Update readme and general cleanup

.pre-commit-config.yaml

@@ -1,9 +1,7 @@
 repos:
 - repo: https://github.com/k8s-at-home/sops-pre-commit
-  rev: v2.1.0
+  rev: v2.1.1
-  hooks:
+  hooks: [{id: forbid-secrets}]
-  - id: forbid-secrets
 - repo: https://github.com/zricethezav/gitleaks
-  rev: v8.6.1
+  rev: v8.8.11
-  hooks:
+  hooks: [{id: gitleaks}]
-    - id: gitleaks

readme.md

@@ -3,9 +3,8 @@
 [🖥️ Upstream][upstream] • [🐙 GitHub Mirror][github]
 
 This repository contains the configuration, scripts, and other goodies for
-building and managing my kubernetes clusters (right now, that's just my home
+building and managing my kubernetes clusters. I share the source with you so
-cluster). I share the source with you so you can make exactly the same mistakes
+you can make exactly the same mistakes as I do.
-as I do.
 
 ## Setup
 
@@ -17,87 +16,9 @@ pre-commit install --install-hooks
 pre-commit autoupdate
 ```
 
-## Provision Machines
+## Usage
 
-Before we interact with the cluster, we have some manual work to do.
+See each cluster's individual readme for more information
-
-### Manual Preparation
-
-- Currently, my nodes are Arch Linux machines on bare metal
-  - Nodes must be ready to be controlled via Ansible
-    - Have `python3` installed
-    - Need to be `ssh`-able from a controller (my workstation)
-      - `curl -L files.lyte.dev/key.pub >> ~/.ssh/authorized_keys`
-  - Nodes must support Longhorn: https://longhorn.io/docs/1.2.3/deploy/install/#installation-requirements
-    - Nodes must be running on a host filesystem that supports `file extents`
-    - Provisioning takes care of the rest
-
-### Automated Provisioning
-
-- Setup Ansible on the controller (from `./ansible`)
-  - `pushd ansible && ansible-galaxy install -r requirements.yml --force`
-- Verify Ansible can reach hosts (from `./ansible`)
-  - `pushd ansible && ansible all -i inventory/hosts.yml --list-hosts`
-  - `pushd ansible && ansible all -i inventory/hosts.yml -m ping`
-- Use Ansible to build the cluster as configured on all nodes (from `./ansible`)
-  - `pushd ansible && ansible-playbook -i inventory/hosts.yml ./build-k3s-cluster.yml`
-
-And the cluster is up! If you want to interact with it from your controller,
-you can do this:
-
-- Copy the cluster information from the `./k3s-cluster-config.kubeconfig.yaml` file into
-  your existing `~/.kube/config` (or just copy it there if it doesn't exist)
-  - You will need to edit the host from `localhost`/`127.0.0.1` to the correct host
-
-```bash
-ansible -i ansible/inventory/hosts.yml $REMOTE_HOST -m fetch \
-  -a "src=/etc/rancher/k3s/k3s.yaml dest=./k3s-cluster-config.kubeconfig.yaml flat=yes"
-# TODO: this did not work for me
-# env KUBECONFIG="~/.kube/config:./k3s-cluster-config.kubeconfig.yaml" \
-# kubectl config view --flatten | sed "s/127.0.0.1/$REMOTE_HOST/" > ~/.kube/new-config
-sed -i 's/127\.0\.0\.1/10.0.0.87' ~/.kube/config
-```
-
-### Automated Teardown
-
-```bash
-cd ansible
-ansible-playbook -i inventory/hosts.yml ./nuke-k3s-cluster.yml
-```
-
-## Setting up Flux
-
-- Install the `flux` CLI on a machine that can `kubectl` into the shiny, new cluster
-  - `paru -S flux-bin`
-  - or `curl -s https://fluxcd.io/install.sh | sudo bash`
-  - https://fluxcd.io/docs/installation/
-- Run the pre-flight check (you must have `~/.kube/config` setup!)
-  - `flux check --pre`
-- Create the `flux-system` namespace
-  - `kubectl create namespace flux-system --dry-run=client -o yaml | kubectl apply -f -`
-- Add the `sops-age` encryption key to the namespace
-  ```bash
-  pass k8s-clusters | grep age-secret-key | awk '{printf $2}' | \
-    kubectl --namespace flux-system create secret generic sops-age \
-    --from-file=age.agekey=/dev/stdin
-  ```
-- Install Flux (note the `fish`-isms here, so you may need to translate to `bash`-isms)
-  ```
-  flux bootstrap git --url=(git remote get-url origin) --branch=master \
-    --path=./cluster/home --private-key-file=$HOME/.ssh/flux-k8s-clusters
-  ```
-
-### Troubleshooting
-
-If you screw something up here, here are some things you can do:
-
-- `flux uninstall` will nuke flux from the cluster so you can retry from the
-  beginning of this section
-- If you get something like `sync path configuration ... would overwrite path ... of existing Kustomization`, you can edit the `path: ...` field in the
-  `flux-system/gotk-sync.yaml` file in whatever you're passing as `--path`,
-  commit, and try the bootstrap again
-- You can pretty easily nuke the entire cluster and start from scratch as
-  a last resort?
 
 # To Do & Status
 
@@ -118,10 +39,6 @@ If you screw something up here, here are some things you can do:
     - Change dummy domain to real domain
     - Change domain's nameserver to cloudflare
     - Should be all set!
-- I want to look into Talos/Sidero + PXEBoot, since that could remove a lot of the ansible stuff?
-- `k3s` has a decent amount of magic AFAICT, so I'd like to learn more about it
-  and all its components so I better understand what my system is actually
-  _doing_
 
 [upstream]: https://git.lyte.dev/lytedev/k8s-clusters
 [github]: https://github.com/lytedev/k8s-clusters