This repo contains assets to ease deploying openshift on kubernetes/openshift using kubevirt vms, all through a tekton pipeline

Requisites

kubernetes/openshift cluster (with admin creds)
tekton/openshift pipelines and kubevirt/openshift virtualization deployed
storage in place (for the disks of the vms)

Configuration

On kubernetes

PROJECT=myproject
kubectl config set-context --current --namespace=$PROJECT
kubectl create -f tekton.yml

On openshift

PROJECT=myproject
oc new-project $PROJECT
oc adm policy add-cluster-role-to-user cluster-admin -z pipeline -n $PROJECT
oc adm policy add-scc-to-user anyuid -z pipeline

On both

mkdir /tmp/creds
cp ~/.ssh/*pub /tmp/creds
cp openshift_pull.json /tmp/creds
kubectl create configmap credentials --from-file=/tmp/creds
rm -rf /tmp/creds
kubectl create -f pipeline.yml

Note that the pipeline can easily be extended with the parameters available here

Launch a deployment

oc create -f pipelinerun.yml

Parameters

Parameter	Default Value
cluster	testk
version	stable
tag	4.9
masters	3
workers	0
numcpus	8
memory	8192
disk_size	30
network_type	OVNKubernetes
async	false

Retrieve credentials

Credentials will be available shortly after the pipeline runs, before completion.

Gathering task output

The following command will print all the information needed to access the cluster:

/etc/hosts entry
kubeadmin password
full kubeconfig

POD=$(oc get pod --sort-by={'.metadata.creationTimestamp'} -o custom-columns=NAME:.metadata.name --no-headers | grep deploy-openshift | tail -1)
oc logs $POD

Gathering task results

You can also gather kubeadmin password and the /etc/hosts entry as results of the pipeline

PIPELINERUN=$(oc get pipelinerun --sort-by={'.metadata.creationTimestamp'} -o name | tail -1)
KUBEADMIN_PASSWORD=$(oc get $PIPELINERUN -o jsonpath='{.status.pipelineResults[0].value}')
echo $KUBEADMIN_PASSWORD

PIPELINERUN=$(oc get pipelinerun --sort-by={'.metadata.creationTimestamp'} -o name | tail -1
oc get $PIPELINERUN -o jsonpath='{.status.pipelineResults[1].value}'

Kubeconfig can't be gathered this way (because of the size of the data preventing to store this way)

Note that during a run, if not running in async mode, you can copy kubeconfig with the following command:

CLUSTER=magic
POD=$(oc get pod --sort-by={'.metadata.creationTimestamp'} -o custom-columns=NAME:.metadata.name --no-headers | grep deploy-openshift | tail -1)
oc cp $POD:/root/.kcli/clusters/$CLUSTER/auth/kubeconfig kubeconfig.$CLUSTER

yuvalk / openshift-tekton-pipeline