This repo is a guide for setting up consensus client and execution client for Zhejiang testnet on Kubernetes.
Any Kubernetes platform should be fine, nevertheless performance of storage speed is still crucial. For my case, I run on GKE with node pool using e2-highcpu-8, along with premium-rwo storage class.
- Clone the repo and
cdinto it:
git clone https://github.com/108356037/zhejiang-testnet-k8s.git
cd zhejiang-testnet-k8s
- Create a dedicated namespace:
kubectl create -f ./k8s-yamls/namespace.yaml
- Create the JWT password for communication between consensus client and execution client:
JWT=$(openssl rand -hex 32)
kubectl -n zhejiang-testnet create secret generic el-cl-jwt-secret --from-literal password=$JWT
- Deploy consensus client and execution client:
# For exectution client (using Besu)
kubectl create -f ./k8s-yamls/besu-depl.yaml
# For consensus client (using Lighthouse)
kubectl create -f ./k8s-yamls/lighthouse-depl.yamlIt should not take too long to sync zhejiang testnet nodes, in my case it took about 20~30 minutes for both clients to fully synchronized. From my experiments, if Besu is unresponsive even after Lighthouse is fully synchronized, try manually triggering a restart on the Besu pod.
I initially tried out Geth+Lighthouse, but as many reported, there currently seems to be a bug for this combo on Zhejiang testnet: ethereum/go-ethereum#26647. Therefore I switched to Besu+Lighthouse and discovered this combo works pretty well on Kubernetes.
An exploration with the new clients and withdrawals can be found here: https://mirror.xyz/daishioubu.eth/hyIn44oruz7FtezwhZV6EA5jiPSaq5lvOZQ4OY8VfGY