问题
I installed kubeadm to deploy multi node kubernetes cluster. Added two nodes. Those are ready. I am able to run my app using node port service. While i am trying yo access the dashboard facing an issue. I am following the steps to install dashboard in this link
kubectl create -f https://raw.githubusercontent.com/kubernetes/dashboard/master/src/deploy/recommended/kubernetes-dashboard.yaml
dash-admin.yaml:
apiVersion: rbac.authorization.k8s.io/v1beta1
kind: ClusterRoleBinding
metadata:
name: kubernetes-dashboard
labels:
k8s-app: kubernetes-dashboard
roleRef:
apiGroup: rbac.authorization.k8s.io
kind: ClusterRole
name: cluster-admin
subjects:
- kind: ServiceAccount
name: kubernetes-dashboard
namespace: kube-system
kubectl create -f dashboard-admin.yaml
nohup kubectl proxy --address="172.20.22.101" -p 443 --accept-hosts='^*$' &
Its running well and saving the output in nohup.out
When i try to access the site using the url: 172.20.22.101:443/api/v1/namespaces/kube-system/services/…. it's showing connection refused.
I observed the output in the nohup.out, it's showing the below error:
I1203 12:28:05.880828 15591 log.go:172] http: proxy error: dial tcp [::1]:8080: connect: connection refused –
回答1:
You are not running it with root or sudo permission.
I have encountered this issue, and after running using root. I was able to access it with no errors.
回答2:
log.go:172] http: proxy error: dial tcp [::1]:8080: connect: connection refused –
In case anyone comes across the above issue, know that this error occurs when one tries to hit Kubernetes API without proper permission.
Note: Its nothing to do with RBAC.
To solve this issue, I took below steps
- Check access. Execute as root.
- In case you are using kubectl proxy to connect to Kubernetes API, make sure kubeconfig file is properly configured. or try kubectl proxy --kubeconfig=/path/to/dashboard-user.kubeconfig
回答3:
I had similar problem this days and root cause of problem is that in cluster deployment (3 nodes), kuberenetes dashboard pod was up on the slave (non master) node. Issue is that proxy is provided only locally (security reasons), so dashboard console could not be started on master, neither node 3!
On master node browser error (kubectl proxy & executed on this node):
"http: proxy error: dial tcp 10.32.0.2:8001: connect: connection refused"
On slave node error (kubectl proxy & executed on this node):
"http: proxy error: dial tcp [::1]:8080: connect: connection refused"
Solution:
Status of the cluster pods showed that dashboard pod kubernetes-dashboard-7b544877d5-lj4xq
was on node 3:
namespace kubernetes-dashboard
pod kubernetes-dashboard-7b544877d5-lj4xq
node pb-kn-node03
[root@PB-KN-Node01 ~]# kubectl get pods --all-namespaces -o wide|more <br>
NAMESPACE NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES <br>
kube-system coredns-66bff467f8-ph7cc 1/1 Running 1 3d17h 10.32.0.3 pb-kn-node01 <none> <none> <br>
kube-system coredns-66bff467f8-x22cv 1/1 Running 1 3d17h 10.32.0.2 pb-kn-node01 <none> <none> <br>
kube-system etcd-pb-kn-node01 1/1 Running 2 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-apiserver-pb-kn-node01 1/1 Running 2 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-controller-manager-pb-kn-node01 1/1 Running 3 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-proxy-4ngd2 1/1 Running 2 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-proxy-7qvbj 1/1 Running 0 3d12h 10.13.40.202 pb-kn-node02 <none> <none> <br>
kube-system kube-proxy-fgrcp 1/1 Running 0 3d12h 10.13.40.203 pb-kn-node03 <none> <none> <br>
kube-system kube-scheduler-pb-kn-node01 1/1 Running 3 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system weave-net-fm2kd 2/2 Running 5 3d12h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system weave-net-l6rmw 2/2 Running 1 3d12h 10.13.40.203 pb-kn-node03 <none> <none> <br>
kube-system weave-net-r56xk 2/2 Running 1 3d12h 10.13.40.202 pb-kn-node02 <none> <none> <br>
kubernetes-dashboard dashboard-metrics-scraper-6b4884c9d5-v2gqp 1/1 Running 0 2d22h 10.40.0.1 pb-kn-node02 <none> <none>
kubernetes-dashboard kubernetes-dashboard-7b544877d5-lj4xq 1/1 Running 15 2d22h 10.32.0.2 pb-kn-node03 <none> <none>
So all active global pods are reallocated from node 3 (included dashboard) to master node after draining the node
[root@PB-KN-Node01 ~]# kubectl drain --delete-local-data --ignore-daemonsets pb-kn-node03
node/pb-kn-node03 already cordoned
WARNING: ignoring DaemonSet-managed Pods: kube-system/kube-proxy-fgrcp, kube-system/weave-net-l6rmw
node/pb-kn-node03 drained
After 2 miniutes ...
[root@PB-KN-Node01 ~]# kubectl get pods --all-namespaces -o wide|more <br>
NAMESPACE NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES <br>
kube-system coredns-66bff467f8-ph7cc 1/1 Running 1 3d17h 10.32.0.3 pb-kn-node01 <none> <none> <br>
kube-system coredns-66bff467f8-x22cv 1/1 Running 1 3d17h 10.32.0.2 pb-kn-node01 <none> <none> <br>
kube-system etcd-pb-kn-node01 1/1 Running 2 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-apiserver-pb-kn-node01 1/1 Running 2 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-controller-manager-pb-kn-node01 1/1 Running 3 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-proxy-4ngd2 1/1 Running 2 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system kube-proxy-7qvbj 1/1 Running 0 3d12h 10.13.40.202 pb-kn-node02 <none> <none> <br>
kube-system kube-proxy-fgrcp 1/1 Running 0 3d12h 10.13.40.203 pb-kn-node03 <none> <none> <br>
kube-system kube-scheduler-pb-kn-node01 1/1 Running 3 3d17h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system weave-net-fm2kd 2/2 Running 5 3d12h 10.13.40.201 pb-kn-node01 <none> <none> <br>
kube-system weave-net-l6rmw 2/2 Running 1 3d12h 10.13.40.203 pb-kn-node03 <none> <none> <br>
kube-system weave-net-r56xk 2/2 Running 1 3d12h 10.13.40.202 pb-kn-node02 <none> <none>
kubernetes-dashboard dashboard-metrics-scraper-6b4884c9d5-v2gqp 1/1 Running 0 2d22h 10.40.0.1 pb-kn-node02 <none> <none> <br>
<b>kubernetes-dashboard kubernetes-dashboard-7b544877d5-8ln2n 1/1 Running 0 89s 10.32.0.4 pb-kn-node01 <none> <none> </b><br>
And the problem was solved, kubernetes dashboard was available on master node.
来源：https://stackoverflow.com/questions/53590248/kubernetes-dashboard-showing-http-proxy-error-dial-tcp-18080-connect-co