Hello all! I'm new to the Suse/Rancher ecosystem and want to say thanks for this ecosystem and the Suse community at large for your contributions here. I would like to set up a reproducible homelab bootstrap process that I can do for me and then a bunch of folks in my neighborhood so they can run their own redundant homelabs (and maybe eventually we could add some community mirror redundancy fun later). I've wanted to use a simplified k8s toolset that would take the place of homelab staples like openmediavault, proxmox, and truenas which typically used Docker to deploy services to solve for single point of failure, inconsistent configurations, backups suck, and and RAID configurations have a terrible recovery rate when a disk crashes. So I'm glad the homelab community is moving towards more k8s under-the-hood options, as well as ZFS or Longhorn mirroring given how cheap storage is now. I knew about Rancher, but just learned about Harvester and so happy to find out that folks are already using this as their homelab foundation! So, my goal is to set up a reproducible way to install harvester + rancher (and likely other ecosystem tools) onto a single "beefy" (motherboard, cpus, mem, 5 x 18TB object storage disks, 4 x 2TB longhorn storage) homelab node that will serve as my initial virtualization and storage node bootstrap various physical edge clusters. These phsyical edge clusters will have various properties I won't get into, but some are SBC clusters that may be portable (temporarily airgapped), while others can be full blown servers on prem (below is a video of a pi cluster unit with poe I/O boards). Considering the Harvester + Rancher setup, I would really hope to approach building a guide and automate some pieces. 1. Am I thinking about these tools in the right way or am I losing the plot for how the community generally uses Rancher/Harvester together? 2. Are there circular dependency issues or production concerns with using Harvester VMs as the main k3s cluster nodes, to then have Rancher deployed on that cluster, which then manages the main cluster it runs on and any new cluster that gets created? a. If this does work, is there an order of operations that must be followed during backups for running this configuration? 3. Would it make sense to have a boot script that modifies the Harvester install to start with a local cluster of K3S that will pre-install a Rancher installation? a. Does this already exist and is maintained by Suse or a community member? 4. Any particularly good reading that might help me ask better questions or some recipe setups for deploying these from scratch? My hope is to treat Rancher as my single pane of glass that runs over my single-node Harvester cluster. I see this as a nice way to get the homelab community introduced to many of the kubernetes concepts without needing to do a lot of toiling with multiple node setups. I know HA is the ultimate way to ensure redundancy and availability, but I want to enable my guide in a series of steps that meets these communities where they're at and slowly encourage them to add physical redundancy with various nodes they can acquire over time and as different needs and use cases arise. I hope my goals and questions make sense. Looking forward to dig in more!

Hiya Brian, thanks for this post! While I'm sure many folks here have recipes to share, inspired by your message, I've created a separate new #C09NGJF8BFB channel to enable better interaction. We'd love for everyone on the channel to join and share stories there!

Thanks Divya! Regarding question 2, I asked this in a more concise way here and got some great answers. I'm diving into question 3/4 now and will follow up on #C09NGJF8BFB as suggested.