dragonflight/services/mam-api/src/index.js

314 lines
13 KiB
JavaScript
Raw Normal View History

2026-04-07 21:58:25 -04:00
import 'dotenv/config';
import express from 'express';
import cors from 'cors';
import session from 'express-session';
import ConnectPgSimple from 'connect-pg-simple';
import os from 'node:os';
import { exec } from 'node:child_process';
2026-04-07 21:58:25 -04:00
import pool from './db/pool.js';
import { errorHandler } from './middleware/errors.js';
import { loadS3ConfigFromDb } from './s3/client.js';
2026-04-07 21:58:25 -04:00
// Routes
import authRouter from './routes/auth.js';
2026-04-07 21:58:25 -04:00
import assetsRouter from './routes/assets.js';
import projectsRouter from './routes/projects.js';
import binsRouter from './routes/bins.js';
import jobsRouter from './routes/jobs.js';
import captureRouter from './routes/capture.js';
2026-04-07 22:05:39 -04:00
import uploadRouter from './routes/upload.js';
import recordersRouter from './routes/recorders.js';
import settingsRouter from './routes/settings.js';
import amppRouter from './routes/ampp.js';
import usersRouter from './routes/users.js';
import groupsRouter from './routes/groups.js';
import tokensRouter from './routes/tokens.js';
2026-05-18 19:54:41 -04:00
import sequencesRouter from './routes/sequences.js';
import systemRouter from './routes/system.js';
import clusterRouter from './routes/cluster.js';
import sdkRouter from './routes/sdk.js';
import schedulesRouter from './routes/schedules.js';
import metricsRouter from './routes/metrics.js';
import commentsRouter from './routes/comments.js';
import importsRouter from './routes/imports.js';
import storageRouter from './routes/storage.js';
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
import { startSchedulerLoop, stopSchedulerLoop } from './scheduler.js';
import { startCleanupLoop } from './tasks/cleanupTempSegments.js';
2026-04-07 21:58:25 -04:00
const app = express();
2026-04-07 21:58:25 -04:00
const PORT = process.env.PORT || 3000;
// ── Middleware ────────────────────────────────────────────────────────────────
fix(auth): make AUTH_ENABLED=true workable end-to-end Three concrete issues kept the login flow broken on dragonflight.live: 1. mam-api trusted no proxy headers, so behind nginx/Cloudflare the session cookie's `secure` flag and the rate-limiter's IP keying both saw the wrong values. Now sets `app.set('trust proxy', 1)`. 2. Session config was tied to NODE_ENV and lacked sameSite/name. Now: - SESSION_COOKIE_SECURE env (default: true when AUTH_ENABLED) so a site behind HTTPS gets Secure cookies regardless of NODE_ENV. - `sameSite: 'lax'` for predictable post-login redirects. - Renamed to `df.sid` so it's obvious in DevTools. - `rolling: true` extends the 7-day TTL on active use. - SESSION_SECRET is now required when AUTH_ENABLED=true; the server refuses to start with a dev default in prod. 3. login.html silently showed the sign-in panel even when no users exist or auth is off: - New GET /auth/setup-status reports {needs_setup, user_count, auth_enabled}. - login.html calls it on load and auto-flips into setup mode when needs_setup is true, or shows an explicit "auth is off" flash when auth_enabled is false (the previous symptom: logout button did nothing because /auth/me returned a synthetic admin no matter what). - Added a `.flash.info` style for the new neutral notice. 4. Sidebar logout used to call /auth/logout then `window.location .reload()`. With auth off that reload landed back on the synthetic- admin app and looked like nothing happened. It now redirects to /login.html in all states so the operator sees feedback (and the server-side messaging about auth being off) instead of a no-op. Deploy notes for zampp1: - Set AUTH_ENABLED=true and a random SESSION_SECRET in the mam-api environment (e.g. /opt/wild-dragon/.env). - Restart mam-api. - First load of /login.html will auto-route to the setup form so you can create the first admin.
