Deployment setup - the parts the workflow deliberately left out

Host setup checklist, real domains and ports, the database backup script the deploy workflow only ever invokes, and a rollback plan. Also drafted the FTPS switch procedure for whenever production moves off the Pi, with a note that shared hosting is likely IIS-based - that's a bigger change than swapping the transport.
2026-07-28 19:07:26 +02:00
parent 9ab30fe2a6
commit cf80f827ae
6 changed files with 550 additions and 0 deletions
@@ -38,6 +38,17 @@
 - **Include Operations Phase**: Yes
 - **Decided At**: Requirements Analysis

+### Deployment Setup
+- **Included**: Yes
+- **Method**: CI/CD (Gitea Actions — confirms what U5/U6 already built)
+- **Completed**: 2026-07-28. Real domains: `test.slpsoftware.nl` / `slpsoftware.nl`. Ports 5100/5101
+  chosen for the Pi's local Kestrel bindings. Database backup script drafted. Full FTPS future-switch
+  procedure drafted (Q5 = B), with an explicit caveat that shared hosting is very likely IIS-based,
+  so systemd-restart and atomic-symlink-switch do not carry over unchanged — treated as a starting
+  brief for a future Infrastructure Design pass, not a ready-to-execute procedure
+- **Artifacts**: `operations/deployment/deployment-plan.md`, `deployment-instructions.md`,
+  `rollback-plan.md`
+
 ## Scope Decisions (from feature-selection.md)
 - **Public website**: documentation/instructions only — where the website build lands in `wwwroot/`, how it coexists with `wwwroot/admin/`, and what a per-website workspace must deliver. The website's own build/deploy workflow stays out of scope (Q4 = A).
 - **Environments**: local, test, production only.
@@ -1212,3 +1212,48 @@ by Infrastructure Design. U5 carries REF-U3-01's Umami origin gate plus FR-21 an
 **Operations Phase Decision**: Yes (Decided At: Requirements Analysis)

 ---
+## Build and Test — Local Debug Crash Reported and Fixed
+
+**Timestamp**: 2026-07-28T03:10:00Z
+
+**User Input**: "Als ik de Master api start krijg ik de volgende error: K:/Development/Projects/SlpModularCms/src/SlpModularCms.Api/bin/Debug/net10.0/SlpModularCms.Api.exe Unhandled exception. System.IO.DirectoryNotFoundException: K:\Development\Projects\SlpModularCms\src\SlpModularCms.Api\wwwroot\ [full stack trace through WebApplication.CreateBuilder] ... Bij het starten van de slave krijg ik dat niet. Dit is lokaal"
+
+**AI Response**: "Root cause identified: not a defect in the feature's code, but a leftover of my own Build and Test verification — dotnet publish had created wwwroot/ (triggering ASP.NET Core's static web assets manifest to reference it), then my cleanup deleted the directory without a clean rebuild, leaving a stale Debug-config static-web-assets manifest pointing at a now-missing path. Fixed by fully cleaning and rebuilding SlpModularCms.Api's Debug obj/bin (dotnet clean + manual rm -rf + dotnet build). Confirmed the Master host starts cleanly again (/health -> 200). Nothing was tracked in git (gitignored build output), so no commit was needed."
+
+**Context**: Build and Test — post-completion issue found by the user, diagnosed and fixed before proceeding
+
+---
+
+## Build and Test (Confirmed) — Proceeding to Deployment Setup
+
+**Timestamp**: 2026-07-28T03:15:00Z
+
+**User Input**: "Werkt weer, ga verder"
+
+**AI Response**: "Confirmed fixed. Proceeding to the Operations phase, starting with Deployment Setup."
