📂 Datasets created
| Dataset | Usage | Service |
|---|---|---|
| models | Generic AI models | Ollama / HuggingFace |
| ollama | Ollama models (LLM) | Ollama |
| huggingface | HuggingFace cache | Python / LangChain |
| whisper | Whisper STT models | Whisper |
| comfyui | ComfyUI workflows and models | ComfyUI |
| flux | Flux models (image generation) | ComfyUI / Flux |
| rag | RAG source documents | Open WebUI |
| embeddings | Vectors (local cache) | bge-m3 / pgvector |
| qdrant | Fallback vector database | Qdrant (fallback) |
🎯 Skills demonstrated
- ZFS administration (pools, datasets, properties)
- AI-oriented storage architecture (data separation)
- Manual ZFS snapshots (before sensitive operations) — no confirmed scheduled automation
- Quota and compression management
- Disk space and pool health monitoring
- ZFS / Proxmox integration
💻 ZFS commands
bash
# Lister tous les datasets zfs list # Créer un dataset zfs create pool/ollama # Snapshot avant migration zfs snapshot pool/ollama@avant-migration-$(date +%Y%m%d) # Lister les snapshots zfs list -t snapshot # Restaurer un snapshot zfs rollback pool/ollama@avant-migration # Vérifier santé du pool zpool status zpool scrub pool
Proxmox VE
IA Locale / LLM
RAG Vectoriel
Speech AI