2026-05-26 22:47:09 -04:00
// Trust the first proxy (nginx in front of us) so req.ip, req.secure, and
// req.protocol reflect the real client request — required for both the
// login rate-limiter's IP keying and `cookie.secure` cookie issuance.
app.set('trust proxy', 1);
app.use(cors({ origin: true, credentials: true }));
2026-04-07 22:05:39 -04:00
app.use(express.json({ limit: '50mb' }));
2026-04-07 21:58:25 -04:00
const PgSession = ConnectPgSimple(session);
fix(auth): make AUTH_ENABLED=true workable end-to-end Three concrete issues kept the login flow broken on dragonflight.live: 1. mam-api trusted no proxy headers, so behind nginx/Cloudflare the session cookie's `secure` flag and the rate-limiter's IP keying both saw the wrong values. Now sets `app.set('trust proxy', 1)`. 2. Session config was tied to NODE_ENV and lacked sameSite/name. Now: - SESSION_COOKIE_SECURE env (default: true when AUTH_ENABLED) so a site behind HTTPS gets Secure cookies regardless of NODE_ENV. - `sameSite: 'lax'` for predictable post-login redirects. - Renamed to `df.sid` so it's obvious in DevTools. - `rolling: true` extends the 7-day TTL on active use. - SESSION_SECRET is now required when AUTH_ENABLED=true; the server refuses to start with a dev default in prod. 3. login.html silently showed the sign-in panel even when no users exist or auth is off: - New GET /auth/setup-status reports {needs_setup, user_count, auth_enabled}. - login.html calls it on load and auto-flips into setup mode when needs_setup is true, or shows an explicit "auth is off" flash when auth_enabled is false (the previous symptom: logout button did nothing because /auth/me returned a synthetic admin no matter what). - Added a `.flash.info` style for the new neutral notice. 4. Sidebar logout used to call /auth/logout then `window.location .reload()`. With auth off that reload landed back on the synthetic- admin app and looked like nothing happened. It now redirects to /login.html in all states so the operator sees feedback (and the server-side messaging about auth being off) instead of a no-op. Deploy notes for zampp1: - Set AUTH_ENABLED=true and a random SESSION_SECRET in the mam-api environment (e.g. /opt/wild-dragon/.env). - Restart mam-api. - First load of /login.html will auto-route to the setup form so you can create the first admin.
2026-05-26 22:47:09 -04:00
// Session security knobs.
//
// - `secure` is set from SESSION_COOKIE_SECURE (default: true when AUTH_ENABLED).
// `trust proxy` above tells express-session that x-forwarded-proto can be
// trusted, so it issues Secure cookies on HTTPS requests forwarded by
// nginx/Cloudflare even though the proxy → mam-api hop is plain HTTP.
// Set SESSION_COOKIE_SECURE=false explicitly for local-only HTTP testing.
// - `sameSite: 'lax'` ships the cookie on top-level navigations (including
// the post-login redirect from /login.html) but blocks cross-site POSTs.
// - Renamed from default `connect.sid` to `df.sid` so it's obvious in DevTools.
// - `rolling: true` refreshes maxAge on every request so an active user
// doesn't get bounced to login after the 7-day TTL.
const authEnabled = process.env.AUTH_ENABLED === 'true';
const SESSION_SECRET = process.env.SESSION_SECRET
|| (authEnabled
? (() => { throw new Error('SESSION_SECRET is required when AUTH_ENABLED=true'); })()
: 'dev-only-not-for-production');
const SESSION_COOKIE_SECURE = process.env.SESSION_COOKIE_SECURE
? process.env.SESSION_COOKIE_SECURE === 'true'
: authEnabled; // default: secure cookies whenever auth is on
2026-04-07 21:58:25 -04:00
app.use(
session({
fix(auth): make AUTH_ENABLED=true workable end-to-end Three concrete issues kept the login flow broken on dragonflight.live: 1. mam-api trusted no proxy headers, so behind nginx/Cloudflare the session cookie's `secure` flag and the rate-limiter's IP keying both saw the wrong values. Now sets `app.set('trust proxy', 1)`. 2. Session config was tied to NODE_ENV and lacked sameSite/name. Now: - SESSION_COOKIE_SECURE env (default: true when AUTH_ENABLED) so a site behind HTTPS gets Secure cookies regardless of NODE_ENV. - `sameSite: 'lax'` for predictable post-login redirects. - Renamed to `df.sid` so it's obvious in DevTools. - `rolling: true` extends the 7-day TTL on active use. - SESSION_SECRET is now required when AUTH_ENABLED=true; the server refuses to start with a dev default in prod. 3. login.html silently showed the sign-in panel even when no users exist or auth is off: - New GET /auth/setup-status reports {needs_setup, user_count, auth_enabled}. - login.html calls it on load and auto-flips into setup mode when needs_setup is true, or shows an explicit "auth is off" flash when auth_enabled is false (the previous symptom: logout button did nothing because /auth/me returned a synthetic admin no matter what). - Added a `.flash.info` style for the new neutral notice. 4. Sidebar logout used to call /auth/logout then `window.location .reload()`. With auth off that reload landed back on the synthetic- admin app and looked like nothing happened. It now redirects to /login.html in all states so the operator sees feedback (and the server-side messaging about auth being off) instead of a no-op. Deploy notes for zampp1: - Set AUTH_ENABLED=true and a random SESSION_SECRET in the mam-api environment (e.g. /opt/wild-dragon/.env). - Restart mam-api. - First load of /login.html will auto-route to the setup form so you can create the first admin.
