feat: implement comprehensive backup and restore system with S3 support

- Add S3/MinIO storage adapter with multipart upload support
- Implement database backup service for SQLite and PostgreSQL
- Create backup manifest generator for tracking backup contents
- Enhance backup service with S3 integration and incremental backups
- Add restore service with safety measures and rollback capability
- Create comprehensive test suite for all backup functionality
- Add admin API endpoints for backup/restore management
- Implement frontend UI with dashboard, configuration, and restore wizard
- Add roadmap section to README with implemented backup feature

This implementation provides:
- Multiple backup destinations (local, rsync, S3/MinIO)
- Intelligent change detection to minimize backup frequency
- Full database backups with compression
- Manifest-based restore with integrity validation
- Pre-restore safety backups with rollback
- Comprehensive error handling and monitoring
- User-friendly admin interface

🤖 Generated with Claude Code

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
2025-07-22 09:05:52 +02:00
parent 3c6837bd90
commit f6a79c815e
44 changed files with 15839 additions and 14 deletions
@@ -0,0 +1,251 @@
const { DatabaseBackupService } = require('../databaseBackup');
const { db } = require('../../database/db');
const fs = require('fs').promises;
const path = require('path');
const crypto = require('crypto');
// Mock dependencies
jest.mock('../../database/db');
jest.mock('../../utils/logger');
jest.mock('../emailProcessor');
jest.mock('child_process');
describe('DatabaseBackupService', () => {
let service;
let mockExecAsync;
beforeEach(() => {
service = new DatabaseBackupService();
mockExecAsync = jest.fn();
// Reset mocks
jest.clearAllMocks();
// Mock execAsync
const childProcess = require('child_process');
childProcess.exec = jest.fn((cmd, opts, callback) => {
if (callback) {
callback(null, { stdout: 'ok', stderr: '' });
}
});
});
afterEach(async () => {
// Cleanup test files
try {
await fs.rmdir('/tmp/test-backup', { recursive: true });
} catch (e) {
// Ignore
}
});
describe('calculateChecksum', () => {
it('should calculate SHA256 checksum of a file', async () => {
const testFile = '/tmp/test-checksum.txt';
const testContent = 'Hello, World!';
await fs.writeFile(testFile, testContent);
const checksum = await service.calculateChecksum(testFile);
// Expected checksum for "Hello, World!"
const expectedChecksum = crypto
.createHash('sha256')
.update(testContent)
.digest('hex');
expect(checksum).toBe(expectedChecksum);
await fs.unlink(testFile);
});
});
describe('getTableChecksums', () => {
it('should get checksums for all tables', async () => {
// Mock getTables
service.getTables = jest.fn().mockResolvedValue(['events', 'photos']);
// Mock SQLite response
db.raw = jest.fn()
.mockResolvedValueOnce([{ row_count: 10, data_sum: 1000 }])
.mockResolvedValueOnce([{ row_count: 20, data_sum: 2000 }]);
const checksums = await service.getTableChecksums();
expect(checksums).toHaveProperty('events');
expect(checksums).toHaveProperty('photos');
expect(checksums.events.rowCount).toBe(10);
expect(checksums.photos.rowCount).toBe(20);
expect(checksums.events.checksum).toBeDefined();
expect(checksums.photos.checksum).toBeDefined();
});
});
describe('getTables', () => {
it('should get list of tables for SQLite', async () => {
service.dbType = 'sqlite';
db.raw = jest.fn().mockResolvedValue([
{ name: 'events' },
{ name: 'photos' },
{ name: 'admin_users' }
]);
const tables = await service.getTables();
expect(tables).toEqual(['events', 'photos', 'admin_users']);
expect(db.raw).toHaveBeenCalledWith(expect.stringContaining('sqlite_master'));
});
it('should get list of tables for PostgreSQL', async () => {
service.dbType = 'postgresql';
db.raw = jest.fn().mockResolvedValue({
rows: [
{ table_name: 'events' },
{ table_name: 'photos' },
{ table_name: 'admin_users' }
]
});
const tables = await service.getTables();
expect(tables).toEqual(['events', 'photos', 'admin_users']);
expect(db.raw).toHaveBeenCalledWith(expect.stringContaining('information_schema.tables'));
});
});
describe('getDatabaseSize', () => {
it('should get database size for SQLite', async () => {
service.dbType = 'sqlite';
const mockSize = 1024 * 1024 * 10; // 10MB
// Mock fs.stat
const originalStat = fs.stat;
fs.stat = jest.fn().mockResolvedValue({ size: mockSize });
const size = await service.getDatabaseSize();
expect(size).toBe(mockSize);
fs.stat = originalStat;
});
it('should get database size for PostgreSQL', async () => {
service.dbType = 'postgresql';
const mockSize = 1024 * 1024 * 100; // 100MB
db.raw = jest.fn().mockResolvedValue({
rows: [{ size: mockSize.toString() }]
});
const size = await service.getDatabaseSize();
expect(size).toBe(mockSize);
expect(db.raw).toHaveBeenCalledWith(expect.stringContaining('pg_database_size'));
});
});
describe('compressFile', () => {
it('should compress file and return stats', async () => {
const testFile = '/tmp/test-compress.txt';
const compressedFile = '/tmp/test-compress.txt.gz';
// Create test file with repetitive content (compresses well)
const testContent = 'Hello, World! '.repeat(1000);
await fs.writeFile(testFile, testContent);
const stats = await service.compressFile(testFile, compressedFile);
expect(stats.originalSize).toBeGreaterThan(0);
expect(stats.compressedSize).toBeGreaterThan(0);
expect(stats.compressedSize).toBeLessThan(stats.originalSize);
expect(parseFloat(stats.compressionRatio)).toBeGreaterThan(0);
// Cleanup
await fs.unlink(testFile);
await fs.unlink(compressedFile);
});
});
describe('backup configuration', () => {
it('should get backup configuration from database', async () => {
const mockConfig = [
{ setting_key: 'database_backup_enabled', setting_value: 'true' },
{ setting_key: 'database_backup_compress', setting_value: 'true' },
{ setting_key: 'database_backup_retention_days', setting_value: '30' }
];
db.mockReturnValue({
where: jest.fn().mockReturnThis(),
select: jest.fn().mockResolvedValue(mockConfig)
});
const config = await service.getBackupConfig();
expect(config.database_backup_enabled).toBe(true);
expect(config.database_backup_compress).toBe(true);
expect(config.database_backup_retention_days).toBe(30);
});
});
describe('cleanupOldBackups', () => {
it('should delete old backup files and records', async () => {
const oldBackups = [
{ id: 1, file_path: '/backup/old1.sql.gz' },
{ id: 2, file_path: '/backup/old2.sql.gz' }
];
db.mockReturnValue({
where: jest.fn().mockReturnThis(),
select: jest.fn().mockResolvedValue(oldBackups),
delete: jest.fn().mockResolvedValue(1)
});
// Mock fs.unlink
fs.unlink = jest.fn().mockResolvedValue(undefined);
await service.cleanupOldBackups(30);
expect(fs.unlink).toHaveBeenCalledTimes(2);
expect(fs.unlink).toHaveBeenCalledWith('/backup/old1.sql.gz');
expect(fs.unlink).toHaveBeenCalledWith('/backup/old2.sql.gz');
});
});
describe('progress tracking', () => {
it('should update and retrieve progress', () => {
expect(service.getProgress()).toBeNull();
service.updateProgress('Testing...', { step: 1 });
const progress = service.getProgress();
expect(progress.message).toBe('Testing...');
expect(progress.details.step).toBe(1);
expect(progress.timestamp).toBeDefined();
});
});
describe('backup history', () => {
it('should retrieve backup history', async () => {
const mockHistory = [
{
id: 1,
started_at: new Date(),
completed_at: new Date(),
status: 'completed',
file_size_bytes: 1024000
}
];
db.mockReturnValue({
orderBy: jest.fn().mockReturnThis(),
limit: jest.fn().mockResolvedValue(mockHistory)
});
const history = await service.getBackupHistory(10);
expect(history).toEqual(mockHistory);
expect(history.length).toBe(1);
});
});
});
+501
View File
@@ -0,0 +1,501 @@
const fs = require('fs').promises;
const path = require('path');
const crypto = require('crypto');
const yaml = require('js-yaml');
const os = require('os');
const { db } = require('../database/db');
const logger = require('../utils/logger');
/**
* Backup Manifest Generator
*
* Generates comprehensive manifests for backups including:
* - Version information (app, node, OS)
* - File listings with metadata and checksums
* - Database information
* - System state at backup time
* - Support for both JSON and YAML formats
* - Incremental backup support with parent references
*/
class BackupManifestGenerator {
constructor() {
this.appVersion = require('../../package.json').version;
this.nodeVersion = process.version;
this.platform = process.platform;
this.osRelease = os.release();
this.hostname = os.hostname();
}
/**
* Generate a comprehensive backup manifest
* @param {Object} options - Manifest generation options
* @param {string} options.backupType - 'full' or 'incremental'
* @param {string} options.backupPath - Path to the backup directory
* @param {Array} options.files - Array of backed up files with metadata
* @param {Object} options.databaseInfo - Database backup information
* @param {string} options.parentBackupId - For incremental backups, reference to parent
* @param {string} options.format - 'json' or 'yaml' (default: 'json')
* @param {Object} options.customMetadata - Additional metadata to include
* @returns {Object} Generated manifest object
*/
async generateManifest(options) {
const {
backupType = 'full',
backupPath,
files = [],
databaseInfo = {},
parentBackupId = null,
format = 'json',
customMetadata = {}
} = options;
const manifest = {
// Manifest metadata
manifest: {
version: '2.0',
created: new Date().toISOString(),
generator: 'PicPeak Backup Manifest Generator',
format: format
},
// Backup information
backup: {
id: this.generateBackupId(),
type: backupType,
timestamp: new Date().toISOString(),
path: backupPath,
parent_backup_id: parentBackupId,
retention_days: customMetadata.retentionDays || 30
},
// System information
system: {
hostname: this.hostname,
platform: this.platform,
os_release: this.osRelease,
architecture: os.arch(),
cpu_count: os.cpus().length,
total_memory: os.totalmem(),
free_memory: os.freemem(),
uptime: os.uptime()
},
// Application information
application: {
name: 'PicPeak',
version: this.appVersion,
node_version: this.nodeVersion,
environment: process.env.NODE_ENV || 'production',
storage_path: process.env.STORAGE_PATH || path.join(__dirname, '../../../storage')
},
// Files information
files: {
count: files.length,
total_size: files.reduce((sum, file) => sum + (file.size || 0), 0),
checksums: await this.generateFileChecksums(files),
manifest: files.map(file => ({
path: file.relativePath || file.path,
size: file.size,
modified: file.modified,
checksum: file.checksum,
type: this.getFileType(file.path),
permissions: file.permissions
}))
},
// Database information
database: {
type: databaseInfo.type || this.getDatabaseType(),
backup_file: databaseInfo.backupFile,
size: databaseInfo.size,
checksum: databaseInfo.checksum,
tables: databaseInfo.tables || {},
row_counts: databaseInfo.rowCounts || {},
schema_version: await this.getSchemaVersion()
},
// Verification information
verification: {
total_checksum: null, // Will be calculated after manifest is complete
file_count_check: files.length,
size_check: files.reduce((sum, file) => sum + (file.size || 0), 0),
integrity_timestamp: new Date().toISOString()
},
// Custom metadata
metadata: {
...customMetadata,
backup_settings: await this.getBackupSettings(),
active_events_count: await this.getActiveEventsCount(),
archived_events_count: await this.getArchivedEventsCount(),
total_photos_count: await this.getTotalPhotosCount()
}
};
// Calculate total checksum of the manifest
manifest.verification.total_checksum = this.calculateManifestChecksum(manifest);
return manifest;
}
/**
* Save manifest to file
* @param {Object} manifest - Manifest object to save
* @param {string} filePath - Path to save the manifest
* @param {string} format - 'json' or 'yaml'
*/
async saveManifest(manifest, filePath, format = 'json') {
try {
let content;
if (format === 'yaml') {
content = yaml.dump(manifest, {
indent: 2,
lineWidth: -1,
noRefs: true,
sortKeys: true
});
} else {
content = JSON.stringify(manifest, null, 2);
}
await fs.writeFile(filePath, content, 'utf8');
logger.info(`Manifest saved to ${filePath} (format: ${format})`);
return filePath;
} catch (error) {
logger.error('Failed to save manifest:', error);
throw error;
}
