40 lines
1.4 KiB
Markdown
40 lines
1.4 KiB
Markdown
# Rocinante
|
|
|
|
Heavy AI inference server.
|
|
|
|
> Planned direction: Rocinante becomes optional if the upgraded PD build receives the 4090. If PD does not get the 4090, Rocinante remains the heavy-inference specialist.
|
|
|
|
## Hardware
|
|
|
|
| Component | Details |
|
|
|-----------|---------|
|
|
| **IP** | 10.5.30.112 |
|
|
| **GPU** | NVIDIA RTX 4090 (24 GB VRAM) |
|
|
| **Role** | Heavy AI inference (Ollama) |
|
|
|
|
## Services
|
|
|
|
### Ollama
|
|
|
|
Ollama runs on port `11434` and hosts the large-model ("heavy") tier.
|
|
|
|
| Model | Use |
|
|
|-------|-----|
|
|
| `qwen3:latest` | Fast reasoning / general chat |
|
|
| `qwen2.5:32b` | Large-context tasks |
|
|
| `gemma3:27b` | Multimodal / vision |
|
|
|
|
> **LiteLLM tier:** Rocinante is the **"heavy"** tier in LiteLLM `config.yaml`.
|
|
> Heavy-tier requests from OpenWebUI and n8n route to `http://10.5.30.112:11434`.
|
|
|
|
## Current operational note
|
|
|
|
- During Doris validation from N.O.M.A.D. on 2026-05-21, `10.5.30.112` did not answer ping and did not accept connections on `11434` or `8787`.
|
|
- Treat Rocinante-hosted Ollama heavy-tier traffic and CUDA Whisper as currently unreachable from N.O.M.A.D. until host/network reachability is restored.
|
|
- Re-validate from N.O.M.A.D. with `curl http://10.5.30.112:11434/api/tags` and `curl http://10.5.30.112:8787/health` after recovery.
|
|
|
|
## Notes
|
|
|
|
- No SSH config needed for routine use — Ollama API is accessed over LAN only
|
|
- Do not run multiple large VRAM allocations simultaneously on the 4090
|