2026-05-26 22:47:09 -04:00
name: 'df.sid',
2026-04-07 21:58:25 -04:00
store: new PgSession({
pool,
tableName: 'sessions',
pruneSessionInterval: 3600,
2026-04-07 21:58:25 -04:00
}),
fix(auth): make AUTH_ENABLED=true workable end-to-end Three concrete issues kept the login flow broken on dragonflight.live: 1. mam-api trusted no proxy headers, so behind nginx/Cloudflare the session cookie's `secure` flag and the rate-limiter's IP keying both saw the wrong values. Now sets `app.set('trust proxy', 1)`. 2. Session config was tied to NODE_ENV and lacked sameSite/name. Now: - SESSION_COOKIE_SECURE env (default: true when AUTH_ENABLED) so a site behind HTTPS gets Secure cookies regardless of NODE_ENV. - `sameSite: 'lax'` for predictable post-login redirects. - Renamed to `df.sid` so it's obvious in DevTools. - `rolling: true` extends the 7-day TTL on active use. - SESSION_SECRET is now required when AUTH_ENABLED=true; the server refuses to start with a dev default in prod. 3. login.html silently showed the sign-in panel even when no users exist or auth is off: - New GET /auth/setup-status reports {needs_setup, user_count, auth_enabled}. - login.html calls it on load and auto-flips into setup mode when needs_setup is true, or shows an explicit "auth is off" flash when auth_enabled is false (the previous symptom: logout button did nothing because /auth/me returned a synthetic admin no matter what). - Added a `.flash.info` style for the new neutral notice. 4. Sidebar logout used to call /auth/logout then `window.location .reload()`. With auth off that reload landed back on the synthetic- admin app and looked like nothing happened. It now redirects to /login.html in all states so the operator sees feedback (and the server-side messaging about auth being off) instead of a no-op. Deploy notes for zampp1: - Set AUTH_ENABLED=true and a random SESSION_SECRET in the mam-api environment (e.g. /opt/wild-dragon/.env). - Restart mam-api. - First load of /login.html will auto-route to the setup form so you can create the first admin.
2026-05-26 22:47:09 -04:00
secret: SESSION_SECRET,
resave: false,
2026-04-07 21:58:25 -04:00
saveUninitialized: false,
fix(auth): make AUTH_ENABLED=true workable end-to-end Three concrete issues kept the login flow broken on dragonflight.live: 1. mam-api trusted no proxy headers, so behind nginx/Cloudflare the session cookie's `secure` flag and the rate-limiter's IP keying both saw the wrong values. Now sets `app.set('trust proxy', 1)`. 2. Session config was tied to NODE_ENV and lacked sameSite/name. Now: - SESSION_COOKIE_SECURE env (default: true when AUTH_ENABLED) so a site behind HTTPS gets Secure cookies regardless of NODE_ENV. - `sameSite: 'lax'` for predictable post-login redirects. - Renamed to `df.sid` so it's obvious in DevTools. - `rolling: true` extends the 7-day TTL on active use. - SESSION_SECRET is now required when AUTH_ENABLED=true; the server refuses to start with a dev default in prod. 3. login.html silently showed the sign-in panel even when no users exist or auth is off: - New GET /auth/setup-status reports {needs_setup, user_count, auth_enabled}. - login.html calls it on load and auto-flips into setup mode when needs_setup is true, or shows an explicit "auth is off" flash when auth_enabled is false (the previous symptom: logout button did nothing because /auth/me returned a synthetic admin no matter what). - Added a `.flash.info` style for the new neutral notice. 4. Sidebar logout used to call /auth/logout then `window.location .reload()`. With auth off that reload landed back on the synthetic- admin app and looked like nothing happened. It now redirects to /login.html in all states so the operator sees feedback (and the server-side messaging about auth being off) instead of a no-op. Deploy notes for zampp1: - Set AUTH_ENABLED=true and a random SESSION_SECRET in the mam-api environment (e.g. /opt/wild-dragon/.env). - Restart mam-api. - First load of /login.html will auto-route to the setup form so you can create the first admin.