+
+**Context**: Build and Test stage fully closed; entering OPERATIONS phase
+
+---
+## Deployment Setup (Planning)
+
+**Timestamp**: 2026-07-28T03:20:00Z
+
+**User Input**: "(implicit — entering Operations phase)"
+
+**AI Response**: "Created deployment-setup-plan.md with 6 questions. Q1/Q2 (inclusion + method) are confirmations of what was already built in Construction, not new decisions. Q3-Q6 are the genuinely open host-specific facts that infrastructure-design.md deliberately deferred to this stage: real domains, whether to draft the DB backup script, FTPS documentation depth, and the one-time host-setup checklist scope."
+
+**Context**: OPERATIONS — Deployment Setup, Step 2-4 (question generation), awaiting user answers
+
+---
+## Deployment Setup (Complete)
+
+**Timestamp**: 2026-07-28T03:30:00Z
+
+**User Input**: "Ik heb de vragen ingevuld, ga verder" (Q1=A, Q2=A, Q3=A with production=slpsoftware.nl/test=test.slpsoftware.nl, Q4=A, Q5=B, Q6=A)
+
+**AI Response**: "Generated deployment-plan.md, deployment-instructions.md (one-time host setup: enable-linger, directory skeleton, shared/env runtime config file, systemd --user units for both environments on ports 5100/5101, nginx routing snippet, Gitea variables/secrets table, database backup script drafted and documented, full FTPS future-switch procedure with an explicit caveat that shared hosting is likely IIS-based and several mechanisms don't carry over unchanged), and rollback-plan.md (fast symlink-repoint rollback, rebuild-and-redeploy for older commits, database restore, explicit list of what a rollback never touches)."
+
+**Context**: OPERATIONS — Deployment Setup complete, presenting completion message
+
+---
@@ -0,0 +1,268 @@
+# Deployment Instructions
+
+**Target**: single Raspberry Pi (`linux-arm64`), test and production both on it, separated by
+directory, `systemd --user` unit, and local port. Domains: `test.slpsoftware.nl` (test),
+`slpsoftware.nl` (production).
+
+---
+
+## 1. One-Time Host Setup
+
+Do this once, before the first automated deploy. Everything here is host configuration the workflow
+assumes already exists — `deploy-scp.yaml` (U6) never creates any of it.
+
+### 1.1 Prerequisites
+- .NET 10 runtime installed on the Pi (ASM-03) — the publish is framework-dependent
+  (`infrastructure-design.md` § 1), so the Pi needs the runtime, not the full SDK
+- SSH access for the deploy user, password auth for now (matches the reference project;
+  SSH-key migration remains a documented future step, same as there)
+- `sqlcmd` installed, for the backup script (§ 4) — e.g. `mssql-tools18` / `unixodbc` on Debian-based
+  Raspberry Pi OS
+
+### 1.2 Enable Lingering (INFRA-U6-01 — do this first, easy to forget)
+```bash
+sudo loginctl enable-linger <deploy-user>
+```
+Without this, the `systemd --user` service manager is torn down when the deploy SSH session ends,
+killing the just-restarted app a few seconds after every successful deploy. This also fixes the
+common "Failed to connect to bus" error `systemctl --user` can throw when invoked from a
+non-interactive SSH command — lingering keeps the user's systemd instance (and `XDG_RUNTIME_DIR`)
+running independent of any login session.
+
+### 1.3 Directory Skeleton
+Run once per environment (`test`, `production`):
+```bash
+mkdir -p ~/apps/slpmodularcms-<env>/releases
+mkdir -p ~/apps/slpmodularcms-<env>/shared/wwwroot-web
+```
+`current` is created by the first deploy itself (`ln -sfn`) — don't pre-create it.
