Skip to content

Latest commit

 

History

History
141 lines (103 loc) · 3.72 KB

README.md

File metadata and controls

141 lines (103 loc) · 3.72 KB

Running BeaconKit with Kurtosis

What is Kurtosis

Kurtosis is a platform for running distributed systems on Docker / Kubernetes. It provides a simple, powerful framework for spinning up and tearing down distributed systems programmatically.

How to Use

To use BeaconKit with Kurtosis, you'll first need to install the Kurtosis CLI and its dependencies. You can find instructions for doing so here.

Docker/local environment

Once you've installed the Kurtosis CLI, you can use it to spin up a Beacon network with the following command from within the root directory of the beacon-kit repository:

make start-devnet

This will automatically build your beacond Docker image from the local source code and spin up a Kurtosis network based on the config file in kurtosis/beaconkit-local.yaml. Once complete, this will output all the network information for your nodes, like so:

Example Network

When you want to tear down your network, you can do so with the following commands:

make stop-devnet
make rm-devnet

And that's it!

Deploy Devnet to Kubernetes Networks

Deploy to a Google Cloud Network

This will allow you to deploy a network orchestrated in the same way as make start-devnet, but on a cloud environment. A similar approach can be taken for a local Kubernetes environment (alternative commands will be commented with [Docker Desktop K8s]).

  • First, open your Kurtosis config:

    kurtosis config path

# The command will output a path which you need to open in an editor
/Users/.../kurtosis-config.yml

  • Update the Kurtosis config with the following, replacing the entire file:

    config-version: 2
should-send-metrics: true
kurtosis-clusters:
  docker:
    type: "docker"
  docker-desktop:
    type: "kubernetes"
    config:
      kubernetes-cluster-name: "docker-desktop"
      storage-class: "hostpath"
  cloud:
    type: "kubernetes"
    config:
      kubernetes-cluster-name: "cloud"
      storage-class: "premium-rwo"

  • Next, ensure Kurtosis is using the correct config so it deploys to the cloud instead of local Docker Desktop:

    kurtosis cluster set cloud

# [Docker Desktop K8s]: kurtosis config use-context docker-desktop

  • Now ensure your Kubernetes config is using the correct context, i.e., the one context you wish to deploy to:

    kubectl config use-context gke_prj-.....

# [Docker Desktop K8s]: kubectl config use-context docker-desktop

  • Run Kurtosis Gateway. This command will start a local "gateway" to connect your local machine to your remote Kubernetes cluster. Run this in a separate shell:

    kurtosis gateway

  • Cloud-based deployments require a Docker image, as local Docker images cannot be pulled from the remote instance. If you want to update the image, edit:

    # Found in beacon-kit/kurtosis/beaconkit-cloud.yaml
images:
  beaconkit: ghcr.io/berachain/beacon-kit:main

  • Deploy. Note that re-executing the same command twice will start the network from zero again unless you change the enclave name in the Makefile:

    make start-devnet-cloud

  • View your deployment in K9s, navigating to the relevant namespace. It should be named kt-my-cloud-devnet-${whoami}.

Helper Commands

If you want to start from a clean state and remove all existing pods:

# Everything is wrecked
kurtosis clean -a
kurtosis engine restart

If you manually kill a pod and want to restart it:

# Note that for the namespace, you should remove the "kt-" prefix
kurtosis service start {namespace} {podname}