stefankoelle/moonweb-site
1
0
Fork 0
mirror of https://github.com/skoelle/moonweb-site.git synced 2026-09-17 17:00:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

Remove all em-dashes from website content and documentation

Browse Source 
Replace ' — ' with ', ' across 40 files. Hyphens (-) untouched.
timecapsule, stefankoelle, LICENSE, CSS comments left as-is.
This commit is contained in:
stefankoelle
2026-09-11 20:13:28 +02:00
parent 3363702ee6
commit eb0b777418
40 changed files with 311 additions and 311 deletions
Download Patch File Download Diff File Expand all files Collapse all files
AGENTS.md
+19 -19
View File
@@ -1,4 +1,4 @@
# AGENTS.md — moonweb-site # AGENTS.md, moonweb-site
## Projektuebersicht ## Projektuebersicht
@@ -23,8 +23,8 @@ Monorepo fuer 6 statische Websites unter www.moonweb.org + stefankoelle.de, basi
### Andere Sites (nicht im Monorepo) ### Andere Sites (nicht im Monorepo)
- 28k8.moonweb.org — 90er BBS/Scene-Archiv - 28k8.moonweb.org, 90er BBS/Scene-Archiv
- buildbroken.moonweb.org — .NET Open Space Blog Archiv - buildbroken.moonweb.org, .NET Open Space Blog Archiv
## Dateistruktur ## Dateistruktur
@@ -101,23 +101,23 @@ npm run build:moonweb # Moonweb + Timecapsule (fuer Deployment)
## Design-Prinzipien ## Design-Prinzipien
1. **Header konsistent** — Identischer Site-Switcher auf allen Home-Sites 1. **Header konsistent**, Identischer Site-Switcher auf allen Home-Sites
2. **Content flexibel** — Detailseiten duerfen eigenes Layout haben 2. **Content flexibel**, Detailseiten duerfen eigenes Layout haben
3. **Accent-Farben:** hub=#3b6ea5, infra=#99333A, smarthome=#1f8a8a, code=#3E5098, retro=#8a6d3b (inlined in base.njk) 3. **Accent-Farben:** hub=#3b6ea5, infra=#99333A, smarthome=#1f8a8a, code=#3E5098, retro=#8a6d3b (inlined in base.njk)
4. **Englisch** — Alle Sites komplett auf Englisch 4. **Englisch**, Alle Sites komplett auf Englisch
5. **Keine Analytics** — Keine Tracking-Tools 5. **Keine Analytics**, Keine Tracking-Tools
6. **Sensible Daten** — Infra-Content wird manuell redigiert (keine IPs, Keys, Passwoerter) 6. **Sensible Daten**, Infra-Content wird manuell redigiert (keine IPs, Keys, Passwoerter)
## URL-Struktur ## URL-Struktur
Alle Sites sind unter `www.moonweb.org` als Subverzeichnisse erreichbar: Alle Sites sind unter `www.moonweb.org` als Subverzeichnisse erreichbar:
- `www.moonweb.org/` — Hub (Root) - `www.moonweb.org/`, Hub (Root)
- `www.moonweb.org/infra/` — Infra - `www.moonweb.org/infra/`, Infra
- `www.moonweb.org/smarthome/` — Smarthome - `www.moonweb.org/smarthome/`, Smarthome
- `www.moonweb.org/code/` — Code - `www.moonweb.org/code/`, Code
- `www.moonweb.org/retro/` — Retro - `www.moonweb.org/retro/`, Retro
- `www.moonweb.org/timecapsule/` — Timecapsule (2001 Design) - `www.moonweb.org/timecapsule/`, Timecapsule (2001 Design)
- `www.moonweb.org/impressum/` — Impressum - `www.moonweb.org/impressum/`, Impressum
Cloudflare Redirects leiten alte Subdomains weiter: Cloudflare Redirects leiten alte Subdomains weiter:
- `hub.moonweb.org/*` → `www.moonweb.org/*` - `hub.moonweb.org/*` → `www.moonweb.org/*`
@@ -145,10 +145,10 @@ npm run prebuild:cv # Alle Pre-Build Tasks (inkl. CV PDF)
``` ```
Dateien: Dateien:
- `stefankoelle/pdf/cv-style.css` — WeasyPrint-Stylesheet (A4, Typografie) - `stefankoelle/pdf/cv-style.css`, WeasyPrint-Stylesheet (A4, Typografie)
- `stefankoelle/pdf/build-pdf.sh` — Shell-Skript fuer PDF-Generierung - `stefankoelle/pdf/build-pdf.sh`, Shell-Skript fuer PDF-Generierung
- `stefankoelle/cv-print.njk` — Standalone HTML-Template (nur CV-Content) - `stefankoelle/cv-print.njk`, Standalone HTML-Template (nur CV-Content)
- `stefankoelle/pdf/cv.pdf` — Generiertes PDF (Output) - `stefankoelle/pdf/cv.pdf`, Generiertes PDF (Output)
Wichtig: Das PDF wird via Eleventy-Passthrough ins Build-Output kopiert (`dist/stefankoelle/pdf/cv.pdf`). Wichtig: Das PDF wird via Eleventy-Passthrough ins Build-Output kopiert (`dist/stefankoelle/pdf/cv.pdf`).
DESIGN.md
+18 -18
View File
@@ -1,57 +1,57 @@
# moonweb.org Hub-Konzept (v5 — fast vollständig, letzte Lücken vor SPEC.md/PLAN.md) # moonweb.org Hub-Konzept (v5, fast vollständig, letzte Lücken vor SPEC.md/PLAN.md)
Status: Konzeptrunde nahezu abgeschlossen. Noch keine Umsetzung. Status: Konzeptrunde nahezu abgeschlossen. Noch keine Umsetzung.
## 1. Header/Layout-Prinzip — final ## 1. Header/Layout-Prinzip, final
Für alle Home-Bereich-Sites (infra, smarthome, code, retro) gilt ab jetzt ein klares Prinzip: Für alle Home-Bereich-Sites (infra, smarthome, code, retro) gilt ab jetzt ein klares Prinzip:
- **Header ist das Wichtigste** — identisch über alle vier Sites (Site-Switcher, Banner, Farbakzent der jeweiligen Domain). - **Header ist das Wichtigste**, identisch über alle vier Sites (Site-Switcher, Banner, Farbakzent der jeweiligen Domain).
- **Übersichtsseiten** (Domain-Root) folgen dem einheitlichen Card-Grid-Grundlayout. - **Übersichtsseiten** (Domain-Root) folgen dem einheitlichen Card-Grid-Grundlayout.
- **Detailseiten/Subseiten** übernehmen den gleichen Header, dürfen darunter aber ein **eigenes, freieres Layout** haben (wie es bei `/ledmatrix/` auf stefankoelle.de heute schon der Fall ist — Pinbelegung, API-Doku, Fotos in freier Anordnung statt starres Card-Raster). - **Detailseiten/Subseiten** übernehmen den gleichen Header, dürfen darunter aber ein **eigenes, freieres Layout** haben (wie es bei `/ledmatrix/` auf stefankoelle.de heute schon der Fall ist, Pinbelegung, API-Doku, Fotos in freier Anordnung statt starres Card-Raster).
Das ist ein wichtiger Unterschied zu v3/v4: nicht "alles identisch", sondern **Header konsistent, Content-Layout pro Subseite flexibel**. Das ist ein wichtiger Unterschied zu v3/v4: nicht "alles identisch", sondern **Header konsistent, Content-Layout pro Subseite flexibel**.
## 2. LEDMatrix-Migration — erledigt (Phase 5) ## 2. LEDMatrix-Migration, erledigt (Phase 5)
- `/ledmatrix/` wurde vollständig ins Monorepo migriert und liegt unter `stefankoelle/de/ledmatrix/`. - `/ledmatrix/` wurde vollständig ins Monorepo migriert und liegt unter `stefankoelle/de/ledmatrix/`.
- Deploy erfolgt über IONOS SFTP alongside dem stefankoelle.de Onepager. - Deploy erfolgt über IONOS SFTP alongside dem stefankoelle.de Onepager.
- Die URL `stefankoelle.de/ledmatrix/` bleibt unverändert bestehen. - Die URL `stefankoelle.de/ledmatrix/` bleibt unverändert bestehen.
## 3. Detailtiefe pro Domain — final festgelegt ## 3. Detailtiefe pro Domain, final festgelegt
| Domain | Detailseiten? | Regel | | Domain | Detailseiten? | Regel |
|---|---|---| |---|---|---|
| smarthome | Ja, aber nur wenn genug Content da ist | Übersicht zeigt Aufbau des Smarthomes; ein Thema bekommt sofort eine Detailseite, wenn Daten vorhanden sind — sonst wird das Thema vorerst nur in der Übersicht erwähnt (kein Platzhalter-Zwang) | | smarthome | Ja, aber nur wenn genug Content da ist | Übersicht zeigt Aufbau des Smarthomes; ein Thema bekommt sofort eine Detailseite, wenn Daten vorhanden sind, sonst wird das Thema vorerst nur in der Übersicht erwähnt (kein Platzhalter-Zwang) |
| code | **Nie** | Nur Übersichtskarten + Link zu GitHub — bewusst kein Duplizieren von README-Inhalten | | code | **Nie** | Nur Übersichtskarten + Link zu GitHub, bewusst kein Duplizieren von README-Inhalten |
| infra | Kaum, eher übersichtsorientiert | Hauptteil bleibt bewusst oberflächlich, da sensible Daten (IPs, Keys, interne Topologie) nicht öffentlich dürfen — siehe Punkt 4 | | infra | Kaum, eher übersichtsorientiert | Hauptteil bleibt bewusst oberflächlich, da sensible Daten (IPs, Keys, interne Topologie) nicht öffentlich dürfen, siehe Punkt 4 |
| retro | Ja, aber bewusst rudimentär | Thema ist inhaltlich noch unausgereift — nur Grundgerüst anlegen, keine weitere Zeit investieren | | retro | Ja, aber bewusst rudimentär | Thema ist inhaltlich noch unausgereift, nur Grundgerüst anlegen, keine weitere Zeit investieren |
## 4. Infra-Sensitivität — Redaktionsregel (neu, wichtig für SPEC.md) ## 4. Infra-Sensitivität, Redaktionsregel (neu, wichtig für SPEC.md)
Da infra "eher oberflächlich" bleiben soll, braucht es eine einfache Faustregel, was rein darf und was nicht — sonst ist das bei der Migration der bestehenden Docs (Heimnetzwerk-Final etc.) jedes Mal eine Einzelfallentscheidung: Da infra "eher oberflächlich" bleiben soll, braucht es eine einfache Faustregel, was rein darf und was nicht, sonst ist das bei der Migration der bestehenden Docs (Heimnetzwerk-Final etc.) jedes Mal eine Einzelfallentscheidung:
- **Darf rein:** Architektur-Ebene (Proxmox + Synology + Docker-Host, VLAN-Konzept ohne konkrete interne IP-Pläne, welche Diensttypen laufen, welche Tools genutzt werden). - **Darf rein:** Architektur-Ebene (Proxmox + Synology + Docker-Host, VLAN-Konzept ohne konkrete interne IP-Pläne, welche Diensttypen laufen, welche Tools genutzt werden).
- **Darf nicht rein:** konkrete IP-Adressen, WireGuard-Keys/Preshared-Keys, Passwörter, interne Hostnamen mit Rückschlussmöglichkeit, Backup-Ziele mit Zugangsdaten. - **Darf nicht rein:** konkrete IP-Adressen, WireGuard-Keys/Preshared-Keys, Passwörter, interne Hostnamen mit Rückschlussmöglichkeit, Backup-Ziele mit Zugangsdaten.
Diese Regel gilt 1:1 auch für retro/smarthome, ist aber bei infra am relevantesten, weil dort die Rohdokumente (z. B. Heimnetzwerk-Final-v3.2, GL_Flint2_Final_v2) aktuell sehr konkret sind (u. a. reale IPs, WireGuard-Keys) und beim Übertragen aktiv gekürzt werden müssen. Diese Regel gilt 1:1 auch für retro/smarthome, ist aber bei infra am relevantesten, weil dort die Rohdokumente (z. B. Heimnetzwerk-Final-v3.2, GL_Flint2_Final_v2) aktuell sehr konkret sind (u. a. reale IPs, WireGuard-Keys) und beim Übertragen aktiv gekürzt werden müssen.
## 5. Bilder/Assets — final ## 5. Bilder/Assets, final
Alle Bilder bleiben **im GitHub-Repo** versioniert (kein separates Cloudflare R2/Images). Einfachster Weg, passt zum ohnehin geplanten Monorepo-Ansatz. Alle Bilder bleiben **im GitHub-Repo** versioniert (kein separates Cloudflare R2/Images). Einfachster Weg, passt zum ohnehin geplanten Monorepo-Ansatz.
## 6. DNS — bereits bei Cloudflare ## 6. DNS, bereits bei Cloudflare
Zone liegt schon bei Cloudflare, Pages-Anbindung ist damit ohne Zusatzschritt möglich. Offen bleibt nur, wohin die nackte Domain `moonweb.org` zeigt (siehe offene Frage 1 unten). Zone liegt schon bei Cloudflare, Pages-Anbindung ist damit ohne Zusatzschritt möglich. Offen bleibt nur, wohin die nackte Domain `moonweb.org` zeigt (siehe offene Frage 1 unten).
## 7. Übersetzung, Last-Updated, Analytics, Perplexity-Rückkanal — final ## 7. Übersetzung, Last-Updated, Analytics, Perplexity-Rückkanal, final
- **Übersetzung:** einmaliger KI-gestützter Durchlauf pro Dokument bei der Migration, kein wiederkehrender Prozess. - **Übersetzung:** einmaliger KI-gestützter Durchlauf pro Dokument bei der Migration, kein wiederkehrender Prozess.
- **Last-Updated-Anzeige:** einfachste Lösung — im Zweifel **gar keine Anzeige**. Stattdessen soll der Anspruch sein, dass Inhalte grundsätzlich aktuell gehalten werden, statt ein Datum zu zeigen, das dann veraltet wirkt. - **Last-Updated-Anzeige:** einfachste Lösung, im Zweifel **gar keine Anzeige**. Stattdessen soll der Anspruch sein, dass Inhalte grundsätzlich aktuell gehalten werden, statt ein Datum zu zeigen, das dann veraltet wirkt.
- **Analytics:** keine Statistik-Erfassung, weder Cloudflare Web Analytics noch anderes. - **Analytics:** keine Statistik-Erfassung, weder Cloudflare Web Analytics noch anderes.
- **Perplexity-Rückkanal:** zurückgestellt, wird später erneut betrachtet, keine Zeit jetzt investieren. - **Perplexity-Rückkanal:** zurückgestellt, wird später erneut betrachtet, keine Zeit jetzt investieren.
## 8. Offene Kleinigkeiten — beantwortet ## 8. Offene Kleinigkeiten, beantwortet
Die vier ursprünglich offenen Punkte wurden inzwischen in SPEC.md/PLAN.md beantwortet: Die vier ursprünglich offenen Punkte wurden inzwischen in SPEC.md/PLAN.md beantwortet:
@@ -60,7 +60,7 @@ Die vier ursprünglich offenen Punkte wurden inzwischen in SPEC.md/PLAN.md beant
3. **Start-Reihenfolge:** Alle Sites gleichzeitig (PLAN.md §1: "launch simultaneously") 3. **Start-Reihenfolge:** Alle Sites gleichzeitig (PLAN.md §1: "launch simultaneously")
4. **Platzhalter-Verhalten:** Keine Platzhalter, ehrliche kurze Seiten + WIP-Hinweise (SPEC.md §5, retro/ hat WIP-Badge) 4. **Platzhalter-Verhalten:** Keine Platzhalter, ehrliche kurze Seiten + WIP-Hinweise (SPEC.md §5, retro/ hat WIP-Badge)
## 9. Gesamtstand — Konzept vollständig ## 9. Gesamtstand, Konzept vollständig
Alle Entscheidungsgrundlagen sind in SPEC.md (was) und PLAN.md (wie/wann) umgesetzt: Alle Entscheidungsgrundlagen sind in SPEC.md (was) und PLAN.md (wie/wann) umgesetzt:
PLAN.md
+12 -12
View File
@@ -1,30 +1,30 @@
# moonweb.org — Implementation Plan # moonweb.org, Implementation Plan
Companion to `SPEC.md`. Documents the completed migration from separate subdomain sites to a unified `www.moonweb.org` subdirectory structure. Companion to `SPEC.md`. Documents the completed migration from separate subdomain sites to a unified `www.moonweb.org` subdirectory structure.
## Phase 0 — Repo & tooling setup ## Phase 0, Repo & tooling setup
1. Created the `moonweb-site` monorepo. 1. Created the `moonweb-site` monorepo.
2. Initialized Eleventy project structure per `SPEC.md §10`. 2. Initialized Eleventy project structure per `SPEC.md §10`.
3. Set up `shared/_includes/base.njk` (header + card-grid) with accent colors inlined. 3. Set up `shared/_includes/base.njk` (header + card-grid) with accent colors inlined.
4. Verified local dev server works per site. 4. Verified local dev server works per site.
## Phase 1 — Content migration & authoring (per domain) ## Phase 1, Content migration & authoring (per domain)
1. **code** — GitHub aggregator, `_data/repos.json`, overview page grouped by subcategory. 1. **code**, GitHub aggregator, `_data/repos.json`, overview page grouped by subcategory.
2. **smarthome** — overview + detail pages where content exists. 2. **smarthome**, overview + detail pages where content exists.
3. **infra** — shallow overview with redaction pass (no IPs, keys, passwords). 3. **infra**, shallow overview with redaction pass (no IPs, keys, passwords).
4. **retro** — minimal overview with WIP notices. 4. **retro**, minimal overview with WIP notices.
5. **hub** — central index linking all sites, including redirect `.htm` files for old www.moonweb.org paths. 5. **hub**, central index linking all sites, including redirect `.htm` files for old www.moonweb.org paths.
6. **timecapsule** — integrated from moonweb-www, preserves original 2001 design under `/timecapsule/`. 6. **timecapsule**, integrated from moonweb-www, preserves original 2001 design under `/timecapsule/`.
## Phase 2 — CI/CD ## Phase 2, CI/CD
1. GitHub Actions workflow `build-deploy-moonweb.yml`: single Eleventy build for all moonweb sites + timecapsule build, deploy via IONOS SFTP. 1. GitHub Actions workflow `build-deploy-moonweb.yml`: single Eleventy build for all moonweb sites + timecapsule build, deploy via IONOS SFTP.
2. GitHub Actions workflow `deploy-stefankoelle.yml`: stefankoelle.de build, deploy via IONOS SFTP. 2. GitHub Actions workflow `deploy-stefankoelle.yml`: stefankoelle.de build, deploy via IONOS SFTP.
3. Cloudflare redirect rules for old subdomains (`scripts/cloudflare/`). 3. Cloudflare redirect rules for old subdomains (`scripts/cloudflare/`).
## Phase 3 — Consolidation to www.moonweb.org ## Phase 3, Consolidation to www.moonweb.org
All moonweb.org sites now live under `www.moonweb.org` as subdirectories: All moonweb.org sites now live under `www.moonweb.org` as subdirectories:
@@ -55,7 +55,7 @@ Cloudflare redirects forward old subdomains:
- `.eleventyignore` excludes `stefankoelle/` and `timecapsule/` (built separately). - `.eleventyignore` excludes `stefankoelle/` and `timecapsule/` (built separately).
- `build-pdf.sh` temporarily renames `.eleventyignore` for stefankoelle build. - `build-pdf.sh` temporarily renames `.eleventyignore` for stefankoelle build.
## Definition of done — Achieved ## Definition of done, Achieved
- All sites live under `www.moonweb.org` as subdirectories, deployed via IONOS SFTP. - All sites live under `www.moonweb.org` as subdirectories, deployed via IONOS SFTP.
