Reindex Knowledge Base / reindex (push) Successful in 3s

Details

docs: add YAML frontmatter to all 151 markdown files

Adds title, description, type, domain, and tags frontmatter to every
doc for improved KB semantic search. The description field is prepended
to every search chunk, and domain/type/tags enable filtered queries.

Type values: context, guide, runbook, reference, troubleshooting
Domain values match directory structure (networking, docker, etc.)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-03-12 09:00:44 -05:00

5.7 KiB

Raw Blame History

title

description

type

domain

Tdarr Node Container Configurations

Overview

Complete examples for running Tdarr transcoding nodes in containers, covering both CPU-only and GPU-accelerated setups.

CPU-Only Configuration (Docker Compose)

For systems without GPU or when GPU isn't needed:

version: "3.4"
services:
  tdarr-node:
    container_name: tdarr-node-cpu
    image: ghcr.io/haveagitgat/tdarr_node:latest
    restart: unless-stopped
    environment:
      - TZ=America/Chicago
      - UMASK_SET=002
      - nodeName=local-workstation-cpu
      - serverIP=YOUR_TDARR_SERVER_IP  # Replace with your tdarr server IP
      - serverPort=8266
      - inContainer=true
      - ffmpegVersion=6
    volumes:
      # Mount your media from the same NAS share as the server
      - /path/to/your/media:/media  # Replace with your local media mount
      # Temp directory for transcoding cache
      - ./temp:/temp

Use case:

CPU-only transcoding
Testing Tdarr functionality
Systems without dedicated GPU
When GPU drivers aren't available

GPU-Accelerated Configuration (Podman)

Recommended for Fedora/RHEL/CentOS/Nobara systems:

Mapped Node (Direct Media Access)

podman run -d --name tdarr-node-gpu-mapped \
    --gpus all \
    --restart unless-stopped \
    -e TZ=America/Chicago \
    -e UMASK_SET=002 \
    -e nodeName=local-workstation-gpu-mapped \
    -e serverIP=10.10.0.43 \
    -e serverPort=8266 \
    -e inContainer=true \
    -e ffmpegVersion=6 \
    -e NVIDIA_DRIVER_CAPABILITIES=all \
    -e NVIDIA_VISIBLE_DEVICES=all \
    -v /mnt/NV2/tdarr-cache:/cache \
    -v /mnt/media/TV:/media/TV \
    -v /mnt/media/Movies:/media/Movies \
    -v /mnt/media/tdarr/tdarr-cache-clean:/temp \
    ghcr.io/haveagitgat/tdarr_node:latest

Unmapped Node (Downloads Files)

podman run -d --name tdarr-node-gpu-unmapped \
    --gpus all \
    --restart unless-stopped \
    -e TZ=America/Chicago \
    -e UMASK_SET=002 \
    -e nodeName=local-workstation-gpu-unmapped \
    -e serverIP=10.10.0.43 \
    -e serverPort=8266 \
    -e inContainer=true \
    -e ffmpegVersion=6 \
    -e NVIDIA_DRIVER_CAPABILITIES=all \
    -e NVIDIA_VISIBLE_DEVICES=all \
    -v /mnt/NV2/tdarr-cache:/cache \
    -v /mnt/media:/media \
    -v /mnt/media/tdarr/tdarr-cache-clean:/temp \
    ghcr.io/haveagitgat/tdarr_node:latest

Use cases:

Mapped: Direct media access, faster processing, no file downloads
Unmapped: Works when network shares aren't available locally
Hardware video encoding/decoding (NVENC/NVDEC)
High-performance transcoding with NVMe cache
Multiple concurrent streams
Fedora-based systems where Podman works better than Docker

GPU-Accelerated Configuration (Docker)

For Ubuntu/Debian systems where Docker GPU support works:

version: "3.4"
services:
  tdarr-node:
    container_name: tdarr-node-gpu
    image: ghcr.io/haveagitgat/tdarr_node:latest
    restart: unless-stopped
    environment:
      - TZ=America/Chicago
      - UMASK_SET=002
      - nodeName=local-workstation-gpu
      - serverIP=YOUR_TDARR_SERVER_IP
      - serverPort=8266
      - inContainer=true
      - ffmpegVersion=6
      - NVIDIA_DRIVER_CAPABILITIES=all
      - NVIDIA_VISIBLE_DEVICES=all
    volumes:
      - /path/to/your/media:/media
      - ./temp:/temp
    deploy:
      resources:
        reservations:
          devices:
          - driver: nvidia
            count: all
            capabilities: [gpu]

Configuration Parameters

Required Environment Variables

TZ: Timezone (e.g., America/Chicago)
nodeName: Unique identifier for this node
serverIP: IP address of Tdarr server
serverPort: Tdarr server port (typically 8266)
inContainer: Set to true for containerized deployments
ffmpegVersion: FFmpeg version to use (6 recommended)

GPU-Specific Variables

NVIDIA_DRIVER_CAPABILITIES: Set to all for full GPU access
NVIDIA_VISIBLE_DEVICES: all for all GPUs, or specific GPU IDs

Volume Mounts

/media: Mount point for media files (must match server configuration)
/temp: Temporary directory for transcoding cache

Platform-Specific Recommendations

Fedora/RHEL/CentOS/Nobara

GPU: Use Podman (Docker Desktop has GPU issues)
CPU: Docker or Podman both work fine

Ubuntu/Debian

GPU: Use Docker with nvidia-container-toolkit
CPU: Docker recommended

Testing GPU Functionality

Verify GPU access inside container:

# For Podman
podman exec tdarr-node-gpu nvidia-smi

# For Docker  
docker exec tdarr-node-gpu nvidia-smi

Test NVENC encoding:

# For Podman
podman exec tdarr-node-gpu /usr/local/bin/tdarr-ffmpeg -f lavfi -i testsrc2=duration=5:size=1920x1080:rate=30 -c:v h264_nvenc -t 5 /tmp/test.mp4

# For Docker
docker exec tdarr-node-gpu /usr/local/bin/tdarr-ffmpeg -f lavfi -i testsrc2=duration=5:size=1920x1080:rate=30 -c:v h264_nvenc -t 5 /tmp/test.mp4

Troubleshooting

GPU not detected: See reference/docker/nvidia-gpu-troubleshooting.md
Permission issues: Ensure proper UMASK_SET and volume permissions
Connection issues: Verify serverIP and firewall settings
Performance issues: Monitor CPU/GPU utilization during transcoding

patterns/docker/gpu-acceleration.md - GPU acceleration patterns
reference/docker/nvidia-gpu-troubleshooting.md - Detailed GPU troubleshooting
start-tdarr-gpu-podman.sh - Ready-to-use Podman startup script

5.7 KiB Raw Blame History