k8s clusters in QEMU in docker

Using gocli

gocli is a tiny go binary which helps managing the containerized clusters. It can be used from a docker images, so no need to install it. You can for instance use a bash alias:

alias gocli="docker run --net=host --privileged --rm -it -v /var/run/docker.sock:/var/run/docker.sock kubevirtci/gocli:latest"
gocli help

Quickstart Kubernetes

Start the cluster

Start a k8s cluster which contains of one control-plane and two nodes:

gocli run --random-ports --nodes 3 --background kubevirtci/k8s-1.13.3

Connect to the cluster

Find out the connection details of the cluster:

$ gocli ports k8s
33396
$ gocli scp /etc/kubernetes/admin.conf - > ./kubeconfig
$ kubectl --kubeconfig ./kubeconfig --insecure-skip-tls-verify --server https://localhost:33396 get pods -n kube-system
NAME                             READY     STATUS    RESTARTS   AGE
etcd-node01                      1/1       Running   0          14m
kube-apiserver-node01            1/1       Running   0          13m
kube-controller-manager-node01   1/1       Running   0          14m
kube-dns-6f4fd4bdf-mh6nb         3/3       Running   0          14m
kube-flannel-ds-4bk76            1/1       Running   0          14m
kube-flannel-ds-5zgmt            1/1       Running   1          14m
kube-flannel-ds-qbm2r            1/1       Running   1          14m
kube-proxy-gtvpb                 1/1       Running   0          14m
kube-proxy-knc6p                 1/1       Running   0          14m
kube-proxy-vx9t6                 1/1       Running   0          14m
kube-scheduler-node01            1/1       Running   0          13m

or to permamently edit kubeconfig:

$ gocli scp /etc/kubernetes/admin.conf - > ./kubeconfig
$ kubectl --kubeconfig=./kubeconfig config set-cluster kubernetes --server=https://127.0.0.1:$(gocli ports k8s|tr -d '\r\n')
$ kubectl --kubeconfig=./kubeconfig config set-cluster kubernetes --insecure-skip-tls-verify=true
$ kubectl --kubeconfig ./kubeconfig get pods -n kube-system
NAME                             READY     STATUS    RESTARTS   AGE
etcd-node01                      1/1       Running   0          14m
kube-apiserver-node01            1/1       Running   0          13m
kube-controller-manager-node01   1/1       Running   0          14m
kube-dns-6f4fd4bdf-mh6nb         3/3       Running   0          14m
kube-flannel-ds-4bk76            1/1       Running   0          14m
kube-flannel-ds-5zgmt            1/1       Running   1          14m
kube-flannel-ds-qbm2r            1/1       Running   1          14m
kube-proxy-gtvpb                 1/1       Running   0          14m
kube-proxy-knc6p                 1/1       Running   0          14m
kube-proxy-vx9t6                 1/1       Running   0          14m
kube-scheduler-node01            1/1       Running   0          13m

Destroy the cluster

$ gocli rm

Accessing the webconsole

Make sure that node01 resolves to 127.0.0.1 and that you added --ocp-port 8443 when creatin the cluster. If you did that, you can simply access the webconsole at https://127.0.0.1:8443. The login credentials are admin:admin.

The two preconditions are necessary to make the authentication redirects work.

Destroy the cluster

$ gocli rm