fix/node-modules-named-volumes (#94)

Co-authored-by: openhands <openhands@all-hands.dev>
Reviewed-on: https://codeberg.org/johba/harb/pulls/94

.dockerignore

@@ -0,0 +1,60 @@
+# Node.js dependencies (should be in named volumes, not copied to build context)
+**/node_modules/
+node_modules/
+
+# Build outputs
+**/dist/
+**/build/
+**/.next/
+**/.nuxt/
+
+# Caches
+**/.cache/
+**/.vite/
+**/.ponder/
+**/.turbo/
+
+# Logs
+*.log
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+pnpm-debug.log*
+
+# Environment files (should be generated in containers)
+**/.env.local
+**/.env.*.local
+
+# Testing
+**/coverage/
+**/.nyc_output/
+
+# Git
+.git/
+.gitignore
+.gitattributes
+
+# IDE
+.vscode/
+.idea/
+*.swp
+*.swo
+*~
+
+# OS
+.DS_Store
+Thumbs.db
+
+# Docker
+Dockerfile
+docker-compose*.yml
+.dockerignore
+
+# Documentation
+*.md
+!README.md
+
+# Temporary files
+tmp/
+temp/
+*.tmp

AGENTS.md

@@ -50,7 +50,17 @@
 - **Container Orchestration**: `docker-compose.yml` has NO `depends_on` declarations. All service ordering is handled in `scripts/dev.sh` via phased startup with explicit health checks.
 - **Startup Phases**: (1) Create all containers, (2) Start anvil+postgres and wait for healthy, (3) Start bootstrap and wait for completion, (4) Start ponder and wait for healthy, (5) Start webapp/landing/txn-bot, (6) Start caddy.
 - **Logging Configuration**: All services have log rotation configured (max 10MB per file, 3 files max = 30MB per container) to prevent disk bloat. Logs are automatically rotated by Docker.
-- **Disk Management**: `./scripts/dev.sh stop` automatically prunes unused Docker resources. For aggressive cleanup, run `./scripts/cleanup-disk.sh`.
+- **Disk Management** (Portable, No Per-Machine Setup Required):
+  - **20GB Hard Limit**: The stack enforces a 20GB total Docker disk usage limit (images + containers + volumes + build cache).
+  - **Pre-flight Checks**: `./scripts/dev.sh start` checks Docker disk usage before starting and refuses to start if over 20GB.
+  - **Aggressive Auto-Cleanup on Stop**: `./scripts/dev.sh stop` automatically prunes ALL unused Docker resources including build cache (the primary cause of bloat).
+  - **Named Volumes for node_modules**: All Node.js services (ponder, webapp, landing, txnBot) use named Docker volumes for `node_modules/` instead of writing to the host filesystem. This prevents host pollution (20-30GB savings) and ensures `docker system prune --volumes` cleans them up.
+  - **npm Best Practices**: All entrypoints use `npm ci` (not `npm install`) for reproducible builds and `npm cache clean --force` to remove ~50-100MB of cache per service.
+  - **PostgreSQL WAL Limits**: Postgres configured with `wal_level=minimal`, `max_wal_size=128MB`, and `archive_mode=off` to prevent unbounded WAL file growth in the postgres volume.
+  - **Log Rotation**: All containers limited to 30MB logs (10MB × 3 files) via docker-compose logging configuration.
+  - **.dockerignore**: Excludes `node_modules/`, caches, and build outputs from Docker build context to speed up builds and reduce image size.
+  - **Monitoring**: The stack displays current Docker disk usage on startup and warns at 80% (16GB).
+  - **Note**: Docker has no built-in portable disk quotas. All limits are enforced via aggressive pruning, bounded configurations, and isolation of dependencies to Docker volumes.
 
 ## Guardrails & Tips
 - `token0isWeth` flips amount semantics; confirm ordering before seeding or interpreting liquidity.

containers/landing-dev-entrypoint.sh

@@ -31,16 +31,16 @@ if [[ ! -f "$REQUIRED_DIST" ]]; then
 fi
 
 cd "$LANDING_DIR"
-DEPS_MARKER="/tmp/.landing-deps-installed"
+
-if [[ ! -d node_modules || ! -f "$DEPS_MARKER" ]]; then
+# Check if node_modules is populated (named volume may be empty on first run)
+if [[ ! -d node_modules/.bin ]]; then
   echo "[landing-entrypoint] Installing dependencies..."
-  npm install --no-save --loglevel error 2>&1 || {
+  npm ci --loglevel error && npm cache clean --force 2>&1 || {
-    echo "[landing-entrypoint] npm install failed, trying with --force"
+    echo "[landing-entrypoint] npm ci failed, trying npm install"
-    npm install --force --no-save --loglevel error
+    npm install --no-save --loglevel error && npm cache clean --force
   }
-  touch "$DEPS_MARKER" || true
 else
-  echo "[landing-entrypoint] Using cached node_modules"
+  echo "[landing-entrypoint] Using cached node_modules from volume"
 fi
 