2026-05-26 22:47:09 -04:00
rolling: true,
2026-04-07 21:58:25 -04:00
cookie: {
fix(auth): make AUTH_ENABLED=true workable end-to-end Three concrete issues kept the login flow broken on dragonflight.live: 1. mam-api trusted no proxy headers, so behind nginx/Cloudflare the session cookie's `secure` flag and the rate-limiter's IP keying both saw the wrong values. Now sets `app.set('trust proxy', 1)`. 2. Session config was tied to NODE_ENV and lacked sameSite/name. Now: - SESSION_COOKIE_SECURE env (default: true when AUTH_ENABLED) so a site behind HTTPS gets Secure cookies regardless of NODE_ENV. - `sameSite: 'lax'` for predictable post-login redirects. - Renamed to `df.sid` so it's obvious in DevTools. - `rolling: true` extends the 7-day TTL on active use. - SESSION_SECRET is now required when AUTH_ENABLED=true; the server refuses to start with a dev default in prod. 3. login.html silently showed the sign-in panel even when no users exist or auth is off: - New GET /auth/setup-status reports {needs_setup, user_count, auth_enabled}. - login.html calls it on load and auto-flips into setup mode when needs_setup is true, or shows an explicit "auth is off" flash when auth_enabled is false (the previous symptom: logout button did nothing because /auth/me returned a synthetic admin no matter what). - Added a `.flash.info` style for the new neutral notice. 4. Sidebar logout used to call /auth/logout then `window.location .reload()`. With auth off that reload landed back on the synthetic- admin app and looked like nothing happened. It now redirects to /login.html in all states so the operator sees feedback (and the server-side messaging about auth being off) instead of a no-op. Deploy notes for zampp1: - Set AUTH_ENABLED=true and a random SESSION_SECRET in the mam-api environment (e.g. /opt/wild-dragon/.env). - Restart mam-api. - First load of /login.html will auto-route to the setup form so you can create the first admin.
2026-05-26 22:47:09 -04:00
secure: SESSION_COOKIE_SECURE,
2026-04-07 21:58:25 -04:00
httpOnly: true,
fix(auth): make AUTH_ENABLED=true workable end-to-end Three concrete issues kept the login flow broken on dragonflight.live: 1. mam-api trusted no proxy headers, so behind nginx/Cloudflare the session cookie's `secure` flag and the rate-limiter's IP keying both saw the wrong values. Now sets `app.set('trust proxy', 1)`. 2. Session config was tied to NODE_ENV and lacked sameSite/name. Now: - SESSION_COOKIE_SECURE env (default: true when AUTH_ENABLED) so a site behind HTTPS gets Secure cookies regardless of NODE_ENV. - `sameSite: 'lax'` for predictable post-login redirects. - Renamed to `df.sid` so it's obvious in DevTools. - `rolling: true` extends the 7-day TTL on active use. - SESSION_SECRET is now required when AUTH_ENABLED=true; the server refuses to start with a dev default in prod. 3. login.html silently showed the sign-in panel even when no users exist or auth is off: - New GET /auth/setup-status reports {needs_setup, user_count, auth_enabled}. - login.html calls it on load and auto-flips into setup mode when needs_setup is true, or shows an explicit "auth is off" flash when auth_enabled is false (the previous symptom: logout button did nothing because /auth/me returned a synthetic admin no matter what). - Added a `.flash.info` style for the new neutral notice. 4. Sidebar logout used to call /auth/logout then `window.location .reload()`. With auth off that reload landed back on the synthetic- admin app and looked like nothing happened. It now redirects to /login.html in all states so the operator sees feedback (and the server-side messaging about auth being off) instead of a no-op. Deploy notes for zampp1: - Set AUTH_ENABLED=true and a random SESSION_SECRET in the mam-api environment (e.g. /opt/wild-dragon/.env). - Restart mam-api. - First load of /login.html will auto-route to the setup form so you can create the first admin.