+
+### 1.4 Runtime Configuration File
+One file per environment, **outside** the release directory so it survives every switch:
+```bash
+touch ~/apps/slpmodularcms-<env>/shared/env
+chmod 600 ~/apps/slpmodularcms-<env>/shared/env
+```
+Contents (fill in real values — this file is never read by the workflow, only by the systemd unit
+below):
+```ini
+ASPNETCORE_ENVIRONMENT=Production
+ASPNETCORE_URLS=http://localhost:<port>
+ConnectionStrings__DefaultConnection=Server=127.0.0.1,1433;User ID=<user>;Password=<password>;Database=SlpModularCms<Env>;TrustServerCertificate=True
+JwtSettings__Secret=<secure-long-random-secret>
+JwtSettings__Issuer=SlpModularCms
+JwtSettings__Audience=SlpModularCmsPortal
+MasterModule__MasterUrl=https://<this-environment-domain>
+Observability__SentryDsn=<sentry-dsn-or-empty>
+SecurityHeaders__AllowedScriptOrigins__0=<umami-script-origin>
+SecurityHeaders__AllowedConnectOrigins__0=<sentry-ingest-origin>
+```
+Use ports **5100** (test) and **5101** (production) unless something else on the Pi already
+occupies them. `ASPNETCORE_ENVIRONMENT=Production` is used for **both** environments deliberately —
+`Development` disables HSTS and exposes the Scalar API explorer (`Program.cs`), neither of which
+should be true for anything reachable at a real domain, including test.
+
+**`SecurityHeaders__AllowedScriptOrigins__0` / `_AllowedConnectOrigins__0` must exactly match** the
+Gitea variables `SECURITY_ALLOWED_SCRIPT_ORIGINS_TEST` / `_PRODUCTION` (see § 3) — REF-U5-01's CI
+gate only catches drift between the frontend build and that Gitea variable; it cannot see this file,
+so keeping the two in sync is a manual discipline, not something enforced automatically.
+
+### 1.5 systemd User Units
+Create `~/.config/systemd/user/slpmodularcms-test.service`:
+```ini
+[Unit]
+Description=SlpModularCms API (test)
+After=network.target
+
+[Service]
+WorkingDirectory=%h/apps/slpmodularcms-test/current
+ExecStart=/usr/bin/dotnet %h/apps/slpmodularcms-test/current/SlpModularCms.Api.dll
+EnvironmentFile=%h/apps/slpmodularcms-test/shared/env
+Restart=on-failure
+RestartSec=5
+KillSignal=SIGINT
+TimeoutStopSec=20
+
+[Install]
+WantedBy=default.target
+```
+And `~/.config/systemd/user/slpmodularcms-production.service` — identical, with `test` replaced by
+`production` throughout (including the port inside `shared/env`).
+
+Enable both (does not start them yet — nothing is deployed there until the first CI run):
+```bash
+systemctl --user daemon-reload
+systemctl --user enable slpmodularcms-test.service
+systemctl --user enable slpmodularcms-production.service
+```
+
+### 1.6 nginx Routing
+The existing reverse proxy (`infrastructure-design.md` § 1, § 4) needs a server block per domain,
+routing to the matching local port:
+```nginx
+server {
+    listen 443 ssl;
+    server_name test.slpsoftware.nl;
+    location / {
+        proxy_pass http://127.0.0.1:5100;
+        proxy_set_header Host $host;
+        proxy_set_header X-Forwarded-Proto $scheme;
+    }
+}
+
+server {
+    listen 443 ssl;
+    server_name slpsoftware.nl;
+    location / {
+        proxy_pass http://127.0.0.1:5101;
+        proxy_set_header Host $host;
+        proxy_set_header X-Forwarded-Proto $scheme;
+    }
+}
+```
+TLS certificate provisioning (e.g. certbot) is existing host operations, out of scope for this
+feature — assumed already handled the same way the reference project's domains are.
+
+### 1.7 Database Backup Credentials (§ 4 depends on this)
+```bash
+touch ~/.config/slpmodularcms-db-backup.env
+chmod 600 ~/.config/slpmodularcms-db-backup.env
+```
+```ini
+DB_SERVER=127.0.0.1,1433
+DB_NAME=SlpModularCmsProduction
+DB_USER=<a-login-with-backup-database-permission>
+DB_PASSWORD=<password>
+```
+Kept **separate** from `shared/env` (§ 1.4) deliberately — the backup script needs its own
+credential, ideally scoped to just `BACKUP DATABASE` permission rather than the application's own
+data-access login.
