VPS Backup Manager

A resilient, multi-site automated backup manager for Linux VPS environments. It handles PostgreSQL database dumps, integrity verification, Cloudflare R2 media sync, local asset folder/CDN synchronization or archiving, automated log rotation, modern public HTML status dashboards for Caddy/Nginx, and dual notifications via Discord and Telegram.

---

Key Features

• Per-Site Rclone Remotes: Every site can define its own dedicated Google Drive, Nextcloud, or R2 remotes (SITE<N>_REMOTE_*), completely isolating credentials and destination buckets across projects.
• Granular Remote Retention: Configure a general retention period (SITE<N>_DB_REMOTE_RETENTION_DAYS) or dedicated per-provider days (SITE<N>_DB_REMOTE_RETENTION_DAYS_GDRIVE, SITE<N>_DB_REMOTE_RETENTION_DAYS_NEXTCLOUD).
• Optional Versioned Trash for Folders & CDN: When syncing local asset folders, versioned date-stamped trash bins (SITE<N>_FOLDER_TRASH_PATH_*) are completely optional. If omitted, standard direct synchronization is performed.
• Multi-Site Discovery: Automatically scans and processes any number of sites configured as SITE<N>_* in .env without modifying code.
• Local Folders & CDN Backups: Supports backing up local asset directories (e.g. CDN folders, uploads) using either high-performance incremental sync with dated trash (rclone sync --backup-dir) or compressed versioned archives (tar.gz or zip).
• PostgreSQL Dumps & Verification: Generates custom compressed PostgreSQL dumps (.dump) and verifies archive integrity using pg_restore --list before starting uploads.
• Credential Protection: Database passwords are automatically parsed from connection URIs and supplied via PGPASSWORD at runtime, preventing plain-text password exposure in ps aux or /proc/<pid>/cmdline.
• Cloudflare R2 Anti-Wipe Protection: Checks bucket size and file counts before performing rclone sync. If the source bucket is empty or unreachable, the sync aborts immediately to prevent wiping remote backups.
• Versioned Trash Retention (--backup-dir): Preserves overwritten or deleted files in dated trash folders before pruning them according to retention policies.
• Automated Log Rotation: Prevents log files from consuming VPS disk space by automatically truncating older entries when exceeding configurable size thresholds (BACKUP_LOG_MAX_MB).
• Live HTML Status Dashboard: Generates a self-contained, responsive dark-mode HTML dashboard (generate_report.py) with real-time log search and system health indicators, ready to be served publicly in read-only mode by Caddy or Nginx.
• Dual Alerts (Discord & Telegram): Delivers rich embed status reports to Discord channels and instant alerts + full log document delivery to Telegram bots.
• Concurrency Lock (flock): Prevents overlapping backup cron executions using Linux file descriptors.

---

Requirements

Ensure the following tools are installed on your Linux system:

Tool	Purpose
bash (>= 4.4)	Script execution with set -Eeuo pipefail
rclone	Cloud storage sync and copy operations
pg_dump & pg_restore	PostgreSQL database dumping and integrity checks
python3 (>= 3.8)	HTML status dashboard rendering
curl	Discord and Telegram webhook delivery
jq	Building JSON payloads for Discord embeds
flock	Concurrency locking
find	Local dump cleanup and retention management
tar & gzip	Standard Unix compression for folder archive mode
zip	Optional zip compression for folder archive mode

On Debian/Ubuntu, install the core dependencies via:

sudo apt-get update
sudo apt-get install -y postgresql-client rclone curl jq util-linux python3 tar zip

---

Directory Structure

backup_manager/
├── run_backup.sh        # Main backup orchestrator script
├── generate_report.py   # Self-contained HTML status dashboard generator
├── .env.example         # Environment configuration template
├── .env                 # Private configuration file (git-ignored)
└── logs/
    ├── backup.log       # System log file (auto-rotated)
    └── index.html       # Public read-only dashboard served by Caddy

---

Configuration (.env)

Copy the configuration template:

cp .env.example .env
chmod 600 .env
chmod +x run_backup.sh generate_report.py

1. General & Dashboard Settings

BACKUP_LOG_FILE="/home/felt/backup_manager/logs/backup.log"
BACKUP_LOG_MAX_MB=10
BACKUP_HTML_FILE="/home/felt/backup_manager/logs/index.html"

# Telegram Notifications
TELEGRAM_BOT_TOKEN="your_bot_token_here"
TELEGRAM_CHAT_ID="your_chat_id_here"
TELEGRAM_SEND_LOG_DOCUMENT=true

2. Defining Sites (SITE1, SITE2, SITE3, ...)

Sites are declared dynamically with numerical suffixes (SITE1, SITE2, etc.):

# Site 1: Production Database + Cloudflare R2
SITE1_ENABLE=true
SITE1_NAME="production_site"
SITE1_DB_URL="postgres://user:password@127.0.0.1:5432/production_db"
SITE1_REMOTE_GDRIVE="production_gdrive"
SITE1_REMOTE_NEXTCLOUD="nextcloud"
SITE1_REMOTE_R2="production_r2"
SITE1_DB_LOCAL_DIR="/home/felt/backup_manager/dumps/production_site"
SITE1_DB_PATH_GDRIVE="backups/production/database"
SITE1_DB_PATH_NEXTCLOUD="backups/production/database"
SITE1_R2_ENABLE=true
SITE1_R2_BUCKET="production-bucket"
SITE1_R2_PATH_GDRIVE="backups/production/r2"
SITE1_R2_TRASH_PATH_GDRIVE="backups/production/r2_trash"
SITE1_DISCORD_WEBHOOK_URL="https://discord.com/api/webhooks/..."

# Site 2: Database + Local CDN / Assets Folder
SITE2_ENABLE=true
SITE2_NAME="assets_site"
SITE2_DB_URL="postgres://user:password@127.0.0.1:5432/assets_db"
SITE2_REMOTE_GDRIVE="assets_drive"
SITE2_REMOTE_NEXTCLOUD="nextcloud"
SITE2_DB_LOCAL_DIR="/home/felt/backup_manager/dumps/assets_site"
SITE2_DB_PATH_GDRIVE="backups/assets/database"
SITE2_DB_PATH_NEXTCLOUD="backups/assets/database"
SITE2_FOLDER_ENABLE=true
SITE2_FOLDER_SRC="/var/www/assets_site/cdn"
SITE2_FOLDER_PATH_GDRIVE="backups/assets/cdn"
SITE2_FOLDER_PATH_NEXTCLOUD="backups/assets/cdn"
SITE2_FOLDER_MODE="sync" # "sync" for incremental mirror or "archive" for tar.gz/zip
SITE2_FOLDER_TRASH_PATH_GDRIVE="backups/assets/cdn_trash"
SITE2_DISCORD_WEBHOOK_URL="https://discord.com/api/webhooks/..."

---

Serving the Public Status Page with Caddy

To publish the read-only status dashboard publicly, add a simple virtual host to your Caddyfile:

status.yourdomain.com {
    root * /home/felt/backup_manager/logs
    file_server
}

Then reload Caddy:

sudo systemctl reload caddy

---

Running the Backup

Manual execution:

./run_backup.sh

Scheduled execution via crontab:

crontab -e

Add a daily schedule (e.g. every night at 03:00 AM):

0 3 * * * /home/felt/backup_manager/run_backup.sh >/dev/null 2>&1

---

License

MIT License. Free for personal and commercial use.