2026-05-26 22:47:09 -04:00
sameSite: 'lax',
maxAge: 1000 * 60 * 60 * 24 * 7, // 7 days
2026-04-07 21:58:25 -04:00
},
})
);
// ── Health (no auth) ──────────────────────────────────────────────────────────
app.get('/health', (_req, res) => res.json({ status: 'ok' }));
// ── API Routes ────────────────────────────────────────────────────────────────
app.use('/api/v1/auth', authRouter);
app.use('/api/v1/assets', assetsRouter);
app.use('/api/v1/projects', projectsRouter);
app.use('/api/v1/bins', binsRouter);
app.use('/api/v1/jobs', jobsRouter);
app.use('/api/v1/capture', captureRouter);
app.use('/api/v1/upload', uploadRouter);
2026-04-07 22:05:39 -04:00
app.use('/api/v1/recorders', recordersRouter);
app.use('/api/v1/settings', settingsRouter);
app.use('/api/v1/ampp', amppRouter);
app.use('/api/v1/users', usersRouter);
app.use('/api/v1/groups', groupsRouter);
app.use('/api/v1/tokens', tokensRouter);
app.use('/api/v1/sequences', sequencesRouter);
app.use('/api/v1/system', systemRouter);
app.use('/api/v1/cluster', clusterRouter);
app.use('/api/v1/sdk', sdkRouter);
app.use('/api/v1/schedules', schedulesRouter);
app.use('/api/v1/metrics', metricsRouter);
app.use('/api/v1/assets/:assetId/comments', commentsRouter);
app.use('/api/v1/imports', importsRouter);
app.use('/api/v1/storage', storageRouter);
2026-04-07 21:58:25 -04:00
// ── Error handler ─────────────────────────────────────────────────────────────
2026-04-07 21:58:25 -04:00
app.use(errorHandler);
// ── Start ────────────────────────────────────────────────────────────────────
import { readdirSync, readFileSync } from 'node:fs';
import { fileURLToPath } from 'node:url';
import { dirname, join } from 'node:path';
const __dirnameMig = dirname(fileURLToPath(import.meta.url));
async function runMigrations() {
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
// Issue #107 — previously the loop swallowed errors and let the server boot
// on a half-migrated schema. Now: track applied migrations in a table, run
// every pending one inside a transaction, and exit non-zero on failure so
// the orchestrator restarts (and so an operator notices) instead of serving
// 500s for the next month.
const dir = join(__dirnameMig, 'db', 'migrations');
let files = [];
try { files = readdirSync(dir).filter(f => f.endsWith('.sql')).sort(); } catch { return; }
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
await pool.query(`
CREATE TABLE IF NOT EXISTS schema_migrations (
filename TEXT PRIMARY KEY,
applied_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
checksum_sha TEXT
)
`);
// Allow forcing a re-run via env when iterating locally.
const force = process.env.MIGRATIONS_FORCE === '1';
const allowFailures = process.env.MIGRATIONS_ALLOW_FAILURES === '1';
const appliedRes = await pool.query('SELECT filename FROM schema_migrations');
const applied = new Set(appliedRes.rows.map(r => r.filename));
for (const f of files) {
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
if (!force && applied.has(f)) continue;
const sql = readFileSync(join(dir, f), 'utf8');
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
const client = await pool.connect();
try {
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
await client.query('BEGIN');
await client.query(sql);
await client.query(
`INSERT INTO schema_migrations (filename) VALUES ($1)
ON CONFLICT (filename) DO UPDATE SET applied_at = NOW()`,
[f]
);
await client.query('COMMIT');
console.log('[migration] applied ' + f);
} catch (err) {
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
await client.query('ROLLBACK').catch(() => {});
console.error('[migration] FAILED ' + f + ': ' + err.message);
client.release();
if (allowFailures) continue;
// Hard fail — better to crash now than serve traffic on a broken schema.
console.error('[migration] aborting startup. Set MIGRATIONS_ALLOW_FAILURES=1 to override.');
process.exit(1);
}
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
client.release();
}
}
await runMigrations();
// Load S3 config from DB so any settings saved via the Settings page override env vars
await loadS3ConfigFromDb();
// ── Cluster self-heartbeat ────────────────────────────────────────────────────
function getLocalIp() {
// Prefer an explicit override — useful when running inside Docker where
// os.networkInterfaces() returns container bridge IPs, not the host LAN IP.
if (process.env.NODE_IP) return process.env.NODE_IP;
const ifaces = os.networkInterfaces();
for (const name of Object.keys(ifaces)) {
for (const iface of (ifaces[name] || [])) {
if (iface.family === 'IPv4' && !iface.internal) return iface.address;
}
}
return '127.0.0.1';
}
// Detect NVIDIA GPUs available to this container via nvidia-smi.
