2026-05-20 13:48:27 -04:00
|
|
|
# Wild Dragon MAM — Worker Node Stack
|
|
|
|
|
# ─────────────────────────────────────
|
|
|
|
|
# Deploy on any machine you want to join the cluster as a worker.
|
|
|
|
|
# The primary stack (mam-api, db, redis) continues running on TrueNAS.
|
|
|
|
|
#
|
|
|
|
|
# Required env vars (set in .env.worker or export before running):
|
|
|
|
|
# MAM_API_URL URL of the primary MAM API e.g. http://10.0.0.25:47432
|
|
|
|
|
# NODE_TOKEN Bearer token from the primary's Tokens page
|
2026-05-21 00:14:33 -04:00
|
|
|
# NODE_IP Host LAN IP to report (set by onboard-node.sh)
|
2026-05-20 13:48:27 -04:00
|
|
|
#
|
2026-05-20 14:19:21 -04:00
|
|
|
# Optional hardware overrides (if Docker can't see /dev directly):
|
|
|
|
|
# GPU_COUNT Number of NVIDIA GPUs on this node (default: auto-detect from /dev/nvidia*)
|
2026-05-21 00:14:33 -04:00
|
|
|
# BMD_COUNT Number of Blackmagic DeckLink cards (default: auto-detect from /dev/blackmagic/)
|
|
|
|
|
# BMD_MODEL Marketed card name (e.g. "DeckLink Duo 2") — drives the port-diagram UI
|
2026-05-20 14:19:21 -04:00
|
|
|
#
|
|
|
|
|
# Optional env vars (needed only if starting the worker or capture profiles):
|
2026-05-20 13:48:27 -04:00
|
|
|
# REDIS_URL, DATABASE_URL, S3_ENDPOINT, S3_BUCKET, S3_ACCESS_KEY, S3_SECRET_KEY
|
2026-05-20 14:19:21 -04:00
|
|
|
# BMD_DEVICE_0 DeckLink device path (default: /dev/blackmagic/dv0)
|
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
|
|
|
# (DeckLink IO / Quad cards expose /dev/blackmagic/io* instead — set BMD_DEVICE_PREFIX=io)
|
2026-05-20 14:19:21 -04:00
|
|
|
# BMD_DEVICE_1 DeckLink device path (default: /dev/blackmagic/dv1)
|
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
|
|
|
# BMD_DEVICE_PREFIX Naming prefix for synthesized BMD_COUNT-based devices (default: dv). Use 'io' for IO/Quad.
|
2026-05-21 18:51:11 -04:00
|
|
|
# LIVE_DIR Host path for HLS live segments (default: /mnt/NVME/MAM/wild-dragon-live)
|
2026-05-20 14:19:21 -04:00
|
|
|
#
|
|
|
|
|
# Profiles:
|
|
|
|
|
# (default) node-agent only — cluster visibility + hardware heartbeat
|
2026-05-21 00:14:33 -04:00
|
|
|
# --profile worker + CPU/GPU job worker (proxy generation, transcoding)
|
2026-05-20 14:19:21 -04:00
|
|
|
# --profile capture + SDI capture service (requires Blackmagic DeckLink card)
|
|
|
|
|
#
|
|
|
|
|
# To enable GPU transcoding, also apply docker-compose.gpu.yml:
|
|
|
|
|
# docker compose -f docker-compose.worker.yml -f docker-compose.gpu.yml --profile worker up -d
|
2026-05-21 18:51:11 -04:00
|
|
|
#
|
|
|
|
|
# NOTE: The node-agent mounts /var/run/docker.sock to spawn on-demand SDI
|
|
|
|
|
# capture sidecars when the primary mam-api routes a recorder to this node.
|
|
|
|
|
# Build the capture image before first use:
|
|
|
|
|
# docker compose -f docker-compose.worker.yml build capture
|
2026-05-20 13:48:27 -04:00
|
|
|
|
|
|
|
|
services:
