home-cluster/.taskfiles/talos/Taskfile.yaml

---
# yaml-language-server: $schema=https://taskfile.dev/schema.json
version: '3'

vars:
  TALHELPER_CLUSTER_DIR: '{{.KUBERNETES_DIR}}/bootstrap/talos/clusterconfig'
  TALHELPER_CONFIG_FILE: '{{.KUBERNETES_DIR}}/bootstrap/talos/talconfig.yaml'
  TALHELPER_SECRET_FILE: '{{.KUBERNETES_DIR}}/bootstrap/talos/talsecret.sops.yaml'
  TALOSCONFIG: '{{.TALHELPER_CLUSTER_DIR}}/talosconfig'

env:
  TALOSCONFIG: '{{.TALOSCONFIG}}'

tasks:

  generate-config:
    desc: Generate Talos configuration
    cmd: talhelper genconfig --config-file {{.TALHELPER_CONFIG_FILE}} --secret-file {{.TALHELPER_SECRET_FILE}} --out-dir {{.TALHELPER_CLUSTER_DIR}}
    preconditions:
      - msg: Missing talhelper config file
        sh: test -f {{.TALHELPER_CONFIG_FILE}}
      - msg: Missing Sops config file
        sh: test -f {{.SOPS_CONFIG_FILE}}
      - msg: Missing Sops Age key file
        sh: test -f {{.SOPS_AGE_KEY_FILE}}

  apply-config:
    desc: Apply Talos configuration to a node
    cmds:
      - '{{if eq .MODE "reboot"}}kubectl drain {{.HOSTNAME}} --ignore-daemonsets --delete-emptydir-data --force{{end}}'
      - talosctl apply-config --mode={{.MODE}} --nodes {{.HOSTNAME}} --file {{.TALHELPER_CLUSTER_DIR}}/{{.CLUSTER_NAME}}-{{.HOSTNAME}}.yaml
      - talosctl --nodes {{.HOSTNAME}} health --wait-timeout=10m --server=false
      - '{{if eq .MODE "reboot"}}kubectl uncordon {{.HOSTNAME}}{{end}}'
    vars:
      MODE: '{{.MODE | default "no-reboot"}}'
      CLUSTER_NAME:
        sh: yq '.clusterName' {{.TALHELPER_CONFIG_FILE}}
    requires:
      vars: [HOSTNAME]
    preconditions:
      - msg: Missing talosconfig
        sh: test -f {{.TALOSCONFIG}}
      - msg: Unable to retrieve Talos config
        sh: talosctl config info &>/dev/null
      - msg: Node not found
        sh: talosctl --nodes {{.HOSTNAME}} get machineconfig &>/dev/null
      - msg: Talos config for node not found
        sh: test -f {{.TALHELPER_CLUSTER_DIR}}/{{.CLUSTER_NAME}}-{{.HOSTNAME}}.yaml

  upgrade-node:
    desc: Upgrade Talos on a single node
    cmds:
      - task: '{{if ne .ROLLOUT true}}down{{else}}noop{{end}}'
      - talosctl --nodes {{.HOSTNAME}} upgrade --image="factory.talos.dev/installer/{{.TALOS_SCHEMATIC_ID}}:{{.TALOS_VERSION}}" --timeout=10m
      - talosctl --nodes {{.HOSTNAME}} health --wait-timeout=10m --server=false
      - task: '{{if ne .ROLLOUT true}}up{{else}}noop{{end}}'
    vars:
      TALOS_SCHEMATIC_ID:
        sh: kubectl get node {{.HOSTNAME}} --output=jsonpath='{.metadata.annotations.extensions\.talos\.dev/schematic}'
      TALOS_VERSION:
        sh: yq '.talosVersion' {{.TALHELPER_CONFIG_FILE}}
    requires:
      vars: [HOSTNAME]
    preconditions:
      - msg: Missing talosconfig
        sh: test -f {{.TALOSCONFIG}}
      - msg: Unable to retrieve Talos config
        sh: talosctl config info &>/dev/null
      - msg: Node not found
        sh: talosctl --nodes {{.HOSTNAME}} get machineconfig &>/dev/null
      - msg: Upstream Talos version not found
        sh: curl -fsSL -o /dev/null --fail https://github.com/siderolabs/talos/releases/tag/{{.TALOS_VERSION}}

  upgrade-cluster:
    desc: Upgrade Talos on the whole cluster
    cmds:
      - task: down
      - for: { var: HOSTNAMES }
        task: upgrade-node
        vars:
          HOSTNAME: '{{.ITEM}}'
          ROLLOUT: true
      - task: up
    vars:
      HOSTNAMES:
        sh: kubectl get nodes --output=jsonpath='{.items[*].metadata.name}'

  upgrade-k8s:
    desc: Upgrade Kubernetes
    cmd: talosctl --nodes {{.KUBERNETES_CONTROLLER}} upgrade-k8s --to {{.KUBERNETES_VERSION}}
    vars:
      KUBERNETES_CONTROLLER:
        sh: talosctl config info --output json | jq --raw-output '.endpoints[]' | shuf -n 1
      KUBERNETES_VERSION:
        sh: yq '.kubernetesVersion' {{.TALHELPER_CONFIG_FILE}}
    preconditions:
      - msg: Missing talosconfig
        sh: test -f {{.TALOSCONFIG}}
      - msg: Unable to retrieve Talos config
        sh: talosctl config info &>/dev/null
      - msg: Node not found
        sh: talosctl --nodes {{.KUBERNETES_CONTROLLER}} get machineconfig &>/dev/null
      - msg: Upstream Kubernetes version not found
        sh: curl -fsSL -o /dev/null --fail https://github.com/siderolabs/kubelet/releases/tag/{{.KUBERNETES_VERSION}}

  reset:
    desc: Resets nodes back to maintenance mode
    dir: "{{.KUBERNETES_DIR}}/bootstrap/talos"
    prompt: This will destroy your cluster and reset the nodes back to maintenance mode... continue?
    cmd: talhelper gencommand reset --config-file {{.TALHELPER_CONFIG_FILE}} --out-dir {{.TALHELPER_CLUSTER_DIR}} --extra-flags="--reboot {{- if eq .CLI_FORCE false }} --system-labels-to-wipe STATE --system-labels-to-wipe EPHEMERAL{{ end }} --graceful=false --wait=false" | bash

  down:
    internal: true
    cmd: flux --namespace flux-system suspend kustomization --all

  up:
    internal: true
    cmd: flux --namespace flux-system resume kustomization --all

  # Ref: https://github.com/go-task/task/issues/608
  noop:
    internal: true
    silent: true
    cmd: noop() { :; }