heapster, kube-dns, kubernetes-dashboard, tiller-deploy always stuck at 'ContainerCreating' state

[Sakib37]

I choose this repository as it is providing the opportunity to deploy latest Kubernetes. However, I tried this but the heapster, kube-dns, kubernetes-dashboard, tiller-deploy pod always stuck at 'ContainerCreating' state. I tried with increased memory and cpu for master and node. Also, each time I waited for around 2 hours for those pods to be running. But it did not work.

[coolpalani]

I request you check the logs kubectl logs pods/kube-dns-6f4fd4bdf-brhd7 -n kube-system

[Sakib37]

I tried, but the problem is, in "ContainerCreating" state there is no log available from a pod.

[xbernpa]

@Sakib37 could you provide the output of kubectl describe po of the containers stuck in ContainerCreating state : kubectl describe po -n kube-system kube-dns-6f4fd4bdf-brhd7 as I wasn't able to recreate the issue locally, this will give the reason the pods are not created in the events section of the output.

[Sakib37]

Please click each link to see the output

$kubectl describe po -n kube-system kube-dns-86f4d74b45-flnvh

$ kubectl get pod --all-namespaces -o wide

[stheodosis]

Hi to all,

I'm facing the same problem. I think that the problem is that all pods have assigned the 10.0.2.15 IP, which is the default host only IP of vagrant deployment.

I tried to specify the interface name in the Vagrant file , with no success Please also search the https://kubernetes.io/docs/setup/independent/troubleshooting-kubeadm/ page for "Default NIC When using flannel as the pod network in Vagrant" to find more details in this problem

[stheodosis]

Using the default network plugin, weaver the following events appear in the describe pod command:

kubectl --namespace=kube-system describe pod kube-dns-86f4d74b45-twfq7

Events:
  Type     Reason                  Age                From                 Message
  ----     ------                  ----               ----                 -------
  Warning  FailedScheduling        9m (x15 over 12m)  default-scheduler    0/1 nodes are available: 1 node(s) were not ready.
  Warning  FailedScheduling        6m (x11 over 8m)   default-scheduler    0/2 nodes are available: 2 node(s) were not ready.
  Normal   SuccessfulMountVolume   5m                 kubelet, k8sworker1  MountVolume.SetUp succeeded for volume "kube-dns-config"
  Normal   SuccessfulMountVolume   5m                 kubelet, k8sworker1  MountVolume.SetUp succeeded for volume "kube-dns-token-c5xqw"
  Warning  FailedCreatePodSandBox  1m                 kubelet, k8sworker1  Failed create pod sandbox: rpc error: code = DeadlineExceeded desc = context deadline exceeded
  Normal   SandboxChanged          1m                 kubelet, k8sworker1  Pod sandbox changed, it will be killed and re-created.

[stheodosis]

The issue has been resolved by installing the vagrant-hostmanager plugin.

[YiannisGkoufas]

Thanks for the info @stheodosis ! I guess you have to give vagrant hostmanager after vagrant up right?

[stheodosis]

Hi Yiannis,

You need to install the plugin first. Afterwards no other action is needed. Perhaps you should specify the VM's IP. All hostnames map there will be in place

[YiannisGkoufas]

Hi Spyro, still getting the same error even after installing the hostmanager plugin :(