-5b6dd75459-7wzvr   4m           22Mi