+
+### 1.8 Gitea Actions Variables and Secrets
+Set once, in this repository's Gitea Actions settings:
+
+| Name | Kind | Value |
+|---|---|---|
+| `PI_MAIN_ADDRESS` | secret | the Pi's address |
+| `PI_MAIN_PORT` | secret | SSH port |
+| `PI_MAIN_USERNAME` | secret | deploy user |
+| `PI_MAIN_PASSWORD` | secret | deploy user's password |
+| `DEPLOY_PATH_TEST` | variable | `/home/<user>/apps/slpmodularcms-test` |
+| `DEPLOY_PATH_PRODUCTION` | variable | `/home/<user>/apps/slpmodularcms-production` |
+| `SERVICE_NAME_TEST` | variable | `slpmodularcms-test.service` |
+| `SERVICE_NAME_PRODUCTION` | variable | `slpmodularcms-production.service` |
+| `HEALTH_CHECK_URL_TEST` | variable | `https://test.slpsoftware.nl/health` |
+| `HEALTH_CHECK_URL_PRODUCTION` | variable | `https://slpsoftware.nl/health` |
+| `VITE_SENTRY_DSN` | variable | Sentry DSN (shared, not sensitive — safe in the client bundle) |
+| `VITE_UMAMI_SCRIPT_URL` | variable | Umami script host (shared) |
+| `VITE_UMAMI_WEBSITE_ID_TEST` / `_PRODUCTION` | variable | per-environment Umami website ID |
+| `SECURITY_ALLOWED_SCRIPT_ORIGINS_TEST` / `_PRODUCTION` | variable | **must match** § 1.4's `SecurityHeaders__AllowedScriptOrigins__0` for that environment |
+
+---
+
+## 2. Deploy Sequence (What Actually Happens on a Run)
+
+Already built (U5/U6) — this is the read-only walkthrough for whoever operates it:
+
+1. Push to `master`, or a manual `workflow_dispatch` → the six gates run
+2. `publish-test` runs (and `publish-production`, only if `workflow_dispatch` with the flag)
+3. `deploy-test` (always, if gates pass) calls `deploy-scp.yaml`, which uploads into a new
+   `releases/{timestamp}/`, links `shared/wwwroot-web` in, switches `current`, restarts
+   `slpmodularcms-test.service`, verifies `https://test.slpsoftware.nl/health`, then prunes old
+   releases (only test — `run_db_backup: false`)
+4. `deploy-production` (only with the flag) does the same, plus a database backup first
+   (`run_db_backup: true`) — see § 4
+
+## 3. First-Ever Deploy Notes
+
+- `wwwroot/web/` will be empty until a website workspace deploys into it — `/` serves the built-in
+  placeholder until then (this is expected, not a failure)
+- The very first run has no "previous release" to keep — pruning naturally has nothing to prune
+- Verify manually after the first run: `curl https://test.slpsoftware.nl/health` and
+  `curl https://slpsoftware.nl/health` (after the first production run) both return `200`
+
+## 4. Database Backup Script
+
+Create `~/scripts/backup-slpmodularcms-db.sh` on the Pi (this script is host-side by design — never
+part of this repository, so no DB credential ever reaches Gitea):
+
+```bash
+#!/usr/bin/env bash
+set -euo pipefail
+
+ENVIRONMENT="${1:?Usage: backup-slpmodularcms-db.sh <environment>}"
+CREDENTIALS_FILE="$HOME/.config/slpmodularcms-db-backup.env"
+
+if [[ ! -f "$CREDENTIALS_FILE" ]]; then
+  echo "Missing $CREDENTIALS_FILE — see deployment-instructions.md § 1.7" >&2
+  exit 1
+fi
+# shellcheck source=/dev/null
+source "$CREDENTIALS_FILE"
+: "${DB_SERVER:?}" "${DB_NAME:?}" "${DB_USER:?}" "${DB_PASSWORD:?}"
+
+BACKUP_DIR="$HOME/backups/slpmodularcms-${ENVIRONMENT}"
+mkdir -p "$BACKUP_DIR"
+TIMESTAMP=$(date -u +%Y%m%d%H%M%S)
+BACKUP_FILE="$BACKUP_DIR/${DB_NAME}-${TIMESTAMP}.bak"
+
+sqlcmd -S "$DB_SERVER" -U "$DB_USER" -P "$DB_PASSWORD" -C -Q \
+  "BACKUP DATABASE [$DB_NAME] TO DISK = N'$BACKUP_FILE' WITH INIT, COMPRESSION"
+
+echo "Backup written to $BACKUP_FILE"
+