- `www.moonweb.org/code/` reflects the current GitHub repos via the `.moonweb.yml` aggregator. - `www.moonweb.org/code/` reflects the current GitHub repos via the `.moonweb.yml` aggregator.
README.md
+1 -1
View File
@@ -1,6 +1,6 @@
# moonweb-site # moonweb-site
Monorepo for the **moonweb.org** homelab — six static sites built with [Eleventy](https://www.11ty.dev/), deployed to [IONOS SFTP](https://www.ionos.de/). Monorepo for the **moonweb.org** homelab, six static sites built with [Eleventy](https://www.11ty.dev/), deployed to [IONOS SFTP](https://www.ionos.de/).
``` ```
www.moonweb.org/ -> Central index & gateway www.moonweb.org/ -> Central index & gateway
SPEC.md
+20 -20
View File
@@ -1,4 +1,4 @@
# moonweb.org — Specification # moonweb.org, Specification
Status: Migration complete. All sites live under `www.moonweb.org` as subdirectories, deployed via IONOS SFTP. Status: Migration complete. All sites live under `www.moonweb.org` as subdirectories, deployed via IONOS SFTP.
@@ -36,9 +36,9 @@ Old subdomain redirects (via Cloudflare):
|---|---|---|---| |---|---|---|---|
| hub | `/` | Gateway, links to everything, one-line description per destination | Minimal | | hub | `/` | Gateway, links to everything, one-line description per destination | Minimal |
| infra | `/infra/` | Shallow, structured overview of the stack: Proxmox, Synology, VLANs, Docker hosting, dev environment | Reference, high-level only | | infra | `/infra/` | Shallow, structured overview of the stack: Proxmox, Synology, VLANs, Docker hosting, dev environment | Reference, high-level only |
| smarthome | `/smarthome/` | Why the homelab exists — sensors, automation, calendar/contacts sync, dashboards, media | Project storytelling | | smarthome | `/smarthome/` | Why the homelab exists, sensors, automation, calendar/contacts sync, dashboards, media | Project storytelling |
| code | `/code/` | Curated, sorted GitHub catalog — overview only, always linking out to GitHub | Portfolio | | code | `/code/` | Curated, sorted GitHub catalog, overview only, always linking out to GitHub | Portfolio |
| retro | `/retro/` | Physical retro hardware collection (not software/demos — that's 28k8's domain) | Simple, factual | | retro | `/retro/` | Physical retro hardware collection (not software/demos, that's 28k8's domain) | Simple, factual |
| timecapsule | `/timecapsule/` | Original www.moonweb.org content from 2001, preserved as-is | Retro 2001 design | | timecapsule | `/timecapsule/` | Original www.moonweb.org content from 2001, preserved as-is | Retro 2001 design |
## 4. Design system ## 4. Design system
@@ -49,7 +49,7 @@ Old subdomain redirects (via Cloudflare):
- **Overview (index) pages** use a shared card-grid layout (clean card-grid with banner header, grouped card sections, sans-serif, generous whitespace, light theme only, no heavy JS). - **Overview (index) pages** use a shared card-grid layout (clean card-grid with banner header, grouped card sections, sans-serif, generous whitespace, light theme only, no heavy JS).
- **Detail/sub-pages** keep the same header but may use a freer layout below it. - **Detail/sub-pages** keep the same header but may use a freer layout below it.
Note: stefankoelle.de uses its own independent design — it does not follow this header-consistent principle. Note: stefankoelle.de uses its own independent design, it does not follow this header-consistent principle.
### 4.2 Accent colors per domain ### 4.2 Accent colors per domain
@@ -63,31 +63,31 @@ Note: stefankoelle.de uses its own independent design — it does not follow thi
### 4.3 URL convention ### 4.3 URL convention
`/section/slug/` — lowercase, hyphenated, trailing slash. All sites share a single Eleventy build with computed `pathPrefix` per section. `/section/slug/`, lowercase, hyphenated, trailing slash. All sites share a single Eleventy build with computed `pathPrefix` per section.
## 5. Content depth rules (per domain) ## 5. Content depth rules (per domain)
| Domain | Detail pages? | Rule | | Domain | Detail pages? | Rule |
|---|---|---| |---|---|---|
| smarthome | Yes, when there's enough content | Index shows how the smart home is structured; a topic gets a detail page immediately if enough data exists — otherwise it's mentioned in the overview only, no placeholder required | | smarthome | Yes, when there's enough content | Index shows how the smart home is structured; a topic gets a detail page immediately if enough data exists, otherwise it's mentioned in the overview only, no placeholder required |
| code | Never | Overview cards + link to GitHub only. No duplicating README content. | | code | Never | Overview cards + link to GitHub only. No duplicating README content. |
| infra | Rarely | Deliberately shallow — most of the raw material is sensitive (see §6) | | infra | Rarely | Deliberately shallow, most of the raw material is sensitive (see §6) |
| retro | Yes, but minimal | Topic is still immature; create only a rudimentary overview, don't over-invest time | | retro | Yes, but minimal | Topic is still immature; create only a rudimentary overview, don't over-invest time |
General rule across all domains: **if content is too thin for a good detail page, skip the detail page — don't create a placeholder.** General rule across all domains: **if content is too thin for a good detail page, skip the detail page, don't create a placeholder.**
## 6. Infra content redaction rule ## 6. Infra content redaction rule
Because infra must stay shallow and public-safe: Because infra must stay shallow and public-safe:
- **Allowed:** architecture level — Proxmox + Synology + Docker host, VLAN concept without concrete internal IP plans, which service types run, which tools are used. - **Allowed:** architecture level, Proxmox + Synology + Docker host, VLAN concept without concrete internal IP plans, which service types run, which tools are used.
- **Not allowed:** concrete IP addresses, WireGuard keys/preshared keys, passwords, internal hostnames that allow inference, backup targets with credentials. - **Not allowed:** concrete IP addresses, WireGuard keys/preshared keys, passwords, internal hostnames that allow inference, backup targets with credentials.
This rule applies to any domain but is most relevant for infra, since the source documents currently contain real IPs and keys that must be actively stripped during migration. This rule applies to any domain but is most relevant for infra, since the source documents currently contain real IPs and keys that must be actively stripped during migration.
## 7. Language ## 7. Language
All sites in the monorepo are written **entirely in English**. Existing German source documents are translated once during migration via a single AI-assisted pass — not a recurring process. New `.moonweb.yml` metadata and generated content are authored in English from the start. All sites in the monorepo are written **entirely in English**. Existing German source documents are translated once during migration via a single AI-assisted pass, not a recurring process. New `.moonweb.yml` metadata and generated content are authored in English from the start.
## 8. GitHub automation (www.moonweb.org/code/) ## 8. GitHub automation (www.moonweb.org/code/)
@@ -104,13 +104,13 @@ summary: "Compact MVG/S-Bahn departure monitor with configurable stations."
repo_url: "https://github.com/skoelle/mvg-departures" repo_url: "https://github.com/skoelle/mvg-departures"
``` ```
A local aggregator script reads `.moonweb.yml` from all repos via the GitHub API, and the result is committed into `_data/repos.json` inside the monorepo. This runs manually, on demand — no scheduled automation for now. A local aggregator script reads `.moonweb.yml` from all repos via the GitHub API, and the result is committed into `_data/repos.json` inside the monorepo. This runs manually, on demand, no scheduled automation for now.
## 9. Content maintenance ## 9. Content maintenance
- infra, smarthome, retro, stefankoelle: **fully manual**, edited in vim, committed via git push. No automation. - infra, smarthome, retro, stefankoelle: **fully manual**, edited in vim, committed via git push. No automation.
- code: the only automated piece is the GitHub aggregator described in §8. - code: the only automated piece is the GitHub aggregator described in §8.
- No "last updated" timestamps are shown anywhere — the goal is that content is simply kept current, not that staleness is displayed. - No "last updated" timestamps are shown anywhere, the goal is that content is simply kept current, not that staleness is displayed.
- No analytics/tracking of any kind on any site. - No analytics/tracking of any kind on any site.
- Images/assets live versioned directly in the monorepo (no external asset host). - Images/assets live versioned directly in the monorepo (no external asset host).
@@ -159,19 +159,19 @@ moonweb-site/
## 11. Technical stack ## 11. Technical stack
- **Static site generator:** Eleventy (11ty) v3.1.6 — single config at root, computed `pathPrefix` per section, per-section collections. - **Static site generator:** Eleventy (11ty) v3.1.6, single config at root, computed `pathPrefix` per section, per-section collections.
- **Static site generator (timecapsule):** Eleventy v2.0.1 — separate config, preserves original 2001 design. - **Static site generator (timecapsule):** Eleventy v2.0.1, separate config, preserves original 2001 design.
- **Templates:** Nunjucks (.njk) — shared `base.njk` layout, `card-grid.njk` for index pages, `sitemap.njk` for central sitemap. - **Templates:** Nunjucks (.njk), shared `base.njk` layout, `card-grid.njk` for index pages, `sitemap.njk` for central sitemap.
- **Styling:** Custom CSS (`base.css`) with accent colors inlined as `<style>` in `base.njk`. No per-site theme CSS files. - **Styling:** Custom CSS (`base.css`) with accent colors inlined as `<style>` in `base.njk`. No per-site theme CSS files.
- **Build:** entirely in GitHub Actions. - **Build:** entirely in GitHub Actions.
- **Deploy:** IONOS SFTP — all sites deployed to `/websites/moonweb/`, stefankoelle.de to `/websites/stefankoelle/`. - **Deploy:** IONOS SFTP, all sites deployed to `/websites/moonweb/`, stefankoelle.de to `/websites/stefankoelle/`.
- **DNS:** Cloudflare — DNS management + redirects from old subdomains (hub.moonweb.org, etc.). - **DNS:** Cloudflare, DNS management + redirects from old subdomains (hub.moonweb.org, etc.).
- **Runtime:** fully static, no server-side code, no containers for the website itself. - **Runtime:** fully static, no server-side code, no containers for the website itself.
- **Local preview:** Eleventy's built-in dev server with live reload (`npm run dev` for all moonweb sites, `npm run dev:stefankoelle`, `npm run dev:timecapsule`). - **Local preview:** Eleventy's built-in dev server with live reload (`npm run dev` for all moonweb sites, `npm run dev:stefankoelle`, `npm run dev:timecapsule`).
## 12. Explicitly out of scope ## 12. Explicitly out of scope
- Migrating 28k8.moonweb.org into the monorepo — undecided, revisit later, likely never. - Migrating 28k8.moonweb.org into the monorepo, undecided, revisit later, likely never.
- Any Perplexity-backchannel mechanism — deferred. - Any Perplexity-backchannel mechanism, deferred.
- Analytics of any kind. - Analytics of any kind.
- Automated content generation/translation pipelines beyond the one-time GitHub aggregator for code. - Automated content generation/translation pipelines beyond the one-time GitHub aggregator for code.
code/index.njk
+2 -2
View File
@@ -1,8 +1,8 @@
--- ---
title: "Open Source Projects — GitHub Repos & Code | moonweb" title: "Open Source Projects, GitHub Repos & Code | moonweb"
section: "code" section: "code"
tags: "code" tags: "code"
description: "Curated catalog of my public GitHub projects — automation, dev tools, web apps, and firmware." description: "Curated catalog of my public GitHub projects, automation, dev tools, web apps, and firmware."
layout: base.njk layout: base.njk
subcategories: subcategories:
- "Smart Home Apps" - "Smart Home Apps"
hub/impressum.njk
+1 -1
View File
@@ -104,7 +104,7 @@ layout: base.njk
<h3>7. Changes to this Policy</h3> <h3>7. Changes to this Policy</h3>
<p> <p>
I may update this policy if the setup of any website in this network changes — for example, if a visit counter or analytics tool is added to stefankoelle.de or to any moonweb.org subdomain. Any such change will be documented here before it goes live. I may update this policy if the setup of any website in this network changes, for example, if a visit counter or analytics tool is added to stefankoelle.de or to any moonweb.org subdomain. Any such change will be documented here before it goes live.
</p> </p>
<p><em>Last updated: August 2026</em></p> <p><em>Last updated: August 2026</em></p>
hub/index.njk
+3 -3
View File
@@ -1,9 +1,9 @@
--- ---
title: "Stefan Koelle — Homelab, Smart Home, Code & Retro Hardware | moonweb" title: "Stefan Koelle, Homelab, Smart Home, Code & Retro Hardware | moonweb"
section: "hub" section: "hub"
tags: "hub" tags: "hub"
permalink: /index.html permalink: /index.html
description: "Stefan Koelle — Senior Software Developer & Architect in Munich. Homelab projects, smart home automation, open-source code, and retro hardware." description: "Stefan Koelle, Senior Software Developer & Architect in Munich. Homelab projects, smart home automation, open-source code, and retro hardware."
layout: base.njk layout: base.njk
ogImage: "/stefan-koelle-foto.jpg" ogImage: "/stefan-koelle-foto.jpg"
sections: sections:
@@ -38,7 +38,7 @@ sections:
href: "https://28k8.moonweb.org" href: "https://28k8.moonweb.org"
emoji: "💾" emoji: "💾"
- title: "moonweb.org 2001" - title: "moonweb.org 2001"
summary: "The original www.moonweb.org as it was in 2001 — preserved as a time capsule." summary: "The original www.moonweb.org as it was in 2001, preserved as a time capsule."
href: "/timecapsule/" href: "/timecapsule/"
emoji: "🌐" emoji: "🌐"
- title: "inseco.de" - title: "inseco.de"
infra/backup-strategy/index.njk
+7 -7
View File
@@ -11,10 +11,10 @@ layout: base.njk
<p>The homelab uses a layered backup approach, split into two independent <p>The homelab uses a layered backup approach, split into two independent
tracks: NAS user content, and Proxmox VM/LXC backups. Both follow a tracks: NAS user content, and Proxmox VM/LXC backups. Both follow a
3-2-1-style rule — multiple copies, multiple media types, at least one 3-2-1-style rule, multiple copies, multiple media types, at least one
copy on a different machine.</p> copy on a different machine.</p>
<h2>NAS user content — four layers</h2> <h2>NAS user content, four layers</h2>
<table> <table>
<tr><th>Layer</th><th>Target</th><th>Frequency</th><th>Purpose</th></tr> <tr><th>Layer</th><th>Target</th><th>Frequency</th><th>Purpose</th></tr>
<tr><td>Snapshots</td><td>Synology, built-in</td><td>Hourly</td><td>Fast recovery of accidentally deleted files</td></tr> <tr><td>Snapshots</td><td>Synology, built-in</td><td>Hourly</td><td>Fast recovery of accidentally deleted files</td></tr>
@@ -23,17 +23,17 @@ copy on a different machine.</p>
<tr><td>Cloud archive</td><td>Cloud storage</td><td>Continuous</td><td>Off-site copy for worst-case scenarios (fire, theft)</td></tr> <tr><td>Cloud archive</td><td>Cloud storage</td><td>Continuous</td><td>Off-site copy for worst-case scenarios (fire, theft)</td></tr>
</table> </table>
<h2>Proxmox VM/LXC backups — three copies, two machines</h2> <h2>Proxmox VM/LXC backups, three copies, two machines</h2>
<table> <table>
<tr><th>Copy</th><th>Target</th><th>Frequency</th></tr> <tr><th>Copy</th><th>Target</th><th>Frequency</th></tr>
<tr><td>1 — original</td><td>Proxmox host itself</td><td>Daily</td></tr> <tr><td>1, original</td><td>Proxmox host itself</td><td>Daily</td></tr>
<tr><td>2 — primary offsite</td><td>Synology NAS (different machine)</td><td>Daily, right after copy 1</td></tr> <tr><td>2, primary offsite</td><td>Synology NAS (different machine)</td><td>Daily, right after copy 1</td></tr>
<tr><td>3 — secondary offsite</td><td>Raspberry Pi backup target, different room</td><td>Daily (skipped one day/week to avoid network contention with the weekly NAS sync)</td></tr> <tr><td>3, secondary offsite</td><td>Raspberry Pi backup target, different room</td><td>Daily (skipped one day/week to avoid network contention with the weekly NAS sync)</td></tr>
</table> </table>
<h2>Design principles</h2> <h2>Design principles</h2>
<ul> <ul>
<li>Backups always cross at least one machine boundary — nothing is considered "backed up" if it only lives on the source machine's local disk.</li> <li>Backups always cross at least one machine boundary, nothing is considered "backed up" if it only lives on the source machine's local disk.</li>
<li>Weekly and daily jobs are scheduled to avoid saturating the internal power-line network link used by the offsite backup target.</li> <li>Weekly and daily jobs are scheduled to avoid saturating the internal power-line network link used by the offsite backup target.</li>
<li>All scheduled jobs ping a dead-man's-switch monitoring service; a missed backup triggers an alert.</li> <li>All scheduled jobs ping a dead-man's-switch monitoring service; a missed backup triggers an alert.</li>
<li>Power to the offsite drive is only switched on for the duration of its weekly sync, to save power and extend disk life.</li> <li>Power to the offsite drive is only switched on for the duration of its weekly sync, to save power and extend disk life.</li>
infra/dev-environment/index.njk
+12 -12
View File
@@ -9,29 +9,29 @@ layout: base.njk
<h1>Xubuntu Dev VM</h1> <h1>Xubuntu Dev VM</h1>
<div class="detail-content"> <div class="detail-content">
<p>This is a Xubuntu virtual machine on <a href="/proxmox/">Proxmox pve2</a> in the study that serves as my main development environment, replacing the need for a dedicated physical workstation. It's designed to be lightweight, remotely accessible from anywhere, and — most importantly — completely disposable, which is what makes it a safe place to let AI coding agents operate with far more autonomy than I'd otherwise allow.</p> <p>This is a Xubuntu virtual machine on <a href="/proxmox/">Proxmox pve2</a> in the study that serves as my main development environment, replacing the need for a dedicated physical workstation. It's designed to be lightweight, remotely accessible from anywhere, and, most importantly, completely disposable, which is what makes it a safe place to let AI coding agents operate with far more autonomy than I'd otherwise allow.</p>
<h2>Why a disposable VM is the biggest advantage</h2> <h2>Why a disposable VM is the biggest advantage</h2>
<ul> <ul>
<li><strong>Daily backups as a safety net</strong> — the VM is backed up once a day through Proxmox. If an AI agent breaks the system, misconfigures something critical, or deletes the wrong files, I simply restore yesterday's snapshot rather than manually debugging the damage.</li> <li><strong>Daily backups as a safety net</strong>, the VM is backed up once a day through Proxmox. If an AI agent breaks the system, misconfigures something critical, or deletes the wrong files, I simply restore yesterday's snapshot rather than manually debugging the damage.</li>
<li><strong>No irreplaceable local data</strong> — nothing important lives only on this VM. All actual project code lives in Git repositories, so after a restore I just <code>git pull</code> everything back and I'm exactly where I left off, minus whatever the agent broke.</li> <li><strong>No irreplaceable local data</strong>, nothing important lives only on this VM. All actual project code lives in Git repositories, so after a restore I just <code>git pull</code> everything back and I'm exactly where I left off, minus whatever the agent broke.</li>
<li><strong>Higher trust, more autonomy for AI agents</strong> — because the worst-case outcome is "restore a snapshot and re-pull," I can give coding agents like OpenCode much broader permissions on this machine than I would on a system holding unique, unbacked-up state. The blast radius of a mistake is capped at "lose a few hours of uncommitted work," never at "lose data."</li> <li><strong>Higher trust, more autonomy for AI agents</strong>, because the worst-case outcome is "restore a snapshot and re-pull," I can give coding agents like OpenCode much broader permissions on this machine than I would on a system holding unique, unbacked-up state. The blast radius of a mistake is capped at "lose a few hours of uncommitted work," never at "lose data."</li>
<li><strong>Cheap to rebuild from scratch</strong> — if the VM ever gets into a truly bad state, standing up a fresh Xubuntu VM and re-cloning repos is a short, well-understood process rather than an emergency.</li> <li><strong>Cheap to rebuild from scratch</strong>, if the VM ever gets into a truly bad state, standing up a fresh Xubuntu VM and re-cloning repos is a short, well-understood process rather than an emergency.</li>
</ul> </ul>
<h2>Remote workflow via SSH</h2> <h2>Remote workflow via SSH</h2>
<ul> <ul>
<li><strong><a href="https://herdr.dev/">Herdr</a></strong> — manages my development workspaces over SSH, letting me organize and switch between multiple project contexts on the VM without manually juggling terminal sessions.</li> <li><strong><a href="https://herdr.dev/">Herdr</a></strong>, manages my development workspaces over SSH, letting me organize and switch between multiple project contexts on the VM without manually juggling terminal sessions.</li>
<li><strong><a href="https://opencode.ai/de">OpenCode</a></strong> — the AI coding agent I run inside these workspaces; it's the main reason the VM's disposability matters, since it can act with elevated freedom knowing any damage is trivially reversible.</li> <li><strong><a href="https://opencode.ai/de">OpenCode</a></strong>, the AI coding agent I run inside these workspaces; it's the main reason the VM's disposability matters, since it can act with elevated freedom knowing any damage is trivially reversible.</li>
