XinFin-K8S

Prerequisite

• Kubernetes Cluster (e.g 1 Master, 2 Workers) - Tutorial - Digital Ocean
• Docker

Pull image on worker node(s)

sudo docker pull xinfinorg/quorum:istanbul-tools-k8s

Clone repository & execute the following:

Create genesis as ConfigMap

kubectl create -f XinFin-K8S/genesis-configmap.yaml

Create nodekeys as secret (base64 encoded)

kubectl create -f XinFin-K8S/secret-minernode.yaml

Pre-provision volumes for our nodes to store chaindata (hostpath)

kubectl create -f XinFin-K8S/geth-persistentvolume.yaml

Deploy bootnode

kubectl create -f XinFin-K8S/bootnode-deployment.yaml

Copy bootnode POD_IP using Dashboard or CLI

kubectl proxy

Copy bootnode POD_IP address & update enode address of --bootnodes argument in minernode-statefulset.yaml & membernode-statefulset.yaml (Can be improved to automatically fetch bootnode IP)

Launch minernodes (replicas: 2)

kubectl create -f XinFin-K8S/minernode-statefulset.yaml

Launch membernode (replicas: 1)

kubectl create -f XinFin-K8S/membernode-statefulset.yaml

We have defined 4 validators in extradata field so we need to scale the minernode(s) to start mining.

Scaling

Minernode

kubectl scale statefulset minernode --replicas=4

Membernode

kubectl scale statefulset membernode --replicas=2

Deploy constellation-bootnode which can be used for autodiscovery of other constellation nodes.

kubectl create -f XinFin-K8S/constellation-bootnode-deployment.yaml

Accessing Geth console

kubectl port-forward minernode-0 8545:8545

geth attach http://localhost:8545

To stop the network use "delete" instead of "create". (Please note launching a clean network would require manual cleanup of content in /data directory on worker hosts)