From 1a55167234d3c398a9611b6d8ccccd9ba36975a7 Mon Sep 17 00:00:00 2001 From: ZdenekSrotyr Date: Tue, 21 Apr 2026 20:24:31 +0200 Subject: [PATCH] docs: workflow-driven VM recreate for startup-script propagation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - ONBOARDING.md: replace 'propagating module changes' section with two explicit options — workflow_dispatch with recreate_targets (recommended, CI audit trail), or local terraform apply -replace (emergency). Adds a 'do not' section banning manual .env edits on VMs. - deployment-log.md: iteration 4 summary (version badge + module v1.5.0 + workflow_dispatch). --- docs/ONBOARDING.md | 37 +++++++++++-------- .../plans/2026-04-21-deployment-log.md | 7 ++++ 2 files changed, 29 insertions(+), 15 deletions(-) diff --git a/docs/ONBOARDING.md b/docs/ONBOARDING.md index c01368f..54f5b36 100644 --- a/docs/ONBOARDING.md +++ b/docs/ONBOARDING.md @@ -182,29 +182,36 @@ curl -X POST "http://$PROD_IP:8000/api/sync/trigger" \ ## Propagating module (startup-script) changes -**Important gotcha:** The `customer-instance` module has `lifecycle { ignore_changes = [metadata_startup_script] }` on VMs — this is intentional so `terraform apply` doesn't reboot VMs on every rerun. The consequence is that **changes inside the startup script are not picked up on a normal `terraform apply`**. +**Important gotcha:** The `customer-instance` module has `lifecycle { ignore_changes = [metadata_startup_script] }` on VMs — intentional, so `terraform apply` doesn't reboot VMs on every rerun. The consequence is that **startup-script changes are not picked up on a normal `terraform apply`**. -To propagate a startup-script change (for example, after bumping `ref=infra-v1.3.0`): +After bumping the module ref (e.g. `ref=infra-v1.5.0` → `infra-v1.6.0`), do one of: + +### Option A — Workflow dispatch with `recreate_targets` (recommended) + +`apply.yml` has a `workflow_dispatch` input `recreate_targets` that takes a comma-separated list of TF resource addresses and passes each as `-replace=` to `terraform apply`. Use this to destroy + recreate VMs with the new startup script, without any SSH. + +``` +Actions → Terraform Apply → Run workflow → recreate_targets: + module.agnes.google_compute_instance.vm["agnes-dev"],module.agnes.google_compute_instance.vm["agnes-prod"] +``` + +The workflow routes dev targets to `apply-dev` and prod targets to `apply-prod`, so the usual dev-first + prod-reviewer gate still applies. Persistent data disks and static IPs are separate resources and are **preserved** across replacement — only the VM (and its fresh boot disk) is recreated. + +Downtime: ~2 min per VM, sequential. Data loss: none (persistent disk keeps `/data`; static IP keeps URL stable). + +### Option B — Local terraform (emergency) ```bash -# VM is recreated; boot disk is fresh; persistent data disk is preserved +export GOOGLE_APPLICATION_CREDENTIALS=~/.agnes-keys/agnes-deploy--key.json +cd terraform terraform apply -replace='module.agnes.google_compute_instance.vm["agnes-prod"]' ``` -Downtime: ~2 minutes. The persistent data disk (where `/data` lives) is *not* recreated — only the VM. Startup script re-runs on the new VM with the latest template content, and your data is still there. +Same semantics as Option A, but no CI audit trail. Use only when CI is broken. -Alternative (less disruptive): hot-patch the VM via SSH: +### Do NOT -```bash -gcloud compute ssh agnes-prod --zone=... --project=... --command="sudo bash -c ' - cd /opt/agnes - curl -fsSL https://raw.githubusercontent.com/keboola/agnes-the-ai-analyst/main/docker-compose.prod.yml -o docker-compose.prod.yml - curl -fsSL https://raw.githubusercontent.com/keboola/agnes-the-ai-analyst/main/docker-compose.host-mount.yml -o docker-compose.host-mount.yml - docker compose -f docker-compose.yml -f docker-compose.prod.yml -f docker-compose.host-mount.yml up -d -'" -``` - -This preserves container state but won't re-install cron / rebuild persistent disk layout. +Do not manually edit `/opt/agnes/.env` or the docker-compose overlay files on a running VM. Any such change is lost on the next VM recreate, and it drifts from Terraform state. If a value needs changing, route it through a module variable or a module upgrade. ## Restoring from backup diff --git a/docs/superpowers/plans/2026-04-21-deployment-log.md b/docs/superpowers/plans/2026-04-21-deployment-log.md index 95db796..cf1bec3 100644 --- a/docs/superpowers/plans/2026-04-21-deployment-log.md +++ b/docs/superpowers/plans/2026-04-21-deployment-log.md @@ -157,6 +157,13 @@ Migrace dat zkopírovala users table, takže heslo je platné i na novém prod. 9. **Auth v2 → v3 action bump** v obou workflow repech (silences Node 20 deprecation warning). 10. **Prod apply-dev úspěšně proběhl** po manuálním triggeru (initial apply měl race s timing secret creation). apply-prod čeká na reviewera. +### Iterace 4 — version badge + workflow-driven recreate + +1. **Version badge v UI**: `/api/version` endpoint + footer badge v `base.html` loadí asynchronně, zobrazuje `- · · deployed ()` s commit SHA v tooltipu. +2. **Module `infra-v1.5.0`**: startup-script odvozuje `AGNES_VERSION` a `RELEASE_CHANNEL` z image tagu (stable-YYYY.MM.N / dev-…) a zjistí `AGNES_COMMIT_SHA` z `docker pull` digest. Tyto vars jdou do `.env` → app čte → `/api/version` vrací → badge renderuje. +3. **`workflow_dispatch` s `recreate_targets`** v `apply.yml` (oba repa): manuálně spustitelný workflow input s comma-separated TF resource addresses → `-replace=` předáno `terraform apply`. Řeší `ignore_changes = [metadata_startup_script]` gotchu. Dev targets routed do `apply-dev`, prod do `apply-prod`. +4. **Dokumentace propagation přepsána** v `docs/ONBOARDING.md` — Option A (workflow_dispatch, recommended) vs Option B (local TF), plus explicit DO NOT sekce proti ručnímu SSH zásahu. + ### Iterace 3 — code review + bootstrap fix + doc sweep 1. **Code review** dispatched přes `superpowers:requesting-code-review` subagent. Nálezy: 7 critical, 9 important, 10 minor.