kubefirst VS rke2

this isn't just any typical friday for us though, this is the day that we've been awaiting forever!!! we're excited to announce that not only did we release a new awesome UI for the kubefirst instant gitops platform, and not only is it all still free and installs gitops platforms in minutes, and not only is it an incredible new user experience, but we even let you play our new video game flappy-kray during the cluster provisioning operation. 😱🚀🎮

[cofounder alert] Would love for you to consider our kubefirst instant gitops FOSS platforms. Fully managed K8S: our platform provisions managed kubernetes clusters in our cloud versions of the platform - or k3d clusters when running the platform locally Support for cross cloud and bare metal: we support aws, civo, vultr*, digitalocean*, and apply our k3d platform to bare metal stories, but our homelabs community is going in some neat new directions for bare metal k8s as well. Installation on new machines should be fully automatic: 100% - single command Terraform: all infra (terraform) and app config (argocd) is powered by a gitops repository that we give you, the tf is wired up and automated with atlantis, and your changes to the platform are a pull request away. love most of our tools, but hate a couple and want a bunch more - no problem, start here. --- stuff you didn't ask for: - application delivery with argo workflows preintegrated with github or gitlab with self hosted runners - secrets management, user management, and an oidc provider with hasicorp vault that's automatically configured throughout the platform. vault is our single source of truth for every secret throughout the platform (apps, iac, ci, etc) - cluster management: management of workload clusters (rancher like) will be release in 2.2 in a few weeks. we have to release our new ui first in 2.1 and that's expected in the next week or two. --- it seems like with the immediate cross cloud / hybrid needs you have, you may need more out of cluster management than we can offer today, but it's the focus for the next 2 releases. we're an open source free solution that's trying to solve a lot of the problems that you're up against, we have an active community and would love to help support your use case.

nothing blew up accidentally this week, but our team at kubefirst is falling more and more in love with aws-nuke. it's an open source command line tool that lets you basically reset an aws account back to an empty state. if you have an environment where you regularly practice your platform provisioning, you probably know that failed destroys while iterating on orchestration can leave junk behind pretty easily. aws-nuke has been so nice to be able to blow away everything in an aws account - and then we just run terraform in the account to get all our core infra back afterward. nice allowlist filters and dryrun detail work too. check them out.

we'll be looking into leveraging this technique at kubefirst - wondering if anyone here has other thoughts on the tech used in this piece or any FOSS alternatives we should be considering for container sig validation? this seems just about as frictionless as the discipline can get - but don't know what other gems might be out there in this space that folks may be flipping over.

at kubefirst we internally love both helm and kustomize. to build our instant oss gitops platforms we use both.

at kubefirst we build containers a lot - gitops ci pipelines are part of our instant oss platforms. i ran into this issue a few years ago that blew my mind and i haven't been able to reproduce until yesterday.

check us out if it sounds neat ⭐ https://github.com/kubefirst/kubefirst

Kubernetes: launch a service of your choosing, like Pritunl VPN. Dockerize it, create some helm charts for it, set it up in a CICD pipeline of your choosing. There's also a project called nebulous you might want to check out that aims to demonstrate k8s capabilities with a live env but it's very early stages. If you can do the former task in, say, DigitalOcean, you'll have a good head start. I can send you some additional "homework" you can work on if you'd like as well and the solution to the first task I mentioned just DM me.

In this essay, we showed how to use Rancher rke2 to deploy a Kubernetes cluster with 6 Debian nodes with firewall enabled. We've also covered deploying Cilium as a CNI for our cluster and have it completely replace kube-proxy so as to increase speed and gain more observability via Cilium tools. This article also showed how to deploy Metallb to manage IP pools and load balance traffic for those IP pools. Throughout this guide, we assumed that we have an external load balancer that will distribute traffic to our workload and control plane nodes. For further information please visit rke2 official documents: "https://docs.rke2.io/".

In this blog post, we will demonstrate how easy and fast it is to deploy Sveltos on an RKE2 cluster with the help of ArgoCD, register two RKE2 Cluster API (CAPI) clusters and create a ClusterProfile to deploy Prometheus and Grafana Helm charts down the managed CAPI clusters.

Did something happen to the Apache 2 rancher? https://github.com/rancher/rancher/blob/v2.7.5/LICENSE RKE2 is similarly Apache 2: https://github.com/rancher/rke2/blob/v1.26.7%2Brke2r1/LICENS...

To provision all of my clusters, I use Rancher with RKE2. The primary Rancher server is hosted on a bootstrapped RKE2 cluster running on a VPS.

Golang has burned me more than once with bizarre design decisions that break things in a user hostile way.

The last one we ran into was a change in Go 1.15 where servers that presented a TLS certificate with the hostname encoded into the CN field instead of the more appropriate SAN field always fail validation.

The behavior could be disabled however that functionality was removed in 1.18 with no way to opt back into the old behavior. I understand why SAN is the right way to do it but in this case I didn’t control the server.

Developers at Google probably never have to deal with 3rd parties with shitty infrastructure but a lot of us do.

Here’s a bug in rke that’s related https://github.com/rancher/rke2/issues/775

just looking at this myself. I think k3s has more support for arm, but looking through the github repo there are a lot of bugs indicating its a mess. RKE2 seems to be their big push, they also have a github issue open that has been open for the last 2 releases that they are going to add a update path from k3s to rke2. https://github.com/rancher/rke2/issues/881

RKE (https://rancher.com/docs/rke) and RKE2 (https://docs.rke2.io/) from Rancher folks