bin-there-done-that/cheatsheets/server_hardening_disaster_recovery.md

🛡️ Server Hardening & Disaster Recovery Cheat Sheet

🔐 Server Hardening Checklist

🔒 OS & User Security

• ✅ Use key-based SSH authentication (~/.ssh/authorized_keys)
• ✅ Disable root login:

  sudo sed -i 's/^PermitRootLogin.*/PermitRootLogin no/' /etc/ssh/sshd_config
  sudo systemctl restart sshd
  

• ✅ Change default SSH port and rate-limit with Fail2Ban or UFW
• ✅ Set strong password policies:

  sudo apt install libpam-pwquality
  sudo nano /etc/security/pwquality.conf
  

• ✅ Lock down /etc/sudoers, remove unnecessary sudo privileges

🔧 Kernel & System Hardening

• ✅ Install and configure ufw or iptables:

  sudo ufw default deny incoming
  sudo ufw allow ssh
  sudo ufw enable
  

• ✅ Disable unused filesystems:

  echo "install cramfs /bin/true" >> /etc/modprobe.d/disable-filesystems.conf
  

• ✅ Set kernel parameters:

  sudo nano /etc/sysctl.d/99-sysctl.conf
  # Example: net.ipv4.ip_forward = 0
  sudo sysctl -p
  

🧾 Logging & Monitoring

• ✅ Enable and configure auditd:

  sudo apt install auditd audispd-plugins
  sudo systemctl enable auditd
  

• ✅ Centralize logs using rsyslog, logrotate, or Fluentbit
• ✅ Use fail2ban, CrowdSec, or Wazuh for intrusion detection

💾 Disaster Recovery Checklist

📦 Backups

• ✅ Automate daily database dumps (e.g., pg_dump, mysqldump)
• ✅ Use ZFS snapshots for versioned backups
• ✅ Sync offsite via rclone, rsync, or cloud storage
• ✅ Encrypt backups using gpg or age

🔁 Testing & Recovery

• ✅ Verify backup integrity regularly:

  gpg --verify backup.sql.gpg
  pg_restore --list backup.dump
  

• ✅ Practice bare-metal restores in a test environment
• ✅ Use PITR (Point-In-Time Recovery) for PostgreSQL

🛑 Emergency Scripts

• ✅ Create service restart scripts:

  systemctl restart mastodon
  docker restart azuracast
  

• ✅ Pre-stage rescue.sh to rebuild key systems
• ✅ Include Mastodon/Gitea/etc. reconfig tools

🗂️ Documentation

• ✅ Maintain a runbook with:
  • Service recovery steps
  • IPs, ports, login methods
  • Admin contacts and escalation

🧪 Chaos Testing

• ✅ Simulate failure of:
  • A disk or volume (use zpool offline)
  • A network link (iptables -A OUTPUT ...)
  • A database node (use Patroni/pg_auto_failover tools)

---

✅ Pro Tip: Integrate all hardening and backup tasks into your Ansible playbooks for consistency and redeployability.