Hoping someone can help me find a bit more definitive answer for part of my lab. I recently installed a 3 host Harvester cluster on some mini PC (HP T740) and it’s working. But I want some access control, and eventually some containers. It seems like I need Rancher installed somewhere and I’m trying to puzzle this out.
One option seems to be put it in docker for very light systems and probably the first thing I’m going to try. But if there are other suggestions, I’d like to hear them. Especially ways to emulate how it would be in a real vm+container harvester/rancher workflow.
And yes it is complex sorting through all the kubernetes parts of how some of these things can be set up. Makes learning kind of a steep curve.
I think you are starting out hard with Harvester. That’s jumping right in the deep end. AND, just my 2 cents, but if you want to learn Kubernetes, it seems to me you really need to learn to set it up with infrastructure as code. In my case, that was Ansible. I used AI to help me build ansible playbooks to fully install 3 control nodes, 3 worker nodes, Longhorn, the NFS CSI driver and Rancher onto 3 VMs. I don’t have three separate physical devices to play with so I spun up three VMs, to emulate using three separate NUC boxes as my cluster. I added an extra 200gb virtual drive to each VM to act as my long horn storage. Again, just trying to enulate what it will look like when I have the budget to buy three physical boxes. In the mean time I have plenty of resources on my Proxmox node to support this. I probably should learn how to use Terraform to spin up the VMs, but that is for a future experiment.
To solve your problem I would just run rancher in another VM or Docker outside of your Harvester cluster, spin it up then import the cluster.
NFS CSI and SMB CSI are two things I need to do so I can save a little space on my Longhorn drives. Even if I can only use these for ISO storage it will save space.
I use the NFS driver for backups. But I don’t run any production workloads in K3S. Its mostly just for learning and experimenting for me at this point.
There is a real world case starting near me, and they pay a lot more so it might benefit me to learn about this topic right now. There is a use case where I work, if they would open up another position with enough pay for me to move over, I might be able to reduce our vmware overhead.
But obviously I at step 1 in the process, lots more to read and understand.