// Returns an array like [{ index: 0, name: 'Tesla P4', memory_mb: 7680 }, ...]
// or an empty array if nvidia-smi is unavailable or no GPUs found.
function detectGpus() {
return new Promise(resolve => {
exec(
'nvidia-smi --query-gpu=index,name,memory.total --format=csv,noheader,nounits',
{ timeout: 5000 },
(err, stdout) => {
if (err || !stdout.trim()) return resolve([]);
const gpus = stdout.trim().split('\n').map(line => {
const parts = line.split(',').map(s => s.trim());
return {
index: parseInt(parts[0], 10),
name: parts[1] || 'Unknown GPU',
memory_mb: parseInt(parts[2], 10) || 0,
};
}).filter(g => !isNaN(g.index));
resolve(gpus);
}
);
});
}
async function selfHeartbeat() {
const load = os.loadavg()[0];
const total = os.totalmem();
const used = total - os.freemem();
const gpus = await detectGpus();
const capabilities = { gpus, blackmagic: [] };
pool.query(
`INSERT INTO cluster_nodes
(hostname, ip_address, role, version, api_url,
cpu_usage, mem_used_mb, mem_total_mb, capabilities, last_seen)
VALUES ($1,$2,'primary',$3,$4,$5,$6,$7,$8,NOW())
ON CONFLICT (hostname) DO UPDATE SET
ip_address = EXCLUDED.ip_address,
cpu_usage = EXCLUDED.cpu_usage,
mem_used_mb = EXCLUDED.mem_used_mb,
mem_total_mb = EXCLUDED.mem_total_mb,
capabilities = EXCLUDED.capabilities,
last_seen = NOW()`,
[
process.env.NODE_HOSTNAME || os.hostname(),
getLocalIp(),
process.env.npm_package_version || null,
`http://${getLocalIp()}:${PORT}`,
parseFloat(load.toFixed(2)),
Math.round(used / 1024 / 1024),
Math.round(total / 1024 / 1024),
JSON.stringify(capabilities),
]
).catch(err => console.error('[cluster] heartbeat failed:', err.message));
}
setInterval(selfHeartbeat, 30_000);
selfHeartbeat();
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
const server = app.listen(PORT, () => {
const authMode = process.env.AUTH_ENABLED === 'true' ? 'ENABLED' : 'DISABLED (set AUTH_ENABLED=true for production)';
2026-04-07 21:58:25 -04:00
console.log(`MAM API listening on port ${PORT}`);
console.log(`Authentication: ${authMode}`);
// Boot the recorder scheduler tick loop after the HTTP server is live so
// the loop's self-calls to /recorders/:id/start|stop reach a ready socket.
startSchedulerLoop();
// Boot the temp-segment cleanup loop (runs hourly).
startCleanupLoop();
2026-04-07 21:58:25 -04:00
});
chore: 1.2 ship-prep sweep — close 38 issues Frontend / UX / a11y - Sidebar collapse/expand toggle with localStorage persistence (#142) - Settings sections wrap inputs in <form> with Enter-to-submit + native validation; password autocomplete=new-password (#141, #138) - Asset thumbnails get descriptive alt text (#140) - Production deploy now precompiles JSX via esbuild and loads the production React UMD instead of dev builds + in-browser Babel (#139, #122) - Search wrapper gets role=search; global search input gets aria-label, role=combobox, aria-controls/aria-expanded/aria-activedescendant wiring (#137, #135) - Dashboard and Library no longer share the same nav icon (#136) - Sidebar collapses off-canvas with a topbar menu button below 768 px; mobile default is collapsed (#134) - --text-3 bumped to #8B92A0 for WCAG AA contrast on --bg-0 (#133) - Schedule and Library routes were rendering empty inside the .main flex container — switched to flex:1 + min-height:0 (#131, #132, editor + asset detail get the same fix) - Jobs nav badge now polls /jobs?status=active every 10 s and reflects the live count (#130, #113) - aria-label sweep on every icon-only button (#126) - Premiere panel release list moved to window.PREMIERE_RELEASES in data.jsx; Editor + Settings read from the same source (#125) - Typo setPgMclips → setPgmClips (#124) - Stray console.error / console.warn calls gated behind window.DF_LOG.