<li><strong><a href="https://github.com/jesseduffield/lazygit">lazygit</a></strong> — terminal UI for Git, used for fast commit/branch/diff workflows directly inside the SSH session without leaving the terminal.</li> <li><strong><a href="https://github.com/jesseduffield/lazygit">lazygit</a></strong>, terminal UI for Git, used for fast commit/branch/diff workflows directly inside the SSH session without leaving the terminal.</li>
</ul> </ul>
<h2>Graphical fallback via Chrome Remote Desktop</h2> <h2>Graphical fallback via Chrome Remote Desktop</h2>
<ul> <ul>
<li><strong>Full desktop access</strong> — when a GUI IDE is preferable to terminal-based editing, Chrome Remote Desktop connects straight into the Xubuntu desktop from any device.</li> <li><strong>Full desktop access</strong>, when a GUI IDE is preferable to terminal-based editing, Chrome Remote Desktop connects straight into the Xubuntu desktop from any device.</li>
<li><strong>WebStorm</strong> — used for TypeScript projects that benefit from a full IDE (refactoring tools, debugging, integrated test runners).</li> <li><strong>WebStorm</strong>, used for TypeScript projects that benefit from a full IDE (refactoring tools, debugging, integrated test runners).</li>
<li><strong>PyCharm</strong> — used the same way for Python projects.</li> <li><strong>PyCharm</strong>, used the same way for Python projects.</li>
</ul> </ul>
<p class="redacted-note">This setup deliberately treats the dev VM as ephemeral infrastructure, not a pet — the philosophy is: keep state in Git, keep the VM replaceable, and let daily backups absorb the risk of giving an AI agent real write access to the system.</p> <p class="redacted-note">This setup deliberately treats the dev VM as ephemeral infrastructure, not a pet, the philosophy is: keep state in Git, keep the VM replaceable, and let daily backups absorb the risk of giving an AI agent real write access to the system.</p>
</div> </div>
infra/docker/index.njk
+29 -29
View File
@@ -9,62 +9,62 @@ layout: base.njk
<h1>Docker Container Strategy</h1> <h1>Docker Container Strategy</h1>
<div class="detail-content"> <div class="detail-content">
<p>My homelab runs almost entirely on Docker, spread across three dedicated Debian VMs instead of installing containers directly on bare metal or on the NAS operating system itself. Each VM is a clean, disposable Docker host that can be snapshotted and backed up as a whole through Proxmox, and every service is defined declaratively as a Docker Compose file living under <code>/docker-data/compose/&lt;service&gt;/</code>. This keeps the setup portable, reproducible, and easy to document — if a host dies, restoring a VM snapshot brings every container definition back with it.</p> <p>My homelab runs almost entirely on Docker, spread across three dedicated Debian VMs instead of installing containers directly on bare metal or on the NAS operating system itself. Each VM is a clean, disposable Docker host that can be snapshotted and backed up as a whole through Proxmox, and every service is defined declaratively as a Docker Compose file living under <code>/docker-data/compose/&lt;service&gt;/</code>. This keeps the setup portable, reproducible, and easy to document, if a host dies, restoring a VM snapshot brings every container definition back with it.</p>
<h2>The three Docker hosts</h2> <h2>The three Docker hosts</h2>
<ul> <ul>
<li><strong>docker-host-pve</strong> — Debian VM on the main Proxmox server (PVE). This is the primary target for almost all container deployments, including my own images pulled from GHCR. Fully covered by Proxmox VM backups, so the whole host can be restored in one shot.</li> <li><strong>docker-host-pve</strong>, Debian VM on the main Proxmox server (PVE). This is the primary target for almost all container deployments, including my own images pulled from GHCR. Fully covered by Proxmox VM backups, so the whole host can be restored in one shot.</li>
<li><strong>docker-host-nas</strong> — Debian VM running directly on the Synology (via Virtual Machine Manager). Identical setup and folder structure to the PVE host, but used specifically for containers that need direct access to NAS storage volumes, such as media-processing tools that read and write large file libraries.</li> <li><strong>docker-host-nas</strong>, Debian VM running directly on the Synology (via Virtual Machine Manager). Identical setup and folder structure to the PVE host, but used specifically for containers that need direct access to NAS storage volumes, such as media-processing tools that read and write large file libraries.</li>
<li><strong>docker-host-pve2</strong> — Debian VM on the secondary Proxmox server (pve2), which lives in the study and is treated as a pure test/dev environment. New containers and compose setups get trialed here before being promoted to docker-host-pve.</li> <li><strong>docker-host-pve2</strong>, Debian VM on the secondary Proxmox server (pve2), which lives in the study and is treated as a pure test/dev environment. New containers and compose setups get trialed here before being promoted to docker-host-pve.</li>
</ul> </ul>
<h2>Standardized layout</h2> <h2>Standardized layout</h2>
<ul> <ul>
<li><strong>/docker-data/compose/</strong> — one subfolder per service, each containing its own <code>docker-compose.yml</code> (and <code>.env</code> where needed).</li> <li><strong>/docker-data/compose/</strong>, one subfolder per service, each containing its own <code>docker-compose.yml</code> (and <code>.env</code> where needed).</li>
<li><strong>/docker-data/volumes/</strong> — persistent container data, kept outside the compose folders so it survives redeploys.</li> <li><strong>/docker-data/volumes/</strong>, persistent container data, kept outside the compose folders so it survives redeploys.</li>
<li><strong>/docker-data/backups/</strong> — local backup staging before data is pulled into the wider backup chain.</li> <li><strong>/docker-data/backups/</strong>, local backup staging before data is pulled into the wider backup chain.</li>
<li><strong>Same convention everywhere</strong> — because all three hosts follow this identical structure, moving a service between hosts or rebuilding a host from scratch is just a matter of copying the compose folder and running <code>docker compose up -d</code>.</li> <li><strong>Same convention everywhere</strong>, because all three hosts follow this identical structure, moving a service between hosts or rebuilding a host from scratch is just a matter of copying the compose folder and running <code>docker compose up -d</code>.</li>
</ul> </ul>
<h2>Own GitHub projects</h2> <h2>Own GitHub projects</h2>
<ul> <ul>
<li><strong>icloud-contacts-sync</strong> — syncs contact data between iCloud and other systems.</li> <li><strong>icloud-contacts-sync</strong>, syncs contact data between iCloud and other systems.</li>
<li><strong>calendar-sync</strong> — keeps calendars synchronized across sources.</li> <li><strong>calendar-sync</strong>, keeps calendars synchronized across sources.</li>
<li><strong>mvg-departures</strong> — pulls Munich public transport (MVG) departure data for local dashboards.</li> <li><strong>mvg-departures</strong>, pulls Munich public transport (MVG) departure data for local dashboards.</li>
<li><strong>wetter-api</strong> — a small weather API/worker pair, built from a custom GHCR image and updated automatically via Watchtower.</li> <li><strong>wetter-api</strong>, a small weather API/worker pair, built from a custom GHCR image and updated automatically via Watchtower.</li>
</ul> </ul>
<h2>Infrastructure containers</h2> <h2>Infrastructure containers</h2>
<ul> <ul>
<li><strong>mosquitto-mqtt</strong> — central MQTT broker for smart-home and sensor messaging.</li> <li><strong>mosquitto-mqtt</strong>, central MQTT broker for smart-home and sensor messaging.</li>
<li><strong>grafana, prometheus, influxdb</strong> — the monitoring and metrics stack for dashboards and time-series data.</li> <li><strong>grafana, prometheus, influxdb</strong>, the monitoring and metrics stack for dashboards and time-series data.</li>
<li><strong>healthchecks, uptime-kuma, cadvisor</strong> — service and container health/uptime monitoring.</li> <li><strong>healthchecks, uptime-kuma, cadvisor</strong>, service and container health/uptime monitoring.</li>
<li><strong>dyndns-updater</strong> — keeps DNS pointed at the home connection; my own project, see <a href="https://github.com/skoelle/dyndns-updater">skoelle/dyndns-updater</a> on GitHub.</li> <li><strong>dyndns-updater</strong>, keeps DNS pointed at the home connection; my own project, see <a href="https://github.com/skoelle/dyndns-updater">skoelle/dyndns-updater</a> on GitHub.</li>
<li><strong>watchtower</strong> — automatically checks for and applies container image updates on a daily schedule.</li> <li><strong>watchtower</strong>, automatically checks for and applies container image updates on a daily schedule.</li>
<li><strong>authelia, nginx-proxy-manager</strong> — authentication layer and reverse proxy for exposing internal services safely.</li> <li><strong>authelia, nginx-proxy-manager</strong>, authentication layer and reverse proxy for exposing internal services safely.</li>
<li><strong>postfix</strong> — internal mail relay used by other containers (e.g. Watchtower notifications) to send email.</li> <li><strong>postfix</strong>, internal mail relay used by other containers (e.g. Watchtower notifications) to send email.</li>
<li><strong>portainer</strong> — web UI for managing containers across hosts.</li> <li><strong>portainer</strong>, web UI for managing containers across hosts.</li>
<li><strong>exporters</strong> — Prometheus exporters for Fritz!Box, Zyxel switches, Nginx, and SNMP devices, feeding the Grafana/Prometheus stack.</li> <li><strong>exporters</strong>, Prometheus exporters for Fritz!Box, Zyxel switches, Nginx, and SNMP devices, feeding the Grafana/Prometheus stack.</li>
</ul> </ul>
<h2>Smart home containers</h2> <h2>Smart home containers</h2>
<ul> <ul>
<li><strong>Home Assistant</strong> — only used for homematicIP integration to mqtt.</li> <li><strong>Home Assistant</strong>, only used for homematicIP integration to mqtt.</li>
</ul> </ul>
<h2>Application containers</h2> <h2>Application containers</h2>
<ul> <ul>
<li><strong>SearXNG</strong> — self-hosted, privacy-respecting metasearch engine that aggregates results from multiple search providers without leaking queries to any of them. Used as the default search backend for local browsing and AI tools.</li> <li><strong>SearXNG</strong>, self-hosted, privacy-respecting metasearch engine that aggregates results from multiple search providers without leaking queries to any of them. Used as the default search backend for local browsing and AI tools.</li>
<li><strong>Open WebUI</strong> — chat interface for locally hosted LLMs.</li> <li><strong>Open WebUI</strong>, chat interface for locally hosted LLMs.</li>
<li><strong>Stirling PDF</strong> — self-hosted PDF toolkit for merging, splitting, converting, compressing, and editing documents without uploading them to external services.</li> <li><strong>Stirling PDF</strong>, self-hosted PDF toolkit for merging, splitting, converting, compressing, and editing documents without uploading them to external services.</li>
<li><strong>ArchiveBox</strong> — personal web archiving of bookmarked pages.</li> <li><strong>ArchiveBox</strong>, personal web archiving of bookmarked pages.</li>
<li><strong>TubeSync / TubeArchivist [NAS]</strong> — download and archive YouTube content directly onto NAS storage; these run on docker-host-nas specifically because they need direct volume access to the Synology's disks.</li> <li><strong>TubeSync / TubeArchivist [NAS]</strong>, download and archive YouTube content directly onto NAS storage; these run on docker-host-nas specifically because they need direct volume access to the Synology's disks.</li>
<li><strong>Music Stretto</strong> — music streaming server for the local MP3 library on the Synology, accessible from the home dashboard.</li> <li><strong>Music Stretto</strong>, music streaming server for the local MP3 library on the Synology, accessible from the home dashboard.</li>
</ul> </ul>
<h2>Development tools</h2> <h2>Development tools</h2>
<ul> <ul>
<li><strong>Gitea</strong> — self-hosted Git server acting as a mirror of my GitHub repositories, giving me a local fallback and faster internal access.</li> <li><strong>Gitea</strong>, self-hosted Git server acting as a mirror of my GitHub repositories, giving me a local fallback and faster internal access.</li>
</ul> </ul>
</div> </div>
infra/index.njk
+1 -1
View File
@@ -1,5 +1,5 @@
--- ---
title: "Homelab Infrastructure — Proxmox, Docker & Networking | moonweb" title: "Homelab Infrastructure, Proxmox, Docker & Networking | moonweb"
section: "infra" section: "infra"
tags: "infra" tags: "infra"
description: "Homelab infrastructure overview: Proxmox hosts, Synology NAS, Docker VMs and Setup-Guides." description: "Homelab infrastructure overview: Proxmox hosts, Synology NAS, Docker VMs and Setup-Guides."
infra/lxc/index.njk
+15 -15
View File
@@ -9,36 +9,36 @@ layout: base.njk
<h1>LXC Container Strategy</h1> <h1>LXC Container Strategy</h1>
<div class="detail-content"> <div class="detail-content">
<p>Alongside the Docker-based setup, I run a small number of Linux Containers (LXC) directly on the main Proxmox host (PVE). An LXC is an OS-level virtualization container — unlike a Docker container, which packages a single application and its dependencies, an LXC behaves like a full lightweight Linux system with its own init process, systemd services and package manager, but without the overhead of a full virtual machine. Proxmox manages LXCs natively as first-class citizens, right next to VMs, with their own snapshotting, backup and resource-limit tooling. I use far fewer LXCs than Docker containers, reserving them for cases where a persistent, OS-like environment or tight integration with Proxmox itself makes more sense than a containerized app.</p> <p>Alongside the Docker-based setup, I run a small number of Linux Containers (LXC) directly on the main Proxmox host (PVE). An LXC is an OS-level virtualization container, unlike a Docker container, which packages a single application and its dependencies, an LXC behaves like a full lightweight Linux system with its own init process, systemd services and package manager, but without the overhead of a full virtual machine. Proxmox manages LXCs natively as first-class citizens, right next to VMs, with their own snapshotting, backup and resource-limit tooling. I use far fewer LXCs than Docker containers, reserving them for cases where a persistent, OS-like environment or tight integration with Proxmox itself makes more sense than a containerized app.</p>
<h2>Why LXC instead of Docker here</h2> <h2>Why LXC instead of Docker here</h2>
<ul> <ul>
<li><strong>Native Proxmox integration</strong> — LXCs show up directly in the Proxmox UI with their own resource graphs, backup jobs and snapshots, without needing a Docker host VM in between.</li> <li><strong>Native Proxmox integration</strong>, LXCs show up directly in the Proxmox UI with their own resource graphs, backup jobs and snapshots, without needing a Docker host VM in between.</li>
<li><strong>Lower overhead than a VM</strong> — an LXC shares the host kernel, so it starts almost instantly and uses less RAM/CPU than a full Debian VM, while still feeling like a real, persistent Linux machine.</li> <li><strong>Lower overhead than a VM</strong>, an LXC shares the host kernel, so it starts almost instantly and uses less RAM/CPU than a full Debian VM, while still feeling like a real, persistent Linux machine.</li>
<li><strong>Better fit for system-level services</strong> — some services (like a DNS resolver or a database server) benefit from being a stable, always-on OS process with predictable networking, rather than an ephemeral, frequently-redeployed container.</li> <li><strong>Better fit for system-level services</strong>, some services (like a DNS resolver or a database server) benefit from being a stable, always-on OS process with predictable networking, rather than an ephemeral, frequently-redeployed container.</li>
<li><strong>Simplicity for infrequently-changed services</strong> — these three services rarely change their configuration, so the extra flexibility of Compose-based redeployment isn't needed; a straightforward apt-installed service is easier to reason about long-term.</li> <li><strong>Simplicity for infrequently-changed services</strong>, these three services rarely change their configuration, so the extra flexibility of Compose-based redeployment isn't needed; a straightforward apt-installed service is easier to reason about long-term.</li>
</ul> </ul>
<h2>Pi-hole</h2> <h2>Pi-hole</h2>
<ul> <ul>
<li><strong>Purpose</strong> — network-wide DNS sinkhole and ad/tracker blocking for every device on the home network.</li> <li><strong>Purpose</strong>, network-wide DNS sinkhole and ad/tracker blocking for every device on the home network.</li>
<li><strong>Why LXC</strong> — DNS resolution needs to be rock-solid and always reachable at a fixed IP; running it as a lean, dedicated LXC avoids depending on the Docker host VM being up, and keeps it isolated from Docker networking quirks that could interfere with DNS.</li> <li><strong>Why LXC</strong>, DNS resolution needs to be rock-solid and always reachable at a fixed IP; running it as a lean, dedicated LXC avoids depending on the Docker host VM being up, and keeps it isolated from Docker networking quirks that could interfere with DNS.</li>
<li><strong>Advantage</strong> — near-zero overhead, boots in seconds after a Proxmox host reboot, and its stability is decoupled from whatever is happening on the Docker hosts.</li> <li><strong>Advantage</strong>, near-zero overhead, boots in seconds after a Proxmox host reboot, and its stability is decoupled from whatever is happening on the Docker hosts.</li>
</ul> </ul>
<h2>Ubuntu-Worker</h2> <h2>Ubuntu-Worker</h2>
<ul> <ul>
<li><strong>Purpose</strong> — a general-purpose Ubuntu LXC used as a flexible worker/utility machine for ad-hoc scripts, cron jobs and small maintenance tasks that don't warrant their own container image.</li> <li><strong>Purpose</strong>, a general-purpose Ubuntu LXC used as a flexible worker/utility machine for ad-hoc scripts, cron jobs and small maintenance tasks that don't warrant their own container image.</li>
<li><strong>Why LXC</strong> — it needs to behave like a normal Linux box (full package manager, cron, shell access) rather than a single-purpose containerized app, which makes an LXC a much better fit than wrapping everything in Docker.</li> <li><strong>Why LXC</strong>, it needs to behave like a normal Linux box (full package manager, cron, shell access) rather than a single-purpose containerized app, which makes an LXC a much better fit than wrapping everything in Docker.</li>
<li><strong>Advantage</strong> — quick to spin up, snapshot and roll back via Proxmox, and useful as a lightweight sandbox for testing shell scripts or automation before productionizing them elsewhere.</li> <li><strong>Advantage</strong>, quick to spin up, snapshot and roll back via Proxmox, and useful as a lightweight sandbox for testing shell scripts or automation before productionizing them elsewhere.</li>
</ul> </ul>
<h2>MariaDB</h2> <h2>MariaDB</h2>
<ul> <ul>
<li><strong>Purpose</strong> — central relational database instance, currently used by services such as nginx-proxy-manager, reachable over the network at a fixed internal IP.</li> <li><strong>Purpose</strong>, central relational database instance, currently used by services such as nginx-proxy-manager, reachable over the network at a fixed internal IP.</li>
<li><strong>Why LXC</strong> — a database benefits from direct, persistent disk access and predictable performance without the extra storage-driver indirection of Docker volumes; running it as an unprivileged LXC keeps it isolated while still feeling like a native MariaDB install.</li> <li><strong>Why LXC</strong>, a database benefits from direct, persistent disk access and predictable performance without the extra storage-driver indirection of Docker volumes; running it as an unprivileged LXC keeps it isolated while still feeling like a native MariaDB install.</li>
<li><strong>Advantage</strong> — simple backup via <code>mysqldump</code>, straightforward remote access configuration (bind-address, dedicated service users per consuming application), and one central database that multiple Docker containers on other hosts can connect to over the network instead of each running their own database container.</li> <li><strong>Advantage</strong>, simple backup via <code>mysqldump</code>, straightforward remote access configuration (bind-address, dedicated service users per consuming application), and one central database that multiple Docker containers on other hosts can connect to over the network instead of each running their own database container.</li>
</ul> </ul>
<p class="redacted-note">All three LXCs run exclusively on the main Proxmox host (PVE), not on pve2 or the Synology — they're intentionally kept centralized since they're few in number and don't need the multi-host redundancy that the Docker setup has. IPs, credentials and specific VMIDs are omitted here; only the reasoning behind the LXC-vs-Docker choice is shown.</p> <p class="redacted-note">All three LXCs run exclusively on the main Proxmox host (PVE), not on pve2 or the Synology, they're intentionally kept centralized since they're few in number and don't need the multi-host redundancy that the Docker setup has. IPs, credentials and specific VMIDs are omitted here; only the reasoning behind the LXC-vs-Docker choice is shown.</p>