}
/**
* Load and validate an existing manifest
* @param {string} filePath - Path to the manifest file
* @returns {Object} Loaded and validated manifest
*/
async loadManifest(filePath) {
try {
const content = await fs.readFile(filePath, 'utf8');
let manifest;
// Detect format and parse
if (filePath.endsWith('.yaml') || filePath.endsWith('.yml')) {
manifest = yaml.load(content);
} else {
manifest = JSON.parse(content);
}
// Validate manifest structure
this.validateManifest(manifest);
return manifest;
} catch (error) {
logger.error('Failed to load manifest:', error);
throw error;
}
}
/**
* Validate manifest structure and integrity
* @param {Object} manifest - Manifest to validate
* @throws {Error} If validation fails
*/
validateManifest(manifest) {
// Check required sections
const requiredSections = ['manifest', 'backup', 'system', 'application', 'files', 'database', 'verification'];
for (const section of requiredSections) {
if (!manifest[section]) {
throw new Error(`Missing required section: ${section}`);
}
}
// Validate manifest version
if (!manifest.manifest.version) {
throw new Error('Missing manifest version');
}
// Validate file checksums
if (manifest.files.count !== manifest.files.manifest.length) {
throw new Error('File count mismatch');
}
// Validate total checksum
const calculatedChecksum = this.calculateManifestChecksum(manifest);
if (manifest.verification.total_checksum !== calculatedChecksum) {
throw new Error('Manifest checksum verification failed');
}
logger.info('Manifest validation passed');
return true;
}
/**
* Compare two manifests for incremental backup
* @param {Object} currentManifest - Current backup manifest
* @param {Object} parentManifest - Parent backup manifest
* @returns {Object} Comparison results
*/
compareManifests(currentManifest, parentManifest) {
const comparison = {
added_files: [],
modified_files: [],
deleted_files: [],
unchanged_files: [],
size_difference: 0,
database_changes: {}
};
// Create file maps for easy comparison
const currentFiles = new Map(
currentManifest.files.manifest.map(f => [f.path, f])
);
const parentFiles = new Map(
parentManifest.files.manifest.map(f => [f.path, f])
);
// Find added and modified files
for (const [path, file] of currentFiles) {
const parentFile = parentFiles.get(path);
if (!parentFile) {
comparison.added_files.push(file);
comparison.size_difference += file.size;
} else if (file.checksum !== parentFile.checksum) {
comparison.modified_files.push(file);
comparison.size_difference += file.size - parentFile.size;
} else {
comparison.unchanged_files.push(file);
}
}
// Find deleted files
for (const [path, file] of parentFiles) {
if (!currentFiles.has(path)) {
comparison.deleted_files.push(file);
comparison.size_difference -= file.size;
}
}
// Compare database info
comparison.database_changes = {
size_difference: currentManifest.database.size - parentManifest.database.size,
checksum_changed: currentManifest.database.checksum !== parentManifest.database.checksum,
schema_version_changed: currentManifest.database.schema_version !== parentManifest.database.schema_version
};
return comparison;
}
/**
* Generate incremental manifest based on parent
* @param {Object} options - Manifest generation options
* @param {Object} parentManifest - Parent backup manifest
* @returns {Object} Incremental manifest
*/
async generateIncrementalManifest(options, parentManifest) {
const fullManifest = await this.generateManifest({
...options,
backupType: 'incremental'
});
const comparison = this.compareManifests(fullManifest, parentManifest);
// Add incremental-specific information
fullManifest.incremental = {
parent_backup_id: parentManifest.backup.id,
parent_timestamp: parentManifest.backup.timestamp,
changes: {
added_files_count: comparison.added_files.length,
modified_files_count: comparison.modified_files.length,
deleted_files_count: comparison.deleted_files.length,
unchanged_files_count: comparison.unchanged_files.length,
size_difference: comparison.size_difference
},
added_files: comparison.added_files.map(f => f.path),
modified_files: comparison.modified_files.map(f => f.path),
deleted_files: comparison.deleted_files.map(f => f.path)
};
return fullManifest;
}
// Helper methods
generateBackupId() {
const timestamp = new Date().toISOString().replace(/[:-]/g, '').replace('T', '-').split('.')[0];
const random = crypto.randomBytes(4).toString('hex');
return `backup-${timestamp}-${random}`;
}
async generateFileChecksums(files) {
const checksums = {};
for (const file of files) {
if (file.checksum) {
checksums[file.relativePath || file.path] = file.checksum;
}
}
return checksums;
}
getFileType(filePath) {
const ext = path.extname(filePath).toLowerCase();
const typeMap = {
'.jpg': 'image',
'.jpeg': 'image',
'.png': 'image',
'.gif': 'image',
'.webp': 'image',
'.zip': 'archive',
'.sql': 'database',
'.db': 'database',
'.json': 'config',
'.yaml': 'config',
'.yml': 'config'
};
return typeMap[ext] || 'other';
}
getDatabaseType() {
return process.env.DB_TYPE === 'postgresql' ? 'postgresql' : 'sqlite';
}
async getSchemaVersion() {
try {
const result = await db('migrations')
.orderBy('run_at', 'desc')
.first();
return result ? result.migration_name : 'unknown';
} catch (error) {
return 'unknown';
}
}
async getBackupSettings() {
try {
const settings = await db('app_settings')
.where('setting_type', 'backup')
.select('setting_key', 'setting_value');
const config = {};
settings.forEach(setting => {
try {
config[setting.setting_key] = JSON.parse(setting.setting_value);
} catch (e) {
config[setting.setting_key] = setting.setting_value;
}
});
return config;
} catch (error) {
return {};
}
}
async getActiveEventsCount() {
try {
const result = await db('events')
.where('status', 'active')
.count('* as count')
.first();
return result ? parseInt(result.count) : 0;
} catch (error) {
return 0;
}
}
async getArchivedEventsCount() {
try {
const result = await db('events')
.where('status', 'archived')
.count('* as count')
.first();
return result ? parseInt(result.count) : 0;
} catch (error) {
return 0;
}
}
async getTotalPhotosCount() {
try {
const result = await db('photos')
.count('* as count')
.first();
return result ? parseInt(result.count) : 0;
} catch (error) {
return 0;
}
}
calculateManifestChecksum(manifest) {
// Create a copy without the checksum field
const manifestCopy = JSON.parse(JSON.stringify(manifest));
if (manifestCopy.verification) {
delete manifestCopy.verification.total_checksum;
}
// Calculate SHA256 of the sorted JSON
const content = JSON.stringify(manifestCopy, Object.keys(manifestCopy).sort());
return crypto.createHash('sha256').update(content).digest('hex');
}
/**
* Generate a summary report from a manifest
* @param {Object} manifest - Manifest to summarize
* @returns {string} Human-readable summary
*/
generateSummaryReport(manifest) {
const report = [];
report.push('=== BACKUP MANIFEST SUMMARY ===');
report.push(`Backup ID: ${manifest.backup.id}`);
report.push(`Type: ${manifest.backup.type}`);
report.push(`Created: ${manifest.backup.timestamp}`);
if (manifest.backup.parent_backup_id) {
report.push(`Parent Backup: ${manifest.backup.parent_backup_id}`);
}
report.push('\n--- System Information ---');
report.push(`Host: ${manifest.system.hostname}`);
report.push(`Platform: ${manifest.system.platform} ${manifest.system.os_release}`);
report.push(`Architecture: ${manifest.system.architecture}`);
report.push('\n--- Application Information ---');
report.push(`App Version: ${manifest.application.version}`);
report.push(`Node Version: ${manifest.application.node_version}`);
report.push(`Environment: ${manifest.application.environment}`);
report.push('\n--- Files Summary ---');
report.push(`Total Files: ${manifest.files.count}`);
report.push(`Total Size: ${(manifest.files.total_size / 1024 / 1024).toFixed(2)} MB`);
if (manifest.incremental) {
report.push('\n--- Incremental Changes ---');
report.push(`Added Files: ${manifest.incremental.changes.added_files_count}`);
report.push(`Modified Files: ${manifest.incremental.changes.modified_files_count}`);
report.push(`Deleted Files: ${manifest.incremental.changes.deleted_files_count}`);
report.push(`Size Difference: ${(manifest.incremental.changes.size_difference / 1024 / 1024).toFixed(2)} MB`);
}
report.push('\n--- Database Information ---');
report.push(`Type: ${manifest.database.type}`);
report.push(`Size: ${manifest.database.size ? (manifest.database.size / 1024 / 1024).toFixed(2) + ' MB' : 'N/A'}`);
report.push(`Schema Version: ${manifest.database.schema_version}`);
report.push('\n--- Content Statistics ---');
report.push(`Active Events: ${manifest.metadata.active_events_count}`);
report.push(`Archived Events: ${manifest.metadata.archived_events_count}`);
report.push(`Total Photos: ${manifest.metadata.total_photos_count}`);
report.push('\n--- Verification ---');
report.push(`Manifest Checksum: ${manifest.verification.total_checksum}`);
report.push(`Integrity Timestamp: ${manifest.verification.integrity_timestamp}`);
return report.join('\n');
}
}
// Export singleton instance
module.exports = new BackupManifestGenerator();
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,720 @@
const cron = require('node-cron');
const path = require('path');
const fs = require('fs').promises;
const crypto = require('crypto');
const { exec } = require('child_process');
const { promisify } = require('util');
const execAsync = promisify(exec);
const { db } = require('../database/db');
const { queueEmail } = require('./emailProcessor');
const logger = require('../utils/logger');
const { formatBoolean } = require('../utils/dbCompat');
const backupManifest = require('./backupManifest');
// Backup job reference
let backupJob = null;
let backupConfig = null;
let isRunning = false;
// Storage paths
const getStoragePath = () => process.env.STORAGE_PATH || path.join(__dirname, '../../../storage');
/**
* Calculate file checksum using SHA256
*/
async function calculateChecksum(filePath) {
const hash = crypto.createHash('sha256');
const stream = require('fs').createReadStream(filePath);
return new Promise((resolve, reject) => {
stream.on('data', data => hash.update(data));
stream.on('end', () => resolve(hash.digest('hex')));
stream.on('error', reject);
});
}
/**
* Get database backup information
*/
async function getDatabaseBackupInfo() {
try {
// Check for recent database backup
const recentDbBackup = await db('database_backup_runs')
.where('status', 'completed')
.orderBy('completed_at', 'desc')
.first();
if (recentDbBackup && recentDbBackup.file_path) {
return {
type: recentDbBackup.backup_type,
backupFile: recentDbBackup.file_path,
size: recentDbBackup.file_size_bytes,
checksum: recentDbBackup.checksum,
tables: recentDbBackup.statistics ? JSON.parse(recentDbBackup.statistics).tables : {},
rowCounts: recentDbBackup.table_checksums ? JSON.parse(recentDbBackup.table_checksums) : {}
};
}
return {
type: process.env.DB_TYPE === 'postgresql' ? 'postgresql' : 'sqlite',
backupFile: null,
size: 0,
checksum: null,
tables: {},
rowCounts: {}
};
} catch (error) {
logger.error('Failed to get database backup info:', error);
return {
type: 'unknown',
backupFile: null,
size: 0,
checksum: null,
tables: {},
rowCounts: {}
};
}
}
/**
* Get backup configuration from database
*/
async function getBackupConfig() {
try {
const settings = await db('app_settings')
.where('setting_type', 'backup')
.select('setting_key', 'setting_value');
const config = {};
settings.forEach(setting => {
try {
config[setting.setting_key] = JSON.parse(setting.setting_value);
} catch (e) {
config[setting.setting_key] = setting.setting_value;
}
});
return config;
} catch (error) {
logger.error('Failed to get backup configuration:', error);
return null;
}
}
/**
* Get list of files to backup
*/
async function getFilesToBackup(includeArchived = true) {
const files = [];
const storagePath = getStoragePath();
try {
// Active events
const activePath = path.join(storagePath, 'events/active');
await scanDirectory(activePath, files, storagePath);
// Archived events (if enabled)
if (includeArchived) {
const archivePath = path.join(storagePath, 'events/archived');
await scanDirectory(archivePath, files, storagePath);
}
// Thumbnails
const thumbsPath = path.join(storagePath, 'thumbnails');
await scanDirectory(thumbsPath, files, storagePath);
// Uploads (logos, favicons, etc.)
