Hello Harvester team! I'm a Container Platform and Platform Engineer at Vanderlande, and I've been working with Harvester in the Company and my homelab setup. I'm trying to use harvester-seeder to automate the provisioning of Harvester nodes on an HPE ML30 machine with ILO 5. I've been running into some issues and I believe it's due to the current state of the

harvester-seeder

repository's

README

, which seems to be outdated. My goal is to get this automated provisioning working, but I keep hitting roadblocks. I'm currently working on a feature branch,

feat-auto-provisioner

, in my own repository. In this branch, I'm using Zarf, an air-gapped solution, to create a K3s cluster and deploy everything. The main goal is to package everything into a single bundle that can be deployed in an air-gapped environment with minimal changes to the configuration files. Here's a quick summary of the problems I've faced: • The

harvester-seeder

pod seems to be looking for a

tinker-stack

service that never becomes available, even after I tried customizing the charts. • I attempted to override the

tink-server

configuration and review the code for

hegel

, but this didn't resolve the issue. • I also tried using the latest Tinkerbell chart, which led to a different problem. While I was able to successfully boot into

hookOS

, the

tink-controller

complained that the workflow was invalid, specifically mentioning that

task.0.worker

was missing. I suspect there might be breaking changes in the latest Tinkerbell charts that are incompatible with the current

harvester-seeder

. I'd appreciate it if you could take a look at my feature branch, especially the folders under

zarf

, and check some of the values or `diff`s I've made. Any insights or guidance on how to move forward would be greatly appreciated. Has anyone else encountered similar issues with

harvester-seeder

and the Tinkerbell integration? Thanks! @great-bear-19718 @bland-farmer-13503 @prehistoric-balloon-31801 @red-king-19196 @salmon-city-57654 @average-energy-22726

i am currently caught up with a few things but i will try and get you the manifests we use

There seems to be a large re-organization/refactoring recently. Almost all of the subcomponent repositories are going to be archived and sourced into the tinkerbell/tinkerbell repository.

I am not sure but smee component now supporting multiple configurations like auto-proxy, proxy which can ease out different deployment scenarios. I am not sure if this configurations are still possible on version v0.4.2. Thanks a lot for heads up!