</div> </div>
infra/monitoring/index.njk
+6 -6
View File
@@ -11,15 +11,15 @@ layout: base.njk
<p>Prometheus and Grafana form the central monitoring stack for the whole <p>Prometheus and Grafana form the central monitoring stack for the whole
homelab, running as Docker containers on the NAS. Every machine in the homelab, running as Docker containers on the NAS. Every machine in the
flat — Proxmox host, Docker VM, Synology NAS, and Raspberry Pi devices — flat, Proxmox host, Docker VM, Synology NAS, and Raspberry Pi devices,
feeds metrics into this central instance.</p> feeds metrics into this central instance.</p>
<h2>What's monitored</h2> <h2>What's monitored</h2>
<ul> <ul>
<li><strong>Proxmox host metrics</strong> — a PVE exporter reports node-level and per-VM/LXC CPU, memory, and disk metrics into a dedicated Grafana dashboard.</li> <li><strong>Proxmox host metrics</strong>, a PVE exporter reports node-level and per-VM/LXC CPU, memory, and disk metrics into a dedicated Grafana dashboard.</li>
<li><strong>Host metrics everywhere</strong> — a node exporter runs on the NAS, the Docker host, and every Raspberry Pi, feeding basic CPU/RAM/disk/network stats.</li> <li><strong>Host metrics everywhere</strong>, a node exporter runs on the NAS, the Docker host, and every Raspberry Pi, feeding basic CPU/RAM/disk/network stats.</li>
<li><strong>Container metrics</strong> — cAdvisor runs on both the Docker host and the NAS, breaking down resource usage per container.</li> <li><strong>Container metrics</strong>, cAdvisor runs on both the Docker host and the NAS, breaking down resource usage per container.</li>
<li><strong>Network hardware via SNMP</strong> — an SNMP exporter polls a network-attached printer for toner level, page count, and status, proving the same pattern works for any SNMP-capable device.</li> <li><strong>Network hardware via SNMP</strong>, an SNMP exporter polls a network-attached printer for toner level, page count, and status, proving the same pattern works for any SNMP-capable device.</li>
</ul> </ul>
<h2>Dashboards</h2> <h2>Dashboards</h2>
@@ -50,7 +50,7 @@ temperature thresholds) are planned but not yet implemented.</p>
<h2>Open items</h2> <h2>Open items</h2>
<ul> <ul>
<li>Alerting rules are not yet configured for most services.</li> <li>Alerting rules are not yet configured for most services.</li>
<li>The secondary (test) Proxmox host isn't monitored yet — it's usually powered off.</li> <li>The secondary (test) Proxmox host isn't monitored yet, it's usually powered off.</li>
<li>Retention policy for Prometheus data hasn't been tuned.</li> <li>Retention policy for Prometheus data hasn't been tuned.</li>
<li>Monitoring data itself isn't currently backed up.</li> <li>Monitoring data itself isn't currently backed up.</li>
</ul> </ul>
infra/openwrt/index.njk
+5 -5
View File
@@ -9,13 +9,13 @@ layout: base.njk
<h1>OpenWrt Multi-WAN Router</h1> <h1>OpenWrt Multi-WAN Router</h1>
<div class="detail-content"> <div class="detail-content">
<p>The study has historically suffered from unreliable network connectivity. To solve this, I set up an OpenWrt router that combines two WAN paths — WiFi and a dedicated Powerline link back to the server room — with automatic failover between them.</p> <p>The study has historically suffered from unreliable network connectivity. To solve this, I set up an OpenWrt router that combines two WAN paths, WiFi and a dedicated Powerline link back to the server room, with automatic failover between them.</p>
<h2>Multi-WAN setup</h2> <h2>Multi-WAN setup</h2>
<ul> <ul>
<li><strong>WAN 1 — WiFi (802.11ac)</strong> — higher throughput but less stable, used for bulk transfers and general traffic.</li> <li><strong>WAN 1, WiFi (802.11ac)</strong>, higher throughput but less stable, used for bulk transfers and general traffic.</li>
<li><strong>WAN 2 — Powerline (Ethernet backhaul)</strong> — lower throughput but rock-solid, carries latency-sensitive traffic like VPN tunnels and Teams calls.</li> <li><strong>WAN 2, Powerline (Ethernet backhaul)</strong>, lower throughput but rock-solid, carries latency-sensitive traffic like VPN tunnels and Teams calls.</li>
<li><strong>Failover</strong> — both paths are active simultaneously; if one drops, traffic automatically shifts to the remaining path. This provides redundancy in both directions.</li> <li><strong>Failover</strong>, both paths are active simultaneously; if one drops, traffic automatically shifts to the remaining path. This provides redundancy in both directions.</li>
</ul> </ul>
<p>Traffic is routed intelligently based on connection quality: VPN and real-time communication prefer the stable Powerline path, while everything else benefits from the higher WiFi throughput when available.</p> <p>Traffic is routed intelligently based on connection quality: VPN and real-time communication prefer the stable Powerline path, while everything else benefits from the higher WiFi throughput when available.</p>
@@ -28,7 +28,7 @@ layout: base.njk
</ul> </ul>
<h2>IoT VLAN 189</h2> <h2>IoT VLAN 189</h2>
<p>A dedicated, isolated network segment for IoT devices, running on a separate small WiFi access point on a different channel. This keeps smart home hardware — including the balcony sensors and light controller — off the main network while maintaining full connectivity to the homelab services.</p> <p>A dedicated, isolated network segment for IoT devices, running on a separate small WiFi access point on a different channel. This keeps smart home hardware, including the balcony sensors and light controller, off the main network while maintaining full connectivity to the homelab services.</p>
<ul> <ul>
<li><strong>VLAN ID:</strong> 189</li> <li><strong>VLAN ID:</strong> 189</li>
<li><strong>Purpose:</strong> Isolate IoT/smart home devices from general network traffic.</li> <li><strong>Purpose:</strong> Isolate IoT/smart home devices from general network traffic.</li>
infra/proxmox/index.njk
+16 -16
View File
@@ -9,35 +9,35 @@ layout: base.njk
<h1>Proxmox Virtualization</h1> <h1>Proxmox Virtualization</h1>
<div class="detail-content"> <div class="detail-content">
<p>Proxmox VE (Virtual Environment) is a free, open-source virtualization platform that combines KVM-based full virtual machines with lightweight LXC containers in a single web-managed system. It's built on top of Debian, includes ZFS storage management, backup/restore, clustering and monitoring out of the box, and is the foundation my entire homelab runs on — both the <a href="/docker/">Docker</a> hosts and the standalone <a href="/lxc/">LXC services</a> live as guests on top of it.</p> <p>Proxmox VE (Virtual Environment) is a free, open-source virtualization platform that combines KVM-based full virtual machines with lightweight LXC containers in a single web-managed system. It's built on top of Debian, includes ZFS storage management, backup/restore, clustering and monitoring out of the box, and is the foundation my entire homelab runs on, both the <a href="/docker/">Docker</a> hosts and the standalone <a href="/lxc/">LXC services</a> live as guests on top of it.</p>
<h2>Hardware</h2> <h2>Hardware</h2>
<ul> <ul>
<li><strong>PVE (main server)</strong> — a Beelink Mini-PC, running as the primary Proxmox host. It has a 500GB SSD for VM and container storage, and connects to the network via the server-room Zyxel switch.</li> <li><strong>PVE (main server)</strong>, a Beelink Mini-PC, running as the primary Proxmox host. It has a 500GB SSD for VM and container storage, and connects to the network via the server-room Zyxel switch.</li>
<li><strong>pve2 (test server)</strong> — a second Proxmox host, physically located in the study and connected via the OpenWRT-side Zyxel switch. It's dedicated purely to testing and development, running the third <a href="/docker/">Docker host VM (docker-host-pve2)</a> plus experimental setups without risking the production environment.</li> <li><strong>pve2 (test server)</strong>, a second Proxmox host, physically located in the study and connected via the OpenWRT-side Zyxel switch. It's dedicated purely to testing and development, running the third <a href="/docker/">Docker host VM (docker-host-pve2)</a> plus experimental setups without risking the production environment.</li>
<li><strong>Backups</strong> — nightly VM backups from PVE run at 00:30 to an NFS share on the Synology (<code>nas:/volume1/backup</code>), with retention of the last 3 daily, 3 monthly and 1 yearly backup.</li> <li><strong>Backups</strong>, nightly VM backups from PVE run at 00:30 to an NFS share on the Synology (<code>nas:/volume1/backup</code>), with retention of the last 3 daily, 3 monthly and 1 yearly backup.</li>
</ul> </ul>
<h2>Advantages</h2> <h2>Advantages</h2>
<ul> <ul>
<li><strong>Mixed virtualization</strong> — full KVM virtual machines and lightweight LXC containers run side by side on the same host, so I can pick whichever fits a service best.</li> <li><strong>Mixed virtualization</strong>, full KVM virtual machines and lightweight LXC containers run side by side on the same host, so I can pick whichever fits a service best.</li>
<li><strong>Web-based management</strong> — the entire cluster, its VMs, containers, storage and backups are managed from one browser UI via Authelia/Nginx, no separate tooling needed.</li> <li><strong>Web-based management</strong>, the entire cluster, its VMs, containers, storage and backups are managed from one browser UI via Authelia/Nginx, no separate tooling needed.</li>
<li><strong>Built-in ZFS support</strong> — snapshots, checksumming and pool management are native, making disk setup and VM protection straightforward.</li> <li><strong>Built-in ZFS support</strong>, snapshots, checksumming and pool management are native, making disk setup and VM protection straightforward.</li>
<li><strong>Native backup/restore</strong> — scheduled backups to an external NFS target are configured directly in Proxmox, without third-party backup software.</li> <li><strong>Native backup/restore</strong>, scheduled backups to an external NFS target are configured directly in Proxmox, without third-party backup software.</li>
<li><strong>No licensing cost</strong> — the platform is fully usable without a subscription by switching to the no-subscription repository, which matters for a homelab that isn't running on enterprise support contracts.</li> <li><strong>No licensing cost</strong>, the platform is fully usable without a subscription by switching to the no-subscription repository, which matters for a homelab that isn't running on enterprise support contracts.</li>
</ul> </ul>
<h2>Open-source edition</h2> <h2>Open-source edition</h2>
<ul> <ul>
<li><strong>Proxmox VE Community/no-subscription</strong> — the enterprise repositories are disabled and the host is switched to the free "no-subscription" package repository, giving full functionality identical to the paid tiers minus vendor support and the enterprise-only repo.</li> <li><strong>Proxmox VE Community/no-subscription</strong>, the enterprise repositories are disabled and the host is switched to the free "no-subscription" package repository, giving full functionality identical to the paid tiers minus vendor support and the enterprise-only repo.</li>
<li><strong>License</strong> — Proxmox VE itself is open source (AGPLv3); the subscription only adds official support and access to the more conservative enterprise update channel.</li> <li><strong>License</strong>, Proxmox VE itself is open source (AGPLv3); the subscription only adds official support and access to the more conservative enterprise update channel.</li>
</ul> </ul>
<h2>Setup process</h2> <h2>Setup process</h2>
<ul> <ul>
<li><strong>Bootstrap</strong> — on new Beelink hardware, an initial OS (Windows) needs to be cleared first; entering the BIOS (DEL key) to set the boot order to USB-first allows booting the Proxmox installer.</li> <li><strong>Bootstrap</strong>, on new Beelink hardware, an initial OS (Windows) needs to be cleared first; entering the BIOS (DEL key) to set the boot order to USB-first allows booting the Proxmox installer.</li>
<li><strong>Base install</strong> — Proxmox VE installed directly from ISO, then updated after disabling the enterprise repo and enabling the no-subscription repo.</li> <li><strong>Base install</strong>, Proxmox VE installed directly from ISO, then updated after disabling the enterprise repo and enabling the no-subscription repo.</li>
<li><strong>Storage</strong> — an additional SSD is partitioned (GPT via <code>parted</code>) and turned into a ZFS pool through the Disks → ZFS section of the web UI.</li> <li><strong>Storage</strong>, an additional SSD is partitioned (GPT via <code>parted</code>) and turned into a ZFS pool through the Disks → ZFS section of the web UI.</li>
<li><strong>Networking/backup target</strong> — an NFS share from the Synology is mounted for backup storage, and a nightly backup job is scheduled per VM.</li> <li><strong>Networking/backup target</strong>, an NFS share from the Synology is mounted for backup storage, and a nightly backup job is scheduled per VM.</li>
<li><strong>Guest creation</strong> — VMs (like the Debian Docker-host VMs) and LXCs (Pi-hole, Ubuntu-Worker, MariaDB) are created through the web UI, with ISO images uploaded to local storage beforehand.</li> <li><strong>Guest creation</strong>, VMs (like the Debian Docker-host VMs) and LXCs (Pi-hole, Ubuntu-Worker, MariaDB) are created through the web UI, with ISO images uploaded to local storage beforehand.</li>
</ul> </ul>
infra/synology/index.njk
+10 -10
View File
@@ -13,24 +13,24 @@ layout: base.njk
<h2>Storage</h2> <h2>Storage</h2>
<ul> <ul>
<li><strong>Bays</strong> — 4 drive slots, populated with a mix of large-capacity HDDs rather than a single uniform set, which SHR was specifically chosen to support.</li> <li><strong>Bays</strong>, 4 drive slots, populated with a mix of large-capacity HDDs rather than a single uniform set, which SHR was specifically chosen to support.</li>
<li><strong>RAID type</strong> — SHR (Synology Hybrid RAID), giving single-drive fault tolerance while allowing disks of different sizes to be mixed and later upgraded one at a time without rebuilding the whole array from scratch.</li> <li><strong>RAID type</strong>, SHR (Synology Hybrid RAID), giving single-drive fault tolerance while allowing disks of different sizes to be mixed and later upgraded one at a time without rebuilding the whole array from scratch.</li>
<li><strong>Snapshots</strong> — Btrfs snapshot replication is used for file versioning instead of the classic recycle bin, taken hourly with a retention schedule of 7 days, 3 weeks, 3 months and 1 year.</li> <li><strong>Snapshots</strong>, Btrfs snapshot replication is used for file versioning instead of the classic recycle bin, taken hourly with a retention schedule of 7 days, 3 weeks, 3 months and 1 year.</li>
</ul> </ul>
<h2>What runs on it today</h2> <h2>What runs on it today</h2>
<ul> <ul>
<li><strong>Docker host VM</strong> — a dedicated Debian VM (docker-host-nas) runs on the NAS's virtualization layer, hosting containers that specifically need direct access to NAS storage volumes.</li> <li><strong>Docker host VM</strong>, a dedicated Debian VM (docker-host-nas) runs on the NAS's virtualization layer, hosting containers that specifically need direct access to NAS storage volumes.</li>
<li><strong>File storage & backup target</strong> — the core role of the NAS remains classic network storage, plus acting as the backup destination for scheduled Proxmox VM backups from the rest of the homelab.</li> <li><strong>File storage & backup target</strong>, the core role of the NAS remains classic network storage, plus acting as the backup destination for scheduled Proxmox VM backups from the rest of the homelab.</li>
<li><strong>Legacy packages removed</strong> — unused DSM packages (like a web/photo/media package, an old drive-sync service and unused indexing) were uninstalled to reduce background load and disk writes.</li> <li><strong>Legacy packages removed</strong>, unused DSM packages (like a web/photo/media package, an old drive-sync service and unused indexing) were uninstalled to reduce background load and disk writes.</li>
</ul> </ul>
<h2>Main strengths</h2> <h2>Main strengths</h2>
<ul> <ul>
<li><strong>Central storage hub</strong> — single source of truth for files and backups that every other host in the network (Proxmox VMs, exporters, workstations) can write to or read from.</li> <li><strong>Central storage hub</strong>, single source of truth for files and backups that every other host in the network (Proxmox VMs, exporters, workstations) can write to or read from.</li>
<li><strong>Built-in snapshotting</strong> — Btrfs-based snapshots give point-in-time recovery without the overhead of a separate versioning package.</li> <li><strong>Built-in snapshotting</strong>, Btrfs-based snapshots give point-in-time recovery without the overhead of a separate versioning package.</li>
<li><strong>Flexible RAID</strong> — SHR allows disks to be swapped and capacity to grow over time without needing matched drive sizes from day one.</li> <li><strong>Flexible RAID</strong>, SHR allows disks to be swapped and capacity to grow over time without needing matched drive sizes from day one.</li>
<li><strong>Lightweight virtualization</strong> — running a single purpose-built Docker VM on the NAS keeps storage-adjacent workloads close to the data, without turning the NAS itself into a general-purpose application server.</li> <li><strong>Lightweight virtualization</strong>, running a single purpose-built Docker VM on the NAS keeps storage-adjacent workloads close to the data, without turning the NAS itself into a general-purpose application server.</li>
</ul> </ul>
</div> </div>
retro/amiga-hardware-landscape/index.njk
+8 -8
View File
@@ -9,14 +9,14 @@ parent: "/retro/"
<h1>Running AmigaOS in 2026: MiSTer FPGA vs Amiberry on Pi</h1> <h1>Running AmigaOS in 2026: MiSTer FPGA vs Amiberry on Pi</h1>
<div class="detail-content"> <div class="detail-content">
<p>AmigaOS 3.2.3 runs beautifully both on FPGA and on modern ARM boards — the question is which trade-offs you're willing to accept. Both approaches have their place, and the choice often comes down to what you're trying to do: run demos at perfect timing, or tinker with software at higher resolutions.</p> <p>AmigaOS 3.2.3 runs beautifully both on FPGA and on modern ARM boards, the question is which trade-offs you're willing to accept. Both approaches have their place, and the choice often comes down to what you're trying to do: run demos at perfect timing, or tinker with software at higher resolutions.</p>
<h2>Two Roads to the Same Workbench</h2> <h2>Two Roads to the Same Workbench</h2>
<ul> <ul>
<li><strong>MiSTer with the Minimig-AGA core</strong> — the most hardware-authentic option. Native resolution tops out around 800×600, with RTG (Picasso96) support pushing that up to 1080p. Timing accuracy and near-zero input latency make it the pick for demos and cycle-sensitive software.</li> <li><strong>MiSTer with the Minimig-AGA core</strong>, the most hardware-authentic option. Native resolution tops out around 800×600, with RTG (Picasso96) support pushing that up to 1080p. Timing accuracy and near-zero input latency make it the pick for demos and cycle-sensitive software.</li>
<li><strong>Amiberry on a Raspberry Pi</strong> — the flexible option. Running headless (no X11 desktop, straight from the console/KMS framebuffer) is the single biggest performance lever — the difference between a laggy desktop and a snappy 60fps Workbench.</li> <li><strong>Amiberry on a Raspberry Pi</strong>, the flexible option. Running headless (no X11 desktop, straight from the console/KMS framebuffer) is the single biggest performance lever, the difference between a laggy desktop and a snappy 60fps Workbench.</li>
<li><strong>The HDF minefield</strong> — hard disk image files either carry an RDB partition table or they don't, and getting the geometry parameters (surfaces/sectors/reserved blocks) wrong produces the classic "Not a DOS disk in device DH0" error. Reformatting through HDToolBox from a fresh install is often faster than debugging geometry math.</li> <li><strong>The HDF minefield</strong>, hard disk image files either carry an RDB partition table or they don't, and getting the geometry parameters (surfaces/sectors/reserved blocks) wrong produces the classic "Not a DOS disk in device DH0" error. Reformatting through HDToolBox from a fresh install is often faster than debugging geometry math.</li>
<li><strong>Picking a monitor</strong> — a 4:3 panel at native 1280×1024 (like an older NEC MultiSync) gives the most period-accurate Workbench experience; anything widescreen needs scaling compromises.</li> <li><strong>Picking a monitor</strong>, a 4:3 panel at native 1280×1024 (like an older NEC MultiSync) gives the most period-accurate Workbench experience; anything widescreen needs scaling compromises.</li>
</ul> </ul>
<h2>When to choose MiSTer</h2> <h2>When to choose MiSTer</h2>
@@ -40,8 +40,8 @@ a Pi to Amiga development without tying up expensive FPGA hardware.</p>
Some demos use raster tricks that only work on real AGA hardware or the Some demos use raster tricks that only work on real AGA hardware or the
MiSTer's FPGA core. Some productivity software depends on specific MiSTer's FPGA core. Some productivity software depends on specific
accelerator card features that neither platform emulates perfectly. accelerator card features that neither platform emulates perfectly.
