[Resolved] Error from server (ServiceUnavailable): the server is currently unable to handle the request (get nodes.metrics.k8s.io)

Issue:- 

When installing the metricserver in the kubernetes getting the following error. 

Error:- 

 Error from server (ServiceUnavailable): the server is currently unable to handle the request (get nodes.metrics.k8s.io)

Effect:-

Due to the above error the metricserver will not work

[centos@kubemaster dashboard]$ kubectl top nodes

W0529 10:18:25.234815   13218 top_node.go:119] Using json format to get metrics. Next release will switch to protocol-buffers, switch early by passing --use-protocol-buffers flag

Error from server (ServiceUnavailable): the server is currently unable to handle the request (get nodes.metrics.k8s.io)

Resolution:-

Follow the steps mentioned below to install the metric server

1. sudo yum install -y wget

2. wget https://github.com/kubernetes-sigs/metrics-server/releases/latest/download/components.yaml

3. vim components.yaml

4.  you need to add the following 1 arguments  with the command

# kubectl edit deployment -n kube-system metrics-server

          - --kubelet-insecure-tls

          - --kubelet-preferred-address-types=InternalIP

Now if you again go to the sublime editor where we have opened the components.yaml file and look fir the containers definition i.e. just search containers with

control+f

Replace

      containers:

      - args:

        - --cert-dir=/tmp

        - --secure-port=443

        - --kubelet-preferred-address-types=InternalIP,ExternalIP,Hostname

        - --kubelet-use-node-status-port

        - --metric-resolution=15s

With

      containers:

      - args:

        - --cert-dir=/tmp

        - --secure-port=443

        - --kubelet-preferred-address-types=InternalIP                     (#this one was added)

        - --kubelet-insecure-tls                                                          (#this one was added)

        - --kubelet-use-node-status-port

        - --metric-resolution=15s

5. Now when you will execute the command it should be successful

 [centos@kubemaster dashboard]$ kubectl top nodes --use-protocol-buffers  
 NAME              CPU(cores)  CPU%  MEMORY(bytes)  MEMORY%  
 kubemaster.unixcloudfusion.in  189m     9%   2087Mi     56%  
 kubenode1.unixcloudfusion.in  77m     3%   904Mi      24%  
 kubenode2.unixcloudfusion.in  90m     4%   1163Mi     31%  

 

Explanation:-

The hostname resolution in this case was happening through the internal DNS system which have pods/service entries but not cluster node entries.