Deploy a Kubernetes Cluster

PREV: Setup <==> NEXT: Deploy a Fabric Network

Ready?


just check-setup

Provision a cloud Kubernetes Instance

Provision an IKS or EKS Kubernetes service at IBM or Amazon Cloud.
- Configure a 3x 4 CPU / 16 GRAM worker pool.
- single region is OK.
Configure your kubectl CLI for access to the remote cluster.
Test CLI access to the cluster:


kubectl cluster-info

Open a new shell and observe the target namespace:


k9s -n test-network

Configuration Options

IBM Cloud / IKS


export WORKSHOP_NAMESPACE="test-network"
export WORKSHOP_CLUSTER_RUNTIME="k3s"
export WORKSHOP_COREDNS_DOMAIN_OVERRIDE="false"
export WORKSHOP_STAGE_DOCKER_IMAGES="false"
export WORKSHOP_STORAGE_CLASS="ibmc-file-gold"

Amazon Web Services / EKS


export WORKSHOP_NAMESPACE="test-network"
export WORKSHOP_CLUSTER_RUNTIME="k3s"
export WORKSHOP_COREDNS_DOMAIN_OVERRIDE="false"
export WORKSHOP_STAGE_DOCKER_IMAGES="false"
export WORKSHOP_STORAGE_CLASS="gp2"

Digital ocean / DOKS


export WORKSHOP_NAMESPACE="test-network"
export WORKSHOP_CLUSTER_RUNTIME="k3s"
export WORKSHOP_COREDNS_DOMAIN_OVERRIDE="false"
export WORKSHOP_STAGE_DOCKER_IMAGES="false"
export WORKSHOP_STORAGE_CLASS="do-block-storage"

Install Nginx Ingress

Install the Nginx controller to the cluster


just nginx

Cluster Ingress DNS Domain

IKS


export INGRESS_IPADDR=$(kubectl -n traefik get svc/traefik -o json | jq -r '.status.loadBalancer.ingress[0].ip')
export WORKSHOP_INGRESS_DOMAIN=$(echo $INGRESS_IPADDR | tr -s '.' '-').nip.io

EKS


export INGRESS_HOSTNAME=$(kubectl -n traefik get svc/traefik -o json  | jq -r '.status.loadBalancer.ingress[0].hostname')
export INGRESS_IPADDR=$(dig $INGRESS_HOSTNAME +short)
export WORKSHOP_INGRESS_DOMAIN=$(echo $INGRESS_IPADDR | tr -s '.' '-').nip.io

Digital ocean


export INGRESS_HOSTNAME=$(kubectl -n traefik get svc/traefik -o json  | jq -r '.status.loadBalancer.ingress[0].ip')
export WORKSHOP_INGRESS_DOMAIN=$(echo $INGRESS_HOSTNAME | tr -s '.' '-').nip.io

Take it Further

During the workshop, one of the steps involves building a chaincode image, tagging the image, and publishing to an insecure docker registry running at localhost:5000. For cloud based clusters, the remote instance will not have access to the local insecure registry.

To upload custom chaincode, configure your local docker client with access to an IBM cloud / public container registry. In addition, make sure that the Fabric target namespace has read access to the repository, allowing the pods created in the cluster with access to your code.

To run the workshop without building and uploading custom code, you can install a chaincode package using the reference asset-transfer smart contract. This reference sample has been made available for public read access, and does not require imagePullSecrets for the chaincode pods to be started in the cluster.

To install the reference smart contract, in the "Install Chaincode" section of the workshop, skip the "build image" sections and install the contract from a CI pipeline.