Running Jenkins master and slaves in a Kubernetes cluster
Kubernetes examples running Jenkins master and slaves
Creating a cluster
Local with Docker Compose
A local testing cluster with one node can be created with Docker Compose
docker-compose up
When using boot2docker or Docker Engine with a remote host, the remote Kubernetes API can be exposed
with docker-machine ssh MACHINE_NAME -L 0.0.0.0:8080:localhost:8080 or boot2docker ssh -L 0.0.0.0:8080:localhost:8080
More info
Google Compute Engine
export KUBERNETES_HOME=~/kubernetes
export KUBERNETES_PROVIDER=gce
export KUBERNETES_NUM_MINIONS=2
export KUBE_GCE_ZONE=us-central1-a
$KUBERNETES_HOME/cluster/kube-up.sh
Creating the pods and services
GKE
gcloud compute disks create --size 20GB jenkins-data-disk
kubectl get nodes
kubectl create -f jenkins-master-gke.yml
kubectl get rc
kubectl get pods
kubectl create -f service-gke.yml
kubectl get services
kubectl create -f jenkins-slaves.yml
kubectl get rc
kubectl get pods
kubectl scale replicationcontrollers --replicas=2 jenkins-slave
kubectl describe services/jenkins
gcloud compute forwarding-rules list
AWS
This assumes a working kubernetes installation. I generate mine with kops.
If kubectl cluster-info gives you output about the location of the API server, you are likely in pretty good shape.
Next, create the working volume for jenkins:
`aws ec2 create-volume --availability-zone us-east-1a --size 20 --volume-type gp2
You'll get a response back that looks something like this:
{
"AvailabilityZone": "us-east-1a",
"Encrypted": false,
"VolumeType": "gp2",
"VolumeId": "vol-002d2b99000000000", # Write this value down
"State": "creating",
"Iops": 100,
"SnapshotId": "",
"CreateTime": "2016-12-24T17:39:34.725Z",
"Size": 20
}
Edit jenkins-master-aws.yml and put the VolumeID in the volumeID field`.
kubectl create -f jenkins-master-aws.yml
kubectl get rc
kubectl get pods
kubectl create -f service-aws.yml
kubectl get services
kubectl describe service jenkins
kubectl create -f jenkins-slaves.yml
kubectl get rc
kubectl get pods
kubectl scale replicationcontrollers --replicas=2 jenkins-slave
kubectl describe services/jenkins
These instructions get you a publically accessible Jenkins dashboard at the load balancer specified in kubectl describe service jenkins. This is likely not ideal for a production environment for a number of reasons to be explored at some future date.
Vagrant
kubectl get nodes
kubectl create -f jenkins-master-vagrant.yml
kubectl get rc
kubectl get pods
kubectl create -f service-vagrant.yml
kubectl get services
kubectl describe services/jenkins
kubectl create -f jenkins-slaves.yml
kubectl get rc
kubectl get pods
kubectl scale replicationcontrollers --replicas=2 jenkins-slave
Rolling update
kubectl rolling-update jenkins-slave --update-period=10s -f replication-v2.yml
Tearing down
kubectl stop replicationcontrollers jenkins-slave
kubectl stop replicationcontrollers jenkins
kubectl delete services jenkins
$KUBERNETES_HOME/cluster/kube-down.sh
Demo
Kubernetes cluster up
Jenkins master and slaves provisioning
Kubernetes cluster teardown
