wiki-llm#

What Was Established#

Dedicated Ubuntu 24.04 VM for hosting the homelab wiki system and Claude Code sessions. Chosen as a VM rather than an LXC for stronger isolation (wiki infrastructure handles credentials and embeddings).

Deployment#

Detail	Value
Hostname	wiki-llm
IP	192.168.1.206
VLAN	Gandalf (192.168.1.x)
OS	Ubuntu 24.04
CPU	2 cores
RAM	4 GB
Type	VM (Proxmox)
SSH user	iluvatar (sudo, PermitRootLogin no)

Access#

VS Code Remote SSH: Primary method for Claude Code sessions — VS Code connects to wiki-llm via remote SSH, giving Claude Code native filesystem access to /opt/wiki/homelab/
Direct SSH: ssh iluvatar@192.168.1.206

Wiki File Structure#

/opt/wiki/
  homelab/        ← git repo, all homelab wiki pages and skills
  work/           ← git repo, work/princelobel wiki + pipeline scripts
  projects/       ← git repo, projects wiki (planned)
  personal/       ← git repo, personal wiki (Gemma-only)
  raw-sources/    ← symlink to /mnt/wiki-nas/LLMWiki
  skills-reference/  ← clone of vanillaflava/llm-wiki-claude-skills (reference only)

Each wiki directory is an independent git repository (git init’d) for clean version history per namespace.

Work wiki NAS storage (symlinks, moved 2026-04-25)#

Raw docs and processed sources for the work wiki were moved from local disk to NAS to free root disk space (torch/pip cache was filling the 32GB root disk):

Path on wiki-llm	NAS target
`/opt/wiki/work/raw/`	`/mnt/wiki-nas/LLMWiki/work/raw/`
`/opt/wiki/work/ingested/`	`/mnt/wiki-nas/LLMWiki/work/ingested/`

Both are symlinks — pipeline scripts follow them transparently with no code changes required.

Python Environment#

Removed packages (2026-04-26)#

All ML inference runs on Legolas (192.168.1.45). Nothing runs locally on this VM.

Package	Size	Why removed
`torch`	~1.2GB	No local ML inference needed
`surya-ocr`	~500MB	Replaced by minicpm-v:8b on Legolas
`marker-pdf`	~500MB	Replaced by PyMuPDF + minicpm-v:8b hybrid

Current dependencies#

pymupdf (~25MB), python-docx, python-pptx, Pillow, psycopg2, requests

pip cache#

~/.cache/pip/ = 2.8GB on this VM. Safe to purge at any time:

pip cache purge

Synology SMB Mount#

Raw source files flow from the Synology NAS (LonelyMountain) to wiki-llm via SMB:

Detail	Value
Share	`//192.168.1.137/Documents`
Mount point	`/mnt/wiki-nas`
LLMWiki path	`/mnt/wiki-nas/LLMWiki`
Symlink	`/opt/wiki/raw-sources` → `/mnt/wiki-nas/LLMWiki`

fstab entry:

//192.168.1.137/Documents /mnt/wiki-nas cifs credentials=/etc/wiki-smb-credentials,uid=1000,gid=1000,_netdev 0 0

Credentials file /etc/wiki-smb-credentials:

username=LLMWiki
password=<password>

Synology setup: User LLMWiki created on Synology with:

SMB app permission enabled
Read/Write on the Documents shared folder

LLMWiki subdirectory structure on Synology:

/mnt/wiki-nas/LLMWiki/
  homelab/    ← raw sources for homelab wiki
  projects/   ← raw sources for projects wiki
  personal/   ← raw sources for personal wiki (Gemma-only)

Monitoring#

Node Exporter#

Installed via apt (Ubuntu package, not manual binary):

sudo apt install prometheus-node-exporter

Runs on port 9100. Prometheus scrape config:

- job_name: 'wiki'
  static_configs:
    - targets: ['192.168.1.206:9100']
      labels:
        host: 'wiki-llm'
        vlan: 'gandalf'

Uptime Kuma: HTTP monitor at http://192.168.1.206:9100 (plain HTTP — node exporter does not serve HTTPS; using https here causes SSL packet length errors).

Skills#

Wiki skills live at /opt/wiki/homelab/skills/. The five skills are:

wiki-ingest — process raw files into wiki pages
wiki-crystallize — distil session transcripts into wiki pages
wiki-lint — health-check the wiki
wiki-integrate — connect a page into the knowledge graph
wiki-query — answer questions from wiki knowledge

Skills were rewritten from the vanillaflava/llm-wiki-claude-skills reference for this homelab’s conventions.

Wiki System - Architecture, AI Infrastructure Overview, Node Exporter Deployment, PostgreSQL, Wiki Pipeline Scripts

Sources#

Homelab AI - 2026-04-20 · ingested/chats/2026-04-20-31-Homelab AI.json