And the perennial question of Kickstart ROM versions — 3.1 for And the perennial question of Kickstart ROM versions, 3.1 for
compatibility, 3.2 for features — applies equally to both setups.</p> compatibility, 3.2 for features, applies equally to both setups.</p>
<p class="redacted-note">Both setups now run side by side without any real winner — MiSTer for authenticity and demos, the Pi for tinkering and RTG resolutions.</p> <p class="redacted-note">Both setups now run side by side without any real winner, MiSTer for authenticity and demos, the Pi for tinkering and RTG resolutions.</p>
</div> </div>
retro/amiga-modern-platforms/index.njk
+6 -6
View File
@@ -9,15 +9,15 @@ parent: "/retro/"
<h1>Vampire V4, A600GS or MiSTer: Amiga Hardware Isn't a Simple Speed Race</h1> <h1>Vampire V4, A600GS or MiSTer: Amiga Hardware Isn't a Simple Speed Race</h1>
<div class="detail-content"> <div class="detail-content">
<p>Trying to compare Amiga hardware options by benchmark numbers alone is a trap. A Dhrystone score from a real FPGA CPU implementation and an ARM chip's MIPS rating measure completely different things — and neither tells you much about which system will actually run your favourite demo correctly.</p> <p>Trying to compare Amiga hardware options by benchmark numbers alone is a trap. A Dhrystone score from a real FPGA CPU implementation and an ARM chip's MIPS rating measure completely different things, and neither tells you much about which system will actually run your favourite demo correctly.</p>
<h2>The Contenders</h2> <h2>The Contenders</h2>
<ul> <ul>
<li><strong>Vampire V4 Standalone</strong> — an FPGA-implemented 68080 CPU that's genuinely blisteringly fast (sub-second Workbench boots), but software compatibility suffers since the 68080 isn't a perfect 68060 clone. Best for games and demos, not for daily productivity.</li> <li><strong>Vampire V4 Standalone</strong>, an FPGA-implemented 68080 CPU that's genuinely blisteringly fast (sub-second Workbench boots), but software compatibility suffers since the 68080 isn't a perfect 68060 clone. Best for games and demos, not for daily productivity.</li>
<li><strong>A600GS / A1200NG</strong> — ARM-based boards running Amiberry underneath, in either a complete mini-PC (A600GS) or a drop-in motherboard replacement with real floppy and CompactFlash support (A1200NG). Cheap, plug-and-play, but with occasional software compatibility quirks.</li> <li><strong>A600GS / A1200NG</strong>, ARM-based boards running Amiberry underneath, in either a complete mini-PC (A600GS) or a drop-in motherboard replacement with real floppy and CompactFlash support (A1200NG). Cheap, plug-and-play, but with occasional software compatibility quirks.</li>
<li><strong>MiSTer with the Minimig core</strong> — "loses" on paper to A600GS in raw MIPS numbers, but wins on emulation accuracy and software compatibility. Its D-Cache-on mode is faster still but introduces glitches, so it's best left off for stability.</li> <li><strong>MiSTer with the Minimig core</strong>, "loses" on paper to A600GS in raw MIPS numbers, but wins on emulation accuracy and software compatibility. Its D-Cache-on mode is faster still but introduces glitches, so it's best left off for stability.</li>
<li><strong>The real decision factor</strong> — not speed, but compatibility, stability, and how much you value hardware-accurate timing versus raw throughput.</li> <li><strong>The real decision factor</strong>, not speed, but compatibility, stability, and how much you value hardware-accurate timing versus raw throughput.</li>
</ul> </ul>
<p class="redacted-note">Conclusion after digging through all three: if you already own a working MiSTer, there's no compelling reason to chase the newer ARM boards for "better" performance — the practical difference is marginal, and MiSTer's broader core library wins in the long run.</p> <p class="redacted-note">Conclusion after digging through all three: if you already own a working MiSTer, there's no compelling reason to chase the newer ARM boards for "better" performance, the practical difference is marginal, and MiSTer's broader core library wins in the long run.</p>
</div> </div>
retro/atari-mega-st-fleet/index.njk
+6 -6
View File
@@ -9,15 +9,15 @@ parent: "/retro/"
<h1>The Atari Mega ST Fleet: Three Machines, Three Jobs</h1> <h1>The Atari Mega ST Fleet: Three Machines, Three Jobs</h1>
<div class="detail-content"> <div class="detail-content">
<p>Somewhere along the way, "one Atari ST" turned into three Mega ST 4 machines plus a Mega ST 2. Rather than let them gather dust, each one got a clearly defined purpose — from untouched nostalgia piece to fully upgraded workstation.</p> <p>Somewhere along the way, "one Atari ST" turned into three Mega ST 4 machines plus a Mega ST 2. Rather than let them gather dust, each one got a clearly defined purpose, from untouched nostalgia piece to fully upgraded workstation.</p>
<h2>The Lineup</h2> <h2>The Lineup</h2>
<ul> <ul>
<li><strong>The heirloom machine</strong> — kept as close to factory-original as possible, running the stock TOS 1.04 and a classic floppy drive. Its job is to be a baseline reference: if something breaks on a modded machine, this one proves the fault isn't in the software.</li> <li><strong>The heirloom machine</strong>, kept as close to factory-original as possible, running the stock TOS 1.04 and a classic floppy drive. Its job is to be a baseline reference: if something breaks on a modded machine, this one proves the fault isn't in the software.</li>
<li><strong>The "Low" machine</strong> — a bargain-bin Mega ST 4 rebuilt with a Cloudy/StormyCloud combo for TOS switching and extra FastRAM, a Gotek floppy emulator, an UltraSatan hard disk emulator, and a SidecarTridge multi-device for ROM, floppy, and Wi-Fi tricks. This is the demo- and game-machine, optimized for fast disk-image swapping rather than serious productivity.</li> <li><strong>The "Low" machine</strong>, a bargain-bin Mega ST 4 rebuilt with a Cloudy/StormyCloud combo for TOS switching and extra FastRAM, a Gotek floppy emulator, an UltraSatan hard disk emulator, and a SidecarTridge multi-device for ROM, floppy, and Wi-Fi tricks. This is the demo- and game-machine, optimized for fast disk-image swapping rather than serious productivity.</li>
<li><strong>The "High" machine</strong> — the workstation build. A MonSTerBoard adds 8MB of AltRAM, IDE storage, and flashable TOS (currently TOS 2.06), running Geneva for real multitasking. A NetUSB adapter and original Atari keyboard/mouse round it out, with an ET4000 graphics card in the pipeline for higher resolutions.</li> <li><strong>The "High" machine</strong>, the workstation build. A MonSTerBoard adds 8MB of AltRAM, IDE storage, and flashable TOS (currently TOS 2.06), running Geneva for real multitasking. A NetUSB adapter and original Atari keyboard/mouse round it out, with an ET4000 graphics card in the pipeline for higher resolutions.</li>
<li><strong>The spare Mega ST 2</strong> — a reserve and donor machine, kept in largely original condition, used to safely test upgrades before they go into the "real" machines.</li> <li><strong>The spare Mega ST 2</strong>, a reserve and donor machine, kept in largely original condition, used to safely test upgrades before they go into the "real" machines.</li>
</ul> </ul>
<p class="redacted-note">Fun fact: two of the "problem" machines turned out to have nothing wrong with the boards at all — a badly soldered CPU socket and a leftover blitter patch bridge were the actual culprits. Moral of the story: check your solder joints twice before blaming the silicon.</p> <p class="redacted-note">Fun fact: two of the "problem" machines turned out to have nothing wrong with the boards at all, a badly soldered CPU socket and a leftover blitter patch bridge were the actual culprits. Moral of the story: check your solder joints twice before blaming the silicon.</p>
</div> </div>
retro/atari-monitor-compatibility/index.njk
+6 -6
View File
@@ -9,15 +9,15 @@ parent: "/retro/"
<h1>The 15kHz Problem: Finding a Monitor That Speaks Atari</h1> <h1>The 15kHz Problem: Finding a Monitor That Speaks Atari</h1>
<div class="detail-content"> <div class="detail-content">
<p>Atari ST colour modes (Low-Res and Medium-Res) run at a horizontal scan frequency of 15kHz — a frequency almost no modern VGA monitor supports, since anything made in the last two decades expects at least 30kHz. High-Res monochrome runs at roughly 35kHz and gets along fine with most displays, but colour gaming is a different story entirely.</p> <p>Atari ST colour modes (Low-Res and Medium-Res) run at a horizontal scan frequency of 15kHz, a frequency almost no modern VGA monitor supports, since anything made in the last two decades expects at least 30kHz. High-Res monochrome runs at roughly 35kHz and gets along fine with most displays, but colour gaming is a different story entirely.</p>
<h2>Panel Technology Matters More Than You'd Think</h2> <h2>Panel Technology Matters More Than You'd Think</h2>
<ul> <ul>
<li><strong>MVA/PVA panels</strong> — the surprise winners. Their flexible crystal alignment tolerates unusual 15kHz signals far better than standard panels. A NEC MultiSync LCD1970NXp with an MVA panel handles Medium-Res colour reliably, once you fine-tune the H.Size setting.</li> <li><strong>MVA/PVA panels</strong>, the surprise winners. Their flexible crystal alignment tolerates unusual 15kHz signals far better than standard panels. A NEC MultiSync LCD1970NXp with an MVA panel handles Medium-Res colour reliably, once you fine-tune the H.Size setting.</li>
<li><strong>IPS panels</strong> — a mixed bag. Some work (a widescreen Lenovo ThinkVision LT1952p reportedly fills the screen beautifully via an ST2VGA adapter), others don't, and there's no way to tell without testing.</li> <li><strong>IPS panels</strong>, a mixed bag. Some work (a widescreen Lenovo ThinkVision LT1952p reportedly fills the screen beautifully via an ST2VGA adapter), others don't, and there's no way to tell without testing.</li>
<li><strong>VA panels</strong> — a dead end. Modern VA displays like the Dell SE2722HX and SE2422HX have a minimum horizontal frequency of 30kHz, ruling out native 15kHz signals entirely.</li> <li><strong>VA panels</strong>, a dead end. Modern VA displays like the Dell SE2722HX and SE2422HX have a minimum horizontal frequency of 30kHz, ruling out native 15kHz signals entirely.</li>
<li><strong>The universal fix</strong> — an Open Source Scan Converter (OSSC) upsamples the 15kHz signal for any HDMI monitor, at the cost of extra hardware and money.</li> <li><strong>The universal fix</strong>, an Open Source Scan Converter (OSSC) upsamples the 15kHz signal for any HDMI monitor, at the cost of extra hardware and money.</li>
</ul> </ul>
<p class="redacted-note">If you're chasing 15kHz authenticity, buy the specific panel model people have already confirmed working in the forums — don't trust a spec sheet that just says "supports VGA."</p> <p class="redacted-note">If you're chasing 15kHz authenticity, buy the specific panel model people have already confirmed working in the forums, don't trust a spec sheet that just says "supports VGA."</p>
</div> </div>
retro/atari-networking-bbs/index.njk
+9 -9
View File
@@ -9,18 +9,18 @@ parent: "/retro/"
<h1>Getting a 40-Year-Old Computer Online: Atari ST Networking and BBS Access</h1> <h1>Getting a 40-Year-Old Computer Online: Atari ST Networking and BBS Access</h1>
<div class="detail-content"> <div class="detail-content">
<p>The Atari ST predates the consumer internet entirely, yet a surprising amount of retrofitted networking hardware exists to get one talking to a modern network — or to a bulletin board system straight out of the dial-up era.</p> <p>The Atari ST predates the consumer internet entirely, yet a surprising amount of retrofitted networking hardware exists to get one talking to a modern network, or to a bulletin board system straight out of the dial-up era.</p>
<h2>The Toolbox</h2> <h2>The Toolbox</h2>
<ul> <ul>
<li><strong>STinG</strong> — the classic TCP/IP stack for TOS, still the backbone for anything network-related on an original machine, from Telnet sessions to FTP transfers.</li> <li><strong>STinG</strong>, the classic TCP/IP stack for TOS, still the backbone for anything network-related on an original machine, from Telnet sessions to FTP transfers.</li>
<li><strong>ESP32-based Wi-Fi modules</strong> — small serial-attached boards that emulate a modem over Wi-Fi. Results vary wildly between firmware projects: one variant handles HTTP downloads fine but can't do Telnet, another supports full modem emulation but drops connections unpredictably during longer BBS sessions.</li> <li><strong>ESP32-based Wi-Fi modules</strong>, small serial-attached boards that emulate a modem over Wi-Fi. Results vary wildly between firmware projects: one variant handles HTTP downloads fine but can't do Telnet, another supports full modem emulation but drops connections unpredictably during longer BBS sessions.</li>
<li><strong>Terminal software quirks</strong> — ANSI art and special characters routinely render incorrectly in period terminal clients, an ongoing annoyance when connecting to text-heavy BBS systems.</li> <li><strong>Terminal software quirks</strong>, ANSI art and special characters routinely render incorrectly in period terminal clients, an ongoing annoyance when connecting to text-heavy BBS systems.</li>
<li><strong>File transfer the easy way</strong> — a plain FTP client on the Atari talking to a NAS on the local network sidesteps most of the internet-connectivity headaches entirely, especially for moving disk images and software archives back and forth.</li> <li><strong>File transfer the easy way</strong>, a plain FTP client on the Atari talking to a NAS on the local network sidesteps most of the internet-connectivity headaches entirely, especially for moving disk images and software archives back and forth.</li>
</ul> </ul>
<h2>Why connect a 40-year-old computer to the internet</h2> <h2>Why connect a 40-year-old computer to the internet</h2>
<p>The obvious question is "why bother?" — the Atari ST was never designed <p>The obvious question is "why bother?", the Atari ST was never designed
for the internet, and the experience of using it online is objectively for the internet, and the experience of using it online is objectively
worse than any modern device. But that's not the point. The real worse than any modern device. But that's not the point. The real
satisfaction comes from making old hardware do things it was never satisfaction comes from making old hardware do things it was never
@@ -31,9 +31,9 @@ original hardware.</p>
<h2>The BBS experience</h2> <h2>The BBS experience</h2>
<p>Bulletin board systems from the dial-up era are still running today, <p>Bulletin board systems from the dial-up era are still running today,
maintained by enthusiasts who keep the old software alive. Connecting to maintained by enthusiasts who keep the old software alive. Connecting to
one from an original Atari ST — via a Wi-Fi modem emulation — is a one from an original Atari ST, via a Wi-Fi modem emulation, is a
genuinely nostalgic experience. The ANSI art login screens, the message genuinely nostalgic experience. The ANSI art login screens, the message
boards, the file sections, the door games — it's all exactly as it was boards, the file sections, the door games, it's all exactly as it was
in the early 1990s. The main difference: instead of tying up a phone in the early 1990s. The main difference: instead of tying up a phone
line, the connection goes over Wi-Fi to a BBS that's accessible from line, the connection goes over Wi-Fi to a BBS that's accessible from
anywhere in the world.</p> anywhere in the world.</p>
@@ -41,7 +41,7 @@ anywhere in the world.</p>
<h2>Lessons learned</h2> <h2>Lessons learned</h2>
<p>The biggest lesson: networking on retro hardware is 90% patience and <p>The biggest lesson: networking on retro hardware is 90% patience and
10% actual configuration. The STinG stack is well-documented but 10% actual configuration. The STinG stack is well-documented but
fragile — wrong TCP/IP settings silently break connections without fragile, wrong TCP/IP settings silently break connections without
useful error messages. The ESP32 Wi-Fi modules are cheap and flexible useful error messages. The ESP32 Wi-Fi modules are cheap and flexible
but require careful firmware selection for the specific use case. And but require careful firmware selection for the specific use case. And
terminal emulation is a rabbit hole of character sets, escape sequences, terminal emulation is a rabbit hole of character sets, escape sequences,
retro/batocera-living-room/index.njk
+5 -5
View File
@@ -9,14 +9,14 @@ parent: "/retro/"
<h1>Batocera in the Living Room: One Box, Every Retro System</h1> <h1>Batocera in the Living Room: One Box, Every Retro System</h1>
<div class="detail-content"> <div class="detail-content">
<p>Not every retro session calls for hardware-accurate FPGA timing. Sometimes you just want to grab a controller on the couch and pick something from a big menu — that's exactly what the Batocera station in the living room is for.</p> <p>Not every retro session calls for hardware-accurate FPGA timing. Sometimes you just want to grab a controller on the couch and pick something from a big menu, that's exactly what the Batocera station in the living room is for.</p>
<h2>What's Under the Hood</h2> <h2>What's Under the Hood</h2>
<ul> <ul>
<li><strong>Software emulation, not FPGA</strong> — Hatari for Atari ST, UAE for Amiga, DOSBox for DOS, MAME/FBA for arcade, plus the usual console emulators (NES, SNES, Mega Drive, PlayStation). It's the go-to place for Atari ST gaming in this setup, since MiSTer's own ST core is unreliable.</li> <li><strong>Software emulation, not FPGA</strong>, Hatari for Atari ST, UAE for Amiga, DOSBox for DOS, MAME/FBA for arcade, plus the usual console emulators (NES, SNES, Mega Drive, PlayStation). It's the go-to place for Atari ST gaming in this setup, since MiSTer's own ST core is unreliable.</li>
<li><strong>ROM organisation</strong> — following the TOSEC standard and cleaning up collections with RomVault DAT files, aiming for "1G1R" (One Game One ROM) sets to keep launcher menus from turning into an unmanageable wall of duplicates.</li> <li><strong>ROM organisation</strong>, following the TOSEC standard and cleaning up collections with RomVault DAT files, aiming for "1G1R" (One Game One ROM) sets to keep launcher menus from turning into an unmanageable wall of duplicates.</li>
<li><strong>Curated sources</strong> — a well-known "scene DVD" collection covers Atari ST releases comprehensively; Amiga has no equally tidy equivalent, since TOSEC's Amiga archive is comprehensive but messy, mixing dozens of crack-group releases per game.</li> <li><strong>Curated sources</strong>, a well-known "scene DVD" collection covers Atari ST releases comprehensively; Amiga has no equally tidy equivalent, since TOSEC's Amiga archive is comprehensive but messy, mixing dozens of crack-group releases per game.</li>