{warn,error} (#123) - Hardcoded /api/v1 paths route through window.ZAMPP_API_PREFIX (#115) - Schedule rows no longer crash on null recorder_id (#117) - EditorKeyboard guards against document.activeElement === null (#116) - Unmount-safe timers for PasswordResetModal, Containers, Editor (#111) - Player seek clamps below totalMs, server-side range clamping + uncached 416 on EOF, client-side EOF-stall watchdog (#143) - Duration badge overlap fix on narrow asset cards (#52) Backend / security / reliability - GET /recorders fixed N+1: single LATERAL JOIN for live_asset_id; Docker inspects bounded to actually-recording rows (#121) - Upload disk-storage (multer.diskStorage) streams parts to S3 instead of buffering 500 MB in RAM (#120) - /assets list clamps limit to MAX_LIMIT=500 to prevent OOM (#119) - SDK upload archive listing + post-extract sanitize block zip-slip / tar-slip and symlink escapes (#118) - Migrations track applied state in schema_migrations, run in a transaction, and exit non-zero on failure (#107) - node-agent BMD_COUNT override uses BMD_DEVICE_PREFIX; filesystem detection wins (#109, #127) - GPU_COUNT override now merges with nvidia-smi enrichment (#108) - /cluster/heartbeat requires a node-bound token or admin user; tokens carry bound_hostname (#106) - /recorders/:id/start error responses no longer echo the Docker create payload — env vars stay out of client responses (#105) - /recorders/probe restricts schemes (srt/rtmp/rtsp/udp/rtp), blocks private + loopback hosts for non-admins, denies common service ports (#104) - Scheduler tick guarded by a Postgres advisory lock; pending/running rows claimed via UPDATE...RETURNING + FOR UPDATE SKIP LOCKED to survive multi-node deploys (#103) - UUID validateUuid('id') param middleware on every /:id route (#102) - Error handler scrubs Postgres error messages and 5xx detail (#101) - Graceful SIGTERM/SIGINT shutdown — stops scheduler, drains the HTTP server, ends the pool, 25 s force-exit watchdog (#100) - AMPP sync moved from fire-and-forget to a persisted retry queue (ampp_sync_status / attempts / next_attempt_at + scheduler retry loop with exponential backoff) (#77) Migrations - 019: api_tokens.bound_hostname (#106) - 020: assets.ampp_sync_status + retry bookkeeping (#77) Other - Defer #92 Growing-files per-upload toggle, #80 Audio tab, #57 Dashboard redesign, #56 Editor SPA polish phase 3, #114 S3 migration tool to v1.3
2026-05-26 22:06:14 -04:00
// Issue #100 — graceful shutdown. Without this, `docker stop` (SIGTERM) killed
// the process mid-scheduler-tick, leaving Redis connections and Docker
// sockets dangling and producing partial DB writes. Now: stop the scheduler,
// finish in-flight HTTP requests, close PG/Redis pools, and exit cleanly
// (or hard-exit after 25 s if something is stuck).
let _shuttingDown = false;
async function gracefulShutdown(signal) {
if (_shuttingDown) return;
_shuttingDown = true;
console.log(`[shutdown] received ${signal} — closing gracefully…`);
// Stop accepting new requests + wind down the scheduler tick.
try { stopSchedulerLoop(); } catch (_) {}
// Force-exit watchdog so a hung connection can't keep us alive forever.
const killSwitch = setTimeout(() => {
console.error('[shutdown] forced exit after 25s timeout');
process.exit(1);
}, 25_000);
killSwitch.unref();
// Stop the HTTP server (waits for in-flight requests to finish).
await new Promise(resolve => server.close(resolve));
// Close DB pool + S3 client + any other resources. Best-effort.
try { await pool.end(); } catch (e) { console.warn('[shutdown] pool.end:', e.message); }
console.log('[shutdown] clean exit');
process.exit(0);
}
process.on('SIGTERM', () => gracefulShutdown('SIGTERM'));
process.on('SIGINT', () => gracefulShutdown('SIGINT'));
process.on('uncaughtException', (err) => {
console.error('[fatal] uncaughtException:', err);
gracefulShutdown('uncaughtException');
});
process.on('unhandledRejection', (reason) => {
console.error('[fatal] unhandledRejection:', reason);
});