Overview

• What you’ll learn: The theory behind backup strategies (full, incremental, differential), Bacula’s multi-component architecture (Director, Storage Daemon, File Daemon), how to define backup jobs in bacula-dir.conf, how to configure rsnapshot for hourly/daily/weekly snapshot rotation, and how to perform reliable restores.
• Prerequisites: System Monitoring Fundamentals (Lesson 61), Log Management (Lesson 63), familiarity with cron scheduling
• Estimated reading time: 20 minutes

Introduction

Data loss is one of the most devastating events in any computing environment. Whether caused by hardware failure, human error, ransomware, or natural disaster, the loss of critical data can halt business operations and in some cases destroy organizations entirely. Backup is the insurance policy against data loss, and like all insurance, it must be planned, tested, and maintained to be effective.

A robust backup strategy goes far beyond simply copying files. It involves understanding recovery point objectives (RPO) — how much data you can afford to lose — and recovery time objectives (RTO) — how quickly you must restore service. These business requirements drive the choice of backup type, frequency, retention policy, and storage medium.

In this lesson, we will explore fundamental backup concepts, then implement two complementary tools on Ubuntu Server: Bacula, an enterprise-grade network backup suite with a client-server architecture, and rsnapshot, a lightweight filesystem snapshot utility built on rsync and hard links. Together, they provide a defense-in-depth approach to data protection.

Backup Strategy Fundamentals

Before choosing tools, you must understand the three primary backup types and their trade-offs:

• Full backup: Copies every file in the backup set. It provides the simplest restore (single backup needed) but consumes the most storage and time. Typically performed weekly or monthly.
• Incremental backup: Copies only files that changed since the last backup of any type. It is fast and storage-efficient, but restoration requires the last full backup plus every subsequent incremental in sequence.
• Differential backup: Copies all files changed since the last full backup. It strikes a balance — larger than incrementals but simpler to restore (only the full plus the latest differential).

The classic enterprise pattern is the Grandfather-Father-Son (GFS) rotation scheme: daily incrementals (sons), weekly differentials or fulls (fathers), and monthly fulls retained long-term (grandfathers). This balances storage cost with restore speed and data retention compliance.

The 3-2-1 rule is a widely accepted best practice: maintain at least 3 copies of your data, on 2 different storage media, with 1 copy stored off-site. This protects against localized failures such as fire, theft, or site-wide outages.

Bacula Architecture

Bacula is a network backup solution composed of several cooperating daemons. Understanding this architecture is critical before writing configuration.

• Bacula Director (bacula-dir): The central control daemon. It schedules and supervises all backup, restore, and verify jobs. Configuration lives in /etc/bacula/bacula-dir.conf.
• Bacula Storage Daemon (bacula-sd): Manages the physical storage media (disk volumes, tape drives). It reads and writes backup data on behalf of the Director. Configuration: /etc/bacula/bacula-sd.conf.
• Bacula File Daemon (bacula-fd): The client agent installed on each machine to be backed up. It reads files from the local filesystem and sends them to the Storage Daemon. Configuration: /etc/bacula/bacula-fd.conf.
• Bacula Catalog: A SQL database (MySQL/PostgreSQL) that stores metadata about all backup jobs, volumes, and file locations. It enables fast file-level restore lookups without scanning media.
• Bacula Console (bconsole): A command-line interface for interacting with the Director — starting jobs, checking status, and performing restores.

# Install Bacula components on Ubuntu
$ sudo apt update
$ sudo apt install -y bacula-director bacula-storage bacula-client 
    bacula-console bacula-common-mysql

# During installation, configure the MySQL/MariaDB catalog database
# Bacula will prompt for database credentials

# Verify all daemons are running
$ sudo systemctl status bacula-director
$ sudo systemctl status bacula-sd
$ sudo systemctl status bacula-fd

Configuring Bacula Director

The Director configuration defines the what, when, and where of backups. Key resource types include Job, FileSet, Schedule, Client, Storage, and Pool.

# Edit the Director configuration
$ sudo nano /etc/bacula/bacula-dir.conf

# Define a backup job
Job {
  Name = "BackupWebServer"
  Type = Backup
  Level = Incremental
  Client = webserver01-fd
  FileSet = "WebServerFiles"
  Schedule = "WeeklyCycle"
  Storage = File1
  Pool = Default
  Messages = Standard
  Priority = 10
  Write Bootstrap = "/var/lib/bacula/%c.bsr"
}

# Define what files to back up
FileSet {
  Name = "WebServerFiles"
  Include {
    Options {
      signature = MD5
      compression = GZIP
    }
    File = /etc
    File = /var/www
    File = /home
    File = /var/lib/mysql
  }
  Exclude {
    File = /var/www/cache
    File = /tmp
    File = /proc
    File = /sys
  }
}