<li><strong>Division of labour with MiSTer/MiST</strong> — Batocera wins on convenience and breadth, MiSTer/MiST win on accuracy and low-latency feel for serious sessions.</li> <li><strong>Division of labour with MiSTer/MiST</strong>, Batocera wins on convenience and breadth, MiSTer/MiST win on accuracy and low-latency feel for serious sessions.</li>
</ul> </ul>
<p class="redacted-note">If you're building a similar box: don't skip the DAT-file cleanup step. An unsorted TOSEC dump makes browsing physically painful once you cross a few thousand entries.</p> <p class="redacted-note">If you're building a similar box: don't skip the DAT-file cleanup step. An unsorted TOSEC dump makes browsing physically painful once you cross a few thousand entries.</p>
retro/dos-486-tower/index.njk
+10 -10
View File
@@ -9,30 +9,30 @@ parent: "/retro/"
<h1>The 486 DX2-66: Rebuilding a DOS Era on MiSTer FPGA</h1> <h1>The 486 DX2-66: Rebuilding a DOS Era on MiSTer FPGA</h1>
<div class="detail-content"> <div class="detail-content">
<p>Long before Atari ST and Amiga took over the retro corner, a Vobis-Highscreen 486 DX2-66 in its distinctive Colani-designed tower case was the main PC. The original hardware is long gone, but the entire setup has been recreated 1:1 on MiSTer FPGA — same DOS version, same drivers, same AUTOEXEC.BAT and CONFIG.SYS, same directory structure.</p> <p>Long before Atari ST and Amiga took over the retro corner, a Vobis-Highscreen 486 DX2-66 in its distinctive Colani-designed tower case was the main PC. The original hardware is long gone, but the entire setup has been recreated 1:1 on MiSTer FPGA, same DOS version, same drivers, same AUTOEXEC.BAT and CONFIG.SYS, same directory structure.</p>
<h2>The original hardware</h2> <h2>The original hardware</h2>
<p>The Vobis-Highscreen was a premium PC in its day — the Colani design gave it a distinctive curved case that still looks futuristic. Inside it ran an Intel 486 DX2 at 66MHz with 8MB of RAM, a Sound Blaster 16, and a Tseng ET4000 graphics card. A 200MB hard drive and a 3.5" floppy drive completed the setup. The machine was the go-to system for early-90s classics like Doom, Wing Commander, Duke Nukem 3D, and Lemmings.</p> <p>The Vobis-Highscreen was a premium PC in its day, the Colani design gave it a distinctive curved case that still looks futuristic. Inside it ran an Intel 486 DX2 at 66MHz with 8MB of RAM, a Sound Blaster 16, and a Tseng ET4000 graphics card. A 200MB hard drive and a 3.5" floppy drive completed the setup. The machine was the go-to system for early-90s classics like Doom, Wing Commander, Duke Nukem 3D, and Lemmings.</p>
<h2>The MiSTer recreation</h2> <h2>The MiSTer recreation</h2>
<p>The AO486 core on MiSTer FPGA reproduces the original hardware with near-perfect accuracy. The key was not just getting the hardware emulation right, but rebuilding the exact software environment: the same MS-DOS version, the same memory managers (HIMEM.SYS, EMM386), the same Sound Blaster drivers, and the same AUTOEXEC.BAT and CONFIG.SYS tuning that made the original machine run smoothly. Even the directory structure and file names match the original setup.</p> <p>The AO486 core on MiSTer FPGA reproduces the original hardware with near-perfect accuracy. The key was not just getting the hardware emulation right, but rebuilding the exact software environment: the same MS-DOS version, the same memory managers (HIMEM.SYS, EMM386), the same Sound Blaster drivers, and the same AUTOEXEC.BAT and CONFIG.SYS tuning that made the original machine run smoothly. Even the directory structure and file names match the original setup.</p>
<h2>Context: DOSMENU</h2> <h2>Context: DOSMENU</h2>
<p>A key part of the preserved environment is <strong><a href="https://28k8.moonweb.org/bbs/pc/dosmenu/">DOSMENU</a></strong>, a boot configuration utility written in Turbo Pascal 6.0 in 1994 — originally developed specifically for this Colani 486 machine. DOSMENU let you choose at boot time which AUTOEXEC.BAT and CONFIG.SYS commands to load — a critical feature in the DOS era, where memory management was manual and games often required different driver configurations. Features included saveable presets, a built-in MOD player, PCX picture display, and a DoubleSpace memory unloader. The original source: ~50k of Turbo Pascal, roughly 1900 lines plus some inline assembler.</p> <p>A key part of the preserved environment is <strong><a href="https://28k8.moonweb.org/bbs/pc/dosmenu/">DOSMENU</a></strong>, a boot configuration utility written in Turbo Pascal 6.0 in 1994, originally developed specifically for this Colani 486 machine. DOSMENU let you choose at boot time which AUTOEXEC.BAT and CONFIG.SYS commands to load, a critical feature in the DOS era, where memory management was manual and games often required different driver configurations. Features included saveable presets, a built-in MOD player, PCX picture display, and a DoubleSpace memory unloader. The original source: ~50k of Turbo Pascal, roughly 1900 lines plus some inline assembler.</p>
<h2>Why a 1:1 recreation</h2> <h2>Why a 1:1 recreation</h2>
<p>The point isn't just to play DOS games — it's to preserve the exact computing experience from that era. The way DOS memory management worked, the way sound drivers had to be configured manually, the way batch files chained together to create a usable environment. This level of detail is lost when you just launch a DOS game from a modern frontend. The MiSTer setup recreates the entire boot process, from BIOS POST to the C:\> prompt, just as it was in the mid-90s.</p> <p>The point isn't just to play DOS games, it's to preserve the exact computing experience from that era. The way DOS memory management worked, the way sound drivers had to be configured manually, the way batch files chained together to create a usable environment. This level of detail is lost when you just launch a DOS game from a modern frontend. The MiSTer setup recreates the entire boot process, from BIOS POST to the C:\> prompt, just as it was in the mid-90s.</p>
<h2>What's preserved</h2> <h2>What's preserved</h2>
<ul> <ul>
<li><strong>DOS boot environment</strong> — the exact AUTOEXEC.BAT and CONFIG.SYS from the original machine, including memory manager configuration and Sound Blaster IRQ settings.</li> <li><strong>DOS boot environment</strong>, the exact AUTOEXEC.BAT and CONFIG.SYS from the original machine, including memory manager configuration and Sound Blaster IRQ settings.</li>
<li><strong>Game installations</strong> — original game installations with their setup utilities, configuration files, and saved games, all in the same directory paths as the original.</li> <li><strong>Game installations</strong>, original game installations with their setup utilities, configuration files, and saved games, all in the same directory paths as the original.</li>
<li><strong>System utilities</strong> — Norton Commander, PKZIP, the usual toolkit of the era, all configured the way they were used daily.</li> <li><strong>System utilities</strong>, Norton Commander, PKZIP, the usual toolkit of the era, all configured the way they were used daily.</li>
<li><strong><a href="https://28k8.moonweb.org/bbs/pc/dosmenu/">DOSMENU</a></strong> — a custom boot configuration utility written specifically for this Colani 486 in 1994. Built in Turbo Pascal, it let you choose at boot time which AUTOEXEC.BAT and CONFIG.SYS commands to load — essential for managing DOS memory across different games and applications.</li> <li><strong><a href="https://28k8.moonweb.org/bbs/pc/dosmenu/">DOSMENU</a></strong>, a custom boot configuration utility written specifically for this Colani 486 in 1994. Built in Turbo Pascal, it let you choose at boot time which AUTOEXEC.BAT and CONFIG.SYS commands to load, essential for managing DOS memory across different games and applications.</li>
<li><strong>The quirks</strong> — the specific workarounds and tweaks that were needed to get everything running together, the kind of institutional knowledge that only exists on a well-used personal machine.</li> <li><strong>The quirks</strong>, the specific workarounds and tweaks that were needed to get everything running together, the kind of institutional knowledge that only exists on a well-used personal machine.</li>
</ul> </ul>
<h2>AO486 vs real hardware</h2> <h2>AO486 vs real hardware</h2>
<p>The MiSTer AO486 core handles the emulation convincingly. Timing accuracy is good enough for games that rely on cycle-sensitive behavior. Sound Blaster emulation covers the standard OPL3 FM synthesis and digital audio. The main trade-off is that some hardware-specific quirks — like the exact timing of the ET4000 graphics card or the behavior of specific ISA bus peripherals — aren't perfectly replicated. But for the purpose of preserving a working DOS environment, it's more than sufficient.</p> <p>The MiSTer AO486 core handles the emulation convincingly. Timing accuracy is good enough for games that rely on cycle-sensitive behavior. Sound Blaster emulation covers the standard OPL3 FM synthesis and digital audio. The main trade-off is that some hardware-specific quirks, like the exact timing of the ET4000 graphics card or the behavior of specific ISA bus peripherals, aren't perfectly replicated. But for the purpose of preserving a working DOS environment, it's more than sufficient.</p>
</div> </div>
retro/et4000-experiment/index.njk
+5 -5
View File
@@ -13,14 +13,14 @@ parent: "/retro/"
<h2>What Went Wrong</h2> <h2>What Went Wrong</h2>
<ul> <ul>
<li><strong>Power supply overload</strong> — the stock Mega ST PSU simply wasn't designed to feed a MonSTerBoard, an ET4000, and a MegaBus adapter at the same time. The 5V and 12V rails ended up right at their instability limit.</li> <li><strong>Power supply overload</strong>, the stock Mega ST PSU simply wasn't designed to feed a MonSTerBoard, an ET4000, and a MegaBus adapter at the same time. The 5V and 12V rails ended up right at their instability limit.</li>
<li><strong>The missing -12V</strong> — the ET4000 needs -12V for its output drivers, usually tapped from the RS-232 port, adding yet another drain on an already stressed supply.</li> <li><strong>The missing -12V</strong>, the ET4000 needs -12V for its output drivers, usually tapped from the RS-232 port, adding yet another drain on an already stressed supply.</li>
<li><strong>Heat with nowhere to go</strong> — the card runs hot, and the Mega ST case has zero provision for active cooling.</li> <li><strong>Heat with nowhere to go</strong>, the card runs hot, and the Mega ST case has zero provision for active cooling.</li>
<li><strong>Signal integrity issues</strong> — the ISA-to-MegaBus adapter introduces reflections at higher frequencies, and access to ET4000 RAM is roughly 65% slower than native ST RAM, with no hardware BitBLT to make up for it.</li> <li><strong>Signal integrity issues</strong>, the ISA-to-MegaBus adapter introduces reflections at higher frequencies, and access to ET4000 RAM is roughly 65% slower than native ST RAM, with no hardware BitBLT to make up for it.</li>
</ul> </ul>
<h2>The Practical Fallback</h2> <h2>The Practical Fallback</h2>
<p>FPGA cores on MiSTer and MiST only emulate the native ST Shifter — there's no ET4000 emulation on the horizon given the FPGA resources it would require. The real answer for higher resolutions turned out to be software: Hatari running in VDI mode on a Raspberry Pi 400, which comfortably delivers 1280×768 with far more colors and zero PSU drama.</p> <p>FPGA cores on MiSTer and MiST only emulate the native ST Shifter, there's no ET4000 emulation on the horizon given the FPGA resources it would require. The real answer for higher resolutions turned out to be software: Hatari running in VDI mode on a Raspberry Pi 400, which comfortably delivers 1280×768 with far more colors and zero PSU drama.</p>
<p class="redacted-note">Lesson learned: "just add a graphics card" on 40-year-old hardware is never "just" anything. The ET4000 project isn't dead, but it's shelved until there's a properly beefed-up power supply and cooling solution to go with it.</p> <p class="redacted-note">Lesson learned: "just add a graphics card" on 40-year-old hardware is never "just" anything. The ET4000 project isn't dead, but it's shelved until there's a properly beefed-up power supply and cooling solution to go with it.</p>
</div> </div>
retro/index.njk
+2 -2
View File
@@ -1,8 +1,8 @@
--- ---
title: "Retro Computing — Vintage Hardware & FPGA | moonweb" title: "Retro Computing, Vintage Hardware & FPGA | moonweb"
section: "retro" section: "retro"
tags: "retro" tags: "retro"
description: "Physical retro hardware collection — Raspberry Pi units, Pi 400, and vintage computing setups." description: "Physical retro hardware collection, Raspberry Pi units, Pi 400, and vintage computing setups."
layout: base.njk layout: base.njk
sections: sections:
- heading: "Atari ST Corner" - heading: "Atari ST Corner"
retro/mini-arcade-handheld-roundup/index.njk
+5 -5
View File
@@ -13,11 +13,11 @@ parent: "/retro/"
<h2>What to Watch For</h2> <h2>What to Watch For</h2>
<ul> <ul>
<li><strong>The bootleg trap</strong> — devices built around VT369-style chips (NES-compatible, not real arcade hardware) often bundle 200+ "games" that turn out to be unlicensed clones and homebrew hacks, cataloged in enthusiast bootleg-game wikis. No Pac-Man, no Donkey Kong, just look-alikes.</li> <li><strong>The bootleg trap</strong>, devices built around VT369-style chips (NES-compatible, not real arcade hardware) often bundle 200+ "games" that turn out to be unlicensed clones and homebrew hacks, cataloged in enthusiast bootleg-game wikis. No Pac-Man, no Donkey Kong, just look-alikes.</li>
<li><strong>Genuinely licensed compact options</strong> — devices built around real Namco or SNK licenses deliver actual classics (Pac-Man, Galaga, Dig Dug, dozens of Neo Geo titles) with premium displays, at a similar price point to the bootleg boxes.</li> <li><strong>Genuinely licensed compact options</strong>, devices built around real Namco or SNK licenses deliver actual classics (Pac-Man, Galaga, Dig Dug, dozens of Neo Geo titles) with premium displays, at a similar price point to the bootleg boxes.</li>
<li><strong>Expandable options</strong> — some cabinet-style devices accept extra MAME ROMs via SD card, letting you add real arcade classics after the fact — the one feature the bootleg boxes almost never offer without heavy soldering.</li> <li><strong>Expandable options</strong>, some cabinet-style devices accept extra MAME ROMs via SD card, letting you add real arcade classics after the fact, the one feature the bootleg boxes almost never offer without heavy soldering.</li>
<li><strong>The maximalist route</strong> — a Raspberry Pi running RetroPie, or a MiSTer FPGA setup, both scale to tens of thousands of games or hardware-accurate arcade cores respectively, at the cost of more setup effort.</li> <li><strong>The maximalist route</strong>, a Raspberry Pi running RetroPie, or a MiSTer FPGA setup, both scale to tens of thousands of games or hardware-accurate arcade cores respectively, at the cost of more setup effort.</li>
</ul> </ul>
<p class="redacted-note">If a mini-arcade's game list reads like a string of titles you've genuinely never heard of, that's the tell. Real arcade classics have instantly recognisable names — bootlegs hide behind vague, generic-sounding ones.</p> <p class="redacted-note">If a mini-arcade's game list reads like a string of titles you've genuinely never heard of, that's the tell. Real arcade classics have instantly recognisable names, bootlegs hide behind vague, generic-sounding ones.</p>
</div> </div>
retro/mist-mister-fpga/index.njk
+5 -5
View File
@@ -9,14 +9,14 @@ parent: "/retro/"
<h1>MiST and MiSTer: Two FPGA Boxes, Two Very Different Personalities</h1> <h1>MiST and MiSTer: Two FPGA Boxes, Two Very Different Personalities</h1>
<div class="detail-content"> <div class="detail-content">
<p>Both boards run classic computers as hardware-accurate FPGA cores rather than software emulation, but living with them side by side reveals they're not interchangeable — each one is simply better at certain jobs.</p> <p>Both boards run classic computers as hardware-accurate FPGA cores rather than software emulation, but living with them side by side reveals they're not interchangeable, each one is simply better at certain jobs.</p>
<h2>Division of Labour</h2> <h2>Division of Labour</h2>
<ul> <ul>
<li><strong>MiST for Atari ST</strong> — older and more limited in scope than MiSTer, but its ST/STE core is rock solid. It's the default choice for anything Atari, paired with a network-capable configuration for file transfers.</li> <li><strong>MiST for Atari ST</strong>, older and more limited in scope than MiSTer, but its ST/STE core is rock solid. It's the default choice for anything Atari, paired with a network-capable configuration for file transfers.</li>
<li><strong>MiSTer for everything else</strong> — Amiga, DOS (via the AO486 core), and a huge library of other systems. Its Atari ST core, oddly, is the less reliable one of the two boards — reason enough to keep MiST around instead of retiring it.</li> <li><strong>MiSTer for everything else</strong>, Amiga, DOS (via the AO486 core), and a huge library of other systems. Its Atari ST core, oddly, is the less reliable one of the two boards, reason enough to keep MiST around instead of retiring it.</li>
<li><strong>Native resolution ceiling</strong> — both boards only emulate the original Shifter chip on Atari ST: Low-Res (320×200), Medium-Res (640×200), and High-Res (640×400 mono). No ET4000-style graphics card emulation is on the table; a Falcon030 core with higher native resolutions has been rumoured for years without shipping.</li> <li><strong>Native resolution ceiling</strong>, both boards only emulate the original Shifter chip on Atari ST: Low-Res (320×200), Medium-Res (640×200), and High-Res (640×400 mono). No ET4000-style graphics card emulation is on the table; a Falcon030 core with higher native resolutions has been rumoured for years without shipping.</li>
<li><strong>Known rough edges</strong> — accessory programs loading but not appearing on the desktop, and disabled XBOOT support, are still open items to chase down on the MiST side.</li> <li><strong>Known rough edges</strong>, accessory programs loading but not appearing on the desktop, and disabled XBOOT support, are still open items to chase down on the MiST side.</li>
</ul> </ul>
<p class="redacted-note">The takeaway after months of use: don't assume the shinier, newer board wins every category. Sometimes the "older" FPGA implementation is simply the more mature one for a specific chipset.</p> <p class="redacted-note">The takeaway after months of use: don't assume the shinier, newer board wins every category. Sometimes the "older" FPGA implementation is simply the more mature one for a specific chipset.</p>
retro/pimiga-journey/index.njk
+6 -6
View File
@@ -2,7 +2,7 @@
title: "PiMiga 3 to 5: Chasing the Perfect Amiga-on-a-Pi Distro" title: "PiMiga 3 to 5: Chasing the Perfect Amiga-on-a-Pi Distro"
section: "retro" section: "retro"
tags: "retro" tags: "retro"
description: "Three generations of PiMiga later, here's what actually changed — and why newer isn't always better." description: "Three generations of PiMiga later, here's what actually changed, and why newer isn't always better."