const uploadsPath = path.join(storagePath, 'uploads');
await scanDirectory(uploadsPath, files, storagePath);
return files;
} catch (error) {
logger.error('Failed to get files to backup:', error);
throw error;
}
}
/**
* Recursively scan directory for files
*/
async function scanDirectory(dirPath, fileList, basePath, excludePatterns = []) {
try {
const entries = await fs.readdir(dirPath, { withFileTypes: true });
for (const entry of entries) {
const fullPath = path.join(dirPath, entry.name);
const relativePath = path.relative(basePath, fullPath);
// Check exclude patterns
if (excludePatterns.some(pattern => {
if (pattern.includes('*')) {
return new RegExp(pattern.replace(/\*/g, '.*')).test(entry.name);
}
return entry.name === pattern;
})) {
continue;
}
if (entry.isDirectory()) {
await scanDirectory(fullPath, fileList, basePath, excludePatterns);
} else if (entry.isFile()) {
const stats = await fs.stat(fullPath);
fileList.push({
path: fullPath,
relativePath: relativePath,
size: stats.size,
modified: stats.mtime
});
}
}
} catch (error) {
if (error.code !== 'ENOENT') {
logger.error(`Failed to scan directory ${dirPath}:`, error);
}
}
}
/**
* Check if file has changed since last backup
*/
async function hasFileChanged(filePath, checksum) {
try {
const fileState = await db('backup_file_states')
.where('file_path', filePath)
.first();
return !fileState || fileState.checksum !== checksum;
} catch (error) {
logger.error('Failed to check file state:', error);
return true; // Assume changed if we can't check
}
}
/**
* Update file state in database
*/
async function updateFileState(filePath, checksum, size, modified) {
try {
const existing = await db('backup_file_states')
.where('file_path', filePath)
.first();
const data = {
file_path: filePath,
checksum: checksum,
size_bytes: size,
last_modified: modified,
last_backed_up: new Date()
};
if (existing) {
await db('backup_file_states')
.where('id', existing.id)
.update(data);
} else {
await db('backup_file_states').insert(data);
}
} catch (error) {
logger.error('Failed to update file state:', error);
}
}
/**
* Perform local directory backup
*/
async function performLocalBackup(config, files) {
const destPath = config.backup_destination_path;
const storagePath = getStoragePath();
let backedUpCount = 0;
let backedUpSize = 0;
const backedUpFiles = [];
// Ensure destination exists
await fs.mkdir(destPath, { recursive: true });
for (const file of files) {
try {
// Skip large files if configured
const maxSizeMB = config.backup_max_file_size_mb || 5000;
if (file.size > maxSizeMB * 1024 * 1024) {
logger.warn(`Skipping large file: ${file.relativePath} (${(file.size / 1024 / 1024).toFixed(2)} MB)`);
continue;
}
// Calculate checksum
const checksum = await calculateChecksum(file.path);
file.checksum = checksum; // Add checksum to file object
// Check if file has changed
const changed = await hasFileChanged(file.relativePath, checksum);
if (!changed) {
continue;
}
// Copy file
const destFilePath = path.join(destPath, file.relativePath);
const destDir = path.dirname(destFilePath);
await fs.mkdir(destDir, { recursive: true });
await fs.copyFile(file.path, destFilePath);
// Update state
await updateFileState(file.relativePath, checksum, file.size, file.modified);
backedUpCount++;
backedUpSize += file.size;
backedUpFiles.push(file.relativePath);
} catch (error) {
logger.error(`Failed to backup file ${file.relativePath}:`, error);
}
}
return { backedUpCount, backedUpSize, backedUpFiles };
}
/**
* Perform rsync backup
*/
async function performRsyncBackup(config, files) {
const storagePath = getStoragePath();
const host = config.backup_rsync_host;
const user = config.backup_rsync_user;
const remotePath = config.backup_rsync_path;
const sshKey = config.backup_rsync_ssh_key;
if (!host || !remotePath) {
throw new Error('Rsync configuration incomplete');
}
// Build rsync command
const rsyncOptions = [
'-avz', // archive, verbose, compress
'--delete', // remove deleted files
'--stats' // show statistics
];
if (sshKey) {
rsyncOptions.push(`-e "ssh -i ${sshKey} -o StrictHostKeyChecking=no"`);
}
// Add exclude patterns
const excludePatterns = config.backup_exclude_patterns || [];
excludePatterns.forEach(pattern => {
rsyncOptions.push(`--exclude="${pattern}"`);
});
const source = `${storagePath}/`;
const destination = user ? `${user}@${host}:${remotePath}` : `${host}:${remotePath}`;
const rsyncCommand = `rsync ${rsyncOptions.join(' ')} "${source}" "${destination}"`;
try {
const { stdout, stderr } = await execAsync(rsyncCommand);
// Parse rsync stats
const stats = parseRsyncStats(stdout);
// Update file states for successfully synced files
for (const file of files) {
try {
const checksum = await calculateChecksum(file.path);
await updateFileState(file.relativePath, checksum, file.size, file.modified);
} catch (error) {
logger.error(`Failed to update state for ${file.relativePath}:`, error);
}
}
return {
backedUpCount: stats.filesTransferred || files.length,
backedUpSize: stats.totalSize || files.reduce((sum, f) => sum + f.size, 0),
backedUpFiles: files.map(f => f.relativePath)
};
} catch (error) {
logger.error('Rsync backup failed:', error);
throw new Error(`Rsync backup failed: ${error.message}`);
}
}
/**
* Parse rsync statistics from output
*/
function parseRsyncStats(output) {
const stats = {};
// Extract files transferred
const filesMatch = output.match(/Number of files transferred: (\d+)/);
if (filesMatch) {
stats.filesTransferred = parseInt(filesMatch[1]);
}
// Extract total size
const sizeMatch = output.match(/Total file size: ([\d,]+) bytes/);
if (sizeMatch) {
stats.totalSize = parseInt(sizeMatch[1].replace(/,/g, ''));
}
return stats;
}
/**
* Perform S3-compatible backup
*/
async function performS3Backup(config, files) {
// This would require AWS SDK or similar
// For now, return a placeholder
throw new Error('S3 backup not implemented yet');
}
/**
* Run backup process
*/
async function runBackup() {
if (isRunning) {
logger.warn('Backup already running, skipping');
return;
}
isRunning = true;
const startTime = new Date();
let backupRun = null;
try {
// Get current configuration
const config = await getBackupConfig();
if (!config.backup_enabled) {
logger.info('Backup is disabled, skipping');
return;
}
// Create backup run record
const [runId] = await db('backup_runs').insert({
started_at: startTime,
status: 'running',
backup_type: 'scheduled'
});
backupRun = { id: runId };
// Get files to backup
const files = await getFilesToBackup(config.backup_include_archived);
logger.info(`Found ${files.length} files to check for backup`);
// Perform backup based on destination type
let result;
switch (config.backup_destination_type) {
case 'local':
result = await performLocalBackup(config, files);
break;
case 'rsync':
result = await performRsyncBackup(config, files);
break;
case 's3':
result = await performS3Backup(config, files);
break;
default:
throw new Error(`Unknown backup destination type: ${config.backup_destination_type}`);
}
// Calculate duration
const endTime = new Date();
const durationSeconds = Math.round((endTime - startTime) / 1000);
// Generate backup manifest
let manifestPath = null;
try {
logger.info('Generating backup manifest...');
// Get database backup info if available
const databaseInfo = await getDatabaseBackupInfo();
// Determine if this is an incremental backup
const lastSuccessfulBackup = await db('backup_runs')
.where('status', 'completed')
.whereNot('id', runId)
.orderBy('completed_at', 'desc')
.first();
let manifest;
const manifestOptions = {
backupType: lastSuccessfulBackup ? 'incremental' : 'full',
backupPath: config.backup_destination_path || config.backup_destination_type,
files: files.filter(f => result.backedUpFiles && result.backedUpFiles.includes(f.relativePath)),
databaseInfo: databaseInfo,
parentBackupId: lastSuccessfulBackup ? lastSuccessfulBackup.manifest_id : null,
format: config.backup_manifest_format || 'json',
customMetadata: {
backup_run_id: runId,
destination_type: config.backup_destination_type,
operator: 'system',
reason: 'scheduled',
retentionDays: config.backup_retention_days || 30
}
};
if (lastSuccessfulBackup && lastSuccessfulBackup.manifest_path) {
try {
const parentManifest = await backupManifest.loadManifest(lastSuccessfulBackup.manifest_path);
manifest = await backupManifest.generateIncrementalManifest(manifestOptions, parentManifest);
} catch (error) {
logger.warn('Failed to load parent manifest, generating full manifest:', error);
manifest = await backupManifest.generateManifest(manifestOptions);
}
} else {
manifest = await backupManifest.generateManifest(manifestOptions);
}
// Save manifest
const manifestDir = config.backup_manifest_path || path.join(config.backup_destination_path || '/backup', 'manifests');
await fs.mkdir(manifestDir, { recursive: true });
const manifestFileName = `backup-manifest-${manifest.backup.id}.${config.backup_manifest_format || 'json'}`;
manifestPath = path.join(manifestDir, manifestFileName);
await backupManifest.saveManifest(manifest, manifestPath, config.backup_manifest_format || 'json');
logger.info(`Backup manifest saved to ${manifestPath}`);
} catch (error) {
logger.error('Failed to generate backup manifest:', error);
// Don't fail the entire backup for manifest generation failure
}
// Update backup run record
await db('backup_runs')
.where('id', runId)
.update({
completed_at: endTime,
status: 'completed',
files_backed_up: result.backedUpCount,
total_size_bytes: result.backedUpSize,
duration_seconds: durationSeconds,
manifest_path: manifestPath,
manifest_id: manifestPath ? path.basename(manifestPath, path.extname(manifestPath)) : null,
statistics: JSON.stringify({
totalFilesChecked: files.length,
filesBackedUp: result.backedUpCount,
totalSize: result.backedUpSize,
averageFileSize: result.backedUpCount > 0 ? Math.round(result.backedUpSize / result.backedUpCount) : 0,
manifestGenerated: !!manifestPath
})
});
logger.info(`Backup completed: ${result.backedUpCount} files, ${(result.backedUpSize / 1024 / 1024).toFixed(2)} MB in ${durationSeconds}s`);
// Send success email if configured
if (config.backup_email_on_success) {
// Get admin emails
const admins = await db('admin_users').where('is_active', formatBoolean(true));
for (const admin of admins) {
await queueEmail(null, admin.email, 'backup_completed', {
start_time: startTime.toISOString(),
duration: `${durationSeconds} seconds`,
files_count: result.backedUpCount.toString(),
total_size: `${(result.backedUpSize / 1024 / 1024).toFixed(2)} MB`,
backup_type: config.backup_destination_type
});
}
}
} catch (error) {
logger.error('Backup failed:', error);
// Update backup run record
if (backupRun) {
await db('backup_runs')
.where('id', backupRun.id)
.update({
completed_at: new Date(),
status: 'failed',
error_message: error.message
});
}
// Send failure email
const config = await getBackupConfig();
if (config && config.backup_email_on_failure) {
const admins = await db('admin_users').where('is_active', formatBoolean(true));
for (const admin of admins) {
await queueEmail(null, admin.email, 'backup_failed', {
start_time: startTime.toISOString(),
backup_type: config.backup_destination_type || 'unknown',
error_message: error.message
});
}
}
} finally {
isRunning = false;
}
}
/**
* Start backup service
*/
async function startBackupService() {
try {
// Get configuration
backupConfig = await getBackupConfig();
if (!backupConfig || !backupConfig.backup_enabled) {
logger.info('Backup service is disabled');
return;
}
// Cancel existing job if any
if (backupJob) {
backupJob.stop();
}
// Schedule backup job
const schedule = backupConfig.backup_schedule || '0 2 * * *'; // Default: 2 AM daily
backupJob = cron.schedule(schedule, async () => {
logger.info('Starting scheduled backup');
await runBackup();
});
logger.info(`Backup service started with schedule: ${schedule}`);
} catch (error) {
logger.error('Failed to start backup service:', error);
}
}
/**
* Stop backup service
*/
function stopBackupService() {
if (backupJob) {
backupJob.stop();
backupJob = null;
logger.info('Backup service stopped');
}
}
/**
* Trigger manual backup
*/
async function triggerManualBackup() {
logger.info('Starting manual backup');
await runBackup();
}
/**
* Get backup status and history
*/
async function getBackupStatus(limit = 10) {
try {
const runs = await db('backup_runs')
.orderBy('started_at', 'desc')
.limit(limit);
const lastRun = runs[0];
const isHealthy = lastRun && lastRun.status === 'completed';
// Validate manifest if exists
let manifestValid = false;
if (lastRun && lastRun.manifest_path) {
try {
const manifest = await backupManifest.loadManifest(lastRun.manifest_path);
backupManifest.validateManifest(manifest);
manifestValid = true;
} catch (error) {
logger.warn('Manifest validation failed:', error);
}
}
return {
isRunning,
isHealthy,
lastRun: lastRun ? {
...lastRun,
manifestValid
} : null,
recentRuns: runs,
nextScheduledRun: backupJob ? getNextScheduledRun() : null
};
} catch (error) {
logger.error('Failed to get backup status:', error);
return {
isRunning,
isHealthy: false,
error: error.message
};
}
}
/**
* Get next scheduled run time
*/
function getNextScheduledRun() {
// This is a simplified version - would need proper cron parsing
const now = new Date();
const tomorrow = new Date(now);
tomorrow.setDate(tomorrow.getDate() + 1);
tomorrow.setHours(2, 0, 0, 0); // Assuming default 2 AM schedule
return tomorrow.toISOString();
}
/**
* Clean up old backup runs
*/
async function cleanupOldBackupRuns(retentionDays = 30) {
try {
const cutoffDate = new Date();
cutoffDate.setDate(cutoffDate.getDate() - retentionDays);
const deleted = await db('backup_runs')
.where('started_at', '<', cutoffDate)
.delete();
if (deleted > 0) {
logger.info(`Cleaned up ${deleted} old backup runs`);
}
} catch (error) {
logger.error('Failed to cleanup old backup runs:', error);
}
}
/**
* Get backup manifest for a specific backup run
*/
async function getBackupManifest(backupRunId) {
try {
const run = await db('backup_runs')
.where('id', backupRunId)
.first();
if (!run || !run.manifest_path) {
throw new Error('Backup manifest not found');
}
const manifest = await backupManifest.loadManifest(run.manifest_path);
return {
manifest,
summary: backupManifest.generateSummaryReport(manifest)
};
} catch (error) {
logger.error('Failed to get backup manifest:', error);
throw error;
}
}
/**
* Validate a backup manifest file
*/
async function validateBackupManifest(manifestPath) {
try {
const manifest = await backupManifest.loadManifest(manifestPath);
backupManifest.validateManifest(manifest);
return { valid: true, manifest };
} catch (error) {
return { valid: false, error: error.message };
}
}
module.exports = {
startBackupService,
stopBackupService,
triggerManualBackup,
getBackupStatus,
runBackup,
cleanupOldBackupRuns,
getBackupManifest,
validateBackupManifest
};
@@ -0,0 +1,147 @@
/**
* Database Backup Service Usage Examples
*
* This service provides comprehensive database backup functionality
* with support for both SQLite and PostgreSQL databases.
*/
const { databaseBackupService } = require('./databaseBackup');
// Example 1: Manual backup with default settings
async function manualBackup() {
try {
const result = await databaseBackupService.backup();
console.log('Backup completed:', result);
// Result includes: path, size, duration, checksum, compressionRatio
} catch (error) {
console.error('Backup failed:', error);
}
}
// Example 2: Backup with custom options
async function customBackup() {
try {
const result = await databaseBackupService.backup({
destinationPath: '/custom/backup/path',
compress: true, // Enable gzip compression
validateIntegrity: true, // Validate backup after creation
includeChecksums: true, // Calculate table checksums
noTransaction: false // Use transaction for consistency (PostgreSQL)
});
console.log('Custom backup completed:', result);
} catch (error) {
console.error('Backup failed:', error);
}
}
// Example 3: Check backup progress (useful for long-running backups)
async function backupWithProgress() {
// Start backup asynchronously
const backupPromise = databaseBackupService.backup();
// Poll for progress
const progressInterval = setInterval(() => {
const progress = databaseBackupService.getProgress();
if (progress) {
console.log(`Progress: ${progress.message}`, progress.details);
}
}, 1000);
try {
const result = await backupPromise;
clearInterval(progressInterval);
console.log('Backup completed:', result);
} catch (error) {
clearInterval(progressInterval);
console.error('Backup failed:', error);
}
}
// Example 4: Get backup history
async function getBackupHistory() {
const history = await databaseBackupService.getBackupHistory(10);
history.forEach(backup => {
console.log(`Backup ${backup.id}:`);
console.log(` Started: ${backup.started_at}`);
console.log(` Status: ${backup.status}`);
console.log(` Size: ${(backup.file_size_bytes / 1024 / 1024).toFixed(2)} MB`);
console.log(` Duration: ${backup.duration_seconds}s`);
});
}
// Example 5: Clean up old backups
async function cleanupBackups() {
// Delete backups older than 30 days
await databaseBackupService.cleanupOldBackups(30);
console.log('Old backups cleaned up');
}
// Example 6: Get table checksums (useful for monitoring changes)
async function getTableChecksums() {
const checksums = await databaseBackupService.getTableChecksums();
console.log('Table Checksums:');
Object.entries(checksums).forEach(([table, info]) => {
console.log(` ${table}: ${info.rowCount} rows, checksum: ${info.checksum}`);
});
}
// Example 7: Using the scheduled backup service
const { startScheduledBackups, stopScheduledBackups } = require('./databaseBackup');
async function setupScheduledBackups() {
// Start scheduled backups (reads schedule from database config)
await startScheduledBackups();
console.log('Scheduled backups started');
// Later, if needed, stop scheduled backups
// stopScheduledBackups();
}
// Example 8: Admin API endpoints available
/*
GET /api/admin/database-backup/status - Get backup status and config
PUT /api/admin/database-backup/config - Update backup configuration
POST /api/admin/database-backup/backup - Trigger manual backup
GET /api/admin/database-backup/progress - Get current backup progress
GET /api/admin/database-backup/history - Get backup history with pagination
DELETE /api/admin/database-backup/cleanup - Delete old backup files
POST /api/admin/database-backup/test - Test backup configuration
GET /api/admin/database-backup/checksums - Get current table checksums
*/
// Example 9: Configuration options stored in database
/*
database_backup_enabled: boolean - Enable/disable scheduled backups
database_backup_schedule: string - Cron schedule (default: '0 3 * * *')
database_backup_destination_path: string - Where to store backups
database_backup_compress: boolean - Enable gzip compression
database_backup_validate_integrity: boolean - Validate after backup
database_backup_include_checksums: boolean - Calculate table checksums
database_backup_retention_days: number - Days to keep old backups
database_backup_email_on_failure: boolean - Send email on failure
database_backup_email_on_success: boolean - Send email on success
*/
// Example 10: Production considerations
/*
1. Ensure destination path has sufficient space
2. For large databases, backups may take significant time
3. PostgreSQL backups use single-transaction mode by default
4. Compression typically reduces size by 70-90%
5. Schedule backups during low-traffic periods
6. Monitor backup history for failures
7. Test restore procedures regularly
8. Consider replication for real-time redundancy
*/
module.exports = {
manualBackup,
customBackup,
backupWithProgress,
getBackupHistory,
cleanupBackups,
getTableChecksums,
setupScheduledBackups
};
+631
View File
@@ -0,0 +1,631 @@
const fs = require('fs').promises;
const path = require('path');
const { exec } = require('child_process');
const { promisify } = require('util');
const execAsync = promisify(exec);
const crypto = require('crypto');
const zlib = require('zlib');
const { pipeline } = require('stream/promises');
const { createReadStream, createWriteStream } = require('fs');
const { db } = require('../database/db');
const knexConfig = require('../../knexfile');
const logger = require('../utils/logger');
const { queueEmail } = require('./emailProcessor');
const { formatBoolean } = require('../utils/dbCompat');
// Constants
const CHUNK_SIZE = 1024 * 1024; // 1MB chunks for streaming
const PROGRESS_INTERVAL = 100; // Report progress every 100 rows
/**
* Database Backup Service
* Supports both SQLite and PostgreSQL with proper escaping,
* compression, checksums, and validation
*/
class DatabaseBackupService {
constructor() {
this.isRunning = false;
this.currentProgress = null;
this.dbType = knexConfig.client === 'pg' ? 'postgresql' : 'sqlite';
}
/**
* Calculate checksum for a file
*/
async calculateChecksum(filePath) {
const hash = crypto.createHash('sha256');
const stream = createReadStream(filePath);
return new Promise((resolve, reject) => {
stream.on('data', data => hash.update(data));
stream.on('end', () => resolve(hash.digest('hex')));
stream.on('error', reject);
});
}
/**
* Compress a file using gzip
*/
async compressFile(inputPath, outputPath) {
const gzip = zlib.createGzip({ level: 6 }); // Balanced compression
const source = createReadStream(inputPath);
const destination = createWriteStream(outputPath);
await pipeline(source, gzip, destination);
// Get compression ratio
const inputStats = await fs.stat(inputPath);
const outputStats = await fs.stat(outputPath);
const ratio = (1 - outputStats.size / inputStats.size) * 100;
return {
originalSize: inputStats.size,
compressedSize: outputStats.size,
compressionRatio: ratio.toFixed(2)
};
}
/**
* Get database size
*/
async getDatabaseSize() {
if (this.dbType === 'sqlite') {
const dbPath = knexConfig.connection.filename;
const stats = await fs.stat(dbPath);
return stats.size;
} else {
// PostgreSQL
const result = await db.raw(`