# Define the backup schedule
Schedule {
  Name = "WeeklyCycle"
  Run = Full 1st sun at 23:05
  Run = Differential 2nd-5th sun at 23:05
  Run = Incremental mon-sat at 23:05
}

# Define the client (File Daemon)
Client {
  Name = webserver01-fd
  Address = 192.168.1.101
  FDPort = 9102
  Catalog = MyCatalog
  Password = "fd-secret-password"
  File Retention = 60 days
  Job Retention = 6 months
  AutoPrune = yes
}

# Define the storage target
Storage {
  Name = File1
  Address = 192.168.1.100
  SDPort = 9103
  Password = "sd-secret-password"
  Device = FileChgr1
  Media Type = File
  Maximum Concurrent Jobs = 5
}

# Validate the configuration
$ sudo bacula-dir -tc /etc/bacula/bacula-dir.conf
Configuration OK

# Reload the Director
$ sudo systemctl reload bacula-director

Configuring rsnapshot

While Bacula handles enterprise backup orchestration, rsnapshot provides a simpler, filesystem-level snapshot approach. Built on rsync and hard links, rsnapshot creates space-efficient point-in-time copies. Unchanged files are hard-linked between snapshots, meaning only modified files consume additional disk space.

# Install rsnapshot
$ sudo apt install -y rsnapshot

# Edit the configuration (note: rsnapshot requires TABS, not spaces)
$ sudo nano /etc/rsnapshot.conf

# Set the snapshot root directory
snapshot_root	/backup/rsnapshot/

# Use rsync with SSH for remote backups
cmd_ssh	/usr/bin/ssh

# Define retention intervals
retain	hourly	6
retain	daily	7
retain	weekly	4
retain	monthly	3

# Define backup sources
backup	/etc/	localhost/
backup	/home/	localhost/
backup	/var/www/	localhost/
backup	/var/log/	localhost/
backup	root@192.168.1.101:/etc/	webserver01/
backup	root@192.168.1.101:/var/www/	webserver01/

# Exclude patterns
exclude	*.tmp
exclude	/var/www/cache/
exclude	.git/

# Test the configuration
$ sudo rsnapshot configtest
Syntax OK

# Run a manual hourly snapshot
$ sudo rsnapshot hourly

# Verify snapshot structure
$ ls -la /backup/rsnapshot/
drwxr-xr-x  4 root root 4096 Mar  2 06:00 hourly.0
drwxr-xr-x  4 root root 4096 Mar  2 05:00 hourly.1
drwxr-xr-x  4 root root 4096 Mar  2 04:00 hourly.2

# Set up cron jobs for automated rotation
$ sudo tee /etc/cron.d/rsnapshot > /dev/null <<'EOF'
0 */4  * * *  root  /usr/bin/rsnapshot hourly
30 3   * * *  root  /usr/bin/rsnapshot daily
0  3   * * 1  root  /usr/bin/rsnapshot weekly
30 2   1 * *  root  /usr/bin/rsnapshot monthly
EOF

Restore Procedures

A backup is only as good as its restore capability. Regular restore testing is essential to validate that your backups are complete and functional.

# --- Bacula restore via bconsole ---
$ sudo bconsole
*restore
# Select option 5: "Select the most recent backup for a client"
# Choose the client and FileSet
# Mark files for restore:
mark /etc/nginx/nginx.conf
mark /var/www/html/
done
# Confirm restore location (default: /tmp/bacula-restores/)
yes

# Monitor restore progress
*status dir

# --- rsnapshot restore (simply copy files) ---
# Restore a single file from the latest hourly snapshot
$ sudo cp /backup/rsnapshot/hourly.0/localhost/etc/nginx/nginx.conf 
    /etc/nginx/nginx.conf

# Restore an entire directory from a daily snapshot
$ sudo rsync -avz /backup/rsnapshot/daily.2/localhost/var/www/ /var/www/

# Compare a file across snapshots to find when a change occurred
$ diff /backup/rsnapshot/hourly.0/localhost/etc/passwd 
       /backup/rsnapshot/hourly.3/localhost/etc/passwd

Key Takeaways

• The three backup types — full, incremental, and differential — serve different trade-offs between storage efficiency and restore complexity; the GFS rotation scheme combines them for enterprise use.
• Bacula’s architecture separates concerns across Director (scheduling), Storage Daemon (media), and File Daemon (client agents), enabling scalable network backup across many hosts.
• rsnapshot uses rsync and hard links to create space-efficient filesystem snapshots with configurable hourly, daily, weekly, and monthly retention.
• The 3-2-1 rule (3 copies, 2 media types, 1 off-site) is the minimum standard for data protection against localized disasters.
• Regular restore testing is non-negotiable — an untested backup is not a backup; validate both file integrity and restore procedures on a scheduled basis.