layout: base.njk layout: base.njk
parent: "/retro/" parent: "/retro/"
--- ---
@@ -13,11 +13,11 @@ parent: "/retro/"
<h2>What Changed Across Versions</h2> <h2>What Changed Across Versions</h2>
<ul> <ul>
<li><strong>PiMiga 3.0 (2022)</strong> — the original recipe: Amiberry under DietPi with an XFCE desktop and the iGame v2 launcher. Now effectively retired, with no further updates.</li> <li><strong>PiMiga 3.0 (2022)</strong>, the original recipe: Amiberry under DietPi with an XFCE desktop and the iGame v2 launcher. Now effectively retired, with no further updates.</li>
<li><strong>PiMiga 4.0 (2024)</strong> — brought multi-architecture support (ARM and Intel/AMD) and swapped in the MegaAGS launcher, with roughly 4,800 games pre-installed. Its biggest strength: MegaAGS ships with strong trainer/cheat support baked into its game profiles.</li> <li><strong>PiMiga 4.0 (2024)</strong>, brought multi-architecture support (ARM and Intel/AMD) and swapped in the MegaAGS launcher, with roughly 4,800 games pre-installed. Its biggest strength: MegaAGS ships with strong trainer/cheat support baked into its game profiles.</li>
<li><strong>PiMiga 5.0 (Dec 2025, final release)</strong> — the creator's last planned version. Switches back to the simpler iGame v2 launcher, drops in 600+ auto-switching backdrops, and adds native support for the newest Pi hardware — but loses much of the trainer-focused depth that made version 4 so useful for casual replay sessions.</li> <li><strong>PiMiga 5.0 (Dec 2025, final release)</strong>, the creator's last planned version. Switches back to the simpler iGame v2 launcher, drops in 600+ auto-switching backdrops, and adds native support for the newest Pi hardware, but loses much of the trainer-focused depth that made version 4 so useful for casual replay sessions.</li>
<li><strong>The real decision criterion</strong> — not "which is newest" but "do I care more about trainers/cheats (stick with v4's MegaAGS) or about the latest hardware support and polish (go with v5's iGame)". Running both on separate SD cards sidesteps the choice entirely.</li> <li><strong>The real decision criterion</strong>, not "which is newest" but "do I care more about trainers/cheats (stick with v4's MegaAGS) or about the latest hardware support and polish (go with v5's iGame)". Running both on separate SD cards sidesteps the choice entirely.</li>
</ul> </ul>
<p class="redacted-note">Also worth knowing: getting the display right on an older 4:3 monitor is almost never a Raspberry Pi config.txt problem — it's nearly always an Amiberry/X11 scaling setting. Check the emulator's Integer Scaling option before touching boot config files.</p> <p class="redacted-note">Also worth knowing: getting the display right on an older 4:3 monitor is almost never a Raspberry Pi config.txt problem, it's nearly always an Amiberry/X11 scaling setting. Check the emulator's Integer Scaling option before touching boot config files.</p>
</div> </div>
retro/retro-corner-snapshot/index.njk
+6 -6
View File
@@ -13,20 +13,20 @@ parent: "/retro/"
<h2>Room by Room</h2> <h2>Room by Room</h2>
<ul> <ul>
<li><strong>The office</strong> — home base for the serious gear: three Atari Mega ST 4 machines plus a Mega ST 2, a MiSTer FPGA covering Amiga/DOS/arcade cores, and a MiST FPGA dedicated to Atari ST. A Raspberry Pi 400 running Hatari and PiMiga rounds things out as the flexible, high-resolution alternative to hardware that can't quite keep up on screen quality.</li> <li><strong>The office</strong>, home base for the serious gear: three Atari Mega ST 4 machines plus a Mega ST 2, a MiSTer FPGA covering Amiga/DOS/arcade cores, and a MiST FPGA dedicated to Atari ST. A Raspberry Pi 400 running Hatari and PiMiga rounds things out as the flexible, high-resolution alternative to hardware that can't quite keep up on screen quality.</li>
<li><strong>The living room</strong> — a Batocera box for casual, couch-friendly sessions across dozens of systems, from Atari ST to arcade to consoles, no FPGA timing precision required.</li> <li><strong>The living room</strong>, a Batocera box for casual, couch-friendly sessions across dozens of systems, from Atari ST to arcade to consoles, no FPGA timing precision required.</li>
<li><strong>On the move</strong> — a small collection of handheld emulation devices for retro gaming away from a screen.</li> <li><strong>On the move</strong>, a small collection of handheld emulation devices for retro gaming away from a screen.</li>
<li><strong>The focus, unmistakably</strong> — Amiga and Atari ST are the heart of the collection; everything else (DOS, arcade, handhelds) exists to fill in the gaps around those two platforms.</li> <li><strong>The focus, unmistakably</strong>, Amiga and Atari ST are the heart of the collection; everything else (DOS, arcade, handhelds) exists to fill in the gaps around those two platforms.</li>
</ul> </ul>
<h2>The office setup in detail</h2> <h2>The office setup in detail</h2>
<p>The office houses the main retro computing workstation. The three Mega ST 4 machines each have a defined role: one runs GEM-based productivity software, one is set up for MIDI music production with a connected synthesizer, and one serves as the development machine for ST software. The Mega ST 2 is the backup and testing machine. The MiSTer FPGA handles Amiga and DOS emulation with near-perfect timing accuracy, while the MiST provides a dedicated Atari ST experience without the overhead of the MiSTer's larger feature set.</p> <p>The office houses the main retro computing workstation. The three Mega ST 4 machines each have a defined role: one runs GEM-based productivity software, one is set up for MIDI music production with a connected synthesizer, and one serves as the development machine for ST software. The Mega ST 2 is the backup and testing machine. The MiSTer FPGA handles Amiga and DOS emulation with near-perfect timing accuracy, while the MiST provides a dedicated Atari ST experience without the overhead of the MiSTer's larger feature set.</p>
<h2>The living room setup</h2> <h2>The living room setup</h2>
<p>The Batocera box in the living room takes a completely different approach — it's not about hardware authenticity, it's about convenience. A compact PC boots directly into the Batocera frontend, providing a console-like experience for casual retro gaming on the TV. Controllers are wireless, the interface is controller-friendly, and the focus is on games that work well from a couch rather than demos or productivity software that needs a keyboard and mouse.</p> <p>The Batocera box in the living room takes a completely different approach, it's not about hardware authenticity, it's about convenience. A compact PC boots directly into the Batocera frontend, providing a console-like experience for casual retro gaming on the TV. Controllers are wireless, the interface is controller-friendly, and the focus is on games that work well from a couch rather than demos or productivity software that needs a keyboard and mouse.</p>
<h2>How it all connects</h2> <h2>How it all connects</h2>
<p>All machines share the same network infrastructure, with FTP access to the NAS for file transfers. Disk images and software archives are stored centrally and served to individual machines as needed. The MiSTer and MiST both use SD cards for storage, while the original hardware uses CompactFlash adapters or SCSI-to-IDE bridges depending on the machine. The Raspberry Pi 400 accesses the same NFS shares as the other Pis in the flat.</p> <p>All machines share the same network infrastructure, with FTP access to the NAS for file transfers. Disk images and software archives are stored centrally and served to individual machines as needed. The MiSTer and MiST both use SD cards for storage, while the original hardware uses CompactFlash adapters or SCSI-to-IDE bridges depending on the machine. The Raspberry Pi 400 accesses the same NFS shares as the other Pis in the flat.</p>
<p class="redacted-note">The honest status update: several open items remain — documenting the exact handheld models, and waiting on the A1200 Mini pre-order. Retro computing, it turns out, is never really "done."</p> <p class="redacted-note">The honest status update: several open items remain, documenting the exact handheld models, and waiting on the A1200 Mini pre-order. Retro computing, it turns out, is never really "done."</p>
</div> </div>
retro/the-a1200/index.njk
+9 -9
View File
@@ -9,20 +9,20 @@ parent: "/retro/"
<h1>The A1200: What We Know Before It Ships</h1> <h1>The A1200: What We Know Before It Ships</h1>
<div class="detail-content"> <div class="detail-content">
<p>Retro Games Ltd. opened pre-orders for The A1200 in November 2025, with delivery slated for June 2026 at a suggested price of 189.99 Euro. As with their previous mini-consoles, official details have trickled out slowly — deliberately, it seems.</p> <p>Retro Games Ltd. opened pre-orders for The A1200 in November 2025, with delivery slated for June 2026 at a suggested price of 189.99 Euro. As with their previous mini-consoles, official details have trickled out slowly, deliberately, it seems.</p>
<h2>What's Confirmed So Far</h2> <h2>What's Confirmed So Far</h2>
<ul> <ul>
<li><strong>Games</strong> — 25 titles pre-installed in total, with only 9 officially confirmed at time of writing: the Turrican trilogy, Defender of the Crown I & II, Beneath a Steel Sky, Lure of the Temptress, Ruff 'n' Tumble, and The Settlers II. The rest are being held back, likely on purpose, as a marketing drip-feed.</li> <li><strong>Games</strong>, 25 titles pre-installed in total, with only 9 officially confirmed at time of writing: the Turrican trilogy, Defender of the Crown I & II, Beneath a Steel Sky, Lure of the Temptress, Ruff 'n' Tumble, and The Settlers II. The rest are being held back, likely on purpose, as a marketing drip-feed.</li>
<li><strong>Hardware essentials</strong> — a fully functional keyboard in the original design, a CD32-style gamepad, a tank-mouse replica, HDMI output, USB-A ports, and USB sideloading support for ADF/HDF game files.</li> <li><strong>Hardware essentials</strong>, a fully functional keyboard in the original design, a CD32-style gamepad, a tank-mouse replica, HDMI output, USB-A ports, and USB sideloading support for ADF/HDF game files.</li>
<li><strong>The mystery SD slot</strong> — a slot spotted at Gamescom 2025 in the position of the original floppy drive, widely speculated to be for SD-card-based game loading, but not officially confirmed.</li> <li><strong>The mystery SD slot</strong>, a slot spotted at Gamescom 2025 in the position of the original floppy drive, widely speculated to be for SD-card-based game loading, but not officially confirmed.</li>
<li><strong>Expected internals</strong> — likely an Allwinner ARM SoC running Amiberry underneath (following the pattern of the earlier A500 Mini), rather than genuine 68EC020 silicon.</li> <li><strong>Expected internals</strong>, likely an Allwinner ARM SoC running Amiberry underneath (following the pattern of the earlier A500 Mini), rather than genuine 68EC020 silicon.</li>
</ul> </ul>
<h2>Context: where this fits in the retro landscape</h2> <h2>Context: where this fits in the retro landscape</h2>
<p>The A1200 sits in a growing market of mini-consoles and FPGA recreations. <p>The A1200 sits in a growing market of mini-consoles and FPGA recreations.
Unlike MiSTer FPGA, which aims for cycle-accurate hardware reproduction, Unlike MiSTer FPGA, which aims for cycle-accurate hardware reproduction,
the A1200 is a software emulation box — similar to the A500 Mini that the A1200 is a software emulation box, similar to the A500 Mini that
Retro Games Ltd. released in 2022. The A500 Mini was well-received but Retro Games Ltd. released in 2022. The A500 Mini was well-received but
had its quirks: limited game selection, no keyboard support, and an had its quirks: limited game selection, no keyboard support, and an
emulation layer that didn't quite match the original hardware in all emulation layer that didn't quite match the original hardware in all
@@ -36,7 +36,7 @@ monitor is the purist approach, but requires maintenance and debugging.
MiSTer FPGA offers near-perfect hardware reproduction but costs MiSTer FPGA offers near-perfect hardware reproduction but costs
significantly more. Amiberry on a Raspberry Pi is the budget option, significantly more. Amiberry on a Raspberry Pi is the budget option,
with good compatibility but occasional timing issues. The A1200 with good compatibility but occasional timing issues. The A1200
positions itself as a plug-and-play alternative — set it up in minutes, positions itself as a plug-and-play alternative, set it up in minutes,
no configuration required, at the cost of some flexibility.</p> no configuration required, at the cost of some flexibility.</p>
<h2>What this means for the Amiga community</h2> <h2>What this means for the Amiga community</h2>
@@ -44,8 +44,8 @@ no configuration required, at the cost of some flexibility.</p>
it could become the default entry point for people curious about the it could become the default entry point for people curious about the
Amiga. The original hardware is getting harder to find in good condition, Amiga. The original hardware is getting harder to find in good condition,
and the capacitor-bomb motherboard issue means that even working units need and the capacitor-bomb motherboard issue means that even working units need
recapping. A modern, reliable reproduction — even if it's software recapping. A modern, reliable reproduction, even if it's software
emulation — lowers the barrier significantly. The 25 pre-installed games emulation, lowers the barrier significantly. The 25 pre-installed games
give newcomers a curated starting point, and the USB sideloading support give newcomers a curated starting point, and the USB sideloading support
means the library can be expanded without hardware modifications.</p> means the library can be expanded without hardware modifications.</p>
scripts/github-aggregator/README.md
+1 -1
View File
@@ -1,7 +1,7 @@
# GitHub aggregator # GitHub aggregator
Reads a `.moonweb.yml` file from each of the `skoelle` GitHub repos and Reads a `.moonweb.yml` file from each of the `skoelle` GitHub repos and
regenerates `code/_data/repos.json`. Run manually — not wired into CI. regenerates `code/_data/repos.json`. Run manually, not wired into CI.
## Usage ## Usage
smarthome/airplay-audio/index.njk
+2 -2
View File
@@ -25,7 +25,7 @@ speaker app.</p>
a software mixer stage (<code>+20&nbsp;dB</code> via an ALSA softvol plugin) sits between a software mixer stage (<code>+20&nbsp;dB</code> via an ALSA softvol plugin) sits between
shairport-sync and the hardware. The hardware mixer itself is deliberately shairport-sync and the hardware. The hardware mixer itself is deliberately
left fixed at 100% / 0&nbsp;dB, and shairport-sync only ever adjusts its own left fixed at 100% / 0&nbsp;dB, and shairport-sync only ever adjusts its own
internal software volume — this avoids the volume jumps and mixer internal software volume, this avoids the volume jumps and mixer
conflicts that show up when multiple layers all try to control loudness.</p> conflicts that show up when multiple layers all try to control loudness.</p>
<h2>Bathroom-specific integration</h2> <h2>Bathroom-specific integration</h2>
@@ -33,7 +33,7 @@ conflicts that show up when multiple layers all try to control loudness.</p>
radio stream automatically when its light sensor detects the light has radio stream automatically when its light sensor detects the light has
been switched on. A simple flag file signals whether AirPlay is currently been switched on. A simple flag file signals whether AirPlay is currently
active, so the automatic radio stream politely stays off while someone is active, so the automatic radio stream politely stays off while someone is
actively AirPlaying — and resumes its normal behavior as soon as the actively AirPlaying, and resumes its normal behavior as soon as the
AirPlay session ends.</p> AirPlay session ends.</p>
<h2>Why this design</h2> <h2>Why this design</h2>
smarthome/balkonpi/index.njk
+4 -4
View File
@@ -13,13 +13,13 @@ layout: base.njk
self-contained smart-balcony controller: light control, an internet self-contained smart-balcony controller: light control, an internet
radio player, sensor readings, and its own backup/monitoring, all on radio player, sensor readings, and its own backup/monitoring, all on
very modest hardware. The goal is to make the balcony a more pleasant very modest hardware. The goal is to make the balcony a more pleasant
space — light control when it gets dark, background music while sitting space, light control when it gets dark, background music while sitting
outside, and temperature readings to know what to expect before stepping outside, and temperature readings to know what to expect before stepping
out.</p> out.</p>
<h2>What it does</h2> <h2>What it does</h2>
<ul> <ul>
<li><strong>Audio output</strong> via a DIY PWM-to-analog circuit on two GPIO pins — the Pi Zero has no built-in audio jack, so this is a well-known community workaround using a small RC filter.</li> <li><strong>Audio output</strong> via a DIY PWM-to-analog circuit on two GPIO pins, the Pi Zero has no built-in audio jack, so this is a well-known community workaround using a small RC filter.</li>
<li><strong>Physical controls</strong>: a push button for the balcony light and a toggle switch for music playback, read directly via GPIO.</li> <li><strong>Physical controls</strong>: a push button for the balcony light and a toggle switch for music playback, read directly via GPIO.</li>
<li><strong>HTTP communication</strong> with an ESP32-based balcony light controller.</li> <li><strong>HTTP communication</strong> with an ESP32-based balcony light controller.</li>
<li><strong>Internet radio</strong> via a lightweight command-line audio player, controlled by a small custom tool ported from an earlier bathroom-Pi project.</li> <li><strong>Internet radio</strong> via a lightweight command-line audio player, controlled by a small custom tool ported from an earlier bathroom-Pi project.</li>
@@ -28,7 +28,7 @@ out.</p>
<h2>Hardware details</h2> <h2>Hardware details</h2>
<p>The Pi Zero W is connected to an ESP32 microcontroller that handles <p>The Pi Zero W is connected to an ESP32 microcontroller that handles
the actual light switching via a relay module. Communication happens the actual light switching via a relay module. Communication happens
over HTTP — the Pi sends a simple request to toggle the relay, and the over HTTP, the Pi sends a simple request to toggle the relay, and the
ESP32 responds with the current light state. The ESP32 also reads a ESP32 responds with the current light state. The ESP32 also reads a
DHT22 temperature and humidity sensor, publishing the values to MQTT DHT22 temperature and humidity sensor, publishing the values to MQTT
for the home dashboard. This split makes sense: the Pi handles the for the home dashboard. This split makes sense: the Pi handles the
@@ -37,7 +37,7 @@ handles the real-time hardware control.</p>
<h2>Monitoring & backup, even on tiny hardware</h2> <h2>Monitoring & backup, even on tiny hardware</h2>
<p>Despite the very limited RAM, this Pi still participates in the same <p>Despite the very limited RAM, this Pi still participates in the same
Prometheus monitoring pattern as the rest of the fleet — with the metrics Prometheus monitoring pattern as the rest of the fleet, with the metrics
collector's default configuration trimmed down to only the essential collector's default configuration trimmed down to only the essential
collectors, since the full default set noticeably overloaded the CPU on collectors, since the full default set noticeably overloaded the CPU on
this specific board. It also runs a nightly rsync backup of its own this specific board. It also runs a nightly rsync backup of its own
smarthome/home-dashboard-controls/index.njk
+15 -15
View File
@@ -13,30 +13,30 @@ parent: "/smarthome/"
<h2>Lighting & ambience</h2> <h2>Lighting & ambience</h2>
<ul> <ul>
<li><strong>Balkon</strong> — LED on/off toggle for the balcony lighting.</li> <li><strong>Balkon</strong>, LED on/off toggle for the balcony lighting.</li>
<li><strong>Beleuchtung (all rooms)</strong> — global "all on"/"all off" shortcut plus per-room toggles for living room and kitchen lighting.</li> <li><strong>Beleuchtung (all rooms)</strong>, global "all on"/"all off" shortcut plus per-room toggles for living room and kitchen lighting.</li>
</ul> </ul>
<h2>Music controls</h2> <h2>Music controls</h2>
<ul> <ul>
<li><strong>Kitchen</strong> — quick buttons for SomaFM, a classical station, and an off switch, all routed to the kitchen speaker.</li> <li><strong>Kitchen</strong>, quick buttons for SomaFM, a classical station, and an off switch, all routed to the kitchen speaker.</li>
<li><strong>Living room</strong> — the same SomaFM/classical options plus dedicated Lounge, 2000s and Deluxe stations, and an off switch.</li> <li><strong>Living room</strong>, the same SomaFM/classical options plus dedicated Lounge, 2000s and Deluxe stations, and an off switch.</li>
</ul> </ul>
<h2>Tasmota smart plugs</h2> <h2>Tasmota smart plugs</h2>
<ul> <ul>
<li><strong>TasmoAdmin</strong> — link straight into the central Tasmota device dashboard for firmware and config management.</li> <li><strong>TasmoAdmin</strong>, link straight into the central Tasmota device dashboard for firmware and config management.</li>