SELECT pg_database_size(current_database()) as size
`);
return parseInt(result.rows[0].size);
}
}
/**
* Get table checksums for change detection
*/
async getTableChecksums() {
const checksums = {};
const tables = await this.getTables();
for (const table of tables) {
if (this.dbType === 'sqlite') {
// SQLite: Use aggregate of all row data
const result = await db.raw(`
SELECT
COUNT(*) as row_count,
COALESCE(SUM(LENGTH(CAST(t.* AS TEXT))), 0) as data_sum
FROM "${table}" t
`);
checksums[table] = {
rowCount: result[0].row_count,
checksum: crypto
.createHash('md5')
.update(`${result[0].row_count}-${result[0].data_sum}`)
.digest('hex')
};
} else {
// PostgreSQL: Use built-in functions
const result = await db.raw(`
SELECT
COUNT(*) as row_count,
MD5(COALESCE(STRING_AGG(MD5(t::text), ''), '')) as checksum
FROM "${table}" t
`);
checksums[table] = {
rowCount: parseInt(result.rows[0].row_count),
checksum: result.rows[0].checksum || 'empty'
};
}
}
return checksums;
}
/**
* Get list of tables
*/
async getTables() {
if (this.dbType === 'sqlite') {
const result = await db.raw(`
SELECT name FROM sqlite_master
WHERE type='table'
AND name NOT LIKE 'sqlite_%'
AND name != 'knex_migrations'
AND name != 'knex_migrations_lock'
ORDER BY name
`);
return result.map(row => row.name);
} else {
// PostgreSQL
const result = await db.raw(`
SELECT table_name
FROM information_schema.tables
WHERE table_schema = 'public'
AND table_type = 'BASE TABLE'
AND table_name NOT IN ('knex_migrations', 'knex_migrations_lock')
ORDER BY table_name
`);
return result.rows.map(row => row.table_name);
}
}
/**
* Create SQLite backup
*/
async createSQLiteBackup(outputPath, options = {}) {
const dbPath = knexConfig.connection.filename;
const tempPath = `${outputPath}.tmp`;
try {
// Use SQLite's backup API for consistency
await execAsync(`sqlite3 "${dbPath}" ".backup '${tempPath}'"`);
// Verify the backup
const verifyResult = await execAsync(`sqlite3 "${tempPath}" "PRAGMA integrity_check"`);
if (!verifyResult.stdout.includes('ok')) {
throw new Error('Backup integrity check failed');
}
// Move temp file to final location
await fs.rename(tempPath, outputPath);
return { success: true };
} catch (error) {
// Cleanup temp file if exists
try {
await fs.unlink(tempPath);
} catch (e) {
// Ignore
}
throw error;
}
}
/**
* Create PostgreSQL backup with proper escaping
*/
async createPostgreSQLBackup(outputPath, options = {}) {
const { host, port, user, password, database } = knexConfig.connection;
// Build connection string with proper escaping
const connectionParts = [
`host=${host}`,
`port=${port}`,
`dbname=${database}`,
`user=${user}`
];
// Set PGPASSWORD environment variable for security
const env = { ...process.env };
if (password) {
env.PGPASSWORD = password;
}
// Build pg_dump command with options
const pgDumpOptions = [
'--verbose',
'--no-owner',
'--no-privileges',
'--clean',
'--if-exists',
'--format=plain',
'--encoding=UTF8'
];
// Add transaction support for consistency
if (!options.noTransaction) {
pgDumpOptions.push('--single-transaction');
}
// Add compression if not doing it separately
if (options.compress && !options.separateCompression) {
pgDumpOptions.push('--compress=6');
}
const command = `pg_dump "${connectionParts.join(' ')}" ${pgDumpOptions.join(' ')} > "${outputPath}"`;
try {
const { stderr } = await execAsync(command, {
env,
maxBuffer: 1024 * 1024 * 100 // 100MB buffer
});
// pg_dump writes progress to stderr, not an error
if (stderr && !stderr.includes('dump complete')) {
logger.warn('pg_dump warnings:', stderr);
}
// Verify the dump file is not empty
const stats = await fs.stat(outputPath);
if (stats.size === 0) {
throw new Error('Backup file is empty');
}
return { success: true, warnings: stderr };
} catch (error) {
throw new Error(`PostgreSQL backup failed: ${error.message}`);
}
}
/**
* Validate backup integrity
*/
async validateBackup(backupPath, originalChecksums) {
const tempDbPath = `${backupPath}.validate`;
try {
if (this.dbType === 'sqlite') {
// For SQLite, we can directly check integrity
const result = await execAsync(`sqlite3 "${backupPath}" "PRAGMA integrity_check"`);
if (!result.stdout.includes('ok')) {
throw new Error('Backup integrity check failed');
}
} else {
// For PostgreSQL, we'd need to restore to a temp database
// This is more complex and might not be feasible in production
logger.info('PostgreSQL backup validation would require restore test');
}
return { valid: true };
} finally {
// Cleanup
try {
await fs.unlink(tempDbPath);
} catch (e) {
// Ignore
}
}
}
/**
* Main backup method
*/
async backup(options = {}) {
if (this.isRunning) {
throw new Error('Backup already in progress');
}
this.isRunning = true;
const startTime = new Date();
let backupRun = null;
try {
// Get configuration
const config = await this.getBackupConfig();
const {
destinationPath = '/backup/database',
compress = true,
validateIntegrity = true,
includeChecksums = true
} = { ...config, ...options };
// Create backup directory
await fs.mkdir(destinationPath, { recursive: true });
// Generate backup filename
const timestamp = new Date().toISOString().replace(/[:.]/g, '-');
const baseName = `picpeak-db-${this.dbType}-${timestamp}`;
const sqlFile = path.join(destinationPath, `${baseName}.sql`);
const finalFile = compress ? path.join(destinationPath, `${baseName}.sql.gz`) : sqlFile;
// Create backup run record
const [runId] = await db('database_backup_runs').insert({
started_at: startTime,
status: 'running',
backup_type: this.dbType,
destination_path: finalFile
});
backupRun = { id: runId };
// Get initial checksums
let tableChecksums = null;
if (includeChecksums) {
this.updateProgress('Calculating table checksums...');
tableChecksums = await this.getTableChecksums();
}
// Get database size
const dbSize = await this.getDatabaseSize();
// Create the backup
this.updateProgress('Creating database backup...');
if (this.dbType === 'sqlite') {
await this.createSQLiteBackup(sqlFile, options);
} else {
await this.createPostgreSQLBackup(sqlFile, options);
}
// Compress if requested
let compressionStats = null;
if (compress) {
this.updateProgress('Compressing backup...');
compressionStats = await this.compressFile(sqlFile, finalFile);
await fs.unlink(sqlFile); // Remove uncompressed file
}
// Calculate checksum
this.updateProgress('Calculating backup checksum...');
const backupChecksum = await this.calculateChecksum(finalFile);
// Validate if requested
if (validateIntegrity && !compress) {
this.updateProgress('Validating backup integrity...');
await this.validateBackup(finalFile, tableChecksums);
}
// Get final file size
const finalStats = await fs.stat(finalFile);
// Calculate duration
const endTime = new Date();
const durationSeconds = Math.round((endTime - startTime) / 1000);
// Update backup run record
await db('database_backup_runs')
.where('id', runId)
.update({
completed_at: endTime,
status: 'completed',
file_path: finalFile,
file_size_bytes: finalStats.size,
original_size_bytes: dbSize,
duration_seconds: durationSeconds,
checksum: backupChecksum,
compression_ratio: compressionStats?.compressionRatio || null,
table_checksums: tableChecksums ? JSON.stringify(tableChecksums) : null,
statistics: JSON.stringify({
dbType: this.dbType,
compressed: compress,
validated: validateIntegrity,
compressionStats,
tableCount: tableChecksums ? Object.keys(tableChecksums).length : null
})
});
logger.info(`Database backup completed: ${finalFile} (${(finalStats.size / 1024 / 1024).toFixed(2)} MB) in ${durationSeconds}s`);
// Send success notification if configured
if (config.emailOnSuccess) {
await this.sendBackupNotification('success', {
duration: durationSeconds,
size: finalStats.size,
compressionRatio: compressionStats?.compressionRatio,
path: finalFile
});
}
return {
success: true,
path: finalFile,
size: finalStats.size,
duration: durationSeconds,
checksum: backupChecksum,
compressionRatio: compressionStats?.compressionRatio
};
} catch (error) {
logger.error('Database backup failed:', error);
// Update backup run record
if (backupRun) {
await db('database_backup_runs')
.where('id', backupRun.id)
.update({
completed_at: new Date(),
status: 'failed',
error_message: error.message
});
}
// Send failure notification
const config = await this.getBackupConfig();
if (config.emailOnFailure) {
await this.sendBackupNotification('failure', {
error: error.message
});
}
throw error;
} finally {
this.isRunning = false;
this.currentProgress = null;
}
}
/**
* Update progress
*/
updateProgress(message, details = {}) {
this.currentProgress = {
message,
details,
timestamp: new Date()
};
logger.info(`Backup progress: ${message}`, details);
}
/**
* Get current progress
*/
getProgress() {
return this.currentProgress;
}
/**
* Get backup configuration
*/
async getBackupConfig() {
const settings = await db('app_settings')
.where('setting_type', 'database_backup')
.select('setting_key', 'setting_value');
const config = {};
settings.forEach(setting => {
try {
config[setting.setting_key] = JSON.parse(setting.setting_value);
} catch (e) {
config[setting.setting_key] = setting.setting_value;
}
});
return config;
}
/**
* Send backup notification email
*/
async sendBackupNotification(type, details) {
const admins = await db('admin_users').where('is_active', formatBoolean(true));
for (const admin of admins) {
if (type === 'success') {
await queueEmail(null, admin.email, 'database_backup_completed', {
backup_type: this.dbType,
duration: `${details.duration} seconds`,
file_size: `${(details.size / 1024 / 1024).toFixed(2)} MB`,
compression_ratio: details.compressionRatio ? `${details.compressionRatio}%` : 'N/A',
file_path: details.path
});
} else {
await queueEmail(null, admin.email, 'database_backup_failed', {
backup_type: this.dbType,
error_message: details.error,
timestamp: new Date().toISOString()
});
}
}
}
/**
* Clean up old backups
*/
async cleanupOldBackups(retentionDays = 30) {
try {
const cutoffDate = new Date();
cutoffDate.setDate(cutoffDate.getDate() - retentionDays);
// Get old backup records
const oldBackups = await db('database_backup_runs')
.where('completed_at', '<', cutoffDate)
.where('status', 'completed')
.select('id', 'file_path');
let deletedCount = 0;
for (const backup of oldBackups) {
try {
// Delete the file
if (backup.file_path) {
await fs.unlink(backup.file_path);
}
// Delete the record
await db('database_backup_runs')
.where('id', backup.id)
.delete();
deletedCount++;
} catch (error) {
logger.error(`Failed to delete old backup ${backup.file_path}:`, error);
}
}
if (deletedCount > 0) {
logger.info(`Cleaned up ${deletedCount} old database backups`);
}
// Also clean up old failed runs
await db('database_backup_runs')
.where('started_at', '<', cutoffDate)
.where('status', 'failed')
.delete();
} catch (error) {
logger.error('Failed to cleanup old database backups:', error);
}
}
/**
* Get backup history
*/
async getBackupHistory(limit = 10) {
return await db('database_backup_runs')
.orderBy('started_at', 'desc')
.limit(limit);
}
/**
* Restore from backup (careful!)