+# Retention: keep the 7 most recent backups for this environment
+ls -1t "$BACKUP_DIR"/*.bak 2>/dev/null | tail -n +8 | xargs -r rm -f
+```
+
+```bash
+chmod +x ~/scripts/backup-slpmodularcms-db.sh
+```
+
+**Verify once, manually**, before relying on it in a real deploy:
+```bash
+~/scripts/backup-slpmodularcms-db.sh production
+```
+Confirm a `.bak` file appears under `~/backups/slpmodularcms-production/` and that `sqlcmd` didn't
+silently fail (the script uses `set -euo pipefail`, so a real SQL error does propagate as a non-zero
+exit — which fails the calling `deploy-scp.yaml` step, correctly blocking the deploy).
+
+## 5. Future: Switching Production to FTPS (Shared Hosting)
+
+D-02/NFR-09 required the workflow's *transport* to be swappable without restructuring — satisfied by
+the `transport` input already on `deploy-scp.yaml` (currently only `scp` is implemented). This
+section documents what actually changes when that day comes (OPEN-04 — not scheduled, drafted now
+per Q5 = B of the deployment setup plan).
+
+### 5.1 What carries over unchanged
+- The CI workflow's gates, the two-build split, the `config` job pattern
+- The overall shape of the interface: `artifact_name`, `environment`, `deploy_path`
+
+### 5.2 What does not carry over — read this before assuming it's a drop-in swap
+Shared .NET hosting is almost always **Windows/IIS-based**, not Linux/systemd. That changes more
+than the transport:
+- **No `systemctl --user` restart** — IIS picks up a new deployment via an app-pool recycle, usually
+  triggered by touching `web.config` or the app-pool's own recycle mechanism, not a service restart
+  command
+- **The atomic release-switch pattern may not be available at all** — many shared hosts expose only
+  a single web root over FTPS, with no ability to create sibling directories and swap a symlink.
+  `wwwroot/web/` persistence (FR-08, ASM-01) would need a **different** mechanism on such a host —
+  e.g. never touching a specific subfolder during upload, rather than linking a persistent directory
+  outside a swapped release tree, since "outside the release tree" may not be an available concept
+- **Database backup** — shared hosting frequently does not expose direct `sqlcmd`/SSH access at all;
+  the backup step in `deploy-scp.yaml` (§ 4 script) would need to become either a provider-specific
+  API call or a documented manual pre-production step (the FR-20 fallback U6 already designed for)
+
+### 5.3 What building `deploy-ftps.yaml` would actually require
+1. A new reusable workflow implementing the **same five required inputs**
+   (`artifact_name`, `environment`, `deploy_path`, `service_name`, `health_check_url`) so
+   `continuous_integration.yaml` calls it identically via the `transport` input
+2. An FTPS upload step (e.g. `lftp` mirror or `curl --ftp-ssl`) replacing the `scp` step
+3. A **new decision, not yet made**: what "atomic switch" and "restart" even mean on the target
+   host — this cannot be answered generically; it depends on which shared host is actually chosen
+4. Re-running this feature's Infrastructure Design step for U6, scoped to the new host, once a
+   specific shared-hosting provider is selected — not a checklist item that can be pre-answered today
+
+**Recommendation**: treat this section as a starting brief for that future Infrastructure Design
+pass, not as a ready-to-execute procedure — the concrete host was unknown at the time this was
+written (OPEN-04), so several of the decisions above are necessarily provider-specific.
