Keep data in host folders instead of Docker named volumes

The database and files now live under ./data (DATA_DIR to move them), so
they can be seen, backed up and restored without going through Docker.
tools/volumes-to-folders.sh copies an existing install's named volumes
across; it only copies, leaving the volumes in place.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

README.md

@@ -57,6 +57,24 @@ A self-hosted web app for organizing your 3D-printing files (STL, 3MF, OBJ, FBX,
 
 Run the unit tests with `npm test`.
 
+### Where your data lives
+
+With Docker, everything sits in folders next to `docker-compose.yml` (set `DATA_DIR` to put them elsewhere):
+
+| Folder | Holds |
+|---|---|
+| `data/pgdata` | The database: models, tags, collections, watches, spools, printers, settings (tokens and printer access codes live here) |
+| `data/uploads` | Your model files, as uploaded or downloaded |
+| `data/thumbnails` | Generated pictures, including plate images |
+| `data/glb` | Converted 3D previews for the viewer |
+| `data/extracted` | Files unpacked from archives — a cache |
+
+A backup needs `data/pgdata` (or a `pg_dump`) and `data/uploads`; the other three rebuild themselves. Everything else in
+the container is replaced on each update, so nothing outside these folders is kept.
+
+Upgrading from a version that used Docker named volumes: stop the stack, run `tools/volumes-to-folders.sh` to copy the
+data across, then start it again. The script only copies — the old volumes stay until you remove them yourself.
+
 ## Configuration
 
 Settings live in the database (editable in the app's **Settings** menu) and fall back to environment variables:

docker-compose.yml

@@ -6,7 +6,7 @@ services:
       POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-printhive}
       POSTGRES_DB: printhive
     volumes:
-      - pgdata:/var/lib/postgresql/data
+      - ${DATA_DIR:-./data}/pgdata:/var/lib/postgresql/data
     healthcheck:
       test: ["CMD-SHELL", "pg_isready -U printhive -d printhive"]
       interval: 5s
@@ -33,10 +33,12 @@ services:
       MAKERWORLD_TOKEN: ${MAKERWORLD_TOKEN:-}
       PRINTABLES_TOKEN: ${PRINTABLES_TOKEN:-}
     volumes:
-      - uploads:/app/uploads
-      - thumbnails:/app/thumbnails
-      - glb:/app/glb
-      - extracted:/app/extracted   # cache of files unpacked from archives
+      # Folders on the host, under ./data (set DATA_DIR to put them elsewhere).
+      # uploads and pgdata hold everything that matters; the other two rebuild.
+      - ${DATA_DIR:-./data}/uploads:/app/uploads
+      - ${DATA_DIR:-./data}/thumbnails:/app/thumbnails
+      - ${DATA_DIR:-./data}/glb:/app/glb
+      - ${DATA_DIR:-./data}/extracted:/app/extracted   # cache of files unpacked from archives
     # No host port published — Traefik auto-discovers the app on the shared network.
     networks:
       - default     # talk to the db
@@ -52,13 +54,6 @@ services:
       - "traefik.http.routers.printhive.tls.certresolver=${TRAEFIK_CERTRESOLVER:-le}"
       - "traefik.http.services.printhive.loadbalancer.server.port=3000"
 
-volumes:
-  pgdata:
-  uploads:
-  thumbnails:
-  glb:
-  extracted:
-
 networks:
   default:
   # The existing Traefik network (the `proxy` network Traefik runs on).

tools/volumes-to-folders.sh

@@ -0,0 +1,53 @@
+#!/usr/bin/env bash
+# Copies Printhive's data out of Docker's named volumes into ./data, for the
+# switch to host folders. Run it on the server, from the folder holding
+# docker-compose.yml, with the stack stopped.
+#
+# It only copies: the named volumes are left exactly as they were, so if
+# anything looks wrong you can put the old docker-compose.yml back and start
+# again with nothing lost.
+set -euo pipefail
+
+DATA_DIR="${DATA_DIR:-./data}"
+VOLUMES=(pgdata uploads thumbnails glb extracted)
+
+# Compose names volumes "<project>_<name>", the project being this folder's name
+# unless it's set elsewhere.
+PROJECT="${COMPOSE_PROJECT_NAME:-$(basename "$(pwd)")}"
+
+if docker compose ps --status running --quiet 2>/dev/null | grep -q .; then
+  echo "Printhive is still running. Stop it first:  docker compose down" >&2
+  exit 1
+fi
+
+echo "Project: $PROJECT"
+echo "Copying into: $DATA_DIR"
+echo
+
+for name in "${VOLUMES[@]}"; do
+  volume="${PROJECT}_${name}"
+  target="$DATA_DIR/$name"
+
+  if ! docker volume inspect "$volume" >/dev/null 2>&1; then
+    echo "· $volume — no such volume, skipping (nothing to copy)"
+    mkdir -p "$target"
+    continue
+  fi
+  if [ -d "$target" ] && [ -n "$(ls -A "$target" 2>/dev/null)" ]; then
+    echo "· $target already has files in it — skipping, so nothing is overwritten" >&2
+    continue
+  fi
+
+  mkdir -p "$target"
+  # -a keeps ownership and timestamps, which Postgres's data folder needs.
+  docker run --rm \
+    -v "$volume":/from:ro \
+    -v "$(cd "$(dirname "$target")" && pwd)/$(basename "$target")":/to \
+    alpine sh -c 'cp -a /from/. /to/'
+  echo "· $volume → $target ($(du -sh "$target" | cut -f1))"
+done
+
+echo
+echo "Done. Start it up again with:  docker compose up -d"
+echo "Once you're happy, the old volumes can be removed with:"
+for name in "${VOLUMES[@]}"; do echo "  docker volume rm ${PROJECT}_${name}"; done