*/
async restore(backupPath, options = {}) {
// This is a dangerous operation and should be used with extreme caution
throw new Error('Restore functionality not implemented for safety. Please restore manually.');
}
}
// Create singleton instance
const databaseBackupService = new DatabaseBackupService();
// Scheduled backup runner
let backupSchedule = null;
/**
* Start scheduled database backups
*/
async function startScheduledBackups() {
const cron = require('node-cron');
try {
const config = await databaseBackupService.getBackupConfig();
if (!config.enabled) {
logger.info('Database backup service is disabled');
return;
}
// Stop existing schedule
if (backupSchedule) {
backupSchedule.stop();
}
// Default schedule: 3 AM daily (offset from file backups at 2 AM)
const schedule = config.schedule || '0 3 * * *';
backupSchedule = cron.schedule(schedule, async () => {
logger.info('Starting scheduled database backup');
try {
await databaseBackupService.backup();
await databaseBackupService.cleanupOldBackups(config.retentionDays || 30);
} catch (error) {
logger.error('Scheduled database backup failed:', error);
}
});
logger.info(`Database backup service started with schedule: ${schedule}`);
} catch (error) {
logger.error('Failed to start database backup service:', error);
}
}
/**
* Stop scheduled database backups
*/
function stopScheduledBackups() {
if (backupSchedule) {
backupSchedule.stop();
backupSchedule = null;
logger.info('Database backup service stopped');
}
}
module.exports = {
databaseBackupService,
startScheduledBackups,
stopScheduledBackups,
DatabaseBackupService // Export class for testing
};
File diff suppressed because it is too large Load Diff
@@ -0,0 +1,311 @@
const S3StorageAdapter = require('../s3Storage');
const { S3Client } = require('@aws-sdk/client-s3');
const { Upload } = require('@aws-sdk/lib-storage');
const fs = require('fs');
const stream = require('stream');
// Mock AWS SDK
jest.mock('@aws-sdk/client-s3');
jest.mock('@aws-sdk/lib-storage');
jest.mock('@aws-sdk/s3-request-presigner');
describe('S3StorageAdapter', () => {
let mockS3Client;
let mockSend;
let s3Storage;
beforeEach(() => {
// Reset mocks
jest.clearAllMocks();
// Mock S3Client
mockSend = jest.fn();
mockS3Client = {
send: mockSend
};
S3Client.mockImplementation(() => mockS3Client);
// Create adapter instance
s3Storage = new S3StorageAdapter({
bucket: 'test-bucket',
region: 'us-east-1',
accessKeyId: 'test-key',
secretAccessKey: 'test-secret'
});
});
describe('constructor', () => {
it('should initialize with required config', () => {
expect(s3Storage.bucket).toBe('test-bucket');
expect(s3Storage.config.region).toBe('us-east-1');
});
it('should throw error if bucket is not provided', () => {
expect(() => {
new S3StorageAdapter({ region: 'us-east-1' });
}).toThrow('S3 bucket name is required');
});
it('should configure for MinIO with path style', () => {
const minioStorage = new S3StorageAdapter({
bucket: 'test-bucket',
endpoint: 'http://localhost:9000',
forcePathStyle: true,
sslEnabled: false
});
expect(S3Client).toHaveBeenCalledWith(
expect.objectContaining({
endpoint: 'http://localhost:9000',
forcePathStyle: true
})
);
});
});
describe('testConnection', () => {
it('should successfully test connection', async () => {
mockSend.mockResolvedValueOnce({});
const result = await s3Storage.testConnection();
expect(result).toBe(true);
expect(mockSend).toHaveBeenCalledWith(
expect.objectContaining({
input: { Bucket: 'test-bucket' }
})
);
});
it('should throw error on connection failure', async () => {
mockSend.mockRejectedValueOnce(new Error('Access Denied'));
await expect(s3Storage.testConnection()).rejects.toThrow('S3 connection test failed');
});
});
describe('upload', () => {
let mockUpload;
let mockDone;
beforeEach(() => {
mockDone = jest.fn().mockResolvedValue({
Location: 'https://test-bucket.s3.amazonaws.com/test-key',
ETag: '"test-etag"'
});
mockUpload = {
on: jest.fn().mockReturnThis(),
done: mockDone
};
Upload.mockImplementation(() => mockUpload);
// Mock fs.stat
jest.spyOn(fs.promises, 'stat').mockResolvedValue({
size: 1024
});
// Mock fs.createReadStream
jest.spyOn(fs, 'createReadStream').mockReturnValue(new stream.Readable());
});
afterEach(() => {
jest.restoreAllMocks();
});
it('should upload file successfully', async () => {
const result = await s3Storage.upload('/path/to/file.jpg', 'test-key');
expect(result.Location).toBe('https://test-bucket.s3.amazonaws.com/test-key');
expect(Upload).toHaveBeenCalledWith(
expect.objectContaining({
client: mockS3Client,
params: expect.objectContaining({
Bucket: 'test-bucket',
Key: 'test-key',
ContentType: 'application/octet-stream'
})
})
);
});
it('should track upload progress', async () => {
const onProgress = jest.fn();
let progressCallback;
mockUpload.on.mockImplementation((event, callback) => {
if (event === 'httpUploadProgress') {
progressCallback = callback;
}
return mockUpload;
});
const uploadPromise = s3Storage.upload('/path/to/file.jpg', 'test-key', {
onProgress
});
// Simulate progress
progressCallback({ loaded: 512, total: 1024 });
await uploadPromise;
expect(onProgress).toHaveBeenCalledWith(512, 1024);
});
it('should emit upload events', async () => {
const uploadStartSpy = jest.fn();
const uploadCompleteSpy = jest.fn();
s3Storage.on('uploadStart', uploadStartSpy);
s3Storage.on('uploadComplete', uploadCompleteSpy);
await s3Storage.upload('/path/to/file.jpg', 'test-key');
expect(uploadStartSpy).toHaveBeenCalledWith({ key: 'test-key', size: 1024 });
expect(uploadCompleteSpy).toHaveBeenCalledWith({
key: 'test-key',
location: 'https://test-bucket.s3.amazonaws.com/test-key'
});
});
});
describe('exists', () => {
it('should return true if object exists', async () => {
mockSend.mockResolvedValueOnce({});
const result = await s3Storage.exists('test-key');
expect(result).toBe(true);
expect(mockSend).toHaveBeenCalledWith(
expect.objectContaining({
input: { Bucket: 'test-bucket', Key: 'test-key' }
})
);
});
it('should return false if object does not exist', async () => {
const error = new Error('Not Found');
error.name = 'NotFound';
error.$metadata = { httpStatusCode: 404 };
mockSend.mockRejectedValueOnce(error);
const result = await s3Storage.exists('test-key');
expect(result).toBe(false);
});
});
describe('generateKey', () => {
it('should generate unique key with timestamp and random string', () => {
const key = s3Storage.generateKey('photo.jpg');
expect(key).toMatch(/^\d+_[a-f0-9]{16}_photo\.jpg$/);
});
it('should add prefix if provided', () => {
const key = s3Storage.generateKey('photo.jpg', 'events/wedding');
expect(key).toMatch(/^events\/wedding\/\d+_[a-f0-9]{16}_photo\.jpg$/);
});
it('should sanitize filename', () => {
const key = s3Storage.generateKey('my photo (1).jpg');
expect(key).toMatch(/^\d+_[a-f0-9]{16}_my_photo__1_\.jpg$/);
});
});
describe('retry logic', () => {
it('should retry on retryable errors', async () => {
const retryableError = new Error('Connection reset');
retryableError.code = 'ECONNRESET';
// First attempt fails, second succeeds
mockSend
.mockRejectedValueOnce(retryableError)
.mockResolvedValueOnce({});
// Mock setTimeout to speed up test
jest.useFakeTimers();
const promise = s3Storage.exists('test-key');
// Advance timers
jest.runAllTimers();
const result = await promise;
expect(result).toBe(true);
expect(mockSend).toHaveBeenCalledTimes(2);
jest.useRealTimers();
});
it('should not retry on non-retryable errors', async () => {
const nonRetryableError = new Error('Invalid credentials');
nonRetryableError.code = 'InvalidCredentials';
mockSend.mockRejectedValueOnce(nonRetryableError);
await expect(s3Storage.exists('test-key')).rejects.toThrow('Invalid credentials');
expect(mockSend).toHaveBeenCalledTimes(1);
});
it('should stop retrying after max attempts', async () => {
const retryableError = new Error('Service unavailable');
retryableError.code = 'ServiceUnavailable';
mockSend.mockRejectedValue(retryableError);
// Mock setTimeout to speed up test
jest.useFakeTimers();
const promise = s3Storage.exists('test-key');
// Advance timers for all retries
for (let i = 0; i < 4; i++) {
jest.runAllTimers();
}
await expect(promise).rejects.toThrow('Service unavailable');
expect(mockSend).toHaveBeenCalledTimes(4); // Initial + 3 retries
jest.useRealTimers();
});
});
describe('getStats', () => {
it('should calculate storage statistics', async () => {
mockSend.mockResolvedValueOnce({
Contents: [
{ Key: 'file1.jpg', Size: 1024 },
{ Key: 'file2.jpg', Size: 2048 }
],
NextContinuationToken: 'token123'
}).mockResolvedValueOnce({
Contents: [
{ Key: 'file3.jpg', Size: 3072 }
]
});
const stats = await s3Storage.getStats('events/');
expect(stats).toEqual({
totalSize: 6144,
totalCount: 3,
totalSizeFormatted: '6 KB'
});
});
});
describe('_formatBytes', () => {
it('should format bytes correctly', () => {
expect(s3Storage._formatBytes(0)).toBe('0 Bytes');
expect(s3Storage._formatBytes(1024)).toBe('1 KB');
expect(s3Storage._formatBytes(1048576)).toBe('1 MB');
expect(s3Storage._formatBytes(1073741824)).toBe('1 GB');
expect(s3Storage._formatBytes(1536, 1)).toBe('1.5 KB');
});
});
});
@@ -0,0 +1,221 @@
/**
* Example usage of the S3StorageAdapter
*
* This file demonstrates how to use the S3 storage adapter for various operations
*/
const S3StorageAdapter = require('./s3Storage');
// Example 1: Basic AWS S3 Configuration
const s3Storage = new S3StorageAdapter({
bucket: 'my-photo-bucket',
region: 'us-east-1',
accessKeyId: process.env.AWS_ACCESS_KEY_ID,
secretAccessKey: process.env.AWS_SECRET_ACCESS_KEY
});
// Example 2: MinIO Configuration (S3-compatible)
const minioStorage = new S3StorageAdapter({
bucket: 'photo-storage',
endpoint: 'http://localhost:9000', // MinIO endpoint
accessKeyId: 'minioadmin',
secretAccessKey: 'minioadmin',
forcePathStyle: true, // Required for MinIO
sslEnabled: false // For local development
});
// Example 3: DigitalOcean Spaces Configuration
const spacesStorage = new S3StorageAdapter({
bucket: 'my-space-name',
endpoint: 'https://nyc3.digitaloceanspaces.com',
region: 'nyc3',
accessKeyId: process.env.DO_SPACES_KEY,
secretAccessKey: process.env.DO_SPACES_SECRET
});
// Usage Examples
async function examples() {
try {
// Test connection
await s3Storage.testConnection();
console.log('Connection successful!');
// Upload a file with progress tracking
const uploadResult = await s3Storage.upload(
'/path/to/local/photo.jpg',
'events/wedding-2024/photo.jpg',
{
contentType: 'image/jpeg',
metadata: {
event: 'wedding-2024',
photographer: 'John Doe'
},
onProgress: (loaded, total) => {
const percentage = Math.round((loaded / total) * 100);
console.log(`Upload progress: ${percentage}%`);
}
}
);
console.log('Upload complete:', uploadResult.Location);
// Upload from stream
const readStream = fs.createReadStream('/path/to/large-video.mp4');
await s3Storage.uploadStream(readStream, 'events/wedding-2024/video.mp4', {
contentType: 'video/mp4',
onProgress: (loaded, total) => {
console.log(`Streamed ${loaded} of ${total} bytes`);
}
});
// Download a file
await s3Storage.download(
'events/wedding-2024/photo.jpg',
'/path/to/downloaded/photo.jpg',
{
onProgress: (loaded, total) => {
const percentage = Math.round((loaded / total) * 100);
console.log(`Download progress: ${percentage}%`);
}
}
);
// Get a download stream
const downloadStream = await s3Storage.downloadStream('events/wedding-2024/photo.jpg');
downloadStream.pipe(fs.createWriteStream('/path/to/output.jpg'));
// List files
const listing = await s3Storage.list('events/wedding-2024/');
console.log(`Found ${listing.Contents.length} files`);
listing.Contents.forEach(file => {