What’s Next

In the next lesson, we will explore DRBD Block Replication — learning how to replicate block devices across servers in real time for high-availability storage that complements backup with continuous data protection.

# 在 Ubuntu 上安裝 Bacula 元件
$ sudo apt install -y bacula-director bacula-storage bacula-client 
    bacula-console bacula-common-mysql

# 驗證所有守護程序正在運行
$ sudo systemctl status bacula-director
$ sudo systemctl status bacula-sd
$ sudo systemctl status bacula-fd

# 定義備份作業
Job {
  Name = "BackupWebServer"
  Type = Backup
  Level = Incremental
  Client = webserver01-fd
  FileSet = "WebServerFiles"
  Schedule = "WeeklyCycle"
  Storage = File1
  Pool = Default
}

# 定義備份排程
Schedule {
  Name = "WeeklyCycle"
  Run = Full 1st sun at 23:05
  Run = Differential 2nd-5th sun at 23:05
  Run = Incremental mon-sat at 23:05
}

# 驗證設定
$ sudo bacula-dir -tc /etc/bacula/bacula-dir.conf

# 安裝 rsnapshot
$ sudo apt install -y rsnapshot

# 編輯設定（注意：rsnapshot 需要 TAB，不是空格）
$ sudo nano /etc/rsnapshot.conf

snapshot_root	/backup/rsnapshot/
retain	hourly	6
retain	daily	7
retain	weekly	4
retain	monthly	3

backup	/etc/	localhost/
backup	/home/	localhost/
backup	/var/www/	localhost/

# 測試設定
$ sudo rsnapshot configtest

# 設定 cron 自動輪替
0 */4  * * *  root  /usr/bin/rsnapshot hourly
30 3   * * *  root  /usr/bin/rsnapshot daily
0  3   * * 1  root  /usr/bin/rsnapshot weekly
30 2   1 * *  root  /usr/bin/rsnapshot monthly

# Bacula 還原
$ sudo bconsole
*restore
# 選擇客戶端和 FileSet，標記要還原的檔案

# rsnapshot 還原（直接複製檔案）
$ sudo cp /backup/rsnapshot/hourly.0/localhost/etc/nginx/nginx.conf 
    /etc/nginx/nginx.conf

# 比較快照間的檔案差異
$ diff /backup/rsnapshot/hourly.0/localhost/etc/passwd 
       /backup/rsnapshot/hourly.3/localhost/etc/passwd

# Ubuntu に Bacula コンポーネントをインストール
$ sudo apt install -y bacula-director bacula-storage bacula-client 
    bacula-console bacula-common-mysql

# すべてのデーモンが実行中であることを確認
$ sudo systemctl status bacula-director
$ sudo systemctl status bacula-sd
$ sudo systemctl status bacula-fd

# バックアップジョブの定義
Job {
  Name = "BackupWebServer"
  Type = Backup
  Level = Incremental
  Client = webserver01-fd
  FileSet = "WebServerFiles"
  Schedule = "WeeklyCycle"
  Storage = File1
  Pool = Default
}

# バックアップスケジュールの定義
Schedule {
  Name = "WeeklyCycle"
  Run = Full 1st sun at 23:05
  Run = Differential 2nd-5th sun at 23:05
  Run = Incremental mon-sat at 23:05
}

# 設定を検証
$ sudo bacula-dir -tc /etc/bacula/bacula-dir.conf

# rsnapshot をインストール
$ sudo apt install -y rsnapshot

# 設定を編集（注意：rsnapshot はタブが必要、スペースではない）
$ sudo nano /etc/rsnapshot.conf

snapshot_root	/backup/rsnapshot/
retain	hourly	6
retain	daily	7
retain	weekly	4
retain	monthly	3

backup	/etc/	localhost/
backup	/home/	localhost/
backup	/var/www/	localhost/

# 設定をテスト
$ sudo rsnapshot configtest

# 自動ローテーションのための cron 設定
0 */4  * * *  root  /usr/bin/rsnapshot hourly
30 3   * * *  root  /usr/bin/rsnapshot daily
0  3   * * 1  root  /usr/bin/rsnapshot weekly
30 2   1 * *  root  /usr/bin/rsnapshot monthly

# Bacula リストア
$ sudo bconsole
*restore
# クライアントと FileSet を選択し、リストアするファイルをマーク

# rsnapshot リストア（単純にファイルをコピー）
$ sudo cp /backup/rsnapshot/hourly.0/localhost/etc/nginx/nginx.conf 
    /etc/nginx/nginx.conf

# スナップショット間のファイル差分を比較
$ diff /backup/rsnapshot/hourly.0/localhost/etc/passwd 
       /backup/rsnapshot/hourly.3/localhost/etc/passwd