-
Notifications
You must be signed in to change notification settings - Fork 4.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
dashboard pod cannot run on kubeadm #1578
Comments
More details: sorry forgot to add my authentication to the Kubernetes API Server troubleshooting: certificate looks ok
List services
check connectivity to API server
It looks to me that pod can't connect tot he APIServer. |
nobody to help? |
Dashboard includes tolerations for running on the master node. What node is dashboard running on? Is the API server accessible on 10.96.0.1:443 from the master node? |
@ianlewis please see below:
So dashboard is not running on the master
I can access 10.96.0.1:443 from the master node:
|
Can you access that IP from kub2 where the dashboard is running? |
kub2 can access the kubernetes service IP on port 443
|
I'm out of ideas. I don't know why it might not be working. Even still it's very likely network or Kubernetes core related rather than a bug in the Dashboard. You should try to actually send a request to the API server and make sure you can get a response. It may be that the API server cannot use etcd. |
Same issue. Latest kubeadm deploy and dashboard can't connect. Dashboard 1.5.0 works on previous kubeadm deploy, but fails on latest kubeadm. Same for Dashboard 1.5.1. Just realized my env has kubeadm 1.6 alpha and kube 1.5.1. That could be the problem.
|
@j0nesin as you using flannel or calico? |
weave |
@Masber I have same issue, kubeadmin 1.6.0-0 kube 1.5.2 Dashboard 1.5.1. |
I have the same issue too, [root@mdinglin09 .kube]# yum list installed | grep kube [root@mdinglin09 .kube]# docker logs -f 66db8728f715 |
I have the same issue too, anyone can help? [root@bongmi-k8s-master ~]# yum list installed | grep kube [root@bongmi-k8s-master ~]# kubectl logs -n kube-system kubernetes-dashboard-4027881251-vpnkq |
BTW: I use flannel network |
@Masber Yes, you're right. |
i have the same issue too, kubectl logs --namespace=kube-system kubernetes-dashboard-3615790904-zmnb4 I use flannel network as well |
Anyone can help on this issue? |
Ok, I am going to check it, now |
So, I used the installation guide on two, clean ubuntu 16 VMs.
The first attempt failed. Some pods could not be scheduled because a single core VM could not satisfy the CPU requests. I increased to two CPU cores, but the master did not start-up properly. A bit strange. My observation was similar to: kubernetes/kubernetes#33671 In the second attempt I started from scretch with a two core VM and everything worked smoothly. It did not matter if Dashboard was scheduled on master or node. Both worked So, I could not find anything Dashboard related. |
I used Weave as suggested in the initial issue description |
+1 |
Things started working for me, and I suspect it was due to using the latest kubernetes-dashboard.yaml with the annotation for running the dashboard on the master. https://github.com/kubernetes/dashboard/blob/master/src/deploy/kubernetes-dashboard.yaml |
@j0nesin I tested deployment on master AND worker node. I could not observe any difference. |
I see this issue too with kubernetes I installed
I see the kubernetes-dashboard in Below are the errors. Has anyone successfully installed kubernetes-dashboard on
|
I am also seeing the same problem on kubeadm. $ kubectl --namespace kube-system logs kubernetes-dashboard-3203962772-vw6sm --follow
Using HTTP port: 9090
Creating API server client for https://10.96.0.1:443
Error while initializing connection to Kubernetes apiserver. This most likely means that the cluster is misconfigured (e.g., it has invalid apiserver certificates or service accounts configuration) or the --apiserver-host param points to a server that does not exist. Reason: Get https://10.96.0.1:443/version: dial tcp 10.96.0.1:443: i/o timeout
Refer to the troubleshooting guide for more information: https://github.com/kubernetes/dashboard/blob/master/docs/user-guide/troubleshooting.md I don't think this is a problem with the dashboard itself because I cannot curl the apiserver from other pods. # curl https://10.96.0.1:443/version
curl: (7) Failed to connect to 10.96.0.1 port 443: Operation timed out |
+1 |
Suffering from the same issue.#1802 |
I'll try to look at this in the next week or two. |
Same issue here. Used kuebadm, looking at the docker logs it seems dashboard doesn't pass the ca cert. |
It is a kubeadm issue not dashboard. Kubernetes is responsible for mounting certs into the pods. Most logs here are related to cluster networking issue. I am sucessfully running kubeadm based cluster on RaspberryPi and my desktop PC with Ubuntu. Kubeadm has just graduated to beta and some people still have issues with dns and networking. |
If you follow the kubeadm instructions to the letter ... Which means install docker, kubernetes (kubeadm, kubectl, & kubelet), and calico with the Kubeadm hosted instructions ... and your computer nodes have a physical ip address in the range of 192.168.X.X then you will end up with the above mentioned non-working dashboard. This is because the node ip addresses clash with the internal calico ip addresses. To fix, do this during the installation: During the master node cluster creation step: When you install the POD network and have chosen calico: wget https://docs.projectcalico.org/v2.5/getting-started/kubernetes/installation/hosted/kubeadm/1.6/calico.yaml -O calico.yaml sed -i "s/192.168.0.0/'"$CALICO_IPV4POOL_CIDR"'/g" calico.yaml kubectl apply -f calico.yaml |
Same issue for me. Followed kubeadm install steps for a master and 2 node cluster on CentOS 7. When adding dashboard, found same issues as here when using flannel, but when recreated cluster with same steps but using Weave, dashboard works. Hope that helps someone narrow down where the issue is. |
@kevinhooke can you please post what you did? |
It doesn't work with flannel. You can use another network, like calico or weave. |
@donutloop I covered most of the steps I followed in a post here: https://www.kevinhooke.com/2017/10/20/deploying-kubernetes-dashboard-to-a-kubeadm-created-cluster/ |
@yarsergio is this a known limitation with kubeadm and/or dashboard, or just what we've discovered through trial and error? If it's a limitation with kubeadm it would be useful if the docs are updated to say not to use flannel, otherwise others will go down this same path and end up stuck too? |
Is there any updates or workaround with flannel network? Same issue. |
What issue? This one is related to very old version of Dashboard. |
My company struggled that in the modern setup (something with RBAC) the dashboard doesn't work. I subscribed only last week to this issue looking for a fix. Since, it looks like something has changed, because I've recently reapplied the I know that's not super helpful, but my team didn't actually come to an understanding why it wasn't working (looked like RBAC issues in the defaults), and we have no explanation why it started working suddenly when someone else tried it again "just incase" |
@flaper87 @leehambley |
@aerobiotic 's solution fixed my problem, there also has a guide for install K8s with kubeadm behind a firewall: https://zhuanlan.zhihu.com/p/40931670 |
Same issue, the dashboard only works when I run it on the master node (apiserver) but not on other nodes while I can access the apiserver (telnet).. |
Issue details
I can't make the dashboard to run. I am using a fresh kubeadm installation + calico.
Steps to reproduce
kubeadm init
kubeadm join ... --> join a new node
kubectl apply -f http://docs.projectcalico.org/v2.0/getting-started/kubernetes/installation/hosted/kubeadm/calico.yaml
kubectl create -f https://rawgit.com/kubernetes/dashboard/master/src/deploy/kubernetes-dashboard.yaml
Observed result
Expected result
dashboard --> Running
Comments
The text was updated successfully, but these errors were encountered: