fix(GRO-2652): reset authInitPromise on failure so retry loop actually retries

Previously the initAuth() function set authInitPromise to the async IIFE's
Promise but never cleared it on rejection. The index.ts retry loop called
initAuth() up to 10 times, but on attempt 2+ the check
`if (authInitPromise) { await authInitPromise; return; }` would immediately
re-throw the original rejection without performing a real retry.

Fix: wrap the final `await authInitPromise` in try/catch and reset
authInitPromise = null on error. This allows the index.ts retry loop to
create a fresh attempt on each call after a failure.

Co-Authored-By: Paperclip <noreply@paperclip.ing>

.gitea/workflows/ci.yml

@@ -102,7 +102,7 @@ jobs:
             git.farh.net/groombook/api:${{ steps.version.outputs.tag }}
             ${{ github.ref == 'refs/heads/main' && 'git.farh.net/groombook/api:latest' || '' }}
           cache-from: type=registry,ref=git.farh.net/groombook/cache:api
-          cache-to: type=registry,ref=git.farh.net/groombook/cache:api,mode=max
+          cache-to: type=registry,ref=git.farh.net/groombook/cache:api,mode=max,ignore-error=true
 
       - name: Build and push Migrate image
         uses: docker/build-push-action@v6
@@ -116,7 +116,7 @@ jobs:
             git.farh.net/groombook/migrate:${{ steps.version.outputs.tag }}
             ${{ github.ref == 'refs/heads/main' && 'git.farh.net/groombook/migrate:latest' || '' }}
           cache-from: type=registry,ref=git.farh.net/groombook/cache:migrate
-          cache-to: type=registry,ref=git.farh.net/groombook/cache:migrate,mode=max
+          cache-to: type=registry,ref=git.farh.net/groombook/cache:migrate,mode=max,ignore-error=true
 
       - name: Smoke test migrate image (blackhole npmjs.org)
         run: |
@@ -141,7 +141,7 @@ jobs:
             git.farh.net/groombook/seed:${{ steps.version.outputs.tag }}
             ${{ github.ref == 'refs/heads/main' && 'git.farh.net/groombook/seed:latest' || '' }}
           cache-from: type=registry,ref=git.farh.net/groombook/cache:seed
-          cache-to: type=registry,ref=git.farh.net/groombook/cache:seed,mode=max
+          cache-to: type=registry,ref=git.farh.net/groombook/cache:seed,mode=max,ignore-error=true
 
       - name: Build and push Reset image
         uses: docker/build-push-action@v6
@@ -155,7 +155,7 @@ jobs:
             git.farh.net/groombook/reset:${{ steps.version.outputs.tag }}
             ${{ github.ref == 'refs/heads/main' && 'git.farh.net/groombook/reset:latest' || '' }}
           cache-from: type=registry,ref=git.farh.net/groombook/cache:reset
-          cache-to: type=registry,ref=git.farh.net/groombook/cache:reset,mode=max
+          cache-to: type=registry,ref=git.farh.net/groombook/cache:reset,mode=max,ignore-error=true
 
       - name: Smoke test seed image (blackhole npmjs.org)
         run: |
@@ -185,3 +185,4 @@ jobs:
             "$IMAGE" \
             sh -c 'set -e; test "$(which pnpm)" = "/usr/local/bin/pnpm"; echo "HOME=$HOME"; pnpm --version'
           echo "reset image: pnpm resolves to /usr/local/bin/pnpm, HOME=/tmp, runs offline ✓"
+