@@ -0,0 +1,57 @@
+# Deployment Plan
+
+**Date**: 2026-07-28
+**Decided at**: Deployment Setup, per `deployment-setup-plan.md` (Q1 = A, Q2 = A)
+
+## Method
+
+Gitea Actions CI/CD, already built in Construction:
+- `.gitea/workflows/continuous_integration.yaml` (U5) — six blocking gates, then a per-environment
+  `dotnet publish`, then invokes the deploy workflow
+- `.gitea/workflows/deploy-scp.yaml` (U6) — reusable `workflow_call` workflow: backup (production
+  only) → upload → link persistent website → atomic switch → restart → health check → prune
+
+This stage does not change that mechanism — it documents the operational side Construction
+deliberately left out: host setup, real domains, the database backup script, and the rollback
+procedure.
+
+## Environments
+
+| Environment | Host | Domain | Trigger |
+|---|---|---|---|
+| Test | Raspberry Pi (`linux-arm64`) | `test.slpsoftware.nl` | Automatic on push to `master`, or any `workflow_dispatch` |
+| Production | Same Pi, separate directory | `slpsoftware.nl` | Only `workflow_dispatch` with `deploy_production = true` |
+
+Same physical host for both (per `infrastructure-design.md` § 1) — separated by directory,
+`systemd --user` unit, and local port, never by anything the workflow manages directly.
+
+## Rationale for What's Documented Here vs. Already Decided
+
+| Already decided (Construction) | Documented here (Operations) |
+|---|---|
+| Atomic switch mechanism, retention count, transport | Real domains, real ports |
+| `deploy-scp.yaml`'s step sequence | The one-time host setup that sequence assumes already exists |
+| That a DB backup step runs before production deploys | The actual backup script's content |
+| That `transport` is reserved for FTPS (D-02, NFR-09) | What actually changes operationally when that day comes |
+
+## Automation Level
+
+Fully automated for test (push to `master` → live on `test.slpsoftware.nl` with no human step).
+Production requires an explicit, manual `workflow_dispatch` run with the flag checked — this manual
+trigger **is** the approval gate (D-09); no separate approval workflow step exists or is needed.
+
+## Rollback Strategy
+
+Summarized here; full procedure in `rollback-plan.md`. Two releases are always retained
+(`infrastructure-design.md` § 3), so rolling back is a re-point-and-restart, never a rebuild, for the
+most recent deploy. Migrations are required to be forward-compatible and non-destructive (D-26), so
+redeploying an older commit remains valid further back than just one release.
+
+## Secrets and Configuration
+
+Per D-16, runtime configuration lives in **host** environment variables that the workflow never
+writes — see `deployment-instructions.md` § "Runtime Configuration" for the full list and where it
+lives on the host (a single `shared/env` file per environment, outside the swapped release
+directory, `chmod 600`). Gitea Actions secrets are limited to what the workflow itself needs to
+reach the host: SSH credentials (`PI_MAIN_*`). No application secret is ever a Gitea secret or
+variable.
@@ -0,0 +1,65 @@
+# Rollback Plan
+
+**Mechanism** (D-26, `infrastructure-design.md` § 3–4): two releases are always retained —
+`current` plus exactly one previous. Pruning only ever runs after a **passing** health check, so a
+failed deploy never leaves fewer than two releases on disk.