 export CHOKIDAR_USEPOLLING=${CHOKIDAR_USEPOLLING:-1}

containers/ponder-dev-entrypoint.sh

@@ -65,16 +65,15 @@ if [[ ! -f "$REQUIRED_DIST" ]]; then
   exit 1
 fi
 
-DEPS_MARKER="/tmp/.ponder-deps-installed"
+# Check if node_modules is populated (named volume may be empty on first run)
-if [[ ! -d node_modules || ! -f "$DEPS_MARKER" ]]; then
+if [[ ! -d node_modules/.bin ]]; then
   echo "[ponder-entrypoint] Installing dependencies..."
-  npm install --no-save --loglevel error 2>&1 || {
+  npm ci --loglevel error && npm cache clean --force 2>&1 || {
-    echo "[ponder-entrypoint] npm install failed, trying with --force"
+    echo "[ponder-entrypoint] npm ci failed, trying npm install"
-    npm install --force --no-save --loglevel error
+    npm install --no-save --loglevel error && npm cache clean --force
   }
-  touch "$DEPS_MARKER" || true
 else
-  echo "[ponder-entrypoint] Using cached node_modules"
+  echo "[ponder-entrypoint] Using cached node_modules from volume"
 fi
 
 # Load and export all environment variables from .env.local

containers/txn-bot-entrypoint.sh

@@ -37,16 +37,16 @@ if [[ ! -f "$REQUIRED_DIST" ]]; then
 fi
 
 cd "$BOT_DIR"
-DEPS_MARKER="/tmp/.txnbot-deps-installed"
+
-if [[ ! -d node_modules || ! -f "$DEPS_MARKER" ]]; then
+# Check if node_modules is populated (named volume may be empty on first run)
+if [[ ! -d node_modules/.bin ]]; then
   echo "[txn-bot-entrypoint] Installing txn-bot dependencies..."
-  npm install --no-save --loglevel error 2>&1 || {
+  npm ci --loglevel error && npm cache clean --force 2>&1 || {
-    echo "[txn-bot-entrypoint] npm install failed, trying with --force"
+    echo "[txn-bot-entrypoint] npm ci failed, trying npm install"
-    npm install --force --no-save --loglevel error
+    npm install --no-save --loglevel error && npm cache clean --force
   }
-  touch "$DEPS_MARKER" || true
 else
-  echo "[txn-bot-entrypoint] Using cached node_modules"
+  echo "[txn-bot-entrypoint] Using cached node_modules from volume"
 fi
 
 echo "[txn-bot-entrypoint] Building TypeScript..."

containers/webapp-dev-entrypoint.sh

@@ -41,16 +41,16 @@ fi
 source "$CONTRACT_ENV"
 
 cd "$APP_DIR"
-DEPS_MARKER="/tmp/.webapp-deps-installed"
+
-if [[ ! -d node_modules || ! -f "$DEPS_MARKER" ]]; then
+# Check if node_modules is populated (named volume may be empty on first run)
+if [[ ! -d node_modules/.bin ]]; then
   echo "[frontend-entrypoint] Installing dependencies..."
-  npm install --no-save --loglevel error 2>&1 || {
+  npm ci --loglevel error && npm cache clean --force 2>&1 || {
-    echo "[frontend-entrypoint] npm install failed, trying with --force"
+    echo "[frontend-entrypoint] npm ci failed, trying npm install"
-    npm install --force --no-save --loglevel error
+    npm install --no-save --loglevel error && npm cache clean --force
   }
-  touch "$DEPS_MARKER" || true
 else
-  echo "[frontend-entrypoint] Using cached node_modules"
+  echo "[frontend-entrypoint] Using cached node_modules from volume"
 fi
 
 export VITE_DEFAULT_CHAIN_ID=${VITE_DEFAULT_CHAIN_ID:-31337}

docker-compose.yml

@@ -34,6 +34,18 @@ services:
 
