Cfrgtkky

I'm doing this tutorial on creating a kubernetes cluster on Ubuntu 16.04 (I'm using 18.04 but there is no tutorial on that version yet). I finished the first three steps and everything went fine. I'm now trying to initialise the cluster with the master node in it, and I'm a bit stuck.

When I run the master.yml playbook with

ansible-playbook -i hosts ~/kube-cluster/master.yml

I get the following output:

    $ ansible-playbook -i hosts master.yml 

    /Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/requests/__init__.py:91: RequestsDependencyWarning: urllib3 (1.24.1) or chardet (3.0.4) doesn't match a supported version!

      RequestsDependencyWarning)



    PLAY [master] *********************************************************************************



    TASK [Gathering Facts] ************************************************************************

    ok: [master]



    TASK [initialize the cluster] *****************************************************************

    changed: [master]



    TASK [create .kube directory] *****************************************************************

     [WARNING]: Module remote_tmp /home/ubuntu/.ansible/tmp did not exist and was created with a

    mode of 0700, this may cause issues when running as another user. To avoid this, create the

    remote_tmp dir with the correct permissions manually



    changed: [master]



    TASK [copy admin.conf to user's kube config] **************************************************

    changed: [master]



    TASK [install Pod network] ********************************************************************

    changed: [master]



    PLAY RECAP ************************************************************************************

    master                     : ok=5    changed=4    unreachable=0    failed=0  

The only thing that's different compared to the tutorial is the warning about the /home/ubuntu/.ansible/tmp directory permissions. When I ssh into the master node server and run

kubectl get nodes

I get the following result:

NAME             STATUS     ROLES    AGE   VERSION

ip-address  NotReady   master   16m   v1.12.2

Instead of the desired

NAME             STATUS     ROLES    AGE   VERSION

master  Ready   master   16m   v1.12.2

I've tried to create the tmp directory with the ubuntu user on the server so that the warning is resolved. Unfortunately this does not change anything about the master node not being ready or having its ip address as NAME.

Question: How do I resolve this problem? How can I correctly initialise the cluster so that the master node is configured properly and is ready?

I went through your problem and created the same instance of the issue using Vagrant to run the nodes.

Repo here, if you want to try orchestrating the node setup with vagrant

Just like you, I ran into the issue you described. It turns out flannel has a couple issues with coredns on ubuntu bionic. Flannel interfers with the coredns setup and causes it to stay in a pending state.

You can use this to check the pod state


ubuntu@ubuntu-bionic:~$ kubectl get pods --namespace=kube-system
NAME READY STATUS RESTARTS AGE
coredns-576cbf47c7-hlvdj 0/1 Pending 0 52m
coredns-576cbf47c7-xmljj 0/1 Pending 0 52m
etcd-ubuntu-bionic 1/1 Running 0 52m
kube-apiserver-ubuntu-bionic 1/1 Running 0 52m
kube-controller-manager-ubuntu-bionic 1/1 Running 0 52m
kube-proxy-gvqk4 1/1 Running 0 52m
kube-scheduler-ubuntu-bionic 1/1 Running 0 51m
kubernetes-dashboard-77fd78f978-5flj8 0/1 Pending 0 4m30s


After a couple searches, I found a link to the fix here on their issues page.

Install a different CNI, they used weave there.

kubectl apply -f "https://cloud.weave.works/k8s/net?k8s-version=$(kubectl version | base64 | tr -d 'n')"

More details here from the docs

From there, your containers should start and the coredns pods should be running.


ubuntu@ubuntu-bionic:~$ kubectl get pods -n kube-system
NAME READY STATUS RESTARTS AGE
coredns-576cbf47c7-jrlbb 1/1 Running 0 11m
coredns-576cbf47c7-nfjq8 1/1 Running 0 11m
etcd-ubuntu-bionic 1/1 Running 0 10m
kube-apiserver-ubuntu-bionic 1/1 Running 0 10m
kube-controller-manager-ubuntu-bionic 1/1 Running 0 10m
kube-proxy-nrbpx 1/1 Running 0 11m
kube-scheduler-ubuntu-bionic 1/1 Running 0 10m
weave-net-459mw 2/2 Running 0 10m


And finally


ubuntu@ubuntu-bionic:~$ kubectl get nodes
NAME STATUS ROLES AGE VERSION
ubuntu-bionic Ready master 14m v1.12.2