Commit Graph
8 Commits
Author SHA1 Message Date
jakman85andClaude Opus 4.8 4cc04147a2 fix(k3s): disable local-storage so Longhorn is the sole default StorageClass
The deployed cluster has TWO default StorageClasses — k3s's built-in local-path
AND longhorn — which is ambiguous for PVCs without an explicit class. Longhorn
(ADR-0003) is the cluster storage, so add local-storage to k3s --disable. Effective
on fresh installs (the k3s install task is idempotent → no restart of the running
cluster); the running dev cluster keeps both defaults harmlessly (all platform PVCs
pin storageClassName: longhorn). Rebuilds get a single default.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-17 07:34:31 -05:00
jakman85andClaude Opus 4.8 95e3eafac6 iac: harden node lifecycle — tofu for_each + self-contained ansible rejoin
Implements T-1.5 findings F1/F2 (author-first, no cluster change).

F1 (tofu): refactor libvirt node resources count -> for_each keyed by
node name so single-node -replace/destroy cannot cascade to siblings.
Add ordered-list/primary helper locals; update outputs.tf off the map;
add moved.tf (12 moved blocks). tofu plan proves a pure state-address
migration: 0 to add, 0 to change, 0 to destroy (all 12 'has moved to').
State-persist via a 0-change apply is operator-permission-gated.

F2 (ansible): k3s_server secondary fetches the join token via
delegate_to the primary (published fact as fast path), so a single-node
rejoin with --limit k3s-dev-N is self-contained and no longer depends on
a fact an earlier play had to set. Syntax-check + CI gate green.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-16 19:42:12 -05:00
jakman85andClaude Opus 4.8 ab0a7977bf Phase 1 T-1.4 DONE: Flux reconciling clusters/dev from host Gitea
host-prereqs.yml opened the firewall (via the operator's ansible-vault become
password); Flux GitRepository READY=True (main@sha1:a2f3bdc). Commit the
encrypted hypervisor vault (ansible_become_password) so rebuild is fully
hands-off; passphrase stays out-of-repo. platform/apps still not reconciled.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-16 15:14:09 -05:00
jakman85andClaude Opus 4.8 4e2ca88bd4 Wire ansible-vault for host become password (unattended host-prereqs)
Where the sudo password goes, codified:
- inventory: new [hypervisor] group (localhost, connection=local).
- host-prereqs.yml targets 'hypervisor'; become password from vaulted
  group_vars/hypervisor/vault.yml (ansible_become_password) — user creates it
  encrypted (safe to commit); README documents the exact commands.
- ansible.cfg vault_password_file = ~/.config/home-k3s/ansible-vault-pass
  (passphrase kept OUTSIDE the repo, never committed).
- ci/yamllint.yaml ignores **/vault.yml (vault blob isn't parseable YAML).
Nodes keep NOPASSWD sudo (group_vars/all) — vault scoped to the host only.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-16 15:04:44 -05:00
jakman85andClaude Opus 4.8 3df5071bd1 Codify host firewall + Gitea bootstrap as IaC (hands-off rebuild)
Per operator: the firewalld rule (VM->host:3000 for the Flux<-Gitea source) must be
IaC, not a manual command (R-ENV-01, R-PLAT-04).
- infra/ansible/playbooks/host-prereqs.yml: idempotent firewalld open on the libvirt
  zone (localhost play, become; no extra collection).
- bootstrap/gitea/bootstrap-gitea.sh: now fully automates admin + public repo + push
  (was manual instructions) — idempotent; creds stay 0600 on host, never committed.
Applying host-prereqs still needs host sudo (I lack non-interactive sudo); with
passwordless become a destroy/rebuild is fully unattended.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-16 14:52:08 -05:00
jakman85andClaude Opus 4.8 125ca428bc fix(phase1): two execution-surfaced bugs bringing up k3s+Cilium
- ansible.cfg: stdout_callback 'yaml' (community.general.yaml) was REMOVED in
  ansible-core 2.20 -> use builtin default + result_format=yaml.
- cilium-values: egressGateway.enabled requires ipv4+eBPF masquerade or the
  agent fatals at startup ('egress gateway requires --enable-ipv4-masquerade
  and --enable-bpf-masquerade'). Add enableIPv4Masquerade + bpf.masquerade=true
  (also the correct pairing with kubeProxyReplacement; [parity] w/ platform/cilium).

Verified live: 3/3 nodes Ready, embedded-etcd HA, Cilium status OK, CoreDNS up.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-16 02:04:50 -05:00
jakman85andClaude Opus 4.8 f2a7e42068 Phase 1 (author-only): node provisioning IaC — tofu + cloud-init + Ansible (T-1.1/1.2/1.3)
OpenTofu libvirt module (k3s-dev NAT 10.73.0.0/24, 3x 6GiB/4vCPU VMs, 20G OS +
40G thin data, host-passthrough), cloud-init baseline (k3s/Cilium/Longhorn
prereqs, static net), Ansible k3s-HA + Cilium (kube-proxy replacement, Gateway
API, Hubble; flannel/kube-proxy/Traefik disabled). PM fix: escape ${...} in a
cloud-init template comment (tofu templatefile parse). tofu fmt/init/validate =
Success; ansible --syntax-check + yamllint pass. NOTHING executed (no apply).

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-16 01:21:00 -05:00
jakman85andClaude Opus 4.8 5ba9238cc1 Scope Phase 1: dev/test cluster bring-up on local KVM
Add infra/PHASE1-SCOPE.md (grounded in local recon: 24c/22GiB-avail/137G,
AMD-V+libvirt ready; OpenTofu/age/sops/flux missing) + infra/ and bootstrap/
structure stubs. Maps T-1.1..T-1.5; resource/network plan; risks; acceptance
gates; operator decisions. No spec edits; gate green.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
2026-08-16 00:48:48 -05:00