+
+## When to Roll Back
+
+- The deploy workflow's health check failed (job already shows red; `current` was left pointing at
+  the new, unhealthy release — no automatic rollback, per `infrastructure-design.md` § 4)
+- The deploy succeeded and passed its health check, but a defect surfaces afterward that only shows
+  up under real traffic
+
+## Fast Rollback (Previous Release Still on Disk) — the Common Case
+
+No rebuild needed. Over SSH, on the Pi:
+
+```bash
+cd ~/apps/slpmodularcms-<env>/releases
+ls -1t                                    # confirm which directory is the previous, working release
+ln -sfn ~/apps/slpmodularcms-<env>/releases/<previous-timestamp> ~/apps/slpmodularcms-<env>/current
+systemctl --user restart slpmodularcms-<env>.service
+curl -f https://<env-domain>/health       # confirm the rollback itself is healthy
+```
+
+This is the same atomic-switch primitive the deploy workflow itself uses — pointing it backward
+instead of forward. `wwwroot/web/` is untouched either way, since it was never part of the switched
+directory to begin with (FR-08).
+
+## Rebuild-and-Redeploy Rollback (Older Than One Release Back)
+
+If the defect predates the retained previous release, redeploy an **earlier commit** through the
+normal pipeline:
+
+1. `git revert` or check out the last-known-good commit on a branch
+2. Push to `master` (test) or run `workflow_dispatch` with `deploy_production: true` (production) —
+   the same gates and deploy sequence run as any other deploy
+3. This only works because migrations are required to be forward-compatible and non-destructive
+   (D-26) — redeploying an older commit's code against a database that has since had newer
+   migrations applied must not break. If a migration since the target commit **was** destructive,
+   this path is not safe and the database backup (§ below) is the actual recovery route instead
+
+## Database Rollback
+
+A backup is taken before every **production** deploy (`operations/deployment/deployment-instructions.md`
+§ 4, invoked by `deploy-scp.yaml` when `run_db_backup: true`). To restore:
+
+```bash
+sqlcmd -S <server> -U <user> -P <password> -C -Q \
+  "RESTORE DATABASE [SlpModularCmsProduction] FROM DISK = N'<path-to-backup>.bak' WITH REPLACE"
+```
+
+Restoring a database backup and rolling back the application release are **independent actions** —
+decide based on the actual failure whether one, the other, or both are needed. Rolling back the app
+without restoring the database is usually sufficient (migrations are non-destructive by design);
+restoring the database without rolling back the app should be rare and deliberate.
+
+## What Is Never Part of a Rollback
+
+- `wwwroot/web/` (the customer's website) — structurally outside every release directory; no
+  rollback action should ever touch it
+- The Data Protection key ring — lives in the database, not the release directory; rolling back the
+  app release does not affect it
+- Gitea Actions variables/secrets — these describe the *target*, not a specific release; nothing
+  about them changes during a rollback
@@ -0,0 +1,104 @@
+# Deployment Setup Plan
+
+## Context already established (not re-asked)
+
+The deployment *mechanism* was already decided and built during Construction — this stage is about
+producing the operational documentation around it, not re-deciding the method:
+
+- **Method**: Gitea Actions CI/CD (`continuous_integration.yaml` → `deploy-scp.yaml`), built in U5/U6
+- **Target**: a single Raspberry Pi (`linux-arm64`), test and production on the same host, split by
+  directory only (`infrastructure-design.md` § 1)
+- **Transport**: SSH/SCP, `sshpass` + plain shell steps (D-05)
+- **Release strategy**: atomic release-directory switch, 2 releases retained, `systemd --user`
+  services, no sudo (`infrastructure-design.md` §§ 2–3)
+- **Rollback mechanism**: re-point the `current` symlink to the previous release and restart the
+  service — already fully specified, this stage just needs to write it up as a runnable procedure
+- **Approval gate for production**: the `workflow_dispatch` + `deploy_production` flag itself is the
+  approval gate (D-09) — no additional gate is needed
+
+## Question 1: Include Deployment Setup?