UAT_PLAYBOOK.md

@@ -108,6 +108,11 @@ Expected: one row, `role = 'groomer'`. If zero rows return, the request hit the
 | TC-API-1.24 | Complete setup creates super user | POST /api/setup with business name (after TC-API-1.23) | First user becomes super user, setup completes | Setup errors, 403 on admin endpoints |
 | TC-API-1.25 | Super user accesses admin features | After TC-API-1.24, GET /api/staff/me and verify isSuperUser: true | isSuperUser: true, admin endpoints accessible | 403 on admin, isSuperUser: false |
 | TC-API-1.26 | Auto-provision skipped during OOBE | During fresh setup (needsSetup: true), complete OIDC login — verify no duplicate staff record created before setup completes | No duplicate staff, OOBE completes successfully | Duplicate staff record, 403 before setup, auto-provision interferes with OOBE |
+| TC-API-1.27 | Multi-origin CORS — demo host sign-in | `POST /api/auth/sign-in/social` with `callbackURL=https://demo.groombook.dev` | 200 OK, no origin-mismatch error | 400/403 "Origin mismatch" |
+| TC-API-1.28 | Multi-origin CORS — farh.net host sign-in | `POST /api/auth/sign-in/social` with `callbackURL=https://groombook.farh.net` | 200 OK, no origin-mismatch error | 400/403 "Origin mismatch" |
+| TC-API-1.29 | CORS — untrusted origin blocked (GRO-2586) | POST /api/auth/sign-in/social with `Origin: https://evil.example.com` header | Response has **no** `Access-Control-Allow-Origin` header — attacker origin is not reflected | `Access-Control-Allow-Origin: https://evil.example.com` present in response |
+| TC-API-1.30 | CORS — trusted origin allowed (GRO-2586) | POST /api/auth/sign-in/social with `Origin: https://uat.groombook.dev` header | `Access-Control-Allow-Origin: https://uat.groombook.dev` + `Access-Control-Allow-Credentials: true` | CORS header absent or trusted origin rejected |
+| TC-API-1.31 | CORS — untrusted preflight blocked (GRO-2586) | `curl -i -X OPTIONS https://uat.groombook.dev/api/auth/sign-in/social -H 'Origin: https://evil.example.com' -H 'Access-Control-Request-Method: POST'` | Response has **no** `Access-Control-Allow-Origin: https://evil.example.com` | Preflight reflects attacker origin |
 
 ### 4.2 Client Management
 
@@ -434,6 +439,38 @@ Both use the stops' stored `latitude`/`longitude` in `stopOrder`: **origin = fir
 | TC-API-18.10 | Groomer cannot export another's route | As groomer, export a route owned by a different groomer | 403 Forbidden (`groomers may only access their own route`) |
 | TC-API-18.11 | Receptionist denied | As **receptionist**, export any route | 403 Forbidden (role not permitted) |
 
+
+### 4.19 Boot Resilience — ECONNRESET Recovery (GRO-2652)
+
+Verifies the API process does not crash on transient boot-time DB connection resets and that auth routes degrade gracefully until initialization succeeds.
+
+| TC | Test Case | Steps | Expected Result |
+|----|-----------|-------|-----------------|
+| TC-API-19.1 | Health endpoint available before auth init | 1. Deploy the image (or restart the api pod)<br>2. `GET /health` immediately (within first 2 s of pod start) | 200 `{"status":"ok"}` — server accepts requests before `initAuth()` completes |
+| TC-API-19.2 | Auth routes return 503 when auth not yet initialized | 1. Temporarily set `OIDC_ISSUER` to an unreachable host so `initAuth()` keeps retrying<br>2. `POST /api/auth/sign-in/email` during the retry window | 503 `{"error":"Authentication not configured"}` — process stays alive, does not exit |
+| TC-API-19.3 | Pod does not crash on first-attempt DB reset | 1. Review pod restart count after normal deployment<br>2. Confirm `kubectl get pod -n groombook` shows `RESTARTS: 0` (or same as before deploy) for the new pod | No new restarts — ECONNRESET causes retry, not process exit |
+| TC-API-19.4 | DB query retry log lines visible | After deploy, `kubectl logs -n groombook <api-pod>` | If any DB retry occurred, log lines matching `[auth] DB query attempt N failed` are present; on clean boot no retry lines appear |
+| TC-API-19.5 | Auth init retry log lines visible | When auth init fails and retries, check pod logs | Log lines matching `[auth] initAuth attempt N failed` present; process continues; no `process.exit` |
+| TC-API-19.6 | Auth succeeds after transient DB hiccup | 1. Allow pod to retry until DB is available<br>2. `POST /api/auth/sign-in/email` with valid credentials after init succeeds | 200 with session cookie — auth recovers without pod restart |
+| TC-API-19.7 | Normal sign-in still works end-to-end | Follow TC-WEB-SSO-3 (SSO sign-in) on UAT | Successful sign-in, staff list visible — no regression from resilience changes |
+| TC-API-19.8 | Public routes unaffected during auth retry | While auth is retrying (TC-API-19.2 setup), `GET /api/branding` | 200 with branding data — public routes bypass auth and serve normally |
+
+### 4.20 Portal OOBE — Create Client from Auth (GRO-2359)
+
+Verifies the `POST /api/portal/clients-from-auth` endpoint that creates a new `clients` row for a first-time SSO user (out-of-box-experience registration). This endpoint requires a valid Better Auth session but does NOT require a portal session; it is the pre-portal step in the new-user OOBE flow.
+
+| TC | Test Case | Steps | Expected Result |
+|----|-----------|-------|-----------------|
+| TC-API-20.1 | Successful client creation | 1. Sign in via SSO to obtain a Better Auth session<br>2. `POST /api/portal/clients-from-auth` with `{ "name": "Test User" }` | 201 `{ "id": "<uuid>", "name": "Test User", "email": "<sso-email>" }` — new `clients` row created |
+| TC-API-20.2 | All optional fields accepted | `POST /api/portal/clients-from-auth` with `{ "name": "Test User", "phone": "555-1234", "address": "1 Main St", "notes": "VIP" }` (authenticated) | 201 with `id`, `name`, `email`; row in DB has all four fields |
+| TC-API-20.3 | Invalid body — missing name | `POST /api/portal/clients-from-auth` with `{}` (authenticated) | 400 (Zod validation failure); no row created |
+| TC-API-20.4 | Invalid body — empty name | `POST /api/portal/clients-from-auth` with `{ "name": "" }` (authenticated) | 400 — name must be at least 1 character |
+| TC-API-20.5 | No session — 401 | `POST /api/portal/clients-from-auth` with a valid body but **no** Better Auth session cookie | 401 `{ "error": "Unauthorized" }` |
+| TC-API-20.6 | Existing email — 409 | 1. Create a client row whose email matches the signed-in SSO user's email<br>2. `POST /api/portal/clients-from-auth` as that user | 409 `{ "error": "A customer record with this email already exists" }` — no duplicate row |
+| TC-API-20.7 | Auth not configured — 503 | Temporarily disable auth (e.g., point `OIDC_ISSUER` to an invalid host) so `getAuth()` throws<br>2. `POST /api/portal/clients-from-auth` | 503 `{ "error": "Authentication not configured" }` — graceful degradation |
+| TC-API-20.8 | Concurrent insert race — 409 | Simulate two near-simultaneous requests from the same SSO user (before any row exists) | At most one request returns 201; the other returns 409 — no duplicate row, no 500 |
+
+
 ## Pass/Fail Criteria
 
 **Pass:**
@@ -455,3 +492,4 @@ Both use the stops' stored `latitude`/`longitude` in `stopOrder`: **origin = fir
 ## Update Policy
 
 Any PR that changes user-facing behaviour MUST update this file. Test cases must be added, modified, or removed to reflect the new behaviour. The PR description must reference which playbook section was updated (e.g., "Updated UAT_PLAYBOOK.md §4.4 — new appointment rescheduling flow").
+

src/__tests__/authCors.test.ts

@@ -0,0 +1,60 @@
+import { describe, it, expect } from "vitest";
+import { enforceAuthCors } from "../lib/auth-cors.js";
+
+const TRUSTED = ["https://uat.groombook.dev", "https://dev.groombook.dev"];
+
+/** Simulates Better Auth reflecting the request Origin (the pre-fix bug). */
+function makeReflectedResponse(origin: string | null): Response {
+  return new Response('{"ok":true}', {
+    status: 200,
+    headers: {
+      "Content-Type": "application/json",
+      ...(origin
+        ? {
+            "Access-Control-Allow-Origin": origin,
+            "Access-Control-Allow-Credentials": "true",
+          }
+        : {}),
+    },
+  });
+}
+
+describe("enforceAuthCors (GRO-2586)", () => {
+  it("passes trusted origin through with credentials", () => {
+    const origin = "https://uat.groombook.dev";
+    const res = enforceAuthCors(origin, TRUSTED, makeReflectedResponse(origin));
+    expect(res.headers.get("Access-Control-Allow-Origin")).toBe(origin);
+    expect(res.headers.get("Access-Control-Allow-Credentials")).toBe("true");
+  });
+
+  it("strips ACAO for attacker origin (credentialed cross-origin read blocked)", () => {
+    const origin = "https://evil.example.com";
+    const res = enforceAuthCors(origin, TRUSTED, makeReflectedResponse(origin));
+    expect(res.headers.get("Access-Control-Allow-Origin")).toBeNull();
+    expect(res.headers.get("Access-Control-Allow-Credentials")).toBeNull();
+  });
+
+  it("strips ACAO when no Origin header (undefined)", () => {
+    const res = enforceAuthCors(undefined, TRUSTED, makeReflectedResponse(null));
+    expect(res.headers.get("Access-Control-Allow-Origin")).toBeNull();
+    expect(res.headers.get("Access-Control-Allow-Credentials")).toBeNull();
+  });
+
+  it("preserves non-CORS response headers and status from Better Auth", () => {
+    const origin = "https://evil.example.com";
+    const res = enforceAuthCors(origin, TRUSTED, makeReflectedResponse(origin));
+    expect(res.headers.get("Content-Type")).toBe("application/json");
+    expect(res.status).toBe(200);
+  });
+
+  it("second trusted origin is also allowed", () => {
+    const origin = "https://dev.groombook.dev";
+    const res = enforceAuthCors(origin, TRUSTED, makeReflectedResponse(origin));
+    expect(res.headers.get("Access-Control-Allow-Origin")).toBe(origin);
+  });
+
+  it("empty string origin is treated as untrusted", () => {
+    const res = enforceAuthCors("", TRUSTED, makeReflectedResponse(""));
+    expect(res.headers.get("Access-Control-Allow-Origin")).toBeNull();
+  });
+});

src/index.ts

@@ -3,6 +3,7 @@ import { Hono } from "hono";
 import { logger } from "hono/logger";
 import { cors } from "hono/cors";
 import { getAuth, initAuth, getActiveProviders } from "./lib/auth.js";
+import { enforceAuthCors } from "./lib/auth-cors.js";
 import { clientsRouter } from "./routes/clients.js";
 import { petsRouter } from "./routes/pets.js";
 import { servicesRouter } from "./routes/services.js";
@@ -200,9 +201,10 @@ api.use("*", resolveStaffMiddleware);
 // Better-Auth handler — mounted as sub-app to handle all /api/auth/* routes
 // authMiddleware and resolveStaffMiddleware both skip /api/auth/ paths
 const authRouter = new Hono();
-authRouter.all("/*", (c) => {
+authRouter.all("/*", async (c) => {
   try {
-    return getAuth().handler(c.req.raw);
+    const res = await getAuth().handler(c.req.raw);
+    return enforceAuthCors(c.req.header("origin"), TRUSTED_ORIGINS, res);
   } catch {
     return c.json({ error: "Authentication not configured" }, 503);
   }
@@ -290,14 +292,34 @@ api.route("/search", searchRouter);
 api.route("/buffer-rules", bufferRulesRouter);
 api.route("/routes", routesRouter);
 
+// Start the HTTP server first so /health and public routes are available immediately.
+// Auth initialization runs afterward with retry — a transient DB ECONNRESET at boot
+// must not crash the process (GRO-2652). Auth routes return 503 until initAuth succeeds.
 const port = Number(process.env.PORT ?? 3000);
-await initAuth();
-console.log(`API server listening on port ${port}`);
 const server = serve({ fetch: app.fetch, port });
+console.log(`API server listening on port ${port}`);
 
 // Start background reminder scheduler (runs every minute to check for upcoming appointments)
 startReminderScheduler();
 