<li><strong>Power strips (Gosund P1)</strong> — two multi-outlet strips covering the home-office network gear and backup drive, and the iMac corner with its peripherals.</li> <li><strong>Power strips (Gosund P1)</strong>, two multi-outlet strips covering the home-office network gear and backup drive, and the iMac corner with its peripherals.</li>
<li><strong>Single sockets (SP112)</strong> — four individually switchable outlets covering hallway, kitchen, dryer and washing machine circuits, each also powering an associated LED matrix or small Pi.</li> <li><strong>Single sockets (SP112)</strong>, four individually switchable outlets covering hallway, kitchen, dryer and washing machine circuits, each also powering an associated LED matrix or small Pi.</li>
<li><strong>Other plug families</strong> — additional Nous, Eightree (ESP32-based), Athom and IDS smart plugs cover spare capacity, living-room seating/desk outlets, storage room, kitchen appliances, and a few legacy TV/PC outlets (several currently marked defective).</li> <li><strong>Other plug families</strong>, additional Nous, Eightree (ESP32-based), Athom and IDS smart plugs cover spare capacity, living-room seating/desk outlets, storage room, kitchen appliances, and a few legacy TV/PC outlets (several currently marked defective).</li>
</ul> </ul>
<p>The plug collection has grown over time as different brands became available at different price points. Gosund P1 multi-socket strips handle areas with multiple devices, while the SP112 single sockets cover dedicated appliances. Eightree plugs are ESP32-based and flashable to Tasmota, while the Athom and IDS models round out the remaining circuits. A Tasmota RF Bridge extends the ecosystem to RF-only devices like the 3D printer plug in the study.</p> <p>The plug collection has grown over time as different brands became available at different price points. Gosund P1 multi-socket strips handle areas with multiple devices, while the SP112 single sockets cover dedicated appliances. Eightree plugs are ESP32-based and flashable to Tasmota, while the Athom and IDS models round out the remaining circuits. A Tasmota RF Bridge extends the ecosystem to RF-only devices like the 3D printer plug in the study.</p>
<h2>Special devices &amp; status</h2> <h2>Special devices &amp; status</h2>
<ul> <ul>
<li><strong>Tasmota RF Bridge / Delock</strong> — bridges RF-only devices (like the home-office 3D printer plug) into the Tasmota/MQTT ecosystem.</li> <li><strong>Tasmota RF Bridge / Delock</strong>, bridges RF-only devices (like the home-office 3D printer plug) into the Tasmota/MQTT ecosystem.</li>
<li><strong>LED Matrix restarts</strong> — one-click restart buttons for each of the six room LED matrix displays (living room, study, bathroom, kitchen, hallway, Nepomuk room), avoiding a manual power-cycle.</li> <li><strong>LED Matrix restarts</strong>, one-click restart buttons for each of the six room LED matrix displays (living room, study, bathroom, kitchen, hallway, Nepomuk room), avoiding a manual power-cycle.</li>
<li><strong>Healthchecks &amp; version info</strong> — direct links into the Healthchecks dashboard and a version/status overview for the connected devices.</li> <li><strong>Healthchecks &amp; version info</strong>, direct links into the Healthchecks dashboard and a version/status overview for the connected devices.</li>
</ul> </ul>
<h2>Calendar &amp; weather</h2> <h2>Calendar &amp; weather</h2>
@@ -47,10 +47,10 @@ parent: "/smarthome/"
<h2>Network shortcuts</h2> <h2>Network shortcuts</h2>
<ul> <ul>
<li><strong>Routers & mesh</strong> — quick links to the main Fritzbox and its mesh repeaters covering different floors and the kitchen/IoT band.</li> <li><strong>Routers & mesh</strong>, quick links to the main Fritzbox and its mesh repeaters covering different floors and the kitchen/IoT band.</li>
<li><strong>Switches</strong> — direct access to each Zyxel switch's admin page (server room, living room, PowerLAN, OpenWRT segment).</li> <li><strong>Switches</strong>, direct access to each Zyxel switch's admin page (server room, living room, PowerLAN, OpenWRT segment).</li>
<li><strong>OpenWRT & mini router</strong> — links into the OpenWRT admin UI and the small travel router used for testing.</li> <li><strong>OpenWRT & mini router</strong>, links into the OpenWRT admin UI and the small travel router used for testing.</li>
<li><strong>Sensors & MQTT</strong> — shortcuts to server-room and balcony sensor readings, bathroom analog values, and background worker/API status pages (HTML and JSON views).</li> <li><strong>Sensors & MQTT</strong>, shortcuts to server-room and balcony sensor readings, bathroom analog values, and background worker/API status pages (HTML and JSON views).</li>
</ul> </ul>
<p class="redacted-note">This dashboard is intentionally simple: static HTML, no login, meant purely for convenience on the local network rather than as a secured control surface. Device names and room assignments are shown for structure; a few outlets are currently unused or marked defective and simply act as placeholders for future devices.</p> <p class="redacted-note">This dashboard is intentionally simple: static HTML, no login, meant purely for convenience on the local network rather than as a secured control surface. Device names and room assignments are shown for structure; a few outlets are currently unused or marked defective and simply act as placeholders for future devices.</p>
smarthome/homematic-mqtt/index.njk
+1 -1
View File
@@ -18,7 +18,7 @@ deliberately low-tech bridge: <strong>Home Assistant → MQTT → InfluxDB</stro
Docker container, but the underlying Python library lost compatibility Docker container, but the underlying Python library lost compatibility
with HomematicIP's cloud API and hasn't been maintained since 2022. Home with HomematicIP's cloud API and hasn't been maintained since 2022. Home
Assistant, on the other hand, ships an actively maintained HomematicIP Assistant, on the other hand, ships an actively maintained HomematicIP
Cloud integration — so instead of chasing a broken exporter, the bridge Cloud integration, so instead of chasing a broken exporter, the bridge
now runs as native Home Assistant automations that simply republish now runs as native Home Assistant automations that simply republish
sensor state changes to MQTT.</p> sensor state changes to MQTT.</p>
smarthome/index.njk
+2 -2
View File
@@ -1,5 +1,5 @@
--- ---
title: "Smart Home Projects — Home Assistant, MQTT & IoT | moonweb" title: "Smart Home Projects, Home Assistant, MQTT & IoT | moonweb"
section: "smarthome" section: "smarthome"
tags: "smarthome" tags: "smarthome"
description: "Smart home projects: Home Assistant, MQTT sensors, AirPlay audio, OctoPrint, and energy monitoring." description: "Smart home projects: Home Assistant, MQTT sensors, AirPlay audio, OctoPrint, and energy monitoring."
@@ -89,7 +89,7 @@ sections:
<p>The home dashboard runs on multiple touch displays around the flat: an ESP32-S3 WT32-SC01 in the living room and an M5Stack in the study. Both show room temperatures, energy usage, and quick-action buttons for lighting and music control.</p> <p>The home dashboard runs on multiple touch displays around the flat: an ESP32-S3 WT32-SC01 in the living room and an M5Stack in the study. Both show room temperatures, energy usage, and quick-action buttons for lighting and music control.</p>
<h2>Automation philosophy</h2> <h2>Automation philosophy</h2>
<p>The goal is subtle automation that improves daily life without being intrusive. Lights turn on when someone enters a room, music starts playing in the bathroom when the light switches on, and LED matrix displays show the time, weather, and laundry status. But nothing talks unless spoken to — no voice assistants, no always-on microphones, just simple sensor-driven automation that runs locally without cloud dependencies.</p> <p>The goal is subtle automation that improves daily life without being intrusive. Lights turn on when someone enters a room, music starts playing in the bathroom when the light switches on, and LED matrix displays show the time, weather, and laundry status. But nothing talks unless spoken to, no voice assistants, no always-on microphones, just simple sensor-driven automation that runs locally without cloud dependencies.</p>
<h2>Self-hosted services</h2> <h2>Self-hosted services</h2>
<p>Beyond hardware automation, several self-hosted services run on the Synology NAS: TubeArchivist and TubeSync for YouTube archiving, OctoPrint for remote 3D printer control, iCloud Contacts Sync for keeping the address book up to date, and a Google Calendar sync for dashboard widgets. Each service is containerized and backed up as part of the NAS backup strategy.</p> <p>Beyond hardware automation, several self-hosted services run on the Synology NAS: TubeArchivist and TubeSync for YouTube archiving, OctoPrint for remote 3D printer control, iCloud Contacts Sync for keeping the address book up to date, and a Google Calendar sync for dashboard widgets. Each service is containerized and backed up as part of the NAS backup strategy.</p>
smarthome/kids-rfid-player/index.njk
+14 -14
View File
@@ -9,43 +9,43 @@ layout: base.njk
<h1>Kids RFID MP3 Player</h1> <h1>Kids RFID MP3 Player</h1>
<div class="detail-content"> <div class="detail-content">
<p>An Arduino-based MP3 player where kids select audio content by holding RFID cards against the device. No screen, no menus, no WiFi — just tap a card and music plays. Built together with my son in 2018 as a cheaper, more flexible alternative to commercial products like the Hoerbert or Tonuino.</p> <p>An Arduino-based MP3 player where kids select audio content by holding RFID cards against the device. No screen, no menus, no WiFi, just tap a card and music plays. Built together with my son in 2018 as a cheaper, more flexible alternative to commercial products like the Hoerbert or Tonuino.</p>
<h2>Why this exists</h2> <h2>Why this exists</h2>
<p>Commercial kids' audio players have trade-offs. The Hoerbert only offers 9 direct-select buttons with color-coded cards that aren't very intuitive for very young children. The Tonuino uses expensive RFID figures and relies on WiFi streaming, which introduces latency and dependency on network availability. This project aimed for the best of both worlds: RFID-based content selection with local-only playback, analog volume control, and minimal buttons.</p> <p>Commercial kids' audio players have trade-offs. The Hoerbert only offers 9 direct-select buttons with color-coded cards that aren't very intuitive for very young children. The Tonuino uses expensive RFID figures and relies on WiFi streaming, which introduces latency and dependency on network availability. This project aimed for the best of both worlds: RFID-based content selection with local-only playback, analog volume control, and minimal buttons.</p>
<h2>Hardware</h2> <h2>Hardware</h2>
<ul> <ul>
<li><strong>Arduino</strong> — the microcontroller running the custom firmware.</li> <li><strong>Arduino</strong>, the microcontroller running the custom firmware.</li>
<li><strong>RFID module</strong> — reads MIFARE cards to trigger audio playback.</li> <li><strong>RFID module</strong>, reads MIFARE cards to trigger audio playback.</li>
<li><strong>DFPlayer Mini</strong> — compact MP3 player module with built-in amplifier, handling audio decoding and playback directly from an SD card.</li> <li><strong>DFPlayer Mini</strong>, compact MP3 player module with built-in amplifier, handling audio decoding and playback directly from an SD card.</li>
<li><strong>Pololu power switch</strong> — handles auto-shutdown to prevent battery drain, replacing a manual toggle switch that a child would forget to turn off.</li> <li><strong>Pololu power switch</strong>, handles auto-shutdown to prevent battery drain, replacing a manual toggle switch that a child would forget to turn off.</li>
<li><strong>Analog volume knob</strong> — a rotary potentiometer for tactile volume control, deliberately chosen over digital buttons.</li> <li><strong>Analog volume knob</strong>, a rotary potentiometer for tactile volume control, deliberately chosen over digital buttons.</li>
<li><strong>3 LEDs</strong> — green (playing), yellow (paused), red (stopped/at end).</li> <li><strong>3 LEDs</strong>, green (playing), yellow (paused), red (stopped/at end).</li>
</ul> </ul>
<h2>How it works</h2> <h2>How it works</h2>
<p>Each RFID card is mapped to a specific audio file on the SD card. Hold a card against the reader and the corresponding story or song starts playing. Hold the RFID keychain against the reader to pause; hold it again to resume. No other buttons exist on the device — the design intentionally minimizes interaction points so even a toddler can use it independently.</p> <p>Each RFID card is mapped to a specific audio file on the SD card. Hold a card against the reader and the corresponding story or song starts playing. Hold the RFID keychain against the reader to pause; hold it again to resume. No other buttons exist on the device, the design intentionally minimizes interaction points so even a toddler can use it independently.</p>
<h2>Auto-shutdown behavior</h2> <h2>Auto-shutdown behavior</h2>
<p>The power management distinguishes between two states to avoid unnecessary battery drain:</p> <p>The power management distinguishes between two states to avoid unnecessary battery drain:</p>
<ul> <ul>
<li><strong>Stopped</strong> (playback reached the end) — the device shuts down after 5 minutes.</li> <li><strong>Stopped</strong> (playback reached the end), the device shuts down after 5 minutes.</li>
<li><strong>Paused</strong> (user paused via RFID) — the device stays on for 1 hour before shutting down, allowing easy resumption.</li> <li><strong>Paused</strong> (user paused via RFID), the device stays on for 1 hour before shutting down, allowing easy resumption.</li>
</ul> </ul>
<h2>The case</h2> <h2>The case</h2>
<p>The enclosure is a custom-built retro-style wooden box, painted by hand. The name "Rocky Box" came from my son, who painted a Paw Patrol character on the front. The design is deliberately quirky and homemade rather than polished — it has character.</p> <p>The enclosure is a custom-built retro-style wooden box, painted by hand. The name "Rocky Box" came from my son, who painted a Paw Patrol character on the front. The design is deliberately quirky and homemade rather than polished, it has character.</p>
<h2>What's missing (by design)</h2> <h2>What's missing (by design)</h2>
<ul> <ul>
<li>No resume after power-off — turning the device off mid-playback means starting the story over from the beginning.</li> <li>No resume after power-off, turning the device off mid-playback means starting the story over from the beginning.</li>
<li>No skip forward/backward within a track — stories play start to finish.</li> <li>No skip forward/backward within a track, stories play start to finish.</li>
</ul> </ul>
<p>Both omissions are intentional. Young children don't need seeking functionality, and keeping the interface to just "tap card, listen" reduces confusion.</p> <p>Both omissions are intentional. Young children don't need seeking functionality, and keeping the interface to just "tap card, listen" reduces confusion.</p>
<h2>Setup</h2> <h2>Setup</h2>
<p>All 100 RFID cards were pre-programmed once and labeled with numbers. No card learning or reprogramming is needed after initial setup — just hand a child a numbered card and they know which story it triggers.</p> <p>All 100 RFID cards were pre-programmed once and labeled with numbers. No card learning or reprogramming is needed after initial setup, just hand a child a numbered card and they know which story it triggers.</p>
<h2>Code</h2> <h2>Code</h2>
<p>The Arduino firmware is not yet published on GitHub. It's a custom implementation (not Tonuino), written in C++ for the Arduino platform. The code may be uploaded in the future.</p> <p>The Arduino firmware is not yet published on GitHub. It's a custom implementation (not Tonuino), written in C++ for the Arduino platform. The code may be uploaded in the future.</p>
smarthome/octoprint/index.njk
+4 -4
View File
@@ -12,12 +12,12 @@ layout: base.njk
<p>A Raspberry Pi 3B runs OctoPrint for remote 3D-printer control, plus a <p>A Raspberry Pi 3B runs OctoPrint for remote 3D-printer control, plus a
live camera stream so print progress can be checked without walking over live camera stream so print progress can be checked without walking over
to the printer. This eliminates the need to stand next to the printer to the printer. This eliminates the need to stand next to the printer
watching the first layer — a quality-of-life improvement that's hard to watching the first layer, a quality-of-life improvement that's hard to
go back from once you've experienced it.</p> go back from once you've experienced it.</p>
<h2>Current setup</h2> <h2>Current setup</h2>
<ul> <ul>
<li><strong>OctoPrint</strong> runs as a Docker container — the main remote control interface for the printer.</li> <li><strong>OctoPrint</strong> runs as a Docker container, the main remote control interface for the printer.</li>
<li><strong>Camera streaming</strong> is handled by a lightweight camera-server project, exposing both an HLS stream and an MJPEG stream, the latter wired directly into OctoPrint's webcam panel.</li> <li><strong>Camera streaming</strong> is handled by a lightweight camera-server project, exposing both an HLS stream and an MJPEG stream, the latter wired directly into OctoPrint's webcam panel.</li>
<li><strong>Container management</strong> (Portainer, cAdvisor, a node exporter) runs alongside, giving the same monitoring pattern as the rest of the fleet.</li> <li><strong>Container management</strong> (Portainer, cAdvisor, a node exporter) runs alongside, giving the same monitoring pattern as the rest of the fleet.</li>
</ul> </ul>
@@ -27,7 +27,7 @@ go back from once you've experienced it.</p>
camera-server project provides two streaming modes: MJPEG for the camera-server project provides two streaming modes: MJPEG for the
OctoPrint dashboard (low latency, works in any browser) and HLS for OctoPrint dashboard (low latency, works in any browser) and HLS for
remote viewing over slower connections. The MJPEG stream is the primary remote viewing over slower connections. The MJPEG stream is the primary
way to check on prints — it loads quickly and updates in real time in the way to check on prints, it loads quickly and updates in real time in the
OctoPrint web interface.</p> OctoPrint web interface.</p>
<h2>History</h2> <h2>History</h2>
@@ -39,7 +39,7 @@ that point to better reflect its now-singular focus on the printer.</p>
<h2>Notable OS lesson</h2> <h2>Notable OS lesson</h2>
<p>After moving this Pi to a newer OS release, boot configuration files <p>After moving this Pi to a newer OS release, boot configuration files
moved to a new path — editing the old path silently does nothing, which moved to a new path, editing the old path silently does nothing, which
is an easy trap when following older notes or tutorials for the same is an easy trap when following older notes or tutorials for the same
hardware. The lesson: always verify that configuration changes actually hardware. The lesson: always verify that configuration changes actually
take effect, especially after OS upgrades on embedded hardware.</p> take effect, especially after OS upgrades on embedded hardware.</p>
smarthome/tasmota-energy/index.njk
+3 -3
View File
@@ -27,7 +27,7 @@ and LED matrix displays.</p>
<h2>Multi-socket power strips</h2> <h2>Multi-socket power strips</h2>
<p>Several rooms use multi-outlet smart power strips (each socket <p>Several rooms use multi-outlet smart power strips (each socket
individually switchable and metered) rather than single smart plugs — this individually switchable and metered) rather than single smart plugs, this
keeps things like "washing machine + dryer" or "office desk cluster" on keeps things like "washing machine + dryer" or "office desk cluster" on
one strip while still tracking each socket's consumption separately.</p> one strip while still tracking each socket's consumption separately.</p>
@@ -35,7 +35,7 @@ one strip while still tracking each socket's consumption separately.</p>
<p>New devices go through a standard conversion flow (from the <p>New devices go through a standard conversion flow (from the
manufacturer's original cloud-dependent firmware to Tasmota) using a manufacturer's original cloud-dependent firmware to Tasmota) using a
well-documented community process. It occasionally fails on the first well-documented community process. It occasionally fails on the first
attempt due to a transient Wi-Fi handshake issue — retrying resolves it attempt due to a transient Wi-Fi handshake issue, retrying resolves it
without any special handling.</p> without any special handling.</p>
<h2>How the data flows</h2> <h2>How the data flows</h2>
@@ -60,7 +60,7 @@ via Flux to display real-time and historical usage charts.</p>
<h2>Why Tasmota over alternatives</h2> <h2>Why Tasmota over alternatives</h2>
<p>Tasmota was chosen over alternatives like ESPHome or Tuya firmware for <p>Tasmota was chosen over alternatives like ESPHome or Tuya firmware for
several reasons. The local-only control means no cloud dependency — the several reasons. The local-only control means no cloud dependency, the
plugs work even if the internet is down. The MQTT integration is mature plugs work even if the internet is down. The MQTT integration is mature
and well-documented. The energy monitoring sensors are accurate enough and well-documented. The energy monitoring sensors are accurate enough
for home use (typically within 5% of a dedicated energy meter). And the for home use (typically within 5% of a dedicated energy meter). And the