+
+Given all of the above was purpose-built in Construction specifically to be deployed, declining this
+stage would leave the pipeline built but undocumented for actual first use. Recommended: A.
+
+A) Yes — produce the deployment documentation (Recommended)
+B) No — deployment is handled elsewhere or not needed
+C) Not sure — suggest an approach and I'll decide
+
+X) Other (please describe after [Answer]: tag below)
+
+[Answer]:A
+
+## Question 2: Deployment Method
+
+Confirms the method already built — asked per the mandatory format, not because it's genuinely open.
+
+A) CI/CD pipeline (already built: Gitea Actions, this is a confirmation, not a new choice) (Recommended)
+B) Something else entirely — describe below
+
+X) Other (please describe after [Answer]: tag below)
+
+[Answer]: A
+
+---
+
+The remaining questions are the genuinely open items — real host facts that only you know, which
+`infrastructure-design.md` deliberately left for this stage.
+
+## Question 3: Real Domains
+
+`HEALTH_CHECK_URL_TEST` / `_PRODUCTION` and the nginx routing (`infrastructure-design.md` § 1) need
+real hostnames to write concrete instructions. What should the deployment instructions use?
+
+A) I'll provide the real domains now — describe them after [Answer]: below (e.g. `test.example.nl`,
+   `example.nl`)
+B) Use placeholders (`<test-domain>`, `<production-domain>`) in the generated instructions; I'll fill
+   them in myself when configuring Gitea variables
+
+X) Other (please describe after [Answer]: tag below)
+
+[Answer]:A, production: slpsoftware.nl, test: test.slpsoftware.nl
+
+## Question 4: Database Backup Script
+
+`deploy-scp.yaml` (U6) invokes `~/scripts/backup-slpmodularcms-db.sh` on the host but does not create
+it — it's host-side, deliberately kept out of the workflow so no DB credentials ever touch Gitea.
+Should this stage draft that script's content for you to place on the Pi?
+
+A) Yes, draft a `sqlcmd`/`BACKUP DATABASE` script now, parameterized by environment (Recommended)
+B) No, I'll write the backup script myself — just document what it must accept/do (arguments,
+   exit-code contract, where the workflow expects it)
+C) Not applicable yet — document it as a TODO, I'll come back to this before the first production deploy
+
+X) Other (please describe after [Answer]: tag below)
+
+[Answer]:A
+
+## Question 5: FTPS Switch Documentation Depth
+
+D-02 requires the *workflow* to allow adding FTPS later without restructuring (already satisfied by
+the `transport` input in `deploy-scp.yaml`) — but how much should the *documentation* say about
+actually switching to it when production eventually moves to shared hosting?
+
+A) Brief note only: confirm the extension point exists in the workflow, don't draft the FTPS steps
+   themselves (Recommended — shared hosting isn't happening yet, OPEN-04)
+B) Draft a full FTPS switch procedure now, even though it won't be used until shared hosting happens
+
+X) Other (please describe after [Answer]: tag below)
+
+[Answer]:B
+
+## Question 6: One-Time Host Setup Checklist
+
+Several one-time, host-side prerequisites were identified during Infrastructure Design and must be
+documented somewhere before the first real deploy can succeed:
+`loginctl enable-linger` (INFRA-U6-01), creating the `systemd --user` unit files, creating the
+`releases`/`shared`/`current` directory skeleton, and the Gitea secrets/variables from
+`infrastructure-design.md` § 6. Confirm this belongs in `deployment-instructions.md`?
+
+A) Yes, include a full one-time host-setup checklist in the deployment instructions (Recommended)
+B) No, I'll assemble the host setup myself from the Infrastructure Design doc directly
+
+X) Other (please describe after [Answer]: tag below)
+
+[Answer]:A