Pins actions/setup-node to v3, skips DB backup when no database exists yet

actions/setup-node@v4's Post step attempts a cache-save that fails hard on
this self-hosted Gitea Actions runner (no GHES cache API support), marking
frontend-prepare as failed and blocking its dependents even though the
actual install/build work succeeds. v3's older cache implementation doesn't
hit this. Also cleans up two stale comments left over from the /admin
MSBuild-target removal.

Also updates the documented backup script to skip cleanly (exit 0) when the
target database doesn't exist yet (MariaDB error 1049) rather than failing
the whole deploy - expected on a brand-new environment's first production
run, before migrations have ever had a chance to create it. Still fails hard
on any other error, so a real backup failure against an existing database
still blocks the deploy as intended.

.gitea/workflows/continuous_integration.yaml

@@ -109,7 +109,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/setup-node@v3
         with:
           node-version: ${{ env.NODE_VERSION }}
       - uses: pnpm/action-setup@v4
@@ -130,16 +130,16 @@ jobs:
 
   # --- Gate 4: frontend build ---
   # Validates the frontend compiles on its own, independent of any environment. The actual
-  # per-environment deployable artifact is produced later by `dotnet publish` (publish-test /
+  # per-environment deployable artifact is produced later by each publish job's own "Build admin
-  # publish-production jobs below), which triggers the same frontend build internally via
+  # frontend" step (publish-test / publish-production, below), which builds frontend/ again with
-  # SlpModularCms.Api.csproj's BuildAndCopyAdminFrontend MSBuild target, with the environment's
+  # that environment's Vite variables set and copies the result into wwwroot/admin before
-  # Vite variables set on the process so that target picks them up.
+  # `dotnet publish` runs — see that step's comment for why this isn't an MSBuild target instead.
   frontend-build:
     needs: frontend-prepare
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/setup-node@v3
         with:
           node-version: ${{ env.NODE_VERSION }}
       - uses: pnpm/action-setup@v4
@@ -167,7 +167,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/setup-node@v3
         with:
           node-version: ${{ env.NODE_VERSION }}
       - uses: pnpm/action-setup@v4
@@ -195,7 +195,7 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - uses: actions/setup-node@v3
         with:
           node-version: ${{ env.NODE_VERSION }}
       - uses: pnpm/action-setup@v4
@@ -217,10 +217,8 @@ jobs:
         working-directory: frontend
         run: pnpm run lint
 
-  # Publishes the Api project for the test environment. The admin SPA is built as part of this
+  # Publishes the Api project for the test environment. The admin SPA is built by this job's own
-  # publish (BuildAndCopyAdminFrontend, BeforeTargets="Publish"), so the VITE_* variables below are
+  # "Build admin frontend" step below, ahead of `dotnet publish` — see that step's comment for why.
-  # set on this step's environment, not passed as CLI arguments — pnpm build (invoked by MSBuild)
-  # reads them the same way `vite build` always does.
   #
   # REF-U5-01: the Umami-origin drift check this gate performs exists because U3's own startup check
   # (BR-U3-22) can't work — the backend can never see VITE_UMAMI_WEBSITE_ID at runtime. It compares
@@ -236,7 +234,7 @@ jobs:
       - uses: actions/setup-dotnet@v4
         with:
           dotnet-version: ${{ env.DOTNET_VERSION }}
-      - uses: actions/setup-node@v4
+      - uses: actions/setup-node@v3
         with:
           node-version: ${{ env.NODE_VERSION }}
       - uses: pnpm/action-setup@v4
@@ -299,7 +297,7 @@ jobs:
       - uses: actions/setup-dotnet@v4
         with:
           dotnet-version: ${{ env.DOTNET_VERSION }}
-      - uses: actions/setup-node@v4
+      - uses: actions/setup-node@v3
         with:
           node-version: ${{ env.NODE_VERSION }}
       - uses: pnpm/action-setup@v4

aidlc-docs/features/gitea-deployment-workflow/operations/deployment/deployment-instructions.md

@@ -556,6 +556,21 @@ fi
 source "$CREDENTIALS_FILE"
 : "${DB_HOST:?}" "${DB_PORT:?}" "${DB_NAME:?}" "${DB_USER:?}" "${DB_PASSWORD:?}"
 
+# A brand-new environment's very first production deploy runs this step before the app has ever
+# started, so before EF Core's migrations have had a chance to create the database — there is
+# nothing to back up yet, and that's fine, not a failure. Distinguish that specific case ("Unknown
+# database", MariaDB error 1049) from every other failure (wrong credentials, network issue,
+# permissions): only the former is safe to skip, since skipping any OTHER error would silently mask
+# a real backup failure against a database that may hold real data.
+USE_ERROR=$(mariadb -h "$DB_HOST" -P "$DB_PORT" -u "$DB_USER" -p"$DB_PASSWORD" -e "USE \`$DB_NAME\`;" 2>&1 >/dev/null) || true
+if echo "$USE_ERROR" | grep -q "Unknown database"; then
+  echo "Database '$DB_NAME' does not exist yet — nothing to back up (expected on a first deploy). Skipping."
+  exit 0
+elif [[ -n "$USE_ERROR" ]]; then
+  echo "Could not verify database '$DB_NAME' exists: $USE_ERROR" >&2
+  exit 1
+fi
+
 BACKUP_DIR="$HOME/backups/slpsoftware/${ENVIRONMENT}"
 mkdir -p "$BACKUP_DIR"
 TIMESTAMP=$(date -u +%Y%m%d%H%M%S)