+let initAttempt = 0;
+while (true) {
+  try {
+    await initAuth();
+    break;
+  } catch (err) {
+    initAttempt++;
+    const delay = Math.min(2 ** initAttempt * 500, 30_000);
+    console.error(`[auth] initAuth attempt ${initAttempt} failed: ${err}`);
+    if (initAttempt >= 10) {
+      console.error("[auth] auth init permanently failed — auth endpoints will serve 503");
+      break;
+    }
+    console.error(`[auth] retrying in ${delay}ms`);
+    await new Promise((r) => setTimeout(r, delay));
+  }
+}
+
 function shutdown() {
   console.log("Shutting down gracefully...");
   // SIGTERM/SIGINT → server.close() → callback → process.exit(0)

src/lib/auth-cors.ts

@@ -0,0 +1,22 @@
+/**
+ * Enforces the trusted-origins CORS allowlist on a raw Response from Better Auth.
+ * Better Auth reflects the request Origin into Access-Control-Allow-Origin
+ * regardless of the trustedOrigins config, allowing credentialed cross-origin reads
+ * from arbitrary attacker origins. This wrapper strips CORS headers for any origin
+ * not in the allowlist. (GRO-2586)
+ */
+export function enforceAuthCors(
+  requestOrigin: string | undefined,
+  trustedOrigins: string[],
+  res: Response
+): Response {
+  const headers = new Headers(res.headers);
+  if (requestOrigin && trustedOrigins.includes(requestOrigin)) {
+    headers.set("Access-Control-Allow-Origin", requestOrigin);
+    headers.set("Access-Control-Allow-Credentials", "true");
+  } else {
+    headers.delete("Access-Control-Allow-Origin");
+    headers.delete("Access-Control-Allow-Credentials");
+  }
+  return new Response(res.body, { status: res.status, statusText: res.statusText, headers });
+}

src/lib/auth.ts

@@ -118,18 +118,34 @@ export async function initAuth(): Promise<void> {
           updateAge: 60 * 60 * 24,
           cookieCache: { enabled: false },
         },
-        trustedOrigins: [process.env.CORS_ORIGIN ?? "http://localhost:5173"],
+        trustedOrigins: (process.env.CORS_ORIGIN ?? "http://localhost:5173")
+          .split(",").map((s) => s.trim()).filter(Boolean),
       });
       return;
     }
 
-    // Step 1: Try to load config from DB
+    // Step 1: Try to load config from DB, with retry-with-backoff for transient ECONNRESET (GRO-2652).
+    // A single connection reset during boot must not abort initialization.
     const db = getDb();
-    const [dbConfig] = await db
-      .select()
-      .from(authProviderConfig)
-      .where(eq(authProviderConfig.enabled, true))
-      .limit(1);
+    let dbQueryRows: (typeof authProviderConfig.$inferSelect)[] = [];
+    let dbAttempt = 0;
+    while (true) {
+      try {
+        dbQueryRows = await db
+          .select()
+          .from(authProviderConfig)
+          .where(eq(authProviderConfig.enabled, true))
+          .limit(1);
+        break;
+      } catch (err) {
+        dbAttempt++;
+        if (dbAttempt >= 5) throw err;
+        const delay = Math.min(1000 * 2 ** (dbAttempt - 1), 8_000);
+        console.warn(`[auth] DB query attempt ${dbAttempt} failed (${err}), retrying in ${delay}ms`);
+        await new Promise((r) => setTimeout(r, delay));
+      }
+    }
+    const [dbConfig] = dbQueryRows;
 
     let providerConfig: {
       providerId: string;
@@ -308,9 +324,15 @@ export async function initAuth(): Promise<void> {
           maxAge: 5 * 60, // 5 minutes
         },
       },
-      trustedOrigins: [process.env.CORS_ORIGIN ?? "http://localhost:5173"],
+      trustedOrigins: (process.env.CORS_ORIGIN ?? "http://localhost:5173")
+        .split(",").map((s) => s.trim()).filter(Boolean),
     });
   })();
 
-  await authInitPromise;
+  try {
+    await authInitPromise;
+  } catch (err) {
+    authInitPromise = null; // allow retry on next call
+    throw err;
+  }
 }