   postgres:
     image: docker.io/library/postgres:16-alpine
+    command:
+      - "postgres"
+      - "-c"
+      - "wal_level=minimal"
+      - "-c"
+      - "max_wal_size=128MB"
+      - "-c"
+      - "max_wal_senders=0"
+      - "-c"
+      - "archive_mode=off"
+      - "-c"
+      - "checkpoint_timeout=30min"
     environment:
       - POSTGRES_USER=ponder
       - POSTGRES_PASSWORD=ponder_local
@@ -82,6 +94,7 @@ services:
       - .:/workspace:z
       - .git:/workspace/.git:ro,z
       - ./kraiken-lib/dist:/workspace/kraiken-lib/dist:ro,z
+      - ponder_node_modules:/workspace/services/ponder/node_modules
     working_dir: /workspace
     environment:
       - CHOKIDAR_USEPOLLING=1
@@ -112,6 +125,7 @@ services:
       - .:/workspace:z
       - .git:/workspace/.git:ro,z
       - ./kraiken-lib/dist:/workspace/kraiken-lib/dist:ro,z
+      - webapp_node_modules:/workspace/web-app/node_modules
     working_dir: /workspace
     environment:
       - CHOKIDAR_USEPOLLING=1
@@ -140,6 +154,7 @@ services:
       - .:/workspace:z
       - .git:/workspace/.git:ro,z
       - ./kraiken-lib/dist:/workspace/kraiken-lib/dist:ro,z
+      - landing_node_modules:/workspace/landing/node_modules
     working_dir: /workspace
     environment:
       - CHOKIDAR_USEPOLLING=1
@@ -166,6 +181,7 @@ services:
       - .:/workspace:z
       - .git:/workspace/.git:ro,z
       - ./kraiken-lib/dist:/workspace/kraiken-lib/dist:ro,z
+      - txnbot_node_modules:/workspace/services/txnBot/node_modules
     working_dir: /workspace
     environment:
       - GIT_BRANCH=${GIT_BRANCH:-}
@@ -199,3 +215,7 @@ services:
 
 volumes:
   postgres-data:
+  ponder_node_modules:
+  webapp_node_modules:
+  landing_node_modules:
+  txnbot_node_modules:

scripts/cleanup-disk.sh

@@ -1,55 +0,0 @@
-#!/usr/bin/env bash
-set -euo pipefail
-
-# Disk cleanup utility for Harb stack
-# Use this to aggressively free up disk space
-
-cd "$(dirname "$0")/.."
-
-echo "=== Harb Stack Disk Cleanup ==="
-echo ""
-
-# Detect container runtime
-if command -v docker &> /dev/null; then
-  RUNTIME_CMD="docker"
-elif command -v podman &> /dev/null; then
-  RUNTIME_CMD="podman"
-else
-  echo "Error: docker/podman not found"
-  exit 1
-fi
-
-echo "Current disk usage:"
-df -h / | tail -1
-echo ""
-
-# Stop the stack first
-if [[ -f "./scripts/dev.sh" ]]; then
-  echo "Stopping stack..."
-  ./scripts/dev.sh stop || true
-fi
-
-echo ""
-echo "Pruning Docker resources..."
-echo "  - Stopped containers"
-echo "  - Unused volumes"
-echo "  - Dangling images"
-echo "  - Build cache"
-echo ""
-
-# Aggressive pruning
-${RUNTIME_CMD} system prune -af --volumes
-
-echo ""
-echo "Cleaning npm caches..."
-rm -rf ~/.npm ~/.cache 2>/dev/null || true
-
-echo ""
-echo "Cleaning journal logs..."
-sudo journalctl --vacuum-size=500M 2>/dev/null || echo "  (skipped: needs sudo)"
-
-echo ""
-echo "Final disk usage:"
-df -h / | tail -1
-echo ""
-echo "[ok] Cleanup complete"

scripts/dev.sh

@@ -11,6 +11,7 @@ readonly PONDER_TIMEOUT=120        # Must index bootstrap events
 readonly WEBAPP_TIMEOUT=120        # npm install + Vite startup
 readonly CADDY_TIMEOUT=20          # Proxy starts instantly
 readonly POLL_INTERVAL=2           # Check health every N seconds
+readonly MAX_DOCKER_DISK_GB=20     # Maximum Docker disk usage in GB
 
 PID_FILE=/tmp/kraiken-watcher.pid
 PROJECT_NAME=${COMPOSE_PROJECT_NAME:-$(basename "$PWD")}
@@ -36,6 +37,58 @@ container_name() {
   echo "${PROJECT_NAME}_${service}_1"
 }
 
