githubixx/ansible-role-kubernetes-controller

Ansible role for installing Kubernetes Controller Cluster

ansible-role-kubernetes-controller's Issues

Question: Purpose of ClusterRole and ClusterRoleBinding

What is the purpose of ClusterRole and ClusterRoleBinding in this ansible role. Are there certs created for the those?

Fails at the 8080 port check

Hey there, I've updated my cluster recently and it seems to get stuck on this:

- name: Wait for kube-apiserver to listen on port 8080 on the host
  wait_for:
    port: 8080
    delay: 5
    host: "{{hostvars[inventory_hostname]['ansible_' + k8s_interface].ipv4.address}}"
  run_once: true
  delegate_to: "{{groups.k8s_controller|first}}"

I wonder if this has anything to do with

removed deprecated kube-apiserver parameter insecure-bind-address (see: #59018)

from your blog post changelog.
kube-apiserver seems to be binding :8080 to localhost for me, so this check is never successful.

Services cannot be enabled on centos 7

Services cannot be enabled on centos 7 due to an apostrophe in the end of some templates, in "WantedBy" directive:
WantedBy=multi-user.target'

error invalid configuration kind "EncryptionConfiguration" provided

In the task TASK [ansible-role-kubernetes-controller : Apply kube-apiserver-to-kubelet ClusterRole]
I am getting

fatal: [scw1.vpn -> scw1.kisiel.net.pl]: FAILED! => {
"changed": true,
"cmd": "kubectl apply --kubeconfig /var/lib/kubernetes/admin.kubeconfig -f /tmp/kube-apiserver-to-kubelet_cluster_role.yaml",
"delta": "0:00:00.157014",
"end": "2019-01-15 22:16:29.965552",
"invocation": {
"module_args": {
"_raw_params": "kubectl apply --kubeconfig /var/lib/kubernetes/admin.kubeconfig -f /tmp/kube-apiserver-to-kubelet_cluster_role.yaml",
"_uses_shell": true,
"argv": null,
"chdir": null,
"creates": null,
"executable": null,
"removes": null,
"stdin": null,
"warn": true
}
},
"msg": "non-zero return code",
"rc": 1,
"start": "2019-01-15 22:16:29.808538",
"stderr": "error: unable to recognize "/tmp/kube-apiserver-to-kubelet_cluster_role.yaml": Get https://10.18.0.111:6443/api?timeout=32s: dial tcp 10.18.0.111:6443: connect: connection refused",
"stderr_lines": [
"error: unable to recognize "/tmp/kube-apiserver-to-kubelet_cluster_role.yaml": Get https://10.18.0.111:6443/api?timeout=32s: dial tcp 10.18.0.111:6443: connect: connection refused"
],
"stdout": "",
"stdout_lines": []
}

Please help
What could have gone wrong?
I mean I can see that tcp 10.18.0.111:6443 is failing
and yes, there is nothing listening on 6443
the host is pingable on 10.18.0.111

Question about scheduler and controller-manager binding

So as you might have guessed, I'm facing this problem.

What if I care about not binding my cluster on all interfaces? Is this only a kubectl status display issue or will this prevent nominal operations going further?

FYI, I have installed k8s v1.9.7 (didn't want to deviate too much from the versions in the tutorials). I'll try the 1.10.x branch later while trying to update my cluster. So I don't know if the problem is fixed in v1.10.x but I guess not since it is still opened.

Thanks again for all the great work and documentation.