backup-recovery – skill DevOps & Hạ tầng cho agent AI

Backup and Recovery

Implement comprehensive backup and recovery strategies using rsync, Restic, and cloud storage backends. Covers the 3-2-1 rule, automated scheduling, S3/B2 backends, encryption, restore procedures, and verification testing.

When to Use

• Designing a backup strategy for servers, databases, or application data
• Setting up Restic for encrypted, deduplicated backups to local or cloud storage
• Automating backups with systemd timers or cron
• Restoring data after accidental deletion, corruption, or disaster
• Migrating data between environments using backup/restore workflows
• Verifying backup integrity and testing recovery procedures

Prerequisites

• rsync installed (included in most Linux distributions)
• restic installed (v0.16+ recommended)
• Cloud CLI configured for the backend: AWS CLI for S3, b2 CLI for Backblaze B2
• Sufficient storage at the backup destination (2-3x source size for retention)
• SSH access for remote rsync targets
• systemd or cron for scheduling

The 3-2-1 Backup Rule

• 3 copies of your data (1 primary + 2 backups)
• 2 different storage media or types (e.g., local disk + cloud)
• 1 copy offsite (cloud storage, remote datacenter)

Confirm source/destination paths, retention ownership and recovery targets. A mirror with --delete propagates deletions and is not an independent historical backup; dry-run before destructive sync/prune. Never copy a running PostgreSQL data directory as a consistent backup: use pg_dump or a supported physical backup/snapshot protocol. Run pipelines in Bash with pipefail so failed dumps cannot become apparently successful empty snapshots. Restore to an isolated new database/path first; --clean deletes target objects. Keep repository credentials and recovery keys in a separate approved location; logs and backup data are sensitive.

rsync Backups

Basic Operations

# Sync a local directory to a backup location
rsync -avz --delete /data/ /backup/data/

# Flags explained:
# -a  archive mode (preserves permissions, ownership, timestamps, symlinks)
# -v  verbose output
# -z  compress data during transfer
# --delete  remove files at destination that no longer exist at source

# Sync to a remote server over SSH
rsync -avz -e "ssh -i ~/.ssh/backup_key" /data/ backup@remote:/backups/server01/

# Exclude patterns
rsync -avz --delete \
  --exclude='*.tmp' \
  --exclude='*.log' \
  --exclude='.cache/' \
  --exclude='node_modules/' \
  /data/ /backup/data/

# Use an exclude file for complex patterns
rsync -avz --delete --exclude-from=/etc/backup-excludes.txt /data/ /backup/data/

# Dry run (preview what would change)
rsync -avzn --delete /data/ /backup/data/

# Limit bandwidth to 50 MB/s and show progress
rsync -avz --bwlimit=50000 --progress /data/ backup@remote:/backups/

Incremental Backups with Hard Links

# Incremental: unchanged files hard-linked to previous backup (saves space)
rsync -avz --delete \
  --link-dest=/backup/daily/latest \
  /data/ /backup/daily/$(date +%Y-%m-%d)/

# Update the 'latest' symlink
ln -snf /backup/daily/$(date +%Y-%m-%d) /backup/daily/latest

# Remove backups older than 30 days
find /backup/daily -maxdepth 1 -type d -name "20*" -mtime +30 -exec rm -rf {} \;

Restic Backup

Installation

# Debian / Ubuntu
apt install -y restic

# RHEL / CentOS
dnf install -y restic

# Alternatively obtain a pinned release for the actual OS/architecture from
# https://github.com/restic/restic/releases and verify its published checksums
# and signature before installing it. Do not mix latest/ URLs with old filenames.

# Verify installation
restic version

Initialize a Repository

# Local repository
restic init --repo /backup/restic-repo

# AWS S3 backend
export AWS_ACCESS_KEY_ID="AKIAEXAMPLE"
export AWS_SECRET_ACCESS_KEY="secretkey"
restic init --repo s3:s3.amazonaws.com/my-backup-bucket

# S3-compatible (MinIO)
export AWS_ACCESS_KEY_ID="minioadmin"
export AWS_SECRET_ACCESS_KEY="miniosecret"
restic init --repo s3:https://minio.example.com:9000/backup-bucket

# Backblaze B2 backend
export B2_ACCOUNT_ID="accountid"
export B2_ACCOUNT_KEY="accountkey"
restic init --repo b2:my-backup-bucket:server01

# SFTP backend
restic init --repo sftp:backup@remote:/backups/server01

# Restic will prompt for a repository password -- store it securely
# Use a password file for automation
umask 077
read -r -s -p "Repository password: " RESTIC_NEW_PASSWORD
printf '%s\n' "$RESTIC_NEW_PASSWORD" > /etc/restic/password.txt
unset RESTIC_NEW_PASSWORD
chmod 600 /etc/restic/password.txt