+# Check Docker disk usage and warn if approaching limits
+check_docker_disk_usage() {
+  if ! command -v docker &> /dev/null; then
+    return 0  # Skip if Docker not available
+  fi
+
+  # Get total Docker disk usage in GB (works on Linux and macOS)
+  local total_size_bytes
+  total_size_bytes=$(docker system df --format '{{.Size}}' 2>/dev/null | \
+    sed 's/[^0-9.]//g' | awk '{sum+=$1} END {print sum}' || echo "0")
+
+  # Parse the actual usage more accurately
+  local docker_df_output
+  docker_df_output=$(docker system df 2>/dev/null || echo "")
+
+  if [[ -z "$docker_df_output" ]]; then
+    return 0  # Docker not running
+  fi
+
+  # Extract total reclaimable space (more accurate than parsing Size)
+  local total_gb
+  total_gb=$(echo "$docker_df_output" | tail -n 1 | awk '{print $NF}' | sed 's/GB//; s/MB/\/1024/; s/KB/\/1048576/' | bc -l 2>/dev/null || echo "0")
+
+  # Alternative: sum up all TYPE sizes (column 3 has the SIZE)
+  local images_size containers_size volumes_size build_cache_size
+  images_size=$(echo "$docker_df_output" | grep "Images" | awk '{print $3}' | sed 's/GB$//; s/MB$/\/1024/; s/KB$/\/1048576/; s/B$/\/1073741824/' | sed 's/^$/0/' | bc -l 2>/dev/null || echo "0")
+  containers_size=$(echo "$docker_df_output" | grep "Containers" | awk '{print $3}' | sed 's/GB$//; s/MB$/\/1024/; s/KB$/\/1048576/; s/B$/\/1073741824/' | sed 's/^$/0/' | bc -l 2>/dev/null || echo "0")
+  volumes_size=$(echo "$docker_df_output" | grep "Local Volumes" | awk '{print $3}' | sed 's/GB$//; s/MB$/\/1024/; s/KB$/\/1048576/; s/B$/\/1073741824/' | sed 's/^$/0/' | bc -l 2>/dev/null || echo "0")
+  build_cache_size=$(echo "$docker_df_output" | grep "Build Cache" | awk '{print $3}' | sed 's/GB$//; s/MB$/\/1024/; s/KB$/\/1048576/; s/B$/\/1073741824/' | sed 's/^$/0/' | bc -l 2>/dev/null || echo "0")
+
+  total_gb=$(echo "$images_size + $containers_size + $volumes_size + $build_cache_size" | bc -l 2>/dev/null || echo "0")
+
+  # Round to 1 decimal place
+  total_gb=$(printf "%.1f" "$total_gb" 2>/dev/null || echo "0")
+
+  echo "  Docker disk usage: ${total_gb}GB / ${MAX_DOCKER_DISK_GB}GB limit"
+
+  # Warn if approaching 80% of limit (16GB)
+  if (( $(echo "$total_gb > 16" | bc -l 2>/dev/null || echo "0") )); then
+    echo "  [!!] WARNING: Docker disk usage is high!"
+    echo "  [!!] Run './scripts/cleanup-disk.sh' to free up space"
+  fi
+
+  # Hard stop if over limit
+  if (( $(echo "$total_gb > $MAX_DOCKER_DISK_GB" | bc -l 2>/dev/null || echo "0") )); then
+    echo ""
+    echo "ERROR: Docker disk usage exceeds ${MAX_DOCKER_DISK_GB}GB limit!"
+    echo "Run './scripts/cleanup-disk.sh' to free up space, then try again."
+    exit 1
+  fi
+}
+
 cleanup_existing() {
   # Kill any existing watch scripts
   pkill -f "watch-kraiken-lib.sh" 2>/dev/null || true
@@ -101,6 +154,9 @@ wait_for_exited() {
 start_stack() {
   local stack_start_time=$(date +%s)
 
+  # Check Docker disk usage before starting
+  check_docker_disk_usage
+
   # Clean up any existing processes first
   cleanup_existing
 
@@ -163,9 +219,14 @@ stop_stack() {
   cleanup_existing
   ${COMPOSE_CMD} down
 
-  # Prune Docker resources to prevent disk bloat
+  # Aggressive pruning to prevent disk bloat
-  echo "  Pruning Docker resources..."
+  echo "  Pruning Docker resources (images, containers, volumes, build cache)..."
-  ${RUNTIME_CMD} system prune -f --volumes 2>&1 | grep -E "Total reclaimed|deleted" || true
+
+  # Prune build cache aggressively (this is usually the biggest culprit)
+  ${RUNTIME_CMD} builder prune -af 2>&1 | grep -E "Total|deleted" || true
+
+  # Prune all unused data (containers, networks, images, volumes)
+  ${RUNTIME_CMD} system prune -af --volumes 2>&1 | grep -E "Total reclaimed|deleted" || true
 
   echo "[ok] Stack stopped and cleaned"
 }