AKGuidebooks Home Server Admin

When things go wrong

Monthly health check#

Fifteen minutes, once a month. Connect with ssh homeserver and run each line.

# Check Command (on the server) Good result
1 Anything broken? systemctl --failed 0 loaded units listed
2 Disk space df -h / /srv/files /srv/backup Use% below 85%
3 Memory free -h available not near 0
4 Temperature sensors below ~70 °C at rest
5 Updates working? sudo tail -5 /var/log/unattended-upgrades/unattended-upgrades.log recent dates, no ERROR
6 Restart needed? ls /var/run/reboot-required "No such file". If it exists, sudo reboot
7 Backups recent? sudo restic -r /srv/backup/restic --password-file /root/restic-password snapshots last line is from last night
8 Remote access tailscale status all your devices listed
9 Firewall sudo ufw status Status: active
10 AI model curl -s http://127.0.0.1:8095/health {"status":"ok"}
11 AI kept offline sudo nft list table inet aiguard table is listed
12 AI model files intact sudo sh -c 'cd /srv/ai/models && sha256sum -c SHA256SUMS' every line OK
13 Fail2ban sudo fail2ban-client status sshd jail running
14 Dashboard private sudo ss -ltn 'sport = :9090' 127.0.0.1:9090 only
15 Clock timedatectl | grep synchronized yes
16 Who can log in cat ~/.ssh/authorized_keys only devices you still own (check the names at the end of each line)

Every 3 months: do a test restore of one file (Chapter 17, step 6).

Every month (laptops): look at the case for battery swelling (Chapter 13).

Made by Anish · 2026