kubectl exec and logs not working.

[sangeethah]

rke version v0.0.3-dev

Steps to reproduce the problem:
Create a cluster using hosts from Digitial Ocean provider.
Create user pods.
kubectl exec and logs do not work.

Sangeethas-MBP:~ sangeethahariharan1$ kubectl exec -it k8testnew-0fbtw  bash
Error from server: error dialing backend: dial tcp: lookup sangee-now-rke-5 on 67.207.67.2:53: no such host
Sangeethas-MBP:~ sangeethahariharan1$ kubectl logs k8testnew-0qrpx
Error from server: Get https://sangee-now-rke-5:10250/containerLogs/default/k8testnew-0qrpx/testcontainer: dial tcp: lookup sangee-now-rke-5 on 67.207.67.2:53: no such host

[galal-hussein]

@sangeethah In DO hostnames are not resolvable so kubeapi will not be able to reach the node by hostname, it will cause exec/logs to fail, to solve this you can just specify the following extra arguments for kubeapi service:

extra_args:
  kubelet-preferred-address-types: "InternalIP,ExternalIP,Hostname"

[sangeethah]

rke version v0.0.4-dev

Steps to reproduce the problem:

used the following extra_args for kube-api service:

extra_args:
  kubelet-preferred-address-types: "InternalIP,ExternalIP,Hostname"

Create a cluster using hosts from Digitial Ocean provider.
Create user pods.
kubectl exec and logs works.

[sangeethah]

Reopening this issue to support the case where when the advertised_hostname provided is not resolvable there should be no need for user to manually override the default order for kube-api service by providing kubelet-preferred-address-types

[sangeethah]

rke version v0.0.6-dev

Following configurations works fine:
Able to create K8s cluster successfully.
Able to deploy pods.
Able to launch dashboard addon.
Able to exec/log to containers.

1. Specify host , internal_address and hostname_override.
   In this case hostname_override is not resolvable:
   nodes:

• address: ec2-.us-east-2.compute.amazonaws.com
  internal_address:
  hostname_override: mynode1
  user: ubuntu
  role: [controlplane]

nodes get registered with the "hostname_override" value.

There was no extra parameters that were set for kubeapi service:

services:
  etcd:
    image: quay.io/coreos/etcd:latest
  kube-api:
    image: rancher/k8s:v1.8.3-rancher2
  kube-controller:
    image: rancher/k8s:v1.8.3-rancher2
  scheduler:
    image: rancher/k8s:v1.8.3-rancher2
  kubelet:
    image: rancher/k8s:v1.8.3-rancher2
  kubeproxy:
    image: rancher/k8s:v1.8.3-rancher2