kubeadm（七）——安装metrics-server

一、下载

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

cd metrics-server-master/1.8+

二、修改yaml文件（镜像被墙）

vim metrics-server-deployment.yaml

修改内容：image：registry.cn-beijing.aliyuncs.com/xingshulin/metrics-server-amd64:v0.2.1

三、创建

kubectl create -f deploy/1.8+/

四、效果

刚搭建完会没有数据，过几分钟就能收集到数据了

[root@master 1.8+]# kubectl top nodes
NAME      CPU(cores)   CPU%   MEMORY(bytes)   MEMORY%
master    153m         7%     2587Mi          33%
node001   34m          3%     616Mi           69%
node002   31m          3%     584Mi           65%
node003   39m          3%     574Mi           64%

五、解决错误

遇到的错误信息

E0725 10:12:12.679030       1 reststorage.go:128] unable to fetch node metrics for node "node002": no metrics known for node
E0725 10:12:12.679056       1 reststorage.go:128] unable to fetch node metrics for node "node003": no metrics known for node
E0725 10:12:12.679061       1 reststorage.go:128] unable to fetch node metrics for node "master": no metrics known for node
E0725 10:12:12.679066       1 reststorage.go:128] unable to fetch node metrics for node "node001": no metrics known for node
E0725 10:12:14.152632       1 reststorage.go:128] unable to fetch node metrics for node "node002": no metrics known for node
E0725 10:12:14.152656       1 reststorage.go:128] unable to fetch node metrics for node "node003": no metrics known for node
E0725 10:12:14.152662       1 reststorage.go:128] unable to fetch node metrics for node "master": no metrics known for node
E0725 10:12:14.152666       1 reststorage.go:128] unable to fetch node metrics for node "node001": no metrics known for node
E0725 10:12:15.344225       1 reststorage.go:128] unable to fetch node metrics for node "node003": no metrics known for node
E0725 10:12:15.344252       1 reststorage.go:128] unable to fetch node metrics for node "master": no metrics known for node
E0725 10:12:15.344257       1 reststorage.go:128] unable to fetch node metrics for node "node001": no metrics known for node
E0725 10:12:15.344261       1 reststorage.go:128] unable to fetch node metrics for node "node002": no metrics known for node

解决办法，修改metrics-server-deployment.yaml

--- a/deploy/1.8+/metrics-server-deployment.yaml
+++ b/deploy/1.8+/metrics-server-deployment.yaml
@@ -29,6 +29,10 @@ spec:
         emptyDir: {}
       containers:
       - name: metrics-server
+        command:
+        - /metrics-server
+        - --kubelet-preferred-address-types=InternalIP
+        - --kubelet-insecure-tls
         image: k8s.gcr.io/metrics-server-amd64:v0.3.3
         imagePullPolicy: Always
         volumeMounts:

重新部署

kubectl apply -f deploy/1.8+/

下面这篇文章写的不错
https://www.cnblogs.com/ding2016/p/10786252.html

遗憾的是，当前Kubernetes Dashboard还不支持metrics-server。因此如果使用metrics-server替代了heapster，将无法在dashboard中以图形展示Pod的内存和CPU情况。