I'm diving into Kubernetes and using the ansible-k3s-argocd-renovate repo to build a test SCADA system infrastructure on vSphere, with plans to later switch to Proxmox for a pre-production setup. I'm facing quite a few challenges, like setting up ZFS pools on the VMs, debugging issues with Ansible, and learning a lot of new concepts. I come from a background of over 35 years in PLC controls and I'm wondering if this repository is a suitable platform to learn from. It has plenty of features like Grafana dashboards and CI/CD integration with ArgoCD, but it often feels frustrating instead of straightforward. Am I on the right track here? Also, I'd love to connect with someone experienced who can mentor me through this.
2 Answers
Honestly, I wouldn't have started with the ansible-k3s-argocd-renovate repo right away. It’s better to focus on the basics and gradually get into the technologies that align specifically with your project needs. You might want to check out 'Kubernetes the Hard Way' by Kelsey Hightower to solidify your understanding of the fundamentals. There are also great learning paths on KodeKloud and other Kubernetes resources that can help. Once you’re comfortable with kubectl and running simpler deployments, then you can explore GitOps and tools like ArgoCD with confidence. If you're set on using Ansible, definitely look into the Kubernetes Collection for Ansible as well. Starting with standard setups will make it easier when you move on to more complex solutions like Longhorn or ArgoCD.
Using scripts that automate everything can lead to big headaches if something goes wrong. The ansible-k3s-argocd-renovate repo might sound appealing at first, but it’s often better to install k3s and ArgoCD from scratch while only including the features you really need. Believe me, once you get the hang of it, you can set it all up fairly quickly—probably within an hour or two. Regarding the ZFS pools, if you’re just setting up persistent volumes, consider using NFS for the time being. You can always transition to something more suitable later on.
An hour sounds way too optimistic! With all the Ansible complexities, it feels like it could easily take days. About the ZFS pools, I initially wanted to stick with Longhorn too, but there are dependencies that make it a hassle to remove from the playbooks. I would really appreciate any guidance you could offer, even just a quick rundown over Teams!