Ask HN: Tips to get started on my own server

Hacker News Discussion ↗

Aspiring self-hosters are encouraged to learn server administration by starting small: use a cheap VPS or a local machine/VM, pick a mainstream Linux distro like Ubuntu or Debian, and bring up one service at a time (often a simple web app behind nginx). Contributors stress basic hygiene—SSH keys, firewalls, fail2ban, monitoring, snapshots/backups—and meticulous notes so setups are reproducible and recoverable. Over time, many suggest moving toward automation and declarative setup with tools like Ansible, Terraform, or NixOS, and optionally experimenting with home hardware, containers, and more advanced services such as mail or databases.

Getting started environment

  • Many suggest starting cheap and disposable: a $5–10/month VPS or local VM, expect to break/reinstall multiple times.
  • Others recommend a home box (old laptop, mini PC, Raspberry Pi, NAS) for physical access and deep networking practice.
  • Some propose first mirroring services you already run on Windows in a Linux VM to reduce unknowns.
  • There’s a minority view that “real” bare-metal/colocation is more educational but higher effort.

Where to host: VPS vs home server

  • Popular VPS options mentioned: DigitalOcean, Vultr, Hetzner (praised as good value), Linode, OVH, BuyVM, Scaleway, lowendbox offers.
  • Cloud: AWS EC2/Lightsail and Azure burstable instances are suggested; some warn about IO and billing surprises.
  • Oracle Cloud “always free” gets both praise (great free box) and strong criticism (random bans, data/2FA concerns, idle-instance archiving).
  • Home hosting seen as cheap and empowering; critics warn about security, ISP limits, and uptime; proponents say it’s the best way to really learn.
  • Dynamic DNS and Cloudflare tunnels are suggested for exposing home services.

OS, tools, and automation

  • Beginner-friendly: Ubuntu or Debian server are most recommended due to documentation and community; Debian seen as “uncontroversial.”
  • Others suggest Arch, Gentoo, Slackware, NixOS, Alpine, FreeBSD, or Illumos for deeper understanding, but as second steps.
  • Use local hypervisors (VirtualBox, Orbstack, UTM) to spin up multiple test VMs.
  • Heavy emphasis on automation and reproducibility: shell scripts, Ansible, Terraform, cloud-init, Proxmox containers; “executable notes.”
  • Docker and higher-level platforms (CapRover, Yunohost) are proposed to quickly run apps, then inspect under the hood.

Security and reliability

  • Core practices: firewall (UFW/iptables or provider firewall), SSH keys only, disable root login, possibly move SSH off port 22, use strong passwords elsewhere.
  • Tools like fail2ban or CrowdSec to block brute-force attacks; VPNs (e.g., Tailscale) to reduce open ports.
  • Use snapshots/backups (Proxmox, provider snapshots, rsync, external disks) and learn to restore often.
  • Stick to LTS distros; consider CIS benchmarks and basic monitoring/perf tools as you advance.

Projects and learning approach

  • Start with one service at a time: simple website (nginx/Apache, Let’s Encrypt), then app stack (e.g., Django + gunicorn + nginx), then databases, DNS, mail, backup, monitoring.
  • Keep meticulous notes (or Obsidian/Zettelkasten-style docs); use shell history when you forget.
  • Philosophy: don’t chase the “perfect” stack; build something small, run it for months, fix issues permanently, then iterate.