dev day #80

DOCS/DOCKER_BUILD_MEMORY_FIX.md

@@ -0,0 +1,165 @@
+# Docker Build Memory Fix Guide
+
+## Problem
+Frontend Docker build fails with "ResourceExhausted: cannot allocate memory" during React/webpack build.
+
+## Applied Fixes
+
+### 1. Dockerfile Optimizations ✅
+**File:** `frontend/Dockerfile`
+
+- Reduced Node memory from 4GB to 2GB (`--max-old-space-size=2048`)
+- Added Node GC optimizations: `--optimize-for-size --max-semi-space-size=1`
+- Set `CI=true` to limit webpack parallelism
+- Added `npm cache clean` before build to free memory
+
+### 2. Docker Compose Updates ✅
+**File:** `docker-compose.yml`
+
+- Increased frontend memory limit: 512M → 1GB
+- Increased CPU limit: 1.0 → 2.0 cores
+- Added `shm_size: 256m` for build stage
+
+## How to Apply
+
+### Method 1: Standard Build (Recommended)
+```bash
+# Clean previous build artifacts
+docker compose down -v
+docker system prune -f
+
+# Rebuild with new settings
+docker compose build frontend --no-cache
+docker compose up -d
+```
+
+### Method 2: If Still Out of Memory
+
+#### Option A: Increase Docker Desktop Memory
+1. Open Docker Desktop Settings
+2. Go to Resources → Advanced
+3. Increase Memory to at least **6GB** (recommended 8GB)
+4. Click "Apply & Restart"
+5. Retry build
+
+#### Option B: Build Outside Docker (Fastest)
+```bash
+cd frontend
+
+# Install dependencies
+npm install
+
+# Build locally
+npm run build
+
+# Then use the pre-built files with Docker
+docker compose up -d
+```
+
+#### Option C: Use Docker BuildKit with More Memory
+```bash
+# Set Docker BuildKit memory limit
+export DOCKER_BUILDKIT=1
+export BUILDKIT_STEP_LOG_MAX_SIZE=50000000
+
+# Build with explicit memory limit
+docker buildx build \
+  --memory 4g \
+  --memory-swap 6g \
+  -t myclub-frontend:latest \
+  ./frontend
+```
+
+## Verification
+
+### Check Build Success
+```bash
+# View build logs
+docker compose logs frontend
+
+# Verify container is running
+docker compose ps
+
+# Test frontend access
+curl http://localhost:3000
+```
+
+### Monitor Memory During Build
+```bash
+# In another terminal, watch Docker stats during build
+docker stats --no-stream
+```
+
+## Troubleshooting
+
+### Error: "Still running out of memory"
+**Solutions:**
+1. **Close other applications** to free system RAM
+2. **Increase Docker Desktop memory** to 8GB
+3. **Use local build** (Option B above)
+4. **Enable swap memory** on your system
+
+### Error: "webpack: Compilation failed"
+**Solutions:**
+1. Check `frontend/package.json` dependencies
+2. Clear npm cache: `npm cache clean --force`
+3. Delete `node_modules` and reinstall: `rm -rf node_modules && npm install`
+
+### Error: "Cannot find ESLint plugin"
+This is **expected** - ESLint is disabled during build with `DISABLE_ESLINT_PLUGIN=true` to save memory.
+
+## Performance Tips
+
+### Speed Up Rebuilds
+```bash
+# Use Docker build cache
+docker compose build frontend
+
+# Or parallel builds
+docker compose build --parallel
+```
+
+### Monitor Build Progress
+```bash
+# Build with verbose output
+docker compose build frontend --progress=plain
+```
+
+## System Requirements
+
+### Minimum for Docker Build
+- **RAM:** 6GB available
+- **CPU:** 2 cores
+- **Disk:** 5GB free space
+
+### Recommended
+- **RAM:** 8GB+ available
+- **CPU:** 4 cores
+- **Disk:** 10GB+ free space
+- **SSD:** For faster builds
+
+## Alternative: Pre-built Images
+
+If memory is consistently an issue, consider:
+
+1. **Build on CI/CD** (GitHub Actions, GitLab CI)
+2. **Use pre-built images** from registry
+3. **Build on more powerful machine** and export image
+
+```bash
+# Export built image
+docker save myclub-frontend:latest | gzip > frontend-image.tar.gz
+
+# Import on target machine
+docker load < frontend-image.tar.gz
+```
+
+## Summary
+
+The applied fixes optimize memory usage during build:
+- **Reduced memory footprint** from 4GB to 2GB
+- **Limited parallel processing** to prevent memory spikes
+- **Cleaned cache** before build
+- **Increased Docker resources** for build stage
+
+Try the standard build first. If it still fails, use Option A (increase Docker memory) or Option B (build locally).