Move the deploy router into the compose- domain prefix to match the rest of the local stack lifecycle (compose-up/down/exec/...). Reorder its block alphabetically between compose-down and compose-entity-purge to keep `make help` and the alphabetical-targets lint happy. Update every active call site: contributor docs, agent-iteration docs, role-level Administration/README/TODO files, the bundles README, the test-environment scripts, the in-repo deploy helpers and main.sh, and the server stage task. The two wildcard prose references (`make deploy-*` in CLAUDE.md and testing/playwright.md) become single `make compose-deploy` mentions since no other deploy-* target exists. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
4.9 KiB
Manage Low-Hardware Resources
Use this guide when you work on a machine with limited CPU, RAM, or disk.
Goal Hardware
A practical long-term target is:
- 64 GB RAM
- a swapfile about the same size as RAM
- an external SSD with more than 500 GB if your computer does not have sufficient storage space
- enough room for Docker images, volumes, build cache, and local repositories
If disk space is tight, you SHOULD move Docker data, caches, and your working directory to the SSD.
On Linux, you MAY move Docker to a different data path:
{
"data-root": "/mnt/ssd/docker"
}
Then restart Docker:
sudo systemctl restart docker
The mount MUST be available before Docker starts. Otherwise the daemon may fall back to the system drive or fail to start cleanly.
On macOS and Windows, you MUST use Docker Desktop settings instead of daemon.json.
A same-size swapfile helps absorb memory spikes on Linux. On Windows and macOS, you SHOULD keep enough free disk space available for virtual memory and local caches.
Load Only What You Need
For a broad stack like the Community Hub, you SHOULD start only the services you actually need. If you are working on Discourse, you do not need Mastodon, Pixelfed, PeerTube, or Friendica in the same session.
You MAY set the INFINITO_SERVICES_DISABLED environment variable before creating the inventory to disable services automatically across all applications without editing any file:
make compose-deploy mode=reinstall apps=web-app-discourse disabled=matomo
This sets enabled: false and shared: false for every listed service in the generated inventory. See variables.md for details.
| Service | Optional | What it provides | Effect of disabling | Safe to disable when |
|---|---|---|---|---|
matomo |
🟢 | Analytics tracking | No usage statistics collected. Usually no functional impact. | You are not testing analytics integration |
oidc |
🟠 | Single sign-on via Keycloak | App falls back to local login | You are not testing SSO/OIDC flows |
ldap |
🟠 | Central user directory via OpenLDAP | App uses its own local user store | You are not testing LDAP/user sync |
css |
🟠 | Custom theming/branding stylesheet | App uses its default upstream theme | You are not testing visual customization |
logout |
🟠 | Shared logout endpoint across apps | Single sign-out does not propagate | You are not testing cross-app logout |
dashboard |
🟠 | Central navigation hub | App is not reachable via the dashboard | You access the app directly by URL |
redis |
🔴 | In-memory cache and session store | Caching and queuing are disabled | The app does not require sessions or queues (rarely safe) |
database |
🔴 | Shared relational database (MariaDB/Postgres) | App cannot persist data | Never disable. Required by almost every app. |
Legend:
- 🟢 Safe to disable. Usually no functional impact.
- 🟠 Optional. Can be disabled, but reduces functionality.
- 🔴 Required. MUST NOT be disabled.
This is a development profile, not a production target.
When running Playwright tests, you SHOULD only disable matomo. All other services (🟠) are REQUIRED for full end-to-end scenario coverage. Disabling them will cause Playwright tests that depend on SSO, LDAP, theming, or logout flows to fail or produce incomplete results.
Test Smarter
On small machines, you SHOULD limit validation to the role you are touching.
For Discourse, start with:
make compose-deploy mode=reinstall apps=web-app-discourse disabled=matomo
If the local inventory and stack already exist, you SHOULD reuse them:
make compose-deploy mode=update apps=web-app-discourse disabled=matomo
You SHOULD use make compose-deploy (full discovery) only when you need broad coverage and have enough time and resources.
Measure Before You Delete
You SHOULD check what is actually consuming space before cleaning up:
make diagnose-disk-usage
That makes it easier to see whether the real issue is Docker, journald, a package cache, or project state.
Cleanup
Run the full cleanup pass to free disk and memory:
make system-purge
On WSL2 or Windows, you MUST pass the additional flag to also configure and run the Windows Disk Cleanup profile. Windows manages its own system caches independently from Linux and Docker:
make system-purge PURGE_WINDOWS_CLEANMGR_SETUP=true
Further Information
- Purge guide: Canonical entry points for cleanup.
- Local purge guide: Local deploy cleanup helpers.
- Local reset guide: Local state reset helpers.
- Makefile commands: All available make targets.
Discussion
Discuss this topic in the related forum article.