|
|
|
|
|
|
2026-05-21 00:14:33 -04:00
|
|
|
# node-agent runs in host network mode so it can see the real host
|
|
|
|
|
# interfaces, GPU devices and DeckLink cards without bridging tricks.
|
|
|
|
|
# The reported IP / hostname will be the host's, not the container's.
|
2026-05-20 13:48:27 -04:00
|
|
|
node-agent:
|
|
|
|
|
build: ./services/node-agent
|
|
|
|
|
restart: unless-stopped
|
2026-05-21 00:14:33 -04:00
|
|
|
network_mode: host
|
2026-05-20 13:48:27 -04:00
|
|
|
environment:
|
|
|
|
|
MAM_API_URL: ${MAM_API_URL}
|
|
|
|
|
NODE_TOKEN: ${NODE_TOKEN:-}
|
|
|
|
|
NODE_ROLE: ${NODE_ROLE:-worker}
|
2026-05-21 00:14:33 -04:00
|
|
|
NODE_IP: ${NODE_IP:-}
|
|
|
|
|
AGENT_PORT: ${AGENT_PORT:-7436}
|
2026-05-20 13:48:27 -04:00
|
|
|
HEARTBEAT_MS: ${HEARTBEAT_MS:-30000}
|
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
|
|
|
GPU_COUNT: ${GPU_COUNT:--1}
|
|
|
|
|
BMD_COUNT: ${BMD_COUNT:--1}
|
|
|
|
|
BMD_MODEL: ${BMD_MODEL:-}
|
|
|
|
|
BMD_DEVICE_PREFIX: ${BMD_DEVICE_PREFIX:-dv}
|
2026-05-21 18:51:11 -04:00
|
|
|
LIVE_DIR: ${LIVE_DIR:-/mnt/NVME/MAM/wild-dragon-live}
|
2026-05-21 00:14:33 -04:00
|
|
|
volumes:
|
2026-05-21 18:51:11 -04:00
|
|
|
- /var/run/docker.sock:/var/run/docker.sock
|
2026-05-21 00:14:33 -04:00
|
|
|
- /dev:/dev:ro
|
2026-05-28 23:20:02 -04:00
|
|
|
- /mnt/NVME/MAM/wild-dragon-live:/mnt/NVME/MAM/wild-dragon-live:ro
|
2026-05-21 18:51:11 -04:00
|
|
|
devices:
|
|
|
|
|
- /dev/blackmagic:/dev/blackmagic
|
2026-05-20 13:48:27 -04:00
|
|
|
|
|
|
|
|
worker:
|
|
|
|
|
build: ./services/worker
|
|
|
|
|
profiles: [worker]
|
|
|
|
|
restart: unless-stopped
|
|
|
|
|
environment:
|
|
|
|
|
REDIS_URL: ${REDIS_URL}
|
|
|
|
|
DATABASE_URL: ${DATABASE_URL}
|
|
|
|
|
S3_ENDPOINT: ${S3_ENDPOINT}
|
|
|
|
|
S3_BUCKET: ${S3_BUCKET}
|
|
|
|
|
S3_ACCESS_KEY: ${S3_ACCESS_KEY}
|
|
|
|
|
S3_SECRET_KEY: ${S3_SECRET_KEY}
|
|
|
|
|
S3_REGION: ${S3_REGION:-us-east-1}
|
2026-05-20 14:19:21 -04:00
|
|
|
NVENC_ENABLED: ${NVENC_ENABLED:-false}
|
|
|
|
|
networks:
|
|
|
|
|
- wild-dragon-worker
|
|
|
|
|
|
|
|
|
|
# SDI capture service — only start on nodes with Blackmagic DeckLink cards
|
|
|
|
|
# Set BMD_DEVICE_0 in .env.worker to the actual device path, e.g. /dev/blackmagic/dv0
|
|
|
|
|
capture:
|
|
|
|
|
build: ./services/capture
|
|
|
|
|
profiles: [capture]
|
|
|
|
|
restart: unless-stopped
|
|
|
|
|
environment:
|
|
|
|
|
REDIS_URL: ${REDIS_URL}
|
|
|
|
|
DATABASE_URL: ${DATABASE_URL}
|
|
|
|
|
S3_ENDPOINT: ${S3_ENDPOINT}
|
|
|
|
|
S3_BUCKET: ${S3_BUCKET}
|
|
|
|
|
S3_ACCESS_KEY: ${S3_ACCESS_KEY}
|
|
|
|
|
S3_SECRET_KEY: ${S3_SECRET_KEY}
|
|
|
|
|
CAPTURE_PORT: 3001
|
|
|
|
|
devices:
|
|
|
|
|
- ${BMD_DEVICE_0:-/dev/blackmagic/dv0}:/dev/blackmagic/dv0
|
|
|
|
|
- ${BMD_DEVICE_1:-/dev/blackmagic/dv1}:/dev/blackmagic/dv1
|
|
|
|
|
ports:
|
|
|
|
|
- "${CAPTURE_PORT:-7437}:3001"
|
2026-05-20 13:48:27 -04:00
|
|
|
networks:
|
|
|
|
|
- wild-dragon-worker
|
|
|
|
|
|
|
|
|
|
networks:
|
|
|
|
|
wild-dragon-worker:
|
|
|
|
|
driver: bridge
|