Backup Operations

# Basic backup
restic backup /data --repo /backup/restic-repo --password-file /etc/restic/password.txt

# Backup multiple directories
restic backup /data /etc \
  --repo s3:s3.amazonaws.com/my-backup-bucket \
  --password-file /etc/restic/password.txt

# Backup with exclusions
restic backup /data \
  --exclude='*.tmp' \
  --exclude='*.log' \
  --exclude-file=/etc/restic/excludes.txt \
  --repo /backup/restic-repo \
  --password-file /etc/restic/password.txt

# Backup with tags (useful for filtering snapshots later)
restic backup /data \
  --tag server01 --tag production --tag daily \
  --repo /backup/restic-repo \
  --password-file /etc/restic/password.txt

# Backup stdin (e.g., database dump)
pg_dump -U postgres mydb | restic backup --stdin --stdin-filename mydb.sql \
  --repo s3:s3.amazonaws.com/my-backup-bucket \
  --password-file /etc/restic/password.txt

# Verbose output showing files processed
restic backup /data -v \
  --repo /backup/restic-repo \
  --password-file /etc/restic/password.txt

Snapshot Management

# List all snapshots (add --tag <tag> to filter)
restic snapshots --repo /backup/restic-repo --password-file /etc/restic/password.txt

# Browse files in the latest snapshot
restic ls latest --repo /backup/restic-repo --password-file /etc/restic/password.txt

# Compare two snapshots
restic diff abc123 def456 --repo /backup/restic-repo --password-file /etc/restic/password.txt

Retention Policy (forget + prune)

# Apply retention policy and reclaim space
restic forget \
  --keep-daily 7 --keep-weekly 4 --keep-monthly 12 --keep-yearly 3 \
  --prune \
  --repo /backup/restic-repo \
  --password-file /etc/restic/password.txt

# Dry run to preview what would be removed
restic forget --keep-daily 7 --keep-weekly 4 --keep-monthly 12 \
  --dry-run --repo /backup/restic-repo --password-file /etc/restic/password.txt

Restore Procedures

# Restore the latest snapshot to a target directory
restic restore latest --target /restore \
  --repo /backup/restic-repo \
  --password-file /etc/restic/password.txt

# Restore a specific snapshot by ID
restic restore abc123 --target /restore \
  --repo /backup/restic-repo \
  --password-file /etc/restic/password.txt

# Restore only specific files or directories
restic restore latest --target /restore --include "/data/config" \
  --repo /backup/restic-repo \
  --password-file /etc/restic/password.txt

# Mount a snapshot as a FUSE filesystem (browse and copy individual files)
mkdir -p /mnt/restic
restic mount /mnt/restic --repo /backup/restic-repo --password-file /etc/restic/password.txt &
# Browse: ls /mnt/restic/snapshots/latest/
# Unmount when done: umount /mnt/restic

Verification

# Verify repository metadata/structure (does not read every data pack)
restic check --repo /backup/restic-repo --password-file /etc/restic/password.txt

# Full data verification (reads all pack files -- slow but thorough)
restic check --read-data --repo /backup/restic-repo --password-file /etc/restic/password.txt

# Verify a random subset of data (faster than full read)
restic check --read-data-subset=5% --repo /backup/restic-repo --password-file /etc/restic/password.txt

Automated Backup with Environment File

/etc/restic/env

# Repository configuration
export RESTIC_REPOSITORY="s3:s3.amazonaws.com/my-backup-bucket"
export RESTIC_PASSWORD_FILE="/etc/restic/password.txt"
export AWS_ACCESS_KEY_ID="AKIAEXAMPLE"
export AWS_SECRET_ACCESS_KEY="secretkey"

# Optional: set cache directory
export RESTIC_CACHE_DIR="/var/cache/restic"

Backup Script

#!/bin/bash
# /usr/local/bin/restic-backup.sh
set -euo pipefail
source /etc/restic/env
LOG="/var/log/restic-backup.log"

echo "$(date): Starting backup" >> "$LOG"

restic backup /data /etc \
  --exclude-file=/etc/restic/excludes.txt \
  --tag "$(hostname)" --tag daily \
  --verbose >> "$LOG" 2>&1

restic forget --keep-daily 7 --keep-weekly 4 --keep-monthly 12 \
  --prune >> "$LOG" 2>&1

# Weekly integrity check (Sundays)
[ "$(date +%u)" -eq 7 ] && restic check --read-data-subset=10% >> "$LOG" 2>&1

echo "$(date): Backup completed" >> "$LOG"

chmod +x /usr/local/bin/restic-backup.sh

Scheduled Backups

Systemd Timer (Recommended)

# /etc/systemd/system/restic-backup.service
[Unit]
Description=Re

… (nội dung đầy đủ có trong gói tải về)