console.log(`- ${file.Key} (${file.Size} bytes)`);
});
// Generate pre-signed URL for temporary access
const downloadUrl = await s3Storage.getSignedUrl('getObject', 'events/wedding-2024/photo.jpg', {
expiresIn: 3600 // 1 hour
});
console.log('Pre-signed download URL:', downloadUrl);
// Generate pre-signed upload URL
const uploadUrl = await s3Storage.getSignedUrl('putObject', 'events/wedding-2024/new-photo.jpg', {
expiresIn: 1800, // 30 minutes
params: {
ContentType: 'image/jpeg'
}
});
console.log('Pre-signed upload URL:', uploadUrl);
// Check if file exists
const exists = await s3Storage.exists('events/wedding-2024/photo.jpg');
console.log('File exists:', exists);
// Get metadata
const metadata = await s3Storage.getMetadata('events/wedding-2024/photo.jpg');
console.log('File metadata:', metadata);
// Copy file
await s3Storage.copy(
'events/wedding-2024/photo.jpg',
'events/wedding-2024/photo-copy.jpg'
);
// Move file
await s3Storage.move(
'events/wedding-2024/photo-copy.jpg',
'events/wedding-2024/archived/photo.jpg'
);
// Delete file
await s3Storage.delete('events/wedding-2024/temp-photo.jpg');
// Delete multiple files
const deleteResult = await s3Storage.deleteMany([
'events/wedding-2024/temp1.jpg',
'events/wedding-2024/temp2.jpg',
'events/wedding-2024/temp3.jpg'
]);
console.log(`Deleted ${deleteResult.Deleted.length} files`);
// Get storage statistics
const stats = await s3Storage.getStats('events/');
console.log(`Total files: ${stats.totalCount}`);
console.log(`Total size: ${stats.totalSizeFormatted}`);
// Listen to events
s3Storage.on('uploadProgress', (data) => {
console.log(`Uploading ${data.key}: ${data.loaded}/${data.total}`);
});
s3Storage.on('uploadComplete', (data) => {
console.log(`Upload completed: ${data.key}`);
});
s3Storage.on('uploadError', (data) => {
console.error(`Upload failed for ${data.key}:`, data.error);
});
} catch (error) {
console.error('Error:', error);
}
}
// Integration with existing photo upload workflow
async function integrateWithPhotoUpload(eventId, files) {
const storage = new S3StorageAdapter({
bucket: process.env.S3_BUCKET,
region: process.env.AWS_REGION,
accessKeyId: process.env.AWS_ACCESS_KEY_ID,
secretAccessKey: process.env.AWS_SECRET_ACCESS_KEY
});
const uploadedPhotos = [];
for (const file of files) {
try {
// Generate unique S3 key
const s3Key = storage.generateKey(file.originalname, `events/${eventId}`);
// Upload to S3
const result = await storage.upload(file.path, s3Key, {
contentType: file.mimetype,
metadata: {
eventId: eventId,
originalName: file.originalname,
uploadedAt: new Date().toISOString()
}
});
uploadedPhotos.push({
filename: s3Key,
originalName: file.originalname,
size: file.size,
mimeType: file.mimetype,
s3Location: result.Location,
s3Key: s3Key
});
// Clean up local temp file
await fs.promises.unlink(file.path);
} catch (error) {
console.error(`Failed to upload ${file.originalname}:`, error);
throw error;
}
}
return uploadedPhotos;
}
// Environment variables needed:
// AWS_ACCESS_KEY_ID=your-access-key
// AWS_SECRET_ACCESS_KEY=your-secret-key
// AWS_REGION=us-east-1
// S3_BUCKET=your-bucket-name
// For MinIO:
// MINIO_ENDPOINT=http://localhost:9000
// MINIO_ACCESS_KEY=minioadmin
// MINIO_SECRET_KEY=minioadmin
// MINIO_BUCKET=photo-storage
module.exports = { examples, integrateWithPhotoUpload };
+733
View File
@@ -0,0 +1,733 @@
const { S3Client, HeadBucketCommand, HeadObjectCommand, GetObjectCommand, PutObjectCommand, DeleteObjectCommand, DeleteObjectsCommand, ListObjectsV2Command, CopyObjectCommand, CreateMultipartUploadCommand, UploadPartCommand, CompleteMultipartUploadCommand, AbortMultipartUploadCommand } = require('@aws-sdk/client-s3');
const { Upload } = require('@aws-sdk/lib-storage');
const { getSignedUrl } = require('@aws-sdk/s3-request-presigner');
const fs = require('fs');
const fsPromises = require('fs').promises;
const path = require('path');
const stream = require('stream');
const crypto = require('crypto');
const logger = require('../../utils/logger');
/**
* S3 Storage Adapter for handling file uploads to S3 and S3-compatible services
*
* Features:
* - Support for S3 and S3-compatible services (MinIO, DigitalOcean Spaces, etc.)
* - Multipart upload for large files (>100MB)
* - Progress tracking with event emitter
* - Retry logic with exponential backoff
* - Stream support for memory efficiency
* - Path-style URL support for MinIO
* - Connection testing
* - Comprehensive error handling
*
* @class S3StorageAdapter
*/
class S3StorageAdapter extends stream.EventEmitter {
/**
* Creates an instance of S3StorageAdapter
*
* @param {Object} config - Configuration object
* @param {string} config.bucket - S3 bucket name
* @param {string} [config.region='us-east-1'] - AWS region
* @param {string} [config.endpoint] - Custom endpoint URL for S3-compatible services
* @param {string} [config.accessKeyId] - AWS access key ID
* @param {string} [config.secretAccessKey] - AWS secret access key
* @param {boolean} [config.forcePathStyle=false] - Force path-style URLs (required for MinIO)
* @param {boolean} [config.sslEnabled=true] - Enable SSL for connections
* @param {number} [config.multipartThreshold=104857600] - Threshold for multipart upload (default 100MB)
* @param {number} [config.partSize=10485760] - Part size for multipart upload (default 10MB)
* @param {number} [config.maxRetries=3] - Maximum number of retry attempts
* @param {number} [config.retryDelay=1000] - Initial retry delay in milliseconds
*/
constructor(config) {
super();
// Validate required config
if (!config.bucket) {
throw new Error('S3 bucket name is required');
}
// Set defaults
this.config = {
region: 'us-east-1',
forcePathStyle: false,
sslEnabled: true,
multipartThreshold: 100 * 1024 * 1024, // 100MB
partSize: 10 * 1024 * 1024, // 10MB
maxRetries: 3,
retryDelay: 1000,
...config
};
// Initialize S3 client
const s3Config = {
region: this.config.region,
forcePathStyle: this.config.forcePathStyle
};
// Add credentials if provided
if (this.config.accessKeyId && this.config.secretAccessKey) {
s3Config.credentials = {
accessKeyId: this.config.accessKeyId,
secretAccessKey: this.config.secretAccessKey
};
}
// Add custom endpoint if provided (for S3-compatible services)
if (this.config.endpoint) {
s3Config.endpoint = this.config.endpoint;
// For MinIO and other S3-compatible services
if (!this.config.endpoint.startsWith('https://') && this.config.sslEnabled) {
s3Config.endpoint = `https://${this.config.endpoint}`;
} else if (!this.config.endpoint.startsWith('http://') && !this.config.sslEnabled) {
s3Config.endpoint = `http://${this.config.endpoint}`;
}
}
this.s3Client = new S3Client(s3Config);
this.bucket = this.config.bucket;
// Bind methods to preserve context
this.upload = this.upload.bind(this);
this.uploadStream = this.uploadStream.bind(this);
this.download = this.download.bind(this);
this.downloadStream = this.downloadStream.bind(this);
this.delete = this.delete.bind(this);
this.exists = this.exists.bind(this);
this.list = this.list.bind(this);
this.copy = this.copy.bind(this);
this.move = this.move.bind(this);
this.getSignedUrl = this.getSignedUrl.bind(this);
this.testConnection = this.testConnection.bind(this);
}
/**
* Test connection to S3 bucket
*
* @returns {Promise<boolean>} - True if connection successful
* @throws {Error} - If connection fails
*/
async testConnection() {
try {
await this.s3Client.send(new HeadBucketCommand({ Bucket: this.bucket }));
logger.info(`Successfully connected to S3 bucket: ${this.bucket}`);
return true;
} catch (error) {
logger.error(`Failed to connect to S3 bucket ${this.bucket}:`, error);
throw new Error(`S3 connection test failed: ${error.message}`);
}
}
/**
* Upload a file to S3 with automatic multipart for large files
*
* @param {string} localPath - Local file path to upload
* @param {string} s3Key - S3 object key (path in bucket)
* @param {Object} [options={}] - Additional options
* @param {Object} [options.metadata] - Object metadata
* @param {string} [options.contentType] - Content type
* @param {string} [options.cacheControl] - Cache control header
* @param {Function} [options.onProgress] - Progress callback function(loaded, total)
* @returns {Promise<Object>} - Upload result with Location, ETag, etc.
*/
async upload(localPath, s3Key, options = {}) {
try {
const stats = await fsPromises.stat(localPath);
const fileSize = stats.size;
// Emit upload start event
this.emit('uploadStart', { key: s3Key, size: fileSize });
// Create file stream
const fileStream = fs.createReadStream(localPath);
// Prepare upload parameters
const uploadParams = {
Bucket: this.bucket,
Key: s3Key,
Body: fileStream,
ContentType: options.contentType || 'application/octet-stream',
Metadata: options.metadata || {},
CacheControl: options.cacheControl
};
// Remove undefined values
Object.keys(uploadParams).forEach(key => uploadParams[key] === undefined && delete uploadParams[key]);
// Use AWS SDK v3 Upload for automatic multipart handling
const parallelUploads3 = new Upload({
client: this.s3Client,
params: uploadParams,
queueSize: 4, // Optional: concurrent uploads
partSize: this.config.partSize,
leavePartsOnError: false
});
// Track upload progress
parallelUploads3.on('httpUploadProgress', (progress) => {
if (options.onProgress) {
options.onProgress(progress.loaded, progress.total);
}
this.emit('uploadProgress', { key: s3Key, loaded: progress.loaded, total: progress.total });
});
// Perform upload with retry
const result = await this._retryOperation(() => parallelUploads3.done());
this.emit('uploadComplete', { key: s3Key, location: result.Location });
return result;
} catch (error) {
logger.error(`Failed to upload file ${localPath} to S3:`, error);
this.emit('uploadError', { key: s3Key, error });
throw error;
}
}
/**
* Upload a stream to S3
*
* @param {stream.Readable} readStream - Readable stream to upload
* @param {string} s3Key - S3 object key
* @param {Object} [options={}] - Additional options
* @returns {Promise<Object>} - Upload result
*/
async uploadStream(readStream, s3Key, options = {}) {
const uploadParams = {
Bucket: this.bucket,
Key: s3Key,
Body: readStream,
ContentType: options.contentType || 'application/octet-stream',
Metadata: options.metadata || {},
CacheControl: options.cacheControl
};
// Remove undefined values
Object.keys(uploadParams).forEach(key => uploadParams[key] === undefined && delete uploadParams[key]);
const parallelUploads3 = new Upload({
client: this.s3Client,
params: uploadParams,
queueSize: 4,
partSize: this.config.partSize,
leavePartsOnError: false
});
// Track progress if callback provided
if (options.onProgress) {
parallelUploads3.on('httpUploadProgress', (progress) => {
options.onProgress(progress.loaded, progress.total);
this.emit('uploadProgress', { key: s3Key, ...progress });
});
}
return await this._retryOperation(() => parallelUploads3.done());
}
/**
* Download a file from S3
*
* @param {string} s3Key - S3 object key
* @param {string} localPath - Local file path to save to
* @param {Object} [options={}] - Additional options
* @param {Function} [options.onProgress] - Progress callback
* @returns {Promise<void>}
*/
async download(s3Key, localPath, options = {}) {
try {
// Ensure directory exists
await fsPromises.mkdir(path.dirname(localPath), { recursive: true });
// Get object metadata first for progress tracking
const headResult = await this._retryOperation(() =>
this.s3Client.send(new HeadObjectCommand({
Bucket: this.bucket,
Key: s3Key
}))
);
const fileSize = headResult.ContentLength;
this.emit('downloadStart', { key: s3Key, size: fileSize });
// Get object
const getObjectResult = await this._retryOperation(() =>
this.s3Client.send(new GetObjectCommand({
Bucket: this.bucket,
Key: s3Key
}))
);
// Create write stream
const writeStream = fs.createWriteStream(localPath);
// Download with progress tracking
await new Promise((resolve, reject) => {
let downloaded = 0;
const bodyStream = getObjectResult.Body;
bodyStream.on('data', (chunk) => {
downloaded += chunk.length;
if (options.onProgress) {
options.onProgress(downloaded, fileSize);
}
this.emit('downloadProgress', { key: s3Key, loaded: downloaded, total: fileSize });
});
bodyStream.on('error', reject);
writeStream.on('error', reject);
writeStream.on('finish', resolve);
bodyStream.pipe(writeStream);
});
this.emit('downloadComplete', { key: s3Key });
} catch (error) {
logger.error(`Failed to download file ${s3Key} from S3:`, error);
this.emit('downloadError', { key: s3Key, error });
throw error;
}
}
/**
* Get a download stream from S3
*
* @param {string} s3Key - S3 object key
* @param {Object} [options={}] - Additional options
* @param {string} [options.range] - Byte range to download (e.g., 'bytes=0-1023')
* @returns {Promise<stream.Readable>} - Readable stream
*/
async downloadStream(s3Key, options = {}) {
const params = {
Bucket: this.bucket,
Key: s3Key,
Range: options.range
};
// Remove undefined values
Object.keys(params).forEach(key => params[key] === undefined && delete params[key]);
const result = await this.s3Client.send(new GetObjectCommand(params));
return result.Body;
}
/**
* Delete a file from S3
*
* @param {string} s3Key - S3 object key
* @returns {Promise<void>}
*/
async delete(s3Key) {
return await this._retryOperation(() =>
this.s3Client.send(new DeleteObjectCommand({
Bucket: this.bucket,
Key: s3Key
}))
);
}
/**
* Delete multiple files from S3
*
* @param {string[]} s3Keys - Array of S3 object keys
* @returns {Promise<Object>} - Delete result
*/
async deleteMany(s3Keys) {
if (!s3Keys || s3Keys.length === 0) {
return { Deleted: [], Errors: [] };
}
// S3 deleteObjects has a limit of 1000 keys per request
const chunks = [];
for (let i = 0; i < s3Keys.length; i += 1000) {
chunks.push(s3Keys.slice(i, i + 1000));
}
const results = await Promise.all(
chunks.map(chunk =>
this._retryOperation(() =>
this.s3Client.send(new DeleteObjectsCommand({
Bucket: this.bucket,
Delete: {
Objects: chunk.map(key => ({ Key: key })),
Quiet: false
}
}))
)
)
);
// Combine results
return {
Deleted: results.flatMap(r => r.Deleted || []),
Errors: results.flatMap(r => r.Errors || [])
};
}
/**
* Check if a file exists in S3
*
* @param {string} s3Key - S3 object key
* @returns {Promise<boolean>} - True if exists
*/
async exists(s3Key) {
try {
await this.s3Client.send(new HeadObjectCommand({
Bucket: this.bucket,
Key: s3Key
}));
return true;
} catch (error) {
if (error.name === 'NotFound' || error.$metadata?.httpStatusCode === 404) {
return false;
}
throw error;
}
}
/**
* List files in S3
*
* @param {string} prefix - S3 prefix to list
* @param {Object} [options={}] - Additional options
* @param {number} [options.maxKeys=1000] - Maximum number of keys to return
* @param {string} [options.continuationToken] - Continuation token for pagination
* @returns {Promise<Object>} - List result with Contents array and NextContinuationToken
*/
async list(prefix, options = {}) {
const params = {
Bucket: this.bucket,
Prefix: prefix,
MaxKeys: options.maxKeys || 1000,
ContinuationToken: options.continuationToken
};
return await this._retryOperation(() =>
this.s3Client.send(new ListObjectsV2Command(params))
);
}
/**
* Copy a file within S3
*
* @param {string} sourceKey - Source S3 object key
* @param {string} targetKey - Target S3 object key
* @param {Object} [options={}] - Additional options
* @returns {Promise<Object>} - Copy result
*/
async copy(sourceKey, targetKey, options = {}) {
const copySource = `${this.bucket}/${sourceKey}`;
const params = {
Bucket: this.bucket,
CopySource: copySource,
Key: targetKey,
MetadataDirective: options.metadata ? 'REPLACE' : 'COPY',
Metadata: options.metadata,
ContentType: options.contentType,
CacheControl: options.cacheControl
};
// Remove undefined values
Object.keys(params).forEach(key => params[key] === undefined && delete params[key]);
return await this._retryOperation(() =>
this.s3Client.send(new CopyObjectCommand(params))
);
}
/**
* Move a file within S3 (copy then delete)
*
* @param {string} sourceKey - Source S3 object key
* @param {string} targetKey - Target S3 object key
* @param {Object} [options={}] - Additional options
* @returns {Promise<Object>} - Move result
*/
async move(sourceKey, targetKey, options = {}) {
// First copy the object
const copyResult = await this.copy(sourceKey, targetKey, options);
// Then delete the original
await this.delete(sourceKey);
return copyResult;
}
/**
* Get a pre-signed URL for downloading or uploading
*
* @param {string} operation - Operation type ('getObject' or 'putObject')
* @param {string} s3Key - S3 object key
* @param {Object} [options={}] - Additional options
* @param {number} [options.expiresIn=3600] - URL expiration in seconds
* @param {Object} [options.params] - Additional parameters for the operation
* @returns {Promise<string>} - Pre-signed URL
*/
async getSignedUrl(operation, s3Key, options = {}) {
const params = {
Bucket: this.bucket,
Key: s3Key,
...options.params
};
let command;
switch (operation.toLowerCase()) {
case 'getobject':
command = new GetObjectCommand(params);
break;
case 'putobject':
command = new PutObjectCommand(params);
break;
default:
throw new Error(`Unsupported operation: ${operation}`);
}
return await getSignedUrl(this.s3Client, command, {
expiresIn: options.expiresIn || 3600
});
}
/**
* Get object metadata
*
* @param {string} s3Key - S3 object key
* @returns {Promise<Object>} - Object metadata
*/
async getMetadata(s3Key) {
return await this._retryOperation(() =>
this.s3Client.send(new HeadObjectCommand({
Bucket: this.bucket,
Key: s3Key
}))
);
}
/**
* Update object metadata
*
* @param {string} s3Key - S3 object key
* @param {Object} metadata - New metadata
* @returns {Promise<Object>} - Update result
*/
async updateMetadata(s3Key, metadata) {
// S3 requires copying the object to itself to update metadata
return await this.copy(s3Key, s3Key, { metadata });
}
/**
* Manual multipart upload for advanced use cases
*
* @param {string} localPath - Local file path
* @param {string} s3Key - S3 object key
* @param {number} fileSize - File size in bytes
* @param {Object} options - Upload options
* @returns {Promise<Object>} - Upload result
* @private
*/
async _manualMultipartUpload(localPath, s3Key, fileSize, options) {
logger.info(`Starting manual multipart upload for ${s3Key} (${fileSize} bytes)`);
// Initiate multipart upload
const multipartParams = {
Bucket: this.bucket,
Key: s3Key,
ContentType: options.contentType || 'application/octet-stream',
Metadata: options.metadata || {},
CacheControl: options.cacheControl
};
// Remove undefined values
Object.keys(multipartParams).forEach(key => multipartParams[key] === undefined && delete multipartParams[key]);
const multipart = await this._retryOperation(() =>
this.s3Client.send(new CreateMultipartUploadCommand(multipartParams))
);
const uploadId = multipart.UploadId;
const partSize = this.config.partSize;
const numParts = Math.ceil(fileSize / partSize);
let uploaded = 0;
const parts = [];
try {
// Upload parts
for (let partNum = 1; partNum <= numParts; partNum++) {
const start = (partNum - 1) * partSize;
const end = Math.min(start + partSize, fileSize);
const partStream = fs.createReadStream(localPath, {
start,
end: end - 1
});
const partParams = {
Bucket: this.bucket,
Key: s3Key,
PartNumber: partNum,
UploadId: uploadId,
Body: partStream
};
// Upload part with retry
const partResult = await this._retryOperation(() =>
this.s3Client.send(new UploadPartCommand(partParams))
);
parts.push({
ETag: partResult.ETag,
PartNumber: partNum
});
uploaded += (end - start);
if (options.onProgress) {
options.onProgress(uploaded, fileSize);
}
this.emit('uploadProgress', { key: s3Key, loaded: uploaded, total: fileSize });
logger.info(`Uploaded part ${partNum}/${numParts} for ${s3Key}`);
}
// Complete multipart upload
const completeParams = {
Bucket: this.bucket,
Key: s3Key,
UploadId: uploadId,
MultipartUpload: { Parts: parts }
};
const result = await this._retryOperation(() =>
this.s3Client.send(new CompleteMultipartUploadCommand(completeParams))
);
this.emit('uploadComplete', { key: s3Key, location: result.Location });
logger.info(`Completed multipart upload for ${s3Key}`);
return result;
} catch (error) {
// Abort multipart upload on error
logger.error(`Multipart upload failed for ${s3Key}, aborting:`, error);
try {
await this.s3Client.send(new AbortMultipartUploadCommand({
Bucket: this.bucket,
Key: s3Key,
UploadId: uploadId
}));
} catch (abortError) {
logger.error(`Failed to abort multipart upload:`, abortError);
}
throw error;
}
}
/**
* Retry operation with exponential backoff
* @private
*/
async _retryOperation(operation, retryCount = 0) {
try {
return await operation();
} catch (error) {
if (retryCount >= this.config.maxRetries) {
throw error;
}
// Check if error is retryable
const retryableErrors = ['ECONNRESET', 'ETIMEDOUT', 'ENOTFOUND', 'ESOCKETTIMEDOUT', 'RequestTimeout', 'SlowDown', 'ServiceUnavailable', 'InternalError'];
const isRetryable = retryableErrors.some(code =>
error.code === code ||
error.name === code ||
error.$metadata?.httpStatusCode === 503 ||
error.$metadata?.httpStatusCode === 500
);
if (!isRetryable) {
throw error;
}
// Calculate delay with exponential backoff and jitter
const delay = Math.min(
this.config.retryDelay * Math.pow(2, retryCount) + Math.random() * 1000,
30000 // Max 30 seconds
);
logger.warn(`Retrying operation after ${delay}ms (attempt ${retryCount + 1}/${this.config.maxRetries}):`, error.message);
await new Promise(resolve => setTimeout(resolve, delay));
return this._retryOperation(operation, retryCount + 1);
}
}
/**
* Generate a unique S3 key for a file
*
* @param {string} originalName - Original filename
* @param {string} [prefix=''] - Optional prefix for the key
* @returns {string} - Generated S3 key
*/
generateKey(originalName, prefix = '') {
const timestamp = Date.now();
const randomStr = crypto.randomBytes(8).toString('hex');
const ext = path.extname(originalName);
const basename = path.basename(originalName, ext);
// Sanitize basename
const safeName = basename.replace(/[^a-zA-Z0-9-_]/g, '_');
const key = `${timestamp}_${randomStr}_${safeName}${ext}`;
return prefix ? path.posix.join(prefix, key) : key;
}
/**
* Get storage statistics
*
* @param {string} [prefix=''] - Optional prefix to filter
* @returns {Promise<Object>} - Storage statistics
*/
async getStats(prefix = '') {
let totalSize = 0;
let totalCount = 0;
let continuationToken;
do {
const result = await this.list(prefix, { continuationToken });
if (result.Contents) {
totalCount += result.Contents.length;
totalSize += result.Contents.reduce((sum, obj) => sum + (obj.Size || 0), 0);
}
continuationToken = result.NextContinuationToken;
} while (continuationToken);
return {
totalSize,
totalCount,
totalSizeFormatted: this._formatBytes(totalSize)
};
}
/**
* Format bytes to human readable format
* @private
*/
_formatBytes(bytes, decimals = 2) {
if (bytes === 0) return '0 Bytes';
const k = 1024;
const dm = decimals < 0 ? 0 : decimals;
const sizes = ['Bytes', 'KB', 'MB', 'GB', 'TB'];
const i = Math.floor(Math.log(bytes) / Math.log(k));
return parseFloat((bytes / Math.pow(k, i)).toFixed(dm)) + ' ' + sizes[i];
}
}
module.exports = S3StorageAdapter;